/** * Authors.....: Jens Steube * Gabriele Gristina * * License.....: MIT */ #define _LOTUS8_ #include "inc_vendor.cl" #include "inc_hash_constants.h" #include "inc_hash_functions.cl" #undef _SHA1_ #include "inc_types.cl" #include "inc_common.cl" #define COMPARE_S "inc_comp_single.cl" #define COMPARE_M "inc_comp_multi.cl" __constant char lotus64_table[] = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz+/"; #define uint_to_hex_upper8(i) l_bin2asc[(i)] __constant u32 lotus_magic_table[256] = { 0xbd, 0x56, 0xea, 0xf2, 0xa2, 0xf1, 0xac, 0x2a, 0xb0, 0x93, 0xd1, 0x9c, 0x1b, 0x33, 0xfd, 0xd0, 0x30, 0x04, 0xb6, 0xdc, 0x7d, 0xdf, 0x32, 0x4b, 0xf7, 0xcb, 0x45, 0x9b, 0x31, 0xbb, 0x21, 0x5a, 0x41, 0x9f, 0xe1, 0xd9, 0x4a, 0x4d, 0x9e, 0xda, 0xa0, 0x68, 0x2c, 0xc3, 0x27, 0x5f, 0x80, 0x36, 0x3e, 0xee, 0xfb, 0x95, 0x1a, 0xfe, 0xce, 0xa8, 0x34, 0xa9, 0x13, 0xf0, 0xa6, 0x3f, 0xd8, 0x0c, 0x78, 0x24, 0xaf, 0x23, 0x52, 0xc1, 0x67, 0x17, 0xf5, 0x66, 0x90, 0xe7, 0xe8, 0x07, 0xb8, 0x60, 0x48, 0xe6, 0x1e, 0x53, 0xf3, 0x92, 0xa4, 0x72, 0x8c, 0x08, 0x15, 0x6e, 0x86, 0x00, 0x84, 0xfa, 0xf4, 0x7f, 0x8a, 0x42, 0x19, 0xf6, 0xdb, 0xcd, 0x14, 0x8d, 0x50, 0x12, 0xba, 0x3c, 0x06, 0x4e, 0xec, 0xb3, 0x35, 0x11, 0xa1, 0x88, 0x8e, 0x2b, 0x94, 0x99, 0xb7, 0x71, 0x74, 0xd3, 0xe4, 0xbf, 0x3a, 0xde, 0x96, 0x0e, 0xbc, 0x0a, 0xed, 0x77, 0xfc, 0x37, 0x6b, 0x03, 0x79, 0x89, 0x62, 0xc6, 0xd7, 0xc0, 0xd2, 0x7c, 0x6a, 0x8b, 0x22, 0xa3, 0x5b, 0x05, 0x5d, 0x02, 0x75, 0xd5, 0x61, 0xe3, 0x18, 0x8f, 0x55, 0x51, 0xad, 0x1f, 0x0b, 0x5e, 0x85, 0xe5, 0xc2, 0x57, 0x63, 0xca, 0x3d, 0x6c, 0xb4, 0xc5, 0xcc, 0x70, 0xb2, 0x91, 0x59, 0x0d, 0x47, 0x20, 0xc8, 0x4f, 0x58, 0xe0, 0x01, 0xe2, 0x16, 0x38, 0xc4, 0x6f, 0x3b, 0x0f, 0x65, 0x46, 0xbe, 0x7e, 0x2d, 0x7b, 0x82, 0xf9, 0x40, 0xb5, 0x1d, 0x73, 0xf8, 0xeb, 0x26, 0xc7, 0x87, 0x97, 0x25, 0x54, 0xb1, 0x28, 0xaa, 0x98, 0x9d, 0xa5, 0x64, 0x6d, 0x7a, 0xd4, 0x10, 0x81, 0x44, 0xef, 0x49, 0xd6, 0xae, 0x2e, 0xdd, 0x76, 0x5c, 0x2f, 0xa7, 0x1c, 0xc9, 0x09, 0x69, 0x9a, 0x83, 0xcf, 0x29, 0x39, 0xb9, 0xe9, 0x4c, 0xff, 0x43, 0xab, }; #if VECT_SIZE == 1 #define BOX1(S,i) (S)[(i)] #elif VECT_SIZE == 2 #define BOX1(S,i) (u32x) ((S)[(i).s0], (S)[(i).s1]) #elif VECT_SIZE == 4 #define BOX1(S,i) (u32x) ((S)[(i).s0], (S)[(i).s1], (S)[(i).s2], (S)[(i).s3]) #elif VECT_SIZE == 8 #define BOX1(S,i) (u32x) ((S)[(i).s0], (S)[(i).s1], (S)[(i).s2], (S)[(i).s3], (S)[(i).s4], (S)[(i).s5], (S)[(i).s6], (S)[(i).s7]) #elif VECT_SIZE == 16 #define BOX1(S,i) (u32x) ((S)[(i).s0], (S)[(i).s1], (S)[(i).s2], (S)[(i).s3], (S)[(i).s4], (S)[(i).s5], (S)[(i).s6], (S)[(i).s7], (S)[(i).s8], (S)[(i).s9], (S)[(i).sa], (S)[(i).sb], (S)[(i).sc], (S)[(i).sd], (S)[(i).se], (S)[(i).sf]) #endif void lotus_mix (u32x *in, __local u32 *s_lotus_magic_table) { u32x p = 0; for (int i = 0; i < 18; i++) { u32 s = 48; for (int j = 0; j < 12; j++) { u32x tmp_in = in[j]; u32x tmp_out = 0; p = (p + s--) & 0xff; p = ((tmp_in >> 0) & 0xff) ^ BOX1 (s_lotus_magic_table, p); tmp_out |= p << 0; p = (p + s--) & 0xff; p = ((tmp_in >> 8) & 0xff) ^ BOX1 (s_lotus_magic_table, p); tmp_out |= p << 8; p = (p + s--) & 0xff; p = ((tmp_in >> 16) & 0xff) ^ BOX1 (s_lotus_magic_table, p); tmp_out |= p << 16; p = (p + s--) & 0xff; p = ((tmp_in >> 24) & 0xff) ^ BOX1 (s_lotus_magic_table, p); tmp_out |= p << 24; in[j] = tmp_out; } } } void lotus_transform_password (u32x in[4], u32x out[4], __local u32 *s_lotus_magic_table) { u32x t = out[3] >> 24; u32x c; #ifdef _unroll #pragma unroll #endif for (int i = 0; i < 4; i++) { t ^= (in[i] >> 0) & 0xff; c = BOX1 (s_lotus_magic_table, t); out[i] ^= c << 0; t = ((out[i] >> 0) & 0xff); t ^= (in[i] >> 8) & 0xff; c = BOX1 (s_lotus_magic_table, t); out[i] ^= c << 8; t = ((out[i] >> 8) & 0xff); t ^= (in[i] >> 16) & 0xff; c = BOX1 (s_lotus_magic_table, t); out[i] ^= c << 16; t = ((out[i] >> 16) & 0xff); t ^= (in[i] >> 24) & 0xff; c = BOX1 (s_lotus_magic_table, t); out[i] ^= c << 24; t = ((out[i] >> 24) & 0xff); } } void pad (u32 w[4], const u32 len) { const u32 val = 16 - len; const u32 mask1 = val << 24; const u32 mask2 = val << 16 | val << 24; const u32 mask3 = val << 8 | val << 16 | val << 24; const u32 mask4 = val << 0 | val << 8 | val << 16 | val << 24; switch (len) { case 0: w[0] = mask4; w[1] = mask4; w[2] = mask4; w[3] = mask4; break; case 1: w[0] |= mask3; w[1] = mask4; w[2] = mask4; w[3] = mask4; break; case 2: w[0] |= mask2; w[1] = mask4; w[2] = mask4; w[3] = mask4; break; case 3: w[0] |= mask1; w[1] = mask4; w[2] = mask4; w[3] = mask4; break; case 4: w[1] = mask4; w[2] = mask4; w[3] = mask4; break; case 5: w[1] |= mask3; w[2] = mask4; w[3] = mask4; break; case 6: w[1] |= mask2; w[2] = mask4; w[3] = mask4; break; case 7: w[1] |= mask1; w[2] = mask4; w[3] = mask4; break; case 8: w[2] = mask4; w[3] = mask4; break; case 9: w[2] |= mask3; w[3] = mask4; break; case 10: w[2] |= mask2; w[3] = mask4; break; case 11: w[2] |= mask1; w[3] = mask4; break; case 12: w[3] = mask4; break; case 13: w[3] |= mask3; break; case 14: w[3] |= mask2; break; case 15: w[3] |= mask1; break; } } void mdtransform_norecalc (u32x state[4], u32x block[4], __local u32 *s_lotus_magic_table) { u32x x[12]; x[ 0] = state[0]; x[ 1] = state[1]; x[ 2] = state[2]; x[ 3] = state[3]; x[ 4] = block[0]; x[ 5] = block[1]; x[ 6] = block[2]; x[ 7] = block[3]; x[ 8] = state[0] ^ block[0]; x[ 9] = state[1] ^ block[1]; x[10] = state[2] ^ block[2]; x[11] = state[3] ^ block[3]; lotus_mix (x, s_lotus_magic_table); state[0] = x[0]; state[1] = x[1]; state[2] = x[2]; state[3] = x[3]; } void mdtransform (u32x state[4], u32x checksum[4], u32x block[4], __local u32 *s_lotus_magic_table) { mdtransform_norecalc (state, block, s_lotus_magic_table); lotus_transform_password (block, checksum, s_lotus_magic_table); } void domino_big_md (const u32x saved_key[16], const u32 size, u32x state[4], __local u32 *s_lotus_magic_table) { u32x checksum[4]; checksum[0] = 0; checksum[1] = 0; checksum[2] = 0; checksum[3] = 0; u32x block[4]; block[0] = 0; block[1] = 0; block[2] = 0; block[3] = 0; u32 curpos; u32 idx; for (curpos = 0, idx = 0; curpos + 16 < size; curpos += 16, idx += 4) { block[0] = saved_key[idx + 0]; block[1] = saved_key[idx + 1]; block[2] = saved_key[idx + 2]; block[3] = saved_key[idx + 3]; mdtransform (state, checksum, block, s_lotus_magic_table); } block[0] = saved_key[idx + 0]; block[1] = saved_key[idx + 1]; block[2] = saved_key[idx + 2]; block[3] = saved_key[idx + 3]; mdtransform (state, checksum, block, s_lotus_magic_table); mdtransform_norecalc (state, checksum, s_lotus_magic_table); } void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5]) { u32 A = digest[0]; u32 B = digest[1]; u32 C = digest[2]; u32 D = digest[3]; u32 E = digest[4]; u32 w0_t = w0[0]; u32 w1_t = w0[1]; u32 w2_t = w0[2]; u32 w3_t = w0[3]; u32 w4_t = w1[0]; u32 w5_t = w1[1]; u32 w6_t = w1[2]; u32 w7_t = w1[3]; u32 w8_t = w2[0]; u32 w9_t = w2[1]; u32 wa_t = w2[2]; u32 wb_t = w2[3]; u32 wc_t = w3[0]; u32 wd_t = w3[1]; u32 we_t = w3[2]; u32 wf_t = w3[3]; #undef K #define K SHA1C00 SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w3_t); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w4_t); SHA1_STEP (SHA1_F0o, A, B, C, D, E, w5_t); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w6_t); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w7_t); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w8_t); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w9_t); SHA1_STEP (SHA1_F0o, A, B, C, D, E, wa_t); SHA1_STEP (SHA1_F0o, E, A, B, C, D, wb_t); SHA1_STEP (SHA1_F0o, D, E, A, B, C, wc_t); SHA1_STEP (SHA1_F0o, C, D, E, A, B, wd_t); SHA1_STEP (SHA1_F0o, B, C, D, E, A, we_t); SHA1_STEP (SHA1_F0o, A, B, C, D, E, wf_t); w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w0_t); w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w1_t); w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w2_t); w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w3_t); #undef K #define K SHA1C01 w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w4_t); w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w5_t); w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w6_t); w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w7_t); w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w8_t); w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w9_t); wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wa_t); wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wb_t); wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wc_t); wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wd_t); we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, we_t); wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wf_t); w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w0_t); w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w1_t); w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w2_t); w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w3_t); w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w4_t); w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w5_t); w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w6_t); w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w7_t); #undef K #define K SHA1C02 w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w8_t); w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w9_t); wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wa_t); wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wb_t); wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wc_t); wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, wd_t); we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, we_t); wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wf_t); w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w0_t); w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w1_t); w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w2_t); w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w3_t); w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w4_t); w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w5_t); w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w6_t); w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w7_t); w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w8_t); w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w9_t); wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wa_t); wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wb_t); #undef K #define K SHA1C03 wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wc_t); wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wd_t); we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, we_t); wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wf_t); w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w0_t); w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w1_t); w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w2_t); w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w3_t); w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w4_t); w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w5_t); w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w6_t); w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w7_t); w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w8_t); w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w9_t); wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wa_t); wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wb_t); wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wc_t); wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wd_t); we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, we_t); wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wf_t); digest[0] += A; digest[1] += B; digest[2] += C; digest[3] += D; digest[4] += E; } void hmac_sha1_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5]) { w0[0] = w0[0] ^ 0x36363636; w0[1] = w0[1] ^ 0x36363636; w0[2] = w0[2] ^ 0x36363636; w0[3] = w0[3] ^ 0x36363636; w1[0] = w1[0] ^ 0x36363636; w1[1] = w1[1] ^ 0x36363636; w1[2] = w1[2] ^ 0x36363636; w1[3] = w1[3] ^ 0x36363636; w2[0] = w2[0] ^ 0x36363636; w2[1] = w2[1] ^ 0x36363636; w2[2] = w2[2] ^ 0x36363636; w2[3] = w2[3] ^ 0x36363636; w3[0] = w3[0] ^ 0x36363636; w3[1] = w3[1] ^ 0x36363636; w3[2] = w3[2] ^ 0x36363636; w3[3] = w3[3] ^ 0x36363636; ipad[0] = SHA1M_A; ipad[1] = SHA1M_B; ipad[2] = SHA1M_C; ipad[3] = SHA1M_D; ipad[4] = SHA1M_E; sha1_transform (w0, w1, w2, w3, ipad); w0[0] = w0[0] ^ 0x6a6a6a6a; w0[1] = w0[1] ^ 0x6a6a6a6a; w0[2] = w0[2] ^ 0x6a6a6a6a; w0[3] = w0[3] ^ 0x6a6a6a6a; w1[0] = w1[0] ^ 0x6a6a6a6a; w1[1] = w1[1] ^ 0x6a6a6a6a; w1[2] = w1[2] ^ 0x6a6a6a6a; w1[3] = w1[3] ^ 0x6a6a6a6a; w2[0] = w2[0] ^ 0x6a6a6a6a; w2[1] = w2[1] ^ 0x6a6a6a6a; w2[2] = w2[2] ^ 0x6a6a6a6a; w2[3] = w2[3] ^ 0x6a6a6a6a; w3[0] = w3[0] ^ 0x6a6a6a6a; w3[1] = w3[1] ^ 0x6a6a6a6a; w3[2] = w3[2] ^ 0x6a6a6a6a; w3[3] = w3[3] ^ 0x6a6a6a6a; opad[0] = SHA1M_A; opad[1] = SHA1M_B; opad[2] = SHA1M_C; opad[3] = SHA1M_D; opad[4] = SHA1M_E; sha1_transform (w0, w1, w2, w3, opad); } void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5], u32 digest[5]) { digest[0] = ipad[0]; digest[1] = ipad[1]; digest[2] = ipad[2]; digest[3] = ipad[3]; digest[4] = ipad[4]; sha1_transform (w0, w1, w2, w3, digest); w0[0] = digest[0]; w0[1] = digest[1]; w0[2] = digest[2]; w0[3] = digest[3]; w1[0] = digest[4]; w1[1] = 0x80000000; w1[2] = 0; w1[3] = 0; w2[0] = 0; w2[1] = 0; w2[2] = 0; w2[3] = 0; w3[0] = 0; w3[1] = 0; w3[2] = 0; w3[3] = (64 + 20) * 8; digest[0] = opad[0]; digest[1] = opad[1]; digest[2] = opad[2]; digest[3] = opad[3]; digest[4] = opad[4]; sha1_transform (w0, w1, w2, w3, digest); } void base64_encode (u8 *base64_hash, const u32 len, const u8 *base64_plain) { u8 *out_ptr = (u8 *) base64_hash; u8 *in_ptr = (u8 *) base64_plain; u32 i; for (i = 0; i < len; i += 3) { char out_val0 = lotus64_table [ ((in_ptr[0] >> 2) & 0x3f)]; char out_val1 = lotus64_table [((in_ptr[0] << 4) & 0x30) | ((in_ptr[1] >> 4) & 0x0f)]; char out_val2 = lotus64_table [((in_ptr[1] << 2) & 0x3c) | ((in_ptr[2] >> 6) & 0x03)]; char out_val3 = lotus64_table [ ((in_ptr[2] >> 0) & 0x3f)]; out_ptr[0] = out_val0 & 0x7f; out_ptr[1] = out_val1 & 0x7f; out_ptr[2] = out_val2 & 0x7f; out_ptr[3] = out_val3 & 0x7f; in_ptr += 3; out_ptr += 4; } } void lotus6_base64_encode (u8 base64_hash[24], const u32 salt0, const u32 salt1, u32 a, u32 b, u32 c) { uchar4 salt0c = as_uchar4 (salt0); uchar4 salt1c = as_uchar4 (salt1); uchar4 ac; uchar4 bc; uchar4 cc; ac = as_uchar4 (a); bc = as_uchar4 (b); cc = as_uchar4 (c); u8 tmp[24]; // size 22 (=pw_len) is needed but base64 needs size divisible by 4 /* * Copy $salt.$digest to a tmp buffer */ u8 base64_plain[16]; base64_plain[ 0] = salt0c.s0; base64_plain[ 1] = salt0c.s1; base64_plain[ 2] = salt0c.s2; base64_plain[ 3] = salt0c.s3; base64_plain[ 3] -= -4; // dont ask! base64_plain[ 4] = salt1c.s0; base64_plain[ 5] = ac.s0; base64_plain[ 6] = ac.s1; base64_plain[ 7] = ac.s2; base64_plain[ 8] = ac.s3; base64_plain[ 9] = bc.s0; base64_plain[10] = bc.s1; base64_plain[11] = bc.s2; base64_plain[12] = bc.s3; base64_plain[13] = cc.s0; base64_plain[14] = cc.s1; base64_plain[15] = cc.s2; /* * base64 encode the $salt.$digest string */ base64_encode (tmp + 2, 14, base64_plain); base64_hash[ 0] = '('; base64_hash[ 1] = 'G'; base64_hash[ 2] = tmp[ 2]; base64_hash[ 3] = tmp[ 3]; base64_hash[ 4] = tmp[ 4]; base64_hash[ 5] = tmp[ 5]; base64_hash[ 6] = tmp[ 6]; base64_hash[ 7] = tmp[ 7]; base64_hash[ 8] = tmp[ 8]; base64_hash[ 9] = tmp[ 9]; base64_hash[10] = tmp[10]; base64_hash[11] = tmp[11]; base64_hash[12] = tmp[12]; base64_hash[13] = tmp[13]; base64_hash[14] = tmp[14]; base64_hash[15] = tmp[15]; base64_hash[16] = tmp[16]; base64_hash[17] = tmp[17]; base64_hash[18] = tmp[18]; base64_hash[19] = tmp[19]; base64_hash[20] = tmp[20]; base64_hash[21] = ')'; } __kernel void m09100_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global lotus8_tmp_t *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 wpa_t *wpa_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { /** * base */ const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); const u32 lsz = get_local_size (0); /** * sbox */ __local u32 s_lotus_magic_table[256]; for (u32 i = lid; i < 256; i += lsz) { s_lotus_magic_table[i] = lotus_magic_table[i]; } __local u32 l_bin2asc[256]; for (u32 i = lid; i < 256; i += lsz) { const u32 i0 = (i >> 0) & 15; const u32 i1 = (i >> 4) & 15; l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'A' - 10 + i0) << 8 | ((i1 < 10) ? '0' + i1 : 'A' - 10 + i1) << 0; } barrier (CLK_LOCAL_MEM_FENCE); if (gid >= gid_max) return; /** * base */ u32 w[16]; w[ 0] = pws[gid].i[ 0]; w[ 1] = pws[gid].i[ 1]; w[ 2] = pws[gid].i[ 2]; w[ 3] = pws[gid].i[ 3]; w[ 4] = pws[gid].i[ 4]; w[ 5] = pws[gid].i[ 5]; w[ 6] = pws[gid].i[ 6]; w[ 7] = pws[gid].i[ 7]; w[ 8] = pws[gid].i[ 8]; w[ 9] = pws[gid].i[ 9]; w[10] = pws[gid].i[10]; w[11] = pws[gid].i[11]; w[12] = pws[gid].i[12]; w[13] = pws[gid].i[13]; w[14] = pws[gid].i[14]; w[15] = pws[gid].i[15]; /** * pad */ u32 pw_len = pws[gid].pw_len; if (pw_len < 16) { pad (&w[ 0], pw_len & 0xf); } else if (pw_len < 32) { pad (&w[ 4], pw_len & 0xf); } else if (pw_len < 48) { pad (&w[ 8], pw_len & 0xf); } else if (pw_len < 64) { pad (&w[12], pw_len & 0xf); } /** * salt */ u32 salt_len = salt_bufs[salt_pos].salt_len; u32 salt_buf0[4]; salt_buf0[0] = salt_bufs[salt_pos].salt_buf[ 0]; salt_buf0[1] = salt_bufs[salt_pos].salt_buf[ 1]; salt_buf0[2] = salt_bufs[salt_pos].salt_buf[ 2]; salt_buf0[3] = salt_bufs[salt_pos].salt_buf[ 3]; u32 salt_buf1[4]; salt_buf1[0] = 0x01000000; salt_buf1[1] = 0x00000080; salt_buf1[2] = 0; salt_buf1[3] = 0; u32 salt_buf2[4]; salt_buf2[0] = 0; salt_buf2[1] = 0; salt_buf2[2] = 0; salt_buf2[3] = 0; u32 salt_buf3[4]; salt_buf3[0] = 0; salt_buf3[1] = 0; salt_buf3[2] = 0; salt_buf3[3] = 0; const u32 salt0 = salt_buf0[0]; const u32 salt1 = (salt_buf0[1] & 0xff) | ('(' << 8); /** * Lotus 6 hash - SEC_pwddigest_V2 */ u32 w_tmp[16]; w_tmp[ 0] = w[ 0]; w_tmp[ 1] = w[ 1]; w_tmp[ 2] = w[ 2]; w_tmp[ 3] = w[ 3]; w_tmp[ 4] = w[ 4]; w_tmp[ 5] = w[ 5]; w_tmp[ 6] = w[ 6]; w_tmp[ 7] = w[ 7]; w_tmp[ 8] = w[ 8]; w_tmp[ 9] = w[ 9]; w_tmp[10] = w[10]; w_tmp[11] = w[11]; w_tmp[12] = w[12]; w_tmp[13] = w[13]; w_tmp[14] = w[14]; w_tmp[15] = w[15]; u32 state[4]; state[0] = 0; state[1] = 0; state[2] = 0; state[3] = 0; domino_big_md (w_tmp, pw_len, state, s_lotus_magic_table); const u32 w0_t = uint_to_hex_upper8 ((state[0] >> 0) & 255) << 0 | uint_to_hex_upper8 ((state[0] >> 8) & 255) << 16; const u32 w1_t = uint_to_hex_upper8 ((state[0] >> 16) & 255) << 0 | uint_to_hex_upper8 ((state[0] >> 24) & 255) << 16; const u32 w2_t = uint_to_hex_upper8 ((state[1] >> 0) & 255) << 0 | uint_to_hex_upper8 ((state[1] >> 8) & 255) << 16; const u32 w3_t = uint_to_hex_upper8 ((state[1] >> 16) & 255) << 0 | uint_to_hex_upper8 ((state[1] >> 24) & 255) << 16; const u32 w4_t = uint_to_hex_upper8 ((state[2] >> 0) & 255) << 0 | uint_to_hex_upper8 ((state[2] >> 8) & 255) << 16; const u32 w5_t = uint_to_hex_upper8 ((state[2] >> 16) & 255) << 0 | uint_to_hex_upper8 ((state[2] >> 24) & 255) << 16; const u32 w6_t = uint_to_hex_upper8 ((state[3] >> 0) & 255) << 0 | uint_to_hex_upper8 ((state[3] >> 8) & 255) << 16; const u32 pade = 0x0e0e0e0e; w_tmp[ 0] = salt0; w_tmp[ 1] = salt1 | w0_t << 16; w_tmp[ 2] = w0_t >> 16 | w1_t << 16; w_tmp[ 3] = w1_t >> 16 | w2_t << 16; w_tmp[ 4] = w2_t >> 16 | w3_t << 16; w_tmp[ 5] = w3_t >> 16 | w4_t << 16; w_tmp[ 6] = w4_t >> 16 | w5_t << 16; w_tmp[ 7] = w5_t >> 16 | w6_t << 16; w_tmp[ 8] = w6_t >> 16 | pade << 16; w_tmp[ 9] = pade; w_tmp[10] = pade; w_tmp[11] = pade; w_tmp[12] = 0; w_tmp[13] = 0; w_tmp[14] = 0; w_tmp[15] = 0; state[0] = 0; state[1] = 0; state[2] = 0; state[3] = 0; domino_big_md (w_tmp, 34, state, s_lotus_magic_table); u32 a = state[0]; u32 b = state[1]; u32 c = state[2]; /** * Base64 encode */ pw_len = 22; u8 base64_hash[22]; lotus6_base64_encode (base64_hash, salt_buf0[0], salt_buf0[1], a, b, c); /** * PBKDF2 - HMACSHA1 - 1st iteration */ u32 w0[4]; u32 w1[4]; u32 w2[4]; u32 w3[4]; w0[0] = (base64_hash[ 0] << 24) | (base64_hash[ 1] << 16) | (base64_hash[ 2] << 8) | base64_hash[ 3]; w0[1] = (base64_hash[ 4] << 24) | (base64_hash[ 5] << 16) | (base64_hash[ 6] << 8) | base64_hash[ 7]; w0[2] = (base64_hash[ 8] << 24) | (base64_hash[ 9] << 16) | (base64_hash[10] << 8) | base64_hash[11]; w0[3] = (base64_hash[12] << 24) | (base64_hash[13] << 16) | (base64_hash[14] << 8) | base64_hash[15]; w1[0] = (base64_hash[16] << 24) | (base64_hash[17] << 16) | (base64_hash[18] << 8) | base64_hash[19]; w1[1] = (base64_hash[20] << 24) | (base64_hash[21] << 16); w1[2] = 0; w1[3] = 0; w2[0] = 0; w2[1] = 0; w2[2] = 0; w2[3] = 0; w3[0] = 0; w3[1] = 0; w3[2] = 0; w3[3] = 0; /** * pads */ u32 ipad[5]; u32 opad[5]; hmac_sha1_pad (w0, w1, w2, w3, ipad, opad); tmps[gid].ipad[0] = ipad[0]; tmps[gid].ipad[1] = ipad[1]; tmps[gid].ipad[2] = ipad[2]; tmps[gid].ipad[3] = ipad[3]; tmps[gid].ipad[4] = ipad[4]; tmps[gid].opad[0] = opad[0]; tmps[gid].opad[1] = opad[1]; tmps[gid].opad[2] = opad[2]; tmps[gid].opad[3] = opad[3]; tmps[gid].opad[4] = opad[4]; w0[0] = salt_buf0[0]; w0[1] = salt_buf0[1]; w0[2] = salt_buf0[2]; w0[3] = salt_buf0[3]; w1[0] = salt_buf1[0]; w1[1] = salt_buf1[1]; w1[2] = salt_buf1[2]; w1[3] = salt_buf1[3]; w2[0] = salt_buf2[0]; w2[1] = salt_buf2[1]; w2[2] = salt_buf2[2]; w2[3] = salt_buf2[3]; w3[0] = salt_buf3[0]; w3[1] = salt_buf3[1]; w3[2] = salt_buf3[2]; //w3[3] = salt_buf3[3]; w0[0] = swap32 (w0[0]); w0[1] = swap32 (w0[1]); w0[2] = swap32 (w0[2]); w0[3] = swap32 (w0[3]); w1[0] = swap32 (w1[0]); w1[1] = swap32 (w1[1]); w1[2] = swap32 (w1[2]); w1[3] = swap32 (w1[3]); w2[0] = swap32 (w2[0]); w2[1] = swap32 (w2[1]); w2[2] = swap32 (w2[2]); w2[3] = swap32 (w2[3]); w3[0] = swap32 (w3[0]); w3[1] = swap32 (w3[1]); w3[2] = swap32 (w3[2]); w3[3] = (64 + salt_len + 4) * 8; u32 dgst[5]; hmac_sha1_run (w0, w1, w2, w3, ipad, opad, dgst); tmps[gid].dgst[0] = dgst[0]; tmps[gid].dgst[1] = dgst[1]; tmps[gid].dgst[2] = dgst[2]; tmps[gid].dgst[3] = dgst[3]; tmps[gid].dgst[4] = dgst[4]; tmps[gid].out[0] = dgst[0]; tmps[gid].out[1] = dgst[1]; tmps[gid].out[2] = dgst[2]; tmps[gid].out[3] = dgst[3]; tmps[gid].out[4] = dgst[4]; } __kernel void m09100_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global lotus8_tmp_t *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 wpa_t *wpa_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { const u32 gid = get_global_id (0); if (gid >= gid_max) return; u32 ipad[5]; u32 opad[5]; ipad[0] = tmps[gid].ipad[0]; ipad[1] = tmps[gid].ipad[1]; ipad[2] = tmps[gid].ipad[2]; ipad[3] = tmps[gid].ipad[3]; ipad[4] = tmps[gid].ipad[4]; opad[0] = tmps[gid].opad[0]; opad[1] = tmps[gid].opad[1]; opad[2] = tmps[gid].opad[2]; opad[3] = tmps[gid].opad[3]; opad[4] = tmps[gid].opad[4]; u32 dgst[5]; u32 out[5]; dgst[0] = tmps[gid].dgst[0]; dgst[1] = tmps[gid].dgst[1]; dgst[2] = tmps[gid].dgst[2]; dgst[3] = tmps[gid].dgst[3]; dgst[4] = tmps[gid].dgst[4]; out[0] = tmps[gid].out[0]; out[1] = tmps[gid].out[1]; out[2] = tmps[gid].out[2]; out[3] = tmps[gid].out[3]; out[4] = tmps[gid].out[4]; for (u32 j = 0; j < loop_cnt; j++) { u32 w0[4]; u32 w1[4]; u32 w2[4]; u32 w3[4]; w0[0] = dgst[0]; w0[1] = dgst[1]; w0[2] = dgst[2]; w0[3] = dgst[3]; w1[0] = dgst[4]; w1[1] = 0x80000000; w1[2] = 0; w1[3] = 0; w2[0] = 0; w2[1] = 0; w2[2] = 0; w2[3] = 0; w3[0] = 0; w3[1] = 0; w3[2] = 0; w3[3] = (64 + 20) * 8; hmac_sha1_run (w0, w1, w2, w3, ipad, opad, dgst); out[0] ^= dgst[0]; out[1] ^= dgst[1]; out[2] ^= dgst[2]; out[3] ^= dgst[3]; out[4] ^= dgst[4]; } tmps[gid].dgst[0] = dgst[0]; tmps[gid].dgst[1] = dgst[1]; tmps[gid].dgst[2] = dgst[2]; tmps[gid].dgst[3] = dgst[3]; tmps[gid].dgst[4] = dgst[4]; tmps[gid].out[0] = out[0]; tmps[gid].out[1] = out[1]; tmps[gid].out[2] = out[2]; tmps[gid].out[3] = out[3]; tmps[gid].out[4] = out[4]; } __kernel void m09100_comp (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global lotus8_tmp_t *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 wpa_t *wpa_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { /** * base */ const u32 gid = get_global_id (0); if (gid >= gid_max) return; const u32 lid = get_local_id (0); /** * digest */ const u32 r0 = tmps[gid].out[DGST_R0]; const u32 r1 = tmps[gid].out[DGST_R1]; const u32 r2 = 0; const u32 r3 = 0; #define il_pos 0 #include COMPARE_M }