/** * Authors......: Jens Steube * Fist0urs * Gabriele Gristina * * License.....: MIT */ #define _KRB5TGS_ //too much register pressure //#define NEW_SIMD_CODE #include "include/constants.h" #include "include/kernel_vendor.h" #define DGST_R0 0 #define DGST_R1 1 #define DGST_R2 2 #define DGST_R3 3 #include "include/kernel_functions.c" #include "OpenCL/types_ocl.c" #include "OpenCL/common.c" #include "include/rp_kernel.h" #include "OpenCL/rp.c" #include "OpenCL/simd.c" typedef struct { u8 S[256]; u32 wtf_its_faster; } RC4_KEY; void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j) { u8 tmp; tmp = rc4_key->S[i]; rc4_key->S[i] = rc4_key->S[j]; rc4_key->S[j] = tmp; } void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4]) { u32 v = 0x03020100; u32 a = 0x04040404; __local u32 *ptr = (__local u32 *) rc4_key->S; #ifdef _unroll #pragma unroll #endif for (u32 i = 0; i < 64; i++) { *ptr++ = v; v += a; } u32 j = 0; for (u32 i = 0; i < 16; i++) { u32 idx = i * 16; u32 v; v = data[0]; j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++; v = data[1]; j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++; v = data[2]; j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++; v = data[3]; j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++; j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++; } } u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __global u32 *in, u32 out[4]) { #ifdef _unroll #pragma unroll #endif for (u32 k = 0; k < 4; k++) { u32 xor4 = 0; u8 idx; i += 1; j += rc4_key->S[i]; swap (rc4_key, i, j); idx = rc4_key->S[i] + rc4_key->S[j]; xor4 |= rc4_key->S[idx] << 0; i += 1; j += rc4_key->S[i]; swap (rc4_key, i, j); idx = rc4_key->S[i] + rc4_key->S[j]; xor4 |= rc4_key->S[idx] << 8; i += 1; j += rc4_key->S[i]; swap (rc4_key, i, j); idx = rc4_key->S[i] + rc4_key->S[j]; xor4 |= rc4_key->S[idx] << 16; i += 1; j += rc4_key->S[i]; swap (rc4_key, i, j); idx = rc4_key->S[i] + rc4_key->S[j]; xor4 |= rc4_key->S[idx] << 24; out[k] = in[k] ^ xor4; } return j; } void md4_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4]) { u32 a = digest[0]; u32 b = digest[1]; u32 c = digest[2]; u32 d = digest[3]; MD4_STEP (MD4_Fo, a, b, c, d, w0[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w0[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w0[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w0[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w1[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w1[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w1[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w1[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w2[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w2[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w2[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w2[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w3[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w3[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w3[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w3[3], MD4C00, MD4S03); MD4_STEP (MD4_Go, a, b, c, d, w0[0], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[0], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[0], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[0], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[1], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[1], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[1], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[1], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[2], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[2], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[2], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[2], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[3], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[3], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[3], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[3], MD4C01, MD4S13); MD4_STEP (MD4_H , a, b, c, d, w0[0], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[0], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[0], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[0], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[2], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[2], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[2], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[2], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[1], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[1], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[1], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[1], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[3], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[3], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[3], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[3], MD4C02, MD4S23); digest[0] += a; digest[1] += b; digest[2] += c; digest[3] += d; } void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4]) { u32 a = digest[0]; u32 b = digest[1]; u32 c = digest[2]; u32 d = digest[3]; 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]; MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03); MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13); MD5_STEP (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20); MD5_STEP (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21); MD5_STEP (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22); MD5_STEP (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23); MD5_STEP (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20); MD5_STEP (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21); MD5_STEP (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22); MD5_STEP (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23); MD5_STEP (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20); MD5_STEP (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21); MD5_STEP (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22); MD5_STEP (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23); MD5_STEP (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20); MD5_STEP (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21); MD5_STEP (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22); MD5_STEP (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23); MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31); MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33); MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31); MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30); MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32); MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30); MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33); digest[0] += a; digest[1] += b; digest[2] += c; digest[3] += d; } void hmac_md5_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32 opad[4]) { 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] = MD5M_A; ipad[1] = MD5M_B; ipad[2] = MD5M_C; ipad[3] = MD5M_D; md5_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] = MD5M_A; opad[1] = MD5M_B; opad[2] = MD5M_C; opad[3] = MD5M_D; md5_transform (w0, w1, w2, w3, opad); } void hmac_md5_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32 opad[4], u32 digest[4]) { digest[0] = ipad[0]; digest[1] = ipad[1]; digest[2] = ipad[2]; digest[3] = ipad[3]; md5_transform (w0, w1, w2, w3, digest); w0[0] = digest[0]; w0[1] = digest[1]; w0[2] = digest[2]; w0[3] = digest[3]; w1[0] = 0x80; w1[1] = 0; 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] = (64 + 16) * 8; w3[3] = 0; digest[0] = opad[0]; digest[1] = opad[1]; digest[2] = opad[2]; digest[3] = opad[3]; md5_transform (w0, w1, w2, w3, digest); } int decrypt_and_check (__local RC4_KEY *rc4_key, u32 data[4], __global u32 *edata2, const u32 edata2_len, const u32 K2[4], const u32 checksum[4]) { rc4_init_16 (rc4_key, data); u32 out0[4]; u32 out1[4]; u8 i = 0; u8 j = 0; /* 8 first bytes are nonce, then ASN1 structs (DER encoding: type-length-data) if length >= 128 bytes: length is on 2 bytes and type is \x63\x82 (encode_krb5_enc_tkt_part) and data is an ASN1 sequence \x30\x82 else: length is on 1 byte and type is \x63\x81 and data is an ASN1 sequence \x30\x81 next headers follow the same ASN1 "type-length-data" scheme */ j = rc4_next_16 (rc4_key, i, j, edata2 + 0, out0); i += 16; if (((out0[2] & 0xff00ffff) != 0x30008163) && ((out0[2] & 0x0000ffff) != 0x00008263)) return 0; j = rc4_next_16 (rc4_key, i, j, edata2 + 4, out1); i += 16; if (((out1[0] & 0x00ffffff) != 0x00000503) && (out1[0] != 0x050307A0)) return 0; rc4_init_16 (rc4_key, data); i = 0; j = 0; // init hmac u32 w0[4]; u32 w1[4]; u32 w2[4]; u32 w3[4]; w0[0] = K2[0]; w0[1] = K2[1]; w0[2] = K2[2]; w0[3] = K2[3]; w1[0] = 0; w1[1] = 0; 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; u32 ipad[4]; u32 opad[4]; hmac_md5_pad (w0, w1, w2, w3, ipad, opad); int edata2_left; for (edata2_left = edata2_len; edata2_left >= 64; edata2_left -= 64) { j = rc4_next_16 (rc4_key, i, j, edata2, w0); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w1); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w2); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w3); i += 16; edata2 += 4; md5_transform (w0, w1, w2, w3, ipad); } w0[0] = 0; w0[1] = 0; w0[2] = 0; w0[3] = 0; w1[0] = 0; w1[1] = 0; 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; if (edata2_left < 16) { j = rc4_next_16 (rc4_key, i, j, edata2, w0); i += 16; edata2 += 4; truncate_block (w0, edata2_left & 0xf); append_0x80_1x4 (w0, edata2_left & 0xf); w3[2] = (64 + edata2_len) * 8; w3[3] = 0; md5_transform (w0, w1, w2, w3, ipad); } else if (edata2_left < 32) { j = rc4_next_16 (rc4_key, i, j, edata2, w0); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w1); i += 16; edata2 += 4; truncate_block (w1, edata2_left & 0xf); append_0x80_1x4 (w1, edata2_left & 0xf); w3[2] = (64 + edata2_len) * 8; w3[3] = 0; md5_transform (w0, w1, w2, w3, ipad); } else if (edata2_left < 48) { j = rc4_next_16 (rc4_key, i, j, edata2, w0); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w1); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w2); i += 16; edata2 += 4; truncate_block (w2, edata2_left & 0xf); append_0x80_1x4 (w2, edata2_left & 0xf); w3[2] = (64 + edata2_len) * 8; w3[3] = 0; md5_transform (w0, w1, w2, w3, ipad); } else { j = rc4_next_16 (rc4_key, i, j, edata2, w0); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w1); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w2); i += 16; edata2 += 4; j = rc4_next_16 (rc4_key, i, j, edata2, w3); i += 16; edata2 += 4; truncate_block (w3, edata2_left & 0xf); append_0x80_1x4 (w3, edata2_left & 0xf); if (edata2_left < 56) { w3[2] = (64 + edata2_len) * 8; w3[3] = 0; md5_transform (w0, w1, w2, w3, ipad); } else { md5_transform (w0, w1, w2, w3, ipad); w0[0] = 0; w0[1] = 0; w0[2] = 0; w0[3] = 0; w1[0] = 0; w1[1] = 0; 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] = (64 + edata2_len) * 8; w3[3] = 0; md5_transform (w0, w1, w2, w3, ipad); } } w0[0] = ipad[0]; w0[1] = ipad[1]; w0[2] = ipad[2]; w0[3] = ipad[3]; w1[0] = 0x80; w1[1] = 0; 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] = (64 + 16) * 8; w3[3] = 0; md5_transform (w0, w1, w2, w3, opad); if (checksum[0] != opad[0]) return 0; if (checksum[1] != opad[1]) return 0; if (checksum[2] != opad[2]) return 0; if (checksum[3] != opad[3]) return 0; return 1; } void kerb_prepare (const u32 w0[4], const u32 w1[4], const u32 pw_len, const u32 checksum[4], u32 digest[4], u32 K2[4]) { /** * pads */ u32 w0_t[4]; u32 w1_t[4]; u32 w2_t[4]; u32 w3_t[4]; w0_t[0] = w0[0]; 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] = 0; w2_t[1] = 0; w2_t[2] = 0; w2_t[3] = 0; w3_t[0] = 0; w3_t[1] = 0; w3_t[2] = 0; w3_t[3] = 0; // K=MD4(Little_indian(UNICODE(pwd)) append_0x80_2x4 (w0_t, w1_t, pw_len); make_unicode (w1_t, w2_t, w3_t); make_unicode (w0_t, w0_t, w1_t); w3_t[2] = pw_len * 8 * 2; w3_t[3] = 0; digest[0] = MD4M_A; digest[1] = MD4M_B; digest[2] = MD4M_C; digest[3] = MD4M_D; md4_transform (w0_t, w1_t, w2_t, w3_t, digest); // K1=MD5_HMAC(K,1); with 2 encoded as little indian on 4 bytes (02000000 in hexa); w0_t[0] = digest[0]; w0_t[1] = digest[1]; w0_t[2] = digest[2]; w0_t[3] = digest[3]; w1_t[0] = 0; w1_t[1] = 0; w1_t[2] = 0; w1_t[3] = 0; w2_t[0] = 0; w2_t[1] = 0; w2_t[2] = 0; w2_t[3] = 0; w3_t[0] = 0; w3_t[1] = 0; w3_t[2] = 0; w3_t[3] = 0; u32 ipad[4]; u32 opad[4]; hmac_md5_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad); w0_t[0] = 2; w0_t[1] = 0x80; w0_t[2] = 0; w0_t[3] = 0; w1_t[0] = 0; w1_t[1] = 0; w1_t[2] = 0; w1_t[3] = 0; w2_t[0] = 0; w2_t[1] = 0; w2_t[2] = 0; w2_t[3] = 0; w3_t[0] = 0; w3_t[1] = 0; w3_t[2] = (64 + 4) * 8; w3_t[3] = 0; hmac_md5_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest); // K2 = K1; K2[0] = digest[0]; K2[1] = digest[1]; K2[2] = digest[2]; K2[3] = digest[3]; // K3=MD5_HMAC(K1,checksum); w0_t[0] = digest[0]; w0_t[1] = digest[1]; w0_t[2] = digest[2]; w0_t[3] = digest[3]; w1_t[0] = 0; w1_t[1] = 0; w1_t[2] = 0; w1_t[3] = 0; w2_t[0] = 0; w2_t[1] = 0; w2_t[2] = 0; w2_t[3] = 0; w3_t[0] = 0; w3_t[1] = 0; w3_t[2] = 0; w3_t[3] = 0; hmac_md5_pad (w0_t, w1_t, w2_t, w3_t, ipad, opad); w0_t[0] = checksum[0]; w0_t[1] = checksum[1]; w0_t[2] = checksum[2]; w0_t[3] = checksum[3]; w1_t[0] = 0x80; w1_t[1] = 0; w1_t[2] = 0; w1_t[3] = 0; w2_t[0] = 0; w2_t[1] = 0; w2_t[2] = 0; w2_t[3] = 0; w3_t[0] = 0; w3_t[1] = 0; w3_t[2] = (64 + 16) * 8; w3_t[3] = 0; hmac_md5_run (w0_t, w1_t, w2_t, w3_t, ipad, opad, digest); } void m13100 (__local RC4_KEY *rc4_keys, u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_len, __global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset) { /** * modifier */ const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); /** * salt */ u32 checksum[4]; checksum[0] = krb5tgs_bufs[salt_pos].checksum[0]; checksum[1] = krb5tgs_bufs[salt_pos].checksum[1]; checksum[2] = krb5tgs_bufs[salt_pos].checksum[2]; checksum[3] = krb5tgs_bufs[salt_pos].checksum[3]; /** * loop */ u32 w0l = w0[0]; for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) { const u32 w0r = bfs_buf[il_pos].i; w0[0] = w0l | w0r; u32 digest[4]; u32 K2[4]; kerb_prepare (w0, w1, pw_len, checksum, digest, K2); u32 tmp[4]; tmp[0] = digest[0]; tmp[1] = digest[1]; tmp[2] = digest[2]; tmp[3] = digest[3]; if (decrypt_and_check (&rc4_keys[lid], tmp, krb5tgs_bufs[salt_pos].edata2, krb5tgs_bufs[salt_pos].edata2_len, K2, checksum) == 1) { mark_hash (plains_buf, d_return_buf, salt_pos, 0, digests_offset + 0, gid, il_pos); d_return_buf[lid] = 1; } } } __kernel void m13100_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { /** * modifier */ const u32 lid = get_local_id (0); const u32 gid = get_global_id (0); if (gid >= gid_max) return; /** * base */ u32 pw_buf0[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]; pw_buf1[3] = pws[gid].i[ 7]; const u32 pw_len = pws[gid].pw_len; /** * shared */ __local RC4_KEY rc4_keys[64]; /** * salt */ u32 checksum[4]; checksum[0] = krb5tgs_bufs[salt_pos].checksum[0]; checksum[1] = krb5tgs_bufs[salt_pos].checksum[1]; checksum[2] = krb5tgs_bufs[salt_pos].checksum[2]; checksum[3] = krb5tgs_bufs[salt_pos].checksum[3]; /** * loop */ for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { u32x w0[4] = { 0 }; u32x w1[4] = { 0 }; u32x w2[4] = { 0 }; u32x w3[4] = { 0 }; const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1); /** * kerberos */ u32 digest[4]; u32 K2[4]; kerb_prepare (w0, w1, out_len, checksum, digest, K2); u32 tmp[4]; tmp[0] = digest[0]; tmp[1] = digest[1]; tmp[2] = digest[2]; tmp[3] = digest[3]; if (decrypt_and_check (&rc4_keys[lid], tmp, krb5tgs_bufs[salt_pos].edata2, krb5tgs_bufs[salt_pos].edata2_len, K2, checksum) == 1) { mark_hash (plains_buf, d_return_buf, salt_pos, 0, digests_offset + 0, gid, il_pos); d_return_buf[lid] = 1; } } } __kernel void m13100_m08 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { } __kernel void m13100_m16 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { } __kernel void m13100_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { /** * modifier */ const u32 lid = get_local_id (0); const u32 gid = get_global_id (0); if (gid >= gid_max) return; /** * base */ u32 pw_buf0[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]; pw_buf1[3] = pws[gid].i[ 7]; const u32 pw_len = pws[gid].pw_len; /** * shared */ __local RC4_KEY rc4_keys[64]; /** * salt */ u32 checksum[4]; checksum[0] = krb5tgs_bufs[salt_pos].checksum[0]; checksum[1] = krb5tgs_bufs[salt_pos].checksum[1]; checksum[2] = krb5tgs_bufs[salt_pos].checksum[2]; checksum[3] = krb5tgs_bufs[salt_pos].checksum[3]; /** * loop */ for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { u32x w0[4] = { 0 }; u32x w1[4] = { 0 }; u32x w2[4] = { 0 }; u32x w3[4] = { 0 }; const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1); /** * kerberos */ u32 digest[4]; u32 K2[4]; kerb_prepare (w0, w1, out_len, checksum, digest, K2); u32 tmp[4]; tmp[0] = digest[0]; tmp[1] = digest[1]; tmp[2] = digest[2]; tmp[3] = digest[3]; if (decrypt_and_check (&rc4_keys[lid], tmp, krb5tgs_bufs[salt_pos].edata2, krb5tgs_bufs[salt_pos].edata2_len, K2, checksum) == 1) { mark_hash (plains_buf, d_return_buf, salt_pos, 0, digests_offset + 0, gid, il_pos); d_return_buf[lid] = 1; } } } __kernel void m13100_s08 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { } __kernel void m13100_s16 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global krb5tgs_t *krb5tgs_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { }