You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
hashcat/OpenCL/m10410_a0-optimized.cl

316 lines
5.3 KiB

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
//too much register pressure
//#define NEW_SIMD_CODE
#ifdef KERNEL_STATIC
#include "inc_vendor.h"
#include "inc_types.h"
#include "inc_platform.cl"
#include "inc_common.cl"
#include "inc_rp_optimized.h"
#include "inc_rp_optimized.cl"
#include "inc_simd.cl"
#include "inc_hash_md5.cl"
#endif
CONSTANT_VK u32a padding[8] =
{
0x5e4ebf28,
0x418a754e,
0x564e0064,
0x0801faff,
0xb6002e2e,
0x803e68d0,
0xfea90c2f,
0x7a695364
};
typedef struct pdf
{
int V;
int R;
int P;
int enc_md;
u32 id_buf[8];
u32 u_buf[32];
u32 o_buf[32];
int id_len;
int o_len;
int u_len;
u32 rc4key[2];
u32 rc4data[2];
} pdf_t;
typedef struct
{
u8 S[256];
u32 wtf_its_faster;
} RC4_KEY;
DECLSPEC void swap (LOCAL_AS 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;
}
DECLSPEC void rc4_init_16 (LOCAL_AS RC4_KEY *rc4_key, const u32 *data)
{
u32 v = 0x03020100;
u32 a = 0x04040404;
LOCAL_AS u32 *ptr = (LOCAL_AS u32 *) rc4_key->S;
#ifdef _unroll
#pragma unroll
#endif
for (u32 i = 0; i < 64; i++)
{
ptr[i] = v; v += a;
}
const u32 d0 = data[0] >> 0;
const u32 d1 = data[0] >> 8;
const u32 d2 = data[0] >> 16;
const u32 d3 = data[0] >> 24;
const u32 d4 = data[1] >> 0;
u32 j = 0;
#ifdef _unroll
#pragma unroll
#endif
for (u32 i = 0; i < 255; i += 5)
{
j += rc4_key->S[i + 0] + d0; swap (rc4_key, i + 0, j);
j += rc4_key->S[i + 1] + d1; swap (rc4_key, i + 1, j);
j += rc4_key->S[i + 2] + d2; swap (rc4_key, i + 2, j);
j += rc4_key->S[i + 3] + d3; swap (rc4_key, i + 3, j);
j += rc4_key->S[i + 4] + d4; swap (rc4_key, i + 4, j);
}
j += rc4_key->S[255] + d0; swap (rc4_key, 255, j);
}
DECLSPEC u8 rc4_next_16 (LOCAL_AS RC4_KEY *rc4_key, u8 i, u8 j, CONSTANT_AS u32a *in, u32 *out)
{
#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;
}
KERNEL_FQ void m10410_m04 (KERN_ATTR_RULES_ESALT (pdf_t))
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
/**
* base
*/
const u64 gid = get_global_id (0);
if (gid >= gid_max) return;
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];
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 & 63;
/**
* shared
*/
LOCAL_VK RC4_KEY rc4_keys[64];
LOCAL_AS RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* 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 };
apply_rules_vect_optimized (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
/**
* pdf
*/
rc4_init_16 (rc4_key, w0);
u32 out[4];
rc4_next_16 (rc4_key, 0, 0, padding, out);
COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
}
}
KERNEL_FQ void m10410_m08 (KERN_ATTR_RULES_ESALT (pdf_t))
{
}
KERNEL_FQ void m10410_m16 (KERN_ATTR_RULES_ESALT (pdf_t))
{
}
KERNEL_FQ void m10410_s04 (KERN_ATTR_RULES_ESALT (pdf_t))
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
/**
* base
*/
const u64 gid = get_global_id (0);
if (gid >= gid_max) return;
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];
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 & 63;
/**
* shared
*/
LOCAL_VK RC4_KEY rc4_keys[64];
LOCAL_AS RC4_KEY *rc4_key = &rc4_keys[lid];
/**
* 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]
};
/**
* 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 };
apply_rules_vect_optimized (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
/**
* pdf
*/
rc4_init_16 (rc4_key, w0);
u32 out[4];
rc4_next_16 (rc4_key, 0, 0, padding, out);
COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
}
}
KERNEL_FQ void m10410_s08 (KERN_ATTR_RULES_ESALT (pdf_t))
{
}
KERNEL_FQ void m10410_s16 (KERN_ATTR_RULES_ESALT (pdf_t))
{
}