Added full AES-GCM cipher & hash-mode 27000 - Stargazer Stellar Wallet XLM
This commit is contained in:
Gabriele Gristina
@@ -0,0 +1,261 @@
|
||||
/**
|
||||
* Author......: See docs/credits.txt
|
||||
* License.....: MIT
|
||||
*/
|
||||
|
||||
#include "inc_vendor.h"
|
||||
#include "inc_types.h"
|
||||
#include "inc_platform.h"
|
||||
#include "inc_common.h"
|
||||
#include "inc_cipher_aes.h"
|
||||
#include "inc_cipher_aes-gcm.h"
|
||||
|
||||
DECLSPEC void AES_GCM_shift_right_block(uchar *block)
|
||||
{
|
||||
u32 val;
|
||||
|
||||
uchar16 *v = (uchar16 *) block;
|
||||
uint4 *p = (uint4 *) block;
|
||||
|
||||
val = hc_swap32_S (p[0].w);
|
||||
val >>= 1;
|
||||
if (v[0].sb & 0x01) val |= 0x80000000;
|
||||
p[0].w = hc_swap32_S (val);
|
||||
|
||||
val = hc_swap32_S (p[0].z);
|
||||
val >>= 1;
|
||||
if (v[0].s7 & 0x01) val |= 0x80000000;
|
||||
p[0].z = hc_swap32_S (val);
|
||||
|
||||
val = hc_swap32_S (p[0].y);
|
||||
val >>= 1;
|
||||
if (v[0].s3 & 0x01) val |= 0x80000000;
|
||||
p[0].y = hc_swap32_S (val);
|
||||
|
||||
val = hc_swap32_S (p[0].x);
|
||||
val >>= 1;
|
||||
p[0].x = hc_swap32_S (val);
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_inc32 (u32 *block)
|
||||
{
|
||||
block[3] += 0x00000001;
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_xor_block (u32 *dst, const u32 *src)
|
||||
{
|
||||
*dst++ ^= *src++;
|
||||
*dst++ ^= *src++;
|
||||
*dst++ ^= *src++;
|
||||
*dst++ ^= *src++;
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_gf_mult (const uchar16 *x, const uchar16 *y, uchar16 *z)
|
||||
{
|
||||
u32 i, j, k;
|
||||
|
||||
z[0] = 0;
|
||||
uchar16 v = y[0].s32107654ba98fedc;
|
||||
|
||||
u8 x_char[16] = { x[0].s3, x[0].s2, x[0].s1, x[0].s0, x[0].s7, x[0].s6, x[0].s5, x[0].s4, x[0].sb, x[0].sa, x[0].s9, x[0].s8, x[0].sf, x[0].se, x[0].sd, x[0].sc };
|
||||
|
||||
u8 *v_char = (u8 *) &v;
|
||||
|
||||
for (i = 0; i < 16; i++)
|
||||
{
|
||||
for (j = 0; j < 8; j++)
|
||||
{
|
||||
if (x_char[i] & 1 << (7 - j))
|
||||
{
|
||||
z[0] ^= v;
|
||||
}
|
||||
|
||||
if (v.sf & 0x01)
|
||||
{
|
||||
AES_GCM_shift_right_block(v_char);
|
||||
v.s0 ^= 0xe1;
|
||||
}
|
||||
else
|
||||
{
|
||||
AES_GCM_shift_right_block(v_char);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_ghash (const u32 *subkey, const u32 *in, u32 in_len, u32 *out)
|
||||
{
|
||||
u32 m = in_len / 16;
|
||||
|
||||
u32 *xpos = in;
|
||||
|
||||
u32 tmp[4] = { 0 };
|
||||
|
||||
for (u32 i = 0; i < m; i++)
|
||||
{
|
||||
AES_GCM_xor_block (out, xpos);
|
||||
|
||||
xpos += 4;
|
||||
|
||||
AES_GCM_gf_mult (out, subkey, tmp);
|
||||
|
||||
tmp[0] = hc_swap32_S (tmp[0]);
|
||||
tmp[1] = hc_swap32_S (tmp[1]);
|
||||
tmp[2] = hc_swap32_S (tmp[2]);
|
||||
tmp[3] = hc_swap32_S (tmp[3]);
|
||||
|
||||
out[0] = tmp[0];
|
||||
out[1] = tmp[1];
|
||||
out[2] = tmp[2];
|
||||
out[3] = tmp[3];
|
||||
}
|
||||
|
||||
if (in + (in_len/4) > xpos)
|
||||
{
|
||||
u32 last = in + (in_len/4) - xpos;
|
||||
|
||||
for (u32 i = 0; i < last; i++)
|
||||
{
|
||||
tmp[i] = xpos[i];
|
||||
}
|
||||
|
||||
for (u32 i = last; i < 4; i++)
|
||||
{
|
||||
tmp[i] = 0;
|
||||
}
|
||||
|
||||
AES_GCM_xor_block (out, tmp);
|
||||
|
||||
AES_GCM_gf_mult (out, subkey, tmp);
|
||||
|
||||
out[0] = tmp[0];
|
||||
out[1] = tmp[1];
|
||||
out[2] = tmp[2];
|
||||
out[3] = tmp[3];
|
||||
}
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_Init (const u32 *ukey, u32 key_len, u32 *key, u32 *subkey, SHM_TYPE u32 *s_te0, SHM_TYPE u32 *s_te1, SHM_TYPE u32 *s_te2, SHM_TYPE u32 *s_te3, SHM_TYPE u32 *s_te4)
|
||||
{
|
||||
if (key_len == 128)
|
||||
{
|
||||
AES128_set_encrypt_key (key, ukey, s_te0, s_te1, s_te2, s_te3);
|
||||
|
||||
AES192_encrypt (key, subkey, subkey, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
}
|
||||
else if (key_len == 192)
|
||||
{
|
||||
AES192_set_encrypt_key (key, ukey, s_te0, s_te1, s_te2, s_te3);
|
||||
|
||||
AES192_encrypt (key, subkey, subkey, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
}
|
||||
else if (key_len == 256)
|
||||
{
|
||||
AES256_set_encrypt_key (key, ukey, s_te0, s_te1, s_te2, s_te3);
|
||||
|
||||
AES256_encrypt (key, subkey, subkey, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
}
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_Prepare_J0 (const u32 *iv, u32 iv_len, const u32 *subkey, u32 *J0)
|
||||
{
|
||||
if (iv_len == 12)
|
||||
{
|
||||
J0[0] = iv[0];
|
||||
J0[1] = iv[1];
|
||||
J0[2] = iv[2];
|
||||
J0[3] = 0x00000001;
|
||||
}
|
||||
else
|
||||
{
|
||||
J0[0] = iv[0];
|
||||
J0[1] = iv[1];
|
||||
J0[2] = iv[2];
|
||||
J0[3] = iv[3];
|
||||
|
||||
u32 len_buf[4] = { 0 };
|
||||
|
||||
len_buf[3] = iv_len * 8;
|
||||
|
||||
AES_GCM_ghash (subkey, len_buf, 16, J0);
|
||||
}
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_gctr (const u32 *key, const u32 *iv, const u32 *in, u32 in_len, u32 *out, SHM_TYPE u32 *s_te0, SHM_TYPE u32 *s_te1, SHM_TYPE u32 *s_te2, SHM_TYPE u32 *s_te3, SHM_TYPE u32 *s_te4)
|
||||
{
|
||||
const u32 *xpos = in;
|
||||
u32 *ypos = out;
|
||||
|
||||
u32 n = in_len / 16;
|
||||
|
||||
u32 iv_buf[4] = { iv[0], iv[1], iv[2], iv[3] };
|
||||
|
||||
for (u32 i = 0; i < n; i++)
|
||||
{
|
||||
AES256_encrypt (key, iv_buf, ypos, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
|
||||
AES_GCM_xor_block (ypos, xpos);
|
||||
|
||||
xpos += 4;
|
||||
ypos += 4;
|
||||
|
||||
AES_GCM_inc32 (iv_buf);
|
||||
}
|
||||
|
||||
u32 last = in + (in_len/4) - xpos;
|
||||
|
||||
if (last)
|
||||
{
|
||||
u32 tmp[4] = { 0 };
|
||||
|
||||
AES256_encrypt (key, iv_buf, tmp, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
|
||||
if (last >= 1) *ypos++ = *xpos++ ^ tmp[0];
|
||||
if (last >= 2) *ypos++ = *xpos++ ^ tmp[1];
|
||||
if (last >= 3) *ypos++ = *xpos++ ^ tmp[2];
|
||||
}
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_GCTR (u32 *key, u32 *J0, u32 *in, u32 in_len, u32 *out, SHM_TYPE u32 *s_te0, SHM_TYPE u32 *s_te1, SHM_TYPE u32 *s_te2, SHM_TYPE u32 *s_te3, SHM_TYPE u32 *s_te4)
|
||||
{
|
||||
u32 J0_incr[4] = {
|
||||
J0[0],
|
||||
J0[1],
|
||||
J0[2],
|
||||
J0[3],
|
||||
};
|
||||
|
||||
AES_GCM_gctr (key, J0_incr, in, in_len, out, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
}
|
||||
|
||||
DECLSPEC void AES_GCM_GHASH (const u32 *subkey, const u32 *aad_buf, u32 aad_len, u32 *enc_buf, u32 enc_len, u32 *out)
|
||||
{
|
||||
u32 len_buf[4] = { 0 };
|
||||
|
||||
out[0] = 0;
|
||||
out[1] = 0;
|
||||
out[2] = 0;
|
||||
out[3] = 0;
|
||||
|
||||
AES_GCM_ghash (subkey, aad_buf, aad_len, out);
|
||||
|
||||
// untested swap
|
||||
/*
|
||||
out[0] = hc_swap32_S (out[0]);
|
||||
out[1] = hc_swap32_S (out[1]);
|
||||
out[2] = hc_swap32_S (out[2]);
|
||||
out[3] = hc_swap32_S (out[3]);
|
||||
*/
|
||||
|
||||
AES_GCM_ghash (subkey, enc_buf, enc_len, out);
|
||||
|
||||
out[0] = hc_swap32_S (out[0]);
|
||||
out[1] = hc_swap32_S (out[1]);
|
||||
out[2] = hc_swap32_S (out[2]);
|
||||
out[3] = hc_swap32_S (out[3]);
|
||||
|
||||
len_buf[0] = aad_len * 8;
|
||||
len_buf[3] = enc_len * 8;
|
||||
|
||||
AES_GCM_ghash (subkey, len_buf, 16, out);
|
||||
}
|
||||
@@ -0,0 +1,20 @@
|
||||
/**
|
||||
* Author......: See docs/credits.txt
|
||||
* License.....: MIT
|
||||
*/
|
||||
|
||||
#ifndef _INC_CIPHER_AES_GCM_H
|
||||
#define _INC_CIPHER_AES_GCM_H
|
||||
|
||||
DECLSPEC void AES_GCM_shift_right_block(uchar *block);
|
||||
DECLSPEC void AES_GCM_inc32 (u32 *block);
|
||||
DECLSPEC void AES_GCM_xor_block (u32 *dst, const u32 *src);
|
||||
DECLSPEC void AES_GCM_gf_mult (const uchar16 *x, const uchar16 *y, uchar16 *z);
|
||||
DECLSPEC void AES_GCM_ghash (const u32 *subkey, const u32 *in, u32 in_len, u32 *out);
|
||||
DECLSPEC void AES_GCM_Init (const u32 *ukey, u32 key_len, u32 *key, u32 *subkey, SHM_TYPE u32 *s_te0, SHM_TYPE u32 *s_te1, SHM_TYPE u32 *s_te2, SHM_TYPE u32 *s_te3, SHM_TYPE u32 *s_te4);
|
||||
DECLSPEC void AES_GCM_Prepare_J0 (const u32 *iv, u32 iv_len, const u32 *subkey, u32 *J0);
|
||||
DECLSPEC void AES_GCM_gctr (const u32 *key, const u32 *iv, const u32 *in, u32 in_len, u32 *out, SHM_TYPE u32 *s_te0, SHM_TYPE u32 *s_te1, SHM_TYPE u32 *s_te2, SHM_TYPE u32 *s_te3, SHM_TYPE u32 *s_te4);
|
||||
DECLSPEC void AES_GCM_GCTR (u32 *key, u32 *J0, u32 *in, u32 in_len, u32 *out, SHM_TYPE u32 *s_te0, SHM_TYPE u32 *s_te1, SHM_TYPE u32 *s_te2, SHM_TYPE u32 *s_te3, SHM_TYPE u32 *s_te4);
|
||||
DECLSPEC void AES_GCM_GHASH (const u32 *subkey, const u32 *aad_buf, u32 aad_len, u32 *enc_buf, u32 enc_len, u32 *out);
|
||||
|
||||
#endif // _INC_CIPHER_AES_GCM_H
|
||||
@@ -0,0 +1,480 @@
|
||||
/**
|
||||
* 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_sha256.cl"
|
||||
#include "inc_cipher_aes.cl"
|
||||
#include "inc_cipher_aes-gcm.cl"
|
||||
#endif
|
||||
|
||||
#define COMPARE_S "inc_comp_single.cl"
|
||||
#define COMPARE_M "inc_comp_multi.cl"
|
||||
|
||||
typedef struct pbkdf2_sha256_tmp
|
||||
{
|
||||
u32 ipad[8];
|
||||
u32 opad[8];
|
||||
|
||||
u32 dgst[32];
|
||||
u32 out[32];
|
||||
|
||||
} pbkdf2_sha256_tmp_t;
|
||||
|
||||
typedef struct pbkdf2_sha256_aes_gcm
|
||||
{
|
||||
u32 salt_buf[64];
|
||||
u32 iv_buf[4];
|
||||
u32 iv_len;
|
||||
u32 ct_buf[14];
|
||||
u32 ct_len;
|
||||
|
||||
} pbkdf2_sha256_aes_gcm_t;
|
||||
|
||||
DECLSPEC void hmac_sha256_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];
|
||||
digest[5] = ipad[5];
|
||||
digest[6] = ipad[6];
|
||||
digest[7] = ipad[7];
|
||||
|
||||
sha256_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] = digest[5];
|
||||
w1[2] = digest[6];
|
||||
w1[3] = digest[7];
|
||||
w2[0] = 0x80000000;
|
||||
w2[1] = 0;
|
||||
w2[2] = 0;
|
||||
w2[3] = 0;
|
||||
w3[0] = 0;
|
||||
w3[1] = 0;
|
||||
w3[2] = 0;
|
||||
w3[3] = (64 + 32) * 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];
|
||||
|
||||
sha256_transform_vector (w0, w1, w2, w3, digest);
|
||||
}
|
||||
|
||||
KERNEL_FQ void m27000_init (KERN_ATTR_TMPS_ESALT (pbkdf2_sha256_tmp_t, pbkdf2_sha256_aes_gcm_t))
|
||||
{
|
||||
/**
|
||||
* base
|
||||
*/
|
||||
|
||||
const u64 gid = get_global_id (0);
|
||||
|
||||
if (gid >= gid_max) return;
|
||||
|
||||
sha256_hmac_ctx_t sha256_hmac_ctx;
|
||||
|
||||
sha256_hmac_init_global_swap (&sha256_hmac_ctx, pws[gid].i, pws[gid].pw_len);
|
||||
|
||||
tmps[gid].ipad[0] = sha256_hmac_ctx.ipad.h[0];
|
||||
tmps[gid].ipad[1] = sha256_hmac_ctx.ipad.h[1];
|
||||
tmps[gid].ipad[2] = sha256_hmac_ctx.ipad.h[2];
|
||||
tmps[gid].ipad[3] = sha256_hmac_ctx.ipad.h[3];
|
||||
tmps[gid].ipad[4] = sha256_hmac_ctx.ipad.h[4];
|
||||
tmps[gid].ipad[5] = sha256_hmac_ctx.ipad.h[5];
|
||||
tmps[gid].ipad[6] = sha256_hmac_ctx.ipad.h[6];
|
||||
tmps[gid].ipad[7] = sha256_hmac_ctx.ipad.h[7];
|
||||
|
||||
tmps[gid].opad[0] = sha256_hmac_ctx.opad.h[0];
|
||||
tmps[gid].opad[1] = sha256_hmac_ctx.opad.h[1];
|
||||
tmps[gid].opad[2] = sha256_hmac_ctx.opad.h[2];
|
||||
tmps[gid].opad[3] = sha256_hmac_ctx.opad.h[3];
|
||||
tmps[gid].opad[4] = sha256_hmac_ctx.opad.h[4];
|
||||
tmps[gid].opad[5] = sha256_hmac_ctx.opad.h[5];
|
||||
tmps[gid].opad[6] = sha256_hmac_ctx.opad.h[6];
|
||||
tmps[gid].opad[7] = sha256_hmac_ctx.opad.h[7];
|
||||
|
||||
sha256_hmac_update_global_swap (&sha256_hmac_ctx, esalt_bufs[DIGESTS_OFFSET].salt_buf, salt_bufs[SALT_POS].salt_len);
|
||||
|
||||
for (u32 i = 0, j = 1; i < 8; i += 8, j += 1)
|
||||
{
|
||||
sha256_hmac_ctx_t sha256_hmac_ctx2 = sha256_hmac_ctx;
|
||||
|
||||
u32 w0[4];
|
||||
u32 w1[4];
|
||||
u32 w2[4];
|
||||
u32 w3[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;
|
||||
|
||||
sha256_hmac_update_64 (&sha256_hmac_ctx2, w0, w1, w2, w3, 4);
|
||||
|
||||
sha256_hmac_final (&sha256_hmac_ctx2);
|
||||
|
||||
tmps[gid].dgst[i + 0] = sha256_hmac_ctx2.opad.h[0];
|
||||
tmps[gid].dgst[i + 1] = sha256_hmac_ctx2.opad.h[1];
|
||||
tmps[gid].dgst[i + 2] = sha256_hmac_ctx2.opad.h[2];
|
||||
tmps[gid].dgst[i + 3] = sha256_hmac_ctx2.opad.h[3];
|
||||
tmps[gid].dgst[i + 4] = sha256_hmac_ctx2.opad.h[4];
|
||||
tmps[gid].dgst[i + 5] = sha256_hmac_ctx2.opad.h[5];
|
||||
tmps[gid].dgst[i + 6] = sha256_hmac_ctx2.opad.h[6];
|
||||
tmps[gid].dgst[i + 7] = sha256_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 m27000_loop (KERN_ATTR_TMPS_ESALT (pbkdf2_sha256_tmp_t, pbkdf2_sha256_aes_gcm_t))
|
||||
{
|
||||
const u64 gid = get_global_id (0);
|
||||
|
||||
if ((gid * VECT_SIZE) >= gid_max) return;
|
||||
|
||||
u32x ipad[8];
|
||||
u32x opad[8];
|
||||
|
||||
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);
|
||||
ipad[5] = packv (tmps, ipad, gid, 5);
|
||||
ipad[6] = packv (tmps, ipad, gid, 6);
|
||||
ipad[7] = packv (tmps, ipad, gid, 7);
|
||||
|
||||
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);
|
||||
opad[5] = packv (tmps, opad, gid, 5);
|
||||
opad[6] = packv (tmps, opad, gid, 6);
|
||||
opad[7] = packv (tmps, opad, gid, 7);
|
||||
|
||||
for (u32 i = 0; i < 8; i += 8)
|
||||
{
|
||||
u32x dgst[8];
|
||||
u32x out[8];
|
||||
|
||||
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);
|
||||
dgst[5] = packv (tmps, dgst, gid, i + 5);
|
||||
dgst[6] = packv (tmps, dgst, gid, i + 6);
|
||||
dgst[7] = packv (tmps, dgst, gid, i + 7);
|
||||
|
||||
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);
|
||||
out[5] = packv (tmps, out, gid, i + 5);
|
||||
out[6] = packv (tmps, out, gid, i + 6);
|
||||
out[7] = packv (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];
|
||||
|
||||
w0[0] = dgst[0];
|
||||
w0[1] = dgst[1];
|
||||
w0[2] = dgst[2];
|
||||
w0[3] = dgst[3];
|
||||
w1[0] = dgst[4];
|
||||
w1[1] = dgst[5];
|
||||
w1[2] = dgst[6];
|
||||
w1[3] = dgst[7];
|
||||
w2[0] = 0x80000000;
|
||||
w2[1] = 0;
|
||||
w2[2] = 0;
|
||||
w2[3] = 0;
|
||||
w3[0] = 0;
|
||||
w3[1] = 0;
|
||||
w3[2] = 0;
|
||||
w3[3] = (64 + 32) * 8;
|
||||
|
||||
hmac_sha256_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];
|
||||
out[5] ^= dgst[5];
|
||||
out[6] ^= dgst[6];
|
||||
out[7] ^= dgst[7];
|
||||
}
|
||||
|
||||
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, dgst, gid, i + 5, dgst[5]);
|
||||
unpackv (tmps, dgst, gid, i + 6, dgst[6]);
|
||||
unpackv (tmps, dgst, gid, i + 7, dgst[7]);
|
||||
|
||||
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]);
|
||||
unpackv (tmps, out, gid, i + 5, out[5]);
|
||||
unpackv (tmps, out, gid, i + 6, out[6]);
|
||||
unpackv (tmps, out, gid, i + 7, out[7]);
|
||||
}
|
||||
}
|
||||
|
||||
KERNEL_FQ void m27000_comp (KERN_ATTR_TMPS_ESALT (pbkdf2_sha256_tmp_t, pbkdf2_sha256_aes_gcm_t))
|
||||
{
|
||||
/**
|
||||
* base
|
||||
*/
|
||||
|
||||
const u64 gid = get_global_id (0);
|
||||
|
||||
if (gid >= gid_max) return;
|
||||
|
||||
const u64 lid = get_local_id (0);
|
||||
const u64 lsz = get_local_size (0);
|
||||
|
||||
/**
|
||||
* aes shared
|
||||
*/
|
||||
|
||||
#ifdef REAL_SHM
|
||||
|
||||
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_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_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_max) return;
|
||||
|
||||
// keys
|
||||
|
||||
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];
|
||||
|
||||
u32 key_len = 32 * 8;
|
||||
|
||||
u32 key[60] = { 0 };
|
||||
u32 subKey[4] = { 0 };
|
||||
|
||||
AES256_set_encrypt_key (key, ukey, s_te0, s_te1, s_te2, s_te3);
|
||||
|
||||
AES256_encrypt (key, subKey, subKey, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
|
||||
// iv
|
||||
|
||||
const u32 iv[4] = {
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[0],
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[1],
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[2],
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[3]
|
||||
};
|
||||
|
||||
const u32 iv_len = esalt_bufs[DIGESTS_OFFSET].iv_len;
|
||||
|
||||
u32 J0[4] = {
|
||||
iv[0],
|
||||
iv[1],
|
||||
iv[2],
|
||||
0x00000001
|
||||
};
|
||||
|
||||
// ct
|
||||
|
||||
u32 enc[14] = { 0 };
|
||||
|
||||
enc[ 0] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 0];
|
||||
enc[ 1] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 1];
|
||||
enc[ 2] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 2];
|
||||
enc[ 3] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 3];
|
||||
enc[ 4] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 4];
|
||||
enc[ 5] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 5];
|
||||
enc[ 6] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 6];
|
||||
enc[ 7] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 7];
|
||||
enc[ 8] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 8];
|
||||
enc[ 9] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 9];
|
||||
enc[10] = esalt_bufs[DIGESTS_OFFSET].ct_buf[10];
|
||||
enc[11] = esalt_bufs[DIGESTS_OFFSET].ct_buf[11];
|
||||
enc[12] = esalt_bufs[DIGESTS_OFFSET].ct_buf[12];
|
||||
enc[13] = esalt_bufs[DIGESTS_OFFSET].ct_buf[13];
|
||||
|
||||
u32 enc_len = esalt_bufs[DIGESTS_OFFSET].ct_len;
|
||||
|
||||
u32 S[4] = { 0 };
|
||||
|
||||
u32 t[4] = { 0 };
|
||||
|
||||
S[0] ^= enc[0];
|
||||
S[1] ^= enc[1];
|
||||
S[2] ^= enc[2];
|
||||
S[3] ^= enc[3];
|
||||
|
||||
AES_GCM_gf_mult (S, subKey, t);
|
||||
|
||||
t[0] = hc_swap32_S (t[0]);
|
||||
t[1] = hc_swap32_S (t[1]);
|
||||
t[2] = hc_swap32_S (t[2]);
|
||||
t[3] = hc_swap32_S (t[3]);
|
||||
|
||||
S[0] = t[0] ^ enc[4];
|
||||
S[1] = t[1] ^ enc[5];
|
||||
S[2] = t[2] ^ enc[6];
|
||||
S[3] = t[3] ^ enc[7];
|
||||
|
||||
AES_GCM_gf_mult (S, subKey, t);
|
||||
|
||||
t[0] = hc_swap32_S (t[0]);
|
||||
t[1] = hc_swap32_S (t[1]);
|
||||
t[2] = hc_swap32_S (t[2]);
|
||||
t[3] = hc_swap32_S (t[3]);
|
||||
|
||||
S[0] = t[0] ^ enc[8];
|
||||
S[1] = t[1] ^ enc[9];
|
||||
S[2] = t[2] ^ enc[10];
|
||||
S[3] = t[3] ^ enc[11];
|
||||
|
||||
AES_GCM_gf_mult (S, subKey, t);
|
||||
|
||||
t[0] = hc_swap32_S (t[0]);
|
||||
t[1] = hc_swap32_S (t[1]);
|
||||
t[2] = hc_swap32_S (t[2]);
|
||||
t[3] = hc_swap32_S (t[3]);
|
||||
|
||||
S[0] = t[0];
|
||||
S[1] = t[1];
|
||||
S[2] = t[2];
|
||||
S[3] = t[3];
|
||||
|
||||
t[0] = enc[12];
|
||||
t[1] = enc[13];
|
||||
t[2] = 0;
|
||||
t[3] = 0;
|
||||
|
||||
S[0] ^= t[0];
|
||||
S[1] ^= t[1];
|
||||
S[2] ^= t[2];
|
||||
S[3] ^= t[3];
|
||||
|
||||
AES_GCM_gf_mult (S, subKey, t);
|
||||
|
||||
S[0] = hc_swap32_S (t[0]);
|
||||
S[1] = hc_swap32_S (t[1]);
|
||||
S[2] = hc_swap32_S (t[2]);
|
||||
S[3] = hc_swap32_S (t[3]);
|
||||
|
||||
u32 len_buf[4] = { 0 };
|
||||
|
||||
len_buf[0] = 0;
|
||||
len_buf[3] = enc_len * 8;
|
||||
|
||||
S[0] ^= len_buf[0];
|
||||
S[1] ^= len_buf[1];
|
||||
S[2] ^= len_buf[2];
|
||||
S[3] ^= len_buf[3];
|
||||
|
||||
AES_GCM_gf_mult (S, subKey, t);
|
||||
|
||||
S[0] = hc_swap32_S (t[0]);
|
||||
S[1] = hc_swap32_S (t[1]);
|
||||
S[2] = hc_swap32_S (t[2]);
|
||||
S[3] = hc_swap32_S (t[3]);
|
||||
|
||||
J0[3] = 0x00000001;
|
||||
|
||||
u32 T[4] = { 0 };
|
||||
|
||||
AES256_encrypt (key, J0, T, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
|
||||
/* compare tag */
|
||||
|
||||
const u32 r0 = T[0] ^ S[0];
|
||||
const u32 r1 = T[1] ^ S[1];
|
||||
const u32 r2 = T[2] ^ S[2];
|
||||
const u32 r3 = T[3] ^ S[3];
|
||||
|
||||
#define il_pos 0
|
||||
|
||||
#ifdef KERNEL_STATIC
|
||||
#include COMPARE_M
|
||||
#endif
|
||||
}
|
||||
@@ -0,0 +1,408 @@
|
||||
/**
|
||||
* 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_sha256.cl"
|
||||
#include "inc_cipher_aes.cl"
|
||||
#include "inc_cipher_aes-gcm.cl"
|
||||
#endif
|
||||
|
||||
#define COMPARE_S "inc_comp_single.cl"
|
||||
#define COMPARE_M "inc_comp_multi.cl"
|
||||
|
||||
typedef struct pbkdf2_sha256_tmp
|
||||
{
|
||||
u32 ipad[8];
|
||||
u32 opad[8];
|
||||
|
||||
u32 dgst[32];
|
||||
u32 out[32];
|
||||
|
||||
} pbkdf2_sha256_tmp_t;
|
||||
|
||||
typedef struct pbkdf2_sha256_aes_gcm
|
||||
{
|
||||
u32 salt_buf[64];
|
||||
u32 iv_buf[4];
|
||||
u32 iv_len;
|
||||
u32 ct_buf[14];
|
||||
u32 ct_len;
|
||||
|
||||
} pbkdf2_sha256_aes_gcm_t;
|
||||
|
||||
DECLSPEC void hmac_sha256_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];
|
||||
digest[5] = ipad[5];
|
||||
digest[6] = ipad[6];
|
||||
digest[7] = ipad[7];
|
||||
|
||||
sha256_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] = digest[5];
|
||||
w1[2] = digest[6];
|
||||
w1[3] = digest[7];
|
||||
w2[0] = 0x80000000;
|
||||
w2[1] = 0;
|
||||
w2[2] = 0;
|
||||
w2[3] = 0;
|
||||
w3[0] = 0;
|
||||
w3[1] = 0;
|
||||
w3[2] = 0;
|
||||
w3[3] = (64 + 32) * 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];
|
||||
|
||||
sha256_transform_vector (w0, w1, w2, w3, digest);
|
||||
}
|
||||
|
||||
KERNEL_FQ void m27000_init (KERN_ATTR_TMPS_ESALT (pbkdf2_sha256_tmp_t, pbkdf2_sha256_aes_gcm_t))
|
||||
{
|
||||
/**
|
||||
* base
|
||||
*/
|
||||
|
||||
const u64 gid = get_global_id (0);
|
||||
|
||||
if (gid >= gid_max) return;
|
||||
|
||||
sha256_hmac_ctx_t sha256_hmac_ctx;
|
||||
|
||||
sha256_hmac_init_global_swap (&sha256_hmac_ctx, pws[gid].i, pws[gid].pw_len);
|
||||
|
||||
tmps[gid].ipad[0] = sha256_hmac_ctx.ipad.h[0];
|
||||
tmps[gid].ipad[1] = sha256_hmac_ctx.ipad.h[1];
|
||||
tmps[gid].ipad[2] = sha256_hmac_ctx.ipad.h[2];
|
||||
tmps[gid].ipad[3] = sha256_hmac_ctx.ipad.h[3];
|
||||
tmps[gid].ipad[4] = sha256_hmac_ctx.ipad.h[4];
|
||||
tmps[gid].ipad[5] = sha256_hmac_ctx.ipad.h[5];
|
||||
tmps[gid].ipad[6] = sha256_hmac_ctx.ipad.h[6];
|
||||
tmps[gid].ipad[7] = sha256_hmac_ctx.ipad.h[7];
|
||||
|
||||
tmps[gid].opad[0] = sha256_hmac_ctx.opad.h[0];
|
||||
tmps[gid].opad[1] = sha256_hmac_ctx.opad.h[1];
|
||||
tmps[gid].opad[2] = sha256_hmac_ctx.opad.h[2];
|
||||
tmps[gid].opad[3] = sha256_hmac_ctx.opad.h[3];
|
||||
tmps[gid].opad[4] = sha256_hmac_ctx.opad.h[4];
|
||||
tmps[gid].opad[5] = sha256_hmac_ctx.opad.h[5];
|
||||
tmps[gid].opad[6] = sha256_hmac_ctx.opad.h[6];
|
||||
tmps[gid].opad[7] = sha256_hmac_ctx.opad.h[7];
|
||||
|
||||
sha256_hmac_update_global_swap (&sha256_hmac_ctx, esalt_bufs[DIGESTS_OFFSET].salt_buf, salt_bufs[SALT_POS].salt_len);
|
||||
|
||||
for (u32 i = 0, j = 1; i < 8; i += 8, j += 1)
|
||||
{
|
||||
sha256_hmac_ctx_t sha256_hmac_ctx2 = sha256_hmac_ctx;
|
||||
|
||||
u32 w0[4];
|
||||
u32 w1[4];
|
||||
u32 w2[4];
|
||||
u32 w3[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;
|
||||
|
||||
sha256_hmac_update_64 (&sha256_hmac_ctx2, w0, w1, w2, w3, 4);
|
||||
|
||||
sha256_hmac_final (&sha256_hmac_ctx2);
|
||||
|
||||
tmps[gid].dgst[i + 0] = sha256_hmac_ctx2.opad.h[0];
|
||||
tmps[gid].dgst[i + 1] = sha256_hmac_ctx2.opad.h[1];
|
||||
tmps[gid].dgst[i + 2] = sha256_hmac_ctx2.opad.h[2];
|
||||
tmps[gid].dgst[i + 3] = sha256_hmac_ctx2.opad.h[3];
|
||||
tmps[gid].dgst[i + 4] = sha256_hmac_ctx2.opad.h[4];
|
||||
tmps[gid].dgst[i + 5] = sha256_hmac_ctx2.opad.h[5];
|
||||
tmps[gid].dgst[i + 6] = sha256_hmac_ctx2.opad.h[6];
|
||||
tmps[gid].dgst[i + 7] = sha256_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 m27000_loop (KERN_ATTR_TMPS_ESALT (pbkdf2_sha256_tmp_t, pbkdf2_sha256_aes_gcm_t))
|
||||
{
|
||||
const u64 gid = get_global_id (0);
|
||||
|
||||
if ((gid * VECT_SIZE) >= gid_max) return;
|
||||
|
||||
u32x ipad[8];
|
||||
u32x opad[8];
|
||||
|
||||
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);
|
||||
ipad[5] = packv (tmps, ipad, gid, 5);
|
||||
ipad[6] = packv (tmps, ipad, gid, 6);
|
||||
ipad[7] = packv (tmps, ipad, gid, 7);
|
||||
|
||||
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);
|
||||
opad[5] = packv (tmps, opad, gid, 5);
|
||||
opad[6] = packv (tmps, opad, gid, 6);
|
||||
opad[7] = packv (tmps, opad, gid, 7);
|
||||
|
||||
for (u32 i = 0; i < 8; i += 8)
|
||||
{
|
||||
u32x dgst[8];
|
||||
u32x out[8];
|
||||
|
||||
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);
|
||||
dgst[5] = packv (tmps, dgst, gid, i + 5);
|
||||
dgst[6] = packv (tmps, dgst, gid, i + 6);
|
||||
dgst[7] = packv (tmps, dgst, gid, i + 7);
|
||||
|
||||
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);
|
||||
out[5] = packv (tmps, out, gid, i + 5);
|
||||
out[6] = packv (tmps, out, gid, i + 6);
|
||||
out[7] = packv (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];
|
||||
|
||||
w0[0] = dgst[0];
|
||||
w0[1] = dgst[1];
|
||||
w0[2] = dgst[2];
|
||||
w0[3] = dgst[3];
|
||||
w1[0] = dgst[4];
|
||||
w1[1] = dgst[5];
|
||||
w1[2] = dgst[6];
|
||||
w1[3] = dgst[7];
|
||||
w2[0] = 0x80000000;
|
||||
w2[1] = 0;
|
||||
w2[2] = 0;
|
||||
w2[3] = 0;
|
||||
w3[0] = 0;
|
||||
w3[1] = 0;
|
||||
w3[2] = 0;
|
||||
w3[3] = (64 + 32) * 8;
|
||||
|
||||
hmac_sha256_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];
|
||||
out[5] ^= dgst[5];
|
||||
out[6] ^= dgst[6];
|
||||
out[7] ^= dgst[7];
|
||||
}
|
||||
|
||||
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, dgst, gid, i + 5, dgst[5]);
|
||||
unpackv (tmps, dgst, gid, i + 6, dgst[6]);
|
||||
unpackv (tmps, dgst, gid, i + 7, dgst[7]);
|
||||
|
||||
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]);
|
||||
unpackv (tmps, out, gid, i + 5, out[5]);
|
||||
unpackv (tmps, out, gid, i + 6, out[6]);
|
||||
unpackv (tmps, out, gid, i + 7, out[7]);
|
||||
}
|
||||
}
|
||||
|
||||
KERNEL_FQ void m27000_comp (KERN_ATTR_TMPS_ESALT (pbkdf2_sha256_tmp_t, pbkdf2_sha256_aes_gcm_t))
|
||||
{
|
||||
/**
|
||||
* base
|
||||
*/
|
||||
|
||||
const u64 gid = get_global_id (0);
|
||||
|
||||
if (gid >= gid_max) return;
|
||||
|
||||
const u64 lid = get_local_id (0);
|
||||
const u64 lsz = get_local_size (0);
|
||||
|
||||
/**
|
||||
* aes shared
|
||||
*/
|
||||
|
||||
#ifdef REAL_SHM
|
||||
|
||||
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_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_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_max) return;
|
||||
|
||||
// keys
|
||||
|
||||
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];
|
||||
|
||||
u32 key_len = 32 * 8;
|
||||
|
||||
u32 key[60] = { 0 };
|
||||
u32 subKey[4] = { 0 };
|
||||
|
||||
AES_GCM_Init (ukey, key_len, key, subKey, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
|
||||
// iv
|
||||
|
||||
const u32 iv[4] = {
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[0],
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[1],
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[2],
|
||||
esalt_bufs[DIGESTS_OFFSET].iv_buf[3]
|
||||
};
|
||||
|
||||
const u32 iv_len = esalt_bufs[DIGESTS_OFFSET].iv_len;
|
||||
|
||||
u32 J0[4] = { 0 };
|
||||
|
||||
AES_GCM_Prepare_J0 (iv, iv_len, subKey, J0);
|
||||
|
||||
// ct
|
||||
|
||||
u32 enc[14] = { 0 };
|
||||
|
||||
enc[ 0] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 0];
|
||||
enc[ 1] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 1];
|
||||
enc[ 2] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 2];
|
||||
enc[ 3] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 3];
|
||||
enc[ 4] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 4];
|
||||
enc[ 5] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 5];
|
||||
enc[ 6] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 6];
|
||||
enc[ 7] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 7];
|
||||
enc[ 8] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 8];
|
||||
enc[ 9] = esalt_bufs[DIGESTS_OFFSET].ct_buf[ 9];
|
||||
enc[10] = esalt_bufs[DIGESTS_OFFSET].ct_buf[10];
|
||||
enc[11] = esalt_bufs[DIGESTS_OFFSET].ct_buf[11];
|
||||
enc[12] = esalt_bufs[DIGESTS_OFFSET].ct_buf[12];
|
||||
enc[13] = esalt_bufs[DIGESTS_OFFSET].ct_buf[13];
|
||||
|
||||
u32 enc_len = esalt_bufs[DIGESTS_OFFSET].ct_len;
|
||||
|
||||
/*
|
||||
// decrypt buffer is not usefull here, skip
|
||||
u32 dec[14] = { 0 };
|
||||
|
||||
AES_GCM_GCTR (key, J0, enc, enc_len, dec, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
*/
|
||||
|
||||
u32 T[4] = { 0 };
|
||||
u32 S[4] = { 0 };
|
||||
|
||||
u32 S_len = 16;
|
||||
u32 aad_buf = 0;
|
||||
u32 aad_len = 0;
|
||||
|
||||
AES_GCM_GHASH (subKey, aad_buf, aad_len, enc, enc_len, S);
|
||||
|
||||
AES_GCM_GCTR (key, J0, S, S_len, T, s_te0, s_te1, s_te2, s_te3, s_te4);
|
||||
|
||||
/* compare tag */
|
||||
|
||||
const u32 r0 = T[0];
|
||||
const u32 r1 = T[1];
|
||||
const u32 r2 = T[2];
|
||||
const u32 r3 = T[3];
|
||||
|
||||
#define il_pos 0
|
||||
|
||||
#ifdef KERNEL_STATIC
|
||||
#include COMPARE_M
|
||||
#endif
|
||||
}
|
||||
Reference in New Issue
Block a user