1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-26 01:50:10 +00:00

Backport -m 500 pure kernel code to -m 6300 enables using it on AMDGPU driver

This commit is contained in:
jsteube 2019-04-04 21:05:07 +02:00
parent cd88410781
commit 875718fb8e
2 changed files with 29 additions and 148 deletions

View File

@ -12,9 +12,6 @@
#include "inc_hash_md5.cl"
#endif
#define PUTCHAR_LE(a,p,c) ((u8 *)(a))[(p)] = (u8) (c)
#define GETCHAR_LE(a,p) ((u8 *)(a))[(p)]
#define COMPARE_S "inc_comp_single.cl"
#define COMPARE_M "inc_comp_multi.cl"
@ -160,152 +157,50 @@ KERNEL_FQ void m06300_loop (KERN_ATTR_TMPS (md5crypt_tmp_t))
* digest
*/
u32 digest[4];
u32 digest[16] = { 0 };
digest[0] = tmps[gid].digest_buf[0];
digest[1] = tmps[gid].digest_buf[1];
digest[2] = tmps[gid].digest_buf[2];
digest[3] = tmps[gid].digest_buf[3];
u32 wpc_len[8];
wpc_len[0] = 16 + 0 + 0 + pw_len;
wpc_len[1] = pw_len + 0 + 0 + 16;
wpc_len[2] = 16 + salt_len + 0 + pw_len;
wpc_len[3] = pw_len + salt_len + 0 + 16;
wpc_len[4] = 16 + 0 + pw_len + pw_len;
wpc_len[5] = pw_len + 0 + pw_len + 16;
wpc_len[6] = 16 + salt_len + pw_len + pw_len;
wpc_len[7] = pw_len + salt_len + pw_len + 16;
// largest possible wpc_len[7] is not enough because of zero buffer loop
u32 wpc[8][64 + 64 + 64 + 64];
#ifdef _unroll
#pragma unroll
#endif
for (u32 i = 0; i < 8; i++)
{
u32 block_len = 0;
if (i & 1)
{
for (u32 j = 0; j < pw_len; j++)
{
PUTCHAR_LE (wpc[i], block_len++, GETCHAR_LE (w, j));
}
}
else
{
block_len += 16;
}
if (i & 2)
{
for (u32 j = 0; j < salt_len; j++)
{
PUTCHAR_LE (wpc[i], block_len++, GETCHAR_LE (s, j));
}
}
if (i & 4)
{
for (u32 j = 0; j < pw_len; j++)
{
PUTCHAR_LE (wpc[i], block_len++, GETCHAR_LE (w, j));
}
}
if (i & 1)
{
block_len += 16;
}
else
{
for (u32 j = 0; j < pw_len; j++)
{
PUTCHAR_LE (wpc[i], block_len++, GETCHAR_LE (w, j));
}
}
}
#ifdef _unroll
#pragma unroll
#endif
for (u32 i = 0; i < 8; i++)
{
u32 *z = wpc[i] + ((wpc_len[i] / 64) * 16);
truncate_block_16x4_le_S (z + 0, z + 4, z + 8, z + 12, wpc_len[i] & 63);
}
/**
* loop
*/
for (u32 i = 0, j = loop_pos; i < loop_cnt; i++, j++)
{
const u32 j1 = (j & 1) ? 1 : 0;
const u32 j3 = (j % 3) ? 2 : 0;
const u32 j7 = (j % 7) ? 4 : 0;
const u32 pc = j1 + j3 + j7;
if (j1)
{
const u32 off = wpc_len[pc] / 4;
const u32 mod = wpc_len[pc] % 4;
u32 *ptr = wpc[pc] + off - 4;
switch (mod)
{
case 0:
ptr[0] = digest[0];
ptr[1] = digest[1];
ptr[2] = digest[2];
ptr[3] = digest[3];
break;
case 1:
ptr[0] = (ptr[0] & 0xff) | (digest[0] << 8);
ptr[1] = (digest[0] >> 24) | (digest[1] << 8);
ptr[2] = (digest[1] >> 24) | (digest[2] << 8);
ptr[3] = (digest[2] >> 24) | (digest[3] << 8);
ptr[4] = (digest[3] >> 24);
break;
case 2:
ptr[0] = (ptr[0] & 0xffff) | (digest[0] << 16);
ptr[1] = (digest[0] >> 16) | (digest[1] << 16);
ptr[2] = (digest[1] >> 16) | (digest[2] << 16);
ptr[3] = (digest[2] >> 16) | (digest[3] << 16);
ptr[4] = (digest[3] >> 16);
break;
case 3:
ptr[0] = (ptr[0] & 0xffffff) | (digest[0] << 24);
ptr[1] = (digest[0] >> 8) | (digest[1] << 24);
ptr[2] = (digest[1] >> 8) | (digest[2] << 24);
ptr[3] = (digest[2] >> 8) | (digest[3] << 24);
ptr[4] = (digest[3] >> 8);
break;
}
}
else
{
wpc[pc][0] = digest[0];
wpc[pc][1] = digest[1];
wpc[pc][2] = digest[2];
wpc[pc][3] = digest[3];
}
md5_ctx_t md5_ctx;
md5_init (&md5_ctx);
md5_update (&md5_ctx, wpc[pc], wpc_len[pc]);
if (j & 1)
{
md5_update (&md5_ctx, w, pw_len);
}
else
{
md5_update (&md5_ctx, digest, 16);
}
if (j % 3)
{
md5_update (&md5_ctx, s, salt_len);
}
if (j % 7)
{
md5_update (&md5_ctx, w, pw_len);
}
if (j & 1)
{
md5_update (&md5_ctx, digest, 16);
}
else
{
md5_update (&md5_ctx, w, pw_len);
}
md5_final (&md5_ctx);

View File

@ -166,20 +166,6 @@ u32 module_pw_max (MAYBE_UNUSED const hashconfig_t *hashconfig, MAYBE_UNUSED con
return pw_max;
}
bool module_unstable_warning (MAYBE_UNUSED const hashconfig_t *hashconfig, MAYBE_UNUSED const user_options_t *user_options, MAYBE_UNUSED const user_options_extra_t *user_options_extra, MAYBE_UNUSED const hc_device_param_t *device_param)
{
// amdgpu-pro-18.50-708488-ubuntu-18.04: Segmentation fault
if ((device_param->device_vendor_id == VENDOR_ID_AMD) && (device_param->has_vperm == false))
{
if ((hashconfig->opti_type & OPTI_TYPE_OPTIMIZED_KERNEL) == 0)
{
return true;
}
}
return false;
}
int module_hash_decode (MAYBE_UNUSED const hashconfig_t *hashconfig, MAYBE_UNUSED void *digest_buf, MAYBE_UNUSED salt_t *salt, MAYBE_UNUSED void *esalt_buf, MAYBE_UNUSED void *hook_salt_buf, MAYBE_UNUSED hashinfo_t *hash_info, const char *line_buf, MAYBE_UNUSED const int line_len)
{
u32 *digest = (u32 *) digest_buf;
@ -314,6 +300,6 @@ void module_init (module_ctx_t *module_ctx)
module_ctx->module_st_hash = module_st_hash;
module_ctx->module_st_pass = module_st_pass;
module_ctx->module_tmp_size = module_tmp_size;
module_ctx->module_unstable_warning = module_unstable_warning;
module_ctx->module_unstable_warning = MODULE_DEFAULT;
module_ctx->module_warmup_disable = MODULE_DEFAULT;
}