1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-12-23 15:18:16 +00:00
hashcat/OpenCL/m24500-pure.cl

655 lines
16 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
#define NEW_SIMD_CODE
#ifdef KERNEL_STATIC
#include M2S(INCLUDE_PATH/inc_vendor.h)
#include M2S(INCLUDE_PATH/inc_types.h)
#include M2S(INCLUDE_PATH/inc_platform.cl)
#include M2S(INCLUDE_PATH/inc_common.cl)
#include M2S(INCLUDE_PATH/inc_simd.cl)
#include M2S(INCLUDE_PATH/inc_hash_sha1.cl)
#include M2S(INCLUDE_PATH/inc_hash_sha512.cl)
#include M2S(INCLUDE_PATH/inc_cipher_aes.cl)
#endif
typedef struct telegram_tmp
{
u64 ipad[8];
u64 opad[8];
u64 dgst[24];
u64 out [24];
} telegram_tmp_t;
typedef struct telegram
{
u32 data[72];
} telegram_t;
DECLSPEC void hmac_sha512_run_V (PRIVATE_AS u32x *w0, PRIVATE_AS u32x *w1, PRIVATE_AS u32x *w2, PRIVATE_AS u32x *w3, PRIVATE_AS u32x *w4, PRIVATE_AS u32x *w5, PRIVATE_AS u32x *w6, PRIVATE_AS u32x *w7, PRIVATE_AS u64x *ipad, PRIVATE_AS u64x *opad, PRIVATE_AS u64x *digest)
{
digest[0] = ipad[0];
digest[1] = ipad[1];
digest[2] = ipad[2];
digest[3] = ipad[3];
digest[4] = ipad[4];
digest[5] = ipad[5];
digest[6] = ipad[6];
digest[7] = ipad[7];
sha512_transform_vector (w0, w1, w2, w3, w4, w5, w6, w7, digest);
w0[0] = h32_from_64 (digest[0]);
w0[1] = l32_from_64 (digest[0]);
w0[2] = h32_from_64 (digest[1]);
w0[3] = l32_from_64 (digest[1]);
w1[0] = h32_from_64 (digest[2]);
w1[1] = l32_from_64 (digest[2]);
w1[2] = h32_from_64 (digest[3]);
w1[3] = l32_from_64 (digest[3]);
w2[0] = h32_from_64 (digest[4]);
w2[1] = l32_from_64 (digest[4]);
w2[2] = h32_from_64 (digest[5]);
w2[3] = l32_from_64 (digest[5]);
w3[0] = h32_from_64 (digest[6]);
w3[1] = l32_from_64 (digest[6]);
w3[2] = h32_from_64 (digest[7]);
w3[3] = l32_from_64 (digest[7]);
w4[0] = 0x80000000;
w4[1] = 0;
w4[2] = 0;
w4[3] = 0;
w5[0] = 0;
w5[1] = 0;
w5[2] = 0;
w5[3] = 0;
w6[0] = 0;
w6[1] = 0;
w6[2] = 0;
w6[3] = 0;
w7[0] = 0;
w7[1] = 0;
w7[2] = 0;
w7[3] = (128 + 64) * 8;
digest[0] = opad[0];
digest[1] = opad[1];
digest[2] = opad[2];
digest[3] = opad[3];
digest[4] = opad[4];
digest[5] = opad[5];
digest[6] = opad[6];
digest[7] = opad[7];
sha512_transform_vector (w0, w1, w2, w3, w4, w5, w6, w7, digest);
}
DECLSPEC void sha1_run (PRIVATE_AS u32 *w, PRIVATE_AS u32 *res)
{
u32 w0[4];
u32 w1[4];
u32 w2[4];
u32 w3[4];
w0[0] = w[ 0];
w0[1] = w[ 1];
w0[2] = w[ 2];
w0[3] = w[ 3];
w1[0] = w[ 4];
w1[1] = w[ 5];
w1[2] = w[ 6];
w1[3] = w[ 7];
w2[0] = w[ 8];
w2[1] = w[ 9];
w2[2] = w[10];
w2[3] = w[11];
w3[0] = 0x80000000;
w3[1] = 0;
w3[2] = 0;
w3[3] = 48 * 8;
u32 digest[5];
digest[0] = SHA1M_A;
digest[1] = SHA1M_B;
digest[2] = SHA1M_C;
digest[3] = SHA1M_D;
digest[4] = SHA1M_E;
sha1_transform (w0, w1, w2, w3, digest);
res[0] = digest[0];
res[1] = digest[1];
res[2] = digest[2];
res[3] = digest[3];
res[4] = digest[4];
}
KERNEL_FQ void m24500_init (KERN_ATTR_TMPS_ESALT (telegram_tmp_t, telegram_t))
{
/**
* base
*/
const u64 gid = get_global_id (0);
if (gid >= GID_CNT) return;
const u32 salt_len = salt_bufs[SALT_POS_HOST].salt_len; // 32
sha512_ctx_t sha512_ctx;
sha512_init (&sha512_ctx);
sha512_update_global (&sha512_ctx, salt_bufs[SALT_POS_HOST].salt_buf, salt_len);
sha512_update_global_swap (&sha512_ctx, pws[gid].i, pws[gid].pw_len);
sha512_update_global (&sha512_ctx, salt_bufs[SALT_POS_HOST].salt_buf, salt_len);
sha512_final (&sha512_ctx);
u32 w[32] = { 0 };
w[ 0] = h32_from_64_S (sha512_ctx.h[0]);
w[ 1] = l32_from_64_S (sha512_ctx.h[0]);
w[ 2] = h32_from_64_S (sha512_ctx.h[1]);
w[ 3] = l32_from_64_S (sha512_ctx.h[1]);
w[ 4] = h32_from_64_S (sha512_ctx.h[2]);
w[ 5] = l32_from_64_S (sha512_ctx.h[2]);
w[ 6] = h32_from_64_S (sha512_ctx.h[3]);
w[ 7] = l32_from_64_S (sha512_ctx.h[3]);
w[ 8] = h32_from_64_S (sha512_ctx.h[4]);
w[ 9] = l32_from_64_S (sha512_ctx.h[4]);
w[10] = h32_from_64_S (sha512_ctx.h[5]);
w[11] = l32_from_64_S (sha512_ctx.h[5]);
w[12] = h32_from_64_S (sha512_ctx.h[6]);
w[13] = l32_from_64_S (sha512_ctx.h[6]);
w[14] = h32_from_64_S (sha512_ctx.h[7]);
w[15] = l32_from_64_S (sha512_ctx.h[7]);
sha512_hmac_ctx_t sha512_hmac_ctx;
sha512_hmac_init (&sha512_hmac_ctx, w, 64);
tmps[gid].ipad[0] = sha512_hmac_ctx.ipad.h[0];
tmps[gid].ipad[1] = sha512_hmac_ctx.ipad.h[1];
tmps[gid].ipad[2] = sha512_hmac_ctx.ipad.h[2];
tmps[gid].ipad[3] = sha512_hmac_ctx.ipad.h[3];
tmps[gid].ipad[4] = sha512_hmac_ctx.ipad.h[4];
tmps[gid].ipad[5] = sha512_hmac_ctx.ipad.h[5];
tmps[gid].ipad[6] = sha512_hmac_ctx.ipad.h[6];
tmps[gid].ipad[7] = sha512_hmac_ctx.ipad.h[7];
tmps[gid].opad[0] = sha512_hmac_ctx.opad.h[0];
tmps[gid].opad[1] = sha512_hmac_ctx.opad.h[1];
tmps[gid].opad[2] = sha512_hmac_ctx.opad.h[2];
tmps[gid].opad[3] = sha512_hmac_ctx.opad.h[3];
tmps[gid].opad[4] = sha512_hmac_ctx.opad.h[4];
tmps[gid].opad[5] = sha512_hmac_ctx.opad.h[5];
tmps[gid].opad[6] = sha512_hmac_ctx.opad.h[6];
tmps[gid].opad[7] = sha512_hmac_ctx.opad.h[7];
sha512_hmac_update_global (&sha512_hmac_ctx, salt_bufs[SALT_POS_HOST].salt_buf, salt_bufs[SALT_POS_HOST].salt_len);
for (u32 i = 0, j = 1; i < 24; i += 8, j += 1)
{
sha512_hmac_ctx_t sha512_hmac_ctx2 = sha512_hmac_ctx;
u32 w0[4];
u32 w1[4];
u32 w2[4];
u32 w3[4];
u32 w4[4];
u32 w5[4];
u32 w6[4];
u32 w7[4];
w0[0] = j;
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;
w4[0] = 0;
w4[1] = 0;
w4[2] = 0;
w4[3] = 0;
w5[0] = 0;
w5[1] = 0;
w5[2] = 0;
w5[3] = 0;
w6[0] = 0;
w6[1] = 0;
w6[2] = 0;
w6[3] = 0;
w7[0] = 0;
w7[1] = 0;
w7[2] = 0;
w7[3] = 0;
sha512_hmac_update_128 (&sha512_hmac_ctx2, w0, w1, w2, w3, w4, w5, w6, w7, 4);
sha512_hmac_final (&sha512_hmac_ctx2);
tmps[gid].dgst[i + 0] = sha512_hmac_ctx2.opad.h[0];
tmps[gid].dgst[i + 1] = sha512_hmac_ctx2.opad.h[1];
tmps[gid].dgst[i + 2] = sha512_hmac_ctx2.opad.h[2];
tmps[gid].dgst[i + 3] = sha512_hmac_ctx2.opad.h[3];
tmps[gid].dgst[i + 4] = sha512_hmac_ctx2.opad.h[4];
tmps[gid].dgst[i + 5] = sha512_hmac_ctx2.opad.h[5];
tmps[gid].dgst[i + 6] = sha512_hmac_ctx2.opad.h[6];
tmps[gid].dgst[i + 7] = sha512_hmac_ctx2.opad.h[7];
tmps[gid].out[i + 0] = tmps[gid].dgst[i + 0];
tmps[gid].out[i + 1] = tmps[gid].dgst[i + 1];
tmps[gid].out[i + 2] = tmps[gid].dgst[i + 2];
tmps[gid].out[i + 3] = tmps[gid].dgst[i + 3];
tmps[gid].out[i + 4] = tmps[gid].dgst[i + 4];
tmps[gid].out[i + 5] = tmps[gid].dgst[i + 5];
tmps[gid].out[i + 6] = tmps[gid].dgst[i + 6];
tmps[gid].out[i + 7] = tmps[gid].dgst[i + 7];
}
}
KERNEL_FQ void m24500_loop (KERN_ATTR_TMPS_ESALT (telegram_tmp_t, telegram_t))
{
const u64 gid = get_global_id (0);
if ((gid * VECT_SIZE) >= GID_CNT) return;
u64x ipad[8];
u64x opad[8];
ipad[0] = pack64v (tmps, ipad, gid, 0);
ipad[1] = pack64v (tmps, ipad, gid, 1);
ipad[2] = pack64v (tmps, ipad, gid, 2);
ipad[3] = pack64v (tmps, ipad, gid, 3);
ipad[4] = pack64v (tmps, ipad, gid, 4);
ipad[5] = pack64v (tmps, ipad, gid, 5);
ipad[6] = pack64v (tmps, ipad, gid, 6);
ipad[7] = pack64v (tmps, ipad, gid, 7);
opad[0] = pack64v (tmps, opad, gid, 0);
opad[1] = pack64v (tmps, opad, gid, 1);
opad[2] = pack64v (tmps, opad, gid, 2);
opad[3] = pack64v (tmps, opad, gid, 3);
opad[4] = pack64v (tmps, opad, gid, 4);
opad[5] = pack64v (tmps, opad, gid, 5);
opad[6] = pack64v (tmps, opad, gid, 6);
opad[7] = pack64v (tmps, opad, gid, 7);
for (u32 i = 0; i < 24; i += 8)
{
u64x dgst[8];
u64x out[8];
dgst[0] = pack64v (tmps, dgst, gid, i + 0);
dgst[1] = pack64v (tmps, dgst, gid, i + 1);
dgst[2] = pack64v (tmps, dgst, gid, i + 2);
dgst[3] = pack64v (tmps, dgst, gid, i + 3);
dgst[4] = pack64v (tmps, dgst, gid, i + 4);
dgst[5] = pack64v (tmps, dgst, gid, i + 5);
dgst[6] = pack64v (tmps, dgst, gid, i + 6);
dgst[7] = pack64v (tmps, dgst, gid, i + 7);
out[0] = pack64v (tmps, out, gid, i + 0);
out[1] = pack64v (tmps, out, gid, i + 1);
out[2] = pack64v (tmps, out, gid, i + 2);
out[3] = pack64v (tmps, out, gid, i + 3);
out[4] = pack64v (tmps, out, gid, i + 4);
out[5] = pack64v (tmps, out, gid, i + 5);
out[6] = pack64v (tmps, out, gid, i + 6);
out[7] = pack64v (tmps, out, gid, i + 7);
for (u32 j = 0; j < LOOP_CNT; j++)
{
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
u32x w4[4];
u32x w5[4];
u32x w6[4];
u32x w7[4];
w0[0] = h32_from_64 (dgst[0]);
w0[1] = l32_from_64 (dgst[0]);
w0[2] = h32_from_64 (dgst[1]);
w0[3] = l32_from_64 (dgst[1]);
w1[0] = h32_from_64 (dgst[2]);
w1[1] = l32_from_64 (dgst[2]);
w1[2] = h32_from_64 (dgst[3]);
w1[3] = l32_from_64 (dgst[3]);
w2[0] = h32_from_64 (dgst[4]);
w2[1] = l32_from_64 (dgst[4]);
w2[2] = h32_from_64 (dgst[5]);
w2[3] = l32_from_64 (dgst[5]);
w3[0] = h32_from_64 (dgst[6]);
w3[1] = l32_from_64 (dgst[6]);
w3[2] = h32_from_64 (dgst[7]);
w3[3] = l32_from_64 (dgst[7]);
w4[0] = 0x80000000;
w4[1] = 0;
w4[2] = 0;
w4[3] = 0;
w5[0] = 0;
w5[1] = 0;
w5[2] = 0;
w5[3] = 0;
w6[0] = 0;
w6[1] = 0;
w6[2] = 0;
w6[3] = 0;
w7[0] = 0;
w7[1] = 0;
w7[2] = 0;
w7[3] = (128 + 64) * 8;
hmac_sha512_run_V (w0, w1, w2, w3, w4, w5, w6, w7, ipad, opad, dgst);
out[0] ^= dgst[0];
out[1] ^= dgst[1];
out[2] ^= dgst[2];
out[3] ^= dgst[3];
out[4] ^= dgst[4];
out[5] ^= dgst[5];
out[6] ^= dgst[6];
out[7] ^= dgst[7];
}
unpack64v (tmps, dgst, gid, i + 0, dgst[0]);
unpack64v (tmps, dgst, gid, i + 1, dgst[1]);
unpack64v (tmps, dgst, gid, i + 2, dgst[2]);
unpack64v (tmps, dgst, gid, i + 3, dgst[3]);
unpack64v (tmps, dgst, gid, i + 4, dgst[4]);
unpack64v (tmps, dgst, gid, i + 5, dgst[5]);
unpack64v (tmps, dgst, gid, i + 6, dgst[6]);
unpack64v (tmps, dgst, gid, i + 7, dgst[7]);
unpack64v (tmps, out, gid, i + 0, out[0]);
unpack64v (tmps, out, gid, i + 1, out[1]);
unpack64v (tmps, out, gid, i + 2, out[2]);
unpack64v (tmps, out, gid, i + 3, out[3]);
unpack64v (tmps, out, gid, i + 4, out[4]);
unpack64v (tmps, out, gid, i + 5, out[5]);
unpack64v (tmps, out, gid, i + 6, out[6]);
unpack64v (tmps, out, gid, i + 7, out[7]);
}
}
KERNEL_FQ void m24500_comp (KERN_ATTR_TMPS_ESALT (telegram_tmp_t, telegram_t))
{
/**
* base
*/
const u64 gid = get_global_id (0);
const u64 lid = get_local_id (0);
const u64 lsz = get_local_size (0);
/**
* aes shared
*/
#ifdef REAL_SHM
LOCAL_VK u32 s_td0[256];
LOCAL_VK u32 s_td1[256];
LOCAL_VK u32 s_td2[256];
LOCAL_VK u32 s_td3[256];
LOCAL_VK u32 s_td4[256];
LOCAL_VK u32 s_te0[256];
LOCAL_VK u32 s_te1[256];
LOCAL_VK u32 s_te2[256];
LOCAL_VK u32 s_te3[256];
LOCAL_VK u32 s_te4[256];
for (u32 i = lid; i < 256; i += lsz)
{
s_td0[i] = td0[i];
s_td1[i] = td1[i];
s_td2[i] = td2[i];
s_td3[i] = td3[i];
s_td4[i] = td4[i];
s_te0[i] = te0[i];
s_te1[i] = te1[i];
s_te2[i] = te2[i];
s_te3[i] = te3[i];
s_te4[i] = te4[i];
}
SYNC_THREADS ();
#else
CONSTANT_AS u32a *s_td0 = td0;
CONSTANT_AS u32a *s_td1 = td1;
CONSTANT_AS u32a *s_td2 = td2;
CONSTANT_AS u32a *s_td3 = td3;
CONSTANT_AS u32a *s_td4 = td4;
CONSTANT_AS u32a *s_te0 = te0;
CONSTANT_AS u32a *s_te1 = te1;
CONSTANT_AS u32a *s_te2 = te2;
CONSTANT_AS u32a *s_te3 = te3;
CONSTANT_AS u32a *s_te4 = te4;
#endif
if (gid >= GID_CNT) return;
u32 data_key[4];
data_key[0] = esalt_bufs[DIGESTS_OFFSET_HOST].data[0];
data_key[1] = esalt_bufs[DIGESTS_OFFSET_HOST].data[1];
data_key[2] = esalt_bufs[DIGESTS_OFFSET_HOST].data[2];
data_key[3] = esalt_bufs[DIGESTS_OFFSET_HOST].data[3];
u32 data_a[12];
u32 data_b[12];
u32 data_c[12];
u32 data_d[12];
data_a[ 0] = data_key[0];
data_a[ 1] = data_key[1];
data_a[ 2] = data_key[2];
data_a[ 3] = data_key[3];
data_b[ 4] = data_key[0];
data_b[ 5] = data_key[1];
data_b[ 6] = data_key[2];
data_b[ 7] = data_key[3];
data_c[ 8] = data_key[0];
data_c[ 9] = data_key[1];
data_c[10] = data_key[2];
data_c[11] = data_key[3];
data_d[ 0] = data_key[0];
data_d[ 1] = data_key[1];
data_d[ 2] = data_key[2];
data_d[ 3] = data_key[3];
data_a[ 4] = h32_from_64_S (tmps[gid].out[ 1]); // not a bug: out[0] is ignored
data_a[ 5] = l32_from_64_S (tmps[gid].out[ 1]);
data_a[ 6] = h32_from_64_S (tmps[gid].out[ 2]);
data_a[ 7] = l32_from_64_S (tmps[gid].out[ 2]);
data_a[ 8] = h32_from_64_S (tmps[gid].out[ 3]);
data_a[ 9] = l32_from_64_S (tmps[gid].out[ 3]);
data_a[10] = h32_from_64_S (tmps[gid].out[ 4]);
data_a[11] = l32_from_64_S (tmps[gid].out[ 4]);
data_b[ 0] = h32_from_64_S (tmps[gid].out[ 5]);
data_b[ 1] = l32_from_64_S (tmps[gid].out[ 5]);
data_b[ 2] = h32_from_64_S (tmps[gid].out[ 6]);
data_b[ 3] = l32_from_64_S (tmps[gid].out[ 6]);
data_b[ 8] = h32_from_64_S (tmps[gid].out[ 7]);
data_b[ 9] = l32_from_64_S (tmps[gid].out[ 7]);
data_b[10] = h32_from_64_S (tmps[gid].out[ 8]);
data_b[11] = l32_from_64_S (tmps[gid].out[ 8]);
data_c[ 0] = h32_from_64_S (tmps[gid].out[ 9]);
data_c[ 1] = l32_from_64_S (tmps[gid].out[ 9]);
data_c[ 2] = h32_from_64_S (tmps[gid].out[10]);
data_c[ 3] = l32_from_64_S (tmps[gid].out[10]);
data_c[ 4] = h32_from_64_S (tmps[gid].out[11]);
data_c[ 5] = l32_from_64_S (tmps[gid].out[11]);
data_c[ 6] = h32_from_64_S (tmps[gid].out[12]);
data_c[ 7] = l32_from_64_S (tmps[gid].out[12]);
data_d[ 4] = h32_from_64_S (tmps[gid].out[13]);
data_d[ 5] = l32_from_64_S (tmps[gid].out[13]);
data_d[ 6] = h32_from_64_S (tmps[gid].out[14]);
data_d[ 7] = l32_from_64_S (tmps[gid].out[14]);
data_d[ 8] = h32_from_64_S (tmps[gid].out[15]);
data_d[ 9] = l32_from_64_S (tmps[gid].out[15]);
data_d[10] = h32_from_64_S (tmps[gid].out[16]);
data_d[11] = l32_from_64_S (tmps[gid].out[16]);
// hash (SHA1 ()) the data_*:
u32 a[5];
sha1_run (data_a, a);
u32 b[5];
sha1_run (data_b, b);
u32 c[5];
sha1_run (data_c, c);
u32 d[5];
sha1_run (data_d, d);
// set up AES key and AES IV:
u32 key[8];
key[0] = a[0];
key[1] = a[1];
key[2] = b[2];
key[3] = b[3];
key[4] = b[4];
key[5] = c[1];
key[6] = c[2];
key[7] = c[3];
u32 iv[8];
iv[0] = a[2];
iv[1] = a[3];
iv[2] = a[4];
iv[3] = b[0];
iv[4] = b[1];
iv[5] = c[4];
iv[6] = d[0];
iv[7] = d[1];
// decrypt with AES-IGE:
#define KEYLEN 60
u32 ks[KEYLEN];
AES256_set_decrypt_key (ks, key, s_te0, s_te1, s_te2, s_te3, s_td0, s_td1, s_td2, s_td3);
u32 x_prev[4];
x_prev[0] = iv[0];
x_prev[1] = iv[1];
x_prev[2] = iv[2];
x_prev[3] = iv[3];
u32 y_prev[4];
y_prev[0] = iv[4];
y_prev[1] = iv[5];
y_prev[2] = iv[6];
y_prev[3] = iv[7];
u32 out[80] = { 0 }; // 64-byte aligned for SHA1
for (int i = 0; i < 68; i += 4)
{
u32 x[4];
x[0] = esalt_bufs[DIGESTS_OFFSET_HOST].data[4 + i];
x[1] = esalt_bufs[DIGESTS_OFFSET_HOST].data[5 + i];
x[2] = esalt_bufs[DIGESTS_OFFSET_HOST].data[6 + i];
x[3] = esalt_bufs[DIGESTS_OFFSET_HOST].data[7 + i];
u32 y[4];
y[0] = x[0] ^ y_prev[0];
y[1] = x[1] ^ y_prev[1];
y[2] = x[2] ^ y_prev[2];
y[3] = x[3] ^ y_prev[3];
u32 dec[4];
AES256_decrypt (ks, y, dec, s_td0, s_td1, s_td2, s_td3, s_td4);
y_prev[0] = dec[0] ^ x_prev[0];
y_prev[1] = dec[1] ^ x_prev[1];
y_prev[2] = dec[2] ^ x_prev[2];
y_prev[3] = dec[3] ^ x_prev[3];
out[i + 0] = y_prev[0];
out[i + 1] = y_prev[1];
out[i + 2] = y_prev[2];
out[i + 3] = y_prev[3];
x_prev[0] = x[0];
x_prev[1] = x[1];
x_prev[2] = x[2];
x_prev[3] = x[3];
}
// final SHA1 checksum of the decrypted data (out):
sha1_ctx_t ctx;
sha1_init (&ctx);
sha1_update (&ctx, out, 272);
sha1_final (&ctx);
const u32 r0 = ctx.h[0];
const u32 r1 = ctx.h[1];
const u32 r2 = ctx.h[2];
const u32 r3 = ctx.h[3];
// verify:
if (r0 == data_key[0] &&
r1 == data_key[1] &&
r2 == data_key[2] &&
r3 == data_key[3])
{
if (hc_atomic_inc (&hashes_shown[DIGESTS_OFFSET_HOST]) == 0)
{
mark_hash (plains_buf, d_return_buf, SALT_POS_HOST, DIGESTS_CNT, 0, DIGESTS_OFFSET_HOST + 0, gid, 0, 0, 0);
}
}
}