1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-14 03:39:09 +00:00
hashcat/OpenCL/m11900-pure.cl
Jens Steube 04d5e5a119 New Attack-Mode: Association Attack. Like JtR's single mode. Very early
stage. See hashcat Forum for detailed writeup.
2020-09-29 15:56:32 +02:00

239 lines
5.1 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
#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_simd.cl"
#include "inc_hash_md5.cl"
#endif
#define COMPARE_S "inc_comp_single.cl"
#define COMPARE_M "inc_comp_multi.cl"
typedef struct pbkdf2_md5
{
u32 salt_buf[64];
} pbkdf2_md5_t;
typedef struct pbkdf2_md5_tmp
{
u32 ipad[4];
u32 opad[4];
u32 dgst[32];
u32 out[32];
} pbkdf2_md5_tmp_t;
DECLSPEC void hmac_md5_run_V (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad, u32x *digest)
{
digest[0] = ipad[0];
digest[1] = ipad[1];
digest[2] = ipad[2];
digest[3] = ipad[3];
md5_transform_vector (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_vector (w0, w1, w2, w3, digest);
}
KERNEL_FQ void m11900_init (KERN_ATTR_TMPS_ESALT (pbkdf2_md5_tmp_t, pbkdf2_md5_t))
{
/**
* base
*/
const u64 gid = get_global_id (0);
if (gid >= gid_max) return;
md5_hmac_ctx_t md5_hmac_ctx;
md5_hmac_init_global (&md5_hmac_ctx, pws[gid].i, pws[gid].pw_len);
tmps[gid].ipad[0] = md5_hmac_ctx.ipad.h[0];
tmps[gid].ipad[1] = md5_hmac_ctx.ipad.h[1];
tmps[gid].ipad[2] = md5_hmac_ctx.ipad.h[2];
tmps[gid].ipad[3] = md5_hmac_ctx.ipad.h[3];
tmps[gid].opad[0] = md5_hmac_ctx.opad.h[0];
tmps[gid].opad[1] = md5_hmac_ctx.opad.h[1];
tmps[gid].opad[2] = md5_hmac_ctx.opad.h[2];
tmps[gid].opad[3] = md5_hmac_ctx.opad.h[3];
md5_hmac_update_global (&md5_hmac_ctx, esalt_bufs[DIGESTS_OFFSET].salt_buf, salt_bufs[SALT_POS].salt_len);
for (u32 i = 0, j = 1; i < 4; i += 4, j += 1)
{
md5_hmac_ctx_t md5_hmac_ctx2 = md5_hmac_ctx;
u32 w0[4];
u32 w1[4];
u32 w2[4];
u32 w3[4];
w0[0] = j << 24;
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;
md5_hmac_update_64 (&md5_hmac_ctx2, w0, w1, w2, w3, 4);
md5_hmac_final (&md5_hmac_ctx2);
tmps[gid].dgst[i + 0] = md5_hmac_ctx2.opad.h[0];
tmps[gid].dgst[i + 1] = md5_hmac_ctx2.opad.h[1];
tmps[gid].dgst[i + 2] = md5_hmac_ctx2.opad.h[2];
tmps[gid].dgst[i + 3] = md5_hmac_ctx2.opad.h[3];
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];
}
}
KERNEL_FQ void m11900_loop (KERN_ATTR_TMPS_ESALT (pbkdf2_md5_tmp_t, pbkdf2_md5_t))
{
const u64 gid = get_global_id (0);
if ((gid * VECT_SIZE) >= gid_max) return;
u32x ipad[4];
u32x opad[4];
ipad[0] = packv (tmps, ipad, gid, 0);
ipad[1] = packv (tmps, ipad, gid, 1);
ipad[2] = packv (tmps, ipad, gid, 2);
ipad[3] = packv (tmps, ipad, gid, 3);
opad[0] = packv (tmps, opad, gid, 0);
opad[1] = packv (tmps, opad, gid, 1);
opad[2] = packv (tmps, opad, gid, 2);
opad[3] = packv (tmps, opad, gid, 3);
for (u32 i = 0; i < 4; i += 4)
{
u32x dgst[4];
u32x out[4];
dgst[0] = packv (tmps, dgst, gid, i + 0);
dgst[1] = packv (tmps, dgst, gid, i + 1);
dgst[2] = packv (tmps, dgst, gid, i + 2);
dgst[3] = packv (tmps, dgst, gid, i + 3);
out[0] = packv (tmps, out, gid, i + 0);
out[1] = packv (tmps, out, gid, i + 1);
out[2] = packv (tmps, out, gid, i + 2);
out[3] = packv (tmps, out, gid, i + 3);
for (u32 j = 0; j < loop_cnt; j++)
{
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
w0[0] = dgst[0];
w0[1] = dgst[1];
w0[2] = dgst[2];
w0[3] = dgst[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;
hmac_md5_run_V (w0, w1, w2, w3, ipad, opad, dgst);
out[0] ^= dgst[0];
out[1] ^= dgst[1];
out[2] ^= dgst[2];
out[3] ^= dgst[3];
}
unpackv (tmps, dgst, gid, i + 0, dgst[0]);
unpackv (tmps, dgst, gid, i + 1, dgst[1]);
unpackv (tmps, dgst, gid, i + 2, dgst[2]);
unpackv (tmps, dgst, gid, i + 3, dgst[3]);
unpackv (tmps, out, gid, i + 0, out[0]);
unpackv (tmps, out, gid, i + 1, out[1]);
unpackv (tmps, out, gid, i + 2, out[2]);
unpackv (tmps, out, gid, i + 3, out[3]);
}
}
KERNEL_FQ void m11900_comp (KERN_ATTR_TMPS_ESALT (pbkdf2_md5_tmp_t, pbkdf2_md5_t))
{
/**
* base
*/
const u64 gid = get_global_id (0);
if (gid >= gid_max) return;
const u64 lid = get_local_id (0);
const u32 r0 = tmps[gid].out[DGST_R0];
const u32 r1 = tmps[gid].out[DGST_R1];
const u32 r2 = tmps[gid].out[DGST_R2];
const u32 r3 = tmps[gid].out[DGST_R3];
#define il_pos 0
#ifdef KERNEL_STATIC
#include COMPARE_M
#endif
}