aaru-dps/Aaru.Checksums.Native
1
0
Fork 0
mirror of https://github.com/aaru-dps/Aaru.Checksums.Native.git synced 2025-12-16 19:24:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

Refactor code.

Browse Source
This commit is contained in:
Natalia Portillo
2021-10-13 03:46:47 +01:00
parent ca65d12c7e
commit 58f97a93f3
10 changed files with 103 additions and 106 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
adler32_neon.c
View File

@@ -112,14 +112,14 @@ TARGET_WITH_SIMD void adler32_neon(uint16_t* sum1, uint16_t* sum2, const uint8_t
* Multiply-add bytes by [ 32, 31, 30, ... ] for s2. * Multiply-add bytes by [ 32, 31, 30, ... ] for s2.
*/ */
#ifdef _WIN32 #ifdef _WIN32
v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_1), neon_ld1m_16((uint16_t[]) {32, 31, 30, 29})); v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_1), neon_ld1m_16((uint16_t[]){32, 31, 30, 29}));
v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_1), neon_ld1m_16((uint16_t[]){28, 27, 26, 25})); v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_1), neon_ld1m_16((uint16_t[]){28, 27, 26, 25}));
v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_2), neon_ld1m_16((uint16_t[]) {24, 23, 22, 21})); v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_2), neon_ld1m_16((uint16_t[]){24, 23, 22, 21}));
v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_2), neon_ld1m_16((uint16_t[]) {20, 19, 18, 17})); v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_2), neon_ld1m_16((uint16_t[]){20, 19, 18, 17}));
v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_3), neon_ld1m_16((uint16_t[]) {16, 15, 14, 13})); v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_3), neon_ld1m_16((uint16_t[]){16, 15, 14, 13}));
v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_3), neon_ld1m_16((uint16_t[]) {12, 11, 10, 9})); v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_3), neon_ld1m_16((uint16_t[]){12, 11, 10, 9}));
v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_4), neon_ld1m_16((uint16_t[]) {8, 7, 6, 5})); v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_4), neon_ld1m_16((uint16_t[]){8, 7, 6, 5}));
v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_4), neon_ld1m_16((uint16_t[]) {4, 3, 2, 1})); v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_4), neon_ld1m_16((uint16_t[]){4, 3, 2, 1}));
#else #else
v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_1), (uint16x4_t){32, 31, 30, 29}); v_s2 = vmlal_u16(v_s2, vget_low_u16(v_column_sum_1), (uint16x4_t){32, 31, 30, 29});
v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_1), (uint16x4_t){28, 27, 26, 25}); v_s2 = vmlal_u16(v_s2, vget_high_u16(v_column_sum_1), (uint16x4_t){28, 27, 26, 25});

4
crc32_vmull.c
View File

@@ -26,13 +26,13 @@
#if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM) #if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM)
#include <arm_neon.h> #include <arm_neon.h>
#include <stdint.h>
#include <stddef.h> #include <stddef.h>
#include <stdint.h>
#include "library.h" #include "library.h"
#include "arm_vmull.h"
#include "crc32.h" #include "crc32.h"
#include "crc32_simd.h" #include "crc32_simd.h"
#include "arm_vmull.h"
/* /*
* somewhat surprisingly the "naive" way of doing this, ie. with a flag and a cond. branch, * somewhat surprisingly the "naive" way of doing this, ie. with a flag and a cond. branch,

14
crc64_vmull.c
View File

@@ -1,16 +1,16 @@
/* /*
* This file is part of the Aaru Data Preservation Suite. * This file is part of the Aaru Data Preservation Suite.
* Copyright (c) 2019-2021 Natalia Portillo. * Copyright (c) 2019-2021 Natalia Portillo.
* *
* This file is under the public domain: * This file is under the public domain:
* https://github.com/rawrunprotected/crc * https://github.com/rawrunprotected/crc
*/ */
#if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM) #if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM)
#include <arm_neon.h> #include <arm_neon.h>
#include <stdint.h>
#include <stddef.h> #include <stddef.h>
#include <stdint.h>
#include "library.h" #include "library.h"
#include "arm_vmull.h" #include "arm_vmull.h"

32
library.h
View File

@@ -1,20 +1,20 @@
/* /*
* This file is part of the Aaru Data Preservation Suite. * This file is part of the Aaru Data Preservation Suite.
* Copyright (c) 2019-2021 Natalia Portillo. * Copyright (c) 2019-2021 Natalia Portillo.
* *
* This library is free software; you can redistribute it and/or modify * This library is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as * it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of the * published by the Free Software Foundation; either version 2.1 of the
* License, or (at your option) any later version. * License, or (at your option) any later version.
* *
* This library is distributed in the hope that it will be useful, but * This library is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of * WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details. * Lesser General Public License for more details.
* *
* You should have received a copy of the GNU Lesser General Public * You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>. * License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/ */
#ifndef AARU_CHECKSUMS_NATIVE_LIBRARY_H #ifndef AARU_CHECKSUMS_NATIVE_LIBRARY_H
#define AARU_CHECKSUMS_NATIVE_LIBRARY_H #define AARU_CHECKSUMS_NATIVE_LIBRARY_H

47
simd.c
View File

@@ -1,20 +1,20 @@
/* /*
* This file is part of the Aaru Data Preservation Suite. * This file is part of the Aaru Data Preservation Suite.
* Copyright (c) 2019-2021 Natalia Portillo. * Copyright (c) 2019-2021 Natalia Portillo.
* *
* This library is free software; you can redistribute it and/or modify * This library is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as * it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of the * published by the Free Software Foundation; either version 2.1 of the
* License, or (at your option) any later version. * License, or (at your option) any later version.
* *
* This library is distributed in the hope that it will be useful, but * This library is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of * WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details. * Lesser General Public License for more details.
* *
* You should have received a copy of the GNU Lesser General Public * You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>. * License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/ */
#include "library.h" #include "library.h"
#include "simd.h" #include "simd.h"
@@ -112,6 +112,7 @@ int have_avx2(void)
#if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM) #if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM)
#if defined(_WIN32) #if defined(_WIN32)
#include <windows.h> #include <windows.h>
#include <processthreadsapi.h> #include <processthreadsapi.h>
#elif defined(__APPLE__) #elif defined(__APPLE__)
#include <sys/sysctl.h> #include <sys/sysctl.h>
@@ -120,16 +121,14 @@ int have_avx2(void)
#endif #endif
#endif #endif
#if (defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM)) && defined(__APPLE__) #if(defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM)) && defined(__APPLE__)
int have_neon_apple() int have_neon_apple()
{ {
int value; int value;
size_t len = sizeof(int); size_t len = sizeof(int);
int ret= int ret = sysctlbyname("hw.optional.neon", &value, &len, NULL, 0);
sysctlbyname("hw.optional.neon", &value, &len, NULL, 0);
if(ret != 0) if(ret != 0) return 0;
return 0;
return value == 1; return value == 1;
} }
@@ -138,11 +137,9 @@ int have_crc32_apple()
{ {
int value; int value;
size_t len = sizeof(int); size_t len = sizeof(int);
int ret= int ret = sysctlbyname("hw.optional.crc32", &value, &len, NULL, 0);
sysctlbyname("hw.optional.crc32", &value, &len, NULL, 0);
if(ret != 0) if(ret != 0) return 0;
return 0;
return value == 1; return value == 1;
} }

32
simd.h
View File

@@ -1,20 +1,20 @@
/* /*
* This file is part of the Aaru Data Preservation Suite. * This file is part of the Aaru Data Preservation Suite.
* Copyright (c) 2019-2021 Natalia Portillo. * Copyright (c) 2019-2021 Natalia Portillo.
* *
* This library is free software; you can redistribute it and/or modify * This library is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as * it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of the * published by the Free Software Foundation; either version 2.1 of the
* License, or (at your option) any later version. * License, or (at your option) any later version.
* *
* This library is distributed in the hope that it will be useful, but * This library is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of * WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details. * Lesser General Public License for more details.
* *
* You should have received a copy of the GNU Lesser General Public * You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>. * License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/ */
#ifndef AARU_CHECKSUMS_NATIVE_SIMD_H #ifndef AARU_CHECKSUMS_NATIVE_SIMD_H
#define AARU_CHECKSUMS_NATIVE_SIMD_H #define AARU_CHECKSUMS_NATIVE_SIMD_H

26
spamsum.c
View File

@@ -141,7 +141,7 @@ AARU_LOCAL inline void roll_hash(spamsum_ctx* ctx, uint8_t c)
AARU_LOCAL inline void fuzzy_try_reduce_blockhash(spamsum_ctx* ctx) AARU_LOCAL inline void fuzzy_try_reduce_blockhash(spamsum_ctx* ctx)
{ {
//assert(ctx->bh_start < ctx->bh_end); // assert(ctx->bh_start < ctx->bh_end);
if(ctx->bh_end - ctx->bh_start < 2) /* Need at least two working hashes. */ if(ctx->bh_end - ctx->bh_start < 2) /* Need at least two working hashes. */
return; return;
@@ -163,7 +163,7 @@ AARU_LOCAL inline void fuzzy_try_fork_blockhash(spamsum_ctx* ctx)
{ {
if(ctx->bh_end >= NUM_BLOCKHASHES) return; if(ctx->bh_end >= NUM_BLOCKHASHES) return;
//assert(ctx->bh_end != 0); // assert(ctx->bh_end != 0);
uint32_t obh = ctx->bh_end - 1; uint32_t obh = ctx->bh_end - 1;
uint32_t nbh = ctx->bh_end; uint32_t nbh = ctx->bh_end;
@@ -184,7 +184,7 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
if(!result) return -1; if(!result) return -1;
/* Verify that our elimination was not overeager. */ /* Verify that our elimination was not overeager. */
//assert(bi == 0 || (uint64_t)SSDEEP_BS(bi) / 2 * SPAMSUM_LENGTH < ctx->total_size); // assert(bi == 0 || (uint64_t)SSDEEP_BS(bi) / 2 * SPAMSUM_LENGTH < ctx->total_size);
/* Initial blocksize guess. */ /* Initial blocksize guess. */
while((uint64_t)SSDEEP_BS(bi) * SPAMSUM_LENGTH < ctx->total_size) while((uint64_t)SSDEEP_BS(bi) * SPAMSUM_LENGTH < ctx->total_size)
@@ -203,21 +203,21 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
while(bi > ctx->bh_start && ctx->bh[bi].d_len < SPAMSUM_LENGTH / 2) --bi; while(bi > ctx->bh_start && ctx->bh[bi].d_len < SPAMSUM_LENGTH / 2) --bi;
//assert(!(bi > 0 && ctx->bh[bi].d_len < SPAMSUM_LENGTH / 2)); // assert(!(bi > 0 && ctx->bh[bi].d_len < SPAMSUM_LENGTH / 2));
int i = snprintf((char*)result, (size_t)remain, "%lu:", (unsigned long)SSDEEP_BS(bi)); int i = snprintf((char*)result, (size_t)remain, "%lu:", (unsigned long)SSDEEP_BS(bi));
if(i <= 0) /* Maybe snprintf has set errno here? */ if(i <= 0) /* Maybe snprintf has set errno here? */
return -1; return -1;
//assert(i < remain); // assert(i < remain);
remain -= i; remain -= i;
result += i; result += i;
i = (int)ctx->bh[bi].d_len; i = (int)ctx->bh[bi].d_len;
//assert(i <= remain); // assert(i <= remain);
memcpy(result, ctx->bh[bi].digest, (size_t)i); memcpy(result, ctx->bh[bi].digest, (size_t)i);
result += i; result += i;
@@ -225,7 +225,7 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
if(h != 0) if(h != 0)
{ {
//assert(remain > 0); // assert(remain > 0);
*result = b64[ctx->bh[bi].h % 64]; *result = b64[ctx->bh[bi].h % 64];
@@ -237,7 +237,7 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
} }
else if(ctx->bh[bi].digest[i] != 0) else if(ctx->bh[bi].digest[i] != 0)
{ {
//assert(remain > 0); // assert(remain > 0);
*result = ctx->bh[bi].digest[i]; *result = ctx->bh[bi].digest[i];
@@ -248,7 +248,7 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
} }
} }
//assert(remain > 0); // assert(remain > 0);
*result++ = ':'; *result++ = ':';
--remain; --remain;
@@ -267,7 +267,7 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
if(h != 0) if(h != 0)
{ {
//assert(remain > 0); // assert(remain > 0);
h = ctx->bh[bi].half_h; h = ctx->bh[bi].half_h;
*result = b64[h % 64]; *result = b64[h % 64];
@@ -284,7 +284,7 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
if(i != 0) if(i != 0)
{ {
//assert(remain > 0); // assert(remain > 0);
*result = (uint8_t)i; *result = (uint8_t)i;
@@ -298,9 +298,9 @@ AARU_EXPORT int AARU_CALL spamsum_final(spamsum_ctx* ctx, uint8_t* result)
} }
else if(h != 0) else if(h != 0)
{ {
//assert(ctx->bh[bi].d_len == 0); // assert(ctx->bh[bi].d_len == 0);
//assert(remain > 0); // assert(remain > 0);
*result++ = b64[ctx->bh[bi].h % 64]; *result++ = b64[ctx->bh[bi].h % 64];
/* No need to bother with FUZZY_FLAG_ELIMSEQ, because this /* No need to bother with FUZZY_FLAG_ELIMSEQ, because this