ddidderr/zstd-rs
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Move bitwise builtins into bits.h

Browse Source
This commit is contained in:
Elliot Gorokhovsky
2022-02-14 11:16:03 -05:00
parent 970460f67d
commit db2f4a6532
16 changed files with 250 additions and 400 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/common/bits.h
+212
View File
@@ -0,0 +1,212 @@
/*
* Copyright (c) Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
#ifndef ZSTD_BITS_H
#define ZSTD_BITS_H
#include "mem.h"
MEM_STATIC unsigned ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus */
{
assert(val != 0);
{
# if defined(_MSC_VER) /* Visual */
# if STATIC_BMI2 == 1
return _lzcnt_u32(val)^31;
# else
if (val != 0) {
unsigned long r;
_BitScanReverse(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */
return (unsigned)__builtin_clz (val) ^ 31;
# elif defined(__ICCARM__) /* IAR Intrinsic */
return 31 - __CLZ(val);
# else /* Software version */
static const U32 DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };
U32 v = val;
v |= v >> 1;
v |= v >> 2;
v |= v >> 4;
v |= v >> 8;
v |= v >> 16;
return DeBruijnClz[(v * 0x07C4ACDDU) >> 27];
# endif
}
}
MEM_STATIC unsigned ZSTD_countTrailingZeros32(U32 val)
{
assert(val != 0);
# if defined(_MSC_VER)
if (val != 0) {
unsigned long r;
_BitScanForward(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# elif defined(__GNUC__) && (__GNUC__ >= 3)
return (unsigned)__builtin_ctz(val);
# elif defined(__ICCARM__) /* IAR Intrinsic */
return __CTZ(val);
# else
static const int DeBruijnBytePos[32] = { 0, 1, 28, 2, 29, 14, 24, 3,
30, 22, 20, 15, 25, 17, 4, 8,
31, 27, 13, 23, 21, 19, 16, 7,
26, 12, 18, 6, 11, 5, 10, 9 };
return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
# endif
}
MEM_STATIC unsigned ZSTD_countTrailingZeros64(U64 val)
{
assert(val != 0);
# if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2
return _tzcnt_u64(val);
# else
if (val != 0) {
unsigned long r;
_BitScanForward64(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
if (MEM_32bits()) {
U32 mostSignificantWord = (U32)(val >> 32);
U32 leastSignificantWord = (U32)val;
if (leastSignificantWord == 0) {
return 32 + (unsigned)__builtin_ctz(mostSignificantWord);
} else {
return (unsigned)__builtin_ctz(leastSignificantWord);
}
} else {
return (unsigned)__builtin_ctzll(val);
}
# else
static const int DeBruijnBytePos[64] = { 0, 1, 2, 7, 3, 13, 8, 19,
4, 25, 14, 28, 9, 34, 20, 56,
5, 17, 26, 54, 15, 41, 29, 43,
10, 31, 38, 35, 21, 45, 49, 57,
63, 6, 12, 18, 24, 27, 33, 55,
16, 53, 40, 42, 30, 37, 44, 48,
62, 11, 23, 32, 52, 39, 36, 47,
61, 22, 51, 46, 60, 50, 59, 58 };
return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];
# endif
}
MEM_STATIC unsigned ZSTD_NbCommonBytes(size_t val)
{
if (MEM_isLittleEndian()) {
if (MEM_64bits()) {
# if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2
return _tzcnt_u64(val) >> 3;
# else
if (val != 0) {
unsigned long r;
_BitScanForward64(&r, (U64)val);
return (unsigned)(r >> 3);
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)(__builtin_ctzll((U64)val) >> 3);
# else
static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2,
0, 3, 1, 3, 1, 4, 2, 7,
0, 2, 3, 6, 1, 5, 3, 5,
1, 3, 4, 4, 2, 5, 6, 7,
7, 0, 1, 2, 3, 3, 4, 6,
2, 6, 5, 5, 3, 4, 5, 6,
7, 1, 2, 4, 6, 4, 4, 5,
7, 2, 6, 5, 7, 6, 7, 7 };
return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];
# endif
} else { /* 32 bits */
# if defined(_MSC_VER)
if (val != 0) {
unsigned long r;
_BitScanForward(&r, (U32)val);
return (unsigned)(r >> 3);
} else {
/* Should not reach this code path */
__assume(0);
}
# elif defined(__GNUC__) && (__GNUC__ >= 3)
return (unsigned)(__builtin_ctz((U32)val) >> 3);
# else
static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0,
3, 2, 2, 1, 3, 2, 0, 1,
3, 3, 1, 2, 2, 2, 2, 0,
3, 1, 2, 0, 1, 0, 1, 1 };
return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
# endif
}
} else { /* Big Endian CPU */
if (MEM_64bits()) {
# if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2
return _lzcnt_u64(val) >> 3;
# else
if (val != 0) {
unsigned long r;
_BitScanReverse64(&r, (U64)val);
return (unsigned)(r >> 3);
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)(__builtin_clzll(val) >> 3);
# else
unsigned r;
const unsigned n32 = sizeof(size_t)*4; /* calculate this way due to compiler complaining in 32-bits mode */
if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; }
if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
r += (!val);
return r;
# endif
} else { /* 32 bits */
# if defined(_MSC_VER)
if (val != 0) {
unsigned long r;
_BitScanReverse(&r, (unsigned long)val);
return (unsigned)(r >> 3);
} else {
/* Should not reach this code path */
__assume(0);
}
# elif defined(__GNUC__) && (__GNUC__ >= 3)
return (unsigned)(__builtin_clz((U32)val) >> 3);
# else
unsigned r;
if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
r += (!val);
return r;
# endif
} }
}
#endif /* ZSTD_BITS_H */
lib/common/bitstream.h
+3 -44
View File
@@ -30,6 +30,7 @@ extern "C" {
#include "compiler.h" /* UNLIKELY() */
#include "debug.h" /* assert(), DEBUGLOG(), RAWLOG() */
#include "error_private.h" /* error codes and messages */
#include "bits.h" /* ZSTD_highbit32 */
/*=========================================
@@ -132,48 +133,6 @@ MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC);
MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits);
/* faster, but works only if nbBits >= 1 */
/*-**************************************************************
* Internal functions
****************************************************************/
MEM_STATIC unsigned BIT_highbit32 (U32 val)
{
assert(val != 0);
{
# if defined(_MSC_VER) /* Visual */
# if STATIC_BMI2 == 1
return _lzcnt_u32(val) ^ 31;
# else
if (val != 0) {
unsigned long r;
_BitScanReverse(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */
return __builtin_clz (val) ^ 31;
# elif defined(__ICCARM__) /* IAR Intrinsic */
return 31 - __CLZ(val);
# else /* Software version */
static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29,
11, 14, 16, 18, 22, 25, 3, 30,
8, 12, 20, 28, 15, 17, 24, 7,
19, 27, 23, 6, 26, 5, 4, 31 };
U32 v = val;
v |= v >> 1;
v |= v >> 2;
v |= v >> 4;
v |= v >> 8;
v |= v >> 16;
return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27];
# endif
}
}
/*===== Local Constants =====*/
static const unsigned BIT_mask[] = {
0, 1, 3, 7, 0xF, 0x1F,
@@ -291,7 +250,7 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
bitD->ptr = (const char*)srcBuffer + srcSize - sizeof(bitD->bitContainer);
bitD->bitContainer = MEM_readLEST(bitD->ptr);
{ BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */
bitD->bitsConsumed = lastByte ? 8 - ZSTD_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */
if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ }
} else {
bitD->ptr = bitD->start;
@@ -319,7 +278,7 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
default: break;
}
{ BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;
bitD->bitsConsumed = lastByte ? 8 - ZSTD_highbit32(lastByte) : 0;
if (lastByte == 0) return ERROR(corruption_detected); /* endMark not present */
}
bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize)*8;
lib/common/entropy_common.c
+7 -34
View File
@@ -21,6 +21,7 @@
#include "fse.h"
#define HUF_STATIC_LINKING_ONLY /* HUF_TABLELOG_ABSOLUTEMAX */
#include "huf.h"
#include "bits.h" /* ZSDT_highbit32, ZSTD_countTrailingZeros32 */
/*=== Version ===*/
@@ -38,34 +39,6 @@ const char* HUF_getErrorName(size_t code) { return ERR_getErrorName(code); }
/*-**************************************************************
* FSE NCount encoding-decoding
****************************************************************/
static U32 FSE_ctz(U32 val)
{
assert(val != 0);
{
# if defined(_MSC_VER) /* Visual */
if (val != 0) {
unsigned long r;
_BitScanForward(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */
return __builtin_ctz(val);
# elif defined(__ICCARM__) /* IAR Intrinsic */
return __CTZ(val);
# else /* Software version */
U32 count = 0;
while ((val & 1) == 0) {
val >>= 1;
++count;
}
return count;
# endif
}
}
FORCE_INLINE_TEMPLATE
size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
const void* headerBuffer, size_t hbSize)
@@ -113,7 +86,7 @@ size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigne
* repeat.
* Avoid UB by setting the high bit to 1.
*/
int repeats = FSE_ctz(~bitStream | 0x80000000) >> 1;
int repeats = ZSTD_countTrailingZeros32(~bitStream | 0x80000000) >> 1;
while (repeats >= 12) {
charnum += 3 * 12;
if (LIKELY(ip <= iend-7)) {
@@ -124,7 +97,7 @@ size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigne
ip = iend - 4;
}
bitStream = MEM_readLE32(ip) >> bitCount;
repeats = FSE_ctz(~bitStream | 0x80000000) >> 1;
repeats = ZSTD_countTrailingZeros32(~bitStream | 0x80000000) >> 1;
}
charnum += 3 * repeats;
bitStream >>= 2 * repeats;
@@ -189,7 +162,7 @@ size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigne
* know that threshold > 1.
*/
if (remaining <= 1) break;
nbBits = BIT_highbit32(remaining) + 1;
nbBits = ZSTD_highbit32(remaining) + 1;
threshold = 1 << (nbBits - 1);
}
if (charnum >= maxSV1) break;
@@ -312,14 +285,14 @@ HUF_readStats_body(BYTE* huffWeight, size_t hwSize, U32* rankStats,
if (weightTotal == 0) return ERROR(corruption_detected);
/* get last non-null symbol weight (implied, total must be 2^n) */
{ U32 const tableLog = BIT_highbit32(weightTotal) + 1;
{ U32 const tableLog = ZSTD_highbit32(weightTotal) + 1;
if (tableLog > HUF_TABLELOG_MAX) return ERROR(corruption_detected);
*tableLogPtr = tableLog;
/* determine last weight */
{ U32 const total = 1 << tableLog;
U32 const rest = total - weightTotal;
U32 const verif = 1 << BIT_highbit32(rest);
U32 const lastWeight = BIT_highbit32(rest) + 1;
U32 const verif = 1 << ZSTD_highbit32(rest);
U32 const lastWeight = ZSTD_highbit32(rest) + 1;
if (verif != rest) return ERROR(corruption_detected); /* last value must be a clean power of 2 */
huffWeight[oSize] = (BYTE)lastWeight;
rankStats[lastWeight]++;
lib/common/fse_decompress.c
+2 -1
View File
@@ -24,6 +24,7 @@
#include "error_private.h"
#define ZSTD_DEPS_NEED_MALLOC
#include "zstd_deps.h"
#include "bits.h" /* ZSTD_highbit32 */
/* **************************************************************
@@ -166,7 +167,7 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
for (u=0; u<tableSize; u++) {
FSE_FUNCTION_TYPE const symbol = (FSE_FUNCTION_TYPE)(tableDecode[u].symbol);
U32 const nextState = symbolNext[symbol]++;
tableDecode[u].nbBits = (BYTE) (tableLog - BIT_highbit32(nextState) );
tableDecode[u].nbBits = (BYTE) (tableLog - ZSTD_highbit32(nextState) );
tableDecode[u].newState = (U16) ( (nextState << tableDecode[u].nbBits) - tableSize);
} }
lib/common/zstd_internal.h
-92
View File
@@ -360,98 +360,6 @@ void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem);
void ZSTD_customFree(void* ptr, ZSTD_customMem customMem);
MEM_STATIC U32 ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus */
{
assert(val != 0);
{
# if defined(_MSC_VER) /* Visual */
# if STATIC_BMI2 == 1
return _lzcnt_u32(val)^31;
# else
if (val != 0) {
unsigned long r;
_BitScanReverse(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */
return (U32)__builtin_clz (val) ^ 31;
# elif defined(__ICCARM__) /* IAR Intrinsic */
return 31 - __CLZ(val);
# else /* Software version */
static const U32 DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };
U32 v = val;
v |= v >> 1;
v |= v >> 2;
v |= v >> 4;
v |= v >> 8;
v |= v >> 16;
return DeBruijnClz[(v * 0x07C4ACDDU) >> 27];
# endif
}
}
/**
* Counts the number of trailing zeros of a `size_t`.
* Most compilers should support CTZ as a builtin. A backup
* implementation is provided if the builtin isn't supported, but
* it may not be terribly efficient.
*/
MEM_STATIC unsigned ZSTD_countTrailingZeros(size_t val)
{
if (MEM_64bits()) {
# if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2
return _tzcnt_u64(val);
# else
if (val != 0) {
unsigned long r;
_BitScanForward64(&r, (U64)val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)__builtin_ctzll((U64)val);
# else
static const int DeBruijnBytePos[64] = { 0, 1, 2, 7, 3, 13, 8, 19,
4, 25, 14, 28, 9, 34, 20, 56,
5, 17, 26, 54, 15, 41, 29, 43,
10, 31, 38, 35, 21, 45, 49, 57,
63, 6, 12, 18, 24, 27, 33, 55,
16, 53, 40, 42, 30, 37, 44, 48,
62, 11, 23, 32, 52, 39, 36, 47,
61, 22, 51, 46, 60, 50, 59, 58 };
return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];
# endif
} else { /* 32 bits */
# if defined(_MSC_VER)
if (val != 0) {
unsigned long r;
_BitScanForward(&r, (U32)val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# elif defined(__GNUC__) && (__GNUC__ >= 3)
return (unsigned)__builtin_ctz((U32)val);
# else
static const int DeBruijnBytePos[32] = { 0, 1, 28, 2, 29, 14, 24, 3,
30, 22, 20, 15, 25, 17, 4, 8,
31, 27, 13, 23, 21, 19, 16, 7,
26, 12, 18, 6, 11, 5, 10, 9 };
return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
# endif
}
}
/* ZSTD_invalidateRepCodes() :
* ensures next compression will not use repcodes from previous block.
* Note : only works with regular variant;