extzstd 0.3.1 → 0.3.3

data/contrib/zstd/lib/common/mem.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2016-2020, Yann Collet, Facebook, Inc.
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  * All rights reserved.
  *
  * This source code is licensed under both the BSD-style license (found in the
@@ -18,8 +18,10 @@ extern "C" {
 /*-****************************************
 *  Dependencies
 ******************************************/
-#include <stddef.h>     /* size_t, ptrdiff_t */
-#include <string.h>     /* memcpy */
+#include <stddef.h>  /* size_t, ptrdiff_t */
+#include "compiler.h"  /* __has_builtin */
+#include "debug.h"  /* DEBUG_STATIC_ASSERT */
+#include "zstd_deps.h"  /* ZSTD_memcpy */
 
 
 /*-****************************************
@@ -39,94 +41,18 @@ extern "C" {
 #  define MEM_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
 #endif
 
-#ifndef __has_builtin
-#  define __has_builtin(x) 0  /* compat. with non-clang compilers */
-#endif
-
-/* code only tested on 32 and 64 bits systems */
-#define MEM_STATIC_ASSERT(c)   { enum { MEM_static_assert = 1/(int)(!!(c)) }; }
-MEM_STATIC void MEM_check(void) { MEM_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }
-
-/* 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
 *****************************************************************/
 #if  !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
-# include <stdint.h>
+#  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;
@@ -139,6 +65,8 @@ void __asan_unpoison_memory_region(void const volatile *addr, size_t size);
 #  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
@@ -157,25 +85,63 @@ void __asan_unpoison_memory_region(void const volatile *addr, size_t size);
 
 
 /*-**************************************************************
-*  Memory I/O
+*  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 :
- * By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
- * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
- * The below switch allow to select different access method for improved performance.
- * Method 0 (default) : use `memcpy()`. Safe and portable.
- * Method 1 : `__packed` statement. It depends on compiler extension (i.e., not portable).
- *            This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
+/* 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.
- *            In some circumstances, it's the only known way to get the most performance (i.e. GCC + ARMv6)
- * See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
- * Prefer these methods in priority order (0 > 1 > 2)
+ * Default  : method 1 if supported, else method 0
  */
 #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__) || defined(__ICCARM__)
+#  ifdef __GNUC__
 #    define MEM_FORCE_MEMORY_ACCESS 1
 #  endif
 #endif
@@ -185,8 +151,22 @@ 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_AMD64 || _M_IX86)
+    return 1;
+#elif defined(__DMC__) && defined(_M_IX86)
+    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)
@@ -204,30 +184,19 @@ MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; }
 
 #elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
 
-/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
-/* currently only defined for gcc and icc */
-#if defined(_MSC_VER) || (defined(__INTEL_COMPILER) && defined(WIN32))
-    __pragma( pack(push, 1) )
-    typedef struct { U16 v; } unalign16;
-    typedef struct { U32 v; } unalign32;
-    typedef struct { U64 v; } unalign64;
-    typedef struct { size_t v; } unalignArch;
-    __pragma( pack(pop) )
-#else
-    typedef struct { U16 v; } __attribute__((packed)) unalign16;
-    typedef struct { U32 v; } __attribute__((packed)) unalign32;
-    typedef struct { U64 v; } __attribute__((packed)) unalign64;
-    typedef struct { size_t v; } __attribute__((packed)) unalignArch;
-#endif
+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)->v; }
-MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign32*)ptr)->v; }
-MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign64*)ptr)->v; }
-MEM_STATIC size_t MEM_readST(const void* ptr) { return ((const unalignArch*)ptr)->v; }
+MEM_STATIC 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)->v = value; }
-MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign32*)memPtr)->v = value; }
-MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign64*)memPtr)->v = value; }
+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
 
@@ -236,41 +205,49 @@ MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign64*)memPtr)->v =
 
 MEM_STATIC U16 MEM_read16(const void* memPtr)
 {
-    U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
+    U16 val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
 }
 
 MEM_STATIC U32 MEM_read32(const void* memPtr)
 {
-    U32 val; memcpy(&val, memPtr, sizeof(val)); return val;
+    U32 val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
 }
 
 MEM_STATIC U64 MEM_read64(const void* memPtr)
 {
-    U64 val; memcpy(&val, memPtr, sizeof(val)); return val;
+    U64 val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
 }
 
 MEM_STATIC size_t MEM_readST(const void* memPtr)
 {
-    size_t val; memcpy(&val, memPtr, sizeof(val)); return val;
+    size_t val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
 }
 
 MEM_STATIC void MEM_write16(void* memPtr, U16 value)
 {
-    memcpy(memPtr, &value, sizeof(value));
+    ZSTD_memcpy(memPtr, &value, sizeof(value));
 }
 
 MEM_STATIC void MEM_write32(void* memPtr, U32 value)
 {
-    memcpy(memPtr, &value, sizeof(value));
+    ZSTD_memcpy(memPtr, &value, sizeof(value));
 }
 
 MEM_STATIC void MEM_write64(void* memPtr, U64 value)
 {
-    memcpy(memPtr, &value, sizeof(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 */
@@ -279,22 +256,13 @@ MEM_STATIC U32 MEM_swap32(U32 in)
   || (defined(__clang__) && __has_builtin(__builtin_bswap32))
     return __builtin_bswap32(in);
 #else
-    return  ((in << 24) & 0xff000000 ) |
-            ((in <<  8) & 0x00ff0000 ) |
-            ((in >>  8) & 0x0000ff00 ) |
-            ((in >> 24) & 0x000000ff );
+    return MEM_swap32_fallback(in);
 #endif
 }
 
-MEM_STATIC U64 MEM_swap64(U64 in)
+MEM_STATIC U64 MEM_swap64_fallback(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  ((in << 56) & 0xff00000000000000ULL) |
+     return  ((in << 56) & 0xff00000000000000ULL) |
             ((in << 40) & 0x00ff000000000000ULL) |
             ((in << 24) & 0x0000ff0000000000ULL) |
             ((in << 8)  & 0x000000ff00000000ULL) |
@@ -302,6 +270,17 @@ MEM_STATIC U64 MEM_swap64(U64 in)
             ((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
 }
 
@@ -338,7 +317,7 @@ MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val)
 
 MEM_STATIC U32 MEM_readLE24(const void* memPtr)
 {
-    return MEM_readLE16(memPtr) + (((const BYTE*)memPtr)[2] << 16);
+    return (U32)MEM_readLE16(memPtr) + ((U32)(((const BYTE*)memPtr)[2]) << 16);
 }
 
 MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val)
@@ -445,6 +424,9 @@ MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val)
         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)); }
+
 
 #if defined (__cplusplus)
 }

data/contrib/zstd/lib/common/pool.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2016-2020, Yann Collet, Facebook, Inc.
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  * All rights reserved.
  *
  * This source code is licensed under both the BSD-style license (found in the
@@ -10,9 +10,9 @@
 
 
 /* ======   Dependencies   ======= */
-#include <stddef.h>    /* size_t */
+#include "../common/allocations.h"  /* ZSTD_customCalloc, ZSTD_customFree */
+#include "zstd_deps.h" /* size_t */
 #include "debug.h"     /* assert */
-#include "zstd_internal.h"  /* ZSTD_malloc, ZSTD_free */
 #include "pool.h"
 
 /* ======   Compiler specifics   ====== */
@@ -86,7 +86,7 @@ static void* POOL_thread(void* opaque) {
         {   POOL_job const job = ctx->queue[ctx->queueHead];
             ctx->queueHead = (ctx->queueHead + 1) % ctx->queueSize;
             ctx->numThreadsBusy++;
-            ctx->queueEmpty = ctx->queueHead == ctx->queueTail;
+            ctx->queueEmpty = (ctx->queueHead == ctx->queueTail);
             /* Unlock the mutex, signal a pusher, and run the job */
             ZSTD_pthread_cond_signal(&ctx->queuePushCond);
             ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
@@ -96,33 +96,37 @@ static void* POOL_thread(void* opaque) {
             /* If the intended queue size was 0, signal after finishing job */
             ZSTD_pthread_mutex_lock(&ctx->queueMutex);
             ctx->numThreadsBusy--;
-            if (ctx->queueSize == 1) {
-                ZSTD_pthread_cond_signal(&ctx->queuePushCond);
-            }
+            ZSTD_pthread_cond_signal(&ctx->queuePushCond);
             ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
         }
     }  /* for (;;) */
     assert(0);  /* Unreachable */
 }
 
+/* ZSTD_createThreadPool() : public access point */
+POOL_ctx* ZSTD_createThreadPool(size_t numThreads) {
+    return POOL_create (numThreads, 0);
+}
+
 POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {
     return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);
 }
 
 POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
-                               ZSTD_customMem customMem) {
+                               ZSTD_customMem customMem)
+{
     POOL_ctx* ctx;
     /* Check parameters */
     if (!numThreads) { return NULL; }
     /* Allocate the context and zero initialize */
-    ctx = (POOL_ctx*)ZSTD_calloc(sizeof(POOL_ctx), customMem);
+    ctx = (POOL_ctx*)ZSTD_customCalloc(sizeof(POOL_ctx), customMem);
     if (!ctx) { return NULL; }
     /* Initialize the job queue.
      * It needs one extra space since one space is wasted to differentiate
      * empty and full queues.
      */
     ctx->queueSize = queueSize + 1;
-    ctx->queue = (POOL_job*)ZSTD_malloc(ctx->queueSize * sizeof(POOL_job), customMem);
+    ctx->queue = (POOL_job*)ZSTD_customCalloc(ctx->queueSize * sizeof(POOL_job), customMem);
     ctx->queueHead = 0;
     ctx->queueTail = 0;
     ctx->numThreadsBusy = 0;
@@ -136,7 +140,7 @@ POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
     }
     ctx->shutdown = 0;
     /* Allocate space for the thread handles */
-    ctx->threads = (ZSTD_pthread_t*)ZSTD_malloc(numThreads * sizeof(ZSTD_pthread_t), customMem);
+    ctx->threads = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), customMem);
     ctx->threadCapacity = 0;
     ctx->customMem = customMem;
     /* Check for errors */
@@ -169,7 +173,7 @@ static void POOL_join(POOL_ctx* ctx) {
     /* Join all of the threads */
     {   size_t i;
         for (i = 0; i < ctx->threadCapacity; ++i) {
-            ZSTD_pthread_join(ctx->threads[i], NULL);  /* note : could fail */
+            ZSTD_pthread_join(ctx->threads[i]);  /* note : could fail */
     }   }
 }
 
@@ -179,14 +183,27 @@ void POOL_free(POOL_ctx *ctx) {
     ZSTD_pthread_mutex_destroy(&ctx->queueMutex);
     ZSTD_pthread_cond_destroy(&ctx->queuePushCond);
     ZSTD_pthread_cond_destroy(&ctx->queuePopCond);
-    ZSTD_free(ctx->queue, ctx->customMem);
-    ZSTD_free(ctx->threads, ctx->customMem);
-    ZSTD_free(ctx, ctx->customMem);
+    ZSTD_customFree(ctx->queue, ctx->customMem);
+    ZSTD_customFree(ctx->threads, ctx->customMem);
+    ZSTD_customFree(ctx, ctx->customMem);
 }
 
+/*! POOL_joinJobs() :
+ *  Waits for all queued jobs to finish executing.
+ */
+void POOL_joinJobs(POOL_ctx* ctx) {
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+    while(!ctx->queueEmpty || ctx->numThreadsBusy > 0) {
+        ZSTD_pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);
+    }
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+}
 
+void ZSTD_freeThreadPool (ZSTD_threadPool* pool) {
+  POOL_free (pool);
+}
 
-size_t POOL_sizeof(POOL_ctx *ctx) {
+size_t POOL_sizeof(const POOL_ctx* ctx) {
     if (ctx==NULL) return 0;  /* supports sizeof NULL */
     return sizeof(*ctx)
         + ctx->queueSize * sizeof(POOL_job)
@@ -203,11 +220,11 @@ static int POOL_resize_internal(POOL_ctx* ctx, size_t numThreads)
         return 0;
     }
     /* numThreads > threadCapacity */
-    {   ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_malloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
+    {   ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
         if (!threadPool) return 1;
         /* replace existing thread pool */
-        memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));
-        ZSTD_free(ctx->threads, ctx->customMem);
+        ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));
+        ZSTD_customFree(ctx->threads, ctx->customMem);
         ctx->threads = threadPool;
         /* Initialize additional threads */
         {   size_t threadId;
@@ -251,9 +268,12 @@ static int isQueueFull(POOL_ctx const* ctx) {
 }
 
 
-static void POOL_add_internal(POOL_ctx* ctx, POOL_function function, void *opaque)
+static void
+POOL_add_internal(POOL_ctx* ctx, POOL_function function, void *opaque)
 {
-    POOL_job const job = {function, opaque};
+    POOL_job job;
+    job.function = function;
+    job.opaque = opaque;
     assert(ctx != NULL);
     if (ctx->shutdown) return;
 
@@ -301,21 +321,28 @@ int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque)
 struct POOL_ctx_s {
     int dummy;
 };
-static POOL_ctx g_ctx;
+static POOL_ctx g_poolCtx;
 
 POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {
     return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);
 }
 
-POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize, ZSTD_customMem customMem) {
+POOL_ctx*
+POOL_create_advanced(size_t numThreads, size_t queueSize, ZSTD_customMem customMem)
+{
     (void)numThreads;
     (void)queueSize;
     (void)customMem;
-    return &g_ctx;
+    return &g_poolCtx;
 }
 
 void POOL_free(POOL_ctx* ctx) {
-    assert(!ctx || ctx == &g_ctx);
+    assert(!ctx || ctx == &g_poolCtx);
+    (void)ctx;
+}
+
+void POOL_joinJobs(POOL_ctx* ctx){
+    assert(!ctx || ctx == &g_poolCtx);
     (void)ctx;
 }
 
@@ -335,9 +362,9 @@ int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque) {
     return 1;
 }
 
-size_t POOL_sizeof(POOL_ctx* ctx) {
+size_t POOL_sizeof(const POOL_ctx* ctx) {
     if (ctx==NULL) return 0;  /* supports sizeof NULL */
-    assert(ctx == &g_ctx);
+    assert(ctx == &g_poolCtx);
     return sizeof(*ctx);
 }
 

data/contrib/zstd/lib/common/pool.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2016-2020, Yann Collet, Facebook, Inc.
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
  * All rights reserved.
  *
  * This source code is licensed under both the BSD-style license (found in the
@@ -16,7 +16,7 @@ extern "C" {
 #endif
 
 
-#include <stddef.h>   /* size_t */
+#include "zstd_deps.h"
 #define ZSTD_STATIC_LINKING_ONLY   /* ZSTD_customMem */
 #include "../zstd.h"
 
@@ -38,6 +38,12 @@ POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
  */
 void POOL_free(POOL_ctx* ctx);
 
+
+/*! POOL_joinJobs() :
+ *  Waits for all queued jobs to finish executing.
+ */
+void POOL_joinJobs(POOL_ctx* ctx);
+
 /*! POOL_resize() :
  *  Expands or shrinks pool's number of threads.
  *  This is more efficient than releasing + creating a new context,
@@ -53,7 +59,7 @@ int POOL_resize(POOL_ctx* ctx, size_t numThreads);
  * @return threadpool memory usage
  *  note : compatible with NULL (returns 0 in this case)
  */
-size_t POOL_sizeof(POOL_ctx* ctx);
+size_t POOL_sizeof(const POOL_ctx* ctx);
 
 /*! POOL_function :
  *  The function type that can be added to a thread pool.
@@ -70,7 +76,7 @@ void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque);
 
 
 /*! POOL_tryAdd() :
- *  Add the job `function(opaque)` to thread pool _if_ a worker is available.
+ *  Add the job `function(opaque)` to thread pool _if_ a queue slot is available.
  *  Returns immediately even if not (does not block).
  * @return : 1 if successful, 0 if not.
  */