1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-12-23 07:08:19 +00:00
hashcat/OpenCL/m12700-pure.cl
R. Yushaev 5de004103a Replace kernel parameter lists with macros
Substitute long parameter lists in ~2900 kernel function declarations
with macros. This cleans up the code, reduces probability of copy-paste
errors and highlights the differences between kernel functions. Also
reduces the size of the OpenCL folder by ~3 MB.
2018-11-16 11:44:33 +01:00

354 lines
8.0 KiB
Common Lisp

/**
* Author......: See docs/credits.txt
* License.....: MIT
*/
#define NEW_SIMD_CODE
#include "inc_vendor.cl"
#include "inc_hash_constants.h"
#include "inc_hash_functions.cl"
#include "inc_types.cl"
#include "inc_common.cl"
#include "inc_simd.cl"
#include "inc_hash_sha1.cl"
#include "inc_cipher_aes.cl"
#define COMPARE_S "inc_comp_single.cl"
#define COMPARE_M "inc_comp_multi.cl"
DECLSPEC void hmac_sha1_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];
digest[4] = ipad[4];
sha1_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] = 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_vector (w0, w1, w2, w3, digest);
}
__kernel void m12700_init (KERN_ATTR_TMPS (mywallet_tmp_t))
{
/**
* base
*/
const u64 gid = get_global_id (0);
if (gid >= gid_max) return;
sha1_hmac_ctx_t sha1_hmac_ctx;
sha1_hmac_init_global_swap (&sha1_hmac_ctx, pws[gid].i, pws[gid].pw_len);
tmps[gid].ipad[0] = sha1_hmac_ctx.ipad.h[0];
tmps[gid].ipad[1] = sha1_hmac_ctx.ipad.h[1];
tmps[gid].ipad[2] = sha1_hmac_ctx.ipad.h[2];
tmps[gid].ipad[3] = sha1_hmac_ctx.ipad.h[3];
tmps[gid].ipad[4] = sha1_hmac_ctx.ipad.h[4];
tmps[gid].opad[0] = sha1_hmac_ctx.opad.h[0];
tmps[gid].opad[1] = sha1_hmac_ctx.opad.h[1];
tmps[gid].opad[2] = sha1_hmac_ctx.opad.h[2];
tmps[gid].opad[3] = sha1_hmac_ctx.opad.h[3];
tmps[gid].opad[4] = sha1_hmac_ctx.opad.h[4];
u32 w0[4];
u32 w1[4];
u32 w2[4];
u32 w3[4];
w0[0] = salt_bufs[salt_pos].salt_buf[0];
w0[1] = salt_bufs[salt_pos].salt_buf[1];
w0[2] = salt_bufs[salt_pos].salt_buf[2];
w0[3] = salt_bufs[salt_pos].salt_buf[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;
sha1_hmac_update_64 (&sha1_hmac_ctx, w0, w1, w2, w3, 16);
for (u32 i = 0, j = 1; i < 8; i += 5, j += 1)
{
sha1_hmac_ctx_t sha1_hmac_ctx2 = sha1_hmac_ctx;
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;
sha1_hmac_update_64 (&sha1_hmac_ctx2, w0, w1, w2, w3, 4);
sha1_hmac_final (&sha1_hmac_ctx2);
tmps[gid].dgst[i + 0] = sha1_hmac_ctx2.opad.h[0];
tmps[gid].dgst[i + 1] = sha1_hmac_ctx2.opad.h[1];
tmps[gid].dgst[i + 2] = sha1_hmac_ctx2.opad.h[2];
tmps[gid].dgst[i + 3] = sha1_hmac_ctx2.opad.h[3];
tmps[gid].dgst[i + 4] = sha1_hmac_ctx2.opad.h[4];
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];
}
}
__kernel void m12700_loop (KERN_ATTR_TMPS (mywallet_tmp_t))
{
const u64 gid = get_global_id (0);
if ((gid * VECT_SIZE) >= gid_max) return;
u32x ipad[5];
u32x opad[5];
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);
ipad[4] = packv (tmps, ipad, gid, 4);
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);
opad[4] = packv (tmps, opad, gid, 4);
for (u32 i = 0; i < 8; i += 5)
{
u32x dgst[5];
u32x out[5];
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);
dgst[4] = packv (tmps, dgst, gid, i + 4);
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);
out[4] = packv (tmps, out, gid, i + 4);
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] = 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_V (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];
}
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, dgst, gid, i + 4, dgst[4]);
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]);
unpackv (tmps, out, gid, i + 4, out[4]);
}
}
__kernel void m12700_comp (KERN_ATTR_TMPS (mywallet_tmp_t))
{
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 u32 s_td0[256];
__local u32 s_td1[256];
__local u32 s_td2[256];
__local u32 s_td3[256];
__local u32 s_td4[256];
__local u32 s_te0[256];
__local u32 s_te1[256];
__local u32 s_te2[256];
__local u32 s_te3[256];
__local u32 s_te4[256];
for (MAYBE_VOLATILE 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];
}
barrier (CLK_LOCAL_MEM_FENCE);
#else
__constant u32a *s_td0 = td0;
__constant u32a *s_td1 = td1;
__constant u32a *s_td2 = td2;
__constant u32a *s_td3 = td3;
__constant u32a *s_td4 = td4;
__constant u32a *s_te0 = te0;
__constant u32a *s_te1 = te1;
__constant u32a *s_te2 = te2;
__constant u32a *s_te3 = te3;
__constant u32a *s_te4 = te4;
#endif
if (gid >= gid_max) return;
/**
* AES part
*/
u32 ukey[8];
ukey[0] = tmps[gid].out[0];
ukey[1] = tmps[gid].out[1];
ukey[2] = tmps[gid].out[2];
ukey[3] = tmps[gid].out[3];
ukey[4] = tmps[gid].out[4];
ukey[5] = tmps[gid].out[5];
ukey[6] = tmps[gid].out[6];
ukey[7] = tmps[gid].out[7];
#define KEYLEN 60
u32 ks[KEYLEN];
AES256_set_decrypt_key (ks, ukey, s_te0, s_te1, s_te2, s_te3, s_te4, s_td0, s_td1, s_td2, s_td3, s_td4);
u32 data[4];
data[0] = salt_bufs[salt_pos].salt_buf[4];
data[1] = salt_bufs[salt_pos].salt_buf[5];
data[2] = salt_bufs[salt_pos].salt_buf[6];
data[3] = salt_bufs[salt_pos].salt_buf[7];
u32 out[4];
AES256_decrypt (ks, data, out, s_td0, s_td1, s_td2, s_td3, s_td4);
out[0] ^= salt_bufs[salt_pos].salt_buf[0];
out[1] ^= salt_bufs[salt_pos].salt_buf[1];
out[2] ^= salt_bufs[salt_pos].salt_buf[2];
out[3] ^= salt_bufs[salt_pos].salt_buf[3];
out[0] = swap32_S (out[0]);
out[1] = swap32_S (out[1]);
out[2] = swap32_S (out[2]);
out[3] = swap32_S (out[3]);
if ((out[0] & 0xff) != '{') return;
char *pt = (char *) out;
for (int i = 1; i < 16 - 6; i++)
{
if (pt[i + 0] != '"') continue;
if (pt[i + 1] != 'g') continue;
if (pt[i + 2] != 'u') continue;
if (pt[i + 3] != 'i') continue;
if (pt[i + 4] != 'd') continue;
if (pt[i + 5] != '"') continue;
const u32 r0 = data[0];
const u32 r1 = data[1];
const u32 r2 = data[2];
const u32 r3 = data[3];
#define il_pos 0
#include COMPARE_M
}
}