1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-14 03:39:09 +00:00
hashcat/OpenCL/m00140_a3-pure.cl

150 lines
2.7 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)
#endif
KERNEL_FQ void m00140_mxx (KERN_ATTR_VECTOR ())
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
const u64 gid = get_global_id (0);
if (gid >= GID_CNT) return;
/**
* base
*/
const u32 pw_len = pws[gid].pw_len;
u32x w[64] = { 0 };
for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
{
w[idx] = pws[gid].i[idx];
}
sha1_ctx_t ctx0;
sha1_init (&ctx0);
sha1_update_global_swap (&ctx0, salt_bufs[SALT_POS_HOST].salt_buf, salt_bufs[SALT_POS_HOST].salt_len);
/**
* loop
*/
u32x w0l = w[0];
for (u32 il_pos = 0; il_pos < IL_CNT; il_pos += VECT_SIZE)
{
const u32x w0r = words_buf_r[il_pos / VECT_SIZE];
const u32x w0 = w0l | w0r;
w[0] = w0;
sha1_ctx_vector_t ctx;
sha1_init_vector_from_scalar (&ctx, &ctx0);
sha1_update_vector_utf16beN (&ctx, w, pw_len);
sha1_final_vector (&ctx);
const u32x r0 = ctx.h[DGST_R0];
const u32x r1 = ctx.h[DGST_R1];
const u32x r2 = ctx.h[DGST_R2];
const u32x r3 = ctx.h[DGST_R3];
COMPARE_M_SIMD (r0, r1, r2, r3);
}
}
KERNEL_FQ void m00140_sxx (KERN_ATTR_VECTOR ())
{
/**
* modifier
*/
const u64 lid = get_local_id (0);
const u64 gid = get_global_id (0);
if (gid >= GID_CNT) return;
/**
* digest
*/
const u32 search[4] =
{
digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R0],
digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R1],
digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R2],
digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R3]
};
/**
* base
*/
const u32 pw_len = pws[gid].pw_len;
u32x w[64] = { 0 };
for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
{
w[idx] = pws[gid].i[idx];
}
sha1_ctx_t ctx0;
sha1_init (&ctx0);
sha1_update_global_swap (&ctx0, salt_bufs[SALT_POS_HOST].salt_buf, salt_bufs[SALT_POS_HOST].salt_len);
/**
* loop
*/
u32x w0l = w[0];
for (u32 il_pos = 0; il_pos < IL_CNT; il_pos += VECT_SIZE)
{
const u32x w0r = words_buf_r[il_pos / VECT_SIZE];
const u32x w0 = w0l | w0r;
w[0] = w0;
sha1_ctx_vector_t ctx;
sha1_init_vector_from_scalar (&ctx, &ctx0);
sha1_update_vector_utf16beN (&ctx, w, pw_len);
sha1_final_vector (&ctx);
const u32x r0 = ctx.h[DGST_R0];
const u32x r1 = ctx.h[DGST_R1];
const u32x r2 = ctx.h[DGST_R2];
const u32x r3 = ctx.h[DGST_R3];
COMPARE_S_SIMD (r0, r1, r2, r3);
}
}