476 lines
13 KiB
C
476 lines
13 KiB
C
// Copyright (c) 2014, The Monero Project
|
|
//
|
|
// All rights reserved.
|
|
//
|
|
// Redistribution and use in source and binary forms, with or without modification, are
|
|
// permitted provided that the following conditions are met:
|
|
//
|
|
// 1. Redistributions of source code must retain the above copyright notice, this list of
|
|
// conditions and the following disclaimer.
|
|
//
|
|
// 2. 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.
|
|
//
|
|
// 3. Neither the name of the copyright holder nor the names of its contributors may be
|
|
// used to endorse or promote products derived from this software without specific
|
|
// prior written permission.
|
|
//
|
|
// 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.
|
|
//
|
|
// Parts of this file are originally copyright (c) 2012-2013 The Cryptonote developers
|
|
|
|
#include <assert.h>
|
|
#include <stddef.h>
|
|
#include <stdint.h>
|
|
#include <string.h>
|
|
|
|
#include "common/int-util.h"
|
|
#include "hash-ops.h"
|
|
#include "oaes_lib.h"
|
|
|
|
#include <emmintrin.h>
|
|
|
|
#if defined(_MSC_VER)
|
|
#include <intrin.h>
|
|
#include <windows.h>
|
|
#define STATIC
|
|
#define INLINE __inline
|
|
#if !defined(RDATA_ALIGN16)
|
|
#define RDATA_ALIGN16 __declspec(align(16))
|
|
#endif
|
|
#else
|
|
#include <wmmintrin.h>
|
|
#include <sys/mman.h>
|
|
#define STATIC static
|
|
#define INLINE inline
|
|
#if !defined(RDATA_ALIGN16)
|
|
#define RDATA_ALIGN16 __attribute__ ((aligned(16)))
|
|
#endif
|
|
#endif
|
|
|
|
#if defined(__INTEL_COMPILER)
|
|
#define ASM __asm__
|
|
#elif !defined(_MSC_VER)
|
|
#define ASM __asm__
|
|
#else
|
|
#define ASM __asm
|
|
#endif
|
|
|
|
#define MEMORY (1 << 21) // 2MB scratchpad
|
|
#define ITER (1 << 20)
|
|
#define AES_BLOCK_SIZE 16
|
|
#define AES_KEY_SIZE 32
|
|
#define INIT_SIZE_BLK 8
|
|
#define INIT_SIZE_BYTE (INIT_SIZE_BLK * AES_BLOCK_SIZE)
|
|
#define TOTALBLOCKS (MEMORY / AES_BLOCK_SIZE)
|
|
|
|
#define U64(x) ((uint64_t *) (x))
|
|
#define R128(x) ((__m128i *) (x))
|
|
|
|
#define state_index(x) (((*((uint64_t *)x) >> 4) & (TOTALBLOCKS - 1)) << 4)
|
|
#if defined(_MSC_VER)
|
|
#if !defined(_WIN64)
|
|
#define __mul() lo = mul128(c[0], b[0], &hi);
|
|
#else
|
|
#define __mul() lo = _umul128(c[0], b[0], &hi);
|
|
#endif
|
|
#else
|
|
#if defined(__x86_64__)
|
|
#define __mul() ASM("mulq %3\n\t" : "=d"(hi), "=a"(lo) : "%a" (c[0]), "rm" (b[0]) : "cc");
|
|
#else
|
|
#define __mul() lo = mul128(c[0], b[0], &hi);
|
|
#endif
|
|
#endif
|
|
|
|
#define pre_aes() \
|
|
j = state_index(a); \
|
|
_c = _mm_load_si128(R128(&hp_state[j])); \
|
|
_a = _mm_load_si128(R128(a)); \
|
|
|
|
// dga's optimized scratchpad twiddling
|
|
#define post_aes() \
|
|
_mm_store_si128(R128(c), _c); \
|
|
_b = _mm_xor_si128(_b, _c); \
|
|
_mm_store_si128(R128(&hp_state[j]), _b); \
|
|
j = state_index(c); \
|
|
p = U64(&hp_state[j]); \
|
|
b[0] = p[0]; b[1] = p[1]; \
|
|
__mul(); \
|
|
a[0] += hi; a[1] += lo; \
|
|
p = U64(&hp_state[j]); \
|
|
p[0] = a[0]; p[1] = a[1]; \
|
|
a[0] ^= b[0]; a[1] ^= b[1]; \
|
|
_b = _c; \
|
|
|
|
#if defined(_MSC_VER)
|
|
#define THREADV __declspec(thread)
|
|
#else
|
|
#define THREADV __thread
|
|
#endif
|
|
|
|
extern int aesb_single_round(const uint8_t *in, uint8_t*out, const uint8_t *expandedKey);
|
|
extern int aesb_pseudo_round(const uint8_t *in, uint8_t *out, const uint8_t *expandedKey);
|
|
|
|
#pragma pack(push, 1)
|
|
union cn_slow_hash_state
|
|
{
|
|
union hash_state hs;
|
|
struct
|
|
{
|
|
uint8_t k[64];
|
|
uint8_t init[INIT_SIZE_BYTE];
|
|
};
|
|
};
|
|
#pragma pack(pop)
|
|
|
|
THREADV uint8_t *hp_state = NULL;
|
|
THREADV int hp_allocated = 0;
|
|
|
|
#if defined(_MSC_VER)
|
|
#define cpuid(info,x) __cpuidex(info,x,0)
|
|
#else
|
|
void cpuid(int CPUInfo[4], int InfoType)
|
|
{
|
|
ASM __volatile__
|
|
(
|
|
"cpuid":
|
|
"=a" (CPUInfo[0]),
|
|
"=b" (CPUInfo[1]),
|
|
"=c" (CPUInfo[2]),
|
|
"=d" (CPUInfo[3]) :
|
|
"a" (InfoType), "c" (0)
|
|
);
|
|
}
|
|
#endif
|
|
|
|
STATIC INLINE void xor_blocks(uint8_t *a, const uint8_t *b)
|
|
{
|
|
U64(a)[0] ^= U64(b)[0];
|
|
U64(a)[1] ^= U64(b)[1];
|
|
}
|
|
|
|
STATIC INLINE int check_aes_hw(void)
|
|
{
|
|
int cpuid_results[4];
|
|
static int supported = -1;
|
|
|
|
if(supported >= 0)
|
|
return supported;
|
|
|
|
cpuid(cpuid_results,1);
|
|
return supported = cpuid_results[2] & (1 << 25);
|
|
}
|
|
|
|
STATIC INLINE void aes_256_assist1(__m128i* t1, __m128i * t2)
|
|
{
|
|
__m128i t4;
|
|
*t2 = _mm_shuffle_epi32(*t2, 0xff);
|
|
t4 = _mm_slli_si128(*t1, 0x04);
|
|
*t1 = _mm_xor_si128(*t1, t4);
|
|
t4 = _mm_slli_si128(t4, 0x04);
|
|
*t1 = _mm_xor_si128(*t1, t4);
|
|
t4 = _mm_slli_si128(t4, 0x04);
|
|
*t1 = _mm_xor_si128(*t1, t4);
|
|
*t1 = _mm_xor_si128(*t1, *t2);
|
|
}
|
|
|
|
STATIC INLINE void aes_256_assist2(__m128i* t1, __m128i * t3)
|
|
{
|
|
__m128i t2, t4;
|
|
t4 = _mm_aeskeygenassist_si128(*t1, 0x00);
|
|
t2 = _mm_shuffle_epi32(t4, 0xaa);
|
|
t4 = _mm_slli_si128(*t3, 0x04);
|
|
*t3 = _mm_xor_si128(*t3, t4);
|
|
t4 = _mm_slli_si128(t4, 0x04);
|
|
*t3 = _mm_xor_si128(*t3, t4);
|
|
t4 = _mm_slli_si128(t4, 0x04);
|
|
*t3 = _mm_xor_si128(*t3, t4);
|
|
*t3 = _mm_xor_si128(*t3, t2);
|
|
}
|
|
|
|
STATIC INLINE void aes_expand_key(const uint8_t *key, uint8_t *expandedKey)
|
|
{
|
|
__m128i *ek = R128(expandedKey);
|
|
__m128i t1, t2, t3;
|
|
|
|
t1 = _mm_loadu_si128(R128(key));
|
|
t3 = _mm_loadu_si128(R128(key + 16));
|
|
|
|
ek[0] = t1;
|
|
ek[1] = t3;
|
|
|
|
t2 = _mm_aeskeygenassist_si128(t3, 0x01);
|
|
aes_256_assist1(&t1, &t2);
|
|
ek[2] = t1;
|
|
aes_256_assist2(&t1, &t3);
|
|
ek[3] = t3;
|
|
|
|
t2 = _mm_aeskeygenassist_si128(t3, 0x02);
|
|
aes_256_assist1(&t1, &t2);
|
|
ek[4] = t1;
|
|
aes_256_assist2(&t1, &t3);
|
|
ek[5] = t3;
|
|
|
|
t2 = _mm_aeskeygenassist_si128(t3, 0x04);
|
|
aes_256_assist1(&t1, &t2);
|
|
ek[6] = t1;
|
|
aes_256_assist2(&t1, &t3);
|
|
ek[7] = t3;
|
|
|
|
t2 = _mm_aeskeygenassist_si128(t3, 0x08);
|
|
aes_256_assist1(&t1, &t2);
|
|
ek[8] = t1;
|
|
aes_256_assist2(&t1, &t3);
|
|
ek[9] = t3;
|
|
|
|
t2 = _mm_aeskeygenassist_si128(t3, 0x10);
|
|
aes_256_assist1(&t1, &t2);
|
|
ek[10] = t1;
|
|
}
|
|
|
|
STATIC INLINE void aes_pseudo_round(const uint8_t *in, uint8_t *out,
|
|
const uint8_t *expandedKey, int nblocks)
|
|
{
|
|
__m128i *k = R128(expandedKey);
|
|
__m128i d;
|
|
int i;
|
|
|
|
for(i = 0; i < nblocks; i++)
|
|
{
|
|
d = _mm_loadu_si128(R128(in + i * AES_BLOCK_SIZE));
|
|
d = _mm_aesenc_si128(d, *R128(&k[0]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[1]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[2]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[3]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[4]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[5]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[6]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[7]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[8]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[9]));
|
|
_mm_storeu_si128((R128(out + i * AES_BLOCK_SIZE)), d);
|
|
}
|
|
}
|
|
|
|
STATIC INLINE void aes_pseudo_round_xor(const uint8_t *in, uint8_t *out,
|
|
const uint8_t *expandedKey, const uint8_t *xor, int nblocks)
|
|
{
|
|
__m128i *k = R128(expandedKey);
|
|
__m128i *x = R128(xor);
|
|
__m128i d;
|
|
int i;
|
|
|
|
for(i = 0; i < nblocks; i++)
|
|
{
|
|
d = _mm_loadu_si128(R128(in + i * AES_BLOCK_SIZE));
|
|
d = _mm_xor_si128(d, *R128(x++));
|
|
d = _mm_aesenc_si128(d, *R128(&k[0]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[1]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[2]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[3]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[4]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[5]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[6]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[7]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[8]));
|
|
d = _mm_aesenc_si128(d, *R128(&k[9]));
|
|
_mm_storeu_si128((R128(out + i * AES_BLOCK_SIZE)), d);
|
|
}
|
|
}
|
|
|
|
#if defined(_MSC_VER)
|
|
BOOL SetLockPagesPrivilege(HANDLE hProcess, BOOL bEnable)
|
|
{
|
|
struct
|
|
{
|
|
DWORD count;
|
|
LUID_AND_ATTRIBUTES privilege[1];
|
|
} info;
|
|
|
|
HANDLE token;
|
|
if(!OpenProcessToken(hProcess, TOKEN_ADJUST_PRIVILEGES, &token))
|
|
return FALSE;
|
|
|
|
info.count = 1;
|
|
info.privilege[0].Attributes = bEnable ? SE_PRIVILEGE_ENABLED : 0;
|
|
|
|
if(!LookupPrivilegeValue(NULL, SE_LOCK_MEMORY_NAME, &(info.privilege[0].Luid)))
|
|
return FALSE;
|
|
|
|
if(!AdjustTokenPrivileges(token, FALSE, (PTOKEN_PRIVILEGES) &info, 0, NULL, NULL))
|
|
return FALSE;
|
|
|
|
if (GetLastError() != ERROR_SUCCESS)
|
|
return FALSE;
|
|
|
|
CloseHandle(token);
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
#endif
|
|
|
|
void slow_hash_allocate_state(void)
|
|
{
|
|
int state = 0;
|
|
if(hp_state != NULL)
|
|
return;
|
|
|
|
#if defined(_MSC_VER)
|
|
SetLockPagesPrivilege(GetCurrentProcess(), TRUE);
|
|
hp_state = (uint8_t *) VirtualAlloc(hp_state, MEMORY, MEM_LARGE_PAGES |
|
|
MEM_COMMIT | MEM_RESERVE, PAGE_READWRITE);
|
|
#else
|
|
#if defined(__APPLE__) || defined(__FreeBSD__)
|
|
hp_state = mmap(0, MEMORY, PROT_READ | PROT_WRITE,
|
|
MAP_PRIVATE | MAP_ANON, 0, 0);
|
|
#else
|
|
hp_state = mmap(0, MEMORY, PROT_READ | PROT_WRITE,
|
|
MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB, 0, 0);
|
|
#endif
|
|
if(hp_state == MAP_FAILED)
|
|
hp_state = NULL;
|
|
#endif
|
|
hp_allocated = 1;
|
|
if(hp_state == NULL)
|
|
{
|
|
hp_allocated = 0;
|
|
hp_state = (uint8_t *) malloc(MEMORY);
|
|
}
|
|
}
|
|
|
|
void slow_hash_free_state(void)
|
|
{
|
|
if(hp_state == NULL)
|
|
return;
|
|
|
|
if(!hp_allocated)
|
|
free(hp_state);
|
|
else
|
|
{
|
|
#if defined(_MSC_VER)
|
|
VirtualFree(hp_state, MEMORY, MEM_RELEASE);
|
|
#else
|
|
munmap(hp_state, MEMORY);
|
|
#endif
|
|
}
|
|
|
|
hp_state = NULL;
|
|
hp_allocated = 0;
|
|
}
|
|
|
|
void cn_slow_hash(const void *data, size_t length, char *hash)
|
|
{
|
|
RDATA_ALIGN16 uint8_t expandedKey[240];
|
|
|
|
uint8_t text[INIT_SIZE_BYTE];
|
|
RDATA_ALIGN16 uint64_t a[2];
|
|
RDATA_ALIGN16 uint64_t b[2];
|
|
RDATA_ALIGN16 uint64_t c[2];
|
|
union cn_slow_hash_state state;
|
|
__m128i _a, _b, _c;
|
|
uint64_t hi, lo;
|
|
|
|
size_t i, j;
|
|
uint64_t *p = NULL;
|
|
oaes_ctx *aes_ctx;
|
|
int useAes = check_aes_hw();
|
|
|
|
static void (*const extra_hashes[4])(const void *, size_t, char *) =
|
|
{
|
|
hash_extra_blake, hash_extra_groestl, hash_extra_jh, hash_extra_skein
|
|
};
|
|
|
|
// this isn't supposed to happen, but guard against it for now.
|
|
if(hp_state == NULL)
|
|
slow_hash_allocate_state();
|
|
|
|
hash_process(&state.hs, data, length);
|
|
memcpy(text, state.init, INIT_SIZE_BYTE);
|
|
|
|
if(useAes)
|
|
{
|
|
aes_expand_key(state.hs.b, expandedKey);
|
|
for(i = 0; i < MEMORY / INIT_SIZE_BYTE; i++)
|
|
{
|
|
aes_pseudo_round(text, text, expandedKey, INIT_SIZE_BLK);
|
|
memcpy(&hp_state[i * INIT_SIZE_BYTE], text, INIT_SIZE_BYTE);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
aes_ctx = (oaes_ctx *) oaes_alloc();
|
|
oaes_key_import_data(aes_ctx, state.hs.b, AES_KEY_SIZE);
|
|
for(i = 0; i < MEMORY / INIT_SIZE_BYTE; i++)
|
|
{
|
|
for(j = 0; j < INIT_SIZE_BLK; j++)
|
|
aesb_pseudo_round(&text[AES_BLOCK_SIZE * j], &text[AES_BLOCK_SIZE * j], aes_ctx->key->exp_data);
|
|
|
|
memcpy(&hp_state[i * INIT_SIZE_BYTE], text, INIT_SIZE_BYTE);
|
|
}
|
|
}
|
|
|
|
U64(a)[0] = U64(&state.k[0])[0] ^ U64(&state.k[32])[0];
|
|
U64(a)[1] = U64(&state.k[0])[1] ^ U64(&state.k[32])[1];
|
|
U64(b)[0] = U64(&state.k[16])[0] ^ U64(&state.k[48])[0];
|
|
U64(b)[1] = U64(&state.k[16])[1] ^ U64(&state.k[48])[1];
|
|
|
|
_b = _mm_load_si128(R128(b));
|
|
// this is ugly but the branching affects the loop somewhat so put it outside.
|
|
if(useAes)
|
|
{
|
|
for(i = 0; i < ITER / 2; i++)
|
|
{
|
|
pre_aes();
|
|
_c = _mm_aesenc_si128(_c, _a);
|
|
// post_aes(), optimized scratchpad twiddling (credits to dga)
|
|
post_aes();
|
|
}
|
|
}
|
|
else
|
|
{
|
|
for(i = 0; i < ITER / 2; i++)
|
|
{
|
|
pre_aes();
|
|
aesb_single_round((uint8_t *) &_c, (uint8_t *) &_c, (uint8_t *) &_a);
|
|
post_aes();
|
|
}
|
|
}
|
|
|
|
memcpy(text, state.init, INIT_SIZE_BYTE);
|
|
if(useAes)
|
|
{
|
|
aes_expand_key(&state.hs.b[32], expandedKey);
|
|
for(i = 0; i < MEMORY / INIT_SIZE_BYTE; i++)
|
|
{
|
|
// add the xor to the pseudo round
|
|
aes_pseudo_round_xor(text, text, expandedKey, &hp_state[i * INIT_SIZE_BYTE], INIT_SIZE_BLK);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
oaes_key_import_data(aes_ctx, &state.hs.b[32], AES_KEY_SIZE);
|
|
for(i = 0; i < MEMORY / INIT_SIZE_BYTE; i++)
|
|
{
|
|
for(j = 0; j < INIT_SIZE_BLK; j++)
|
|
{
|
|
xor_blocks(&text[j * AES_BLOCK_SIZE], &hp_state[i * INIT_SIZE_BYTE + j * AES_BLOCK_SIZE]);
|
|
aesb_pseudo_round(&text[AES_BLOCK_SIZE * j], &text[AES_BLOCK_SIZE * j], aes_ctx->key->exp_data);
|
|
}
|
|
}
|
|
oaes_free((OAES_CTX **) &aes_ctx);
|
|
}
|
|
|
|
memcpy(state.init, text, INIT_SIZE_BYTE);
|
|
hash_permutation(&state.hs);
|
|
extra_hashes[state.hs.b[0] & 3](&state, 200, hash);
|
|
}
|