zstd-ruby 1.4.2.0 → 1.4.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 69e6ba233f94b8ed930f34217befa3268b0bd94fb8116130066e7976c0b31d23
-  data.tar.gz: 7d01847ddb0a3a5e1996eb4a45bb2c84ad9230f5b9c002dea32d202036be15d6
+  metadata.gz: 53be8076094a9ed214cbc86a174eeb3b587dc3d4781f7fbcf8ee280ffd0ca169
+  data.tar.gz: ebf9bcf8d062447dab0589c1acfe63f2481f9445c78f51653daf1f60750dfdd5
 SHA512:
-  metadata.gz: 7122bbd06a6ff9a52240c8d0b708f083c6567e4b586affd3c48bc17e412cb6beaaf097288627ded338787c4c7a940ec5dab32f704db54c29a7cba7fde848dca8
-  data.tar.gz: 1839b136a9b3c181cfc3f18cce8f07b96eaa2e537d0753d8f085243ea8f9b91019c2c59174a1a2a5ca2419ec39a23542faf175ba685c70b0543597f461c1d937
+  metadata.gz: 7f3ac7650f5ec553aea615d60737bc092bca9d06c998eecc0570b5d542344f4cfc703335154e845c53927878d424ca244fa494f4bddb3e59f01bbd6111de5172
+  data.tar.gz: b96f8df37df844f461fab2cd1b60fe9b6a729a1a411f62f1dcef34c9684672ebd7c42963d79abdf5fe1e8d5d443104509492fe9128a7dbf3773e9db180592cbd

data/README.md

@@ -10,7 +10,7 @@ See https://github.com/facebook/zstd
 Fork from https://github.com/jarredholman/ruby-zstd.
 
 ## Zstd version
-v1.4.2 (https://github.com/facebook/zstd/tree/v1.4.2)
+v1.4.4 (https://github.com/facebook/zstd/tree/v1.4.4)
 
 ## Installation
 

data/ext/zstdruby/libzstd/Makefile

@@ -244,8 +244,6 @@ libzstd.pc:
 libzstd.pc: libzstd.pc.in
 	@echo creating pkgconfig
 	@sed -e 's|@PREFIX@|$(PREFIX)|' \
-             -e 's|@LIBDIR@|$(LIBDIR)|' \
-             -e 's|@INCLUDEDIR@|$(INCLUDEDIR)|' \
              -e 's|@VERSION@|$(VERSION)|' \
              $< >$@
 

data/ext/zstdruby/libzstd/README.md

@@ -27,10 +27,10 @@ Enabling multithreading requires 2 conditions :
 Both conditions are automatically applied when invoking `make lib-mt` target.
 
 When linking a POSIX program with a multithreaded version of `libzstd`,
-note that it's necessary to request the `-pthread` flag during link stage.
+note that it's necessary to invoke the `-pthread` flag during link stage.
 
 Multithreading capabilities are exposed
-via the [advanced API defined in `lib/zstd.h`](https://github.com/facebook/zstd/blob/v1.3.8/lib/zstd.h#L592).
+via the [advanced API defined in `lib/zstd.h`](https://github.com/facebook/zstd/blob/v1.4.3/lib/zstd.h#L351).
 
 
 #### API
@@ -112,6 +112,17 @@ The file structure is designed to make this selection manually achievable for an
   will expose the deprecated `ZSTDMT` API exposed by `zstdmt_compress.h` in
   the shared library, which is now hidden by default.
 
+- The build macro `DYNAMIC_BMI2` can be set to 1 or 0 in order to generate binaries
+  which can detect at runtime the presence of BMI2 instructions, and use them only if present.
+  These instructions contribute to better performance, notably on the decoder side.
+  By default, this feature is automatically enabled on detecting
+  the right instruction set (x64) and compiler (clang or gcc >= 5).
+  It's obviously disabled for different cpus,
+  or when BMI2 instruction set is _required_ by the compiler command line
+  (in this case, only the BMI2 code path is generated).
+  Setting this macro will either force to generate the BMI2 dispatcher (1)
+  or prevent it (0). It overrides automatic detection.
+
 
 #### Windows : using MinGW+MSYS to create DLL
 

data/ext/zstdruby/libzstd/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;

data/ext/zstdruby/libzstd/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

data/ext/zstdruby/libzstd/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 */

data/ext/zstdruby/libzstd/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
 
 
 /* **************************************************************

data/ext/zstdruby/libzstd/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

data/ext/zstdruby/libzstd/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);

data/ext/zstdruby/libzstd/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

data/ext/zstdruby/libzstd/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 */
 

data/ext/zstdruby/libzstd/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
 

data/ext/zstdruby/libzstd/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;