diff options
Diffstat (limited to 'thirdparty/zstd/common')
| -rw-r--r-- | thirdparty/zstd/common/bitstream.h | 9 | ||||
| -rw-r--r-- | thirdparty/zstd/common/compiler.h | 22 | ||||
| -rw-r--r-- | thirdparty/zstd/common/fse.h | 2 | ||||
| -rw-r--r-- | thirdparty/zstd/common/fse_decompress.c | 2 | ||||
| -rw-r--r-- | thirdparty/zstd/common/mem.h | 75 | ||||
| -rw-r--r-- | thirdparty/zstd/common/pool.c | 10 | ||||
| -rw-r--r-- | thirdparty/zstd/common/threading.c | 47 | ||||
| -rw-r--r-- | thirdparty/zstd/common/threading.h | 33 | ||||
| -rw-r--r-- | thirdparty/zstd/common/xxhash.c | 10 | ||||
| -rw-r--r-- | thirdparty/zstd/common/zstd_internal.h | 95 |
10 files changed, 231 insertions, 74 deletions
diff --git a/thirdparty/zstd/common/bitstream.h b/thirdparty/zstd/common/bitstream.h index d955bd677b..1c294b80d1 100644 --- a/thirdparty/zstd/common/bitstream.h +++ b/thirdparty/zstd/common/bitstream.h @@ -57,6 +57,8 @@ extern "C" { =========================================*/ #if defined(__BMI__) && defined(__GNUC__) # include <immintrin.h> /* support for bextr (experimental) */ +#elif defined(__ICCARM__) +# include <intrinsics.h> #endif #define STREAM_ACCUMULATOR_MIN_32 25 @@ -162,7 +164,9 @@ MEM_STATIC unsigned BIT_highbit32 (U32 val) _BitScanReverse ( &r, val ); return (unsigned) r; # elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */ - return 31 - __builtin_clz (val); + return __builtin_clz (val) ^ 31; +# elif defined(__ICCARM__) /* IAR Intrinsic */ + return 31 - __CLZ(val); # else /* Software version */ static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, @@ -240,9 +244,9 @@ 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; - assert(bitC->ptr <= bitC->endPtr); bitC->bitPos &= 7; bitC->bitContainer >>= nbBytes*8; } @@ -256,6 +260,7 @@ 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; diff --git a/thirdparty/zstd/common/compiler.h b/thirdparty/zstd/common/compiler.h index 87bf51ae8c..1877a0c1d9 100644 --- a/thirdparty/zstd/common/compiler.h +++ b/thirdparty/zstd/common/compiler.h @@ -23,7 +23,7 @@ # define INLINE_KEYWORD #endif -#if defined(__GNUC__) +#if defined(__GNUC__) || defined(__ICCARM__) # define FORCE_INLINE_ATTR __attribute__((always_inline)) #elif defined(_MSC_VER) # define FORCE_INLINE_ATTR __forceinline @@ -61,11 +61,18 @@ # define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR #endif +/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */ +#if defined(__GNUC__) +# define UNUSED_ATTR __attribute__((unused)) +#else +# define UNUSED_ATTR +#endif + /* force no inlining */ #ifdef _MSC_VER # define FORCE_NOINLINE static __declspec(noinline) #else -# ifdef __GNUC__ +# if defined(__GNUC__) || defined(__ICCARM__) # define FORCE_NOINLINE static __attribute__((__noinline__)) # else # define FORCE_NOINLINE static @@ -76,7 +83,7 @@ #ifndef __has_attribute #define __has_attribute(x) 0 /* Compatibility with non-clang compilers. */ #endif -#if defined(__GNUC__) +#if defined(__GNUC__) || defined(__ICCARM__) # define TARGET_ATTRIBUTE(target) __attribute__((__target__(target))) #else # define TARGET_ATTRIBUTE(target) @@ -127,9 +134,14 @@ } \ } -/* vectorization */ +/* vectorization + * older GCC (pre gcc-4.3 picked as the cutoff) uses a different syntax */ #if !defined(__clang__) && defined(__GNUC__) -# define DONT_VECTORIZE __attribute__((optimize("no-tree-vectorize"))) +# 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 diff --git a/thirdparty/zstd/common/fse.h b/thirdparty/zstd/common/fse.h index 811c670bdd..a7553e3721 100644 --- a/thirdparty/zstd/common/fse.h +++ b/thirdparty/zstd/common/fse.h @@ -308,7 +308,7 @@ If there is an error, the function will return an error code, which can be teste *******************************************/ /* FSE buffer bounds */ #define FSE_NCOUNTBOUND 512 -#define FSE_BLOCKBOUND(size) (size + (size>>7)) +#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 */ diff --git a/thirdparty/zstd/common/fse_decompress.c b/thirdparty/zstd/common/fse_decompress.c index 72bbead5be..4f07378982 100644 --- a/thirdparty/zstd/common/fse_decompress.c +++ b/thirdparty/zstd/common/fse_decompress.c @@ -52,7 +52,9 @@ #define FSE_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c) /* use only *after* variable declarations */ /* check and forward error code */ +#ifndef CHECK_F #define CHECK_F(f) { size_t const e = f; if (FSE_isError(e)) return e; } +#endif /* ************************************************************** diff --git a/thirdparty/zstd/common/mem.h b/thirdparty/zstd/common/mem.h index 5da248756f..530d30c8f7 100644 --- a/thirdparty/zstd/common/mem.h +++ b/thirdparty/zstd/common/mem.h @@ -47,6 +47,79 @@ extern "C" { #define MEM_STATIC_ASSERT(c) { enum { MEM_static_assert = 1/(int)(!!(c)) }; } MEM_STATIC void MEM_check(void) { MEM_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); } +/* detects whether we are being compiled under msan */ +#if defined (__has_feature) +# if __has_feature(memory_sanitizer) +# define MEMORY_SANITIZER 1 +# endif +#endif + +#if defined (MEMORY_SANITIZER) +/* 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 <stdint.h> /* intptr_t */ + +/* 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); +#endif + +/* detects whether we are being compiled under asan */ +#if defined (__has_feature) +# if __has_feature(address_sanitizer) +# define ADDRESS_SANITIZER 1 +# endif +#elif defined(__SANITIZE_ADDRESS__) +# define ADDRESS_SANITIZER 1 +#endif + +#if defined (ADDRESS_SANITIZER) +/* 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... */ + +/** + * 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 + /*-************************************************************** * Basic Types @@ -102,7 +175,7 @@ MEM_STATIC void MEM_check(void) { MEM_STATIC_ASSERT((sizeof(size_t)==4) || (size #ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ # if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) ) # define MEM_FORCE_MEMORY_ACCESS 2 -# elif defined(__INTEL_COMPILER) || defined(__GNUC__) +# elif defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__) # define MEM_FORCE_MEMORY_ACCESS 1 # endif #endif diff --git a/thirdparty/zstd/common/pool.c b/thirdparty/zstd/common/pool.c index 7a82945432..f575935076 100644 --- a/thirdparty/zstd/common/pool.c +++ b/thirdparty/zstd/common/pool.c @@ -127,9 +127,13 @@ POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize, ctx->queueTail = 0; ctx->numThreadsBusy = 0; ctx->queueEmpty = 1; - (void)ZSTD_pthread_mutex_init(&ctx->queueMutex, NULL); - (void)ZSTD_pthread_cond_init(&ctx->queuePushCond, NULL); - (void)ZSTD_pthread_cond_init(&ctx->queuePopCond, NULL); + { + int error = 0; + error |= ZSTD_pthread_mutex_init(&ctx->queueMutex, NULL); + error |= ZSTD_pthread_cond_init(&ctx->queuePushCond, NULL); + error |= ZSTD_pthread_cond_init(&ctx->queuePopCond, NULL); + if (error) { POOL_free(ctx); return NULL; } + } ctx->shutdown = 0; /* Allocate space for the thread handles */ ctx->threads = (ZSTD_pthread_t*)ZSTD_malloc(numThreads * sizeof(ZSTD_pthread_t), customMem); diff --git a/thirdparty/zstd/common/threading.c b/thirdparty/zstd/common/threading.c index f3d4fa8418..482664bd9a 100644 --- a/thirdparty/zstd/common/threading.c +++ b/thirdparty/zstd/common/threading.c @@ -14,6 +14,8 @@ * This file will hold wrapper for systems, which do not support pthreads */ +#include "threading.h" + /* create fake symbol to avoid empty translation unit warning */ int g_ZSTD_threading_useless_symbol; @@ -28,7 +30,6 @@ int g_ZSTD_threading_useless_symbol; /* === Dependencies === */ #include <process.h> #include <errno.h> -#include "threading.h" /* === Implementation === */ @@ -73,3 +74,47 @@ int ZSTD_pthread_join(ZSTD_pthread_t thread, void **value_ptr) } #endif /* ZSTD_MULTITHREAD */ + +#if defined(ZSTD_MULTITHREAD) && DEBUGLEVEL >= 1 && !defined(_WIN32) + +#include <stdlib.h> + +int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t const* attr) +{ + *mutex = (pthread_mutex_t*)malloc(sizeof(pthread_mutex_t)); + if (!*mutex) + return 1; + return pthread_mutex_init(*mutex, attr); +} + +int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex) +{ + if (!*mutex) + return 0; + { + int const ret = pthread_mutex_destroy(*mutex); + free(*mutex); + return ret; + } +} + +int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const* attr) +{ + *cond = (pthread_cond_t*)malloc(sizeof(pthread_cond_t)); + if (!*cond) + return 1; + return pthread_cond_init(*cond, attr); +} + +int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond) +{ + if (!*cond) + return 0; + { + int const ret = pthread_cond_destroy(*cond); + free(*cond); + return ret; + } +} + +#endif diff --git a/thirdparty/zstd/common/threading.h b/thirdparty/zstd/common/threading.h index d806c89d01..3193ca7db8 100644 --- a/thirdparty/zstd/common/threading.h +++ b/thirdparty/zstd/common/threading.h @@ -13,6 +13,8 @@ #ifndef THREADING_H_938743 #define THREADING_H_938743 +#include "debug.h" + #if defined (__cplusplus) extern "C" { #endif @@ -75,10 +77,12 @@ int ZSTD_pthread_join(ZSTD_pthread_t thread, void** value_ptr); */ -#elif defined(ZSTD_MULTITHREAD) /* posix assumed ; need a better detection method */ +#elif defined(ZSTD_MULTITHREAD) /* posix assumed ; need a better detection method */ /* === POSIX Systems === */ # include <pthread.h> +#if DEBUGLEVEL < 1 + #define ZSTD_pthread_mutex_t pthread_mutex_t #define ZSTD_pthread_mutex_init(a, b) pthread_mutex_init((a), (b)) #define ZSTD_pthread_mutex_destroy(a) pthread_mutex_destroy((a)) @@ -96,6 +100,33 @@ int ZSTD_pthread_join(ZSTD_pthread_t thread, void** value_ptr); #define ZSTD_pthread_create(a, b, c, d) pthread_create((a), (b), (c), (d)) #define ZSTD_pthread_join(a, b) pthread_join((a),(b)) +#else /* DEBUGLEVEL >= 1 */ + +/* Debug implementation of threading. + * In this implementation we use pointers for mutexes and condition variables. + * This way, if we forget to init/destroy them the program will crash or ASAN + * will report leaks. + */ + +#define ZSTD_pthread_mutex_t pthread_mutex_t* +int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t const* attr); +int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex); +#define ZSTD_pthread_mutex_lock(a) pthread_mutex_lock(*(a)) +#define ZSTD_pthread_mutex_unlock(a) pthread_mutex_unlock(*(a)) + +#define ZSTD_pthread_cond_t pthread_cond_t* +int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const* attr); +int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond); +#define ZSTD_pthread_cond_wait(a, b) pthread_cond_wait(*(a), *(b)) +#define ZSTD_pthread_cond_signal(a) pthread_cond_signal(*(a)) +#define ZSTD_pthread_cond_broadcast(a) pthread_cond_broadcast(*(a)) + +#define ZSTD_pthread_t pthread_t +#define ZSTD_pthread_create(a, b, c, d) pthread_create((a), (b), (c), (d)) +#define ZSTD_pthread_join(a, b) pthread_join((a),(b)) + +#endif + #else /* ZSTD_MULTITHREAD not defined */ /* No multithreading support */ diff --git a/thirdparty/zstd/common/xxhash.c b/thirdparty/zstd/common/xxhash.c index 30599aaae4..99d2459621 100644 --- a/thirdparty/zstd/common/xxhash.c +++ b/thirdparty/zstd/common/xxhash.c @@ -53,7 +53,8 @@ # if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) ) # define XXH_FORCE_MEMORY_ACCESS 2 # elif (defined(__INTEL_COMPILER) && !defined(WIN32)) || \ - (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) + (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) || \ + defined(__ICCARM__) # define XXH_FORCE_MEMORY_ACCESS 1 # endif #endif @@ -120,7 +121,7 @@ static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcp # define INLINE_KEYWORD #endif -#if defined(__GNUC__) +#if defined(__GNUC__) || defined(__ICCARM__) # define FORCE_INLINE_ATTR __attribute__((always_inline)) #elif defined(_MSC_VER) # define FORCE_INLINE_ATTR __forceinline @@ -206,7 +207,12 @@ static U64 XXH_read64(const void* memPtr) # define XXH_rotl32(x,r) _rotl(x,r) # define XXH_rotl64(x,r) _rotl64(x,r) #else +#if defined(__ICCARM__) +# include <intrinsics.h> +# define XXH_rotl32(x,r) __ROR(x,(32 - r)) +#else # define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r))) +#endif # define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r))) #endif diff --git a/thirdparty/zstd/common/zstd_internal.h b/thirdparty/zstd/common/zstd_internal.h index 81b16eac2e..dcdcbdb81c 100644 --- a/thirdparty/zstd/common/zstd_internal.h +++ b/thirdparty/zstd/common/zstd_internal.h @@ -56,9 +56,9 @@ extern "C" { /** * 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(). + * 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, ...) \ if (cond) { \ @@ -197,79 +197,56 @@ static void ZSTD_copy8(void* dst, const void* src) { memcpy(dst, src, 8); } static void ZSTD_copy16(void* dst, const void* src) { memcpy(dst, src, 16); } #define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; } -#define WILDCOPY_OVERLENGTH 8 -#define VECLEN 16 +#define WILDCOPY_OVERLENGTH 32 +#define WILDCOPY_VECLEN 16 typedef enum { ZSTD_no_overlap, - ZSTD_overlap_src_before_dst, + ZSTD_overlap_src_before_dst /* ZSTD_overlap_dst_before_src, */ } ZSTD_overlap_e; /*! ZSTD_wildcopy() : - * custom version of memcpy(), can overwrite up to WILDCOPY_OVERLENGTH bytes (if length==0) */ + * Custom version of memcpy(), can over read/write up to WILDCOPY_OVERLENGTH bytes (if length==0) + * @param ovtype controls the overlap detection + * - ZSTD_no_overlap: The source and destination are guaranteed to be at least WILDCOPY_VECLEN bytes apart. + * - ZSTD_overlap_src_before_dst: The src and dst may overlap, but they MUST be at least 8 bytes apart. + * The src buffer must be before the dst buffer. + */ MEM_STATIC FORCE_INLINE_ATTR DONT_VECTORIZE -void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e ovtype) +void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e const ovtype) { ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src; const BYTE* ip = (const BYTE*)src; BYTE* op = (BYTE*)dst; BYTE* const oend = op + length; - assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff < -8)); - if (length < VECLEN || (ovtype == ZSTD_overlap_src_before_dst && diff < VECLEN)) { - do - COPY8(op, ip) - while (op < oend); - } - else { - if ((length & 8) == 0) - COPY8(op, ip); - do { + assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff <= -WILDCOPY_VECLEN)); + + if (ovtype == ZSTD_overlap_src_before_dst && diff < WILDCOPY_VECLEN) { + /* Handle short offset copies. */ + do { + COPY8(op, ip) + } while (op < oend); + } else { + assert(diff >= WILDCOPY_VECLEN || diff <= -WILDCOPY_VECLEN); + /* Separate out the first two COPY16() calls because the copy length is + * almost certain to be short, so the branches have different + * probabilities. + * On gcc-9 unrolling once is +1.6%, twice is +2%, thrice is +1.8%. + * On clang-8 unrolling once is +1.4%, twice is +3.3%, thrice is +3%. + */ COPY16(op, ip); - } - while (op < oend); - } -} - -/*! ZSTD_wildcopy_16min() : - * same semantics as ZSTD_wilcopy() except guaranteed to be able to copy 16 bytes at the start */ -MEM_STATIC FORCE_INLINE_ATTR DONT_VECTORIZE -void ZSTD_wildcopy_16min(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e ovtype) -{ - ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src; - const BYTE* ip = (const BYTE*)src; - BYTE* op = (BYTE*)dst; - BYTE* const oend = op + length; - - assert(length >= 8); - assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff < -8)); - - if (ovtype == ZSTD_overlap_src_before_dst && diff < VECLEN) { - do - COPY8(op, ip) - while (op < oend); - } - else { - if ((length & 8) == 0) - COPY8(op, ip); - do { COPY16(op, ip); - } - while (op < oend); + if (op >= oend) return; + do { + COPY16(op, ip); + COPY16(op, ip); + } + while (op < oend); } } -MEM_STATIC void ZSTD_wildcopy_e(void* dst, const void* src, void* dstEnd) /* should be faster for decoding, but strangely, not verified on all platform */ -{ - const BYTE* ip = (const BYTE*)src; - BYTE* op = (BYTE*)dst; - BYTE* const oend = (BYTE*)dstEnd; - do - COPY8(op, ip) - while (op < oend); -} - /*-******************************************* * Private declarations @@ -323,7 +300,9 @@ MEM_STATIC U32 ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus _BitScanReverse(&r, val); return (unsigned)r; # elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */ - return 31 - __builtin_clz(val); + return __builtin_clz (val) ^ 31; +# elif defined(__ICCARM__) /* IAR Intrinsic */ + return 31 - __CLZ(val); # else /* Software version */ static const U32 DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 }; U32 v = val; |