Ecosystem/xmrig
Ecosystem/xmrig
1
0
Fork 0
mirror of https://github.com/xmrig/xmrig.git synced 2025-01-08 20:09:52 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Much better software AES implementation (--av 4).

Browse source
This commit is contained in:
XMRig 2017-04-19 10:03:40 +03:00
parent 1013aa5004
commit 21c243ed8f
12 changed files with 359 additions and 1921 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
CMakeLists.txt
View file

@ -21,9 +21,6 @@ set(HEADERS_CRYPTO
crypto/c_blake256.h crypto/c_blake256.h
crypto/c_jh.h crypto/c_jh.h
crypto/c_skein.h crypto/c_skein.h
crypto/oaes_lib.h
crypto/oaes_config.h
crypto/aesb.h
) )
set(HEADERS_COMPAT set(HEADERS_COMPAT
@ -53,8 +50,7 @@ set(SOURCES_CRYPTO
crypto/c_blake256.c crypto/c_blake256.c
crypto/c_jh.c crypto/c_jh.c
crypto/c_skein.c crypto/c_skein.c
crypto/oaes_lib.c crypto/soft_aes.c
crypto/aesb.c
) )
set(SOURCES_UTILS set(SOURCES_UTILS
@ -106,7 +102,7 @@ if (CMAKE_SIZEOF_VOID_P EQUAL 8)
set(CRYPTONIGHT64 set(CRYPTONIGHT64
algo/cryptonight/cryptonight_av1_aesni.c algo/cryptonight/cryptonight_av1_aesni.c
algo/cryptonight/cryptonight_av2_aesni_wolf.c algo/cryptonight/cryptonight_av2_aesni_wolf.c
algo/cryptonight/cryptonight_av4_legacy.c algo/cryptonight/cryptonight_av4_softaes.c
algo/cryptonight/cryptonight_av5_aesni_stak.c algo/cryptonight/cryptonight_av5_aesni_stak.c
algo/cryptonight/cryptonight_av6_aesni_experimental.c algo/cryptonight/cryptonight_av6_aesni_experimental.c
) )
@ -116,7 +112,7 @@ if (CMAKE_SIZEOF_VOID_P EQUAL 8)
else() else()
set(CRYPTONIGHT32 set(CRYPTONIGHT32
algo/cryptonight/cryptonight_av1_aesni32.c algo/cryptonight/cryptonight_av1_aesni32.c
algo/cryptonight/cryptonight_av4_legacy.c algo/cryptonight/cryptonight_av4_softaes.c
) )
add_executable(xmrig32 ${HEADERS} ${HEADERS_CRYPTO} ${SOURCES} ${SOURCES_CRYPTO} ${HEADERS_UTILS} ${SOURCES_UTILS} ${HEADERS_COMPAT} ${SOURCES_COMPAT} ${SOURCES_OS} ${CRYPTONIGHT32}) add_executable(xmrig32 ${HEADERS} ${HEADERS_CRYPTO} ${SOURCES} ${SOURCES_CRYPTO} ${HEADERS_UTILS} ${SOURCES_UTILS} ${HEADERS_COMPAT} ${SOURCES_COMPAT} ${SOURCES_OS} ${CRYPTONIGHT32})

43
algo/cryptonight/cryptonight_av3_aesni_bmi2.c
View file

@ -108,9 +108,12 @@ static inline void ExpandAESKey256(char *keybuf)
keys[14] = tmp1; keys[14] = tmp1;
} }
void cryptonight_av3_aesni_bmi2(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx) void cryptonight_av3_aesni_bmi2(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx)
{ {
keccak((const uint8_t *) input, 76, (uint8_t *) &ctx->state.hs, 200); uint64_t* state = ctx->state.hs.w;
keccak((const uint8_t *)input, 76, (uint8_t *) state, 200);
uint8_t ExpandedKey[256]; uint8_t ExpandedKey[256];
size_t i, j; size_t i, j;
@ -146,38 +149,32 @@ void cryptonight_av3_aesni_bmi2(void *restrict output, const void *restrict inpu
_mm_store_si128(&(longoutput[(i >> 4) + 7]), xmminput[7]); _mm_store_si128(&(longoutput[(i >> 4) + 7]), xmminput[7]);
} }
for (i = 0; i < 2; i++) uint64_t a[2] __attribute((aligned(16))) = { state[0] ^ state[4], state[1] ^ state[5] };
{ uint64_t c __attribute((aligned(16)));
ctx->a[i] = ((uint64_t *)ctx->state.k)[i] ^ ((uint64_t *)ctx->state.k)[i+4];
ctx->b[i] = ((uint64_t *)ctx->state.k)[i+2] ^ ((uint64_t *)ctx->state.k)[i+6];
}
__m128i a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
__m128i b_x = _mm_load_si128((__m128i *) ctx->b);
uint64_t c[2] __attribute((aligned(16)));
uint64_t d[2] __attribute((aligned(16))); uint64_t d[2] __attribute((aligned(16)));
uint64_t hi; uint64_t hi;
for (i = 0; __builtin_expect(i < 0x80000, 1); i++) { __m128i a_x = _mm_load_si128((__m128i *) &memory[a[0] & 0x1FFFF0]);
__m128i c_x = _mm_aesenc_si128(a_x, _mm_load_si128((__m128i *) ctx->a)); __m128i b_x = _mm_set_epi64x(state[3] ^ state[7], state[2] ^ state[6]);
_mm_store_si128((__m128i *) c, c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c[0] & 0x1FFFF0]; for (i = 0; __builtin_expect(i < 0x80000, 1); i++) {
_mm_store_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x)); __m128i c_x = _mm_aesenc_si128(a_x, _mm_load_si128((__m128i *) a));
c = _mm_cvtsi128_si64(c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c & 0x1FFFF0];
_mm_store_si128((__m128i *) &memory[a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x));
b_x = c_x; b_x = c_x;
d[0] = d_ptr[0]; d[0] = d_ptr[0];
d[1] = d_ptr[1]; d[1] = d_ptr[1];
d_ptr[1] = ctx->a[1] += _mulx_u64(c[0], d[0], &hi); d_ptr[1] = a[1] += _mulx_u64(c, d[0], &hi);
d_ptr[0] = ctx->a[0] += hi; d_ptr[0] = a[0] += hi;
ctx->a[0] ^= d[0]; a[0] ^= d[0];
ctx->a[1] ^= d[1]; a[1] ^= d[1];
a_x = _mm_load_si128((__m128i *) &memory[ctx->a[0] & 0x1FFFF0]);
a_x = _mm_load_si128((__m128i *) &memory[a[0] & 0x1FFFF0]);
} }
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE); memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
@ -209,6 +206,6 @@ void cryptonight_av3_aesni_bmi2(void *restrict output, const void *restrict inpu
} }
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE); memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24); keccakf((uint64_t *) state, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output); extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
} }

151
algo/cryptonight/cryptonight_av4_legacy.c
View file

@ -1,151 +0,0 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "compat.h"
#include "crypto/c_keccak.h"
#include "crypto/aesb.h"
#include "crypto/oaes_lib.h"
static inline uint64_t mul128(uint64_t multiplier, uint64_t multiplicand, uint64_t *product_hi) {
// multiplier = ab = a * 2^32 + b
// multiplicand = cd = c * 2^32 + d
// ab * cd = a * c * 2^64 + (a * d + b * c) * 2^32 + b * d
uint64_t a = multiplier >> 32;
uint64_t b = multiplier & 0xFFFFFFFF;
uint64_t c = multiplicand >> 32;
uint64_t d = multiplicand & 0xFFFFFFFF;
//uint64_t ac = a * c;
uint64_t ad = a * d;
//uint64_t bc = b * c;
uint64_t bd = b * d;
uint64_t adbc = ad + (b * c);
uint64_t adbc_carry = adbc < ad ? 1 : 0;
// multiplier * multiplicand = product_hi * 2^64 + product_lo
uint64_t product_lo = bd + (adbc << 32);
uint64_t product_lo_carry = product_lo < bd ? 1 : 0;
*product_hi = (a * c) + (adbc >> 32) + (adbc_carry << 32) + product_lo_carry;
return product_lo;
}
static inline void mul_sum_xor_dst(const uint8_t* a, uint8_t* c, uint8_t* dst) {
uint64_t hi, lo = mul128(((uint64_t*) a)[0], ((uint64_t*) dst)[0], &hi) + ((uint64_t*) c)[1];
hi += ((uint64_t*) c)[0];
((uint64_t*) c)[0] = ((uint64_t*) dst)[0] ^ hi;
((uint64_t*) c)[1] = ((uint64_t*) dst)[1] ^ lo;
((uint64_t*) dst)[0] = hi;
((uint64_t*) dst)[1] = lo;
}
static inline void xor_blocks(uint8_t* a, const uint8_t* b) {
((uint64_t*) a)[0] ^= ((uint64_t*) b)[0];
((uint64_t*) a)[1] ^= ((uint64_t*) b)[1];
}
static inline void xor_blocks_dst(const uint8_t* a, const uint8_t* b, uint8_t* dst) {
((uint64_t*) dst)[0] = ((uint64_t*) a)[0] ^ ((uint64_t*) b)[0];
((uint64_t*) dst)[1] = ((uint64_t*) a)[1] ^ ((uint64_t*) b)[1];
}
void cryptonight_av4_legacy(void *restrict output, const void *restrict input, const char *restrict memory, struct cryptonight_ctx *restrict ctx) {
oaes_ctx *aes_ctx = (oaes_ctx*) oaes_alloc();
size_t i, j;
keccak((const uint8_t *)input, 76, (uint8_t *) &ctx->state.hs, 200);
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
oaes_key_import_data(aes_ctx, ctx->state.hs.b, AES_KEY_SIZE);
for (i = 0; likely(i < MEMORY); i += INIT_SIZE_BYTE) {
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 0], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 1], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 2], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 3], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 4], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 5], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 6], aes_ctx->key->exp_data);
aesb_pseudo_round_mut(&ctx->text[AES_BLOCK_SIZE * 7], aes_ctx->key->exp_data);
memcpy((void *) &memory[i], ctx->text, INIT_SIZE_BYTE);
}
xor_blocks_dst(&ctx->state.k[0], &ctx->state.k[32], (uint8_t*) ctx->a);
xor_blocks_dst(&ctx->state.k[16], &ctx->state.k[48], (uint8_t*) ctx->b);
for (i = 0; likely(i < ITER / 4); ++i) {
/* Dependency chain: address -> read value ------+
* written value <-+ hard function (AES or MUL) <+
* next address <-+
*/
/* Iteration 1 */
j = ctx->a[0] & 0x1FFFF0;
aesb_single_round((const uint8_t*) &memory[j], (uint8_t *) ctx->c, (const uint8_t *) ctx->a);
xor_blocks_dst((const uint8_t*) ctx->c, (const uint8_t*) ctx->b, (uint8_t*) &memory[j]);
/* Iteration 2 */
mul_sum_xor_dst((const uint8_t*) ctx->c, (uint8_t*) ctx->a, (uint8_t*) &memory[ctx->c[0] & 0x1FFFF0]);
/* Iteration 3 */
j = ctx->a[0] & 0x1FFFF0;
aesb_single_round(&memory[j], (uint8_t *) ctx->b, (uint8_t *) ctx->a);
xor_blocks_dst((const uint8_t*) ctx->b, (const uint8_t*) ctx->c, (uint8_t*) &memory[j]);
/* Iteration 4 */
mul_sum_xor_dst((const uint8_t*) ctx->b, (uint8_t*) ctx->a, (uint8_t*) &memory[ctx->b[0] & 0x1FFFF0]);
}
memcpy(ctx->text, ctx->state.init, INIT_SIZE_BYTE);
oaes_key_import_data(aes_ctx, &ctx->state.hs.b[32], AES_KEY_SIZE);
for (i = 0; likely(i < MEMORY); i += INIT_SIZE_BYTE) {
xor_blocks(&ctx->text[0 * AES_BLOCK_SIZE], &memory[i + 0 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[0 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[1 * AES_BLOCK_SIZE], &memory[i + 1 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[1 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[2 * AES_BLOCK_SIZE], &memory[i + 2 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[2 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[3 * AES_BLOCK_SIZE], &memory[i + 3 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[3 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[4 * AES_BLOCK_SIZE], &memory[i + 4 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[4 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[5 * AES_BLOCK_SIZE], &memory[i + 5 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[5 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[6 * AES_BLOCK_SIZE], &memory[i + 6 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[6 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
xor_blocks(&ctx->text[7 * AES_BLOCK_SIZE], &memory[i + 7 * AES_BLOCK_SIZE]);
aesb_pseudo_round_mut(&ctx->text[7 * AES_BLOCK_SIZE], aes_ctx->key->exp_data);
}
memcpy(ctx->state.init, ctx->text, INIT_SIZE_BYTE);
keccakf((uint64_t *) &ctx->state.hs, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
oaes_free((OAES_CTX **) &aes_ctx);
}

248
algo/cryptonight/cryptonight_av4_softaes.c Normal file
View file

@ -0,0 +1,248 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2017 fireice-uk <https://github.com/fireice-uk>
* Copyright 2016-2017 XMRig <support@xmrig.com>
*
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <x86intrin.h>
#include <string.h>
#include "cryptonight.h"
#include "crypto/c_keccak.h"
__m128i soft_aesenc(__m128i in, __m128i key);
__m128i soft_aeskeygenassist(__m128i key, uint8_t rcon);
#ifdef __GNUC__
static inline uint64_t _umul128(uint64_t a, uint64_t b, uint64_t* hi)
{
unsigned __int128 r = (unsigned __int128)a * (unsigned __int128)b;
*hi = r >> 64;
return (uint64_t)r;
}
#endif
// This will shift and xor tmp1 into itself as 4 32-bit vals such as
// sl_xor(a1 a2 a3 a4) = a1 (a2^a1) (a3^a2^a1) (a4^a3^a2^a1)
static inline __m128i sl_xor(__m128i tmp1)
{
__m128i tmp4;
tmp4 = _mm_slli_si128(tmp1, 0x04);
tmp1 = _mm_xor_si128(tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
tmp1 = _mm_xor_si128(tmp1, tmp4);
tmp4 = _mm_slli_si128(tmp4, 0x04);
tmp1 = _mm_xor_si128(tmp1, tmp4);
return tmp1;
}
static inline void aes_genkey_sub(__m128i* xout0, __m128i* xout2, uint8_t rcon)
{
__m128i xout1 = soft_aeskeygenassist(*xout2, rcon);
xout1 = _mm_shuffle_epi32(xout1, 0xFF); // see PSHUFD, set all elems to 4th elem
*xout0 = sl_xor(*xout0);
*xout0 = _mm_xor_si128(*xout0, xout1);
xout1 = soft_aeskeygenassist(*xout0, 0x00);
xout1 = _mm_shuffle_epi32(xout1, 0xAA); // see PSHUFD, set all elems to 3rd elem
*xout2 = sl_xor(*xout2);
*xout2 = _mm_xor_si128(*xout2, xout1);
}
static inline void aes_genkey(const __m128i* memory, __m128i* k0, __m128i* k1, __m128i* k2, __m128i* k3, __m128i* k4, __m128i* k5, __m128i* k6, __m128i* k7, __m128i* k8, __m128i* k9)
{
__m128i xout0 = _mm_load_si128(memory);
__m128i xout2 = _mm_load_si128(memory + 1);
*k0 = xout0;
*k1 = xout2;
aes_genkey_sub(&xout0, &xout2, 0x1);
*k2 = xout0;
*k3 = xout2;
aes_genkey_sub(&xout0, &xout2, 0x2);
*k4 = xout0;
*k5 = xout2;
aes_genkey_sub(&xout0, &xout2, 0x4);
*k6 = xout0;
*k7 = xout2;
aes_genkey_sub(&xout0, &xout2, 0x8);
*k8 = xout0;
*k9 = xout2;
}
static inline void aes_round(__m128i key, __m128i* x0, __m128i* x1, __m128i* x2, __m128i* x3, __m128i* x4, __m128i* x5, __m128i* x6, __m128i* x7)
{
*x0 = soft_aesenc(*x0, key);
*x1 = soft_aesenc(*x1, key);
*x2 = soft_aesenc(*x2, key);
*x3 = soft_aesenc(*x3, key);
*x4 = soft_aesenc(*x4, key);
*x5 = soft_aesenc(*x5, key);
*x6 = soft_aesenc(*x6, key);
*x7 = soft_aesenc(*x7, key);
}
static inline void cn_explode_scratchpad(const __m128i* input, __m128i* output)
{
// This is more than we have registers, compiler will assign 2 keys on the stack
__m128i xin0, xin1, xin2, xin3, xin4, xin5, xin6, xin7;
__m128i k0, k1, k2, k3, k4, k5, k6, k7, k8, k9;
aes_genkey(input, &k0, &k1, &k2, &k3, &k4, &k5, &k6, &k7, &k8, &k9);
xin0 = _mm_load_si128(input + 4);
xin1 = _mm_load_si128(input + 5);
xin2 = _mm_load_si128(input + 6);
xin3 = _mm_load_si128(input + 7);
xin4 = _mm_load_si128(input + 8);
xin5 = _mm_load_si128(input + 9);
xin6 = _mm_load_si128(input + 10);
xin7 = _mm_load_si128(input + 11);
for (size_t i = 0; i < MEMORY / sizeof(__m128i); i += 8) {
aes_round(k0, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k1, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k2, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k3, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k4, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k5, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k6, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k7, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k8, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
aes_round(k9, &xin0, &xin1, &xin2, &xin3, &xin4, &xin5, &xin6, &xin7);
_mm_store_si128(output + i + 0, xin0);
_mm_store_si128(output + i + 1, xin1);
_mm_store_si128(output + i + 2, xin2);
_mm_store_si128(output + i + 3, xin3);
_mm_prefetch((const char*)output + i + 0, _MM_HINT_T2);
_mm_store_si128(output + i + 4, xin4);
_mm_store_si128(output + i + 5, xin5);
_mm_store_si128(output + i + 6, xin6);
_mm_store_si128(output + i + 7, xin7);
_mm_prefetch((const char*)output + i + 4, _MM_HINT_T2);
}
}
static inline void cn_implode_scratchpad(const __m128i* input, __m128i* output)
{
// This is more than we have registers, compiler will assign 2 keys on the stack
__m128i xout0, xout1, xout2, xout3, xout4, xout5, xout6, xout7;
__m128i k0, k1, k2, k3, k4, k5, k6, k7, k8, k9;
aes_genkey(output + 2, &k0, &k1, &k2, &k3, &k4, &k5, &k6, &k7, &k8, &k9);
xout0 = _mm_load_si128(output + 4);
xout1 = _mm_load_si128(output + 5);
xout2 = _mm_load_si128(output + 6);
xout3 = _mm_load_si128(output + 7);
xout4 = _mm_load_si128(output + 8);
xout5 = _mm_load_si128(output + 9);
xout6 = _mm_load_si128(output + 10);
xout7 = _mm_load_si128(output + 11);
for (size_t i = 0; i < MEMORY / sizeof(__m128i); i += 8)
{
_mm_prefetch((const char*)input + i + 0, _MM_HINT_NTA);
xout0 = _mm_xor_si128(_mm_load_si128(input + i + 0), xout0);
xout1 = _mm_xor_si128(_mm_load_si128(input + i + 1), xout1);
xout2 = _mm_xor_si128(_mm_load_si128(input + i + 2), xout2);
xout3 = _mm_xor_si128(_mm_load_si128(input + i + 3), xout3);
_mm_prefetch((const char*)input + i + 4, _MM_HINT_NTA);
xout4 = _mm_xor_si128(_mm_load_si128(input + i + 4), xout4);
xout5 = _mm_xor_si128(_mm_load_si128(input + i + 5), xout5);
xout6 = _mm_xor_si128(_mm_load_si128(input + i + 6), xout6);
xout7 = _mm_xor_si128(_mm_load_si128(input + i + 7), xout7);
aes_round(k0, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k1, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k2, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k3, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k4, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k5, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k6, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k7, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k8, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
aes_round(k9, &xout0, &xout1, &xout2, &xout3, &xout4, &xout5, &xout6, &xout7);
}
_mm_store_si128(output + 4, xout0);
_mm_store_si128(output + 5, xout1);
_mm_store_si128(output + 6, xout2);
_mm_store_si128(output + 7, xout3);
_mm_store_si128(output + 8, xout4);
_mm_store_si128(output + 9, xout5);
_mm_store_si128(output + 10, xout6);
_mm_store_si128(output + 11, xout7);
}
void cryptonight_av4_softaes(void *restrict output, const void *restrict input, char *restrict memory, struct cryptonight_ctx *restrict ctx)
{
uint64_t* state = ctx->state.hs.w;
keccak((const uint8_t *) input, 76, (uint8_t *) state, 200);
cn_explode_scratchpad((__m128i*) state, (__m128i*) memory);
uint64_t a[2] __attribute((aligned(16))) = { state[0] ^ state[4], state[1] ^ state[5] };
uint64_t c __attribute((aligned(16)));
uint64_t d[2] __attribute((aligned(16)));
__m128i a_x = _mm_load_si128((__m128i *) &memory[a[0] & 0x1FFFF0]);
__m128i b_x = _mm_set_epi64x(state[3] ^ state[7], state[2] ^ state[6]);
for (size_t i = 0; __builtin_expect(i < 0x80000, 1); i++) {
__m128i c_x = soft_aesenc(a_x, _mm_load_si128((__m128i *) a));
c = _mm_cvtsi128_si64(c_x);
uint64_t *restrict d_ptr = (uint64_t *) &memory[c & 0x1FFFF0];
_mm_store_si128((__m128i *) &memory[a[0] & 0x1FFFF0], _mm_xor_si128(b_x, c_x));
b_x = c_x;
d[0] = d_ptr[0];
d[1] = d_ptr[1];
{
unsigned __int128 res = (unsigned __int128) c * d[0];
d_ptr[0] = a[0] += res >> 64;
d_ptr[1] = a[1] += (uint64_t) res;
}
a[0] ^= d[0];
a[1] ^= d[1];
a_x = _mm_load_si128((__m128i *) &memory[a[0] & 0x1FFFF0]);
}
cn_implode_scratchpad((__m128i*) memory, (__m128i*) state);
keccakf(state, 24);
extra_hashes[ctx->state.hs.b[0] & 3](&ctx->state, 200, output);
}

4
algo/cryptonight/cryptonight_common.c
View file

@ -46,7 +46,7 @@
void cryptonight_av1_aesni32(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av1_aesni32(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
#endif #endif
void cryptonight_av4_legacy(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av4_softaes(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void (*cryptonight_hash_ctx)(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx) = NULL; void (*cryptonight_hash_ctx)(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx) = NULL;
@ -81,7 +81,7 @@ void cryptonight_init(int variant)
#endif #endif
case XMR_VARIANT_LEGACY: case XMR_VARIANT_LEGACY:
cryptonight_hash_ctx = cryptonight_av4_legacy; cryptonight_hash_ctx = cryptonight_av4_softaes;
break; break;
default: default:

10
crypto/aesb.h
View file

@ -1,10 +0,0 @@
#ifndef __AESB_H__
#define __AESB_H__
void aesb_single_round(const uint8_t *in, uint8_t*out, const uint8_t *expandedKey);
void aesb_pseudo_round_mut(uint8_t *val, const uint8_t *expandedKey);
#define fast_aesb_single_round aesb_single_round
#define fast_aesb_pseudo_round_mut aesb_pseudo_round_mut
#endif /* __AESB_H__ */

50
crypto/oaes_config.h
View file

@ -1,50 +0,0 @@
/*
* ---------------------------------------------------------------------------
* OpenAES License
* ---------------------------------------------------------------------------
* Copyright (c) 2012, Nabil S. Al Ramli, www.nalramli.com
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* - Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* - Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
* ---------------------------------------------------------------------------
*/
#ifndef _OAES_CONFIG_H
#define _OAES_CONFIG_H
#ifdef __cplusplus
extern "C" {
#endif
//#ifndef OAES_HAVE_ISAAC
//#define OAES_HAVE_ISAAC 1
//#endif // OAES_HAVE_ISAAC
//#ifndef OAES_DEBUG
//#define OAES_DEBUG 0
//#endif // OAES_DEBUG
#ifdef __cplusplus
}
#endif
#endif // _OAES_CONFIG_H

1417
crypto/oaes_lib.c
View file

File diff suppressed because it is too large Load diff

214
crypto/oaes_lib.h
View file

@ -1,214 +0,0 @@
/*
* ---------------------------------------------------------------------------
* OpenAES License
* ---------------------------------------------------------------------------
* Copyright (c) 2012, Nabil S. Al Ramli, www.nalramli.com
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* - Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* - Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
* ---------------------------------------------------------------------------
*/
#ifndef _OAES_LIB_H
#define _OAES_LIB_H
#include <stdint.h>
#ifdef __cplusplus
extern "C" {
#endif
#ifdef _WIN32
# ifdef OAES_SHARED
# ifdef oaes_lib_EXPORTS
# define OAES_API __declspec(dllexport)
# else
# define OAES_API __declspec(dllimport)
# endif
# else
# define OAES_API
# endif
#else
# define OAES_API
#endif // WIN32
#define OAES_VERSION "0.8.1"
#define OAES_BLOCK_SIZE 16
typedef void OAES_CTX;
typedef enum
{
OAES_RET_FIRST = 0,
OAES_RET_SUCCESS = 0,
OAES_RET_UNKNOWN,
OAES_RET_ARG1,
OAES_RET_ARG2,
OAES_RET_ARG3,
OAES_RET_ARG4,
OAES_RET_ARG5,
OAES_RET_NOKEY,
OAES_RET_MEM,
OAES_RET_BUF,
OAES_RET_HEADER,
OAES_RET_COUNT
} OAES_RET;
/*
* oaes_set_option() takes one of these values for its [option] parameter
* some options accept either an optional or a required [value] parameter
*/
// no option
#define OAES_OPTION_NONE 0
// enable ECB mode, disable CBC mode
#define OAES_OPTION_ECB 1
// enable CBC mode, disable ECB mode
// value is optional, may pass uint8_t iv[OAES_BLOCK_SIZE] to specify
// the value of the initialization vector, iv
#define OAES_OPTION_CBC 2
#ifdef OAES_DEBUG
typedef int ( * oaes_step_cb ) (
const uint8_t state[OAES_BLOCK_SIZE],
const char * step_name,
int step_count,
void * user_data );
// enable state stepping mode
// value is required, must pass oaes_step_cb to receive the state at each step
#define OAES_OPTION_STEP_ON 4
// disable state stepping mode
#define OAES_OPTION_STEP_OFF 8
#endif // OAES_DEBUG
typedef uint16_t OAES_OPTION;
typedef struct _oaes_key
{
size_t data_len;
uint8_t *data;
size_t exp_data_len;
uint8_t *exp_data;
size_t num_keys;
size_t key_base;
} oaes_key;
typedef struct _oaes_ctx
{
#ifdef OAES_HAVE_ISAAC
randctx * rctx;
#endif // OAES_HAVE_ISAAC
#ifdef OAES_DEBUG
oaes_step_cb step_cb;
#endif // OAES_DEBUG
oaes_key * key;
OAES_OPTION options;
uint8_t iv[OAES_BLOCK_SIZE];
} oaes_ctx;
/*
* // usage:
*
* OAES_CTX * ctx = oaes_alloc();
* .
* .
* .
* {
* oaes_gen_key_xxx( ctx );
* {
* oaes_key_export( ctx, _buf, &_buf_len );
* // or
* oaes_key_export_data( ctx, _buf, &_buf_len );\
* }
* }
* // or
* {
* oaes_key_import( ctx, _buf, _buf_len );
* // or
* oaes_key_import_data( ctx, _buf, _buf_len );
* }
* .
* .
* .
* oaes_encrypt( ctx, m, m_len, c, &c_len );
* .
* .
* .
* oaes_decrypt( ctx, c, c_len, m, &m_len );
* .
* .
* .
* oaes_free( &ctx );
*/
OAES_API OAES_CTX * oaes_alloc(void);
OAES_API OAES_RET oaes_free( OAES_CTX ** ctx );
OAES_API OAES_RET oaes_set_option( OAES_CTX * ctx,
OAES_OPTION option, const void * value );
OAES_API OAES_RET oaes_key_gen_128( OAES_CTX * ctx );
OAES_API OAES_RET oaes_key_gen_192( OAES_CTX * ctx );
OAES_API OAES_RET oaes_key_gen_256( OAES_CTX * ctx );
// export key with header information
// set data == NULL to get the required data_len
OAES_API OAES_RET oaes_key_export( OAES_CTX * ctx,
uint8_t * data, size_t * data_len );
// directly export the data from key
// set data == NULL to get the required data_len
OAES_API OAES_RET oaes_key_export_data( OAES_CTX * ctx,
uint8_t * data, size_t * data_len );
// import key with header information
OAES_API OAES_RET oaes_key_import( OAES_CTX * ctx,
const uint8_t * data, size_t data_len );
// directly import data into key
OAES_API OAES_RET oaes_key_import_data( OAES_CTX * ctx,
const uint8_t * data, size_t data_len );
// set c == NULL to get the required c_len
OAES_API OAES_RET oaes_encrypt( OAES_CTX * ctx,
const uint8_t * m, size_t m_len, uint8_t * c, size_t * c_len );
// set m == NULL to get the required m_len
OAES_API OAES_RET oaes_decrypt( OAES_CTX * ctx,
const uint8_t * c, size_t c_len, uint8_t * m, size_t * m_len );
// set buf == NULL to get the required buf_len
OAES_API OAES_RET oaes_sprintf(
char * buf, size_t * buf_len, const uint8_t * data, size_t data_len );
OAES_API OAES_RET oaes_encryption_round( const uint8_t * key, uint8_t * c );
OAES_API OAES_RET oaes_pseudo_encrypt_ecb( OAES_CTX * ctx, uint8_t * c );
#ifdef __cplusplus
}
#endif
#endif // _OAES_LIB_H

120
crypto/aesb.c → crypto/soft_aes.c
View file

@ -1,31 +1,37 @@
/* /*
--------------------------------------------------------------------------- * This program is free software: you can redistribute it and/or modify
Copyright (c) 1998-2013, Brian Gladman, Worcester, UK. All rights reserved. * it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
The redistribution and use of this software (with or without changes) * any later version.
is allowed without the payment of fees or royalties provided that: *
* This program is distributed in the hope that it will be useful,
source code distributions include the above copyright notice, this * but WITHOUT ANY WARRANTY; without even the implied warranty of
list of conditions and the following disclaimer; * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
binary distributions include the above copyright notice, this list *
of conditions and the following disclaimer in their documentation. * You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
This software is provided 'as is' with no explicit or implied warranties *
in respect of its operation, including, but not limited to, correctness * Additional permission under GNU GPL version 3 section 7
and fitness for purpose. *
--------------------------------------------------------------------------- * If you modify this Program, or any covered work, by linking or combining
Issue Date: 20/12/2007 * it with OpenSSL (or a modified version of that library), containing parts
* covered by the terms of OpenSSL License and SSLeay License, the licensors
* of this Program grant you additional permission to convey the resulting work.
*
*/ */
#include <stdint.h> /*
* The orginal author of this AES implementation is Karl Malbrain.
*/
#include "aesb.h" #ifdef __GNUC__
#include <x86intrin.h>
#else
#include <intrin.h>
#endif // __GNUC__
#if defined(__cplusplus) #include <inttypes.h>
extern "C"
{
#endif
#define TABLE_ALIGN 32 #define TABLE_ALIGN 32
#define WPOLY 0x011b #define WPOLY 0x011b
@ -146,25 +152,61 @@ y[3] = (k)[3] ^ (t_fn[0][x[3] & 0xff] ^ t_fn[1][(x[0] >> 8) & 0xff] ^ t_fn[2][(
d_4(uint32_t, t_dec(f,n), sb_data, u0, u1, u2, u3); d_4(uint32_t, t_dec(f,n), sb_data, u0, u1, u2, u3);
inline void aesb_single_round(const uint8_t *restrict in, uint8_t *out, const uint8_t *restrict expandedKey) { __m128i soft_aesenc(__m128i in, __m128i key)
round(((uint32_t*) out), ((uint32_t*) in), ((uint32_t*) expandedKey)); {
uint32_t x0, x1, x2, x3;
x0 = _mm_cvtsi128_si32(in);
x1 = _mm_cvtsi128_si32(_mm_shuffle_epi32(in, 0x55));
x2 = _mm_cvtsi128_si32(_mm_shuffle_epi32(in, 0xAA));
x3 = _mm_cvtsi128_si32(_mm_shuffle_epi32(in, 0xFF));
__m128i out = _mm_set_epi32(
(t_fn[0][x3 & 0xff] ^ t_fn[1][(x0 >> 8) & 0xff] ^ t_fn[2][(x1 >> 16) & 0xff] ^ t_fn[3][x2 >> 24]),
(t_fn[0][x2 & 0xff] ^ t_fn[1][(x3 >> 8) & 0xff] ^ t_fn[2][(x0 >> 16) & 0xff] ^ t_fn[3][x1 >> 24]),
(t_fn[0][x1 & 0xff] ^ t_fn[1][(x2 >> 8) & 0xff] ^ t_fn[2][(x3 >> 16) & 0xff] ^ t_fn[3][x0 >> 24]),
(t_fn[0][x0 & 0xff] ^ t_fn[1][(x1 >> 8) & 0xff] ^ t_fn[2][(x2 >> 16) & 0xff] ^ t_fn[3][x3 >> 24]));
return _mm_xor_si128(out, key);
} }
inline void aesb_pseudo_round_mut(uint8_t *restrict val, const uint8_t *restrict expandedKey) { uint8_t Sbox[256] = { // forward s-box
uint32_t b1[4]; 0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
round(b1, ((uint32_t*) val), ((const uint32_t *) expandedKey)); 0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
round(((uint32_t*) val), b1, ((const uint32_t *) expandedKey) + 1 * N_COLS); 0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
round(b1, ((uint32_t*) val), ((const uint32_t *) expandedKey) + 2 * N_COLS); 0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
round(((uint32_t*) val), b1, ((const uint32_t *) expandedKey) + 3 * N_COLS); 0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
round(b1, ((uint32_t*) val), ((const uint32_t *) expandedKey) + 4 * N_COLS); 0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
round(((uint32_t*) val), b1, ((const uint32_t *) expandedKey) + 5 * N_COLS); 0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
round(b1, ((uint32_t*) val), ((const uint32_t *) expandedKey) + 6 * N_COLS); 0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
round(((uint32_t*) val), b1, ((const uint32_t *) expandedKey) + 7 * N_COLS); 0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
round(b1, ((uint32_t*) val), ((const uint32_t *) expandedKey) + 8 * N_COLS); 0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
round(((uint32_t*) val), b1, ((const uint32_t *) expandedKey) + 9 * N_COLS); 0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16};
static inline void sub_word(uint8_t* key)
{
key[0] = Sbox[key[0]];
key[1] = Sbox[key[1]];
key[2] = Sbox[key[2]];
key[3] = Sbox[key[3]];
} }
#ifdef __clang__
#if defined(__cplusplus) uint32_t _rotr(uint32_t value, uint32_t amount)
{
return (value >> amount) | (value << ((32 - amount) & 31));
} }
#endif #endif
__m128i soft_aeskeygenassist(__m128i key, uint8_t rcon)
{
uint32_t X1 = _mm_cvtsi128_si32(_mm_shuffle_epi32(key, 0x55));
uint32_t X3 = _mm_cvtsi128_si32(_mm_shuffle_epi32(key, 0xFF));
sub_word((uint8_t*)&X1);
sub_word((uint8_t*)&X3);
return _mm_set_epi32(_rotr(X3, 8) ^ rcon, X3,_rotr(X1, 8) ^ rcon, X1);
}

7
test/cryptonight/CMakeLists.txt
View file

@ -1,16 +1,13 @@
set(SOURCES set(SOURCES
../../algo/cryptonight/cryptonight.h ../../algo/cryptonight/cryptonight.h
../../algo/cryptonight/cryptonight_common.c ../../algo/cryptonight/cryptonight_common.c
../../algo/cryptonight/cryptonight_av4_legacy.c ../../algo/cryptonight/cryptonight_av4_softaes.c
../../crypto/c_keccak.c ../../crypto/c_keccak.c
../../crypto/c_blake256.c ../../crypto/c_blake256.c
../../crypto/c_groestl.c ../../crypto/c_groestl.c
../../crypto/c_jh.c ../../crypto/c_jh.c
../../crypto/c_skein.c ../../crypto/c_skein.c
../../crypto/oaes_config.h ../../crypto/soft_aes.c
../../crypto/oaes_lib.h
../../crypto/oaes_lib.c
../../crypto/aesb.c
) )
if (CMAKE_SIZEOF_VOID_P EQUAL 8) if (CMAKE_SIZEOF_VOID_P EQUAL 8)

4
test/cryptonight/cryptonight.c
View file

@ -7,7 +7,7 @@
void cryptonight_av1_aesni(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av1_aesni(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av2_aesni_wolf(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av2_aesni_wolf(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av3_aesni_bmi2(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av3_aesni_bmi2(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av4_legacy(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av4_softaes(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av5_aesni_stak(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av5_aesni_stak(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
void cryptonight_av6_aesni_experimental(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx); void cryptonight_av6_aesni_experimental(void* output, const void* input, const char *memory, struct cryptonight_ctx* ctx);
@ -117,7 +117,7 @@ void test_cryptonight_av4_should_CalcHash(void)
uint8_t *memory = (uint8_t *) malloc(MEMORY); uint8_t *memory = (uint8_t *) malloc(MEMORY);
struct cryptonight_ctx *ctx = (struct cryptonight_ctx*)malloc(sizeof(struct cryptonight_ctx)); struct cryptonight_ctx *ctx = (struct cryptonight_ctx*)malloc(sizeof(struct cryptonight_ctx));
cryptonight_av4_legacy(&hash, data, memory, ctx); cryptonight_av4_softaes(&hash, data, memory, ctx);
free(memory); free(memory);
free(ctx); free(ctx);