/**
 * \file zstd.c
 * Single-file Zstandard library.
 *
 * Generate using:
 * \code
 *    python combine.py -r ../../lib -x legacy/zstd_legacy.h -o zstd.c zstd-in.c
 * \endcode
 */
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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.
 */
/*
 * Settings to bake for the single library file.
 *
 * Note: It's important that none of these affects 'zstd.h' (only the
 * implementation files we're amalgamating).
 *
 * Note: MEM_MODULE stops xxhash redefining BYTE, U16, etc., which are also
 * defined in mem.h (breaking C99 compatibility).
 *
 * Note: the undefs for xxHash allow Zstd's implementation to coincide with
 * standalone xxHash usage (with global defines).
 *
 * Note: if you enable ZSTD_LEGACY_SUPPORT the combine.py script will need
 * re-running without the "-x legacy/zstd_legacy.h" option (it excludes the
 * legacy support at the source level).
 *
 * Note: multithreading is enabled for all platforms apart from Emscripten.
 */
#define DEBUGLEVEL 0
#define MEM_MODULE
#undef  XXH_NAMESPACE
#define XXH_NAMESPACE ZSTD_
#undef  XXH_PRIVATE_API
#define XXH_PRIVATE_API
#undef  XXH_INLINE_ALL
#define XXH_INLINE_ALL
#define ZSTD_LEGACY_SUPPORT 0
#ifndef __EMSCRIPTEN__
#define ZSTD_MULTITHREAD
#endif
#define ZSTD_TRACE 0
/* TODO: Can't amalgamate ASM function */
#define ZSTD_DISABLE_ASM 1

/* >> v_patch start */
#if defined(__TINYC__)

#if defined(_WIN32)
#undef ZSTD_MULTITHREAD
#define ZSTD_NO_INTRINSICS
#endif

/* tcc doesn't support ARM asm */
#if defined(__arm__) || defined(__aarch64__)
#define NO_PREFETCH
#endif

#if defined(__FreeBSD__) || defined(__OpenBSD__)
/* tcc on FreeBSD/OpenBSD define __GNUC__, but it can't work here */
#undef __GNUC__
#endif

#endif /* __TINYC__ */
/* << v_patch end */

/* Include zstd_deps.h first with all the options we need enabled. */
#define ZSTD_DEPS_NEED_MALLOC
#define ZSTD_DEPS_NEED_MATH64
/**** start inlining common/zstd_deps.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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.
 */

/* This file provides common libc dependencies that zstd requires.
 * The purpose is to allow replacing this file with a custom implementation
 * to compile zstd without libc support.
 */

/* Need:
 * NULL
 * INT_MAX
 * UINT_MAX
 * ZSTD_memcpy()
 * ZSTD_memset()
 * ZSTD_memmove()
 */
#ifndef ZSTD_DEPS_COMMON
#define ZSTD_DEPS_COMMON

/* Even though we use qsort_r only for the dictionary builder, the macro
 * _GNU_SOURCE has to be declared *before* the inclusion of any standard
 * header and the script 'combine.sh' combines the whole zstd source code
 * in a single file.
 */
#if defined(__linux) || defined(__linux__) || defined(linux) || defined(__gnu_linux__) || \
    defined(__CYGWIN__) || defined(__MSYS__)
#if !defined(_GNU_SOURCE) && !defined(__ANDROID__) /* NDK doesn't ship qsort_r(). */
#define _GNU_SOURCE
#endif
#endif

#include <limits.h>
#include <stddef.h>
#include <string.h>

#if defined(__GNUC__) && __GNUC__ >= 4
# define ZSTD_memcpy(d,s,l) __builtin_memcpy((d),(s),(l))
# define ZSTD_memmove(d,s,l) __builtin_memmove((d),(s),(l))
# define ZSTD_memset(p,v,l) __builtin_memset((p),(v),(l))
#else
# define ZSTD_memcpy(d,s,l) memcpy((d),(s),(l))
# define ZSTD_memmove(d,s,l) memmove((d),(s),(l))
# define ZSTD_memset(p,v,l) memset((p),(v),(l))
#endif

#endif /* ZSTD_DEPS_COMMON */

/* Need:
 * ZSTD_malloc()
 * ZSTD_free()
 * ZSTD_calloc()
 */
#ifdef ZSTD_DEPS_NEED_MALLOC
#ifndef ZSTD_DEPS_MALLOC
#define ZSTD_DEPS_MALLOC

#include <stdlib.h>

#define ZSTD_malloc(s) malloc(s)
#define ZSTD_calloc(n,s) calloc((n), (s))
#define ZSTD_free(p) free((p))

#endif /* ZSTD_DEPS_MALLOC */
#endif /* ZSTD_DEPS_NEED_MALLOC */

/*
 * Provides 64-bit math support.
 * Need:
 * U64 ZSTD_div64(U64 dividend, U32 divisor)
 */
#ifdef ZSTD_DEPS_NEED_MATH64
#ifndef ZSTD_DEPS_MATH64
#define ZSTD_DEPS_MATH64

#define ZSTD_div64(dividend, divisor) ((dividend) / (divisor))

#endif /* ZSTD_DEPS_MATH64 */
#endif /* ZSTD_DEPS_NEED_MATH64 */

/* Need:
 * assert()
 */
#ifdef ZSTD_DEPS_NEED_ASSERT
#ifndef ZSTD_DEPS_ASSERT
#define ZSTD_DEPS_ASSERT

#include <assert.h>

#endif /* ZSTD_DEPS_ASSERT */
#endif /* ZSTD_DEPS_NEED_ASSERT */

/* Need:
 * ZSTD_DEBUG_PRINT()
 */
#ifdef ZSTD_DEPS_NEED_IO
#ifndef ZSTD_DEPS_IO
#define ZSTD_DEPS_IO

#include <stdio.h>
#define ZSTD_DEBUG_PRINT(...) fprintf(stderr, __VA_ARGS__)

#endif /* ZSTD_DEPS_IO */
#endif /* ZSTD_DEPS_NEED_IO */

/* Only requested when <stdint.h> is known to be present.
 * Need:
 * intptr_t
 */
#ifdef ZSTD_DEPS_NEED_STDINT
#ifndef ZSTD_DEPS_STDINT
#define ZSTD_DEPS_STDINT

#include <stdint.h>

#endif /* ZSTD_DEPS_STDINT */
#endif /* ZSTD_DEPS_NEED_STDINT */
/**** ended inlining common/zstd_deps.h ****/

/**** start inlining common/debug.c ****/
/* ******************************************************************
 * debug
 * Part of FSE library
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 *
 * You can contact the author at :
 * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
 *
 * 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.
****************************************************************** */


/*
 * This module only hosts one global variable
 * which can be used to dynamically influence the verbosity of traces,
 * such as DEBUGLOG and RAWLOG
 */

/**** start inlining debug.h ****/
/* ******************************************************************
 * debug
 * Part of FSE library
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 *
 * You can contact the author at :
 * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
 *
 * 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.
****************************************************************** */


/*
 * The purpose of this header is to enable debug functions.
 * They regroup assert(), DEBUGLOG() and RAWLOG() for run-time,
 * and DEBUG_STATIC_ASSERT() for compile-time.
 *
 * By default, DEBUGLEVEL==0, which means run-time debug is disabled.
 *
 * Level 1 enables assert() only.
 * Starting level 2, traces can be generated and pushed to stderr.
 * The higher the level, the more verbose the traces.
 *
 * It's possible to dynamically adjust level using variable g_debug_level,
 * which is only declared if DEBUGLEVEL>=2,
 * and is a global variable, not multi-thread protected (use with care)
 */

#ifndef DEBUG_H_12987983217
#define DEBUG_H_12987983217


/* static assert is triggered at compile time, leaving no runtime artefact.
 * static assert only works with compile-time constants.
 * Also, this variant can only be used inside a function. */
#define DEBUG_STATIC_ASSERT(c) (void)sizeof(char[(c) ? 1 : -1])


/* DEBUGLEVEL is expected to be defined externally,
 * typically through compiler command line.
 * Value must be a number. */
#ifndef DEBUGLEVEL
#  define DEBUGLEVEL 0
#endif


/* recommended values for DEBUGLEVEL :
 * 0 : release mode, no debug, all run-time checks disabled
 * 1 : enables assert() only, no display
 * 2 : reserved, for currently active debug path
 * 3 : events once per object lifetime (CCtx, CDict, etc.)
 * 4 : events once per frame
 * 5 : events once per block
 * 6 : events once per sequence (verbose)
 * 7+: events at every position (*very* verbose)
 *
 * It's generally inconvenient to output traces > 5.
 * In which case, it's possible to selectively trigger high verbosity levels
 * by modifying g_debug_level.
 */

#if (DEBUGLEVEL>=1)
#  define ZSTD_DEPS_NEED_ASSERT
/**** skipping file: zstd_deps.h ****/
#else
#  ifndef assert   /* assert may be already defined, due to prior #include <assert.h> */
#    define assert(condition) ((void)0)   /* disable assert (default) */
#  endif
#endif

#if (DEBUGLEVEL>=2)
#  define ZSTD_DEPS_NEED_IO
/**** skipping file: zstd_deps.h ****/
extern int g_debuglevel; /* the variable is only declared,
                            it actually lives in debug.c,
                            and is shared by the whole process.
                            It's not thread-safe.
                            It's useful when enabling very verbose levels
                            on selective conditions (such as position in src) */

#  define RAWLOG(l, ...)                   \
    do {                                   \
        if (l<=g_debuglevel) {             \
            ZSTD_DEBUG_PRINT(__VA_ARGS__); \
        }                                  \
    } while (0)

#define STRINGIFY(x) #x
#define TOSTRING(x) STRINGIFY(x)
#define LINE_AS_STRING TOSTRING(__LINE__)

#  define DEBUGLOG(l, ...)                               \
    do {                                                 \
        if (l<=g_debuglevel) {                           \
            ZSTD_DEBUG_PRINT(__FILE__ ":" LINE_AS_STRING ": " __VA_ARGS__); \
            ZSTD_DEBUG_PRINT(" \n");                     \
        }                                                \
    } while (0)
#else
#  define RAWLOG(l, ...)   do { } while (0)    /* disabled */
#  define DEBUGLOG(l, ...) do { } while (0)    /* disabled */
#endif

#endif /* DEBUG_H_12987983217 */
/**** ended inlining debug.h ****/

#if !defined(ZSTD_LINUX_KERNEL) || (DEBUGLEVEL>=2)
/* We only use this when DEBUGLEVEL>=2, but we get -Werror=pedantic errors if a
 * translation unit is empty. So remove this from Linux kernel builds, but
 * otherwise just leave it in.
 */
int g_debuglevel = DEBUGLEVEL;
#endif
/**** ended inlining common/debug.c ****/
/**** start inlining common/entropy_common.c ****/
/* ******************************************************************
 * Common functions of New Generation Entropy library
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 *
 *  You can contact the author at :
 *  - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
 *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
 *
 * 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.
****************************************************************** */

/* *************************************
*  Dependencies
***************************************/
/**** start inlining mem.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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

/*-****************************************
*  Dependencies
******************************************/
#include <stddef.h>  /* size_t, ptrdiff_t */
/**** start inlining compiler.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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_COMPILER_H
#define ZSTD_COMPILER_H

#include <stddef.h>

/**** start inlining portability_macros.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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_PORTABILITY_MACROS_H
#define ZSTD_PORTABILITY_MACROS_H

/**
 * This header file contains macro definitions to support portability.
 * This header is shared between C and ASM code, so it MUST only
 * contain macro definitions. It MUST not contain any C code.
 *
 * This header ONLY defines macros to detect platforms/feature support.
 *
 */


/* compat. with non-clang compilers */
#ifndef __has_attribute
  #define __has_attribute(x) 0
#endif

/* compat. with non-clang compilers */
#ifndef __has_builtin
#  define __has_builtin(x) 0
#endif

/* compat. with non-clang compilers */
#ifndef __has_feature
#  define __has_feature(x) 0
#endif

/* detects whether we are being compiled under msan */
#ifndef ZSTD_MEMORY_SANITIZER
#  if __has_feature(memory_sanitizer)
#    define ZSTD_MEMORY_SANITIZER 1
#  else
#    define ZSTD_MEMORY_SANITIZER 0
#  endif
#endif

/* detects whether we are being compiled under asan */
#ifndef ZSTD_ADDRESS_SANITIZER
#  if __has_feature(address_sanitizer)
#    define ZSTD_ADDRESS_SANITIZER 1
#  elif defined(__SANITIZE_ADDRESS__)
#    define ZSTD_ADDRESS_SANITIZER 1
#  else
#    define ZSTD_ADDRESS_SANITIZER 0
#  endif
#endif

/* detects whether we are being compiled under dfsan */
#ifndef ZSTD_DATAFLOW_SANITIZER
# if __has_feature(dataflow_sanitizer)
#  define ZSTD_DATAFLOW_SANITIZER 1
# else
#  define ZSTD_DATAFLOW_SANITIZER 0
# endif
#endif

/* Mark the internal assembly functions as hidden  */
#ifdef __ELF__
# define ZSTD_HIDE_ASM_FUNCTION(func) .hidden func
#elif defined(__APPLE__)
# define ZSTD_HIDE_ASM_FUNCTION(func) .private_extern func
#else
# define ZSTD_HIDE_ASM_FUNCTION(func)
#endif

/* Compile time determination of BMI2 support */
#ifndef STATIC_BMI2
#  if defined(__BMI2__)
#    define STATIC_BMI2 1
#  elif defined(_MSC_VER) && defined(__AVX2__)
#    define STATIC_BMI2 1 /* MSVC does not have a BMI2 specific flag, but every CPU that supports AVX2 also supports BMI2 */
#  endif
#endif

#ifndef STATIC_BMI2
#  define STATIC_BMI2 0
#endif

/* Enable runtime BMI2 dispatch based on the CPU.
 * Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default.
 */
#ifndef DYNAMIC_BMI2
#  if ((defined(__clang__) && __has_attribute(__target__)) \
      || (defined(__GNUC__) \
          && (__GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)))) \
      && (defined(__i386__) || defined(__x86_64__) || defined(_M_IX86) || defined(_M_X64)) \
      && !defined(__BMI2__)
#    define DYNAMIC_BMI2 1
#  else
#    define DYNAMIC_BMI2 0
#  endif
#endif

/**
 * Only enable assembly for GNU C compatible compilers,
 * because other platforms may not support GAS assembly syntax.
 *
 * Only enable assembly for Linux / MacOS / Win32, other platforms may
 * work, but they haven't been tested. This could likely be
 * extended to BSD systems.
 *
 * Disable assembly when MSAN is enabled, because MSAN requires
 * 100% of code to be instrumented to work.
 */
#if defined(__GNUC__)
#  if defined(__linux__) || defined(__linux) || defined(__APPLE__) || defined(_WIN32)
#    if ZSTD_MEMORY_SANITIZER
#      define ZSTD_ASM_SUPPORTED 0
#    elif ZSTD_DATAFLOW_SANITIZER
#      define ZSTD_ASM_SUPPORTED 0
#    else
#      define ZSTD_ASM_SUPPORTED 1
#    endif
#  else
#    define ZSTD_ASM_SUPPORTED 0
#  endif
#else
#  define ZSTD_ASM_SUPPORTED 0
#endif

/**
 * Determines whether we should enable assembly for x86-64
 * with BMI2.
 *
 * Enable if all of the following conditions hold:
 * - ASM hasn't been explicitly disabled by defining ZSTD_DISABLE_ASM
 * - Assembly is supported
 * - We are compiling for x86-64 and either:
 *   - DYNAMIC_BMI2 is enabled
 *   - BMI2 is supported at compile time
 */
#if !defined(ZSTD_DISABLE_ASM) &&                                 \
    ZSTD_ASM_SUPPORTED &&                                         \
    defined(__x86_64__) &&                                        \
    (DYNAMIC_BMI2 || defined(__BMI2__))
# define ZSTD_ENABLE_ASM_X86_64_BMI2 1
#else
# define ZSTD_ENABLE_ASM_X86_64_BMI2 0
#endif

/*
 * For x86 ELF targets, add .note.gnu.property section for Intel CET in
 * assembly sources when CET is enabled.
 *
 * Additionally, any function that may be called indirectly must begin
 * with ZSTD_CET_ENDBRANCH.
 */
#if defined(__ELF__) && (defined(__x86_64__) || defined(__i386__)) \
    && defined(__has_include)
# if __has_include(<cet.h>)
#  include <cet.h>
#  define ZSTD_CET_ENDBRANCH _CET_ENDBR
# endif
#endif

#ifndef ZSTD_CET_ENDBRANCH
# define ZSTD_CET_ENDBRANCH
#endif

#endif /* ZSTD_PORTABILITY_MACROS_H */
/**** ended inlining portability_macros.h ****/

/*-*******************************************************
*  Compiler specifics
*********************************************************/
/* force inlining */

#if !defined(ZSTD_NO_INLINE)
#if (defined(__GNUC__) && !defined(__STRICT_ANSI__)) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
#  define INLINE_KEYWORD inline
#else
#  define INLINE_KEYWORD
#endif

#if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
#  define FORCE_INLINE_ATTR __attribute__((always_inline))
#elif defined(_MSC_VER)
#  define FORCE_INLINE_ATTR __forceinline
#else
#  define FORCE_INLINE_ATTR
#endif

#else

#define INLINE_KEYWORD
#define FORCE_INLINE_ATTR

#endif

/**
  On MSVC qsort requires that functions passed into it use the __cdecl calling conversion(CC).
  This explicitly marks such functions as __cdecl so that the code will still compile
  if a CC other than __cdecl has been made the default.
*/
#if  defined(_MSC_VER)
#  define WIN_CDECL __cdecl
#else
#  define WIN_CDECL
#endif

/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
#if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
#  define UNUSED_ATTR __attribute__((unused))
#else
#  define UNUSED_ATTR
#endif

/**
 * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
 * parameters. They must be inlined for the compiler to eliminate the constant
 * branches.
 */
#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR UNUSED_ATTR
/**
 * HINT_INLINE is used to help the compiler generate better code. It is *not*
 * used for "templates", so it can be tweaked based on the compilers
 * performance.
 *
 * gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the
 * always_inline attribute.
 *
 * clang up to 5.0.0 (trunk) benefit tremendously from the always_inline
 * attribute.
 */
#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5
#  define HINT_INLINE static INLINE_KEYWORD
#else
#  define HINT_INLINE FORCE_INLINE_TEMPLATE
#endif

/* "soft" inline :
 * The compiler is free to select if it's a good idea to inline or not.
 * The main objective is to silence compiler warnings
 * when a defined function in included but not used.
 *
 * Note : this macro is prefixed `MEM_` because it used to be provided by `mem.h` unit.
 * Updating the prefix is probably preferable, but requires a fairly large codemod,
 * since this name is used everywhere.
 */
#ifndef MEM_STATIC  /* already defined in Linux Kernel mem.h */
#if defined(__GNUC__)
#  define MEM_STATIC static __inline UNUSED_ATTR
#elif defined(__IAR_SYSTEMS_ICC__)
#  define MEM_STATIC static inline UNUSED_ATTR
#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
#endif

/* force no inlining */
#ifdef _MSC_VER
#  define FORCE_NOINLINE static __declspec(noinline)
#else
#  if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
#    define FORCE_NOINLINE static __attribute__((__noinline__))
#  else
#    define FORCE_NOINLINE static
#  endif
#endif


/* target attribute */
#if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
#  define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
#else
#  define TARGET_ATTRIBUTE(target)
#endif

/* Target attribute for BMI2 dynamic dispatch.
 * Enable lzcnt, bmi, and bmi2.
 * We test for bmi1 & bmi2. lzcnt is included in bmi1.
 */
#define BMI2_TARGET_ATTRIBUTE TARGET_ATTRIBUTE("lzcnt,bmi,bmi2")

/* prefetch
 * can be disabled, by declaring NO_PREFETCH build macro */
#if defined(NO_PREFETCH)
#  define PREFETCH_L1(ptr)  do { (void)(ptr); } while (0)  /* disabled */
#  define PREFETCH_L2(ptr)  do { (void)(ptr); } while (0)  /* disabled */
#else
#  if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) && !defined(_M_ARM64EC)  /* _mm_prefetch() is not defined outside of x86/x64 */
#    include <mmintrin.h>   /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */
#    define PREFETCH_L1(ptr)  _mm_prefetch((const char*)(ptr), _MM_HINT_T0)
#    define PREFETCH_L2(ptr)  _mm_prefetch((const char*)(ptr), _MM_HINT_T1)
#  elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) )
#    define PREFETCH_L1(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */)
#    define PREFETCH_L2(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */)
#  elif defined(__aarch64__)
#    define PREFETCH_L1(ptr)  do { __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr))); } while (0)
#    define PREFETCH_L2(ptr)  do { __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr))); } while (0)
#  else
#    define PREFETCH_L1(ptr) do { (void)(ptr); } while (0)  /* disabled */
#    define PREFETCH_L2(ptr) do { (void)(ptr); } while (0)  /* disabled */
#  endif
#endif  /* NO_PREFETCH */

#define CACHELINE_SIZE 64

#define PREFETCH_AREA(p, s)                              \
    do {                                                 \
        const char* const _ptr = (const char*)(p);       \
        size_t const _size = (size_t)(s);                \
        size_t _pos;                                     \
        for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \
            PREFETCH_L2(_ptr + _pos);                    \
        }                                                \
    } while (0)

/* vectorization
 * older GCC (pre gcc-4.3 picked as the cutoff) uses a different syntax,
 * and some compilers, like Intel ICC and MCST LCC, do not support it at all. */
#if !defined(__INTEL_COMPILER) && !defined(__clang__) && defined(__GNUC__) && !defined(__LCC__) && !defined(__TINYC__)
#  if (__GNUC__ == 4 && __GNUC_MINOR__ > 3) || (__GNUC__ >= 5)
#    define DONT_VECTORIZE __attribute__((optimize("no-tree-vectorize")))
#  else
#    define DONT_VECTORIZE _Pragma("GCC optimize(\"no-tree-vectorize\")")
#  endif
#else
#  define DONT_VECTORIZE
#endif

/* Tell the compiler that a branch is likely or unlikely.
 * Only use these macros if it causes the compiler to generate better code.
 * If you can remove a LIKELY/UNLIKELY annotation without speed changes in gcc
 * and clang, please do.
 */
#if defined(__GNUC__)
#define LIKELY(x) (__builtin_expect((x), 1))
#define UNLIKELY(x) (__builtin_expect((x), 0))
#else
#define LIKELY(x) (x)
#define UNLIKELY(x) (x)
#endif

#if __has_builtin(__builtin_unreachable) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 5)))
#  define ZSTD_UNREACHABLE do { assert(0), __builtin_unreachable(); } while (0)
#else
#  define ZSTD_UNREACHABLE do { assert(0); } while (0)
#endif

/* disable warnings */
#ifdef _MSC_VER    /* Visual Studio */
#  include <intrin.h>                    /* For Visual 2005 */
#  pragma warning(disable : 4100)        /* disable: C4100: unreferenced formal parameter */
#  pragma warning(disable : 4127)        /* disable: C4127: conditional expression is constant */
#  pragma warning(disable : 4204)        /* disable: C4204: non-constant aggregate initializer */
#  pragma warning(disable : 4214)        /* disable: C4214: non-int bitfields */
#  pragma warning(disable : 4324)        /* disable: C4324: padded structure */
#endif

/* compile time determination of SIMD support */
#if !defined(ZSTD_NO_INTRINSICS)
#  if defined(__AVX2__)
#    define ZSTD_ARCH_X86_AVX2
#  endif
#  if defined(__SSE2__) || defined(_M_X64) || (defined (_M_IX86) && defined(_M_IX86_FP) && (_M_IX86_FP >= 2))
#    define ZSTD_ARCH_X86_SSE2
#  endif
#  if defined(__ARM_NEON) || defined(_M_ARM64)
#    define ZSTD_ARCH_ARM_NEON
#  endif
#
#  if defined(ZSTD_ARCH_X86_AVX2)
#    include <immintrin.h>
#  endif
#  if defined(ZSTD_ARCH_X86_SSE2)
#    include <emmintrin.h>
#  elif defined(ZSTD_ARCH_ARM_NEON)
#    include <arm_neon.h>
#  endif
#endif

/* C-language Attributes are added in C23. */
#if defined(__STDC_VERSION__) && (__STDC_VERSION__ > 201710L) && defined(__has_c_attribute)
# define ZSTD_HAS_C_ATTRIBUTE(x) __has_c_attribute(x)
#else
# define ZSTD_HAS_C_ATTRIBUTE(x) 0
#endif

/* Only use C++ attributes in C++. Some compilers report support for C++
 * attributes when compiling with C.
 */
#if defined(__cplusplus) && defined(__has_cpp_attribute)
# define ZSTD_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x)
#else
# define ZSTD_HAS_CPP_ATTRIBUTE(x) 0
#endif

/* Define ZSTD_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute.
 * - C23: https://en.cppreference.com/w/c/language/attributes/fallthrough
 * - CPP17: https://en.cppreference.com/w/cpp/language/attributes/fallthrough
 * - Else: __attribute__((__fallthrough__))
 */
#ifndef ZSTD_FALLTHROUGH
# if ZSTD_HAS_C_ATTRIBUTE(fallthrough)
#  define ZSTD_FALLTHROUGH [[fallthrough]]
# elif ZSTD_HAS_CPP_ATTRIBUTE(fallthrough)
#  define ZSTD_FALLTHROUGH [[fallthrough]]
# elif __has_attribute(__fallthrough__)
/* Leading semicolon is to satisfy gcc-11 with -pedantic. Without the semicolon
 * gcc complains about: a label can only be part of a statement and a declaration is not a statement.
 */
#  define ZSTD_FALLTHROUGH ; __attribute__((__fallthrough__))
# else
#  define ZSTD_FALLTHROUGH
# endif
#endif

/*-**************************************************************
*  Alignment
*****************************************************************/

/* @return 1 if @u is a 2^n value, 0 otherwise
 * useful to check a value is valid for alignment restrictions */
MEM_STATIC int ZSTD_isPower2(size_t u) {
    return (u & (u-1)) == 0;
}

/* this test was initially positioned in mem.h,
 * but this file is removed (or replaced) for linux kernel
 * so it's now hosted in compiler.h,
 * which remains valid for both user & kernel spaces.
 */

#ifndef ZSTD_ALIGNOF
# if defined(__GNUC__) || defined(_MSC_VER)
/* covers gcc, clang & MSVC */
/* note : this section must come first, before C11,
 * due to a limitation in the kernel source generator */
#  define ZSTD_ALIGNOF(T) __alignof(T)

# elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)
/* C11 support */
#  include <stdalign.h>
#  define ZSTD_ALIGNOF(T) alignof(T)

# else
/* No known support for alignof() - imperfect backup */
#  define ZSTD_ALIGNOF(T) (sizeof(void*) < sizeof(T) ? sizeof(void*) : sizeof(T))

# endif
#endif /* ZSTD_ALIGNOF */

#ifndef ZSTD_ALIGNED
/* C90-compatible alignment macro (GCC/Clang). Adjust for other compilers if needed. */
# if defined(__GNUC__) || defined(__clang__)
#  define ZSTD_ALIGNED(a) __attribute__((aligned(a)))
# elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */
#  define ZSTD_ALIGNED(a) _Alignas(a)
#elif defined(_MSC_VER)
#  define ZSTD_ALIGNED(n) __declspec(align(n))
# else
   /* this compiler will require its own alignment instruction */
#  define ZSTD_ALIGNED(...)
# endif
#endif /* ZSTD_ALIGNED */


/*-**************************************************************
*  Sanitizer
*****************************************************************/

/**
 * Zstd relies on pointer overflow in its decompressor.
 * We add this attribute to functions that rely on pointer overflow.
 */
#ifndef ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
#  if __has_attribute(no_sanitize)
#    if !defined(__clang__) && defined(__GNUC__) && __GNUC__ < 8
       /* gcc < 8 only has signed-integer-overlow which triggers on pointer overflow */
#      define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR __attribute__((no_sanitize("signed-integer-overflow")))
#    else
       /* older versions of clang [3.7, 5.0) will warn that pointer-overflow is ignored. */
#      define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR __attribute__((no_sanitize("pointer-overflow")))
#    endif
#  else
#    define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
#  endif
#endif

/**
 * Helper function to perform a wrapped pointer difference without triggering
 * UBSAN.
 *
 * @returns lhs - rhs with wrapping
 */
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
ptrdiff_t ZSTD_wrappedPtrDiff(unsigned char const* lhs, unsigned char const* rhs)
{
    return lhs - rhs;
}

/**
 * Helper function to perform a wrapped pointer add without triggering UBSAN.
 *
 * @return ptr + add with wrapping
 */
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
unsigned char const* ZSTD_wrappedPtrAdd(unsigned char const* ptr, ptrdiff_t add)
{
    return ptr + add;
}

/**
 * Helper function to perform a wrapped pointer subtraction without triggering
 * UBSAN.
 *
 * @return ptr - sub with wrapping
 */
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
unsigned char const* ZSTD_wrappedPtrSub(unsigned char const* ptr, ptrdiff_t sub)
{
    return ptr - sub;
}

/**
 * Helper function to add to a pointer that works around C's undefined behavior
 * of adding 0 to NULL.
 *
 * @returns `ptr + add` except it defines `NULL + 0 == NULL`.
 */
MEM_STATIC
unsigned char* ZSTD_maybeNullPtrAdd(unsigned char* ptr, ptrdiff_t add)
{
    return add > 0 ? ptr + add : ptr;
}

/* Issue #3240 reports an ASAN failure on an llvm-mingw build. Out of an
 * abundance of caution, disable our custom poisoning on mingw. */
#ifdef __MINGW32__
#ifndef ZSTD_ASAN_DONT_POISON_WORKSPACE
#define ZSTD_ASAN_DONT_POISON_WORKSPACE 1
#endif
#ifndef ZSTD_MSAN_DONT_POISON_WORKSPACE
#define ZSTD_MSAN_DONT_POISON_WORKSPACE 1
#endif
#endif

#if ZSTD_MEMORY_SANITIZER && !defined(ZSTD_MSAN_DONT_POISON_WORKSPACE)
/* Not all platforms that support msan provide sanitizers/msan_interface.h.
 * We therefore declare the functions we need ourselves, rather than trying to
 * include the header file... */
#include <stddef.h>  /* size_t */
#define ZSTD_DEPS_NEED_STDINT
/**** skipping file: zstd_deps.h ****/

/* Make memory region fully initialized (without changing its contents). */
void __msan_unpoison(const volatile void *a, size_t size);

/* Make memory region fully uninitialized (without changing its contents).
   This is a legacy interface that does not update origin information. Use
   __msan_allocated_memory() instead. */
void __msan_poison(const volatile void *a, size_t size);

/* Returns the offset of the first (at least partially) poisoned byte in the
   memory range, or -1 if the whole range is good. */
intptr_t __msan_test_shadow(const volatile void *x, size_t size);

/* Print shadow and origin for the memory range to stderr in a human-readable
   format. */
void __msan_print_shadow(const volatile void *x, size_t size);
#endif

#if ZSTD_ADDRESS_SANITIZER && !defined(ZSTD_ASAN_DONT_POISON_WORKSPACE)
/* Not all platforms that support asan provide sanitizers/asan_interface.h.
 * We therefore declare the functions we need ourselves, rather than trying to
 * include the header file... */
#include <stddef.h>  /* size_t */

/**
 * Marks a memory region (<c>[addr, addr+size)</c>) as unaddressable.
 *
 * This memory must be previously allocated by your program. Instrumented
 * code is forbidden from accessing addresses in this region until it is
 * unpoisoned. This function is not guaranteed to poison the entire region -
 * it could poison only a subregion of <c>[addr, addr+size)</c> due to ASan
 * alignment restrictions.
 *
 * \note This function is not thread-safe because no two threads can poison or
 * unpoison memory in the same memory region simultaneously.
 *
 * \param addr Start of memory region.
 * \param size Size of memory region. */
void __asan_poison_memory_region(void const volatile *addr, size_t size);

/**
 * Marks a memory region (<c>[addr, addr+size)</c>) as addressable.
 *
 * This memory must be previously allocated by your program. Accessing
 * addresses in this region is allowed until this region is poisoned again.
 * This function could unpoison a super-region of <c>[addr, addr+size)</c> due
 * to ASan alignment restrictions.
 *
 * \note This function is not thread-safe because no two threads can
 * poison or unpoison memory in the same memory region simultaneously.
 *
 * \param addr Start of memory region.
 * \param size Size of memory region. */
void __asan_unpoison_memory_region(void const volatile *addr, size_t size);
#endif

#endif /* ZSTD_COMPILER_H */
/**** ended inlining compiler.h ****/
/**** skipping file: debug.h ****/
/**** skipping file: zstd_deps.h ****/


/*-****************************************
*  Compiler specifics
******************************************/
#if defined(_MSC_VER)   /* Visual Studio */
#   include <stdlib.h>  /* _byteswap_ulong */
#   include <intrin.h>  /* _byteswap_* */
#elif defined(__ICCARM__)
#   include <intrinsics.h>
#endif

/*-**************************************************************
*  Basic Types
*****************************************************************/
#if  !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
#  if defined(_AIX)
#    include <inttypes.h>
#  else
#    include <stdint.h> /* intptr_t */
#  endif
  typedef   uint8_t BYTE;
  typedef   uint8_t U8;
  typedef    int8_t S8;
  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;
  typedef unsigned char      U8;
  typedef   signed char      S8;
#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 API
*****************************************************************/
/*=== Static platform detection ===*/
MEM_STATIC unsigned MEM_32bits(void);
MEM_STATIC unsigned MEM_64bits(void);
MEM_STATIC unsigned MEM_isLittleEndian(void);

/*=== Native unaligned read/write ===*/
MEM_STATIC U16 MEM_read16(const void* memPtr);
MEM_STATIC U32 MEM_read32(const void* memPtr);
MEM_STATIC U64 MEM_read64(const void* memPtr);
MEM_STATIC size_t MEM_readST(const void* memPtr);

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

/*=== Little endian unaligned read/write ===*/
MEM_STATIC U16 MEM_readLE16(const void* memPtr);
MEM_STATIC U32 MEM_readLE24(const void* memPtr);
MEM_STATIC U32 MEM_readLE32(const void* memPtr);
MEM_STATIC U64 MEM_readLE64(const void* memPtr);
MEM_STATIC size_t MEM_readLEST(const void* memPtr);

MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val);
MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val);
MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32);
MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64);
MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val);

/*=== Big endian unaligned read/write ===*/
MEM_STATIC U32 MEM_readBE32(const void* memPtr);
MEM_STATIC U64 MEM_readBE64(const void* memPtr);
MEM_STATIC size_t MEM_readBEST(const void* memPtr);

MEM_STATIC void MEM_writeBE32(void* memPtr, U32 val32);
MEM_STATIC void MEM_writeBE64(void* memPtr, U64 val64);
MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val);

/*=== Byteswap ===*/
MEM_STATIC U32 MEM_swap32(U32 in);
MEM_STATIC U64 MEM_swap64(U64 in);
MEM_STATIC size_t MEM_swapST(size_t in);


/*-**************************************************************
*  Memory I/O Implementation
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS : For accessing unaligned memory:
 * Method 0 : always use `memcpy()`. Safe and portable.
 * Method 1 : Use compiler extension to set unaligned access.
 * Method 2 : direct access. This method is portable but violate C standard.
 *            It can generate buggy code on targets depending on alignment.
 * Default  : method 1 if supported, else method 0
 */
#ifndef MEM_FORCE_MEMORY_ACCESS   /* can be defined externally, on command line for example */
#  ifdef __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)
{
#if defined(__BYTE_ORDER__) && defined(__ORDER_LITTLE_ENDIAN__) && (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)
    return 1;
#elif defined(__BYTE_ORDER__) && defined(__ORDER_BIG_ENDIAN__) && (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__)
    return 0;
#elif defined(__clang__) && __LITTLE_ENDIAN__
    return 1;
#elif defined(__clang__) && __BIG_ENDIAN__
    return 0;
#elif defined(_MSC_VER) && (_M_X64 || _M_IX86)
    return 1;
#elif defined(__DMC__) && defined(_M_IX86)
    return 1;
#elif defined(__IAR_SYSTEMS_ICC__) && __LITTLE_ENDIAN__
    return 1;
#else
    const union { U32 u; BYTE c[4]; } one = { 1 };   /* don't use static : performance detrimental  */
    return one.c[0];
#endif
}

#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)

typedef __attribute__((aligned(1))) U16 unalign16;
typedef __attribute__((aligned(1))) U32 unalign32;
typedef __attribute__((aligned(1))) U64 unalign64;
typedef __attribute__((aligned(1))) size_t unalignArch;

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

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

#else

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

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

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

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

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

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

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

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

#endif /* MEM_FORCE_MEMORY_ACCESS */

MEM_STATIC U32 MEM_swap32_fallback(U32 in)
{
    return  ((in << 24) & 0xff000000 ) |
            ((in <<  8) & 0x00ff0000 ) |
            ((in >>  8) & 0x0000ff00 ) |
            ((in >> 24) & 0x000000ff );
}

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)) \
  || (defined(__clang__) && __has_builtin(__builtin_bswap32))
    return __builtin_bswap32(in);
#elif defined(__ICCARM__)
    return __REV(in);
#else
    return MEM_swap32_fallback(in);
#endif
}

MEM_STATIC U64 MEM_swap64_fallback(U64 in)
{
     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);
}

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)) \
  || (defined(__clang__) && __has_builtin(__builtin_bswap64))
    return __builtin_bswap64(in);
#else
    return MEM_swap64_fallback(in);
#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 (U32)MEM_readLE16(memPtr) + ((U32)(((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);
}

/* code only tested on 32 and 64 bits systems */
MEM_STATIC void MEM_check(void) { DEBUG_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }

#endif /* MEM_H_MODULE */
/**** ended inlining mem.h ****/
/**** start inlining error_private.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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.
 */

/* Note : this module is expected to remain private, do not expose it */

#ifndef ERROR_H_MODULE
#define ERROR_H_MODULE

/* ****************************************
*  Dependencies
******************************************/
/**** start inlining ../zstd_errors.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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_ERRORS_H_398273423
#define ZSTD_ERRORS_H_398273423

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

/* =====   ZSTDERRORLIB_API : control library symbols visibility   ===== */
#ifndef ZSTDERRORLIB_VISIBLE
   /* Backwards compatibility with old macro name */
#  ifdef ZSTDERRORLIB_VISIBILITY
#    define ZSTDERRORLIB_VISIBLE ZSTDERRORLIB_VISIBILITY
#  elif defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
#    define ZSTDERRORLIB_VISIBLE __attribute__ ((visibility ("default")))
#  else
#    define ZSTDERRORLIB_VISIBLE
#  endif
#endif

#ifndef ZSTDERRORLIB_HIDDEN
#  if defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
#    define ZSTDERRORLIB_HIDDEN __attribute__ ((visibility ("hidden")))
#  else
#    define ZSTDERRORLIB_HIDDEN
#  endif
#endif

#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
#  define ZSTDERRORLIB_API __declspec(dllexport) ZSTDERRORLIB_VISIBLE
#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
#  define ZSTDERRORLIB_API __declspec(dllimport) ZSTDERRORLIB_VISIBLE /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
#else
#  define ZSTDERRORLIB_API ZSTDERRORLIB_VISIBLE
#endif

/*-*********************************************
 *  Error codes list
 *-*********************************************
 *  Error codes _values_ are pinned down since v1.3.1 only.
 *  Therefore, don't rely on values if you may link to any version < v1.3.1.
 *
 *  Only values < 100 are considered stable.
 *
 *  note 1 : this API shall be used with static linking only.
 *           dynamic linking is not yet officially supported.
 *  note 2 : Prefer relying on the enum than on its value whenever possible
 *           This is the only supported way to use the error list < v1.3.1
 *  note 3 : ZSTD_isError() is always correct, whatever the library version.
 **********************************************/
typedef enum {
  ZSTD_error_no_error = 0,
  ZSTD_error_GENERIC  = 1,
  ZSTD_error_prefix_unknown                = 10,
  ZSTD_error_version_unsupported           = 12,
  ZSTD_error_frameParameter_unsupported    = 14,
  ZSTD_error_frameParameter_windowTooLarge = 16,
  ZSTD_error_corruption_detected = 20,
  ZSTD_error_checksum_wrong      = 22,
  ZSTD_error_literals_headerWrong = 24,
  ZSTD_error_dictionary_corrupted      = 30,
  ZSTD_error_dictionary_wrong          = 32,
  ZSTD_error_dictionaryCreation_failed = 34,
  ZSTD_error_parameter_unsupported   = 40,
  ZSTD_error_parameter_combination_unsupported = 41,
  ZSTD_error_parameter_outOfBound    = 42,
  ZSTD_error_tableLog_tooLarge       = 44,
  ZSTD_error_maxSymbolValue_tooLarge = 46,
  ZSTD_error_maxSymbolValue_tooSmall = 48,
  ZSTD_error_cannotProduce_uncompressedBlock = 49,
  ZSTD_error_stabilityCondition_notRespected = 50,
  ZSTD_error_stage_wrong       = 60,
  ZSTD_error_init_missing      = 62,
  ZSTD_error_memory_allocation = 64,
  ZSTD_error_workSpace_tooSmall= 66,
  ZSTD_error_dstSize_tooSmall = 70,
  ZSTD_error_srcSize_wrong    = 72,
  ZSTD_error_dstBuffer_null   = 74,
  ZSTD_error_noForwardProgress_destFull = 80,
  ZSTD_error_noForwardProgress_inputEmpty = 82,
  /* following error codes are __NOT STABLE__, they can be removed or changed in future versions */
  ZSTD_error_frameIndex_tooLarge = 100,
  ZSTD_error_seekableIO          = 102,
  ZSTD_error_dstBuffer_wrong     = 104,
  ZSTD_error_srcBuffer_wrong     = 105,
  ZSTD_error_sequenceProducer_failed = 106,
  ZSTD_error_externalSequences_invalid = 107,
  ZSTD_error_maxCode = 120  /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */
} ZSTD_ErrorCode;

ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code);   /**< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */


#if defined (__cplusplus)
}
#endif

#endif /* ZSTD_ERRORS_H_398273423 */
/**** ended inlining ../zstd_errors.h ****/
/**** skipping file: compiler.h ****/
/**** skipping file: debug.h ****/
/**** skipping file: zstd_deps.h ****/

/* ****************************************
*  Compiler-specific
******************************************/
#if defined(__GNUC__)
#  define ERR_STATIC static __attribute__((unused))
#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
#  define ERR_STATIC static inline
#elif defined(_MSC_VER)
#  define ERR_STATIC static __inline
#else
#  define ERR_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
#endif


/*-****************************************
*  Customization (error_public.h)
******************************************/
typedef ZSTD_ErrorCode ERR_enum;
#define PREFIX(name) ZSTD_error_##name


/*-****************************************
*  Error codes handling
******************************************/
#undef ERROR   /* already defined on Visual Studio */
#define ERROR(name) ZSTD_ERROR(name)
#define ZSTD_ERROR(name) ((size_t)-PREFIX(name))

ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); }

ERR_STATIC ERR_enum ERR_getErrorCode(size_t code) { if (!ERR_isError(code)) return (ERR_enum)0; return (ERR_enum) (0-code); }

/* check and forward error code */
#define CHECK_V_F(e, f)     \
    size_t const e = f;     \
    do {                    \
        if (ERR_isError(e)) \
            return e;       \
    } while (0)
#define CHECK_F(f)   do { CHECK_V_F(_var_err__, f); } while (0)


/*-****************************************
*  Error Strings
******************************************/

const char* ERR_getErrorString(ERR_enum code);   /* error_private.c */

ERR_STATIC const char* ERR_getErrorName(size_t code)
{
    return ERR_getErrorString(ERR_getErrorCode(code));
}

/**
 * Ignore: this is an internal helper.
 *
 * This is a helper function to help force C99-correctness during compilation.
 * Under strict compilation modes, variadic macro arguments can't be empty.
 * However, variadic function arguments can be. Using a function therefore lets
 * us statically check that at least one (string) argument was passed,
 * independent of the compilation flags.
 */
static INLINE_KEYWORD UNUSED_ATTR
void _force_has_format_string(const char *format, ...) {
  (void)format;
}

/**
 * Ignore: this is an internal helper.
 *
 * We want to force this function invocation to be syntactically correct, but
 * we don't want to force runtime evaluation of its arguments.
 */
#define _FORCE_HAS_FORMAT_STRING(...)              \
    do {                                           \
        if (0) {                                   \
            _force_has_format_string(__VA_ARGS__); \
        }                                          \
    } while (0)

#define ERR_QUOTE(str) #str

/**
 * Return the specified error if the condition evaluates to true.
 *
 * In debug modes, prints additional information.
 * In order to do that (particularly, printing the conditional that failed),
 * this can't just wrap RETURN_ERROR().
 */
#define RETURN_ERROR_IF(cond, err, ...)                                        \
    do {                                                                       \
        if (cond) {                                                            \
            RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s",          \
                  __FILE__, __LINE__, ERR_QUOTE(cond), ERR_QUOTE(ERROR(err))); \
            _FORCE_HAS_FORMAT_STRING(__VA_ARGS__);                             \
            RAWLOG(3, ": " __VA_ARGS__);                                       \
            RAWLOG(3, "\n");                                                   \
            return ERROR(err);                                                 \
        }                                                                      \
    } while (0)

/**
 * Unconditionally return the specified error.
 *
 * In debug modes, prints additional information.
 */
#define RETURN_ERROR(err, ...)                                               \
    do {                                                                     \
        RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", \
              __FILE__, __LINE__, ERR_QUOTE(ERROR(err)));                    \
        _FORCE_HAS_FORMAT_STRING(__VA_ARGS__);                               \
        RAWLOG(3, ": " __VA_ARGS__);                                         \
        RAWLOG(3, "\n");                                                     \
        return ERROR(err);                                                   \
    } while(0)

/**
 * If the provided expression evaluates to an error code, returns that error code.
 *
 * In debug modes, prints additional information.
 */
#define FORWARD_IF_ERROR(err, ...)                                                 \
    do {                                                                           \
        size_t const err_code = (err);                                             \
        if (ERR_isError(err_code)) {                                               \
            RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s",                 \
                  __FILE__, __LINE__, ERR_QUOTE(err), ERR_getErrorName(err_code)); \
            _FORCE_HAS_FORMAT_STRING(__VA_ARGS__);                                 \
            RAWLOG(3, ": " __VA_ARGS__);                                           \
            RAWLOG(3, "\n");                                                       \
            return err_code;                                                       \
        }                                                                          \
    } while(0)

#endif /* ERROR_H_MODULE */
/**** ended inlining error_private.h ****/
#define FSE_STATIC_LINKING_ONLY  /* FSE_MIN_TABLELOG */
/**** start inlining fse.h ****/
/* ******************************************************************
 * FSE : Finite State Entropy codec
 * Public Prototypes declaration
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 *
 * You can contact the author at :
 * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
 *
 * 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 FSE_H
#define FSE_H


/*-*****************************************
*  Dependencies
******************************************/
/**** skipping file: zstd_deps.h ****/

/*-*****************************************
*  FSE_PUBLIC_API : control library symbols visibility
******************************************/
#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)
#  define FSE_PUBLIC_API __attribute__ ((visibility ("default")))
#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1)   /* Visual expected */
#  define FSE_PUBLIC_API __declspec(dllexport)
#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)
#  define FSE_PUBLIC_API __declspec(dllimport) /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
#else
#  define FSE_PUBLIC_API
#endif

/*------   Version   ------*/
#define FSE_VERSION_MAJOR    0
#define FSE_VERSION_MINOR    9
#define FSE_VERSION_RELEASE  0

#define FSE_LIB_VERSION FSE_VERSION_MAJOR.FSE_VERSION_MINOR.FSE_VERSION_RELEASE
#define FSE_QUOTE(str) #str
#define FSE_EXPAND_AND_QUOTE(str) FSE_QUOTE(str)
#define FSE_VERSION_STRING FSE_EXPAND_AND_QUOTE(FSE_LIB_VERSION)

#define FSE_VERSION_NUMBER  (FSE_VERSION_MAJOR *100*100 + FSE_VERSION_MINOR *100 + FSE_VERSION_RELEASE)
FSE_PUBLIC_API unsigned FSE_versionNumber(void);   /**< library version number; to be used when checking dll version */


/*-*****************************************
*  Tool functions
******************************************/
FSE_PUBLIC_API size_t FSE_compressBound(size_t size);       /* maximum compressed size */

/* Error Management */
FSE_PUBLIC_API unsigned    FSE_isError(size_t code);        /* tells if a return value is an error code */
FSE_PUBLIC_API const char* FSE_getErrorName(size_t code);   /* provides error code string (useful for debugging) */


/*-*****************************************
*  FSE detailed API
******************************************/
/*!
FSE_compress() does the following:
1. count symbol occurrence from source[] into table count[] (see hist.h)
2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog)
3. save normalized counters to memory buffer using writeNCount()
4. build encoding table 'CTable' from normalized counters
5. encode the data stream using encoding table 'CTable'

FSE_decompress() does the following:
1. read normalized counters with readNCount()
2. build decoding table 'DTable' from normalized counters
3. decode the data stream using decoding table 'DTable'

The following API allows targeting specific sub-functions for advanced tasks.
For example, it's possible to compress several blocks using the same 'CTable',
or to save and provide normalized distribution using external method.
*/

/* *** COMPRESSION *** */

/*! FSE_optimalTableLog():
    dynamically downsize 'tableLog' when conditions are met.
    It saves CPU time, by using smaller tables, while preserving or even improving compression ratio.
    @return : recommended tableLog (necessarily <= 'maxTableLog') */
FSE_PUBLIC_API unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);

/*! FSE_normalizeCount():
    normalize counts so that sum(count[]) == Power_of_2 (2^tableLog)
    'normalizedCounter' is a table of short, of minimum size (maxSymbolValue+1).
    useLowProbCount is a boolean parameter which trades off compressed size for
    faster header decoding. When it is set to 1, the compressed data will be slightly
    smaller. And when it is set to 0, FSE_readNCount() and FSE_buildDTable() will be
    faster. If you are compressing a small amount of data (< 2 KB) then useLowProbCount=0
    is a good default, since header deserialization makes a big speed difference.
    Otherwise, useLowProbCount=1 is a good default, since the speed difference is small.
    @return : tableLog,
              or an errorCode, which can be tested using FSE_isError() */
FSE_PUBLIC_API size_t FSE_normalizeCount(short* normalizedCounter, unsigned tableLog,
                    const unsigned* count, size_t srcSize, unsigned maxSymbolValue, unsigned useLowProbCount);

/*! FSE_NCountWriteBound():
    Provides the maximum possible size of an FSE normalized table, given 'maxSymbolValue' and 'tableLog'.
    Typically useful for allocation purpose. */
FSE_PUBLIC_API size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog);

/*! FSE_writeNCount():
    Compactly save 'normalizedCounter' into 'buffer'.
    @return : size of the compressed table,
              or an errorCode, which can be tested using FSE_isError(). */
FSE_PUBLIC_API size_t FSE_writeNCount (void* buffer, size_t bufferSize,
                                 const short* normalizedCounter,
                                 unsigned maxSymbolValue, unsigned tableLog);

/*! Constructor and Destructor of FSE_CTable.
    Note that FSE_CTable size depends on 'tableLog' and 'maxSymbolValue' */
typedef unsigned FSE_CTable;   /* don't allocate that. It's only meant to be more restrictive than void* */

/*! FSE_buildCTable():
    Builds `ct`, which must be already allocated, using FSE_createCTable().
    @return : 0, or an errorCode, which can be tested using FSE_isError() */
FSE_PUBLIC_API size_t FSE_buildCTable(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);

/*! FSE_compress_usingCTable():
    Compress `src` using `ct` into `dst` which must be already allocated.
    @return : size of compressed data (<= `dstCapacity`),
              or 0 if compressed data could not fit into `dst`,
              or an errorCode, which can be tested using FSE_isError() */
FSE_PUBLIC_API size_t FSE_compress_usingCTable (void* dst, size_t dstCapacity, const void* src, size_t srcSize, const FSE_CTable* ct);

/*!
Tutorial :
----------
The first step is to count all symbols. FSE_count() does this job very fast.
Result will be saved into 'count', a table of unsigned int, which must be already allocated, and have 'maxSymbolValuePtr[0]+1' cells.
'src' is a table of bytes of size 'srcSize'. All values within 'src' MUST be <= maxSymbolValuePtr[0]
maxSymbolValuePtr[0] will be updated, with its real value (necessarily <= original value)
FSE_count() will return the number of occurrence of the most frequent symbol.
This can be used to know if there is a single symbol within 'src', and to quickly evaluate its compressibility.
If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()).

The next step is to normalize the frequencies.
FSE_normalizeCount() will ensure that sum of frequencies is == 2 ^'tableLog'.
It also guarantees a minimum of 1 to any Symbol with frequency >= 1.
You can use 'tableLog'==0 to mean "use default tableLog value".
If you are unsure of which tableLog value to use, you can ask FSE_optimalTableLog(),
which will provide the optimal valid tableLog given sourceSize, maxSymbolValue, and a user-defined maximum (0 means "default").

The result of FSE_normalizeCount() will be saved into a table,
called 'normalizedCounter', which is a table of signed short.
'normalizedCounter' must be already allocated, and have at least 'maxSymbolValue+1' cells.
The return value is tableLog if everything proceeded as expected.
It is 0 if there is a single symbol within distribution.
If there is an error (ex: invalid tableLog value), the function will return an ErrorCode (which can be tested using FSE_isError()).

'normalizedCounter' can be saved in a compact manner to a memory area using FSE_writeNCount().
'buffer' must be already allocated.
For guaranteed success, buffer size must be at least FSE_headerBound().
The result of the function is the number of bytes written into 'buffer'.
If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError(); ex : buffer size too small).

'normalizedCounter' can then be used to create the compression table 'CTable'.
The space required by 'CTable' must be already allocated, using FSE_createCTable().
You can then use FSE_buildCTable() to fill 'CTable'.
If there is an error, both functions will return an ErrorCode (which can be tested using FSE_isError()).

'CTable' can then be used to compress 'src', with FSE_compress_usingCTable().
Similar to FSE_count(), the convention is that 'src' is assumed to be a table of char of size 'srcSize'
The function returns the size of compressed data (without header), necessarily <= `dstCapacity`.
If it returns '0', compressed data could not fit into 'dst'.
If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()).
*/


/* *** DECOMPRESSION *** */

/*! FSE_readNCount():
    Read compactly saved 'normalizedCounter' from 'rBuffer'.
    @return : size read from 'rBuffer',
              or an errorCode, which can be tested using FSE_isError().
              maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */
FSE_PUBLIC_API size_t FSE_readNCount (short* normalizedCounter,
                           unsigned* maxSymbolValuePtr, unsigned* tableLogPtr,
                           const void* rBuffer, size_t rBuffSize);

/*! FSE_readNCount_bmi2():
 * Same as FSE_readNCount() but pass bmi2=1 when your CPU supports BMI2 and 0 otherwise.
 */
FSE_PUBLIC_API size_t FSE_readNCount_bmi2(short* normalizedCounter,
                           unsigned* maxSymbolValuePtr, unsigned* tableLogPtr,
                           const void* rBuffer, size_t rBuffSize, int bmi2);

typedef unsigned FSE_DTable;   /* don't allocate that. It's just a way to be more restrictive than void* */

/*!
Tutorial :
----------
(Note : these functions only decompress FSE-compressed blocks.
 If block is uncompressed, use memcpy() instead
 If block is a single repeated byte, use memset() instead )

The first step is to obtain the normalized frequencies of symbols.
This can be performed by FSE_readNCount() if it was saved using FSE_writeNCount().
'normalizedCounter' must be already allocated, and have at least 'maxSymbolValuePtr[0]+1' cells of signed short.
In practice, that means it's necessary to know 'maxSymbolValue' beforehand,
or size the table to handle worst case situations (typically 256).
FSE_readNCount() will provide 'tableLog' and 'maxSymbolValue'.
The result of FSE_readNCount() is the number of bytes read from 'rBuffer'.
Note that 'rBufferSize' must be at least 4 bytes, even if useful information is less than that.
If there is an error, the function will return an error code, which can be tested using FSE_isError().

The next step is to build the decompression tables 'FSE_DTable' from 'normalizedCounter'.
This is performed by the function FSE_buildDTable().
The space required by 'FSE_DTable' must be already allocated using FSE_createDTable().
If there is an error, the function will return an error code, which can be tested using FSE_isError().

`FSE_DTable` can then be used to decompress `cSrc`, with FSE_decompress_usingDTable().
`cSrcSize` must be strictly correct, otherwise decompression will fail.
FSE_decompress_usingDTable() result will tell how many bytes were regenerated (<=`dstCapacity`).
If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small)
*/

#endif  /* FSE_H */


#if defined(FSE_STATIC_LINKING_ONLY) && !defined(FSE_H_FSE_STATIC_LINKING_ONLY)
#define FSE_H_FSE_STATIC_LINKING_ONLY
/**** start inlining bitstream.h ****/
/* ******************************************************************
 * bitstream
 * Part of FSE library
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 *
 * You can contact the author at :
 * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
 *
 * 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 BITSTREAM_H_MODULE
#define BITSTREAM_H_MODULE

/*
*  This API consists of small unitary functions, which must be inlined for best performance.
*  Since link-time-optimization is not available for all compilers,
*  these functions are defined into a .h to be included.
*/

/*-****************************************
*  Dependencies
******************************************/
/**** skipping file: mem.h ****/
/**** skipping file: compiler.h ****/
/**** skipping file: debug.h ****/
/**** skipping file: error_private.h ****/
/**** start inlining bits.h ****/
/*
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 * 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

/**** skipping file: mem.h ****/

MEM_STATIC unsigned ZSTD_countTrailingZeros32_fallback(U32 val)
{
    assert(val != 0);
    {
        static const U32 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];
    }
}

MEM_STATIC unsigned ZSTD_countTrailingZeros32(U32 val)
{
    assert(val != 0);
#if defined(_MSC_VER)
#  if STATIC_BMI2
    return (unsigned)_tzcnt_u32(val);
#  else
    if (val != 0) {
        unsigned long r;
        _BitScanForward(&r, val);
        return (unsigned)r;
    } else {
        __assume(0); /* Should not reach this code path */
    }
#  endif
#elif defined(__GNUC__) && (__GNUC__ >= 4)
    return (unsigned)__builtin_ctz(val);
#elif defined(__ICCARM__)
    return (unsigned)__builtin_ctz(val);
#else
    return ZSTD_countTrailingZeros32_fallback(val);
#endif
}

MEM_STATIC unsigned ZSTD_countLeadingZeros32_fallback(U32 val)
{
    assert(val != 0);
    {
        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};
        val |= val >> 1;
        val |= val >> 2;
        val |= val >> 4;
        val |= val >> 8;
        val |= val >> 16;
        return 31 - DeBruijnClz[(val * 0x07C4ACDDU) >> 27];
    }
}

MEM_STATIC unsigned ZSTD_countLeadingZeros32(U32 val)
{
    assert(val != 0);
#if defined(_MSC_VER)
#  if STATIC_BMI2
    return (unsigned)_lzcnt_u32(val);
#  else
    if (val != 0) {
        unsigned long r;
        _BitScanReverse(&r, val);
        return (unsigned)(31 - r);
    } else {
        __assume(0); /* Should not reach this code path */
    }
#  endif
#elif defined(__GNUC__) && (__GNUC__ >= 4)
    return (unsigned)__builtin_clz(val);
#elif defined(__ICCARM__)
    return (unsigned)__builtin_clz(val);
#else
    return ZSTD_countLeadingZeros32_fallback(val);
#endif
}

MEM_STATIC unsigned ZSTD_countTrailingZeros64(U64 val)
{
    assert(val != 0);
#if defined(_MSC_VER) && defined(_WIN64)
#  if STATIC_BMI2
    return (unsigned)_tzcnt_u64(val);
#  else
    if (val != 0) {
        unsigned long r;
        _BitScanForward64(&r, val);
        return (unsigned)r;
    } else {
        __assume(0); /* Should not reach this code path */
    }
#  endif
#elif defined(__GNUC__) && (__GNUC__ >= 4) && defined(__LP64__)
    return (unsigned)__builtin_ctzll(val);
#elif defined(__ICCARM__)
    return (unsigned)__builtin_ctzll(val);
#else
    {
        U32 mostSignificantWord = (U32)(val >> 32);
        U32 leastSignificantWord = (U32)val;
        if (leastSignificantWord == 0) {
            return 32 + ZSTD_countTrailingZeros32(mostSignificantWord);
        } else {
            return ZSTD_countTrailingZeros32(leastSignificantWord);
        }
    }
#endif
}

MEM_STATIC unsigned ZSTD_countLeadingZeros64(U64 val)
{
    assert(val != 0);
#if defined(_MSC_VER) && defined(_WIN64)
#  if STATIC_BMI2
    return (unsigned)_lzcnt_u64(val);
#  else
    if (val != 0) {
        unsigned long r;
        _BitScanReverse64(&r, val);
        return (unsigned)(63 - r);
    } else {
        __assume(0); /* Should not reach this code path */
    }
#  endif
#elif defined(__GNUC__) && (__GNUC__ >= 4)
    return (unsigned)(__builtin_clzll(val));
#elif defined(__ICCARM__)
    return (unsigned)(__builtin_clzll(val));
#else
    {
        U32 mostSignificantWord = (U32)(val >> 32);
        U32 leastSignificantWord = (U32)val;
        if (mostSignificantWord == 0) {
            return 32 + ZSTD_countLeadingZeros32(leastSignificantWord);
        } else {
            return ZSTD_countLeadingZeros32(mostSignificantWord);
        }
    }
#endif
}

MEM_STATIC unsigned ZSTD_NbCommonBytes(size_t val)
{
    if (MEM_isLittleEndian()) {
        if (MEM_64bits()) {
            return ZSTD_countTrailingZeros64((U64)val) >> 3;
        } else {
            return ZSTD_countTrailingZeros32((U32)val) >> 3;
        }
    } else {  /* Big Endian CPU */
        if (MEM_64bits()) {
            return ZSTD_countLeadingZeros64((U64)val) >> 3;
        } else {
            return ZSTD_countLeadingZeros32((U32)val) >> 3;
        }
    }
}

MEM_STATIC unsigned ZSTD_highbit32(U32 val)   /* compress, dictBuilder, decodeCorpus */
{
    assert(val != 0);
    return 31 - ZSTD_countLeadingZeros32(val);
}

/* ZSTD_rotateRight_*():
 * Rotates a bitfield to the right by "count" bits.
 * https://en.wikipedia.org/w/index.php?title=Circular_shift&oldid=991635599#Implementing_circular_shifts
 */
MEM_STATIC
U64 ZSTD_rotateRight_U64(U64 const value, U32 count) {
    assert(count < 64);
    count &= 0x3F; /* for fickle pattern recognition */
    return (value >> count) | (U64)(value << ((0U - count) & 0x3F));
}

MEM_STATIC
U32 ZSTD_rotateRight_U32(U32 const value, U32 count) {
    assert(count < 32);
    count &= 0x1F; /* for fickle pattern recognition */
    return (value >> count) | (U32)(value << ((0U - count) & 0x1F));
}

MEM_STATIC
U16 ZSTD_rotateRight_U16(U16 const value, U32 count) {
    assert(count < 16);
    count &= 0x0F; /* for fickle pattern recognition */
    return (value >> count) | (U16)(value << ((0U - count) & 0x0F));
}

#endif /* ZSTD_BITS_H */
/**** ended inlining bits.h ****/

/*=========================================
*  Target specific
=========================================*/
#ifndef ZSTD_NO_INTRINSICS
#  if (defined(__BMI__) || defined(__BMI2__)) && defined(__GNUC__)
#    include <immintrin.h>   /* support for bextr (experimental)/bzhi */
#  elif defined(__ICCARM__)
#    include <intrinsics.h>
#  endif
#endif

#define STREAM_ACCUMULATOR_MIN_32  25
#define STREAM_ACCUMULATOR_MIN_64  57
#define STREAM_ACCUMULATOR_MIN    ((U32)(MEM_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64))


/*-******************************************
*  bitStream encoding API (write forward)
********************************************/
typedef size_t BitContainerType;
/* bitStream can mix input from multiple sources.
 * A critical property of these streams is that they encode and decode in **reverse** direction.
 * So the first bit sequence you add will be the last to be read, like a LIFO stack.
 */
typedef struct {
    BitContainerType bitContainer;
    unsigned bitPos;
    char*  startPtr;
    char*  ptr;
    char*  endPtr;
} BIT_CStream_t;

MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC, void* dstBuffer, size_t dstCapacity);
MEM_STATIC void   BIT_addBits(BIT_CStream_t* bitC, BitContainerType value, unsigned nbBits);
MEM_STATIC void   BIT_flushBits(BIT_CStream_t* bitC);
MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC);

/* Start with initCStream, providing the size of buffer to write into.
*  bitStream will never write outside of this buffer.
*  `dstCapacity` must be >= sizeof(bitD->bitContainer), otherwise @return will be an error code.
*
*  bits are first added to a local register.
*  Local register is BitContainerType, 64-bits on 64-bits systems, or 32-bits on 32-bits systems.
*  Writing data into memory is an explicit operation, performed by the flushBits function.
*  Hence keep track how many bits are potentially stored into local register to avoid register overflow.
*  After a flushBits, a maximum of 7 bits might still be stored into local register.
*
*  Avoid storing elements of more than 24 bits if you want compatibility with 32-bits bitstream readers.
*
*  Last operation is to close the bitStream.
*  The function returns the final size of CStream in bytes.
*  If data couldn't fit into `dstBuffer`, it will return a 0 ( == not storable)
*/


/*-********************************************
*  bitStream decoding API (read backward)
**********************************************/
typedef struct {
    BitContainerType bitContainer;
    unsigned bitsConsumed;
    const char* ptr;
    const char* start;
    const char* limitPtr;
} BIT_DStream_t;

typedef enum { BIT_DStream_unfinished = 0,  /* fully refilled */
               BIT_DStream_endOfBuffer = 1, /* still some bits left in bitstream */
               BIT_DStream_completed = 2,   /* bitstream entirely consumed, bit-exact */
               BIT_DStream_overflow = 3     /* user requested more bits than present in bitstream */
    } BIT_DStream_status;  /* result of BIT_reloadDStream() */

MEM_STATIC size_t   BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize);
MEM_STATIC BitContainerType BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits);
MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD);
MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* bitD);


/* Start by invoking BIT_initDStream().
*  A chunk of the bitStream is then stored into a local register.
*  Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (BitContainerType).
*  You can then retrieve bitFields stored into the local register, **in reverse order**.
*  Local register is explicitly reloaded from memory by the BIT_reloadDStream() method.
*  A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished.
*  Otherwise, it can be less than that, so proceed accordingly.
*  Checking if DStream has reached its end can be performed with BIT_endOfDStream().
*/


/*-****************************************
*  unsafe API
******************************************/
MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC, BitContainerType value, unsigned nbBits);
/* faster, but works only if value is "clean", meaning all high bits above nbBits are 0 */

MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC);
/* unsafe version; does not check buffer overflow */

MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits);
/* faster, but works only if nbBits >= 1 */

/*=====    Local Constants   =====*/
static const unsigned BIT_mask[] = {
    0,          1,         3,         7,         0xF,       0x1F,
    0x3F,       0x7F,      0xFF,      0x1FF,     0x3FF,     0x7FF,
    0xFFF,      0x1FFF,    0x3FFF,    0x7FFF,    0xFFFF,    0x1FFFF,
    0x3FFFF,    0x7FFFF,   0xFFFFF,   0x1FFFFF,  0x3FFFFF,  0x7FFFFF,
    0xFFFFFF,   0x1FFFFFF, 0x3FFFFFF, 0x7FFFFFF, 0xFFFFFFF, 0x1FFFFFFF,
    0x3FFFFFFF, 0x7FFFFFFF}; /* up to 31 bits */
#define BIT_MASK_SIZE (sizeof(BIT_mask) / sizeof(BIT_mask[0]))

/*-**************************************************************
*  bitStream encoding
****************************************************************/
/*! BIT_initCStream() :
 *  `dstCapacity` must be > sizeof(size_t)
 *  @return : 0 if success,
 *            otherwise an error code (can be tested using ERR_isError()) */
MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,
                                  void* startPtr, size_t dstCapacity)
{
    bitC->bitContainer = 0;
    bitC->bitPos = 0;
    bitC->startPtr = (char*)startPtr;
    bitC->ptr = bitC->startPtr;
    bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->bitContainer);
    if (dstCapacity <= sizeof(bitC->bitContainer)) return ERROR(dstSize_tooSmall);
    return 0;
}

FORCE_INLINE_TEMPLATE BitContainerType BIT_getLowerBits(BitContainerType bitContainer, U32 const nbBits)
{
#if STATIC_BMI2 && !defined(ZSTD_NO_INTRINSICS)
#  if (defined(__x86_64__) || defined(_M_X64)) && !defined(__ILP32__)
    return _bzhi_u64(bitContainer, nbBits);
#  else
    DEBUG_STATIC_ASSERT(sizeof(bitContainer) == sizeof(U32));
    return _bzhi_u32(bitContainer, nbBits);
#  endif
#else
    assert(nbBits < BIT_MASK_SIZE);
    return bitContainer & BIT_mask[nbBits];
#endif
}

/*! BIT_addBits() :
 *  can add up to 31 bits into `bitC`.
 *  Note : does not check for register overflow ! */
MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
                            BitContainerType value, unsigned nbBits)
{
    DEBUG_STATIC_ASSERT(BIT_MASK_SIZE == 32);
    assert(nbBits < BIT_MASK_SIZE);
    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
    bitC->bitContainer |= BIT_getLowerBits(value, nbBits) << bitC->bitPos;
    bitC->bitPos += nbBits;
}

/*! BIT_addBitsFast() :
 *  works only if `value` is _clean_,
 *  meaning all high bits above nbBits are 0 */
MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,
                                BitContainerType value, unsigned nbBits)
{
    assert((value>>nbBits) == 0);
    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
    bitC->bitContainer |= value << bitC->bitPos;
    bitC->bitPos += nbBits;
}

/*! BIT_flushBitsFast() :
 *  assumption : bitContainer has not overflowed
 *  unsafe version; does not check buffer overflow */
MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC)
{
    size_t const nbBytes = bitC->bitPos >> 3;
    assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8);
    assert(bitC->ptr <= bitC->endPtr);
    MEM_writeLEST(bitC->ptr, bitC->bitContainer);
    bitC->ptr += nbBytes;
    bitC->bitPos &= 7;
    bitC->bitContainer >>= nbBytes*8;
}

/*! BIT_flushBits() :
 *  assumption : bitContainer has not overflowed
 *  safe version; check for buffer overflow, and prevents it.
 *  note : does not signal buffer overflow.
 *  overflow will be revealed later on using BIT_closeCStream() */
MEM_STATIC void BIT_flushBits(BIT_CStream_t* bitC)
{
    size_t const nbBytes = bitC->bitPos >> 3;
    assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8);
    assert(bitC->ptr <= bitC->endPtr);
    MEM_writeLEST(bitC->ptr, bitC->bitContainer);
    bitC->ptr += nbBytes;
    if (bitC->ptr > bitC->endPtr) bitC->ptr = bitC->endPtr;
    bitC->bitPos &= 7;
    bitC->bitContainer >>= nbBytes*8;
}

/*! BIT_closeCStream() :
 *  @return : size of CStream, in bytes,
 *            or 0 if it could not fit into dstBuffer */
MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC)
{
    BIT_addBitsFast(bitC, 1, 1);   /* endMark */
    BIT_flushBits(bitC);
    if (bitC->ptr >= bitC->endPtr) return 0; /* overflow detected */
    return (size_t)(bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0);
}


/*-********************************************************
*  bitStream decoding
**********************************************************/
/*! BIT_initDStream() :
 *  Initialize a BIT_DStream_t.
 * `bitD` : a pointer to an already allocated BIT_DStream_t structure.
 * `srcSize` must be the *exact* size of the bitStream, in bytes.
 * @return : size of stream (== srcSize), or an errorCode if a problem is detected
 */
MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize)
{
    if (srcSize < 1) { ZSTD_memset(bitD, 0, sizeof(*bitD)); return ERROR(srcSize_wrong); }

    bitD->start = (const char*)srcBuffer;
    bitD->limitPtr = bitD->start + sizeof(bitD->bitContainer);

    if (srcSize >=  sizeof(bitD->bitContainer)) {  /* normal case */
        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 - ZSTD_highbit32(lastByte) : 0;  /* ensures bitsConsumed is always set */
          if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ }
    } else {
        bitD->ptr   = bitD->start;
        bitD->bitContainer = *(const BYTE*)(bitD->start);
        switch(srcSize)
        {
        case 7: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);
                ZSTD_FALLTHROUGH;

        case 6: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
                ZSTD_FALLTHROUGH;

        case 5: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
                ZSTD_FALLTHROUGH;

        case 4: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[3]) << 24;
                ZSTD_FALLTHROUGH;

        case 3: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[2]) << 16;
                ZSTD_FALLTHROUGH;

        case 2: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[1]) <<  8;
                ZSTD_FALLTHROUGH;

        default: break;
        }
        {   BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
            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;
    }

    return srcSize;
}

FORCE_INLINE_TEMPLATE BitContainerType BIT_getUpperBits(BitContainerType bitContainer, U32 const start)
{
    return bitContainer >> start;
}

FORCE_INLINE_TEMPLATE BitContainerType BIT_getMiddleBits(BitContainerType bitContainer, U32 const start, U32 const nbBits)
{
    U32 const regMask = sizeof(bitContainer)*8 - 1;
    /* if start > regMask, bitstream is corrupted, and result is undefined */
    assert(nbBits < BIT_MASK_SIZE);
    /* x86 transform & ((1 << nbBits) - 1) to bzhi instruction, it is better
     * than accessing memory. When bmi2 instruction is not present, we consider
     * such cpus old (pre-Haswell, 2013) and their performance is not of that
     * importance.
     */
#if defined(__x86_64__) || defined(_M_X64)
    return (bitContainer >> (start & regMask)) & ((((U64)1) << nbBits) - 1);
#else
    return (bitContainer >> (start & regMask)) & BIT_mask[nbBits];
#endif
}

/*! BIT_lookBits() :
 *  Provides next n bits from local register.
 *  local register is not modified.
 *  On 32-bits, maxNbBits==24.
 *  On 64-bits, maxNbBits==56.
 * @return : value extracted */
FORCE_INLINE_TEMPLATE BitContainerType BIT_lookBits(const BIT_DStream_t*  bitD, U32 nbBits)
{
    /* arbitrate between double-shift and shift+mask */
#if 1
    /* if bitD->bitsConsumed + nbBits > sizeof(bitD->bitContainer)*8,
     * bitstream is likely corrupted, and result is undefined */
    return BIT_getMiddleBits(bitD->bitContainer, (sizeof(bitD->bitContainer)*8) - bitD->bitsConsumed - nbBits, nbBits);
#else
    /* this code path is slower on my os-x laptop */
    U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
    return ((bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> 1) >> ((regMask-nbBits) & regMask);
#endif
}

/*! BIT_lookBitsFast() :
 *  unsafe version; only works if nbBits >= 1 */
MEM_STATIC BitContainerType BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits)
{
    U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
    assert(nbBits >= 1);
    return (bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> (((regMask+1)-nbBits) & regMask);
}

FORCE_INLINE_TEMPLATE void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
{
    bitD->bitsConsumed += nbBits;
}

/*! BIT_readBits() :
 *  Read (consume) next n bits from local register and update.
 *  Pay attention to not read more than nbBits contained into local register.
 * @return : extracted value. */
FORCE_INLINE_TEMPLATE BitContainerType BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
{
    BitContainerType const value = BIT_lookBits(bitD, nbBits);
    BIT_skipBits(bitD, nbBits);
    return value;
}

/*! BIT_readBitsFast() :
 *  unsafe version; only works if nbBits >= 1 */
MEM_STATIC BitContainerType BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
{
    BitContainerType const value = BIT_lookBitsFast(bitD, nbBits);
    assert(nbBits >= 1);
    BIT_skipBits(bitD, nbBits);
    return value;
}

/*! BIT_reloadDStream_internal() :
 *  Simple variant of BIT_reloadDStream(), with two conditions:
 *  1. bitstream is valid : bitsConsumed <= sizeof(bitD->bitContainer)*8
 *  2. look window is valid after shifted down : bitD->ptr >= bitD->start
 */
MEM_STATIC BIT_DStream_status BIT_reloadDStream_internal(BIT_DStream_t* bitD)
{
    assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
    bitD->ptr -= bitD->bitsConsumed >> 3;
    assert(bitD->ptr >= bitD->start);
    bitD->bitsConsumed &= 7;
    bitD->bitContainer = MEM_readLEST(bitD->ptr);
    return BIT_DStream_unfinished;
}

/*! BIT_reloadDStreamFast() :
 *  Similar to BIT_reloadDStream(), but with two differences:
 *  1. bitsConsumed <= sizeof(bitD->bitContainer)*8 must hold!
 *  2. Returns BIT_DStream_overflow when bitD->ptr < bitD->limitPtr, at this
 *     point you must use BIT_reloadDStream() to reload.
 */
MEM_STATIC BIT_DStream_status BIT_reloadDStreamFast(BIT_DStream_t* bitD)
{
    if (UNLIKELY(bitD->ptr < bitD->limitPtr))
        return BIT_DStream_overflow;
    return BIT_reloadDStream_internal(bitD);
}

/*! BIT_reloadDStream() :
 *  Refill `bitD` from buffer previously set in BIT_initDStream() .
 *  This function is safe, it guarantees it will not never beyond src buffer.
 * @return : status of `BIT_DStream_t` internal register.
 *           when status == BIT_DStream_unfinished, internal register is filled with at least 25 or 57 bits */
FORCE_INLINE_TEMPLATE BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
{
    /* note : once in overflow mode, a bitstream remains in this mode until it's reset */
    if (UNLIKELY(bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))) {
        static const BitContainerType zeroFilled = 0;
        bitD->ptr = (const char*)&zeroFilled; /* aliasing is allowed for char */
        /* overflow detected, erroneous scenario or end of stream: no update */
        return BIT_DStream_overflow;
    }

    assert(bitD->ptr >= bitD->start);

    if (bitD->ptr >= bitD->limitPtr) {
        return BIT_reloadDStream_internal(bitD);
    }
    if (bitD->ptr == bitD->start) {
        /* reached end of bitStream => no update */
        if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer;
        return BIT_DStream_completed;
    }
    /* start < ptr < limitPtr => cautious update */
    {   U32 nbBytes = bitD->bitsConsumed >> 3;
        BIT_DStream_status result = BIT_DStream_unfinished;
        if (bitD->ptr - nbBytes < bitD->start) {
            nbBytes = (U32)(bitD->ptr - bitD->start);  /* ptr > start */
            result = BIT_DStream_endOfBuffer;
        }
        bitD->ptr -= nbBytes;
        bitD->bitsConsumed -= nbBytes*8;
        bitD->bitContainer = MEM_readLEST(bitD->ptr);   /* reminder : srcSize > sizeof(bitD->bitContainer), otherwise bitD->ptr == bitD->start */
        return result;
    }
}

/*! BIT_endOfDStream() :
 * @return : 1 if DStream has _exactly_ reached its end (all bits consumed).
 */
MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* DStream)
{
    return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer)*8));
}

#endif /* BITSTREAM_H_MODULE */
/**** ended inlining bitstream.h ****/

/* *****************************************
*  Static allocation
*******************************************/
/* FSE buffer bounds */
#define FSE_NCOUNTBOUND 512
#define FSE_BLOCKBOUND(size) ((size) + ((size)>>7) + 4 /* fse states */ + sizeof(size_t) /* bitContainer */)
#define FSE_COMPRESSBOUND(size) (FSE_NCOUNTBOUND + FSE_BLOCKBOUND(size))   /* Macro version, useful for static allocation */

/* It is possible to statically allocate FSE CTable/DTable as a table of FSE_CTable/FSE_DTable using below macros */
#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue)   (1 + (1<<((maxTableLog)-1)) + (((maxSymbolValue)+1)*2))
#define FSE_DTABLE_SIZE_U32(maxTableLog)                   (1 + (1<<(maxTableLog)))

/* or use the size to malloc() space directly. Pay attention to alignment restrictions though */
#define FSE_CTABLE_SIZE(maxTableLog, maxSymbolValue)   (FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(FSE_CTable))
#define FSE_DTABLE_SIZE(maxTableLog)                   (FSE_DTABLE_SIZE_U32(maxTableLog) * sizeof(FSE_DTable))


/* *****************************************
 *  FSE advanced API
 ***************************************** */

unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus);
/**< same as FSE_optimalTableLog(), which used `minus==2` */

size_t FSE_buildCTable_rle (FSE_CTable* ct, unsigned char symbolValue);
/**< build a fake FSE_CTable, designed to compress always the same symbolValue */

/* FSE_buildCTable_wksp() :
 * Same as FSE_buildCTable(), but using an externally allocated scratch buffer (`workSpace`).
 * `wkspSize` must be >= `FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog)` of `unsigned`.
 * See FSE_buildCTable_wksp() for breakdown of workspace usage.
 */
#define FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog) (((maxSymbolValue + 2) + (1ull << (tableLog)))/2 + sizeof(U64)/sizeof(U32) /* additional 8 bytes for potential table overwrite */)
#define FSE_BUILD_CTABLE_WORKSPACE_SIZE(maxSymbolValue, tableLog) (sizeof(unsigned) * FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog))
size_t FSE_buildCTable_wksp(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);

#define FSE_BUILD_DTABLE_WKSP_SIZE(maxTableLog, maxSymbolValue) (sizeof(short) * (maxSymbolValue + 1) + (1ULL << maxTableLog) + 8)
#define FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) ((FSE_BUILD_DTABLE_WKSP_SIZE(maxTableLog, maxSymbolValue) + sizeof(unsigned) - 1) / sizeof(unsigned))
FSE_PUBLIC_API size_t FSE_buildDTable_wksp(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
/**< Same as FSE_buildDTable(), using an externally allocated `workspace` produced with `FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxSymbolValue)` */

#define FSE_DECOMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) (FSE_DTABLE_SIZE_U32(maxTableLog) + 1 + FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) + (FSE_MAX_SYMBOL_VALUE + 1) / 2 + 1)
#define FSE_DECOMPRESS_WKSP_SIZE(maxTableLog, maxSymbolValue) (FSE_DECOMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(unsigned))
size_t FSE_decompress_wksp_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize, int bmi2);
/**< same as FSE_decompress(), using an externally allocated `workSpace` produced with `FSE_DECOMPRESS_WKSP_SIZE_U32(maxLog, maxSymbolValue)`.
 * Set bmi2 to 1 if your CPU supports BMI2 or 0 if it doesn't */

typedef enum {
   FSE_repeat_none,  /**< Cannot use the previous table */
   FSE_repeat_check, /**< Can use the previous table but it must be checked */
   FSE_repeat_valid  /**< Can use the previous table and it is assumed to be valid */
 } FSE_repeat;

/* *****************************************
*  FSE symbol compression API
*******************************************/
/*!
   This API consists of small unitary functions, which highly benefit from being inlined.
   Hence their body are included in next section.
*/
typedef struct {
    ptrdiff_t   value;
    const void* stateTable;
    const void* symbolTT;
    unsigned    stateLog;
} FSE_CState_t;

static void FSE_initCState(FSE_CState_t* CStatePtr, const FSE_CTable* ct);

static void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* CStatePtr, unsigned symbol);

static void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* CStatePtr);

/**<
These functions are inner components of FSE_compress_usingCTable().
They allow the creation of custom streams, mixing multiple tables and bit sources.

A key property to keep in mind is that encoding and decoding are done **in reverse direction**.
So the first symbol you will encode is the last you will decode, like a LIFO stack.

You will need a few variables to track your CStream. They are :

FSE_CTable    ct;         // Provided by FSE_buildCTable()
BIT_CStream_t bitStream;  // bitStream tracking structure
FSE_CState_t  state;      // State tracking structure (can have several)


The first thing to do is to init bitStream and state.
    size_t errorCode = BIT_initCStream(&bitStream, dstBuffer, maxDstSize);
    FSE_initCState(&state, ct);

Note that BIT_initCStream() can produce an error code, so its result should be tested, using FSE_isError();
You can then encode your input data, byte after byte.
FSE_encodeSymbol() outputs a maximum of 'tableLog' bits at a time.
Remember decoding will be done in reverse direction.
    FSE_encodeByte(&bitStream, &state, symbol);

At any time, you can also add any bit sequence.
Note : maximum allowed nbBits is 25, for compatibility with 32-bits decoders
    BIT_addBits(&bitStream, bitField, nbBits);

The above methods don't commit data to memory, they just store it into local register, for speed.
Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
Writing data to memory is a manual operation, performed by the flushBits function.
    BIT_flushBits(&bitStream);

Your last FSE encoding operation shall be to flush your last state value(s).
    FSE_flushState(&bitStream, &state);

Finally, you must close the bitStream.
The function returns the size of CStream in bytes.
If data couldn't fit into dstBuffer, it will return a 0 ( == not compressible)
If there is an error, it returns an errorCode (which can be tested using FSE_isError()).
    size_t size = BIT_closeCStream(&bitStream);
*/


/* *****************************************
*  FSE symbol decompression API
*******************************************/
typedef struct {
    size_t      state;
    const void* table;   /* precise table may vary, depending on U16 */
} FSE_DState_t;


static void     FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt);

static unsigned char FSE_decodeSymbol(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD);

static unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr);

/**<
Let's now decompose FSE_decompress_usingDTable() into its unitary components.
You will decode FSE-encoded symbols from the bitStream,
and also any other bitFields you put in, **in reverse order**.

You will need a few variables to track your bitStream. They are :

BIT_DStream_t DStream;    // Stream context
FSE_DState_t  DState;     // State context. Multiple ones are possible
FSE_DTable*   DTablePtr;  // Decoding table, provided by FSE_buildDTable()

The first thing to do is to init the bitStream.
    errorCode = BIT_initDStream(&DStream, srcBuffer, srcSize);

You should then retrieve your initial state(s)
(in reverse flushing order if you have several ones) :
    errorCode = FSE_initDState(&DState, &DStream, DTablePtr);

You can then decode your data, symbol after symbol.
For information the maximum number of bits read by FSE_decodeSymbol() is 'tableLog'.
Keep in mind that symbols are decoded in reverse order, like a LIFO stack (last in, first out).
    unsigned char symbol = FSE_decodeSymbol(&DState, &DStream);

You can retrieve any bitfield you eventually stored into the bitStream (in reverse order)
Note : maximum allowed nbBits is 25, for 32-bits compatibility
    size_t bitField = BIT_readBits(&DStream, nbBits);

All above operations only read from local register (which size depends on size_t).
Refueling the register from memory is manually performed by the reload method.
    endSignal = FSE_reloadDStream(&DStream);

BIT_reloadDStream() result tells if there is still some more data to read from DStream.
BIT_DStream_unfinished : there is still some data left into the DStream.
BIT_DStream_endOfBuffer : Dstream reached end of buffer. Its container may no longer be completely filled.
BIT_DStream_completed : Dstream reached its exact end, corresponding in general to decompression completed.
BIT_DStream_tooFar : Dstream went too far. Decompression result is corrupted.

When reaching end of buffer (BIT_DStream_endOfBuffer), progress slowly, notably if you decode multiple symbols per loop,
to properly detect the exact end of stream.
After each decoded symbol, check if DStream is fully consumed using this simple test :
    BIT_reloadDStream(&DStream) >= BIT_DStream_completed

When it's done, verify decompression is fully completed, by checking both DStream and the relevant states.
Checking if DStream has reached its end is performed by :
    BIT_endOfDStream(&DStream);
Check also the states. There might be some symbols left there, if some high probability ones (>50%) are possible.
    FSE_endOfDState(&DState);
*/


/* *****************************************
*  FSE unsafe API
*******************************************/
static unsigned char FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD);
/* faster, but works only if nbBits is always >= 1 (otherwise, result will be corrupted) */


/* *****************************************
*  Implementation of inlined functions
*******************************************/
typedef struct {
    int deltaFindState;
    U32 deltaNbBits;
} FSE_symbolCompressionTransform; /* total 8 bytes */

MEM_STATIC void FSE_initCState(FSE_CState_t* statePtr, const FSE_CTable* ct)
{
    const void* ptr = ct;
    const U16* u16ptr = (const U16*) ptr;
    const U32 tableLog = MEM_read16(ptr);
    statePtr->value = (ptrdiff_t)1<<tableLog;
    statePtr->stateTable = u16ptr+2;
    statePtr->symbolTT = ct + 1 + (tableLog ? (1<<(tableLog-1)) : 1);
    statePtr->stateLog = tableLog;
}


/*! FSE_initCState2() :
*   Same as FSE_initCState(), but the first symbol to include (which will be the last to be read)
*   uses the smallest state value possible, saving the cost of this symbol */
MEM_STATIC void FSE_initCState2(FSE_CState_t* statePtr, const FSE_CTable* ct, U32 symbol)
{
    FSE_initCState(statePtr, ct);
    {   const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
        const U16* stateTable = (const U16*)(statePtr->stateTable);
        U32 nbBitsOut  = (U32)((symbolTT.deltaNbBits + (1<<15)) >> 16);
        statePtr->value = (nbBitsOut << 16) - symbolTT.deltaNbBits;
        statePtr->value = stateTable[(statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
    }
}

MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, unsigned symbol)
{
    FSE_symbolCompressionTransform const symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
    const U16* const stateTable = (const U16*)(statePtr->stateTable);
    U32 const nbBitsOut  = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);
    BIT_addBits(bitC, (BitContainerType)statePtr->value, nbBitsOut);
    statePtr->value = stateTable[ (statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
}

MEM_STATIC void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* statePtr)
{
    BIT_addBits(bitC, (BitContainerType)statePtr->value, statePtr->stateLog);
    BIT_flushBits(bitC);
}


/* FSE_getMaxNbBits() :
 * Approximate maximum cost of a symbol, in bits.
 * Fractional get rounded up (i.e. a symbol with a normalized frequency of 3 gives the same result as a frequency of 2)
 * note 1 : assume symbolValue is valid (<= maxSymbolValue)
 * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */
MEM_STATIC U32 FSE_getMaxNbBits(const void* symbolTTPtr, U32 symbolValue)
{
    const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;
    return (symbolTT[symbolValue].deltaNbBits + ((1<<16)-1)) >> 16;
}

/* FSE_bitCost() :
 * Approximate symbol cost, as fractional value, using fixed-point format (accuracyLog fractional bits)
 * note 1 : assume symbolValue is valid (<= maxSymbolValue)
 * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */
MEM_STATIC U32 FSE_bitCost(const void* symbolTTPtr, U32 tableLog, U32 symbolValue, U32 accuracyLog)
{
    const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;
    U32 const minNbBits = symbolTT[symbolValue].deltaNbBits >> 16;
    U32 const threshold = (minNbBits+1) << 16;
    assert(tableLog < 16);
    assert(accuracyLog < 31-tableLog);  /* ensure enough room for renormalization double shift */
    {   U32 const tableSize = 1 << tableLog;
        U32 const deltaFromThreshold = threshold - (symbolTT[symbolValue].deltaNbBits + tableSize);
        U32 const normalizedDeltaFromThreshold = (deltaFromThreshold << accuracyLog) >> tableLog;   /* linear interpolation (very approximate) */
        U32 const bitMultiplier = 1 << accuracyLog;
        assert(symbolTT[symbolValue].deltaNbBits + tableSize <= threshold);
        assert(normalizedDeltaFromThreshold <= bitMultiplier);
        return (minNbBits+1)*bitMultiplier - normalizedDeltaFromThreshold;
    }
}


/* ======    Decompression    ====== */

typedef struct {
    U16 tableLog;
    U16 fastMode;
} FSE_DTableHeader;   /* sizeof U32 */

typedef struct
{
    unsigned short newState;
    unsigned char  symbol;
    unsigned char  nbBits;
} FSE_decode_t;   /* size == U32 */

MEM_STATIC void FSE_initDState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD, const FSE_DTable* dt)
{
    const void* ptr = dt;
    const FSE_DTableHeader* const DTableH = (const FSE_DTableHeader*)ptr;
    DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);
    BIT_reloadDStream(bitD);
    DStatePtr->table = dt + 1;
}

MEM_STATIC BYTE FSE_peekSymbol(const FSE_DState_t* DStatePtr)
{
    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
    return DInfo.symbol;
}

MEM_STATIC void FSE_updateState(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD)
{
    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
    U32 const nbBits = DInfo.nbBits;
    size_t const lowBits = BIT_readBits(bitD, nbBits);
    DStatePtr->state = DInfo.newState + lowBits;
}

MEM_STATIC BYTE FSE_decodeSymbol(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD)
{
    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
    U32 const nbBits = DInfo.nbBits;
    BYTE const symbol = DInfo.symbol;
    size_t const lowBits = BIT_readBits(bitD, nbBits);

    DStatePtr->state = DInfo.newState + lowBits;
    return symbol;
}

/*! FSE_decodeSymbolFast() :
    unsafe, only works if no symbol has a probability > 50% */
MEM_STATIC BYTE FSE_decodeSymbolFast(FSE_DState_t* DStatePtr, BIT_DStream_t* bitD)
{
    FSE_decode_t const DInfo = ((const FSE_decode_t*)(DStatePtr->table))[DStatePtr->state];
    U32 const nbBits = DInfo.nbBits;
    BYTE const symbol = DInfo.symbol;
    size_t const lowBits = BIT_readBitsFast(bitD, nbBits);

    DStatePtr->state = DInfo.newState + lowBits;
    return symbol;
}

MEM_STATIC unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr)
{
    return DStatePtr->state == 0;
}



#ifndef FSE_COMMONDEFS_ONLY

/* **************************************************************
*  Tuning parameters
****************************************************************/
/*!MEMORY_USAGE :
*  Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
*  Increasing memory usage improves compression ratio
*  Reduced memory usage can improve speed, due to cache effect
*  Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */
#ifndef FSE_MAX_MEMORY_USAGE
#  define FSE_MAX_MEMORY_USAGE 14
#endif
#ifndef FSE_DEFAULT_MEMORY_USAGE
#  define FSE_DEFAULT_MEMORY_USAGE 13
#endif
#if (FSE_DEFAULT_MEMORY_USAGE > FSE_MAX_MEMORY_USAGE)
#  error "FSE_DEFAULT_MEMORY_USAGE must be <= FSE_MAX_MEMORY_USAGE"
#endif

/*!FSE_MAX_SYMBOL_VALUE :
*  Maximum symbol value authorized.
*  Required for proper stack allocation */
#ifndef FSE_MAX_SYMBOL_VALUE
#  define FSE_MAX_SYMBOL_VALUE 255
#endif

/* **************************************************************
*  template functions type & suffix
****************************************************************/
#define FSE_FUNCTION_TYPE BYTE
#define FSE_FUNCTION_EXTENSION
#define FSE_DECODE_TYPE FSE_decode_t


#endif   /* !FSE_COMMONDEFS_ONLY */


/* ***************************************************************
*  Constants
*****************************************************************/
#define FSE_MAX_TABLELOG  (FSE_MAX_MEMORY_USAGE-2)
#define FSE_MAX_TABLESIZE (1U<<FSE_MAX_TABLELOG)
#define FSE_MAXTABLESIZE_MASK (FSE_MAX_TABLESIZE-1)
#define FSE_DEFAULT_TABLELOG (FSE_DEFAULT_MEMORY_USAGE-2)
#define FSE_MIN_TABLELOG 5

#define FSE_TABLELOG_ABSOLUTE_MAX 15
#if FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX
#  error "FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX is not supported"
#endif

#define FSE_TABLESTEP(tableSize) (((tableSize)>>1) + ((tableSize)>>3) + 3)

#endif /* FSE_STATIC_LINKING_ONLY */
/**** ended inlining fse.h ****/
/**** start inlining huf.h ****/
/* ******************************************************************
 * huff0 huffman codec,
 * part of Finite State Entropy library
 * Copyright (c) Meta Platforms, Inc. and affiliates.
 *
 * You can contact the author at :
 * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
 *
 * 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 HUF_H_298734234
#define HUF_H_298734234

/* *** Dependencies *** */
/**** skipping file: zstd_deps.h ****/
/**** skipping file: mem.h ****/
#define FSE_STATIC_LINKING_ONLY
/**** skipping file: fse.h ****/

/* ***   Tool functions *** */
#define HUF_BLOCKSIZE_MAX (128 * 1024)   /**< maximum input size for a single block compressed with HUF_compress */
size_t HUF_compressBound(size_t size);   /**< maximum compressed size (worst case) */

/* Error Management */
unsigned    HUF_isError(size_t code);       /**< tells if a return value is an error code */
const char* HUF_getErrorName(size_t code);  /**< provides error code string (useful for debugging) */


#define HUF_WORKSPACE_SIZE ((8 << 10) + 512 /* sorting scratch space */)
#define HUF_WORKSPACE_SIZE_U64 (HUF_WORKSPACE_SIZE / sizeof(U64))

/* *** Constants *** */
#define HUF_TABLELOG_MAX      12      /* max runtime value of tableLog (due to static allocation); can be modified up to HUF_TABLELOG_ABSOLUTEMAX */
#define HUF_TABLELOG_DEFAULT  11      /* default tableLog value when none specified */
#define HUF_SYMBOLVALUE_MAX  255

#define HUF_TABLELOG_ABSOLUTEMAX  12  /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */
#if (HUF_TABLELOG_MAX > HUF_TABLELOG_ABSOLUTEMAX)
#  error "HUF_TABLELOG_MAX is too large !"
#endif


/* ****************************************
*  Static allocation
******************************************/
/* HUF buffer bounds */
#define HUF_CTABLEBOUND 129
#define HUF_BLOCKBOUND(size) (size + (size>>8) + 8)   /* only true when incompressible is pre-filtered with fast heuristic */
#define HUF_COMPRESSBOUND(size) (HUF_CTABLEBOUND + HUF_BLOCKBOUND(size))   /* Macro version, useful for static allocation */

/* static allocation of HUF's Compression Table */
/* this is a private definition, just exposed for allocation and strict aliasing purpose. never EVER access its members directly */
typedef size_t HUF_CElt;   /* consider it an incomplete type */
#define HUF_CTABLE_SIZE_ST(maxSymbolValue)   ((maxSymbolValue)+2)   /* Use tables of size_t, for proper alignment */
#define HUF_CTABLE_SIZE(maxSymbolValue)       (HUF_CTABLE_SIZE_ST(maxSymbolValue) * sizeof(size_t))
#define HUF_CREATE_STATIC_CTABLE(name, maxSymbolValue) \
    HUF_CElt name[HUF_CTABLE_SIZE_ST(maxSymbolValue)] /* no final ; */

/* static allocation of HUF's DTable */
typedef U32 HUF_DTable;
#define HUF_DTABLE_SIZE(maxTableLog)   (1 + (1<<(maxTableLog)))
#define HUF_CREATE_STATIC_DTABLEX1(DTable, maxTableLog) \
        HUF_DTable DTable[HUF_DTABLE_SIZE((maxTableLog)-1)] = { ((U32)((maxTableLog)-1) * 0x01000001) }
#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) \
        HUF_DTable DTable[HUF_DTABLE_SIZE(maxTableLog)] = { ((U32)(maxTableLog) * 0x01000001) }


/* ****************************************
*  Advanced decompression functions
******************************************/

/**
 * Huffman flags bitset.
 * For all flags, 0 is the default value.
 */
typedef enum {
    /**
     * If compiled with DYNAMIC_BMI2: Set flag only if the CPU supports BMI2 at runtime.
     * Otherwise: Ignored.
     */
    HUF_flags_bmi2 = (1 << 0),
    /**
     * If set: Test possible table depths to find the one that produces the smallest header + encoded size.
     * If unset: Use heuristic to find the table depth.
     */
    HUF_flags_optimalDepth = (1 << 1),
    /**
     * If set: If the previous table can encode the input, always reuse the previous table.
     * If unset: If the previous table can encode the input, reuse the previous table if it results in a smaller output.
     */
    HUF_flags_preferRepeat = (1 << 2),
    /**
     * If set: Sample the input and check if the sample is uncompressible, if it is then don't attempt to compress.
     * If unset: Always histogram the entire input.
     */
    HUF_flags_suspectUncompressible = (1 << 3),
    /**
     * If set: Don't use assembly implementations
     * If unset: Allow using assembly implementations
     */
    HUF_flags_disableAsm = (1 << 4),
    /**
     * If set: Don't use the fast decoding loop, always use the fallback decoding loop.
     * If unset: Use the fast decoding loop when possible.
     */
    HUF_flags_disableFast = (1 << 5)
} HUF_flags_e;


/* ****************************************
 *  HUF detailed API
 * ****************************************/
#define HUF_OPTIMAL_DEPTH_THRESHOLD ZSTD_btultra

/*! HUF_compress() does the following:
 *  1. count symbol occurrence from source[] into table count[] using FSE_count() (exposed within "fse.h")
 *  2. (optional) refine tableLog using HUF_optimalTableLog()
 *  3. build Huffman table from count using HUF_buildCTable()
 *  4. save Huffman table to memory buffer using HUF_writeCTable()
 *  5. encode the data stream using HUF_compress4X_usingCTable()
 *
 *  The following API allows targeting specific sub-functions for advanced tasks.
 *  For example, it's possible to compress several blocks using the same 'CTable',
 *  or to save and regenerate 'CTable' using external methods.
 */
unsigned HUF_minTableLog(unsigned symbolCardinality);
unsigned HUF_cardinality(const unsigned* count, unsigned maxSymbolValue);
unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, void* workSpace,
 size_t wkspSize, HUF_CElt* table, const unsigned* count, int flags); /* table is used as scratch space for building and testing tables, not a return value */
size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog, void* workspace, size_t workspaceSize);
size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int flags);
size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);

typedef enum {
   HUF_repeat_none,  /**< Cannot use the previous table */
   HUF_repeat_check, /**< Can use the previous table but it must be checked. Note : The previous table must have been constructed by HUF_compress{1, 4}X_repeat */
   HUF_repeat_valid  /**< Can use the previous table and it is assumed to be valid */
 } HUF_repeat;

/** HUF_compress4X_repeat() :
 *  Same as HUF_compress4X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
 *  If it uses hufTable it does not modify hufTable or repeat.
 *  If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.
 *  If preferRepeat then the old table will always be used if valid.
 *  If suspectUncompressible then some sampling checks will be run to potentially skip huffman coding */
size_t HUF_compress4X_repeat(void* dst, size_t dstSize,
                       const void* src, size_t srcSize,
                       unsigned maxSymbolValue, unsigned tableLog,
                       void* workSpace, size_t wkspSize,    /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
                       HUF_CElt* hufTable, HUF_repeat* repeat, int flags);

/** HUF_buildCTable_wksp() :
 *  Same as HUF_buildCTable(), but using externally allocated scratch buffer.
 * `workSpace` must be aligned on 4-bytes boundaries, and its size must be >= HUF_CTABLE_WORKSPACE_SIZE.
 */
#define HUF_CTABLE_WORKSPACE_SIZE_U32 ((4 * (HUF_SYMBOLVALUE_MAX + 1)) + 192)
#define HUF_CTABLE_WORKSPACE_SIZE (HUF_CTABLE_WORKSPACE_SIZE_U32 * sizeof(unsigned))
size_t HUF_buildCTable_wksp (HUF_CElt* tree,
                       const unsigned* count, U32 maxSymbolValue, U32 maxNbBits,
                             void* workSpace, size_t wkspSize);

/*! HUF_readStats() :
 *  Read compact Huffman tree, saved by HUF_writeCTable().
 * `huffWeight` is destination buffer.
 * @return : size read from `src` , or an error Code .
 *  Note : Needed by HUF_readCTable() and HUF_readDTableXn() . */
size_t HUF_readStats(BYTE* huffWeight, size_t hwSize,
                     U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr,
                     const void* src, size_t srcSize);

/*! HUF_readStats_wksp() :
 * Same as HUF_readStats() but takes an external workspace which must be
 * 4-byte aligned and its size must be >= HUF_READ_STATS_WORKSPACE_SIZE.
 * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.
 */
#define HUF_READ_STATS_WORKSPACE_SIZE_U32 FSE_DECOMPRESS_WKSP_SIZE_U32(6, HUF_TABLELOG_MAX-1)
#define HUF_READ_STATS_WORKSPACE_SIZE (HUF_READ_STATS_WORKSPACE_SIZE_U32 * sizeof(unsigned))
size_t HUF_readStats_wksp(BYTE* huffWeight, size_t hwSize,
                          U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr,
                          const void* src, size_t srcSize,
                          void* workspace, size_t wkspSize,
                          int flags);

/** HUF_readCTable() :
 *  Loading a CTable saved with HUF_writeCTable() */
size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize, unsigned *hasZeroWeights);

/** HUF_getNbBitsFromCTable() :
 *  Read nbBits from CTable symbolTable, for symbol `symbolValue` presumed <= HUF_SYMBOLVALUE_MAX
 *  Note 1 : If symbolValue > HUF_readCTableHeader(symbolTable).maxSymbolValue, returns 0
 *  Note 2 : is not inlined, as HUF_CElt definition is private
 */
U32 HUF_getNbBitsFromCTable(const HUF_CElt* symbolTable, U32 symbolValue);

typedef struct {
    BYTE tableLog;
    BYTE maxSymbolValue;
    BYTE unused[sizeof(size_t) - 2];
} HUF_CTableHeader;

/** HUF_readCTableHeader() :
 *  @returns The header from the CTable specifying the tableLog and the maxSymbolValue.
 */
HUF_CTableHeader HUF_readCTableHeader(HUF_CElt const* ctable);

/*
 * HUF_decompress() does the following:
 * 1. select the decompression algorithm (X1, X2) based on pre-computed heuristics
 * 2. build Huffman table from save, using HUF_readDTableX?()
 * 3. decode 1 or 4 segments in parallel using HUF_decompress?X?_usingDTable()
 */

/** HUF_selectDecoder() :
 *  Tells which decoder is likely to decode faster,
 *  based on a set of pre-computed metrics.
 * @return : 0==HUF_decompress4X1, 1==HUF_decompress4X2 .
 *  Assumption : 0 < dstSize <= 128 KB */
U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize);

/**
 *  The minimum workspace size for the `workSpace` used in
 *  HUF_readDTableX1_wksp() and HUF_readDTableX2_wksp().
 *
 *  The space u