view contrib/python-zstandard/zstd/common/mem.h @ 40121:73fef626dae3

zstandard: vendor python-zstandard 0.10.1

This was just released.

The upstream source distribution from PyPI was extracted. Unwanted
files were removed.

The clang-format ignore list was updated to reflect the new source
of files.

setup.py was updated to pass a new argument to python-zstandard's
function for returning an Extension instance. Upstream had to change
to use relative paths because Python 3.7's packaging doesn't
seem to like absolute paths when defining sources, includes, etc.
The default relative path calculation is relative to setup_zstd.py
which is different from the directory of Mercurial's setup.py.

The project contains a vendored copy of zstandard 1.3.6. The old
version was 1.3.4.

The API should be backwards compatible and nothing in core should
need adjusted. However, there is a new "chunker" API that we
may find useful in places where we want to emit compressed chunks
of a fixed size.

There are a pair of bug fixes in 0.10.0 with regards to
compressobj() and decompressobj() when block flushing is used. I
actually found these bugs when introducing these APIs in Mercurial!
But existing Mercurial code is not affected because we don't
perform block flushing.

# no-check-commit because 3rd party code has different style guidelines

Differential Revision: https://phab.mercurial-scm.org/D4911

/*
 * Copyright (c) 2016-present, Yann Collet, 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 MEM_H_MODULE
#define MEM_H_MODULE

#if defined (__cplusplus)
extern "C" {
#endif

/*-****************************************
*  Dependencies
******************************************/
#include <stddef.h>     /* size_t, ptrdiff_t */
#include <string.h>     /* memcpy */


/*-****************************************
*  Compiler specifics
******************************************/
#if defined(_MSC_VER)   /* Visual Studio */
#   include <stdlib.h>  /* _byteswap_ulong */
#   include <intrin.h>  /* _byteswap_* */
#endif
#if defined(__GNUC__)
#  define MEM_STATIC static __inline __attribute__((unused))
#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
#  define MEM_STATIC static inline
#elif defined(_MSC_VER)
#  define MEM_STATIC static __inline
#else
#  define MEM_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
#endif

/* code only tested on 32 and 64 bits systems */
#define MEM_STATIC_ASSERT(c)   { enum { MEM_static_assert = 1/(int)(!!(c)) }; }
MEM_STATIC void MEM_check(void) { MEM_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }


/*-**************************************************************
*  Basic Types
*****************************************************************/
#if  !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
# include <stdint.h>
  typedef   uint8_t BYTE;
  typedef  uint16_t U16;
  typedef   int16_t S16;
  typedef  uint32_t U32;
  typedef   int32_t S32;
  typedef  uint64_t U64;
  typedef   int64_t S64;
#else
# include <limits.h>
#if CHAR_BIT != 8
#  error "this implementation requires char to be exactly 8-bit type"
#endif
  typedef unsigned char      BYTE;
#if USHRT_MAX != 65535
#  error "this implementation requires short to be exactly 16-bit type"
#endif
  typedef unsigned short      U16;
  typedef   signed short      S16;
#if UINT_MAX != 4294967295
#  error "this implementation requires int to be exactly 32-bit type"
#endif
  typedef unsigned int        U32;
  typedef   signed int        S32;
/* note : there are no limits defined for long long type in C90.
 * limits exist in C99, however, in such case, <stdint.h> is preferred */
  typedef unsigned long long  U64;
  typedef   signed long long  S64;
#endif


/*-**************************************************************
*  Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS :
 * By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
 * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
 * The below switch allow to select different access method for improved performance.
 * Method 0 (default) : use `memcpy()`. Safe and portable.
 * Method 1 : `__packed` statement. It depends on compiler extension (i.e., not portable).
 *            This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
 * Method 2 : direct access. This method is portable but violate C standard.
 *            It can generate buggy code on targets depending on alignment.
 *            In some circumstances, it's the only known way to get the most performance (i.e. GCC + ARMv6)
 * See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
 * Prefer these methods in priority order (0 > 1 > 2)
 */
#ifndef MEM_FORCE_MEMORY_ACCESS   /* can be defined externally, on command line for example */
#  if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) )
#    define MEM_FORCE_MEMORY_ACCESS 2
#  elif defined(__INTEL_COMPILER) || defined(__GNUC__)
#    define MEM_FORCE_MEMORY_ACCESS 1
#  endif
#endif

MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; }

MEM_STATIC unsigned MEM_isLittleEndian(void)
{
    const union { U32 u; BYTE c[4]; } one = { 1 };   /* don't use static : performance detrimental  */
    return one.c[0];
}

#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)

/* violates C standard, by lying on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC size_t MEM_readST(const void* memPtr) { return *(const size_t*) memPtr; }

MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; }

#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)

/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
#if defined(_MSC_VER) || (defined(__INTEL_COMPILER) && defined(WIN32))
    __pragma( pack(push, 1) )
    typedef struct { U16 v; } unalign16;
    typedef struct { U32 v; } unalign32;
    typedef struct { U64 v; } unalign64;
    typedef struct { size_t v; } unalignArch;
    __pragma( pack(pop) )
#else
    typedef struct { U16 v; } __attribute__((packed)) unalign16;
    typedef struct { U32 v; } __attribute__((packed)) unalign32;
    typedef struct { U64 v; } __attribute__((packed)) unalign64;
    typedef struct { size_t v; } __attribute__((packed)) unalignArch;
#endif

MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign16*)ptr)->v; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign32*)ptr)->v; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign64*)ptr)->v; }
MEM_STATIC size_t MEM_readST(const void* ptr) { return ((const unalignArch*)ptr)->v; }

MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign16*)memPtr)->v = value; }
MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign32*)memPtr)->v = value; }
MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign64*)memPtr)->v = value; }

#else

/* default method, safe and standard.
   can sometimes prove slower */

MEM_STATIC U16 MEM_read16(const void* memPtr)
{
    U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
}

MEM_STATIC U32 MEM_read32(const void* memPtr)
{
    U32 val; memcpy(&val, memPtr, sizeof(val)); return val;
}

MEM_STATIC U64 MEM_read64(const void* memPtr)
{
    U64 val; memcpy(&val, memPtr, sizeof(val)); return val;
}

MEM_STATIC size_t MEM_readST(const void* memPtr)
{
    size_t val; memcpy(&val, memPtr, sizeof(val)); return val;
}

MEM_STATIC void MEM_write16(void* memPtr, U16 value)
{
    memcpy(memPtr, &value, sizeof(value));
}

MEM_STATIC void MEM_write32(void* memPtr, U32 value)
{
    memcpy(memPtr, &value, sizeof(value));
}

MEM_STATIC void MEM_write64(void* memPtr, U64 value)
{
    memcpy(memPtr, &value, sizeof(value));
}

#endif /* MEM_FORCE_MEMORY_ACCESS */

MEM_STATIC U32 MEM_swap32(U32 in)
{
#if defined(_MSC_VER)     /* Visual Studio */
    return _byteswap_ulong(in);
#elif defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)
    return __builtin_bswap32(in);
#else
    return  ((in << 24) & 0xff000000 ) |
            ((in <<  8) & 0x00ff0000 ) |
            ((in >>  8) & 0x0000ff00 ) |
            ((in >> 24) & 0x000000ff );
#endif
}

MEM_STATIC U64 MEM_swap64(U64 in)
{
#if defined(_MSC_VER)     /* Visual Studio */
    return _byteswap_uint64(in);
#elif defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)
    return __builtin_bswap64(in);
#else
    return  ((in << 56) & 0xff00000000000000ULL) |
            ((in << 40) & 0x00ff000000000000ULL) |
            ((in << 24) & 0x0000ff0000000000ULL) |
            ((in << 8)  & 0x000000ff00000000ULL) |
            ((in >> 8)  & 0x00000000ff000000ULL) |
            ((in >> 24) & 0x0000000000ff0000ULL) |
            ((in >> 40) & 0x000000000000ff00ULL) |
            ((in >> 56) & 0x00000000000000ffULL);
#endif
}

MEM_STATIC size_t MEM_swapST(size_t in)
{
    if (MEM_32bits())
        return (size_t)MEM_swap32((U32)in);
    else
        return (size_t)MEM_swap64((U64)in);
}

/*=== Little endian r/w ===*/

MEM_STATIC U16 MEM_readLE16(const void* memPtr)
{
    if (MEM_isLittleEndian())
        return MEM_read16(memPtr);
    else {
        const BYTE* p = (const BYTE*)memPtr;
        return (U16)(p[0] + (p[1]<<8));
    }
}

MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val)
{
    if (MEM_isLittleEndian()) {
        MEM_write16(memPtr, val);
    } else {
        BYTE* p = (BYTE*)memPtr;
        p[0] = (BYTE)val;
        p[1] = (BYTE)(val>>8);
    }
}

MEM_STATIC U32 MEM_readLE24(const void* memPtr)
{
    return MEM_readLE16(memPtr) + (((const BYTE*)memPtr)[2] << 16);
}

MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val)
{
    MEM_writeLE16(memPtr, (U16)val);
    ((BYTE*)memPtr)[2] = (BYTE)(val>>16);
}

MEM_STATIC U32 MEM_readLE32(const void* memPtr)
{
    if (MEM_isLittleEndian())
        return MEM_read32(memPtr);
    else
        return MEM_swap32(MEM_read32(memPtr));
}

MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32)
{
    if (MEM_isLittleEndian())
        MEM_write32(memPtr, val32);
    else
        MEM_write32(memPtr, MEM_swap32(val32));
}

MEM_STATIC U64 MEM_readLE64(const void* memPtr)
{
    if (MEM_isLittleEndian())
        return MEM_read64(memPtr);
    else
        return MEM_swap64(MEM_read64(memPtr));
}

MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64)
{
    if (MEM_isLittleEndian())
        MEM_write64(memPtr, val64);
    else
        MEM_write64(memPtr, MEM_swap64(val64));
}

MEM_STATIC size_t MEM_readLEST(const void* memPtr)
{
    if (MEM_32bits())
        return (size_t)MEM_readLE32(memPtr);
    else
        return (size_t)MEM_readLE64(memPtr);
}

MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val)
{
    if (MEM_32bits())
        MEM_writeLE32(memPtr, (U32)val);
    else
        MEM_writeLE64(memPtr, (U64)val);
}

/*=== Big endian r/w ===*/

MEM_STATIC U32 MEM_readBE32(const void* memPtr)
{
    if (MEM_isLittleEndian())
        return MEM_swap32(MEM_read32(memPtr));
    else
        return MEM_read32(memPtr);
}

MEM_STATIC void MEM_writeBE32(void* memPtr, U32 val32)
{
    if (MEM_isLittleEndian())
        MEM_write32(memPtr, MEM_swap32(val32));
    else
        MEM_write32(memPtr, val32);
}

MEM_STATIC U64 MEM_readBE64(const void* memPtr)
{
    if (MEM_isLittleEndian())
        return MEM_swap64(MEM_read64(memPtr));
    else
        return MEM_read64(memPtr);
}

MEM_STATIC void MEM_writeBE64(void* memPtr, U64 val64)
{
    if (MEM_isLittleEndian())
        MEM_write64(memPtr, MEM_swap64(val64));
    else
        MEM_write64(memPtr, val64);
}

MEM_STATIC size_t MEM_readBEST(const void* memPtr)
{
    if (MEM_32bits())
        return (size_t)MEM_readBE32(memPtr);
    else
        return (size_t)MEM_readBE64(memPtr);
}

MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val)
{
    if (MEM_32bits())
        MEM_writeBE32(memPtr, (U32)val);
    else
        MEM_writeBE64(memPtr, (U64)val);
}


#if defined (__cplusplus)
}
#endif

#endif /* MEM_H_MODULE */