From b6ba4d489ab1e623aecf6c499e5af0321719b8f8 Mon Sep 17 00:00:00 2001 From: Jens Steube Date: Mon, 18 Apr 2016 15:16:23 +0200 Subject: [PATCH] Cleanup -m 114xx kernels to latest standard --- OpenCL/m11400_a0.cl | 93 ++------- OpenCL/m11400_a1.cl | 447 ++++++++++++++++++-------------------------- OpenCL/m11400_a3.cl | 283 ++++------------------------ 3 files changed, 232 insertions(+), 591 deletions(-) diff --git a/OpenCL/m11400_a0.cl b/OpenCL/m11400_a0.cl index 4b428f8b2..2ad222c16 100644 --- a/OpenCL/m11400_a0.cl +++ b/OpenCL/m11400_a0.cl @@ -7,7 +7,8 @@ #define _MD5_ -#define NEW_SIMD_CODE +//incompatible because of brances +//#define NEW_SIMD_CODE #include "include/constants.h" #include "include/kernel_vendor.h" @@ -791,14 +792,12 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, */ 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]; @@ -813,6 +812,7 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 salt_len = esalt_bufs[salt_pos].salt_len; // not a bug, we need to get it from the esalt u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -830,9 +830,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -857,6 +854,8 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -874,9 +873,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -893,9 +889,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -940,6 +933,7 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -957,9 +951,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -979,31 +970,25 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, u32 block_len = 0; - block_len = memcat32_VV (block0, block1, salt_len, w0, w1, w2, w3, out_len); + block_len = memcat32 (block0, block1, salt_len, w0, w1, w2, w3, out_len); u32x w0_t[4]; + u32x w1_t[4]; + u32x w2_t[4]; + u32x w3_t[4]; w0_t[0] = block0[ 0]; w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - - u32x w1_t[4]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - - u32x w2_t[4]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - - u32x w3_t[4]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -1105,17 +1090,14 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -1216,12 +1198,10 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -1319,17 +1299,14 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -1432,17 +1409,14 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -1568,14 +1542,12 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, */ 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]; @@ -1590,6 +1562,7 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 salt_len = esalt_bufs[salt_pos].salt_len; // not a bug, we need to get it from the esalt u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -1607,9 +1580,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -1634,6 +1604,8 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -1651,9 +1623,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -1670,9 +1639,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -1729,6 +1695,7 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -1746,9 +1713,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -1768,31 +1732,25 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, u32 block_len = 0; - block_len = memcat32_VV (block0, block1, salt_len, w0, w1, w2, w3, out_len); + block_len = memcat32 (block0, block1, salt_len, w0, w1, w2, w3, out_len); u32x w0_t[4]; + u32x w1_t[4]; + u32x w2_t[4]; + u32x w3_t[4]; w0_t[0] = block0[ 0]; w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - - u32x w1_t[4]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - - u32x w2_t[4]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - - u32x w3_t[4]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -1894,17 +1852,14 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -2005,12 +1960,10 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -2108,17 +2061,14 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -2221,17 +2171,14 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; diff --git a/OpenCL/m11400_a1.cl b/OpenCL/m11400_a1.cl index b0646f660..b2f632d28 100644 --- a/OpenCL/m11400_a1.cl +++ b/OpenCL/m11400_a1.cl @@ -5,6 +5,9 @@ #define _MD5_ +//incompatible because of brances +//#define NEW_SIMD_CODE + #include "include/constants.h" #include "include/kernel_vendor.h" @@ -16,13 +19,21 @@ #include "include/kernel_functions.c" #include "OpenCL/types_ocl.c" #include "OpenCL/common.c" - -#define COMPARE_S "OpenCL/check_single_comp4.c" -#define COMPARE_M "OpenCL/check_multi_comp4.c" - -#define uint_to_hex_lower8(i) l_bin2asc[(i)] - -static u32 memcat32 (u32 block0[16], u32 block1[16], const u32 block_len, const u32 append0[4], const u32 append1[4], const u32 append2[4], const u32 append3[4], const u32 append_len) +#include "OpenCL/simd.c" + +#if VECT_SIZE == 1 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i)]) +#elif VECT_SIZE == 2 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1]) +#elif VECT_SIZE == 4 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3]) +#elif VECT_SIZE == 8 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7]) +#elif VECT_SIZE == 16 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7], l_bin2asc[(i).s8], l_bin2asc[(i).s9], l_bin2asc[(i).sa], l_bin2asc[(i).sb], l_bin2asc[(i).sc], l_bin2asc[(i).sd], l_bin2asc[(i).se], l_bin2asc[(i).sf]) +#endif + +static u32 memcat32 (u32x block0[16], u32x block1[16], const u32 block_len, const u32x append0[4], const u32x append1[4], const u32x append2[4], const u32x append3[4], const u32 append_len) { const u32 mod = block_len & 3; const u32 div = block_len / 4; @@ -30,35 +41,35 @@ static u32 memcat32 (u32 block0[16], u32 block1[16], const u32 block_len, const #if defined IS_AMD || defined IS_GENERIC const int offset_minus_4 = 4 - mod; - u32 append0_t[4]; + u32x append0_t[4]; append0_t[0] = amd_bytealign (append0[0], 0, offset_minus_4); append0_t[1] = amd_bytealign (append0[1], append0[0], offset_minus_4); append0_t[2] = amd_bytealign (append0[2], append0[1], offset_minus_4); append0_t[3] = amd_bytealign (append0[3], append0[2], offset_minus_4); - u32 append1_t[4]; + u32x append1_t[4]; append1_t[0] = amd_bytealign (append1[0], append0[3], offset_minus_4); append1_t[1] = amd_bytealign (append1[1], append1[0], offset_minus_4); append1_t[2] = amd_bytealign (append1[2], append1[1], offset_minus_4); append1_t[3] = amd_bytealign (append1[3], append1[2], offset_minus_4); - u32 append2_t[4]; + u32x append2_t[4]; append2_t[0] = amd_bytealign (append2[0], append1[3], offset_minus_4); append2_t[1] = amd_bytealign (append2[1], append2[0], offset_minus_4); append2_t[2] = amd_bytealign (append2[2], append2[1], offset_minus_4); append2_t[3] = amd_bytealign (append2[3], append2[2], offset_minus_4); - u32 append3_t[4]; + u32x append3_t[4]; append3_t[0] = amd_bytealign (append3[0], append2[3], offset_minus_4); append3_t[1] = amd_bytealign (append3[1], append3[0], offset_minus_4); append3_t[2] = amd_bytealign (append3[2], append3[1], offset_minus_4); append3_t[3] = amd_bytealign (append3[3], append3[2], offset_minus_4); - u32 append4_t[4]; + u32x append4_t[4]; append4_t[0] = amd_bytealign ( 0, append3[3], offset_minus_4); append4_t[1] = 0; @@ -100,35 +111,35 @@ static u32 memcat32 (u32 block0[16], u32 block1[16], const u32 block_len, const const int selector = (0x76543210 >> (offset_minus_4 * 4)) & 0xffff; - u32 append0_t[4]; + u32x append0_t[4]; append0_t[0] = __byte_perm ( 0, append0[0], selector); append0_t[1] = __byte_perm (append0[0], append0[1], selector); append0_t[2] = __byte_perm (append0[1], append0[2], selector); append0_t[3] = __byte_perm (append0[2], append0[3], selector); - u32 append1_t[4]; + u32x append1_t[4]; append1_t[0] = __byte_perm (append0[3], append1[0], selector); append1_t[1] = __byte_perm (append1[0], append1[1], selector); append1_t[2] = __byte_perm (append1[1], append1[2], selector); append1_t[3] = __byte_perm (append1[2], append1[3], selector); - u32 append2_t[4]; + u32x append2_t[4]; append2_t[0] = __byte_perm (append1[3], append2[0], selector); append2_t[1] = __byte_perm (append2[0], append2[1], selector); append2_t[2] = __byte_perm (append2[1], append2[2], selector); append2_t[3] = __byte_perm (append2[2], append2[3], selector); - u32 append3_t[4]; + u32x append3_t[4]; append3_t[0] = __byte_perm (append2[3], append3[0], selector); append3_t[1] = __byte_perm (append3[0], append3[1], selector); append3_t[2] = __byte_perm (append3[1], append3[2], selector); append3_t[3] = __byte_perm (append3[2], append3[3], selector); - u32 append4_t[4]; + u32x append4_t[4]; append4_t[0] = __byte_perm (append3[3], 0, selector); append4_t[1] = 0; @@ -776,43 +787,20 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * base */ - u32 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]; - - u32 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]; - - u32 wordl2[4]; + u32 pw_buf0[4]; + u32 pw_buf1[4]; - wordl2[0] = 0; - wordl2[1] = 0; - wordl2[2] = 0; - wordl2[3] = 0; - - u32 wordl3[4]; - - wordl3[0] = 0; - wordl3[1] = 0; - wordl3[2] = 0; - wordl3[3] = 0; + 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]; + pw_buf1[0] = pws[gid].i[4]; + pw_buf1[1] = pws[gid].i[5]; + pw_buf1[2] = pws[gid].i[6]; + pw_buf1[3] = pws[gid].i[7]; const u32 pw_l_len = pws[gid].pw_len; - if (combs_mode == COMBINATOR_MODE_BASE_RIGHT) - { - append_0x80_2x4 (wordl0, wordl1, pw_l_len); - - switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len); - } - /** * salt */ @@ -820,6 +808,7 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 salt_len = esalt_bufs[salt_pos].salt_len; // not a bug, we need to get it from the esalt u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -837,9 +826,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -864,6 +850,8 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -881,9 +869,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -900,9 +885,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -927,74 +909,76 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * loop */ - for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { - const u32 pw_r_len = combs_buf[il_pos].pw_len; - - const u32 pw_len = pw_l_len + pw_r_len; - - u32 wordr0[4]; - - wordr0[0] = combs_buf[il_pos].i[0]; - wordr0[1] = combs_buf[il_pos].i[1]; - wordr0[2] = combs_buf[il_pos].i[2]; - wordr0[3] = combs_buf[il_pos].i[3]; - - u32 wordr1[4]; + const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos); - wordr1[0] = combs_buf[il_pos].i[4]; - wordr1[1] = combs_buf[il_pos].i[5]; - wordr1[2] = combs_buf[il_pos].i[6]; - wordr1[3] = combs_buf[il_pos].i[7]; + const u32x pw_len = pw_l_len + pw_r_len; - u32 wordr2[4]; - - wordr2[0] = 0; - wordr2[1] = 0; - wordr2[2] = 0; - wordr2[3] = 0; - - u32 wordr3[4]; + /** + * concat password candidate + */ - wordr3[0] = 0; - wordr3[1] = 0; - wordr3[2] = 0; - wordr3[3] = 0; + u32x wordl0[4] = { 0 }; + u32x wordl1[4] = { 0 }; + u32x wordl2[4] = { 0 }; + u32x wordl3[4] = { 0 }; + + wordl0[0] = pw_buf0[0]; + wordl0[1] = pw_buf0[1]; + wordl0[2] = pw_buf0[2]; + wordl0[3] = pw_buf0[3]; + wordl1[0] = pw_buf1[0]; + wordl1[1] = pw_buf1[1]; + wordl1[2] = pw_buf1[2]; + wordl1[3] = pw_buf1[3]; + + u32x wordr0[4] = { 0 }; + u32x wordr1[4] = { 0 }; + u32x wordr2[4] = { 0 }; + u32x wordr3[4] = { 0 }; + + wordr0[0] = ix_create_combt (combs_buf, il_pos, 0); + wordr0[1] = ix_create_combt (combs_buf, il_pos, 1); + wordr0[2] = ix_create_combt (combs_buf, il_pos, 2); + wordr0[3] = ix_create_combt (combs_buf, il_pos, 3); + wordr1[0] = ix_create_combt (combs_buf, il_pos, 4); + wordr1[1] = ix_create_combt (combs_buf, il_pos, 5); + wordr1[2] = ix_create_combt (combs_buf, il_pos, 6); + wordr1[3] = ix_create_combt (combs_buf, il_pos, 7); if (combs_mode == COMBINATOR_MODE_BASE_LEFT) { - switch_buffer_by_offset_le (wordr0, wordr1, wordr2, wordr3, pw_l_len); + switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len); + } + else + { + switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len); } - u32 w0[4]; + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[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]; - - u32 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]; - - u32 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]; - - u32 w3[4]; - w3[0] = wordl3[0] | wordr3[0]; w3[1] = wordl3[1] | wordr3[1]; w3[2] = wordl3[2] | wordr3[2]; w3[3] = wordl3[3] | wordr3[3]; - const u32 pw_salt_len = salt_len + pw_len; + const u32x pw_salt_len = salt_len + pw_len; /* * HA1 = md5 ($salt . $pass) @@ -1002,7 +986,8 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // append the pass to the salt - u32 block0[16]; + u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -1020,9 +1005,6 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32 block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -1044,29 +1026,23 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, block_len = memcat32 (block0, block1, salt_len, w0, w1, w2, w3, pw_len); - u32 w0_t[4]; + u32x w0_t[4]; + u32x w1_t[4]; + u32x w2_t[4]; + u32x w3_t[4]; w0_t[0] = block0[ 0]; w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - - u32 w1_t[4]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - - u32 w2_t[4]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - - u32 w3_t[4]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -1079,10 +1055,10 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // md5 - u32 a = MD5M_A; - u32 b = MD5M_B; - u32 c = MD5M_C; - u32 d = MD5M_D; + u32x a = MD5M_A; + u32x b = MD5M_B; + u32x c = MD5M_C; + u32x d = MD5M_D; MD5_STEP (MD5_Fo, a, b, c, d, w0_t[0], MD5C00, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w0_t[1], MD5C01, MD5S01); @@ -1159,26 +1135,23 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, if (block_len > 55) { - u32 r_a = a; - u32 r_b = b; - u32 r_c = c; - u32 r_d = d; + u32x r_a = a; + u32x r_b = b; + u32x r_c = c; + u32x r_d = d; w0_t[0] = block1[ 0]; w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -1279,12 +1252,10 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -1371,10 +1342,10 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, c += MD5M_C; d += MD5M_D; - u32 r_a = a; - u32 r_b = b; - u32 r_c = c; - u32 r_d = d; + u32x r_a = a; + u32x r_b = b; + u32x r_c = c; + u32x r_d = d; // 2nd transform @@ -1382,17 +1353,14 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -1495,17 +1463,14 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -1585,12 +1550,7 @@ __kernel void m11400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, c += r_c; d += r_d; - const u32 r0 = a; - const u32 r1 = d; - const u32 r2 = c; - const u32 r3 = b; - - #include COMPARE_M + COMPARE_M_SIMD (a, d, c, b); } } @@ -1635,43 +1595,20 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * base */ - u32 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]; - - u32 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]; - - u32 wordl2[4]; - - wordl2[0] = 0; - wordl2[1] = 0; - wordl2[2] = 0; - wordl2[3] = 0; - - u32 wordl3[4]; + u32 pw_buf0[4]; + u32 pw_buf1[4]; - wordl3[0] = 0; - wordl3[1] = 0; - wordl3[2] = 0; - wordl3[3] = 0; + 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]; + pw_buf1[0] = pws[gid].i[4]; + pw_buf1[1] = pws[gid].i[5]; + pw_buf1[2] = pws[gid].i[6]; + pw_buf1[3] = pws[gid].i[7]; const u32 pw_l_len = pws[gid].pw_len; - if (combs_mode == COMBINATOR_MODE_BASE_RIGHT) - { - append_0x80_2x4 (wordl0, wordl1, pw_l_len); - - switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len); - } - /** * salt */ @@ -1679,6 +1616,7 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 salt_len = esalt_bufs[salt_pos].salt_len; // not a bug, we need to get it from the esalt u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -1696,9 +1634,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -1723,6 +1658,8 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -1740,9 +1677,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -1759,9 +1693,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -1798,74 +1729,76 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * loop */ - for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { - const u32 pw_r_len = combs_buf[il_pos].pw_len; - - const u32 pw_len = pw_l_len + pw_r_len; - - u32 wordr0[4]; - - wordr0[0] = combs_buf[il_pos].i[0]; - wordr0[1] = combs_buf[il_pos].i[1]; - wordr0[2] = combs_buf[il_pos].i[2]; - wordr0[3] = combs_buf[il_pos].i[3]; - - u32 wordr1[4]; - - wordr1[0] = combs_buf[il_pos].i[4]; - wordr1[1] = combs_buf[il_pos].i[5]; - wordr1[2] = combs_buf[il_pos].i[6]; - wordr1[3] = combs_buf[il_pos].i[7]; - - u32 wordr2[4]; + const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos); - wordr2[0] = 0; - wordr2[1] = 0; - wordr2[2] = 0; - wordr2[3] = 0; + const u32x pw_len = pw_l_len + pw_r_len; - u32 wordr3[4]; + /** + * concat password candidate + */ - wordr3[0] = 0; - wordr3[1] = 0; - wordr3[2] = 0; - wordr3[3] = 0; + u32x wordl0[4] = { 0 }; + u32x wordl1[4] = { 0 }; + u32x wordl2[4] = { 0 }; + u32x wordl3[4] = { 0 }; + + wordl0[0] = pw_buf0[0]; + wordl0[1] = pw_buf0[1]; + wordl0[2] = pw_buf0[2]; + wordl0[3] = pw_buf0[3]; + wordl1[0] = pw_buf1[0]; + wordl1[1] = pw_buf1[1]; + wordl1[2] = pw_buf1[2]; + wordl1[3] = pw_buf1[3]; + + u32x wordr0[4] = { 0 }; + u32x wordr1[4] = { 0 }; + u32x wordr2[4] = { 0 }; + u32x wordr3[4] = { 0 }; + + wordr0[0] = ix_create_combt (combs_buf, il_pos, 0); + wordr0[1] = ix_create_combt (combs_buf, il_pos, 1); + wordr0[2] = ix_create_combt (combs_buf, il_pos, 2); + wordr0[3] = ix_create_combt (combs_buf, il_pos, 3); + wordr1[0] = ix_create_combt (combs_buf, il_pos, 4); + wordr1[1] = ix_create_combt (combs_buf, il_pos, 5); + wordr1[2] = ix_create_combt (combs_buf, il_pos, 6); + wordr1[3] = ix_create_combt (combs_buf, il_pos, 7); if (combs_mode == COMBINATOR_MODE_BASE_LEFT) { - switch_buffer_by_offset_le (wordr0, wordr1, wordr2, wordr3, pw_l_len); + switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len); + } + else + { + switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len); } - u32 w0[4]; + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[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]; - - u32 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]; - - u32 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]; - - u32 w3[4]; - w3[0] = wordl3[0] | wordr3[0]; w3[1] = wordl3[1] | wordr3[1]; w3[2] = wordl3[2] | wordr3[2]; w3[3] = wordl3[3] | wordr3[3]; - const u32 pw_salt_len = salt_len + pw_len; + const u32x pw_salt_len = salt_len + pw_len; /* * HA1 = md5 ($salt . $pass) @@ -1873,7 +1806,8 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // append the pass to the salt - u32 block0[16]; + u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -1891,9 +1825,6 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32 block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -1915,29 +1846,23 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, block_len = memcat32 (block0, block1, salt_len, w0, w1, w2, w3, pw_len); - u32 w0_t[4]; + u32x w0_t[4]; + u32x w1_t[4]; + u32x w2_t[4]; + u32x w3_t[4]; w0_t[0] = block0[ 0]; w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - - u32 w1_t[4]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - - u32 w2_t[4]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - - u32 w3_t[4]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -1950,10 +1875,10 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // md5 - u32 a = MD5M_A; - u32 b = MD5M_B; - u32 c = MD5M_C; - u32 d = MD5M_D; + u32x a = MD5M_A; + u32x b = MD5M_B; + u32x c = MD5M_C; + u32x d = MD5M_D; MD5_STEP (MD5_Fo, a, b, c, d, w0_t[0], MD5C00, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w0_t[1], MD5C01, MD5S01); @@ -2030,26 +1955,23 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, if (block_len > 55) { - u32 r_a = a; - u32 r_b = b; - u32 r_c = c; - u32 r_d = d; + u32x r_a = a; + u32x r_b = b; + u32x r_c = c; + u32x r_d = d; w0_t[0] = block1[ 0]; w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -2150,12 +2072,10 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -2242,10 +2162,10 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, c += MD5M_C; d += MD5M_D; - u32 r_a = a; - u32 r_b = b; - u32 r_c = c; - u32 r_d = d; + u32x r_a = a; + u32x r_b = b; + u32x r_c = c; + u32x r_d = d; // 2nd transform @@ -2253,17 +2173,14 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -2366,17 +2283,14 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -2456,12 +2370,7 @@ __kernel void m11400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, c += r_c; d += r_d; - const u32 r0 = a; - const u32 r1 = d; - const u32 r2 = c; - const u32 r3 = b; - - #include COMPARE_S + COMPARE_S_SIMD (a, d, c, b); } } diff --git a/OpenCL/m11400_a3.cl b/OpenCL/m11400_a3.cl index 03664442b..ac985f65d 100644 --- a/OpenCL/m11400_a3.cl +++ b/OpenCL/m11400_a3.cl @@ -8,7 +8,8 @@ #define _MD5_ -#define NEW_SIMD_CODE +//incompatible because of brances +//#define NEW_SIMD_CODE #include "include/constants.h" #include "include/kernel_vendor.h" @@ -765,18 +766,6 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); - /** - * 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 */ @@ -786,6 +775,7 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -803,9 +793,6 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -830,6 +817,7 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -847,9 +835,6 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -888,6 +873,7 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -905,9 +891,6 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -934,17 +917,14 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -956,17 +936,14 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = pw_salt_len * 8; @@ -1073,12 +1050,10 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -1176,17 +1151,14 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -1265,7 +1237,7 @@ static void m11400m_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p c += r_c; d += r_d; - COMPARE_S_SIMD (a, d, c, b); + COMPARE_M_SIMD (a, d, c, b); } } @@ -1278,18 +1250,6 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); - /** - * 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 */ @@ -1299,6 +1259,7 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -1316,9 +1277,6 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -1343,6 +1301,8 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -1360,9 +1320,6 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -1379,9 +1336,6 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -1420,6 +1374,7 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -1437,9 +1392,6 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -1466,17 +1418,14 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -1488,17 +1437,14 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = pw_salt_len * 8; @@ -1605,12 +1551,10 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -1708,17 +1652,14 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -1808,17 +1749,14 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -1897,7 +1835,7 @@ static void m11400m_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p c += r_c; d += r_d; - COMPARE_S_SIMD (a, d, c, b); + COMPARE_M_SIMD (a, d, c, b); } } @@ -1910,18 +1848,6 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); - /** - * 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 */ @@ -1931,6 +1857,7 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -1948,9 +1875,6 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -1975,6 +1899,7 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -1992,9 +1917,6 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -2033,6 +1955,7 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -2050,9 +1973,6 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -2079,17 +1999,14 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -2101,17 +2018,14 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -2206,17 +2120,14 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -2316,12 +2227,10 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -2419,17 +2328,14 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -2508,7 +2414,7 @@ static void m11400m_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p c += r_c; d += r_d; - COMPARE_S_SIMD (a, d, c, b); + COMPARE_M_SIMD (a, d, c, b); } } @@ -2521,18 +2427,6 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); - /** - * 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 */ @@ -2542,6 +2436,7 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -2559,9 +2454,6 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -2586,6 +2478,8 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -2603,9 +2497,6 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -2622,9 +2513,6 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -2663,6 +2551,7 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -2680,9 +2569,6 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32 block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -2709,17 +2595,14 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -2731,17 +2614,14 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -2836,17 +2716,14 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -2946,12 +2823,10 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -3049,17 +2924,14 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -3149,17 +3021,14 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -3238,7 +3107,7 @@ static void m11400m_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p c += r_c; d += r_d; - COMPARE_S_SIMD (a, d, c, b); + COMPARE_M_SIMD (a, d, c, b); } } @@ -3272,6 +3141,7 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -3289,9 +3159,6 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -3316,6 +3183,7 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -3333,9 +3201,6 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -3374,6 +3239,7 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -3391,9 +3257,6 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -3420,17 +3283,14 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -3442,17 +3302,14 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = pw_salt_len * 8; @@ -3559,12 +3416,10 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -3662,17 +3517,14 @@ static void m11400s_0_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -3785,6 +3637,7 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -3802,9 +3655,6 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -3829,6 +3679,8 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -3846,9 +3698,6 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -3865,9 +3714,6 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -3906,6 +3752,7 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -3923,9 +3770,6 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -3952,17 +3796,14 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -3974,17 +3815,14 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = pw_salt_len * 8; @@ -4091,12 +3929,10 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -4194,17 +4030,14 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -4294,17 +4127,14 @@ static void m11400s_0_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -4417,6 +4247,7 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -4434,9 +4265,6 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -4461,6 +4289,7 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -4478,9 +4307,6 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -4519,6 +4345,7 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -4536,9 +4363,6 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -4565,17 +4389,14 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -4587,17 +4408,14 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -4692,17 +4510,14 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -4802,12 +4617,10 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -4905,17 +4718,14 @@ static void m11400s_1_0 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = digest_esalt_len * 8; @@ -5028,6 +4838,7 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 pw_salt_len = salt_len + pw_len; u32 salt_buf0[16]; + u32 salt_buf1[16]; salt_buf0[ 0] = esalt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[ 1] = esalt_bufs[salt_pos].salt_buf[ 1]; @@ -5045,9 +4856,6 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p salt_buf0[13] = esalt_bufs[salt_pos].salt_buf[13]; salt_buf0[14] = esalt_bufs[salt_pos].salt_buf[14]; salt_buf0[15] = esalt_bufs[salt_pos].salt_buf[15]; - - u32 salt_buf1[16]; - salt_buf1[ 0] = esalt_bufs[salt_pos].salt_buf[16]; salt_buf1[ 1] = esalt_bufs[salt_pos].salt_buf[17]; salt_buf1[ 2] = esalt_bufs[salt_pos].salt_buf[18]; @@ -5072,6 +4880,8 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p const u32 esalt_len = esalt_bufs[salt_pos].esalt_len; u32 esalt_buf0[16]; + u32 esalt_buf1[16]; + u32 esalt_buf2[16]; esalt_buf0[ 0] = esalt_bufs[salt_pos].esalt_buf[ 0]; esalt_buf0[ 1] = esalt_bufs[salt_pos].esalt_buf[ 1]; @@ -5089,9 +4899,6 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf0[13] = esalt_bufs[salt_pos].esalt_buf[13]; esalt_buf0[14] = esalt_bufs[salt_pos].esalt_buf[14]; esalt_buf0[15] = esalt_bufs[salt_pos].esalt_buf[15]; - - u32 esalt_buf1[16]; - esalt_buf1[ 0] = esalt_bufs[salt_pos].esalt_buf[16]; esalt_buf1[ 1] = esalt_bufs[salt_pos].esalt_buf[17]; esalt_buf1[ 2] = esalt_bufs[salt_pos].esalt_buf[18]; @@ -5108,9 +4915,6 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p esalt_buf1[13] = esalt_bufs[salt_pos].esalt_buf[29]; esalt_buf1[14] = esalt_bufs[salt_pos].esalt_buf[30]; esalt_buf1[15] = esalt_bufs[salt_pos].esalt_buf[31]; - - u32 esalt_buf2[16]; - esalt_buf2[ 0] = esalt_bufs[salt_pos].esalt_buf[32]; esalt_buf2[ 1] = esalt_bufs[salt_pos].esalt_buf[33]; esalt_buf2[ 2] = esalt_bufs[salt_pos].esalt_buf[34]; @@ -5149,6 +4953,7 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p // append the pass to the salt u32x block0[16]; + u32x block1[16]; block0[ 0] = salt_buf0[ 0]; block0[ 1] = salt_buf0[ 1]; @@ -5166,9 +4971,6 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p block0[13] = salt_buf0[13]; block0[14] = salt_buf0[14]; block0[15] = salt_buf0[15]; - - u32x block1[16]; - block1[ 0] = salt_buf1[ 0]; block1[ 1] = salt_buf1[ 1]; block1[ 2] = salt_buf1[ 2]; @@ -5195,17 +4997,14 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = w0[1]; w0_t[2] = w0[2]; w0_t[3] = w0[3]; - w1_t[0] = w1[0]; w1_t[1] = w1[1]; w1_t[2] = w1[2]; w1_t[3] = w1[3]; - w2_t[0] = w2[0]; w2_t[1] = w2[1]; w2_t[2] = w2[2]; w2_t[3] = w2[3]; - w3_t[0] = w3[0]; w3_t[1] = w3[1]; w3_t[2] = w3[2]; @@ -5217,17 +5016,14 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block0[ 1]; w0_t[2] = block0[ 2]; w0_t[3] = block0[ 3]; - w1_t[0] = block0[ 4]; w1_t[1] = block0[ 5]; w1_t[2] = block0[ 6]; w1_t[3] = block0[ 7]; - w2_t[0] = block0[ 8]; w2_t[1] = block0[ 9]; w2_t[2] = block0[10]; w2_t[3] = block0[11]; - w3_t[0] = block0[12]; w3_t[1] = block0[13]; w3_t[2] = block0[14]; @@ -5322,17 +5118,14 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = block1[ 1]; w0_t[2] = block1[ 2]; w0_t[3] = block1[ 3]; - w1_t[0] = block1[ 4]; w1_t[1] = block1[ 5]; w1_t[2] = block1[ 6]; w1_t[3] = block1[ 7]; - w2_t[0] = block1[ 8]; w2_t[1] = block1[ 9]; w2_t[2] = block1[10]; w2_t[3] = block1[11]; - w3_t[0] = block1[12]; w3_t[1] = block1[13]; w3_t[2] = pw_salt_len * 8; @@ -5432,12 +5225,10 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p | uint_to_hex_lower8 ((d >> 8) & 255) << 16; w1_t[3] = uint_to_hex_lower8 ((d >> 16) & 255) << 0 | uint_to_hex_lower8 ((d >> 24) & 255) << 16; - w2_t[0] = esalt_buf0[0]; w2_t[1] = esalt_buf0[1]; w2_t[2] = esalt_buf0[2]; w2_t[3] = esalt_buf0[3]; - w3_t[0] = esalt_buf0[4]; w3_t[1] = esalt_buf0[5]; w3_t[2] = esalt_buf0[6]; @@ -5535,17 +5326,14 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf0[ 9]; w0_t[2] = esalt_buf0[10]; w0_t[3] = esalt_buf0[11]; - w1_t[0] = esalt_buf0[12]; w1_t[1] = esalt_buf0[13]; w1_t[2] = esalt_buf0[14]; w1_t[3] = esalt_buf0[15]; - w2_t[0] = esalt_buf1[ 0]; w2_t[1] = esalt_buf1[ 1]; w2_t[2] = esalt_buf1[ 2]; w2_t[3] = esalt_buf1[ 3]; - w3_t[0] = esalt_buf1[ 4]; w3_t[1] = esalt_buf1[ 5]; w3_t[2] = esalt_buf1[ 6]; @@ -5635,17 +5423,14 @@ static void m11400s_1_1 (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 p w0_t[1] = esalt_buf1[ 9]; w0_t[2] = esalt_buf1[10]; w0_t[3] = esalt_buf1[11]; - w1_t[0] = esalt_buf1[12]; w1_t[1] = esalt_buf1[13]; w1_t[2] = esalt_buf1[14]; w1_t[3] = esalt_buf1[15]; - w2_t[0] = esalt_buf2[ 0]; w2_t[1] = esalt_buf2[ 1]; w2_t[2] = esalt_buf2[ 2]; w2_t[3] = esalt_buf2[ 3]; - w3_t[0] = esalt_buf2[ 4]; w3_t[1] = esalt_buf2[ 5]; w3_t[2] = digest_esalt_len * 8;