Update zstd to 1.3.2.

34 files changed, 5440 insertions, 3439 deletions

diff --git a/core/io/SCsub b/core/io/SCsub
index 4efc902717..948153019d 100644
--- a/core/io/SCsub
+++ b/core/io/SCsub
@@ -5,3 +5,4 @@ Import('env')
 env.add_source_files(env.core_sources, "*.cpp")
 
 Export('env')
+env.Append(CCFLAGS="-DZSTD_STATIC_LINKING_ONLY")
diff --git a/thirdparty/zstd/SCsub b/thirdparty/zstd/SCsub
index e8be1aaf44..899a18e1cf 100644
--- a/thirdparty/zstd/SCsub
+++ b/thirdparty/zstd/SCsub
@@ -14,10 +14,16 @@ thirdparty_zstd_sources = [
 	"compress/fse_compress.c",
 	"compress/huf_compress.c",
 	"compress/zstd_compress.c",
+	"compress/zstd_double_fast.c",
+	"compress/zstd_fast.c",
+	"compress/zstd_lazy.c",
+	"compress/zstd_ldm.c",
 	"compress/zstdmt_compress.c",
+	"compress/zstd_opt.c",
 	"decompress/huf_decompress.c",
 	"decompress/zstd_decompress.c",
 ]
 thirdparty_zstd_sources = [thirdparty_zstd_dir + file for file in thirdparty_zstd_sources]
 env.add_source_files(env.core_sources, thirdparty_zstd_sources)
 env.Append(CPPPATH=["#thirdparty/zstd", "#thirdparty/zstd/common"])
+env.Append(CCFLAGS="-DZSTD_STATIC_LINKING_ONLY")
diff --git a/thirdparty/zstd/common/bitstream.h b/thirdparty/zstd/common/bitstream.h
index 06121f21c5..2094823fe2 100644
--- a/thirdparty/zstd/common/bitstream.h
+++ b/thirdparty/zstd/common/bitstream.h
@@ -169,33 +169,39 @@ MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits);
 ****************************************************************/
 MEM_STATIC unsigned BIT_highbit32 (register U32 val)
 {
+    assert(val != 0);
+    {
 #   if defined(_MSC_VER)   /* Visual */
-    unsigned long r=0;
-    _BitScanReverse ( &r, val );
-    return (unsigned) r;
+        unsigned long r=0;
+        _BitScanReverse ( &r, val );
+        return (unsigned) r;
 #   elif defined(__GNUC__) && (__GNUC__ >= 3)   /* Use GCC Intrinsic */
-    return 31 - __builtin_clz (val);
+        return 31 - __builtin_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,
-                                              8, 12, 20, 28, 15, 17, 24,  7,
-                                             19, 27, 23,  6, 26,  5,  4, 31 };
-    U32 v = val;
-    v |= v >> 1;
-    v |= v >> 2;
-    v |= v >> 4;
-    v |= v >> 8;
-    v |= v >> 16;
-    return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27];
+        static const unsigned 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;
+        v |= v >> 1;
+        v |= v >> 2;
+        v |= v >> 4;
+        v |= v >> 8;
+        v |= v >> 16;
+        return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27];
 #   endif
+    }
 }
 
 /*=====    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 };   /* up to 26 bits */
-
+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
@@ -217,11 +223,14 @@ MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,
 }
 
 /*! BIT_addBits() :
- *  can add up to 26 bits into `bitC`.
+ *  can add up to 31 bits into `bitC`.
  *  Note : does not check for register overflow ! */
 MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
                             size_t value, unsigned nbBits)
 {
+    MEM_STATIC_ASSERT(BIT_MASK_SIZE == 32);
+    assert(nbBits < BIT_MASK_SIZE);
+    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
     bitC->bitContainer |= (value & BIT_mask[nbBits]) << bitC->bitPos;
     bitC->bitPos += nbBits;
 }
@@ -232,6 +241,7 @@ MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,
                                 size_t value, unsigned nbBits)
 {
     assert((value>>nbBits) == 0);
+    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
     bitC->bitContainer |= value << bitC->bitPos;
     bitC->bitPos += nbBits;
 }
@@ -242,7 +252,7 @@ MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,
 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->bitPos < sizeof(bitC->bitContainer) * 8);
     MEM_writeLEST(bitC->ptr, bitC->bitContainer);
     bitC->ptr += nbBytes;
     assert(bitC->ptr <= bitC->endPtr);
@@ -258,7 +268,7 @@ MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC)
 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->bitPos < sizeof(bitC->bitContainer) * 8);
     MEM_writeLEST(bitC->ptr, bitC->bitContainer);
     bitC->ptr += nbBytes;
     if (bitC->ptr > bitC->endPtr) bitC->ptr = bitC->endPtr;
@@ -350,12 +360,14 @@ MEM_STATIC size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 co
 #  endif
         return _bextr_u32(bitContainer, start, nbBits);
 #else
+    assert(nbBits < BIT_MASK_SIZE);
     return (bitContainer >> start) & BIT_mask[nbBits];
 #endif
 }
 
 MEM_STATIC size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
 {
+    assert(nbBits < BIT_MASK_SIZE);
     return bitContainer & BIT_mask[nbBits];
 }
 
diff --git a/thirdparty/zstd/common/compiler.h b/thirdparty/zstd/common/compiler.h
index d7225c443e..3a7553c380 100644
--- a/thirdparty/zstd/common/compiler.h
+++ b/thirdparty/zstd/common/compiler.h
@@ -5,6 +5,7 @@
  * 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
diff --git a/thirdparty/zstd/common/error_private.c b/thirdparty/zstd/common/error_private.c
index b5b14b509c..11f7cdab1c 100644
--- a/thirdparty/zstd/common/error_private.c
+++ b/thirdparty/zstd/common/error_private.c
@@ -5,6 +5,7 @@
  * 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 file is to have a single list of error strings embedded in binary */
@@ -29,14 +30,15 @@ const char* ERR_getErrorString(ERR_enum code)
     case PREFIX(init_missing): return "Context should be init first";
     case PREFIX(memory_allocation): return "Allocation error : not enough memory";
     case PREFIX(stage_wrong): return "Operation not authorized at current processing stage";
-    case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";
-    case PREFIX(srcSize_wrong): return "Src size is incorrect";
     case PREFIX(tableLog_tooLarge): return "tableLog requires too much memory : unsupported";
     case PREFIX(maxSymbolValue_tooLarge): return "Unsupported max Symbol Value : too large";
     case PREFIX(maxSymbolValue_tooSmall): return "Specified maxSymbolValue is too small";
     case PREFIX(dictionary_corrupted): return "Dictionary is corrupted";
     case PREFIX(dictionary_wrong): return "Dictionary mismatch";
     case PREFIX(dictionaryCreation_failed): return "Cannot create Dictionary from provided samples";
+    case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";
+    case PREFIX(srcSize_wrong): return "Src size is incorrect";
+        /* following error codes are not stable and may be removed or changed in a future version */
     case PREFIX(frameIndex_tooLarge): return "Frame index is too large";
     case PREFIX(seekableIO): return "An I/O error occurred when reading/seeking";
     case PREFIX(maxCode):
diff --git a/thirdparty/zstd/common/error_private.h b/thirdparty/zstd/common/error_private.h
index 9dd9a87cfa..0d2fa7e34b 100644
--- a/thirdparty/zstd/common/error_private.h
+++ b/thirdparty/zstd/common/error_private.h
@@ -5,6 +5,7 @@
  * 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 */
@@ -48,10 +49,9 @@ typedef ZSTD_ErrorCode ERR_enum;
 /*-****************************************
 *  Error codes handling
 ******************************************/
-#ifdef ERROR
-#  undef ERROR   /* reported already defined on VS 2015 (Rich Geldreich) */
-#endif
-#define ERROR(name) ((size_t)-PREFIX(name))
+#undef ERROR   /* reported already defined on VS 2015 (Rich Geldreich) */
+#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)); }
 
diff --git a/thirdparty/zstd/common/fse.h b/thirdparty/zstd/common/fse.h
index 1c44f83750..afd7801963 100644
--- a/thirdparty/zstd/common/fse.h
+++ b/thirdparty/zstd/common/fse.h
@@ -184,7 +184,7 @@ FSE_PUBLIC_API size_t FSE_writeNCount (void* buffer, size_t bufferSize, const sh
 /*! 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_PUBLIC_API FSE_CTable* FSE_createCTable (unsigned tableLog, unsigned maxSymbolValue);
+FSE_PUBLIC_API FSE_CTable* FSE_createCTable (unsigned maxSymbolValue, unsigned tableLog);
 FSE_PUBLIC_API void        FSE_freeCTable (FSE_CTable* ct);
 
 /*! FSE_buildCTable():
diff --git a/thirdparty/zstd/common/huf.h b/thirdparty/zstd/common/huf.h
index 2b3015a84c..522bf9b6c0 100644
--- a/thirdparty/zstd/common/huf.h
+++ b/thirdparty/zstd/common/huf.h
@@ -242,7 +242,7 @@ size_t HUF_readStats(BYTE* huffWeight, size_t hwSize, U32* rankStats,
 
 /** HUF_readCTable() :
 *   Loading a CTable saved with HUF_writeCTable() */
-size_t HUF_readCTable (HUF_CElt* CTable, unsigned maxSymbolValue, const void* src, size_t srcSize);
+size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize);
 
 
 /*
diff --git a/thirdparty/zstd/common/mem.h b/thirdparty/zstd/common/mem.h
index df85404fb8..23335c3146 100644
--- a/thirdparty/zstd/common/mem.h
+++ b/thirdparty/zstd/common/mem.h
@@ -5,6 +5,7 @@
  * 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
diff --git a/thirdparty/zstd/common/pool.c b/thirdparty/zstd/common/pool.c
index a227044f7f..1b0fe1035d 100644
--- a/thirdparty/zstd/common/pool.c
+++ b/thirdparty/zstd/common/pool.c
@@ -5,6 +5,7 @@
  * 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.
  */
 
 
@@ -25,13 +26,14 @@
 
 /* A job is a function and an opaque argument */
 typedef struct POOL_job_s {
-  POOL_function function;
-  void *opaque;
+    POOL_function function;
+    void *opaque;
 } POOL_job;
 
 struct POOL_ctx_s {
+    ZSTD_customMem customMem;
     /* Keep track of the threads */
-    pthread_t *threads;
+    ZSTD_pthread_t *threads;
     size_t numThreads;
 
     /* The queue is a circular buffer */
@@ -46,11 +48,11 @@ struct POOL_ctx_s {
     int queueEmpty;
 
     /* The mutex protects the queue */
-    pthread_mutex_t queueMutex;
+    ZSTD_pthread_mutex_t queueMutex;
     /* Condition variable for pushers to wait on when the queue is full */
-    pthread_cond_t queuePushCond;
+    ZSTD_pthread_cond_t queuePushCond;
     /* Condition variables for poppers to wait on when the queue is empty */
-    pthread_cond_t queuePopCond;
+    ZSTD_pthread_cond_t queuePopCond;
     /* Indicates if the queue is shutting down */
     int shutdown;
 };
@@ -65,14 +67,14 @@ static void* POOL_thread(void* opaque) {
     if (!ctx) { return NULL; }
     for (;;) {
         /* Lock the mutex and wait for a non-empty queue or until shutdown */
-        pthread_mutex_lock(&ctx->queueMutex);
+        ZSTD_pthread_mutex_lock(&ctx->queueMutex);
 
         while (ctx->queueEmpty && !ctx->shutdown) {
-            pthread_cond_wait(&ctx->queuePopCond, &ctx->queueMutex);
+            ZSTD_pthread_cond_wait(&ctx->queuePopCond, &ctx->queueMutex);
         }
         /* empty => shutting down: so stop */
         if (ctx->queueEmpty) {
-            pthread_mutex_unlock(&ctx->queueMutex);
+            ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
             return opaque;
         }
         /* Pop a job off the queue */
@@ -81,28 +83,32 @@ static void* POOL_thread(void* opaque) {
             ctx->numThreadsBusy++;
             ctx->queueEmpty = ctx->queueHead == ctx->queueTail;
             /* Unlock the mutex, signal a pusher, and run the job */
-            pthread_mutex_unlock(&ctx->queueMutex);
-            pthread_cond_signal(&ctx->queuePushCond);
+            ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+            ZSTD_pthread_cond_signal(&ctx->queuePushCond);
 
             job.function(job.opaque);
 
             /* If the intended queue size was 0, signal after finishing job */
             if (ctx->queueSize == 1) {
-                pthread_mutex_lock(&ctx->queueMutex);
+                ZSTD_pthread_mutex_lock(&ctx->queueMutex);
                 ctx->numThreadsBusy--;
-                pthread_mutex_unlock(&ctx->queueMutex);
-                pthread_cond_signal(&ctx->queuePushCond);
+                ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+                ZSTD_pthread_cond_signal(&ctx->queuePushCond);
         }   }
     }  /* for (;;) */
     /* Unreachable */
 }
 
-POOL_ctx *POOL_create(size_t numThreads, size_t queueSize) {
-    POOL_ctx *ctx;
+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* ctx;
     /* Check the parameters */
     if (!numThreads) { return NULL; }
     /* Allocate the context and zero initialize */
-    ctx = (POOL_ctx *)calloc(1, sizeof(POOL_ctx));
+    ctx = (POOL_ctx*)ZSTD_calloc(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
@@ -114,19 +120,20 @@ POOL_ctx *POOL_create(size_t numThreads, size_t queueSize) {
     ctx->queueTail = 0;
     ctx->numThreadsBusy = 0;
     ctx->queueEmpty = 1;
-    (void)pthread_mutex_init(&ctx->queueMutex, NULL);
-    (void)pthread_cond_init(&ctx->queuePushCond, NULL);
-    (void)pthread_cond_init(&ctx->queuePopCond, NULL);
+    (void)ZSTD_pthread_mutex_init(&ctx->queueMutex, NULL);
+    (void)ZSTD_pthread_cond_init(&ctx->queuePushCond, NULL);
+    (void)ZSTD_pthread_cond_init(&ctx->queuePopCond, NULL);
     ctx->shutdown = 0;
     /* Allocate space for the thread handles */
-    ctx->threads = (pthread_t*)malloc(numThreads * sizeof(pthread_t));
+    ctx->threads = (ZSTD_pthread_t*)ZSTD_malloc(numThreads * sizeof(ZSTD_pthread_t), customMem);
     ctx->numThreads = 0;
+    ctx->customMem = customMem;
     /* Check for errors */
     if (!ctx->threads || !ctx->queue) { POOL_free(ctx); return NULL; }
     /* Initialize the threads */
     {   size_t i;
         for (i = 0; i < numThreads; ++i) {
-            if (pthread_create(&ctx->threads[i], NULL, &POOL_thread, ctx)) {
+            if (ZSTD_pthread_create(&ctx->threads[i], NULL, &POOL_thread, ctx)) {
                 ctx->numThreads = i;
                 POOL_free(ctx);
                 return NULL;
@@ -139,37 +146,37 @@ POOL_ctx *POOL_create(size_t numThreads, size_t queueSize) {
 /*! POOL_join() :
     Shutdown the queue, wake any sleeping threads, and join all of the threads.
 */
-static void POOL_join(POOL_ctx *ctx) {
+static void POOL_join(POOL_ctx* ctx) {
     /* Shut down the queue */
-    pthread_mutex_lock(&ctx->queueMutex);
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
     ctx->shutdown = 1;
-    pthread_mutex_unlock(&ctx->queueMutex);
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
     /* Wake up sleeping threads */
-    pthread_cond_broadcast(&ctx->queuePushCond);
-    pthread_cond_broadcast(&ctx->queuePopCond);
+    ZSTD_pthread_cond_broadcast(&ctx->queuePushCond);
+    ZSTD_pthread_cond_broadcast(&ctx->queuePopCond);
     /* Join all of the threads */
     {   size_t i;
         for (i = 0; i < ctx->numThreads; ++i) {
-            pthread_join(ctx->threads[i], NULL);
+            ZSTD_pthread_join(ctx->threads[i], NULL);
     }   }
 }
 
 void POOL_free(POOL_ctx *ctx) {
     if (!ctx) { return; }
     POOL_join(ctx);
-    pthread_mutex_destroy(&ctx->queueMutex);
-    pthread_cond_destroy(&ctx->queuePushCond);
-    pthread_cond_destroy(&ctx->queuePopCond);
-    if (ctx->queue) free(ctx->queue);
-    if (ctx->threads) free(ctx->threads);
-    free(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);
 }
 
 size_t POOL_sizeof(POOL_ctx *ctx) {
     if (ctx==NULL) return 0;  /* supports sizeof NULL */
     return sizeof(*ctx)
         + ctx->queueSize * sizeof(POOL_job)
-        + ctx->numThreads * sizeof(pthread_t);
+        + ctx->numThreads * sizeof(ZSTD_pthread_t);
 }
 
 /**
@@ -191,12 +198,12 @@ void POOL_add(void* ctxVoid, POOL_function function, void *opaque) {
     POOL_ctx* const ctx = (POOL_ctx*)ctxVoid;
     if (!ctx) { return; }
 
-    pthread_mutex_lock(&ctx->queueMutex);
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
     {   POOL_job const job = {function, opaque};
 
         /* Wait until there is space in the queue for the new job */
         while (isQueueFull(ctx) && !ctx->shutdown) {
-          pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);
+          ZSTD_pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);
         }
         /* The queue is still going => there is space */
         if (!ctx->shutdown) {
@@ -205,8 +212,8 @@ void POOL_add(void* ctxVoid, POOL_function function, void *opaque) {
             ctx->queueTail = (ctx->queueTail + 1) % ctx->queueSize;
         }
     }
-    pthread_mutex_unlock(&ctx->queueMutex);
-    pthread_cond_signal(&ctx->queuePopCond);
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+    ZSTD_pthread_cond_signal(&ctx->queuePopCond);
 }
 
 #else  /* ZSTD_MULTITHREAD  not defined */
@@ -214,26 +221,34 @@ void POOL_add(void* ctxVoid, POOL_function function, void *opaque) {
 
 /* We don't need any data, but if it is empty malloc() might return NULL. */
 struct POOL_ctx_s {
-  int data;
+    int dummy;
 };
+static POOL_ctx g_ctx;
 
 POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {
-  (void)numThreads;
-  (void)queueSize;
-  return (POOL_ctx*)malloc(sizeof(POOL_ctx));
+    return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);
+}
+
+POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize, ZSTD_customMem customMem) {
+    (void)numThreads;
+    (void)queueSize;
+    (void)customMem;
+    return &g_ctx;
 }
 
 void POOL_free(POOL_ctx* ctx) {
-    free(ctx);
+    assert(!ctx || ctx == &g_ctx);
+    (void)ctx;
 }
 
 void POOL_add(void* ctx, POOL_function function, void* opaque) {
-  (void)ctx;
-  function(opaque);
+    (void)ctx;
+    function(opaque);
 }
 
 size_t POOL_sizeof(POOL_ctx* ctx) {
     if (ctx==NULL) return 0;  /* supports sizeof NULL */
+    assert(ctx == &g_ctx);
     return sizeof(*ctx);
 }
 
diff --git a/thirdparty/zstd/common/pool.h b/thirdparty/zstd/common/pool.h
index 264c5c9ca7..08c63715aa 100644
--- a/thirdparty/zstd/common/pool.h
+++ b/thirdparty/zstd/common/pool.h
@@ -5,6 +5,7 @@
  * 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 POOL_H
@@ -16,6 +17,7 @@ extern "C" {
 
 
 #include <stddef.h>   /* size_t */
+#include "zstd_internal.h"   /* ZSTD_customMem */
 
 typedef struct POOL_ctx_s POOL_ctx;
 
@@ -27,6 +29,8 @@ typedef struct POOL_ctx_s POOL_ctx;
 */
 POOL_ctx *POOL_create(size_t numThreads, size_t queueSize);
 
+POOL_ctx *POOL_create_advanced(size_t numThreads, size_t queueSize, ZSTD_customMem customMem);
+
 /*! POOL_free() :
     Free a thread pool returned by POOL_create().
 */
diff --git a/thirdparty/zstd/common/threading.c b/thirdparty/zstd/common/threading.c
index 141376c561..8be8c8da94 100644
--- a/thirdparty/zstd/common/threading.c
+++ b/thirdparty/zstd/common/threading.c
@@ -2,9 +2,9 @@
  * Copyright (c) 2016 Tino Reichardt
  * All rights reserved.
  *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of this source tree. An additional grant
- * of patent rights can be found in the PATENTS file in the same directory.
+ * 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 can contact the author at:
  * - zstdmt source repository: https://github.com/mcmilk/zstdmt
@@ -14,12 +14,8 @@
  * This file will hold wrapper for systems, which do not support pthreads
  */
 
-/* When ZSTD_MULTITHREAD is not defined, this file would become an empty translation unit.
-* Include some ISO C header code to prevent this and portably avoid related warnings.
-* (Visual C++: C4206 / GCC: -Wpedantic / Clang: -Wempty-translation-unit)
-*/
-#include <stddef.h>
-
+/* create fake symbol to avoid empty trnaslation unit warning */
+int g_ZSTD_threading_useles_symbol;
 
 #if defined(ZSTD_MULTITHREAD) && defined(_WIN32)
 
@@ -39,12 +35,12 @@
 
 static unsigned __stdcall worker(void *arg)
 {
-    pthread_t* const thread = (pthread_t*) arg;
+    ZSTD_pthread_t* const thread = (ZSTD_pthread_t*) arg;
     thread->arg = thread->start_routine(thread->arg);
     return 0;
 }
 
-int pthread_create(pthread_t* thread, const void* unused,
+int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
             void* (*start_routine) (void*), void* arg)
 {
     (void)unused;
@@ -58,16 +54,16 @@ int pthread_create(pthread_t* thread, const void* unused,
         return 0;
 }
 
-int _pthread_join(pthread_t * thread, void **value_ptr)
+int ZSTD_pthread_join(ZSTD_pthread_t thread, void **value_ptr)
 {
     DWORD result;
 
-    if (!thread->handle) return 0;
+    if (!thread.handle) return 0;
 
-    result = WaitForSingleObject(thread->handle, INFINITE);
+    result = WaitForSingleObject(thread.handle, INFINITE);
     switch (result) {
     case WAIT_OBJECT_0:
-        if (value_ptr) *value_ptr = thread->arg;
+        if (value_ptr) *value_ptr = thread.arg;
         return 0;
     case WAIT_ABANDONED:
         return EINVAL;
diff --git a/thirdparty/zstd/common/threading.h b/thirdparty/zstd/common/threading.h
index ab09977a86..197770db27 100644
--- a/thirdparty/zstd/common/threading.h
+++ b/thirdparty/zstd/common/threading.h
@@ -2,9 +2,9 @@
  * Copyright (c) 2016 Tino Reichardt
  * All rights reserved.
  *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of this source tree. An additional grant
- * of patent rights can be found in the PATENTS file in the same directory.
+ * 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 can contact the author at:
  * - zstdmt source repository: https://github.com/mcmilk/zstdmt
@@ -37,35 +37,38 @@ extern "C" {
 #  define WIN32_LEAN_AND_MEAN
 #endif
 
+#undef ERROR   /* reported already defined on VS 2015 (Rich Geldreich) */
 #include <windows.h>
+#undef ERROR
+#define ERROR(name) ZSTD_ERROR(name)
+
 
 /* mutex */
-#define pthread_mutex_t           CRITICAL_SECTION
-#define pthread_mutex_init(a,b)   (InitializeCriticalSection((a)), 0)
-#define pthread_mutex_destroy(a)  DeleteCriticalSection((a))
-#define pthread_mutex_lock(a)     EnterCriticalSection((a))
-#define pthread_mutex_unlock(a)   LeaveCriticalSection((a))
+#define ZSTD_pthread_mutex_t           CRITICAL_SECTION
+#define ZSTD_pthread_mutex_init(a, b)  (InitializeCriticalSection((a)), 0)
+#define ZSTD_pthread_mutex_destroy(a)  DeleteCriticalSection((a))
+#define ZSTD_pthread_mutex_lock(a)     EnterCriticalSection((a))
+#define ZSTD_pthread_mutex_unlock(a)   LeaveCriticalSection((a))
 
 /* condition variable */
-#define pthread_cond_t             CONDITION_VARIABLE
-#define pthread_cond_init(a, b)    (InitializeConditionVariable((a)), 0)
-#define pthread_cond_destroy(a)    /* No delete */
-#define pthread_cond_wait(a, b)    SleepConditionVariableCS((a), (b), INFINITE)
-#define pthread_cond_signal(a)     WakeConditionVariable((a))
-#define pthread_cond_broadcast(a)  WakeAllConditionVariable((a))
-
-/* pthread_create() and pthread_join() */
+#define ZSTD_pthread_cond_t             CONDITION_VARIABLE
+#define ZSTD_pthread_cond_init(a, b)    (InitializeConditionVariable((a)), 0)
+#define ZSTD_pthread_cond_destroy(a)    /* No delete */
+#define ZSTD_pthread_cond_wait(a, b)    SleepConditionVariableCS((a), (b), INFINITE)
+#define ZSTD_pthread_cond_signal(a)     WakeConditionVariable((a))
+#define ZSTD_pthread_cond_broadcast(a)  WakeAllConditionVariable((a))
+
+/* ZSTD_pthread_create() and ZSTD_pthread_join() */
 typedef struct {
     HANDLE handle;
     void* (*start_routine)(void*);
     void* arg;
-} pthread_t;
+} ZSTD_pthread_t;
 
-int pthread_create(pthread_t* thread, const void* unused,
+int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
                    void* (*start_routine) (void*), void* arg);
 
-#define pthread_join(a, b) _pthread_join(&(a), (b))
-int _pthread_join(pthread_t* thread, void** value_ptr);
+int ZSTD_pthread_join(ZSTD_pthread_t thread, void** value_ptr);
 
 /**
  * add here more wrappers as required
@@ -76,23 +79,40 @@ int _pthread_join(pthread_t* thread, void** value_ptr);
 /* ===   POSIX Systems   === */
 #  include <pthread.h>
 
+#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))
+#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
+#define ZSTD_pthread_cond_init(a, b)    pthread_cond_init((a), (b))
+#define ZSTD_pthread_cond_destroy(a)    pthread_cond_destroy((a))
+#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))
+
 #else  /* ZSTD_MULTITHREAD not defined */
 /* No multithreading support */
 
-#define pthread_mutex_t int   /* #define rather than typedef, because sometimes pthread support is implicit, resulting in duplicated symbols */
-#define pthread_mutex_init(a,b)    ((void)a, 0)
-#define pthread_mutex_destroy(a)
-#define pthread_mutex_lock(a)
-#define pthread_mutex_unlock(a)
+typedef int ZSTD_pthread_mutex_t;
+#define ZSTD_pthread_mutex_init(a, b)   ((void)a, 0)
+#define ZSTD_pthread_mutex_destroy(a)
+#define ZSTD_pthread_mutex_lock(a)
+#define ZSTD_pthread_mutex_unlock(a)
 
-#define pthread_cond_t int
-#define pthread_cond_init(a,b)     ((void)a, 0)
-#define pthread_cond_destroy(a)
-#define pthread_cond_wait(a,b)
-#define pthread_cond_signal(a)
-#define pthread_cond_broadcast(a)
+typedef int ZSTD_pthread_cond_t;
+#define ZSTD_pthread_cond_init(a, b)    ((void)a, 0)
+#define ZSTD_pthread_cond_destroy(a)
+#define ZSTD_pthread_cond_wait(a, b)
+#define ZSTD_pthread_cond_signal(a)
+#define ZSTD_pthread_cond_broadcast(a)
 
-/* do not use pthread_t */
+/* do not use ZSTD_pthread_t */
 
 #endif /* ZSTD_MULTITHREAD */
 
diff --git a/thirdparty/zstd/common/zstd_common.c b/thirdparty/zstd/common/zstd_common.c
index 08384cabf5..c2041053be 100644
--- a/thirdparty/zstd/common/zstd_common.c
+++ b/thirdparty/zstd/common/zstd_common.c
@@ -5,6 +5,7 @@
  * 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.
  */
 
 
@@ -15,8 +16,7 @@
 #include <stdlib.h>      /* malloc, calloc, free */
 #include <string.h>      /* memset */
 #include "error_private.h"
-#define ZSTD_STATIC_LINKING_ONLY
-#include "zstd.h"
+#include "zstd_internal.h"
 
 
 /*-****************************************
diff --git a/thirdparty/zstd/common/zstd_errors.h b/thirdparty/zstd/common/zstd_errors.h
index a69387b714..4bcb7769fe 100644
--- a/thirdparty/zstd/common/zstd_errors.h
+++ b/thirdparty/zstd/common/zstd_errors.h
@@ -5,6 +5,7 @@
  * 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
@@ -62,9 +63,10 @@ typedef enum {
   ZSTD_error_memory_allocation = 64,
   ZSTD_error_dstSize_tooSmall = 70,
   ZSTD_error_srcSize_wrong    = 72,
+  /* following error codes are not stable and may be removed or changed in a future version */
   ZSTD_error_frameIndex_tooLarge = 100,
   ZSTD_error_seekableIO          = 102,
-  ZSTD_error_maxCode = 120  /* never EVER use this value directly, it may change in future versions! Use ZSTD_isError() instead */
+  ZSTD_error_maxCode = 120  /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */
 } ZSTD_ErrorCode;
 
 /*! ZSTD_getErrorCode() :
diff --git a/thirdparty/zstd/common/zstd_internal.h b/thirdparty/zstd/common/zstd_internal.h
index 2610528608..e91cd20baa 100644
--- a/thirdparty/zstd/common/zstd_internal.h
+++ b/thirdparty/zstd/common/zstd_internal.h
@@ -5,6 +5,7 @@
  * 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_CCOMMON_H_MODULE
@@ -29,6 +30,11 @@
 #include "xxhash.h"                /* XXH_reset, update, digest */
 
 
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
 /*-*************************************
 *  Debug
 ***************************************/
@@ -96,9 +102,13 @@ static const U32 repStartValue[ZSTD_REP_NUM] = { 1, 4, 8 };
 #define BIT0   1
 
 #define ZSTD_WINDOWLOG_ABSOLUTEMIN 10
+#define ZSTD_WINDOWLOG_DEFAULTMAX 27 /* Default maximum allowed window log */
 static const size_t ZSTD_fcs_fieldSize[4] = { 0, 2, 4, 8 };
 static const size_t ZSTD_did_fieldSize[4] = { 0, 1, 2, 4 };
 
+#define ZSTD_FRAMEIDSIZE 4
+static const size_t ZSTD_frameIdSize = ZSTD_FRAMEIDSIZE;  /* magic number size */
+
 #define ZSTD_BLOCKHEADERSIZE 3   /* C standard doesn't allow `static const` variable to be init using another `static const` variable */
 static const size_t ZSTD_blockHeaderSize = ZSTD_BLOCKHEADERSIZE;
 typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e;
@@ -117,7 +127,8 @@ typedef enum { set_basic, set_rle, set_compressed, set_repeat } symbolEncodingTy
 #define MaxLit ((1<<Litbits) - 1)
 #define MaxML  52
 #define MaxLL  35
-#define MaxOff 28
+#define DefaultMaxOff 28
+#define MaxOff 31
 #define MaxSeq MAX(MaxLL, MaxML)   /* Assumption : MaxOff < MaxLL,MaxML */
 #define MLFSELog    9
 #define LLFSELog    9
@@ -143,8 +154,8 @@ static const S16 ML_defaultNorm[MaxML+1] = { 1, 4, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1,
 #define ML_DEFAULTNORMLOG 6  /* for static allocation */
 static const U32 ML_defaultNormLog = ML_DEFAULTNORMLOG;
 
-static const S16 OF_defaultNorm[MaxOff+1] = { 1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1,
-                                              1, 1, 1, 1, 1, 1, 1, 1,-1,-1,-1,-1,-1 };
+static const S16 OF_defaultNorm[DefaultMaxOff+1] = { 1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1,
+                                                     1, 1, 1, 1, 1, 1, 1, 1,-1,-1,-1,-1,-1 };
 #define OF_DEFAULTNORMLOG 5  /* for static allocation */
 static const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG;
 
@@ -244,6 +255,26 @@ typedef struct {
 } optState_t;
 
 typedef struct {
+    U32 offset;
+    U32 checksum;
+} ldmEntry_t;
+
+typedef struct {
+    ldmEntry_t* hashTable;
+    BYTE* bucketOffsets;    /* Next position in bucket to insert entry */
+    U64 hashPower;          /* Used to compute the rolling hash.
+                             * Depends on ldmParams.minMatchLength */
+} ldmState_t;
+
+typedef struct {
+    U32 enableLdm;          /* 1 if enable long distance matching */
+    U32 hashLog;            /* Log size of hashTable */
+    U32 bucketSizeLog;      /* Log bucket size for collision resolution, at most 8 */
+    U32 minMatchLength;     /* Minimum match length */
+    U32 hashEveryLog;       /* Log number of entries to skip */
+} ldmParams_t;
+
+typedef struct {
     U32 hufCTable[HUF_CTABLE_SIZE_U32(255)];
     FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];
     FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];
@@ -255,6 +286,28 @@ typedef struct {
     FSE_repeat litlength_repeatMode;
 } ZSTD_entropyCTables_t;
 
+struct ZSTD_CCtx_params_s {
+    ZSTD_format_e format;
+    ZSTD_compressionParameters cParams;
+    ZSTD_frameParameters fParams;
+
+    int compressionLevel;
+    U32 forceWindow;           /* force back-references to respect limit of
+                                * 1<<wLog, even for dictionary */
+
+    /* Multithreading: used to pass parameters to mtctx */
+    U32 nbThreads;
+    unsigned jobSize;
+    unsigned overlapSizeLog;
+
+    /* Long distance matching parameters */
+    ldmParams_t ldmParams;
+
+    /* For use with createCCtxParams() and freeCCtxParams() only */
+    ZSTD_customMem customMem;
+
+};  /* typedef'd to ZSTD_CCtx_params within "zstd.h" */
+
 const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx);
 void ZSTD_seqToCodes(const seqStore_t* seqStorePtr);
 
@@ -268,24 +321,27 @@ void ZSTD_free(void* ptr, ZSTD_customMem customMem);
 
 MEM_STATIC U32 ZSTD_highbit32(U32 val)
 {
+    assert(val != 0);
+    {
 #   if defined(_MSC_VER)   /* Visual */
-    unsigned long r=0;
-    _BitScanReverse(&r, val);
-    return (unsigned)r;
+        unsigned long r=0;
+        _BitScanReverse(&r, val);
+        return (unsigned)r;
 #   elif defined(__GNUC__) && (__GNUC__ >= 3)   /* GCC Intrinsic */
-    return 31 - __builtin_clz(val);
+        return 31 - __builtin_clz(val);
 #   else   /* Software version */
-    static const int 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;
-    int r;
-    v |= v >> 1;
-    v |= v >> 2;
-    v |= v >> 4;
-    v |= v >> 8;
-    v |= v >> 16;
-    r = DeBruijnClz[(U32)(v * 0x07C4ACDDU) >> 27];
-    return r;
+        static const int 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;
+        int r;
+        v |= v >> 1;
+        v |= v >> 2;
+        v |= v >> 4;
+        v |= v >> 8;
+        v |= v >> 16;
+        r = DeBruijnClz[(U32)(v * 0x07C4ACDDU) >> 27];
+        return r;
 #   endif
+    }
 }
 
 
@@ -306,7 +362,7 @@ void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx);
 size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
                      const void* dict, size_t dictSize,
                      const ZSTD_CDict* cdict,
-                     ZSTD_parameters params, unsigned long long pledgedSrcSize);
+                     ZSTD_CCtx_params  params, unsigned long long pledgedSrcSize);
 
 /*! ZSTD_compressStream_generic() :
  *  Private use only. To be called from zstdmt_compress.c in single-thread mode. */
@@ -315,10 +371,25 @@ size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
                                    ZSTD_inBuffer* input,
                                    ZSTD_EndDirective const flushMode);
 
-/*! ZSTD_getParamsFromCDict() :
+/*! ZSTD_getCParamsFromCDict() :
  *  as the name implies */
-ZSTD_parameters ZSTD_getParamsFromCDict(const ZSTD_CDict* cdict);
-
+ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict);
+
+/* ZSTD_compressBegin_advanced_internal() :
+ * Private use only. To be called from zstdmt_compress.c. */
+size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,
+                                    const void* dict, size_t dictSize,
+                                    ZSTD_dictMode_e dictMode,
+                                    ZSTD_CCtx_params params,
+                                    unsigned long long pledgedSrcSize);
+
+/* ZSTD_compress_advanced_internal() :
+ * Private use only. To be called from zstdmt_compress.c. */
+size_t ZSTD_compress_advanced_internal(ZSTD_CCtx* cctx,
+                                       void* dst, size_t dstCapacity,
+                                 const void* src, size_t srcSize,
+                                 const void* dict,size_t dictSize,
+                                 ZSTD_CCtx_params params);
 
 typedef struct {
     blockType_e blockType;
@@ -331,5 +402,8 @@ typedef struct {
 size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
                           blockProperties_t* bpPtr);
 
+#if defined (__cplusplus)
+}
+#endif
 
 #endif   /* ZSTD_CCOMMON_H_MODULE */
diff --git a/thirdparty/zstd/compress/fse_compress.c b/thirdparty/zstd/compress/fse_compress.c
index cc9fa73514..549c115d42 100644
--- a/thirdparty/zstd/compress/fse_compress.c
+++ b/thirdparty/zstd/compress/fse_compress.c
@@ -461,6 +461,7 @@ static unsigned FSE_minTableLog(size_t srcSize, unsigned maxSymbolValue)
     U32 minBitsSrc = BIT_highbit32((U32)(srcSize - 1)) + 1;
     U32 minBitsSymbols = BIT_highbit32(maxSymbolValue) + 2;
     U32 minBits = minBitsSrc < minBitsSymbols ? minBitsSrc : minBitsSymbols;
+    assert(srcSize > 1); /* Not supported, RLE should be used instead */
     return minBits;
 }
 
@@ -469,6 +470,7 @@ unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsi
     U32 maxBitsSrc = BIT_highbit32((U32)(srcSize - 1)) - minus;
     U32 tableLog = maxTableLog;
     U32 minBits = FSE_minTableLog(srcSize, maxSymbolValue);
+    assert(srcSize > 1); /* Not supported, RLE should be used instead */
     if (tableLog==0) tableLog = FSE_DEFAULT_TABLELOG;
     if (maxBitsSrc < tableLog) tableLog = maxBitsSrc;   /* Accuracy can be reduced */
     if (minBits > tableLog) tableLog = minBits;   /* Need a minimum to safely represent all symbol values */
@@ -580,7 +582,7 @@ size_t FSE_normalizeCount (short* normalizedCounter, unsigned tableLog,
     if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge);   /* Unsupported size */
     if (tableLog < FSE_minTableLog(total, maxSymbolValue)) return ERROR(GENERIC);   /* Too small tableLog, compression potentially impossible */
 
-    {   U32 const rtbTable[] = {     0, 473195, 504333, 520860, 550000, 700000, 750000, 830000 };
+    {   static U32 const rtbTable[] = {     0, 473195, 504333, 520860, 550000, 700000, 750000, 830000 };
         U64 const scale = 62 - tableLog;
         U64 const step = ((U64)1<<62) / total;   /* <== here, one division ! */
         U64 const vStep = 1ULL<<(scale-20);
diff --git a/thirdparty/zstd/compress/huf_compress.c b/thirdparty/zstd/compress/huf_compress.c
index 2a47c18205..5692d56e00 100644
--- a/thirdparty/zstd/compress/huf_compress.c
+++ b/thirdparty/zstd/compress/huf_compress.c
@@ -167,7 +167,7 @@ size_t HUF_writeCTable (void* dst, size_t maxDstSize,
 }
 
 
-size_t HUF_readCTable (HUF_CElt* CTable, U32 maxSymbolValue, const void* src, size_t srcSize)
+size_t HUF_readCTable (HUF_CElt* CTable, U32* maxSymbolValuePtr, const void* src, size_t srcSize)
 {
     BYTE huffWeight[HUF_SYMBOLVALUE_MAX + 1];   /* init not required, even though some static analyzer may complain */
     U32 rankVal[HUF_TABLELOG_ABSOLUTEMAX + 1];   /* large enough for values from 0 to 16 */
@@ -179,7 +179,7 @@ size_t HUF_readCTable (HUF_CElt* CTable, U32 maxSymbolValue, const void* src, si
 
     /* check result */
     if (tableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);
-    if (nbSymbols > maxSymbolValue+1) return ERROR(maxSymbolValue_tooSmall);
+    if (nbSymbols > *maxSymbolValuePtr+1) return ERROR(maxSymbolValue_tooSmall);
 
     /* Prepare base value per rank */
     {   U32 n, nextRankStart = 0;
@@ -208,9 +208,10 @@ size_t HUF_readCTable (HUF_CElt* CTable, U32 maxSymbolValue, const void* src, si
                 min >>= 1;
         }   }
         /* assign value within rank, symbol order */
-        { U32 n; for (n=0; n<=maxSymbolValue; n++) CTable[n].val = valPerRank[CTable[n].nbBits]++; }
+        { U32 n; for (n=0; n<nbSymbols; n++) CTable[n].val = valPerRank[CTable[n].nbBits]++; }
     }
 
+    *maxSymbolValuePtr = nbSymbols - 1;
     return readSize;
 }
 
diff --git a/thirdparty/zstd/compress/zstd_compress.c b/thirdparty/zstd/compress/zstd_compress.c
index 0322c03eb3..2c46c79f1c 100644
--- a/thirdparty/zstd/compress/zstd_compress.c
+++ b/thirdparty/zstd/compress/zstd_compress.c
@@ -5,6 +5,7 @@
  * 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.
  */
 
 
@@ -25,25 +26,19 @@
 #include "fse.h"
 #define HUF_STATIC_LINKING_ONLY
 #include "huf.h"
-#include "zstd_internal.h"  /* includes zstd.h */
-#include "zstdmt_compress.h"
-
-
-/*-*************************************
-*  Constants
-***************************************/
-static const U32 g_searchStrength = 8;   /* control skip over incompressible data */
-#define HASH_READ_SIZE 8
-typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;
+#include "zstd_compress.h"
+#include "zstd_fast.h"
+#include "zstd_double_fast.h"
+#include "zstd_lazy.h"
+#include "zstd_opt.h"
+#include "zstd_ldm.h"
 
 
 /*-*************************************
 *  Helper functions
 ***************************************/
 size_t ZSTD_compressBound(size_t srcSize) {
-    size_t const lowLimit = 256 KB;
-    size_t const margin = (srcSize < lowLimit) ? (lowLimit-srcSize) >> 12 : 0;  /* from 64 to 0 */
-    return srcSize + (srcSize >> 8) + margin;
+    return ZSTD_COMPRESSBOUND(srcSize);
 }
 
 
@@ -61,8 +56,6 @@ static void ZSTD_resetSeqStore(seqStore_t* ssPtr)
 /*-*************************************
 *  Context memory management
 ***************************************/
-typedef enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage;
-
 struct ZSTD_CDict_s {
     void* dictBuffer;
     const void* dictContent;
@@ -70,65 +63,6 @@ struct ZSTD_CDict_s {
     ZSTD_CCtx* refContext;
 };  /* typedef'd to ZSTD_CDict within "zstd.h" */
 
-struct ZSTD_CCtx_s {
-    const BYTE* nextSrc;    /* next block here to continue on current prefix */
-    const BYTE* base;       /* All regular indexes relative to this position */
-    const BYTE* dictBase;   /* extDict indexes relative to this position */
-    U32   dictLimit;        /* below that point, need extDict */
-    U32   lowLimit;         /* below that point, no more data */
-    U32   nextToUpdate;     /* index from which to continue dictionary update */
-    U32   nextToUpdate3;    /* index from which to continue dictionary update */
-    U32   hashLog3;         /* dispatch table : larger == faster, more memory */
-    U32   loadedDictEnd;    /* index of end of dictionary */
-    U32   forceWindow;      /* force back-references to respect limit of 1<<wLog, even for dictionary */
-    ZSTD_compressionStage_e stage;
-    U32   dictID;
-    int   compressionLevel;
-    ZSTD_parameters requestedParams;
-    ZSTD_parameters appliedParams;
-    void* workSpace;
-    size_t workSpaceSize;
-    size_t blockSize;
-    U64 pledgedSrcSizePlusOne;  /* this way, 0 (default) == unknown */
-    U64 consumedSrcSize;
-    XXH64_state_t xxhState;
-    ZSTD_customMem customMem;
-    size_t staticSize;
-
-    seqStore_t seqStore;    /* sequences storage ptrs */
-    optState_t optState;
-    U32* hashTable;
-    U32* hashTable3;
-    U32* chainTable;
-    ZSTD_entropyCTables_t* entropy;
-
-    /* streaming */
-    char*  inBuff;
-    size_t inBuffSize;
-    size_t inToCompress;
-    size_t inBuffPos;
-    size_t inBuffTarget;
-    char*  outBuff;
-    size_t outBuffSize;
-    size_t outBuffContentSize;
-    size_t outBuffFlushedSize;
-    ZSTD_cStreamStage streamStage;
-    U32    frameEnded;
-
-    /* Dictionary */
-    ZSTD_dictMode_e dictMode; /* select restricting dictionary to "rawContent" or "fullDict" only */
-    U32 dictContentByRef;
-    ZSTD_CDict* cdictLocal;
-    const ZSTD_CDict* cdict;
-    const void* prefix;
-    size_t prefixSize;
-
-    /* Multi-threading */
-    U32 nbThreads;
-    ZSTDMT_CCtx* mtctx;
-};
-
-
 ZSTD_CCtx* ZSTD_createCCtx(void)
 {
     return ZSTD_createCCtx_advanced(ZSTD_defaultCMem);
@@ -143,7 +77,7 @@ ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem)
     cctx = (ZSTD_CCtx*) ZSTD_calloc(sizeof(ZSTD_CCtx), customMem);
     if (!cctx) return NULL;
     cctx->customMem = customMem;
-    cctx->compressionLevel = ZSTD_CLEVEL_DEFAULT;
+    cctx->requestedParams.compressionLevel = ZSTD_CLEVEL_DEFAULT;
     ZSTD_STATIC_ASSERT(zcss_init==0);
     ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN==(0ULL - 1));
     return cctx;
@@ -151,7 +85,7 @@ ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem)
 
 ZSTD_CCtx* ZSTD_initStaticCCtx(void *workspace, size_t workspaceSize)
 {
-    ZSTD_CCtx* cctx = (ZSTD_CCtx*) workspace;
+    ZSTD_CCtx* const cctx = (ZSTD_CCtx*) workspace;
     if (workspaceSize <= sizeof(ZSTD_CCtx)) return NULL;  /* minimum size */
     if ((size_t)workspace & 7) return NULL;  /* must be 8-aligned */
     memset(workspace, 0, workspaceSize);   /* may be a bit generous, could memset be smaller ? */
@@ -161,7 +95,7 @@ ZSTD_CCtx* ZSTD_initStaticCCtx(void *workspace, size_t workspaceSize)
 
     /* entropy space (never moves) */
     if (cctx->workSpaceSize < sizeof(ZSTD_entropyCTables_t)) return NULL;
-    assert(((size_t)cctx->workSpace & 7) == 0);   /* ensure correct alignment */
+    assert(((size_t)cctx->workSpace & (sizeof(void*)-1)) == 0);   /* ensure correct alignment */
     cctx->entropy = (ZSTD_entropyCTables_t*)cctx->workSpace;
 
     return cctx;
@@ -175,23 +109,36 @@ size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)
     cctx->workSpace = NULL;
     ZSTD_freeCDict(cctx->cdictLocal);
     cctx->cdictLocal = NULL;
+#ifdef ZSTD_MULTITHREAD
     ZSTDMT_freeCCtx(cctx->mtctx);
     cctx->mtctx = NULL;
+#endif
     ZSTD_free(cctx, cctx->customMem);
     return 0;   /* reserved as a potential error code in the future */
 }
 
+
+static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx)
+{
+#ifdef ZSTD_MULTITHREAD
+    return ZSTDMT_sizeof_CCtx(cctx->mtctx);
+#else
+    (void) cctx;
+    return 0;
+#endif
+}
+
+
 size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx)
 {
     if (cctx==NULL) return 0;   /* support sizeof on NULL */
-    DEBUGLOG(5, "sizeof(*cctx) : %u", (U32)sizeof(*cctx));
-    DEBUGLOG(5, "workSpaceSize : %u", (U32)cctx->workSpaceSize);
-    DEBUGLOG(5, "streaming buffers : %u", (U32)(cctx->outBuffSize + cctx->inBuffSize));
-    DEBUGLOG(5, "inner MTCTX : %u", (U32)ZSTDMT_sizeof_CCtx(cctx->mtctx));
+    DEBUGLOG(3, "sizeof(*cctx) : %u", (U32)sizeof(*cctx));
+    DEBUGLOG(3, "workSpaceSize (including streaming buffers): %u", (U32)cctx->workSpaceSize);
+    DEBUGLOG(3, "inner cdict : %u", (U32)ZSTD_sizeof_CDict(cctx->cdictLocal));
+    DEBUGLOG(3, "inner MTCTX : %u", (U32)ZSTD_sizeof_mtctx(cctx));
     return sizeof(*cctx) + cctx->workSpaceSize
            + ZSTD_sizeof_CDict(cctx->cdictLocal)
-           + cctx->outBuffSize + cctx->inBuffSize
-           + ZSTDMT_sizeof_CCtx(cctx->mtctx);
+           + ZSTD_sizeof_mtctx(cctx);
 }
 
 size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)
@@ -202,29 +149,99 @@ size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)
 /* private API call, for dictBuilder only */
 const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx) { return &(ctx->seqStore); }
 
-static ZSTD_parameters ZSTD_getParamsFromCCtx(const ZSTD_CCtx* cctx) { return cctx->appliedParams; }
+#define ZSTD_CLEVEL_CUSTOM 999
 
-/* older variant; will be deprecated */
-size_t ZSTD_setCCtxParameter(ZSTD_CCtx* cctx, ZSTD_CCtxParameter param, unsigned value)
+static ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(
+        ZSTD_CCtx_params params, U64 srcSizeHint, size_t dictSize)
 {
-    switch(param)
-    {
-    case ZSTD_p_forceWindow : cctx->forceWindow = value>0; cctx->loadedDictEnd = 0; return 0;
-    ZSTD_STATIC_ASSERT(ZSTD_dm_auto==0);
-    ZSTD_STATIC_ASSERT(ZSTD_dm_rawContent==1);
-    case ZSTD_p_forceRawDict : cctx->dictMode = (ZSTD_dictMode_e)(value>0); return 0;
-    default: return ERROR(parameter_unsupported);
-    }
+    return (params.compressionLevel == ZSTD_CLEVEL_CUSTOM ?
+                    params.cParams :
+                    ZSTD_getCParams(params.compressionLevel, srcSizeHint, dictSize));
 }
 
+static void ZSTD_cLevelToCCtxParams_srcSize(ZSTD_CCtx_params* params, U64 srcSize)
+{
+    params->cParams = ZSTD_getCParamsFromCCtxParams(*params, srcSize, 0);
+    params->compressionLevel = ZSTD_CLEVEL_CUSTOM;
+}
 
-#define ZSTD_CLEVEL_CUSTOM 999
 static void ZSTD_cLevelToCParams(ZSTD_CCtx* cctx)
 {
-    if (cctx->compressionLevel==ZSTD_CLEVEL_CUSTOM) return;
-    cctx->requestedParams.cParams = ZSTD_getCParams(cctx->compressionLevel,
-                                            cctx->pledgedSrcSizePlusOne-1, 0);
-    cctx->compressionLevel = ZSTD_CLEVEL_CUSTOM;
+    ZSTD_cLevelToCCtxParams_srcSize(
+            &cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1);
+}
+
+static void ZSTD_cLevelToCCtxParams(ZSTD_CCtx_params* params)
+{
+    ZSTD_cLevelToCCtxParams_srcSize(params, 0);
+}
+
+static ZSTD_CCtx_params ZSTD_makeCCtxParamsFromCParams(
+        ZSTD_compressionParameters cParams)
+{
+    ZSTD_CCtx_params cctxParams;
+    memset(&cctxParams, 0, sizeof(cctxParams));
+    cctxParams.cParams = cParams;
+    cctxParams.compressionLevel = ZSTD_CLEVEL_CUSTOM;
+    return cctxParams;
+}
+
+static ZSTD_CCtx_params* ZSTD_createCCtxParams_advanced(
+        ZSTD_customMem customMem)
+{
+    ZSTD_CCtx_params* params;
+    if (!customMem.customAlloc ^ !customMem.customFree) return NULL;
+    params = (ZSTD_CCtx_params*)ZSTD_calloc(
+            sizeof(ZSTD_CCtx_params), customMem);
+    if (!params) { return NULL; }
+    params->customMem = customMem;
+    params->compressionLevel = ZSTD_CLEVEL_DEFAULT;
+    return params;
+}
+
+ZSTD_CCtx_params* ZSTD_createCCtxParams(void)
+{
+    return ZSTD_createCCtxParams_advanced(ZSTD_defaultCMem);
+}
+
+size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params)
+{
+    if (params == NULL) { return 0; }
+    ZSTD_free(params, params->customMem);
+    return 0;
+}
+
+size_t ZSTD_resetCCtxParams(ZSTD_CCtx_params* params)
+{
+    return ZSTD_initCCtxParams(params, ZSTD_CLEVEL_DEFAULT);
+}
+
+size_t ZSTD_initCCtxParams(ZSTD_CCtx_params* cctxParams, int compressionLevel) {
+    if (!cctxParams) { return ERROR(GENERIC); }
+    memset(cctxParams, 0, sizeof(*cctxParams));
+    cctxParams->compressionLevel = compressionLevel;
+    return 0;
+}
+
+size_t ZSTD_initCCtxParams_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params)
+{
+    if (!cctxParams) { return ERROR(GENERIC); }
+    CHECK_F( ZSTD_checkCParams(params.cParams) );
+    memset(cctxParams, 0, sizeof(*cctxParams));
+    cctxParams->cParams = params.cParams;
+    cctxParams->fParams = params.fParams;
+    cctxParams->compressionLevel = ZSTD_CLEVEL_CUSTOM;
+    return 0;
+}
+
+static ZSTD_CCtx_params ZSTD_assignParamsToCCtxParams(
+        ZSTD_CCtx_params cctxParams, ZSTD_parameters params)
+{
+    ZSTD_CCtx_params ret = cctxParams;
+    ret.cParams = params.cParams;
+    ret.fParams = params.fParams;
+    ret.compressionLevel = ZSTD_CLEVEL_CUSTOM;
+    return ret;
 }
 
 #define CLAMPCHECK(val,min,max) {            \
@@ -238,167 +255,285 @@ size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, unsigned v
 
     switch(param)
     {
-    case ZSTD_p_compressionLevel :
-        if ((int)value > ZSTD_maxCLevel()) value = ZSTD_maxCLevel();   /* cap max compression level */
+    case ZSTD_p_format :
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_compressionLevel:
         if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
         if (cctx->cdict) return ERROR(stage_wrong);
-        cctx->compressionLevel = value;
-        return 0;
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_windowLog:
+    case ZSTD_p_hashLog:
+    case ZSTD_p_chainLog:
+    case ZSTD_p_searchLog:
+    case ZSTD_p_minMatch:
+    case ZSTD_p_targetLength:
+    case ZSTD_p_compressionStrategy:
+        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
+        if (cctx->cdict) return ERROR(stage_wrong);
+        ZSTD_cLevelToCParams(cctx);  /* Can optimize if srcSize is known */
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
 
-    case ZSTD_p_windowLog :
-        DEBUGLOG(5, "setting ZSTD_p_windowLog = %u (cdict:%u)",
-                    value, (cctx->cdict!=NULL));
+    case ZSTD_p_contentSizeFlag:
+    case ZSTD_p_checksumFlag:
+    case ZSTD_p_dictIDFlag:
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_forceMaxWindow :  /* Force back-references to remain < windowSize,
+                                   * even when referencing into Dictionary content
+                                   * default : 0 when using a CDict, 1 when using a Prefix */
+        cctx->loadedDictEnd = 0;
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_nbThreads:
+        if (value==0) return 0;
+        DEBUGLOG(5, " setting nbThreads : %u", value);
+        if (value > 1 && cctx->staticSize) {
+            return ERROR(parameter_unsupported);  /* MT not compatible with static alloc */
+        }
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_jobSize:
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_overlapSizeLog:
+        DEBUGLOG(5, " setting overlap with nbThreads == %u", cctx->requestedParams.nbThreads);
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_enableLongDistanceMatching:
+        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value != 0) {
+            ZSTD_cLevelToCParams(cctx);
+        }
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_ldmHashLog:
+    case ZSTD_p_ldmMinMatch:
         if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
         if (cctx->cdict) return ERROR(stage_wrong);
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    case ZSTD_p_ldmBucketSizeLog:
+    case ZSTD_p_ldmHashEveryLog:
+        if (cctx->cdict) return ERROR(stage_wrong);
+        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+
+    default: return ERROR(parameter_unsupported);
+    }
+}
+
+size_t ZSTD_CCtxParam_setParameter(
+        ZSTD_CCtx_params* params, ZSTD_cParameter param, unsigned value)
+{
+    switch(param)
+    {
+    case ZSTD_p_format :
+        if (value > (unsigned)ZSTD_f_zstd1_magicless)
+            return ERROR(parameter_unsupported);
+        params->format = (ZSTD_format_e)value;
+        return 0;
+
+    case ZSTD_p_compressionLevel :
+        if ((int)value > ZSTD_maxCLevel()) value = ZSTD_maxCLevel();
+        if (value == 0) return 0;
+        params->compressionLevel = value;
+        return 0;
+
+    case ZSTD_p_windowLog :
+        if (value == 0) return 0;
         CLAMPCHECK(value, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.windowLog = value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.windowLog = value;
         return 0;
 
     case ZSTD_p_hashLog :
-        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
-        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value == 0) return 0;
         CLAMPCHECK(value, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.hashLog = value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.hashLog = value;
         return 0;
 
     case ZSTD_p_chainLog :
-        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
-        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value == 0) return 0;
         CLAMPCHECK(value, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.chainLog = value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.chainLog = value;
         return 0;
 
     case ZSTD_p_searchLog :
-        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
-        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value == 0) return 0;
         CLAMPCHECK(value, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.searchLog = value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.searchLog = value;
         return 0;
 
     case ZSTD_p_minMatch :
-        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
-        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value == 0) return 0;
         CLAMPCHECK(value, ZSTD_SEARCHLENGTH_MIN, ZSTD_SEARCHLENGTH_MAX);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.searchLength = value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.searchLength = value;
         return 0;
 
     case ZSTD_p_targetLength :
-        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
-        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value == 0) return 0;
         CLAMPCHECK(value, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.targetLength = value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.targetLength = value;
         return 0;
 
     case ZSTD_p_compressionStrategy :
-        if (value == 0) return 0;  /* special value : 0 means "don't change anything" */
-        if (cctx->cdict) return ERROR(stage_wrong);
+        if (value == 0) return 0;
         CLAMPCHECK(value, (unsigned)ZSTD_fast, (unsigned)ZSTD_btultra);
-        ZSTD_cLevelToCParams(cctx);
-        cctx->requestedParams.cParams.strategy = (ZSTD_strategy)value;
+        ZSTD_cLevelToCCtxParams(params);
+        params->cParams.strategy = (ZSTD_strategy)value;
         return 0;
 
     case ZSTD_p_contentSizeFlag :
-        DEBUGLOG(5, "set content size flag = %u", (value>0));
         /* Content size written in frame header _when known_ (default:1) */
-        cctx->requestedParams.fParams.contentSizeFlag = value>0;
+        DEBUGLOG(5, "set content size flag = %u", (value>0));
+        params->fParams.contentSizeFlag = value > 0;
         return 0;
 
     case ZSTD_p_checksumFlag :
         /* A 32-bits content checksum will be calculated and written at end of frame (default:0) */
-        cctx->requestedParams.fParams.checksumFlag = value>0;
+        params->fParams.checksumFlag = value > 0;
         return 0;
 
     case ZSTD_p_dictIDFlag : /* When applicable, dictionary's dictID is provided in frame header (default:1) */
         DEBUGLOG(5, "set dictIDFlag = %u", (value>0));
-        cctx->requestedParams.fParams.noDictIDFlag = (value==0);
+        params->fParams.noDictIDFlag = (value == 0);
         return 0;
 
-    /* Dictionary parameters */
-    case ZSTD_p_dictMode :
-        if (cctx->cdict) return ERROR(stage_wrong);  /* must be set before loading */
-        /* restrict dictionary mode, to "rawContent" or "fullDict" only */
-        ZSTD_STATIC_ASSERT((U32)ZSTD_dm_fullDict > (U32)ZSTD_dm_rawContent);
-        if (value > (unsigned)ZSTD_dm_fullDict)
-            return ERROR(parameter_outOfBound);
-        cctx->dictMode = (ZSTD_dictMode_e)value;
+    case ZSTD_p_forceMaxWindow :
+        params->forceWindow = value > 0;
         return 0;
 
-    case ZSTD_p_refDictContent :
-        if (cctx->cdict) return ERROR(stage_wrong);  /* must be set before loading */
-        /* dictionary content will be referenced, instead of copied */
-        cctx->dictContentByRef = value>0;
+    case ZSTD_p_nbThreads :
+        if (value == 0) return 0;
+#ifndef ZSTD_MULTITHREAD
+        if (value > 1) return ERROR(parameter_unsupported);
         return 0;
+#else
+        return ZSTDMT_initializeCCtxParameters(params, value);
+#endif
 
-    case ZSTD_p_forceMaxWindow :  /* Force back-references to remain < windowSize,
-                                   * even when referencing into Dictionary content
-                                   * default : 0 when using a CDict, 1 when using a Prefix */
-        cctx->forceWindow = value>0;
-        cctx->loadedDictEnd = 0;
-        return 0;
+    case ZSTD_p_jobSize :
+#ifndef ZSTD_MULTITHREAD
+        return ERROR(parameter_unsupported);
+#else
+        if (params->nbThreads <= 1) return ERROR(parameter_unsupported);
+        return ZSTDMT_CCtxParam_setMTCtxParameter(params, ZSTDMT_p_sectionSize, value);
+#endif
 
-    case ZSTD_p_nbThreads:
-        if (value==0) return 0;
-        DEBUGLOG(5, " setting nbThreads : %u", value);
+    case ZSTD_p_overlapSizeLog :
 #ifndef ZSTD_MULTITHREAD
-        if (value > 1) return ERROR(parameter_unsupported);
+        return ERROR(parameter_unsupported);
+#else
+        if (params->nbThreads <= 1) return ERROR(parameter_unsupported);
+        return ZSTDMT_CCtxParam_setMTCtxParameter(params, ZSTDMT_p_overlapSectionLog, value);
 #endif
-        if ((value>1) && (cctx->nbThreads != value)) {
-            if (cctx->staticSize)  /* MT not compatible with static alloc */
-                return ERROR(parameter_unsupported);
-            ZSTDMT_freeCCtx(cctx->mtctx);
-            cctx->nbThreads = 1;
-            cctx->mtctx = ZSTDMT_createCCtx_advanced(value, cctx->customMem);
-            if (cctx->mtctx == NULL) return ERROR(memory_allocation);
+
+    case ZSTD_p_enableLongDistanceMatching :
+        if (value != 0) {
+            ZSTD_cLevelToCCtxParams(params);
+            params->cParams.windowLog = ZSTD_LDM_DEFAULT_WINDOW_LOG;
         }
-        cctx->nbThreads = value;
+        return ZSTD_ldm_initializeParameters(&params->ldmParams, value);
+
+    case ZSTD_p_ldmHashLog :
+        if (value == 0) return 0;
+        CLAMPCHECK(value, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX);
+        params->ldmParams.hashLog = value;
         return 0;
 
-    case ZSTD_p_jobSize:
-        if (cctx->nbThreads <= 1) return ERROR(parameter_unsupported);
-        assert(cctx->mtctx != NULL);
-        return ZSTDMT_setMTCtxParameter(cctx->mtctx, ZSTDMT_p_sectionSize, value);
+    case ZSTD_p_ldmMinMatch :
+        if (value == 0) return 0;
+        CLAMPCHECK(value, ZSTD_LDM_MINMATCH_MIN, ZSTD_LDM_MINMATCH_MAX);
+        params->ldmParams.minMatchLength = value;
+        return 0;
 
-    case ZSTD_p_overlapSizeLog:
-        DEBUGLOG(5, " setting overlap with nbThreads == %u", cctx->nbThreads);
-        if (cctx->nbThreads <= 1) return ERROR(parameter_unsupported);
-        assert(cctx->mtctx != NULL);
-        return ZSTDMT_setMTCtxParameter(cctx->mtctx, ZSTDMT_p_overlapSectionLog, value);
+    case ZSTD_p_ldmBucketSizeLog :
+        if (value > ZSTD_LDM_BUCKETSIZELOG_MAX) {
+            return ERROR(parameter_outOfBound);
+        }
+        params->ldmParams.bucketSizeLog = value;
+        return 0;
+
+    case ZSTD_p_ldmHashEveryLog :
+        if (value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN) {
+            return ERROR(parameter_outOfBound);
+        }
+        params->ldmParams.hashEveryLog = value;
+        return 0;
 
     default: return ERROR(parameter_unsupported);
     }
 }
 
+/**
+ * This function should be updated whenever ZSTD_CCtx_params is updated.
+ * Parameters are copied manually before the dictionary is loaded.
+ * The multithreading parameters jobSize and overlapSizeLog are set only if
+ * nbThreads > 1.
+ *
+ * Pledged srcSize is treated as unknown.
+ */
+size_t ZSTD_CCtx_setParametersUsingCCtxParams(
+        ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params)
+{
+    if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
+    if (cctx->cdict) return ERROR(stage_wrong);
+
+    /* Assume the compression and frame parameters are validated */
+    cctx->requestedParams.cParams = params->cParams;
+    cctx->requestedParams.fParams = params->fParams;
+    cctx->requestedParams.compressionLevel = params->compressionLevel;
+
+    /* Set force window explicitly since it sets cctx->loadedDictEnd */
+    CHECK_F( ZSTD_CCtx_setParameter(
+                   cctx, ZSTD_p_forceMaxWindow, params->forceWindow) );
+
+    /* Set multithreading parameters explicitly */
+    CHECK_F( ZSTD_CCtx_setParameter(cctx, ZSTD_p_nbThreads, params->nbThreads) );
+    if (params->nbThreads > 1) {
+        CHECK_F( ZSTD_CCtx_setParameter(cctx, ZSTD_p_jobSize, params->jobSize) );
+        CHECK_F( ZSTD_CCtx_setParameter(
+                    cctx, ZSTD_p_overlapSizeLog, params->overlapSizeLog) );
+    }
+
+    /* Copy long distance matching parameters */
+    cctx->requestedParams.ldmParams = params->ldmParams;
+
+    /* customMem is used only for create/free params and can be ignored */
+    return 0;
+}
+
 ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize)
 {
-    DEBUGLOG(5, " setting pledgedSrcSize to %u", (U32)pledgedSrcSize);
+    DEBUGLOG(4, " setting pledgedSrcSize to %u", (U32)pledgedSrcSize);
     if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
     cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;
     return 0;
 }
 
-ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
+size_t ZSTD_CCtx_loadDictionary_advanced(
+        ZSTD_CCtx* cctx, const void* dict, size_t dictSize,
+        ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode)
 {
     if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
     if (cctx->staticSize) return ERROR(memory_allocation);  /* no malloc for static CCtx */
-    DEBUGLOG(5, "load dictionary of size %u", (U32)dictSize);
+    DEBUGLOG(4, "load dictionary of size %u", (U32)dictSize);
     ZSTD_freeCDict(cctx->cdictLocal);  /* in case one already exists */
     if (dict==NULL || dictSize==0) {   /* no dictionary mode */
         cctx->cdictLocal = NULL;
         cctx->cdict = NULL;
     } else {
         ZSTD_compressionParameters const cParams =
-                cctx->compressionLevel == ZSTD_CLEVEL_CUSTOM ?
-                cctx->requestedParams.cParams :
-                ZSTD_getCParams(cctx->compressionLevel, 0, dictSize);
+                ZSTD_getCParamsFromCCtxParams(cctx->requestedParams, 0, dictSize);
         cctx->cdictLocal = ZSTD_createCDict_advanced(
                                 dict, dictSize,
-                                cctx->dictContentByRef, cctx->dictMode,
+                                dictLoadMethod, dictMode,
                                 cParams, cctx->customMem);
         cctx->cdict = cctx->cdictLocal;
         if (cctx->cdictLocal == NULL)
@@ -407,21 +542,41 @@ ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, s
     return 0;
 }
 
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(
+      ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
+{
+    return ZSTD_CCtx_loadDictionary_advanced(
+            cctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dm_auto);
+}
+
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
+{
+    return ZSTD_CCtx_loadDictionary_advanced(
+            cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dm_auto);
+}
+
+
 size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
 {
     if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
     cctx->cdict = cdict;
-    cctx->prefix = NULL;   /* exclusive */
-    cctx->prefixSize = 0;
+    memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));  /* exclusive */
     return 0;
 }
 
 size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize)
 {
+    return ZSTD_CCtx_refPrefix_advanced(cctx, prefix, prefixSize, ZSTD_dm_rawContent);
+}
+
+size_t ZSTD_CCtx_refPrefix_advanced(
+        ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictMode_e dictMode)
+{
     if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
     cctx->cdict = NULL;   /* prefix discards any prior cdict */
-    cctx->prefix = prefix;
-    cctx->prefixSize = prefixSize;
+    cctx->prefixDict.dict = prefix;
+    cctx->prefixDict.dictSize = prefixSize;
+    cctx->prefixDict.dictMode = dictMode;
     return 0;
 }
 
@@ -484,28 +639,37 @@ static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)
 
 /** ZSTD_adjustCParams_internal() :
     optimize `cPar` for a given input (`srcSize` and `dictSize`).
-    mostly downsizing to reduce memory consumption and initialization.
-    Both `srcSize` and `dictSize` are optional (use 0 if unknown),
-    but if both are 0, no optimization can be done.
-    Note : cPar is considered validated at this stage. Use ZSTD_checkParams() to ensure that. */
+    mostly downsizing to reduce memory consumption and initialization latency.
+    Both `srcSize` and `dictSize` are optional (use 0 if unknown).
+    Note : cPar is considered validated at this stage. Use ZSTD_checkCParams() to ensure that condition. */
 ZSTD_compressionParameters ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize)
 {
+    static const U64 minSrcSize = 513; /* (1<<9) + 1 */
+    static const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1);
     assert(ZSTD_checkCParams(cPar)==0);
-    if (srcSize+dictSize == 0) return cPar;   /* no size information available : no adjustment */
-
-    /* resize params, to use less memory when necessary */
-    {   U32 const minSrcSize = (srcSize==0) ? 500 : 0;
-        U64 const rSize = srcSize + dictSize + minSrcSize;
-        if (rSize < ((U64)1<<ZSTD_WINDOWLOG_MAX)) {
-            U32 const srcLog = MAX(ZSTD_HASHLOG_MIN, ZSTD_highbit32((U32)(rSize)-1) + 1);
-            if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;
-    }   }
+
+    if (dictSize && (srcSize+1<2) /* srcSize unknown */ )
+        srcSize = minSrcSize;  /* presumed small when there is a dictionary */
+    else if (srcSize == 0)
+        srcSize = ZSTD_CONTENTSIZE_UNKNOWN;  /* 0 == unknown : presumed large */
+
+    /* resize windowLog if input is small enough, to use less memory */
+    if ( (srcSize < maxWindowResize)
+      && (dictSize < maxWindowResize) )  {
+        U32 const tSize = (U32)(srcSize + dictSize);
+        static U32 const hashSizeMin = 1 << ZSTD_HASHLOG_MIN;
+        U32 const srcLog = (tSize < hashSizeMin) ? ZSTD_HASHLOG_MIN :
+                            ZSTD_highbit32(tSize-1) + 1;
+        if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;
+    }
     if (cPar.hashLog > cPar.windowLog) cPar.hashLog = cPar.windowLog;
     {   U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);
-        if (cycleLog > cPar.windowLog) cPar.chainLog -= (cycleLog - cPar.windowLog);
+        if (cycleLog > cPar.windowLog)
+            cPar.chainLog -= (cycleLog - cPar.windowLog);
     }
 
-    if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN) cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN;  /* required for frame header */
+    if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)
+        cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN;  /* required for frame header */
 
     return cPar;
 }
@@ -516,56 +680,81 @@ ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, u
     return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize);
 }
 
-
-size_t ZSTD_estimateCCtxSize_advanced(ZSTD_compressionParameters cParams)
+size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params)
 {
-    size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
-    U32    const divider = (cParams.searchLength==3) ? 3 : 4;
-    size_t const maxNbSeq = blockSize / divider;
-    size_t const tokenSpace = blockSize + 11*maxNbSeq;
+    /* Estimate CCtx size is supported for single-threaded compression only. */
+    if (params->nbThreads > 1) { return ERROR(GENERIC); }
+    {   ZSTD_compressionParameters const cParams =
+                ZSTD_getCParamsFromCCtxParams(*params, 0, 0);
+        size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
+        U32    const divider = (cParams.searchLength==3) ? 3 : 4;
+        size_t const maxNbSeq = blockSize / divider;
+        size_t const tokenSpace = blockSize + 11*maxNbSeq;
+        size_t const chainSize =
+                (cParams.strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams.chainLog);
+        size_t const hSize = ((size_t)1) << cParams.hashLog;
+        U32    const hashLog3 = (cParams.searchLength>3) ?
+                                0 : MIN(ZSTD_HASHLOG3_MAX, cParams.windowLog);
+        size_t const h3Size = ((size_t)1) << hashLog3;
+        size_t const entropySpace = sizeof(ZSTD_entropyCTables_t);
+        size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
+
+        size_t const optBudget =
+                ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits))*sizeof(U32)
+                + (ZSTD_OPT_NUM+1)*(sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));
+        size_t const optSpace = ((cParams.strategy == ZSTD_btopt) || (cParams.strategy == ZSTD_btultra)) ? optBudget : 0;
 
-    size_t const chainSize = (cParams.strategy == ZSTD_fast) ? 0 : (1 << cParams.chainLog);
-    size_t const hSize = ((size_t)1) << cParams.hashLog;
-    U32    const hashLog3 = (cParams.searchLength>3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, cParams.windowLog);
-    size_t const h3Size = ((size_t)1) << hashLog3;
-    size_t const entropySpace = sizeof(ZSTD_entropyCTables_t);
-    size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
+        size_t const ldmSpace = params->ldmParams.enableLdm ?
+            ZSTD_ldm_getTableSize(params->ldmParams.hashLog,
+                                  params->ldmParams.bucketSizeLog) : 0;
 
-    size_t const optBudget = ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits))*sizeof(U32)
-                          + (ZSTD_OPT_NUM+1)*(sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));
-    size_t const optSpace = ((cParams.strategy == ZSTD_btopt) || (cParams.strategy == ZSTD_btultra)) ? optBudget : 0;
-    size_t const neededSpace = entropySpace + tableSpace + tokenSpace + optSpace;
+        size_t const neededSpace = entropySpace + tableSpace + tokenSpace +
+                                   optSpace + ldmSpace;
 
-    DEBUGLOG(5, "sizeof(ZSTD_CCtx) : %u", (U32)sizeof(ZSTD_CCtx));
-    DEBUGLOG(5, "estimate workSpace : %u", (U32)neededSpace);
-    return sizeof(ZSTD_CCtx) + neededSpace;
+        DEBUGLOG(5, "sizeof(ZSTD_CCtx) : %u", (U32)sizeof(ZSTD_CCtx));
+        DEBUGLOG(5, "estimate workSpace : %u", (U32)neededSpace);
+        return sizeof(ZSTD_CCtx) + neededSpace;
+    }
+}
+
+size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams)
+{
+    ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams);
+    return ZSTD_estimateCCtxSize_usingCCtxParams(&params);
 }
 
 size_t ZSTD_estimateCCtxSize(int compressionLevel)
 {
     ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, 0);
-    return ZSTD_estimateCCtxSize_advanced(cParams);
+    return ZSTD_estimateCCtxSize_usingCParams(cParams);
 }
 
-size_t ZSTD_estimateCStreamSize_advanced(ZSTD_compressionParameters cParams)
+size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)
 {
-    size_t const CCtxSize = ZSTD_estimateCCtxSize_advanced(cParams);
-    size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
-    size_t const inBuffSize = ((size_t)1 << cParams.windowLog) + blockSize;
-    size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1;
-    size_t const streamingSize = inBuffSize + outBuffSize;
+    if (params->nbThreads > 1) { return ERROR(GENERIC); }
+    {   size_t const CCtxSize = ZSTD_estimateCCtxSize_usingCCtxParams(params);
+        size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << params->cParams.windowLog);
+        size_t const inBuffSize = ((size_t)1 << params->cParams.windowLog) + blockSize;
+        size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1;
+        size_t const streamingSize = inBuffSize + outBuffSize;
 
-    return CCtxSize + streamingSize;
+        return CCtxSize + streamingSize;
+    }
+}
+
+size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams)
+{
+    ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams);
+    return ZSTD_estimateCStreamSize_usingCCtxParams(&params);
 }
 
 size_t ZSTD_estimateCStreamSize(int compressionLevel) {
     ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, 0);
-    return ZSTD_estimateCStreamSize_advanced(cParams);
+    return ZSTD_estimateCStreamSize_usingCParams(cParams);
 }
 
-
-static U32 ZSTD_equivalentParams(ZSTD_compressionParameters cParams1,
-                                 ZSTD_compressionParameters cParams2)
+static U32 ZSTD_equivalentCParams(ZSTD_compressionParameters cParams1,
+                                  ZSTD_compressionParameters cParams2)
 {
     U32 bslog1 = MIN(cParams1.windowLog, ZSTD_BLOCKSIZELOG_MAX);
     U32 bslog2 = MIN(cParams2.windowLog, ZSTD_BLOCKSIZELOG_MAX);
@@ -576,18 +765,39 @@ static U32 ZSTD_equivalentParams(ZSTD_compressionParameters cParams1,
          & ((cParams1.searchLength==3) == (cParams2.searchLength==3));  /* hashlog3 space */
 }
 
+/** The parameters are equivalent if ldm is not enabled in both sets or
+ *  all the parameters are equivalent. */
+static U32 ZSTD_equivalentLdmParams(ldmParams_t ldmParams1,
+                                    ldmParams_t ldmParams2)
+{
+    return (!ldmParams1.enableLdm && !ldmParams2.enableLdm) ||
+           (ldmParams1.enableLdm == ldmParams2.enableLdm &&
+            ldmParams1.hashLog == ldmParams2.hashLog &&
+            ldmParams1.bucketSizeLog == ldmParams2.bucketSizeLog &&
+            ldmParams1.minMatchLength == ldmParams2.minMatchLength &&
+            ldmParams1.hashEveryLog == ldmParams2.hashEveryLog);
+}
+
+/** Equivalence for resetCCtx purposes */
+static U32 ZSTD_equivalentParams(ZSTD_CCtx_params params1,
+                                 ZSTD_CCtx_params params2)
+{
+    return ZSTD_equivalentCParams(params1.cParams, params2.cParams) &&
+           ZSTD_equivalentLdmParams(params1.ldmParams, params2.ldmParams);
+}
+
 /*! ZSTD_continueCCtx() :
  *  reuse CCtx without reset (note : requires no dictionary) */
-static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_parameters params, U64 pledgedSrcSize)
+static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_CCtx_params params, U64 pledgedSrcSize)
 {
     U32 const end = (U32)(cctx->nextSrc - cctx->base);
-    DEBUGLOG(5, "continue mode");
+    DEBUGLOG(4, "continue mode");
     cctx->appliedParams = params;
     cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;
     cctx->consumedSrcSize = 0;
     if (pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN)
         cctx->appliedParams.fParams.contentSizeFlag = 0;
-    DEBUGLOG(5, "pledged content size : %u ; flag : %u",
+    DEBUGLOG(4, "pledged content size : %u ; flag : %u",
         (U32)pledgedSrcSize, cctx->appliedParams.fParams.contentSizeFlag);
     cctx->lowLimit = end;
     cctx->dictLimit = end;
@@ -607,15 +817,19 @@ typedef enum { ZSTDb_not_buffered, ZSTDb_buffered } ZSTD_buffered_policy_e;
 /*! ZSTD_resetCCtx_internal() :
     note : `params` are assumed fully validated at this stage */
 static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
-                                      ZSTD_parameters params, U64 pledgedSrcSize,
+                                      ZSTD_CCtx_params params, U64 pledgedSrcSize,
                                       ZSTD_compResetPolicy_e const crp,
                                       ZSTD_buffered_policy_e const zbuff)
 {
+    DEBUGLOG(4, "ZSTD_resetCCtx_internal");
     assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
+    DEBUGLOG(4, "pledgedSrcSize: %u", (U32)pledgedSrcSize);
 
     if (crp == ZSTDcrp_continue) {
-        if (ZSTD_equivalentParams(params.cParams, zc->appliedParams.cParams)) {
-            DEBUGLOG(5, "ZSTD_equivalentParams()==1");
+        if (ZSTD_equivalentParams(params, zc->appliedParams)) {
+            DEBUGLOG(4, "ZSTD_equivalentParams()==1");
+            assert(!(params.ldmParams.enableLdm &&
+                     params.ldmParams.hashEveryLog == ZSTD_LDM_HASHEVERYLOG_NOTSET));
             zc->entropy->hufCTable_repeatMode = HUF_repeat_none;
             zc->entropy->offcode_repeatMode = FSE_repeat_none;
             zc->entropy->matchlength_repeatMode = FSE_repeat_none;
@@ -623,12 +837,21 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
             return ZSTD_continueCCtx(zc, params, pledgedSrcSize);
     }   }
 
+    if (params.ldmParams.enableLdm) {
+        /* Adjust long distance matching parameters */
+        ZSTD_ldm_adjustParameters(&params.ldmParams, params.cParams.windowLog);
+        assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog);
+        assert(params.ldmParams.hashEveryLog < 32);
+        zc->ldmState.hashPower =
+                ZSTD_ldm_getHashPower(params.ldmParams.minMatchLength);
+    }
+
     {   size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << params.cParams.windowLog);
         U32    const divider = (params.cParams.searchLength==3) ? 3 : 4;
         size_t const maxNbSeq = blockSize / divider;
         size_t const tokenSpace = blockSize + 11*maxNbSeq;
         size_t const chainSize = (params.cParams.strategy == ZSTD_fast) ?
-                                0 : (1 << params.cParams.chainLog);
+                                0 : ((size_t)1 << params.cParams.chainLog);
         size_t const hSize = ((size_t)1) << params.cParams.hashLog;
         U32    const hashLog3 = (params.cParams.searchLength>3) ?
                                 0 : MIN(ZSTD_HASHLOG3_MAX, params.cParams.windowLog);
@@ -646,10 +869,13 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
                                     || (params.cParams.strategy == ZSTD_btultra)) ?
                                     optPotentialSpace : 0;
             size_t const bufferSpace = buffInSize + buffOutSize;
-            size_t const neededSpace = entropySpace + optSpace + tableSpace
-                                     + tokenSpace + bufferSpace;
+            size_t const ldmSpace = params.ldmParams.enableLdm
+                ? ZSTD_ldm_getTableSize(params.ldmParams.hashLog, params.ldmParams.bucketSizeLog)
+                : 0;
+            size_t const neededSpace = entropySpace + optSpace + ldmSpace +
+                                       tableSpace + tokenSpace + bufferSpace;
 
-            if (zc->workSpaceSize < neededSpace) {  /* too small : resize /*/
+            if (zc->workSpaceSize < neededSpace) {  /* too small : resize */
                 DEBUGLOG(5, "Need to update workSpaceSize from %uK to %uK \n",
                             (unsigned)zc->workSpaceSize>>10,
                             (unsigned)neededSpace>>10);
@@ -714,6 +940,16 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
             ptr = zc->optState.priceTable + ZSTD_OPT_NUM+1;
         }
 
+        /* ldm hash table */
+        /* initialize bucketOffsets table later for pointer alignment */
+        if (params.ldmParams.enableLdm) {
+            size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog;
+            memset(ptr, 0, ldmHSize * sizeof(ldmEntry_t));
+            assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */
+            zc->ldmState.hashTable = (ldmEntry_t*)ptr;
+            ptr = zc->ldmState.hashTable + ldmHSize;
+        }
+
         /* table Space */
         if (crp!=ZSTDcrp_noMemset) memset(ptr, 0, tableSpace);   /* reset tables only */
         assert(((size_t)ptr & 3) == 0);  /* ensure ptr is properly aligned */
@@ -731,6 +967,16 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
         zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq;
         ptr = zc->seqStore.litStart + blockSize;
 
+        /* ldm bucketOffsets table */
+        if (params.ldmParams.enableLdm) {
+            size_t const ldmBucketSize =
+                  ((size_t)1) << (params.ldmParams.hashLog -
+                                  params.ldmParams.bucketSizeLog);
+            memset(ptr, 0, ldmBucketSize);
+            zc->ldmState.bucketOffsets = (BYTE*)ptr;
+            ptr = zc->ldmState.bucketOffsets + ldmBucketSize;
+        }
+
         /* buffers */
         zc->inBuffSize = buffInSize;
         zc->inBuff = (char*)ptr;
@@ -753,6 +999,8 @@ void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx) {
 
 /*! ZSTD_copyCCtx_internal() :
  *  Duplicate an existing context `srcCCtx` into another one `dstCCtx`.
+ *  The "context", in this case, refers to the hash and chain tables, entropy
+ *  tables, and dictionary offsets.
  *  Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).
  *  pledgedSrcSize=0 means "empty" if fParams.contentSizeFlag=1
  *  @return : 0, or an error code */
@@ -766,14 +1014,16 @@ static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx,
     if (srcCCtx->stage!=ZSTDcs_init) return ERROR(stage_wrong);
 
     memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
-    {   ZSTD_parameters params = srcCCtx->appliedParams;
+    {   ZSTD_CCtx_params params = dstCCtx->requestedParams;
+        /* Copy only compression parameters related to tables. */
+        params.cParams = srcCCtx->appliedParams.cParams;
         params.fParams = fParams;
         ZSTD_resetCCtx_internal(dstCCtx, params, pledgedSrcSize,
                                 ZSTDcrp_noMemset, zbuff);
     }
 
     /* copy tables */
-    {   size_t const chainSize = (srcCCtx->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : (1 << srcCCtx->appliedParams.cParams.chainLog);
+    {   size_t const chainSize = (srcCCtx->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : ((size_t)1 << srcCCtx->appliedParams.cParams.chainLog);
         size_t const hSize =  (size_t)1 << srcCCtx->appliedParams.cParams.hashLog;
         size_t const h3Size = (size_t)1 << srcCCtx->hashLog3;
         size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
@@ -826,18 +1076,36 @@ static void ZSTD_reduceTable (U32* const table, U32 const size, U32 const reduce
     }
 }
 
+/*! ZSTD_ldm_reduceTable() :
+ *  reduce table indexes by `reducerValue` */
+static void ZSTD_ldm_reduceTable(ldmEntry_t* const table, U32 const size,
+                                 U32 const reducerValue)
+{
+    U32 u;
+    for (u = 0; u < size; u++) {
+        if (table[u].offset < reducerValue) table[u].offset = 0;
+        else table[u].offset -= reducerValue;
+    }
+}
+
 /*! ZSTD_reduceIndex() :
 *   rescale all indexes to avoid future overflow (indexes are U32) */
 static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue)
 {
-    { U32 const hSize = 1 << zc->appliedParams.cParams.hashLog;
+    { U32 const hSize = (U32)1 << zc->appliedParams.cParams.hashLog;
       ZSTD_reduceTable(zc->hashTable, hSize, reducerValue); }
 
-    { U32 const chainSize = (zc->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : (1 << zc->appliedParams.cParams.chainLog);
+    { U32 const chainSize = (zc->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : ((U32)1 << zc->appliedParams.cParams.chainLog);
       ZSTD_reduceTable(zc->chainTable, chainSize, reducerValue); }
 
-    { U32 const h3Size = (zc->hashLog3) ? 1 << zc->hashLog3 : 0;
+    { U32 const h3Size = (zc->hashLog3) ? (U32)1 << zc->hashLog3 : 0;
       ZSTD_reduceTable(zc->hashTable3, h3Size, reducerValue); }
+
+    { if (zc->appliedParams.ldmParams.enableLdm) {
+          U32 const ldmHSize = (U32)1 << zc->appliedParams.ldmParams.hashLog;
+          ZSTD_ldm_reduceTable(zc->ldmState.hashTable, ldmHSize, reducerValue);
+      }
+    }
 }
 
 
@@ -976,24 +1244,6 @@ static size_t ZSTD_compressLiterals (ZSTD_entropyCTables_t * entropy,
     return lhSize+cLitSize;
 }
 
-static const BYTE LL_Code[64] = {  0,  1,  2,  3,  4,  5,  6,  7,
-                                   8,  9, 10, 11, 12, 13, 14, 15,
-                                  16, 16, 17, 17, 18, 18, 19, 19,
-                                  20, 20, 20, 20, 21, 21, 21, 21,
-                                  22, 22, 22, 22, 22, 22, 22, 22,
-                                  23, 23, 23, 23, 23, 23, 23, 23,
-                                  24, 24, 24, 24, 24, 24, 24, 24,
-                                  24, 24, 24, 24, 24, 24, 24, 24 };
-
-static const BYTE ML_Code[128] = { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15,
-                                  16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,
-                                  32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37,
-                                  38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39,
-                                  40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,
-                                  41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41,
-                                  42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42,
-                                  42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 };
-
 
 void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)
 {
@@ -1018,20 +1268,30 @@ void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)
         mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
 }
 
-MEM_STATIC symbolEncodingType_e ZSTD_selectEncodingType(FSE_repeat* repeatMode,
-        size_t const mostFrequent, size_t nbSeq, U32 defaultNormLog)
+typedef enum {
+    ZSTD_defaultDisallowed = 0,
+    ZSTD_defaultAllowed = 1
+} ZSTD_defaultPolicy_e;
+
+MEM_STATIC symbolEncodingType_e ZSTD_selectEncodingType(
+        FSE_repeat* repeatMode, size_t const mostFrequent, size_t nbSeq,
+        U32 defaultNormLog, ZSTD_defaultPolicy_e const isDefaultAllowed)
 {
 #define MIN_SEQ_FOR_DYNAMIC_FSE   64
 #define MAX_SEQ_FOR_STATIC_FSE  1000
-
-    if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
+    ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0);
+    if ((mostFrequent == nbSeq) && (!isDefaultAllowed || nbSeq > 2)) {
+        /* Prefer set_basic over set_rle when there are 2 or less symbols,
+         * since RLE uses 1 byte, but set_basic uses 5-6 bits per symbol.
+         * If basic encoding isn't possible, always choose RLE.
+         */
         *repeatMode = FSE_repeat_check;
         return set_rle;
     }
-    if ((*repeatMode == FSE_repeat_valid) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
+    if (isDefaultAllowed && (*repeatMode == FSE_repeat_valid) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
         return set_repeat;
     }
-    if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (defaultNormLog-1)))) {
+    if (isDefaultAllowed && ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (defaultNormLog-1))))) {
         *repeatMode = FSE_repeat_valid;
         return set_basic;
     }
@@ -1067,6 +1327,7 @@ MEM_STATIC size_t ZSTD_buildCTable(void* dst, size_t dstCapacity,
             count[codeTable[nbSeq-1]]--;
             nbSeq_1--;
         }
+        assert(nbSeq_1 > 1);
         CHECK_F(FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max));
         {   size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog);   /* overflow protected */
             if (FSE_isError(NCountSize)) return NCountSize;
@@ -1131,7 +1392,7 @@ MEM_STATIC size_t ZSTD_encodeSequences(void* dst, size_t dstCapacity,
             BIT_addBits(&blockStream, sequences[n].litLength, llBits);
             if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);
             BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
-            if (MEM_32bits()) BIT_flushBits(&blockStream);                  /* (7)*/
+            if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream);
             if (longOffsets) {
                 int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);
                 if (extraBits) {
@@ -1204,7 +1465,7 @@ MEM_STATIC size_t ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,
     /* CTable for Literal Lengths */
     {   U32 max = MaxLL;
         size_t const mostFrequent = FSE_countFast_wksp(count, &max, llCodeTable, nbSeq, entropy->workspace);
-        LLtype = ZSTD_selectEncodingType(&entropy->litlength_repeatMode, mostFrequent, nbSeq, LL_defaultNormLog);
+        LLtype = ZSTD_selectEncodingType(&entropy->litlength_repeatMode, mostFrequent, nbSeq, LL_defaultNormLog, ZSTD_defaultAllowed);
         {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
                     count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL,
                     entropy->workspace, sizeof(entropy->workspace));
@@ -1214,9 +1475,11 @@ MEM_STATIC size_t ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,
     /* CTable for Offsets */
     {   U32 max = MaxOff;
         size_t const mostFrequent = FSE_countFast_wksp(count, &max, ofCodeTable, nbSeq, entropy->workspace);
-        Offtype = ZSTD_selectEncodingType(&entropy->offcode_repeatMode, mostFrequent, nbSeq, OF_defaultNormLog);
+        /* We can only use the basic table if max <= DefaultMaxOff, otherwise the offsets are too large */
+        ZSTD_defaultPolicy_e const defaultPolicy = max <= DefaultMaxOff ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed;
+        Offtype = ZSTD_selectEncodingType(&entropy->offcode_repeatMode, mostFrequent, nbSeq, OF_defaultNormLog, defaultPolicy);
         {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
-                    count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, MaxOff,
+                    count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
                     entropy->workspace, sizeof(entropy->workspace));
             if (ZSTD_isError(countSize)) return countSize;
             op += countSize;
@@ -1224,7 +1487,7 @@ MEM_STATIC size_t ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,
     /* CTable for MatchLengths */
     {   U32 max = MaxML;
         size_t const mostFrequent = FSE_countFast_wksp(count, &max, mlCodeTable, nbSeq, entropy->workspace);
-        MLtype = ZSTD_selectEncodingType(&entropy->matchlength_repeatMode, mostFrequent, nbSeq, ML_defaultNormLog);
+        MLtype = ZSTD_selectEncodingType(&entropy->matchlength_repeatMode, mostFrequent, nbSeq, ML_defaultNormLog, ZSTD_defaultAllowed);
         {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
                     count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML,
                     entropy->workspace, sizeof(entropy->workspace));
@@ -1279,1528 +1542,11 @@ MEM_STATIC size_t ZSTD_compressSequences(seqStore_t* seqStorePtr,
     return cSize;
 }
 
-
-/*! ZSTD_storeSeq() :
-    Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.
-    `offsetCode` : distance to match, or 0 == repCode.
-    `matchCode` : matchLength - MINMATCH
-*/
-MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t matchCode)
-{
-#if defined(ZSTD_DEBUG) && (ZSTD_DEBUG >= 6)
-    static const BYTE* g_start = NULL;
-    U32 const pos = (U32)((const BYTE*)literals - g_start);
-    if (g_start==NULL) g_start = (const BYTE*)literals;
-    if ((pos > 0) && (pos < 1000000000))
-        DEBUGLOG(6, "Cpos %6u :%5u literals & match %3u bytes at distance %6u",
-               pos, (U32)litLength, (U32)matchCode+MINMATCH, (U32)offsetCode);
-#endif
-    /* copy Literals */
-    assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + 128 KB);
-    ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);
-    seqStorePtr->lit += litLength;
-
-    /* literal Length */
-    if (litLength>0xFFFF) {
-        seqStorePtr->longLengthID = 1;
-        seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-    }
-    seqStorePtr->sequences[0].litLength = (U16)litLength;
-
-    /* match offset */
-    seqStorePtr->sequences[0].offset = offsetCode + 1;
-
-    /* match Length */
-    if (matchCode>0xFFFF) {
-        seqStorePtr->longLengthID = 2;
-        seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-    }
-    seqStorePtr->sequences[0].matchLength = (U16)matchCode;
-
-    seqStorePtr->sequences++;
-}
-
-
-/*-*************************************
-*  Match length counter
-***************************************/
-static unsigned ZSTD_NbCommonBytes (register size_t val)
-{
-    if (MEM_isLittleEndian()) {
-        if (MEM_64bits()) {
-#       if defined(_MSC_VER) && defined(_WIN64)
-            unsigned long r = 0;
-            _BitScanForward64( &r, (U64)val );
-            return (unsigned)(r>>3);
-#       elif defined(__GNUC__) && (__GNUC__ >= 3)
-            return (__builtin_ctzll((U64)val) >> 3);
-#       else
-            static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2,
-                                                     0, 3, 1, 3, 1, 4, 2, 7,
-                                                     0, 2, 3, 6, 1, 5, 3, 5,
-                                                     1, 3, 4, 4, 2, 5, 6, 7,
-                                                     7, 0, 1, 2, 3, 3, 4, 6,
-                                                     2, 6, 5, 5, 3, 4, 5, 6,
-                                                     7, 1, 2, 4, 6, 4, 4, 5,
-                                                     7, 2, 6, 5, 7, 6, 7, 7 };
-            return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];
-#       endif
-        } else { /* 32 bits */
-#       if defined(_MSC_VER)
-            unsigned long r=0;
-            _BitScanForward( &r, (U32)val );
-            return (unsigned)(r>>3);
-#       elif defined(__GNUC__) && (__GNUC__ >= 3)
-            return (__builtin_ctz((U32)val) >> 3);
-#       else
-            static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0,
-                                                     3, 2, 2, 1, 3, 2, 0, 1,
-                                                     3, 3, 1, 2, 2, 2, 2, 0,
-                                                     3, 1, 2, 0, 1, 0, 1, 1 };
-            return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
-#       endif
-        }
-    } else {  /* Big Endian CPU */
-        if (MEM_64bits()) {
-#       if defined(_MSC_VER) && defined(_WIN64)
-            unsigned long r = 0;
-            _BitScanReverse64( &r, val );
-            return (unsigned)(r>>3);
-#       elif defined(__GNUC__) && (__GNUC__ >= 3)
-            return (__builtin_clzll(val) >> 3);
-#       else
-            unsigned r;
-            const unsigned n32 = sizeof(size_t)*4;   /* calculate this way due to compiler complaining in 32-bits mode */
-            if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; }
-            if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
-            r += (!val);
-            return r;
-#       endif
-        } else { /* 32 bits */
-#       if defined(_MSC_VER)
-            unsigned long r = 0;
-            _BitScanReverse( &r, (unsigned long)val );
-            return (unsigned)(r>>3);
-#       elif defined(__GNUC__) && (__GNUC__ >= 3)
-            return (__builtin_clz((U32)val) >> 3);
-#       else
-            unsigned r;
-            if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
-            r += (!val);
-            return r;
-#       endif
-    }   }
-}
-
-
-static size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit)
-{
-    const BYTE* const pStart = pIn;
-    const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1);
-
-    while (pIn < pInLoopLimit) {
-        size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);
-        if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; }
-        pIn += ZSTD_NbCommonBytes(diff);
-        return (size_t)(pIn - pStart);
-    }
-    if (MEM_64bits()) if ((pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; }
-    if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; }
-    if ((pIn<pInLimit) && (*pMatch == *pIn)) pIn++;
-    return (size_t)(pIn - pStart);
-}
-
-/** ZSTD_count_2segments() :
-*   can count match length with `ip` & `match` in 2 different segments.
-*   convention : on reaching mEnd, match count continue starting from iStart
-*/
-static size_t ZSTD_count_2segments(const BYTE* ip, const BYTE* match, const BYTE* iEnd, const BYTE* mEnd, const BYTE* iStart)
-{
-    const BYTE* const vEnd = MIN( ip + (mEnd - match), iEnd);
-    size_t const matchLength = ZSTD_count(ip, match, vEnd);
-    if (match + matchLength != mEnd) return matchLength;
-    return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd);
-}
-
-
-/*-*************************************
-*  Hashes
-***************************************/
-static const U32 prime3bytes = 506832829U;
-static U32    ZSTD_hash3(U32 u, U32 h) { return ((u << (32-24)) * prime3bytes)  >> (32-h) ; }
-MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h); } /* only in zstd_opt.h */
-
-static const U32 prime4bytes = 2654435761U;
-static U32    ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32-h) ; }
-static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_read32(ptr), h); }
-
-static const U64 prime5bytes = 889523592379ULL;
-static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u  << (64-40)) * prime5bytes) >> (64-h)) ; }
-static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h); }
-
-static const U64 prime6bytes = 227718039650203ULL;
-static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u  << (64-48)) * prime6bytes) >> (64-h)) ; }
-static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }
-
-static const U64 prime7bytes = 58295818150454627ULL;
-static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u  << (64-56)) * prime7bytes) >> (64-h)) ; }
-static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h); }
-
-static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
-static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }
-static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }
-
-static size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)
-{
-    switch(mls)
-    {
-    default:
-    case 4: return ZSTD_hash4Ptr(p, hBits);
-    case 5: return ZSTD_hash5Ptr(p, hBits);
-    case 6: return ZSTD_hash6Ptr(p, hBits);
-    case 7: return ZSTD_hash7Ptr(p, hBits);
-    case 8: return ZSTD_hash8Ptr(p, hBits);
-    }
-}
-
-
-/*-*************************************
-*  Fast Scan
-***************************************/
-static void ZSTD_fillHashTable (ZSTD_CCtx* zc, const void* end, const U32 mls)
-{
-    U32* const hashTable = zc->hashTable;
-    U32  const hBits = zc->appliedParams.cParams.hashLog;
-    const BYTE* const base = zc->base;
-    const BYTE* ip = base + zc->nextToUpdate;
-    const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
-    const size_t fastHashFillStep = 3;
-
-    while(ip <= iend) {
-        hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);
-        ip += fastHashFillStep;
-    }
-}
-
-
-FORCE_INLINE_TEMPLATE
-void ZSTD_compressBlock_fast_generic(ZSTD_CCtx* cctx,
-                               const void* src, size_t srcSize,
-                               const U32 mls)
-{
-    U32* const hashTable = cctx->hashTable;
-    U32  const hBits = cctx->appliedParams.cParams.hashLog;
-    seqStore_t* seqStorePtr = &(cctx->seqStore);
-    const BYTE* const base = cctx->base;
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const U32   lowestIndex = cctx->dictLimit;
-    const BYTE* const lowest = base + lowestIndex;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - HASH_READ_SIZE;
-    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
-    U32 offsetSaved = 0;
-
-    /* init */
-    ip += (ip==lowest);
-    {   U32 const maxRep = (U32)(ip-lowest);
-        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
-        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
-    }
-
-    /* Main Search Loop */
-    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
-        size_t mLength;
-        size_t const h = ZSTD_hashPtr(ip, hBits, mls);
-        U32 const current = (U32)(ip-base);
-        U32 const matchIndex = hashTable[h];
-        const BYTE* match = base + matchIndex;
-        hashTable[h] = current;   /* update hash table */
-
-        if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {
-            mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
-            ip++;
-            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
-        } else {
-            U32 offset;
-            if ( (matchIndex <= lowestIndex) || (MEM_read32(match) != MEM_read32(ip)) ) {
-                ip += ((ip-anchor) >> g_searchStrength) + 1;
-                continue;
-            }
-            mLength = ZSTD_count(ip+4, match+4, iend) + 4;
-            offset = (U32)(ip-match);
-            while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
-            offset_2 = offset_1;
-            offset_1 = offset;
-
-            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
-        }
-
-        /* match found */
-        ip += mLength;
-        anchor = ip;
-
-        if (ip <= ilimit) {
-            /* Fill Table */
-            hashTable[ZSTD_hashPtr(base+current+2, hBits, mls)] = current+2;  /* here because current+2 could be > iend-8 */
-            hashTable[ZSTD_hashPtr(ip-2, hBits, mls)] = (U32)(ip-2-base);
-            /* check immediate repcode */
-            while ( (ip <= ilimit)
-                 && ( (offset_2>0)
-                 & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
-                /* store sequence */
-                size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
-                { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; }  /* swap offset_2 <=> offset_1 */
-                hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip-base);
-                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);
-                ip += rLength;
-                anchor = ip;
-                continue;   /* faster when present ... (?) */
-    }   }   }
-
-    /* save reps for next block */
-    seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
-    seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-static void ZSTD_compressBlock_fast(ZSTD_CCtx* ctx,
-                       const void* src, size_t srcSize)
-{
-    const U32 mls = ctx->appliedParams.cParams.searchLength;
-    switch(mls)
-    {
-    default: /* includes case 3 */
-    case 4 :
-        ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 4); return;
-    case 5 :
-        ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 5); return;
-    case 6 :
-        ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 6); return;
-    case 7 :
-        ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 7); return;
-    }
-}
-
-
-static void ZSTD_compressBlock_fast_extDict_generic(ZSTD_CCtx* ctx,
-                                 const void* src, size_t srcSize,
-                                 const U32 mls)
-{
-    U32* hashTable = ctx->hashTable;
-    const U32 hBits = ctx->appliedParams.cParams.hashLog;
-    seqStore_t* seqStorePtr = &(ctx->seqStore);
-    const BYTE* const base = ctx->base;
-    const BYTE* const dictBase = ctx->dictBase;
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const U32   lowestIndex = ctx->lowLimit;
-    const BYTE* const dictStart = dictBase + lowestIndex;
-    const U32   dictLimit = ctx->dictLimit;
-    const BYTE* const lowPrefixPtr = base + dictLimit;
-    const BYTE* const dictEnd = dictBase + dictLimit;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - 8;
-    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
-
-    /* Search Loop */
-    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
-        const size_t h = ZSTD_hashPtr(ip, hBits, mls);
-        const U32 matchIndex = hashTable[h];
-        const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;
-        const BYTE* match = matchBase + matchIndex;
-        const U32 current = (U32)(ip-base);
-        const U32 repIndex = current + 1 - offset_1;   /* offset_1 expected <= current +1 */
-        const BYTE* repBase = repIndex < dictLimit ? dictBase : base;
-        const BYTE* repMatch = repBase + repIndex;
-        size_t mLength;
-        hashTable[h] = current;   /* update hash table */
-
-        if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))
-           && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
-            const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
-            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, lowPrefixPtr) + 4;
-            ip++;
-            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
-        } else {
-            if ( (matchIndex < lowestIndex) ||
-                 (MEM_read32(match) != MEM_read32(ip)) ) {
-                ip += ((ip-anchor) >> g_searchStrength) + 1;
-                continue;
-            }
-            {   const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;
-                const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
-                U32 offset;
-                mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, lowPrefixPtr) + 4;
-                while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
-                offset = current - matchIndex;
-                offset_2 = offset_1;
-                offset_1 = offset;
-                ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
-        }   }
-
-        /* found a match : store it */
-        ip += mLength;
-        anchor = ip;
-
-        if (ip <= ilimit) {
-            /* Fill Table */
-            hashTable[ZSTD_hashPtr(base+current+2, hBits, mls)] = current+2;
-            hashTable[ZSTD_hashPtr(ip-2, hBits, mls)] = (U32)(ip-2-base);
-            /* check immediate repcode */
-            while (ip <= ilimit) {
-                U32 const current2 = (U32)(ip-base);
-                U32 const repIndex2 = current2 - offset_2;
-                const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
-                if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex))  /* intentional overflow */
-                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
-                    const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
-                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, lowPrefixPtr) + 4;
-                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
-                    ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);
-                    hashTable[ZSTD_hashPtr(ip, hBits, mls)] = current2;
-                    ip += repLength2;
-                    anchor = ip;
-                    continue;
-                }
-                break;
-    }   }   }
-
-    /* save reps for next block */
-    seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-static void ZSTD_compressBlock_fast_extDict(ZSTD_CCtx* ctx,
-                         const void* src, size_t srcSize)
-{
-    U32 const mls = ctx->appliedParams.cParams.searchLength;
-    switch(mls)
-    {
-    default: /* includes case 3 */
-    case 4 :
-        ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 4); return;
-    case 5 :
-        ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 5); return;
-    case 6 :
-        ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 6); return;
-    case 7 :
-        ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 7); return;
-    }
-}
-
-
-/*-*************************************
-*  Double Fast
-***************************************/
-static void ZSTD_fillDoubleHashTable (ZSTD_CCtx* cctx, const void* end, const U32 mls)
-{
-    U32* const hashLarge = cctx->hashTable;
-    U32  const hBitsL = cctx->appliedParams.cParams.hashLog;
-    U32* const hashSmall = cctx->chainTable;
-    U32  const hBitsS = cctx->appliedParams.cParams.chainLog;
-    const BYTE* const base = cctx->base;
-    const BYTE* ip = base + cctx->nextToUpdate;
-    const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
-    const size_t fastHashFillStep = 3;
-
-    while(ip <= iend) {
-        hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
-        hashLarge[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
-        ip += fastHashFillStep;
-    }
-}
-
-
-FORCE_INLINE_TEMPLATE
-void ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx* cctx,
-                                 const void* src, size_t srcSize,
-                                 const U32 mls)
-{
-    U32* const hashLong = cctx->hashTable;
-    const U32 hBitsL = cctx->appliedParams.cParams.hashLog;
-    U32* const hashSmall = cctx->chainTable;
-    const U32 hBitsS = cctx->appliedParams.cParams.chainLog;
-    seqStore_t* seqStorePtr = &(cctx->seqStore);
-    const BYTE* const base = cctx->base;
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const U32 lowestIndex = cctx->dictLimit;
-    const BYTE* const lowest = base + lowestIndex;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - HASH_READ_SIZE;
-    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
-    U32 offsetSaved = 0;
-
-    /* init */
-    ip += (ip==lowest);
-    {   U32 const maxRep = (U32)(ip-lowest);
-        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
-        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
-    }
-
-    /* Main Search Loop */
-    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
-        size_t mLength;
-        size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);
-        size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);
-        U32 const current = (U32)(ip-base);
-        U32 const matchIndexL = hashLong[h2];
-        U32 const matchIndexS = hashSmall[h];
-        const BYTE* matchLong = base + matchIndexL;
-        const BYTE* match = base + matchIndexS;
-        hashLong[h2] = hashSmall[h] = current;   /* update hash tables */
-
-        assert(offset_1 <= current);   /* supposed guaranteed by construction */
-        if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {
-            /* favor repcode */
-            mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
-            ip++;
-            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
-        } else {
-            U32 offset;
-            if ( (matchIndexL > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip)) ) {
-                mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
-                offset = (U32)(ip-matchLong);
-                while (((ip>anchor) & (matchLong>lowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
-            } else if ( (matchIndexS > lowestIndex) && (MEM_read32(match) == MEM_read32(ip)) ) {
-                size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
-                U32 const matchIndexL3 = hashLong[hl3];
-                const BYTE* matchL3 = base + matchIndexL3;
-                hashLong[hl3] = current + 1;
-                if ( (matchIndexL3 > lowestIndex) && (MEM_read64(matchL3) == MEM_read64(ip+1)) ) {
-                    mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;
-                    ip++;
-                    offset = (U32)(ip-matchL3);
-                    while (((ip>anchor) & (matchL3>lowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */
-                } else {
-                    mLength = ZSTD_count(ip+4, match+4, iend) + 4;
-                    offset = (U32)(ip-match);
-                    while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
-                }
-            } else {
-                ip += ((ip-anchor) >> g_searchStrength) + 1;
-                continue;
-            }
-
-            offset_2 = offset_1;
-            offset_1 = offset;
-
-            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
-        }
-
-        /* match found */
-        ip += mLength;
-        anchor = ip;
-
-        if (ip <= ilimit) {
-            /* Fill Table */
-            hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] =
-                hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;  /* here because current+2 could be > iend-8 */
-            hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] =
-                hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
-
-            /* check immediate repcode */
-            while ( (ip <= ilimit)
-                 && ( (offset_2>0)
-                 & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
-                /* store sequence */
-                size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
-                { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */
-                hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);
-                hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);
-                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);
-                ip += rLength;
-                anchor = ip;
-                continue;   /* faster when present ... (?) */
-    }   }   }
-
-    /* save reps for next block */
-    seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
-    seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-static void ZSTD_compressBlock_doubleFast(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    const U32 mls = ctx->appliedParams.cParams.searchLength;
-    switch(mls)
-    {
-    default: /* includes case 3 */
-    case 4 :
-        ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 4); return;
-    case 5 :
-        ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 5); return;
-    case 6 :
-        ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 6); return;
-    case 7 :
-        ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 7); return;
-    }
-}
-
-
-static void ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx* ctx,
-                                 const void* src, size_t srcSize,
-                                 const U32 mls)
-{
-    U32* const hashLong = ctx->hashTable;
-    U32  const hBitsL = ctx->appliedParams.cParams.hashLog;
-    U32* const hashSmall = ctx->chainTable;
-    U32  const hBitsS = ctx->appliedParams.cParams.chainLog;
-    seqStore_t* seqStorePtr = &(ctx->seqStore);
-    const BYTE* const base = ctx->base;
-    const BYTE* const dictBase = ctx->dictBase;
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const U32   lowestIndex = ctx->lowLimit;
-    const BYTE* const dictStart = dictBase + lowestIndex;
-    const U32   dictLimit = ctx->dictLimit;
-    const BYTE* const lowPrefixPtr = base + dictLimit;
-    const BYTE* const dictEnd = dictBase + dictLimit;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - 8;
-    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
-
-    /* Search Loop */
-    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
-        const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
-        const U32 matchIndex = hashSmall[hSmall];
-        const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;
-        const BYTE* match = matchBase + matchIndex;
-
-        const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
-        const U32 matchLongIndex = hashLong[hLong];
-        const BYTE* matchLongBase = matchLongIndex < dictLimit ? dictBase : base;
-        const BYTE* matchLong = matchLongBase + matchLongIndex;
-
-        const U32 current = (U32)(ip-base);
-        const U32 repIndex = current + 1 - offset_1;   /* offset_1 expected <= current +1 */
-        const BYTE* repBase = repIndex < dictLimit ? dictBase : base;
-        const BYTE* repMatch = repBase + repIndex;
-        size_t mLength;
-        hashSmall[hSmall] = hashLong[hLong] = current;   /* update hash table */
-
-        if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))
-           && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
-            const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
-            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, lowPrefixPtr) + 4;
-            ip++;
-            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
-        } else {
-            if ((matchLongIndex > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
-                const BYTE* matchEnd = matchLongIndex < dictLimit ? dictEnd : iend;
-                const BYTE* lowMatchPtr = matchLongIndex < dictLimit ? dictStart : lowPrefixPtr;
-                U32 offset;
-                mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, lowPrefixPtr) + 8;
-                offset = current - matchLongIndex;
-                while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; }   /* catch up */
-                offset_2 = offset_1;
-                offset_1 = offset;
-                ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
-
-            } else if ((matchIndex > lowestIndex) && (MEM_read32(match) == MEM_read32(ip))) {
-                size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
-                U32 const matchIndex3 = hashLong[h3];
-                const BYTE* const match3Base = matchIndex3 < dictLimit ? dictBase : base;
-                const BYTE* match3 = match3Base + matchIndex3;
-                U32 offset;
-                hashLong[h3] = current + 1;
-                if ( (matchIndex3 > lowestIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {
-                    const BYTE* matchEnd = matchIndex3 < dictLimit ? dictEnd : iend;
-                    const BYTE* lowMatchPtr = matchIndex3 < dictLimit ? dictStart : lowPrefixPtr;
-                    mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, lowPrefixPtr) + 8;
-                    ip++;
-                    offset = current+1 - matchIndex3;
-                    while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */
-                } else {
-                    const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;
-                    const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
-                    mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, lowPrefixPtr) + 4;
-                    offset = current - matchIndex;
-                    while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
-                }
-                offset_2 = offset_1;
-                offset_1 = offset;
-                ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
-
-            } else {
-                ip += ((ip-anchor) >> g_searchStrength) + 1;
-                continue;
-        }   }
-
-        /* found a match : store it */
-        ip += mLength;
-        anchor = ip;
-
-        if (ip <= ilimit) {
-            /* Fill Table */
-            hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;
-            hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2;
-            hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
-            hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
-            /* check immediate repcode */
-            while (ip <= ilimit) {
-                U32 const current2 = (U32)(ip-base);
-                U32 const repIndex2 = current2 - offset_2;
-                const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
-                if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex))  /* intentional overflow */
-                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
-                    const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
-                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, lowPrefixPtr) + 4;
-                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
-                    ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);
-                    hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
-                    hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
-                    ip += repLength2;
-                    anchor = ip;
-                    continue;
-                }
-                break;
-    }   }   }
-
-    /* save reps for next block */
-    seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-static void ZSTD_compressBlock_doubleFast_extDict(ZSTD_CCtx* ctx,
-                         const void* src, size_t srcSize)
-{
-    U32 const mls = ctx->appliedParams.cParams.searchLength;
-    switch(mls)
-    {
-    default: /* includes case 3 */
-    case 4 :
-        ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 4); return;
-    case 5 :
-        ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 5); return;
-    case 6 :
-        ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 6); return;
-    case 7 :
-        ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 7); return;
-    }
-}
-
-
-/*-*************************************
-*  Binary Tree search
-***************************************/
-/** ZSTD_insertBt1() : add one or multiple positions to tree.
-*   ip : assumed <= iend-8 .
-*   @return : nb of positions added */
-static U32 ZSTD_insertBt1(ZSTD_CCtx* zc, const BYTE* const ip, const U32 mls, const BYTE* const iend, U32 nbCompares,
-                          U32 extDict)
-{
-    U32*   const hashTable = zc->hashTable;
-    U32    const hashLog = zc->appliedParams.cParams.hashLog;
-    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
-    U32*   const bt = zc->chainTable;
-    U32    const btLog  = zc->appliedParams.cParams.chainLog - 1;
-    U32    const btMask = (1 << btLog) - 1;
-    U32 matchIndex = hashTable[h];
-    size_t commonLengthSmaller=0, commonLengthLarger=0;
-    const BYTE* const base = zc->base;
-    const BYTE* const dictBase = zc->dictBase;
-    const U32 dictLimit = zc->dictLimit;
-    const BYTE* const dictEnd = dictBase + dictLimit;
-    const BYTE* const prefixStart = base + dictLimit;
-    const BYTE* match;
-    const U32 current = (U32)(ip-base);
-    const U32 btLow = btMask >= current ? 0 : current - btMask;
-    U32* smallerPtr = bt + 2*(current&btMask);
-    U32* largerPtr  = smallerPtr + 1;
-    U32 dummy32;   /* to be nullified at the end */
-    U32 const windowLow = zc->lowLimit;
-    U32 matchEndIdx = current+8;
-    size_t bestLength = 8;
-#ifdef ZSTD_C_PREDICT
-    U32 predictedSmall = *(bt + 2*((current-1)&btMask) + 0);
-    U32 predictedLarge = *(bt + 2*((current-1)&btMask) + 1);
-    predictedSmall += (predictedSmall>0);
-    predictedLarge += (predictedLarge>0);
-#endif /* ZSTD_C_PREDICT */
-
-    hashTable[h] = current;   /* Update Hash Table */
-
-    while (nbCompares-- && (matchIndex > windowLow)) {
-        U32* const nextPtr = bt + 2*(matchIndex & btMask);
-        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
-
-#ifdef ZSTD_C_PREDICT   /* note : can create issues when hlog small <= 11 */
-        const U32* predictPtr = bt + 2*((matchIndex-1) & btMask);   /* written this way, as bt is a roll buffer */
-        if (matchIndex == predictedSmall) {
-            /* no need to check length, result known */
-            *smallerPtr = matchIndex;
-            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
-            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
-            predictedSmall = predictPtr[1] + (predictPtr[1]>0);
-            continue;
-        }
-        if (matchIndex == predictedLarge) {
-            *largerPtr = matchIndex;
-            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            largerPtr = nextPtr;
-            matchIndex = nextPtr[0];
-            predictedLarge = predictPtr[0] + (predictPtr[0]>0);
-            continue;
-        }
-#endif
-        if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
-            match = base + matchIndex;
-            if (match[matchLength] == ip[matchLength])
-                matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
-        } else {
-            match = dictBase + matchIndex;
-            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
-            if (matchIndex+matchLength >= dictLimit)
-                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
-        }
-
-        if (matchLength > bestLength) {
-            bestLength = matchLength;
-            if (matchLength > matchEndIdx - matchIndex)
-                matchEndIdx = matchIndex + (U32)matchLength;
-        }
-
-        if (ip+matchLength == iend)   /* equal : no way to know if inf or sup */
-            break;   /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt the tree */
-
-        if (match[matchLength] < ip[matchLength]) {  /* necessarily within correct buffer */
-            /* match is smaller than current */
-            *smallerPtr = matchIndex;             /* update smaller idx */
-            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
-            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
-            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
-        } else {
-            /* match is larger than current */
-            *largerPtr = matchIndex;
-            commonLengthLarger = matchLength;
-            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            largerPtr = nextPtr;
-            matchIndex = nextPtr[0];
-    }   }
-
-    *smallerPtr = *largerPtr = 0;
-    if (bestLength > 384) return MIN(192, (U32)(bestLength - 384));   /* speed optimization */
-    if (matchEndIdx > current + 8) return matchEndIdx - current - 8;
-    return 1;
-}
-
-
-static size_t ZSTD_insertBtAndFindBestMatch (
-                        ZSTD_CCtx* zc,
-                        const BYTE* const ip, const BYTE* const iend,
-                        size_t* offsetPtr,
-                        U32 nbCompares, const U32 mls,
-                        U32 extDict)
-{
-    U32*   const hashTable = zc->hashTable;
-    U32    const hashLog = zc->appliedParams.cParams.hashLog;
-    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
-    U32*   const bt = zc->chainTable;
-    U32    const btLog  = zc->appliedParams.cParams.chainLog - 1;
-    U32    const btMask = (1 << btLog) - 1;
-    U32 matchIndex  = hashTable[h];
-    size_t commonLengthSmaller=0, commonLengthLarger=0;
-    const BYTE* const base = zc->base;
-    const BYTE* const dictBase = zc->dictBase;
-    const U32 dictLimit = zc->dictLimit;
-    const BYTE* const dictEnd = dictBase + dictLimit;
-    const BYTE* const prefixStart = base + dictLimit;
-    const U32 current = (U32)(ip-base);
-    const U32 btLow = btMask >= current ? 0 : current - btMask;
-    const U32 windowLow = zc->lowLimit;
-    U32* smallerPtr = bt + 2*(current&btMask);
-    U32* largerPtr  = bt + 2*(current&btMask) + 1;
-    U32 matchEndIdx = current+8;
-    U32 dummy32;   /* to be nullified at the end */
-    size_t bestLength = 0;
-
-    hashTable[h] = current;   /* Update Hash Table */
-
-    while (nbCompares-- && (matchIndex > windowLow)) {
-        U32* const nextPtr = bt + 2*(matchIndex & btMask);
-        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
-        const BYTE* match;
-
-        if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
-            match = base + matchIndex;
-            if (match[matchLength] == ip[matchLength])
-                matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
-        } else {
-            match = dictBase + matchIndex;
-            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
-            if (matchIndex+matchLength >= dictLimit)
-                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
-        }
-
-        if (matchLength > bestLength) {
-            if (matchLength > matchEndIdx - matchIndex)
-                matchEndIdx = matchIndex + (U32)matchLength;
-            if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )
-                bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;
-            if (ip+matchLength == iend)   /* equal : no way to know if inf or sup */
-                break;   /* drop, to guarantee consistency (miss a little bit of compression) */
-        }
-
-        if (match[matchLength] < ip[matchLength]) {
-            /* match is smaller than current */
-            *smallerPtr = matchIndex;             /* update smaller idx */
-            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
-            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
-            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
-        } else {
-            /* match is larger than current */
-            *largerPtr = matchIndex;
-            commonLengthLarger = matchLength;
-            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            largerPtr = nextPtr;
-            matchIndex = nextPtr[0];
-    }   }
-
-    *smallerPtr = *largerPtr = 0;
-
-    zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;
-    return bestLength;
-}
-
-
-static void ZSTD_updateTree(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls)
-{
-    const BYTE* const base = zc->base;
-    const U32 target = (U32)(ip - base);
-    U32 idx = zc->nextToUpdate;
-
-    while(idx < target)
-        idx += ZSTD_insertBt1(zc, base+idx, mls, iend, nbCompares, 0);
-}
-
-/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */
-static size_t ZSTD_BtFindBestMatch (
-                        ZSTD_CCtx* zc,
-                        const BYTE* const ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 mls)
-{
-    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
-    ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
-    return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 0);
-}
-
-
-static size_t ZSTD_BtFindBestMatch_selectMLS (
-                        ZSTD_CCtx* zc,   /* Index table will be updated */
-                        const BYTE* ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 matchLengthSearch)
-{
-    switch(matchLengthSearch)
-    {
-    default : /* includes case 3 */
-    case 4 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
-    case 5 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
-    case 7 :
-    case 6 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
-    }
-}
-
-
-static void ZSTD_updateTree_extDict(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls)
-{
-    const BYTE* const base = zc->base;
-    const U32 target = (U32)(ip - base);
-    U32 idx = zc->nextToUpdate;
-
-    while (idx < target) idx += ZSTD_insertBt1(zc, base+idx, mls, iend, nbCompares, 1);
-}
-
-
-/** Tree updater, providing best match */
-static size_t ZSTD_BtFindBestMatch_extDict (
-                        ZSTD_CCtx* zc,
-                        const BYTE* const ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 mls)
-{
-    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
-    ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
-    return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 1);
-}
-
-
-static size_t ZSTD_BtFindBestMatch_selectMLS_extDict (
-                        ZSTD_CCtx* zc,   /* Index table will be updated */
-                        const BYTE* ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 matchLengthSearch)
-{
-    switch(matchLengthSearch)
-    {
-    default : /* includes case 3 */
-    case 4 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
-    case 5 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
-    case 7 :
-    case 6 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
-    }
-}
-
-
-
-/* *********************************
-*  Hash Chain
-***********************************/
-#define NEXT_IN_CHAIN(d, mask)   chainTable[(d) & mask]
-
-/* Update chains up to ip (excluded)
-   Assumption : always within prefix (i.e. not within extDict) */
-FORCE_INLINE_TEMPLATE
-U32 ZSTD_insertAndFindFirstIndex (ZSTD_CCtx* zc, const BYTE* ip, U32 mls)
-{
-    U32* const hashTable  = zc->hashTable;
-    const U32 hashLog = zc->appliedParams.cParams.hashLog;
-    U32* const chainTable = zc->chainTable;
-    const U32 chainMask = (1 << zc->appliedParams.cParams.chainLog) - 1;
-    const BYTE* const base = zc->base;
-    const U32 target = (U32)(ip - base);
-    U32 idx = zc->nextToUpdate;
-
-    while(idx < target) { /* catch up */
-        size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);
-        NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
-        hashTable[h] = idx;
-        idx++;
-    }
-
-    zc->nextToUpdate = target;
-    return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
-}
-
-
-/* inlining is important to hardwire a hot branch (template emulation) */
-FORCE_INLINE_TEMPLATE
-size_t ZSTD_HcFindBestMatch_generic (
-                        ZSTD_CCtx* zc,   /* Index table will be updated */
-                        const BYTE* const ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 mls, const U32 extDict)
-{
-    U32* const chainTable = zc->chainTable;
-    const U32 chainSize = (1 << zc->appliedParams.cParams.chainLog);
-    const U32 chainMask = chainSize-1;
-    const BYTE* const base = zc->base;
-    const BYTE* const dictBase = zc->dictBase;
-    const U32 dictLimit = zc->dictLimit;
-    const BYTE* const prefixStart = base + dictLimit;
-    const BYTE* const dictEnd = dictBase + dictLimit;
-    const U32 lowLimit = zc->lowLimit;
-    const U32 current = (U32)(ip-base);
-    const U32 minChain = current > chainSize ? current - chainSize : 0;
-    int nbAttempts=maxNbAttempts;
-    size_t ml=4-1;
-
-    /* HC4 match finder */
-    U32 matchIndex = ZSTD_insertAndFindFirstIndex (zc, ip, mls);
-
-    for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) {
-        const BYTE* match;
-        size_t currentMl=0;
-        if ((!extDict) || matchIndex >= dictLimit) {
-            match = base + matchIndex;
-            if (match[ml] == ip[ml])   /* potentially better */
-                currentMl = ZSTD_count(ip, match, iLimit);
-        } else {
-            match = dictBase + matchIndex;
-            if (MEM_read32(match) == MEM_read32(ip))   /* assumption : matchIndex <= dictLimit-4 (by table construction) */
-                currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4;
-        }
-
-        /* save best solution */
-        if (currentMl > ml) {
-            ml = currentMl;
-            *offsetPtr = current - matchIndex + ZSTD_REP_MOVE;
-            if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
-        }
-
-        if (matchIndex <= minChain) break;
-        matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
-    }
-
-    return ml;
-}
-
-
-FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_selectMLS (
-                        ZSTD_CCtx* zc,
-                        const BYTE* ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 matchLengthSearch)
-{
-    switch(matchLengthSearch)
-    {
-    default : /* includes case 3 */
-    case 4 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 0);
-    case 5 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 0);
-    case 7 :
-    case 6 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 0);
-    }
-}
-
-
-FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_extDict_selectMLS (
-                        ZSTD_CCtx* zc,
-                        const BYTE* ip, const BYTE* const iLimit,
-                        size_t* offsetPtr,
-                        const U32 maxNbAttempts, const U32 matchLengthSearch)
-{
-    switch(matchLengthSearch)
-    {
-    default : /* includes case 3 */
-    case 4 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 1);
-    case 5 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 1);
-    case 7 :
-    case 6 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 1);
-    }
-}
-
-
-/* *******************************
-*  Common parser - lazy strategy
-*********************************/
-FORCE_INLINE_TEMPLATE
-void ZSTD_compressBlock_lazy_generic(ZSTD_CCtx* ctx,
-                                     const void* src, size_t srcSize,
-                                     const U32 searchMethod, const U32 depth)
-{
-    seqStore_t* seqStorePtr = &(ctx->seqStore);
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - 8;
-    const BYTE* const base = ctx->base + ctx->dictLimit;
-
-    U32 const maxSearches = 1 << ctx->appliedParams.cParams.searchLog;
-    U32 const mls = ctx->appliedParams.cParams.searchLength;
-
-    typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
-                        size_t* offsetPtr,
-                        U32 maxNbAttempts, U32 matchLengthSearch);
-    searchMax_f const searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS;
-    U32 offset_1 = seqStorePtr->rep[0], offset_2 = seqStorePtr->rep[1], savedOffset=0;
-
-    /* init */
-    ip += (ip==base);
-    ctx->nextToUpdate3 = ctx->nextToUpdate;
-    {   U32 const maxRep = (U32)(ip-base);
-        if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;
-        if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0;
-    }
-
-    /* Match Loop */
-    while (ip < ilimit) {
-        size_t matchLength=0;
-        size_t offset=0;
-        const BYTE* start=ip+1;
-
-        /* check repCode */
-        if ((offset_1>0) & (MEM_read32(ip+1) == MEM_read32(ip+1 - offset_1))) {
-            /* repcode : we take it */
-            matchLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
-            if (depth==0) goto _storeSequence;
-        }
-
-        /* first search (depth 0) */
-        {   size_t offsetFound = 99999999;
-            size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
-            if (ml2 > matchLength)
-                matchLength = ml2, start = ip, offset=offsetFound;
-        }
-
-        if (matchLength < 4) {
-            ip += ((ip-anchor) >> g_searchStrength) + 1;   /* jump faster over incompressible sections */
-            continue;
-        }
-
-        /* let's try to find a better solution */
-        if (depth>=1)
-        while (ip<ilimit) {
-            ip ++;
-            if ((offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {
-                size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;
-                int const gain2 = (int)(mlRep * 3);
-                int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
-                if ((mlRep >= 4) && (gain2 > gain1))
-                    matchLength = mlRep, offset = 0, start = ip;
-            }
-            {   size_t offset2=99999999;
-                size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-                int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
-                int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);
-                if ((ml2 >= 4) && (gain2 > gain1)) {
-                    matchLength = ml2, offset = offset2, start = ip;
-                    continue;   /* search a better one */
-            }   }
-
-            /* let's find an even better one */
-            if ((depth==2) && (ip<ilimit)) {
-                ip ++;
-                if ((offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {
-                    size_t const ml2 = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;
-                    int const gain2 = (int)(ml2 * 4);
-                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
-                    if ((ml2 >= 4) && (gain2 > gain1))
-                        matchLength = ml2, offset = 0, start = ip;
-                }
-                {   size_t offset2=99999999;
-                    size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-                    int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
-                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);
-                    if ((ml2 >= 4) && (gain2 > gain1)) {
-                        matchLength = ml2, offset = offset2, start = ip;
-                        continue;
-            }   }   }
-            break;  /* nothing found : store previous solution */
-        }
-
-        /* NOTE:
-         * start[-offset+ZSTD_REP_MOVE-1] is undefined behavior.
-         * (-offset+ZSTD_REP_MOVE-1) is unsigned, and is added to start, which
-         * overflows the pointer, which is undefined behavior.
-         */
-        /* catch up */
-        if (offset) {
-            while ( (start > anchor)
-                 && (start > base+offset-ZSTD_REP_MOVE)
-                 && (start[-1] == (start-offset+ZSTD_REP_MOVE)[-1]) )  /* only search for offset within prefix */
-                { start--; matchLength++; }
-            offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);
-        }
-        /* store sequence */
-_storeSequence:
-        {   size_t const litLength = start - anchor;
-            ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength-MINMATCH);
-            anchor = ip = start + matchLength;
-        }
-
-        /* check immediate repcode */
-        while ( (ip <= ilimit)
-             && ((offset_2>0)
-             & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
-            /* store sequence */
-            matchLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
-            offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap repcodes */
-            ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength-MINMATCH);
-            ip += matchLength;
-            anchor = ip;
-            continue;   /* faster when present ... (?) */
-    }   }
-
-    /* Save reps for next block */
-    seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : savedOffset;
-    seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : savedOffset;
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-static void ZSTD_compressBlock_btlazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 1, 2);
-}
-
-static void ZSTD_compressBlock_lazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 2);
-}
-
-static void ZSTD_compressBlock_lazy(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 1);
-}
-
-static void ZSTD_compressBlock_greedy(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 0);
-}
-
-
-FORCE_INLINE_TEMPLATE
-void ZSTD_compressBlock_lazy_extDict_generic(ZSTD_CCtx* ctx,
-                                     const void* src, size_t srcSize,
-                                     const U32 searchMethod, const U32 depth)
-{
-    seqStore_t* seqStorePtr = &(ctx->seqStore);
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - 8;
-    const BYTE* const base = ctx->base;
-    const U32 dictLimit = ctx->dictLimit;
-    const U32 lowestIndex = ctx->lowLimit;
-    const BYTE* const prefixStart = base + dictLimit;
-    const BYTE* const dictBase = ctx->dictBase;
-    const BYTE* const dictEnd  = dictBase + dictLimit;
-    const BYTE* const dictStart  = dictBase + ctx->lowLimit;
-
-    const U32 maxSearches = 1 << ctx->appliedParams.cParams.searchLog;
-    const U32 mls = ctx->appliedParams.cParams.searchLength;
-
-    typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
-                        size_t* offsetPtr,
-                        U32 maxNbAttempts, U32 matchLengthSearch);
-    searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS_extDict : ZSTD_HcFindBestMatch_extDict_selectMLS;
-
-    U32 offset_1 = seqStorePtr->rep[0], offset_2 = seqStorePtr->rep[1];
-
-    /* init */
-    ctx->nextToUpdate3 = ctx->nextToUpdate;
-    ip += (ip == prefixStart);
-
-    /* Match Loop */
-    while (ip < ilimit) {
-        size_t matchLength=0;
-        size_t offset=0;
-        const BYTE* start=ip+1;
-        U32 current = (U32)(ip-base);
-
-        /* check repCode */
-        {   const U32 repIndex = (U32)(current+1 - offset_1);
-            const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
-            const BYTE* const repMatch = repBase + repIndex;
-            if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))   /* intentional overflow */
-            if (MEM_read32(ip+1) == MEM_read32(repMatch)) {
-                /* repcode detected we should take it */
-                const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
-                matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repEnd, prefixStart) + 4;
-                if (depth==0) goto _storeSequence;
-        }   }
-
-        /* first search (depth 0) */
-        {   size_t offsetFound = 99999999;
-            size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
-            if (ml2 > matchLength)
-                matchLength = ml2, start = ip, offset=offsetFound;
-        }
-
-         if (matchLength < 4) {
-            ip += ((ip-anchor) >> g_searchStrength) + 1;   /* jump faster over incompressible sections */
-            continue;
-        }
-
-        /* let's try to find a better solution */
-        if (depth>=1)
-        while (ip<ilimit) {
-            ip ++;
-            current++;
-            /* check repCode */
-            if (offset) {
-                const U32 repIndex = (U32)(current - offset_1);
-                const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
-                const BYTE* const repMatch = repBase + repIndex;
-                if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))  /* intentional overflow */
-                if (MEM_read32(ip) == MEM_read32(repMatch)) {
-                    /* repcode detected */
-                    const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
-                    size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
-                    int const gain2 = (int)(repLength * 3);
-                    int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
-                    if ((repLength >= 4) && (gain2 > gain1))
-                        matchLength = repLength, offset = 0, start = ip;
-            }   }
-
-            /* search match, depth 1 */
-            {   size_t offset2=99999999;
-                size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-                int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
-                int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);
-                if ((ml2 >= 4) && (gain2 > gain1)) {
-                    matchLength = ml2, offset = offset2, start = ip;
-                    continue;   /* search a better one */
-            }   }
-
-            /* let's find an even better one */
-            if ((depth==2) && (ip<ilimit)) {
-                ip ++;
-                current++;
-                /* check repCode */
-                if (offset) {
-                    const U32 repIndex = (U32)(current - offset_1);
-                    const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
-                    const BYTE* const repMatch = repBase + repIndex;
-                    if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))  /* intentional overflow */
-                    if (MEM_read32(ip) == MEM_read32(repMatch)) {
-                        /* repcode detected */
-                        const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
-                        size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
-                        int const gain2 = (int)(repLength * 4);
-                        int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
-                        if ((repLength >= 4) && (gain2 > gain1))
-                            matchLength = repLength, offset = 0, start = ip;
-                }   }
-
-                /* search match, depth 2 */
-                {   size_t offset2=99999999;
-                    size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-                    int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
-                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);
-                    if ((ml2 >= 4) && (gain2 > gain1)) {
-                        matchLength = ml2, offset = offset2, start = ip;
-                        continue;
-            }   }   }
-            break;  /* nothing found : store previous solution */
-        }
-
-        /* catch up */
-        if (offset) {
-            U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));
-            const BYTE* match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;
-            const BYTE* const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;
-            while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; }  /* catch up */
-            offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);
-        }
-
-        /* store sequence */
-_storeSequence:
-        {   size_t const litLength = start - anchor;
-            ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength-MINMATCH);
-            anchor = ip = start + matchLength;
-        }
-
-        /* check immediate repcode */
-        while (ip <= ilimit) {
-            const U32 repIndex = (U32)((ip-base) - offset_2);
-            const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
-            const BYTE* const repMatch = repBase + repIndex;
-            if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))  /* intentional overflow */
-            if (MEM_read32(ip) == MEM_read32(repMatch)) {
-                /* repcode detected we should take it */
-                const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
-                matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
-                offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset;   /* swap offset history */
-                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength-MINMATCH);
-                ip += matchLength;
-                anchor = ip;
-                continue;   /* faster when present ... (?) */
-            }
-            break;
-    }   }
-
-    /* Save reps for next block */
-    seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-void ZSTD_compressBlock_greedy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 0);
-}
-
-static void ZSTD_compressBlock_lazy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 1);
-}
-
-static void ZSTD_compressBlock_lazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 2);
-}
-
-static void ZSTD_compressBlock_btlazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-    ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 1, 2);
-}
-
-
-/* The optimal parser */
-#include "zstd_opt.h"
-
-static void ZSTD_compressBlock_btopt(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-    ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 0);
-#else
-    (void)ctx; (void)src; (void)srcSize;
-    return;
-#endif
-}
-
-static void ZSTD_compressBlock_btultra(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-    ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 1);
-#else
-    (void)ctx; (void)src; (void)srcSize;
-    return;
-#endif
-}
-
-static void ZSTD_compressBlock_btopt_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-    ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 0);
-#else
-    (void)ctx; (void)src; (void)srcSize;
-    return;
-#endif
-}
-
-static void ZSTD_compressBlock_btultra_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
-{
-#ifdef ZSTD_OPT_H_91842398743
-    ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 1);
-#else
-    (void)ctx; (void)src; (void)srcSize;
-    return;
-#endif
-}
-
-
 /* ZSTD_selectBlockCompressor() :
+ * Not static, but internal use only (used by long distance matcher)
  * assumption : strat is a valid strategy */
-typedef void (*ZSTD_blockCompressor) (ZSTD_CCtx* ctx, const void* src, size_t srcSize);
-static ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict)
+typedef size_t (*ZSTD_blockCompressor) (ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict)
 {
     static const ZSTD_blockCompressor blockCompressor[2][(unsigned)ZSTD_btultra+1] = {
         { ZSTD_compressBlock_fast  /* default for 0 */,
@@ -2819,18 +1565,37 @@ static ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int
     return blockCompressor[extDict!=0][(U32)strat];
 }
 
+static void ZSTD_storeLastLiterals(seqStore_t* seqStorePtr,
+                                   const BYTE* anchor, size_t lastLLSize)
+{
+    memcpy(seqStorePtr->lit, anchor, lastLLSize);
+    seqStorePtr->lit += lastLLSize;
+}
 
 static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
 {
-    ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, zc->lowLimit < zc->dictLimit);
     const BYTE* const base = zc->base;
     const BYTE* const istart = (const BYTE*)src;
     const U32 current = (U32)(istart-base);
+    size_t lastLLSize;
+    const BYTE* anchor;
+    U32 const extDict = zc->lowLimit < zc->dictLimit;
+    const ZSTD_blockCompressor blockCompressor =
+        zc->appliedParams.ldmParams.enableLdm
+            ? (extDict ? ZSTD_compressBlock_ldm_extDict : ZSTD_compressBlock_ldm)
+            : ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, extDict);
+
     if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) return 0;   /* don't even attempt compression below a certain srcSize */
     ZSTD_resetSeqStore(&(zc->seqStore));
     if (current > zc->nextToUpdate + 384)
         zc->nextToUpdate = current - MIN(192, (U32)(current - zc->nextToUpdate - 384));   /* limited update after finding a very long match */
-    blockCompressor(zc, src, srcSize);
+
+    lastLLSize = blockCompressor(zc, src, srcSize);
+
+    /* Last literals */
+    anchor = (const BYTE*)src + srcSize - lastLLSize;
+    ZSTD_storeLastLiterals(&zc->seqStore, anchor, lastLLSize);
+
     return ZSTD_compressSequences(&zc->seqStore, zc->entropy, &zc->appliedParams.cParams, dst, dstCapacity, srcSize);
 }
 
@@ -2852,7 +1617,7 @@ static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,
     const BYTE* ip = (const BYTE*)src;
     BYTE* const ostart = (BYTE*)dst;
     BYTE* op = ostart;
-    U32 const maxDist = 1 << cctx->appliedParams.cParams.windowLog;
+    U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog;
 
     if (cctx->appliedParams.fParams.checksumFlag && srcSize)
         XXH64_update(&cctx->xxhState, src, srcSize);
@@ -2865,13 +1630,33 @@ static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,
             return ERROR(dstSize_tooSmall);   /* not enough space to store compressed block */
         if (remaining < blockSize) blockSize = remaining;
 
-        /* preemptive overflow correction */
+        /* preemptive overflow correction:
+         * 1. correction is large enough:
+         *    lowLimit > (3<<29) ==> current > 3<<29 + 1<<windowLog - blockSize
+         *    1<<windowLog <= newCurrent < 1<<chainLog + 1<<windowLog
+         *
+         *    current - newCurrent
+         *    > (3<<29 + 1<<windowLog - blockSize) - (1<<windowLog + 1<<chainLog)
+         *    > (3<<29 - blockSize) - (1<<chainLog)
+         *    > (3<<29 - blockSize) - (1<<30)             (NOTE: chainLog <= 30)
+         *    > 1<<29 - 1<<17
+         *
+         * 2. (ip+blockSize - cctx->base) doesn't overflow:
+         *    In 32 bit mode we limit windowLog to 30 so we don't get
+         *    differences larger than 1<<31-1.
+         * 3. cctx->lowLimit < 1<<32:
+         *    windowLog <= 31 ==> 3<<29 + 1<<windowLog < 7<<29 < 1<<32.
+         */
         if (cctx->lowLimit > (3U<<29)) {
-            U32 const cycleMask = (1 << ZSTD_cycleLog(cctx->appliedParams.cParams.hashLog, cctx->appliedParams.cParams.strategy)) - 1;
+            U32 const cycleMask = ((U32)1 << ZSTD_cycleLog(cctx->appliedParams.cParams.chainLog, cctx->appliedParams.cParams.strategy)) - 1;
             U32 const current = (U32)(ip - cctx->base);
-            U32 const newCurrent = (current & cycleMask) + (1 << cctx->appliedParams.cParams.windowLog);
+            U32 const newCurrent = (current & cycleMask) + ((U32)1 << cctx->appliedParams.cParams.windowLog);
             U32 const correction = current - newCurrent;
-            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_64 <= 30);
+            ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);
+            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);
+            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);
+            assert(current > newCurrent);
+            assert(correction > 1<<28); /* Loose bound, should be about 1<<29 */
             ZSTD_reduceIndex(cctx, correction);
             cctx->base += correction;
             cctx->dictBase += correction;
@@ -2879,6 +1664,7 @@ static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,
             cctx->dictLimit -= correction;
             if (cctx->nextToUpdate < correction) cctx->nextToUpdate = 0;
             else cctx->nextToUpdate -= correction;
+            DEBUGLOG(4, "Correction of 0x%x bytes to lowLimit=0x%x\n", correction, cctx->lowLimit);
         }
 
         if ((U32)(ip+blockSize - cctx->base) > cctx->loadedDictEnd + maxDist) {
@@ -2915,25 +1701,29 @@ static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,
 
 
 static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity,
-                                    ZSTD_parameters params, U64 pledgedSrcSize, U32 dictID)
+                                    ZSTD_CCtx_params params, U64 pledgedSrcSize, U32 dictID)
 {   BYTE* const op = (BYTE*)dst;
     U32   const dictIDSizeCodeLength = (dictID>0) + (dictID>=256) + (dictID>=65536);   /* 0-3 */
     U32   const dictIDSizeCode = params.fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength;   /* 0-3 */
     U32   const checksumFlag = params.fParams.checksumFlag>0;
-    U32   const windowSize = 1U << params.cParams.windowLog;
+    U32   const windowSize = (U32)1 << params.cParams.windowLog;
     U32   const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);
     BYTE  const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);
     U32   const fcsCode = params.fParams.contentSizeFlag ?
                      (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : 0;  /* 0-3 */
     BYTE  const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );
-    size_t pos;
+    size_t pos=0;
 
     if (dstCapacity < ZSTD_frameHeaderSize_max) return ERROR(dstSize_tooSmall);
-    DEBUGLOG(5, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u",
+    DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u",
                 !params.fParams.noDictIDFlag, dictID,  dictIDSizeCode);
 
-    MEM_writeLE32(dst, ZSTD_MAGICNUMBER);
-    op[4] = frameHeaderDecriptionByte; pos=5;
+    if (params.format == ZSTD_f_zstd1) {
+        DEBUGLOG(4, "writing zstd magic number");
+        MEM_writeLE32(dst, ZSTD_MAGICNUMBER);
+        pos = 4;
+    }
+    op[pos++] = frameHeaderDecriptionByte;
     if (!singleSegment) op[pos++] = windowLogByte;
     switch(dictIDSizeCode)
     {
@@ -2969,7 +1759,7 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,
 
     if (frame && (cctx->stage==ZSTDcs_init)) {
         fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams,
-                                cctx->pledgedSrcSizePlusOne-1, cctx->dictID);
+                                       cctx->pledgedSrcSizePlusOne-1, cctx->dictID);
         if (ZSTD_isError(fhSize)) return fhSize;
         dstCapacity -= fhSize;
         dst = (char*)dst + fhSize;
@@ -3018,11 +1808,9 @@ size_t ZSTD_compressContinue (ZSTD_CCtx* cctx,
 
 size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx)
 {
-    U32 const cLevel = cctx->compressionLevel;
-    ZSTD_compressionParameters cParams = (cLevel == ZSTD_CLEVEL_CUSTOM) ?
-                                        cctx->appliedParams.cParams :
-                                        ZSTD_getCParams(cLevel, 0, 0);
-    return MIN (ZSTD_BLOCKSIZE_MAX, 1 << cParams.windowLog);
+    ZSTD_compressionParameters const cParams =
+            ZSTD_getCParamsFromCCtxParams(cctx->appliedParams, 0, 0);
+    return MIN (ZSTD_BLOCKSIZE_MAX, (U32)1 << cParams.windowLog);
 }
 
 size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
@@ -3046,7 +1834,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t
     zc->dictBase = zc->base;
     zc->base += ip - zc->nextSrc;
     zc->nextToUpdate = zc->dictLimit;
-    zc->loadedDictEnd = zc->forceWindow ? 0 : (U32)(iend - zc->base);
+    zc->loadedDictEnd = zc->appliedParams.forceWindow ? 0 : (U32)(iend - zc->base);
 
     zc->nextSrc = iend;
     if (srcSize <= HASH_READ_SIZE) return 0;
@@ -3056,7 +1844,6 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t
     case ZSTD_fast:
         ZSTD_fillHashTable (zc, iend, zc->appliedParams.cParams.searchLength);
         break;
-
     case ZSTD_dfast:
         ZSTD_fillDoubleHashTable (zc, iend, zc->appliedParams.cParams.searchLength);
         break;
@@ -3072,7 +1859,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t
     case ZSTD_btopt:
     case ZSTD_btultra:
         if (srcSize >= HASH_READ_SIZE)
-            ZSTD_updateTree(zc, iend-HASH_READ_SIZE, iend, 1 << zc->appliedParams.cParams.searchLog, zc->appliedParams.cParams.searchLength);
+            ZSTD_updateTree(zc, iend-HASH_READ_SIZE, iend, (U32)1 << zc->appliedParams.cParams.searchLog, zc->appliedParams.cParams.searchLength);
         break;
 
     default:
@@ -3120,8 +1907,10 @@ static size_t ZSTD_loadZstdDictionary(ZSTD_CCtx* cctx, const void* dict, size_t
     cctx->dictID = cctx->appliedParams.fParams.noDictIDFlag ? 0 :  MEM_readLE32(dictPtr);
     dictPtr += 4;
 
-    {   size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)cctx->entropy->hufCTable, 255, dictPtr, dictEnd-dictPtr);
+    {   unsigned maxSymbolValue = 255;
+        size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)cctx->entropy->hufCTable, &maxSymbolValue, dictPtr, dictEnd-dictPtr);
         if (HUF_isError(hufHeaderSize)) return ERROR(dictionary_corrupted);
+        if (maxSymbolValue < 255) return ERROR(dictionary_corrupted);
         dictPtr += hufHeaderSize;
     }
 
@@ -3221,12 +2010,10 @@ static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx,
                              const void* dict, size_t dictSize,
                              ZSTD_dictMode_e dictMode,
                              const ZSTD_CDict* cdict,
-                                   ZSTD_parameters params, U64 pledgedSrcSize,
+                                   ZSTD_CCtx_params params, U64 pledgedSrcSize,
                                    ZSTD_buffered_policy_e zbuff)
 {
     DEBUGLOG(4, "ZSTD_compressBegin_internal");
-    DEBUGLOG(4, "dict ? %s", dict ? "dict" : (cdict ? "cdict" : "none"));
-    DEBUGLOG(4, "dictMode : %u", (U32)dictMode);
     /* params are supposed to be fully validated at this point */
     assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
     assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
@@ -3242,6 +2029,19 @@ static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx,
     return ZSTD_compress_insertDictionary(cctx, dict, dictSize, dictMode);
 }
 
+size_t ZSTD_compressBegin_advanced_internal(
+                                    ZSTD_CCtx* cctx,
+                                    const void* dict, size_t dictSize,
+                                    ZSTD_dictMode_e dictMode,
+                                    ZSTD_CCtx_params params,
+                                    unsigned long long pledgedSrcSize)
+{
+    /* compression parameters verification and optimization */
+    CHECK_F( ZSTD_checkCParams(params.cParams) );
+    return ZSTD_compressBegin_internal(cctx, dict, dictSize, dictMode, NULL,
+                                       params, pledgedSrcSize,
+                                       ZSTDb_not_buffered);
+}
 
 /*! ZSTD_compressBegin_advanced() :
 *   @return : 0, or an error code */
@@ -3249,21 +2049,22 @@ size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx,
                              const void* dict, size_t dictSize,
                                    ZSTD_parameters params, unsigned long long pledgedSrcSize)
 {
-    /* compression parameters verification and optimization */
-    CHECK_F(ZSTD_checkCParams(params.cParams));
-    return ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dm_auto, NULL,
-                                    params, pledgedSrcSize, ZSTDb_not_buffered);
+    ZSTD_CCtx_params const cctxParams =
+            ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);
+    return ZSTD_compressBegin_advanced_internal(cctx, dict, dictSize, ZSTD_dm_auto,
+                                                cctxParams,
+                                                pledgedSrcSize);
 }
 
-
 size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)
 {
     ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);
+    ZSTD_CCtx_params const cctxParams =
+            ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);
     return ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dm_auto, NULL,
-                                       params, 0, ZSTDb_not_buffered);
+                                       cctxParams, 0, ZSTDb_not_buffered);
 }
 
-
 size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel)
 {
     return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel);
@@ -3324,9 +2125,9 @@ size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,
     endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize);
     if (ZSTD_isError(endResult)) return endResult;
     if (cctx->appliedParams.fParams.contentSizeFlag) {  /* control src size */
-        DEBUGLOG(5, "end of frame : controlling src size");
+        DEBUGLOG(4, "end of frame : controlling src size");
         if (cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1) {
-            DEBUGLOG(5, "error : pledgedSrcSize = %u, while realSrcSize = %u",
+            DEBUGLOG(4, "error : pledgedSrcSize = %u, while realSrcSize = %u",
                 (U32)cctx->pledgedSrcSizePlusOne-1, (U32)cctx->consumedSrcSize);
             return ERROR(srcSize_wrong);
     }   }
@@ -3340,9 +2141,13 @@ static size_t ZSTD_compress_internal (ZSTD_CCtx* cctx,
                          const void* dict,size_t dictSize,
                                ZSTD_parameters params)
 {
-    CHECK_F( ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dm_auto, NULL,
-                                         params, srcSize, ZSTDb_not_buffered) );
-    return ZSTD_compressEnd(cctx, dst,  dstCapacity, src, srcSize);
+    ZSTD_CCtx_params const cctxParams =
+            ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);
+    return ZSTD_compress_advanced_internal(cctx,
+                                          dst, dstCapacity,
+                                          src, srcSize,
+                                          dict, dictSize,
+                                          cctxParams);
 }
 
 size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx,
@@ -3355,6 +2160,19 @@ size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx,
     return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params);
 }
 
+/* Internal */
+size_t ZSTD_compress_advanced_internal(
+        ZSTD_CCtx* cctx,
+        void* dst, size_t dstCapacity,
+        const void* src, size_t srcSize,
+        const void* dict,size_t dictSize,
+        ZSTD_CCtx_params params)
+{
+    CHECK_F( ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dm_auto, NULL,
+                                         params, srcSize, ZSTDb_not_buffered) );
+    return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
+}
+
 size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize,
                                const void* dict, size_t dictSize, int compressionLevel)
 {
@@ -3384,18 +2202,21 @@ size_t ZSTD_compress(void* dst, size_t dstCapacity, const void* src, size_t srcS
 
 /*! ZSTD_estimateCDictSize_advanced() :
  *  Estimate amount of memory that will be needed to create a dictionary with following arguments */
-size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, unsigned byReference)
+size_t ZSTD_estimateCDictSize_advanced(
+        size_t dictSize, ZSTD_compressionParameters cParams,
+        ZSTD_dictLoadMethod_e dictLoadMethod)
 {
     DEBUGLOG(5, "sizeof(ZSTD_CDict) : %u", (U32)sizeof(ZSTD_CDict));
-    DEBUGLOG(5, "CCtx estimate : %u", (U32)ZSTD_estimateCCtxSize_advanced(cParams));
-    return sizeof(ZSTD_CDict) + ZSTD_estimateCCtxSize_advanced(cParams)
-           + (byReference ? 0 : dictSize);
+    DEBUGLOG(5, "CCtx estimate : %u",
+             (U32)ZSTD_estimateCCtxSize_usingCParams(cParams));
+    return sizeof(ZSTD_CDict) + ZSTD_estimateCCtxSize_usingCParams(cParams)
+           + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
 }
 
 size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel)
 {
     ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, dictSize);
-    return ZSTD_estimateCDictSize_advanced(dictSize, cParams, 0);
+    return ZSTD_estimateCDictSize_advanced(dictSize, cParams, ZSTD_dlm_byCopy);
 }
 
 size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict)
@@ -3406,22 +2227,15 @@ size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict)
     return ZSTD_sizeof_CCtx(cdict->refContext) + (cdict->dictBuffer ? cdict->dictContentSize : 0) + sizeof(*cdict);
 }
 
-static ZSTD_parameters ZSTD_makeParams(ZSTD_compressionParameters cParams, ZSTD_frameParameters fParams)
-{
-    ZSTD_parameters params;
-    params.cParams = cParams;
-    params.fParams = fParams;
-    return params;
-}
-
 static size_t ZSTD_initCDict_internal(
                     ZSTD_CDict* cdict,
               const void* dictBuffer, size_t dictSize,
-                    unsigned byReference, ZSTD_dictMode_e dictMode,
+                    ZSTD_dictLoadMethod_e dictLoadMethod,
+                    ZSTD_dictMode_e dictMode,
                     ZSTD_compressionParameters cParams)
 {
     DEBUGLOG(5, "ZSTD_initCDict_internal, mode %u", (U32)dictMode);
-    if ((byReference) || (!dictBuffer) || (!dictSize)) {
+    if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dictBuffer) || (!dictSize)) {
         cdict->dictBuffer = NULL;
         cdict->dictContent = dictBuffer;
     } else {
@@ -3433,13 +2247,12 @@ static size_t ZSTD_initCDict_internal(
     }
     cdict->dictContentSize = dictSize;
 
-    {   ZSTD_frameParameters const fParams = { 0 /* contentSizeFlag */,
-                    0 /* checksumFlag */, 0 /* noDictIDFlag */ };  /* dummy */
-        ZSTD_parameters const params = ZSTD_makeParams(cParams, fParams);
+    {   ZSTD_CCtx_params cctxParams = cdict->refContext->requestedParams;
+        cctxParams.cParams = cParams;
         CHECK_F( ZSTD_compressBegin_internal(cdict->refContext,
                                         cdict->dictContent, dictSize, dictMode,
                                         NULL,
-                                        params, ZSTD_CONTENTSIZE_UNKNOWN,
+                                        cctxParams, ZSTD_CONTENTSIZE_UNKNOWN,
                                         ZSTDb_not_buffered) );
     }
 
@@ -3447,7 +2260,8 @@ static size_t ZSTD_initCDict_internal(
 }
 
 ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize,
-                                      unsigned byReference, ZSTD_dictMode_e dictMode,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_dictMode_e dictMode,
                                       ZSTD_compressionParameters cParams, ZSTD_customMem customMem)
 {
     DEBUGLOG(5, "ZSTD_createCDict_advanced, mode %u", (U32)dictMode);
@@ -3462,10 +2276,9 @@ ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize,
             return NULL;
         }
         cdict->refContext = cctx;
-
         if (ZSTD_isError( ZSTD_initCDict_internal(cdict,
                                         dictBuffer, dictSize,
-                                        byReference, dictMode,
+                                        dictLoadMethod, dictMode,
                                         cParams) )) {
             ZSTD_freeCDict(cdict);
             return NULL;
@@ -3479,7 +2292,7 @@ ZSTD_CDict* ZSTD_createCDict(const void* dict, size_t dictSize, int compressionL
 {
     ZSTD_compressionParameters cParams = ZSTD_getCParams(compressionLevel, 0, dictSize);
     return ZSTD_createCDict_advanced(dict, dictSize,
-                                     0 /* byReference */, ZSTD_dm_auto,
+                                     ZSTD_dlm_byCopy, ZSTD_dm_auto,
                                      cParams, ZSTD_defaultCMem);
 }
 
@@ -3487,7 +2300,7 @@ ZSTD_CDict* ZSTD_createCDict_byReference(const void* dict, size_t dictSize, int
 {
     ZSTD_compressionParameters cParams = ZSTD_getCParams(compressionLevel, 0, dictSize);
     return ZSTD_createCDict_advanced(dict, dictSize,
-                                     1 /* byReference */, ZSTD_dm_auto,
+                                     ZSTD_dlm_byRef, ZSTD_dm_auto,
                                      cParams, ZSTD_defaultCMem);
 }
 
@@ -3517,11 +2330,12 @@ size_t ZSTD_freeCDict(ZSTD_CDict* cdict)
  */
 ZSTD_CDict* ZSTD_initStaticCDict(void* workspace, size_t workspaceSize,
                            const void* dict, size_t dictSize,
-                                 unsigned byReference, ZSTD_dictMode_e dictMode,
+                                 ZSTD_dictLoadMethod_e dictLoadMethod,
+                                 ZSTD_dictMode_e dictMode,
                                  ZSTD_compressionParameters cParams)
 {
-    size_t const cctxSize = ZSTD_estimateCCtxSize_advanced(cParams);
-    size_t const neededSize = sizeof(ZSTD_CDict) + (byReference ? 0 : dictSize)
+    size_t const cctxSize = ZSTD_estimateCCtxSize_usingCParams(cParams);
+    size_t const neededSize = sizeof(ZSTD_CDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize)
                             + cctxSize;
     ZSTD_CDict* const cdict = (ZSTD_CDict*) workspace;
     void* ptr;
@@ -3531,7 +2345,7 @@ ZSTD_CDict* ZSTD_initStaticCDict(void* workspace, size_t workspaceSize,
         (U32)workspaceSize, (U32)neededSize, (U32)(workspaceSize < neededSize));
     if (workspaceSize < neededSize) return NULL;
 
-    if (!byReference) {
+    if (dictLoadMethod == ZSTD_dlm_byCopy) {
         memcpy(cdict+1, dict, dictSize);
         dict = cdict+1;
         ptr = (char*)workspace + sizeof(ZSTD_CDict) + dictSize;
@@ -3542,15 +2356,15 @@ ZSTD_CDict* ZSTD_initStaticCDict(void* workspace, size_t workspaceSize,
 
     if (ZSTD_isError( ZSTD_initCDict_internal(cdict,
                                               dict, dictSize,
-                                              1 /* byReference */, dictMode,
+                                              ZSTD_dlm_byRef, dictMode,
                                               cParams) ))
         return NULL;
 
     return cdict;
 }
 
-ZSTD_parameters ZSTD_getParamsFromCDict(const ZSTD_CDict* cdict) {
-    return ZSTD_getParamsFromCCtx(cdict->refContext);
+ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict) {
+    return cdict->refContext->appliedParams.cParams;
 }
 
 /* ZSTD_compressBegin_usingCDict_advanced() :
@@ -3560,7 +2374,8 @@ size_t ZSTD_compressBegin_usingCDict_advanced(
     ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize)
 {
     if (cdict==NULL) return ERROR(dictionary_wrong);
-    {   ZSTD_parameters params = cdict->refContext->appliedParams;
+    {   ZSTD_CCtx_params params = cctx->requestedParams;
+        params.cParams = ZSTD_getCParamsFromCDict(cdict);
         params.fParams = fParams;
         DEBUGLOG(5, "ZSTD_compressBegin_usingCDict_advanced");
         return ZSTD_compressBegin_internal(cctx,
@@ -3644,7 +2459,7 @@ size_t ZSTD_CStreamOutSize(void)
 static size_t ZSTD_resetCStream_internal(ZSTD_CStream* zcs,
                     const void* dict, size_t dictSize, ZSTD_dictMode_e dictMode,
                     const ZSTD_CDict* cdict,
-                    ZSTD_parameters params, unsigned long long pledgedSrcSize)
+                    const ZSTD_CCtx_params params, unsigned long long pledgedSrcSize)
 {
     DEBUGLOG(4, "ZSTD_resetCStream_internal");
     /* params are supposed to be fully validated at this point */
@@ -3668,13 +2483,11 @@ static size_t ZSTD_resetCStream_internal(ZSTD_CStream* zcs,
 
 size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
 {
-    ZSTD_parameters params = zcs->requestedParams;
+    ZSTD_CCtx_params params = zcs->requestedParams;
     params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-    DEBUGLOG(5, "ZSTD_resetCStream");
-    if (zcs->compressionLevel != ZSTD_CLEVEL_CUSTOM) {
-        params.cParams = ZSTD_getCParams(zcs->compressionLevel, pledgedSrcSize, 0 /* dictSize */);
-    }
-    return ZSTD_resetCStream_internal(zcs, NULL, 0, zcs->dictMode, zcs->cdict, params, pledgedSrcSize);
+    params.cParams = ZSTD_getCParamsFromCCtxParams(params, pledgedSrcSize, 0);
+    DEBUGLOG(4, "ZSTD_resetCStream");
+    return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dm_auto, zcs->cdict, params, pledgedSrcSize);
 }
 
 /*! ZSTD_initCStream_internal() :
@@ -3683,9 +2496,9 @@ size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
  *  Assumption 2 : either dict, or cdict, is defined, not both */
 size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
                     const void* dict, size_t dictSize, const ZSTD_CDict* cdict,
-                    ZSTD_parameters params, unsigned long long pledgedSrcSize)
+                    ZSTD_CCtx_params params, unsigned long long pledgedSrcSize)
 {
-    DEBUGLOG(5, "ZSTD_initCStream_internal");
+    DEBUGLOG(4, "ZSTD_initCStream_internal");
     assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
     assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
 
@@ -3697,23 +2510,23 @@ size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
         }
         ZSTD_freeCDict(zcs->cdictLocal);
         zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,
-                                            zcs->dictContentByRef, zcs->dictMode,
+                                            ZSTD_dlm_byCopy, ZSTD_dm_auto,
                                             params.cParams, zcs->customMem);
         zcs->cdict = zcs->cdictLocal;
         if (zcs->cdictLocal == NULL) return ERROR(memory_allocation);
     } else {
         if (cdict) {
-            ZSTD_parameters const cdictParams = ZSTD_getParamsFromCDict(cdict);
-            params.cParams = cdictParams.cParams;  /* cParams are enforced from cdict */
+            params.cParams = ZSTD_getCParamsFromCDict(cdict);  /* cParams are enforced from cdict */
         }
         ZSTD_freeCDict(zcs->cdictLocal);
         zcs->cdictLocal = NULL;
         zcs->cdict = cdict;
     }
 
+    params.compressionLevel = ZSTD_CLEVEL_CUSTOM;
     zcs->requestedParams = params;
-    zcs->compressionLevel = ZSTD_CLEVEL_CUSTOM;
-    return ZSTD_resetCStream_internal(zcs, NULL, 0, zcs->dictMode, zcs->cdict, params, pledgedSrcSize);
+
+    return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dm_auto, zcs->cdict, params, pledgedSrcSize);
 }
 
 /* ZSTD_initCStream_usingCDict_advanced() :
@@ -3724,7 +2537,8 @@ size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
                                             unsigned long long pledgedSrcSize)
 {   /* cannot handle NULL cdict (does not know what to do) */
     if (!cdict) return ERROR(dictionary_wrong);
-    {   ZSTD_parameters params = ZSTD_getParamsFromCDict(cdict);
+    {   ZSTD_CCtx_params params = zcs->requestedParams;
+        params.cParams = ZSTD_getCParamsFromCDict(cdict);
         params.fParams = fParams;
         return ZSTD_initCStream_internal(zcs,
                                 NULL, 0, cdict,
@@ -3743,30 +2557,32 @@ size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
                                  const void* dict, size_t dictSize,
                                  ZSTD_parameters params, unsigned long long pledgedSrcSize)
 {
+    ZSTD_CCtx_params const cctxParams =
+            ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
     CHECK_F( ZSTD_checkCParams(params.cParams) );
-    zcs->requestedParams = params;
-    zcs->compressionLevel = ZSTD_CLEVEL_CUSTOM;
-    return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, params, pledgedSrcSize);
+    return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, cctxParams, pledgedSrcSize);
 }
 
 size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)
 {
     ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);
-    zcs->compressionLevel = compressionLevel;
-    return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, params, 0);
+    ZSTD_CCtx_params const cctxParams =
+            ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
+    return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, cctxParams, 0);
 }
 
 size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize)
 {
-    ZSTD_parameters params = ZSTD_getParams(compressionLevel, pledgedSrcSize, 0);
-    params.fParams.contentSizeFlag = (pledgedSrcSize>0);
-    return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, params, pledgedSrcSize);
+    ZSTD_CCtx_params cctxParams;
+    ZSTD_parameters const params = ZSTD_getParams(compressionLevel, pledgedSrcSize, 0);
+    cctxParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
+    cctxParams.fParams.contentSizeFlag = (pledgedSrcSize>0);
+    return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, cctxParams, pledgedSrcSize);
 }
 
 size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
 {
-    ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, 0);
-    return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, params, 0);
+    return ZSTD_initCStream_srcSize(zcs, compressionLevel, 0);
 }
 
 /*======   Compression   ======*/
@@ -3781,6 +2597,7 @@ MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity,
 
 /** ZSTD_compressStream_generic():
  *  internal function for all *compressStream*() variants and *compress_generic()
+ *  non-static, because can be called from zstdmt.c
  * @return : hint size for next input */
 size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
                                    ZSTD_outBuffer* output,
@@ -3934,21 +2751,13 @@ size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuf
     return ZSTD_compressStream_generic(zcs, output, input, ZSTD_e_continue);
 }
 
-/*! ZSTDMT_initCStream_internal() :
- *  Private use only. Init streaming operation.
- *  expects params to be valid.
- *  must receive dict, or cdict, or none, but not both.
- *  @return : 0, or an error code */
-size_t ZSTDMT_initCStream_internal(ZSTDMT_CCtx* zcs,
-                    const void* dict, size_t dictSize, const ZSTD_CDict* cdict,
-                    ZSTD_parameters params, unsigned long long pledgedSrcSize);
-
 
 size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,
                               ZSTD_outBuffer* output,
                               ZSTD_inBuffer* input,
                               ZSTD_EndDirective endOp)
 {
+    DEBUGLOG(5, "ZSTD_compress_generic");
     /* check conditions */
     if (output->pos > output->size) return ERROR(GENERIC);
     if (input->pos  > input->size)  return ERROR(GENERIC);
@@ -3956,31 +2765,42 @@ size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,
 
     /* transparent initialization stage */
     if (cctx->streamStage == zcss_init) {
-        const void* const prefix = cctx->prefix;
-        size_t const prefixSize = cctx->prefixSize;
-        ZSTD_parameters params = cctx->requestedParams;
-        if (cctx->compressionLevel != ZSTD_CLEVEL_CUSTOM)
-            params.cParams = ZSTD_getCParams(cctx->compressionLevel,
-                                cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/);
-        cctx->prefix = NULL; cctx->prefixSize = 0;   /* single usage */
-        assert(prefix==NULL || cctx->cdict==NULL);   /* only one can be set */
+        ZSTD_prefixDict const prefixDict = cctx->prefixDict;
+        ZSTD_CCtx_params params = cctx->requestedParams;
+        params.cParams = ZSTD_getCParamsFromCCtxParams(
+                cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/);
+        memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));  /* single usage */
+        assert(prefixDict.dict==NULL || cctx->cdict==NULL);   /* only one can be set */
+        DEBUGLOG(4, "ZSTD_compress_generic : transparent init stage");
 
 #ifdef ZSTD_MULTITHREAD
-        if (cctx->nbThreads > 1) {
-            DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbThreads=%u", cctx->nbThreads);
-            CHECK_F( ZSTDMT_initCStream_internal(cctx->mtctx, prefix, prefixSize, cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) );
+        if (params.nbThreads > 1) {
+            if (cctx->mtctx == NULL || cctx->appliedParams.nbThreads != params.nbThreads) {
+                ZSTDMT_freeCCtx(cctx->mtctx);
+                cctx->mtctx = ZSTDMT_createCCtx_advanced(params.nbThreads, cctx->customMem);
+                if (cctx->mtctx == NULL) return ERROR(memory_allocation);
+            }
+            DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbThreads=%u", params.nbThreads);
+            CHECK_F( ZSTDMT_initCStream_internal(
+                             cctx->mtctx,
+                             prefixDict.dict, prefixDict.dictSize, ZSTD_dm_rawContent,
+                             cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) );
             cctx->streamStage = zcss_load;
+            cctx->appliedParams.nbThreads = params.nbThreads;
         } else
 #endif
         {
-            CHECK_F( ZSTD_resetCStream_internal(cctx, prefix, prefixSize, cctx->dictMode, cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) );
+            CHECK_F( ZSTD_resetCStream_internal(
+                             cctx, prefixDict.dict, prefixDict.dictSize,
+                             prefixDict.dictMode, cctx->cdict, params,
+                             cctx->pledgedSrcSizePlusOne-1) );
     }   }
 
     /* compression stage */
 #ifdef ZSTD_MULTITHREAD
-    if (cctx->nbThreads > 1) {
+    if (cctx->appliedParams.nbThreads > 1) {
         size_t const flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp);
-        DEBUGLOG(5, "ZSTDMT_compressStream_generic : %u", (U32)flushMin);
+        DEBUGLOG(5, "ZSTDMT_compressStream_generic result : %u", (U32)flushMin);
         if ( ZSTD_isError(flushMin)
           || (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */
             ZSTD_startNewCompression(cctx);
@@ -3988,7 +2808,6 @@ size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,
         return flushMin;
     }
 #endif
-
     CHECK_F( ZSTD_compressStream_generic(cctx, output, input, endOp) );
     DEBUGLOG(5, "completed ZSTD_compress_generic");
     return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */
@@ -4189,6 +3008,7 @@ ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long l
     if (compressionLevel > ZSTD_MAX_CLEVEL) compressionLevel = ZSTD_MAX_CLEVEL;
     { ZSTD_compressionParameters const cp = ZSTD_defaultCParameters[tableID][compressionLevel];
       return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize); }
+
 }
 
 /*! ZSTD_getParams() :
diff --git a/thirdparty/zstd/compress/zstd_compress.h b/thirdparty/zstd/compress/zstd_compress.h
new file mode 100644
index 0000000000..94606edc93
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_compress.h
@@ -0,0 +1,307 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+#ifndef ZSTD_COMPRESS_H
+#define ZSTD_COMPRESS_H
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#include "zstd_internal.h"
+#ifdef ZSTD_MULTITHREAD
+#  include "zstdmt_compress.h"
+#endif
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+/*-*************************************
+*  Constants
+***************************************/
+static const U32 g_searchStrength = 8;
+#define HASH_READ_SIZE 8
+
+
+/*-*************************************
+*  Context memory management
+***************************************/
+typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;
+typedef enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage;
+
+typedef struct ZSTD_prefixDict_s {
+    const void* dict;
+    size_t dictSize;
+    ZSTD_dictMode_e dictMode;
+} ZSTD_prefixDict;
+
+struct ZSTD_CCtx_s {
+    const BYTE* nextSrc;    /* next block here to continue on current prefix */
+    const BYTE* base;       /* All regular indexes relative to this position */
+    const BYTE* dictBase;   /* extDict indexes relative to this position */
+    U32   dictLimit;        /* below that point, need extDict */
+    U32   lowLimit;         /* below that point, no more data */
+    U32   nextToUpdate;     /* index from which to continue dictionary update */
+    U32   nextToUpdate3;    /* index from which to continue dictionary update */
+    U32   hashLog3;         /* dispatch table : larger == faster, more memory */
+    U32   loadedDictEnd;    /* index of end of dictionary */
+    ZSTD_compressionStage_e stage;
+    U32   dictID;
+    ZSTD_CCtx_params requestedParams;
+    ZSTD_CCtx_params appliedParams;
+    void* workSpace;
+    size_t workSpaceSize;
+    size_t blockSize;
+    U64 pledgedSrcSizePlusOne;  /* this way, 0 (default) == unknown */
+    U64 consumedSrcSize;
+    XXH64_state_t xxhState;
+    ZSTD_customMem customMem;
+    size_t staticSize;
+
+    seqStore_t seqStore;    /* sequences storage ptrs */
+    optState_t optState;
+    ldmState_t ldmState;    /* long distance matching state */
+    U32* hashTable;
+    U32* hashTable3;
+    U32* chainTable;
+    ZSTD_entropyCTables_t* entropy;
+
+    /* streaming */
+    char*  inBuff;
+    size_t inBuffSize;
+    size_t inToCompress;
+    size_t inBuffPos;
+    size_t inBuffTarget;
+    char*  outBuff;
+    size_t outBuffSize;
+    size_t outBuffContentSize;
+    size_t outBuffFlushedSize;
+    ZSTD_cStreamStage streamStage;
+    U32    frameEnded;
+
+    /* Dictionary */
+    ZSTD_CDict* cdictLocal;
+    const ZSTD_CDict* cdict;
+    ZSTD_prefixDict prefixDict;   /* single-usage dictionary */
+
+    /* Multi-threading */
+#ifdef ZSTD_MULTITHREAD
+    ZSTDMT_CCtx* mtctx;
+#endif
+};
+
+
+static const BYTE LL_Code[64] = {  0,  1,  2,  3,  4,  5,  6,  7,
+                                   8,  9, 10, 11, 12, 13, 14, 15,
+                                  16, 16, 17, 17, 18, 18, 19, 19,
+                                  20, 20, 20, 20, 21, 21, 21, 21,
+                                  22, 22, 22, 22, 22, 22, 22, 22,
+                                  23, 23, 23, 23, 23, 23, 23, 23,
+                                  24, 24, 24, 24, 24, 24, 24, 24,
+                                  24, 24, 24, 24, 24, 24, 24, 24 };
+
+static const BYTE ML_Code[128] = { 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15,
+                                  16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,
+                                  32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37,
+                                  38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39,
+                                  40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,
+                                  41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41,
+                                  42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42,
+                                  42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 };
+
+/*! ZSTD_storeSeq() :
+    Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.
+    `offsetCode` : distance to match, or 0 == repCode.
+    `matchCode` : matchLength - MINMATCH
+*/
+MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t matchCode)
+{
+#if defined(ZSTD_DEBUG) && (ZSTD_DEBUG >= 6)
+    static const BYTE* g_start = NULL;
+    U32 const pos = (U32)((const BYTE*)literals - g_start);
+    if (g_start==NULL) g_start = (const BYTE*)literals;
+    if ((pos > 0) && (pos < 1000000000))
+        DEBUGLOG(6, "Cpos %6u :%5u literals & match %3u bytes at distance %6u",
+               pos, (U32)litLength, (U32)matchCode+MINMATCH, (U32)offsetCode);
+#endif
+    /* copy Literals */
+    assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + 128 KB);
+    ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);
+    seqStorePtr->lit += litLength;
+
+    /* literal Length */
+    if (litLength>0xFFFF) {
+        seqStorePtr->longLengthID = 1;
+        seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
+    }
+    seqStorePtr->sequences[0].litLength = (U16)litLength;
+
+    /* match offset */
+    seqStorePtr->sequences[0].offset = offsetCode + 1;
+
+    /* match Length */
+    if (matchCode>0xFFFF) {
+        seqStorePtr->longLengthID = 2;
+        seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
+    }
+    seqStorePtr->sequences[0].matchLength = (U16)matchCode;
+
+    seqStorePtr->sequences++;
+}
+
+
+/*-*************************************
+*  Match length counter
+***************************************/
+static unsigned ZSTD_NbCommonBytes (register size_t val)
+{
+    if (MEM_isLittleEndian()) {
+        if (MEM_64bits()) {
+#       if defined(_MSC_VER) && defined(_WIN64)
+            unsigned long r = 0;
+            _BitScanForward64( &r, (U64)val );
+            return (unsigned)(r>>3);
+#       elif defined(__GNUC__) && (__GNUC__ >= 4)
+            return (__builtin_ctzll((U64)val) >> 3);
+#       else
+            static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2,
+                                                     0, 3, 1, 3, 1, 4, 2, 7,
+                                                     0, 2, 3, 6, 1, 5, 3, 5,
+                                                     1, 3, 4, 4, 2, 5, 6, 7,
+                                                     7, 0, 1, 2, 3, 3, 4, 6,
+                                                     2, 6, 5, 5, 3, 4, 5, 6,
+                                                     7, 1, 2, 4, 6, 4, 4, 5,
+                                                     7, 2, 6, 5, 7, 6, 7, 7 };
+            return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];
+#       endif
+        } else { /* 32 bits */
+#       if defined(_MSC_VER)
+            unsigned long r=0;
+            _BitScanForward( &r, (U32)val );
+            return (unsigned)(r>>3);
+#       elif defined(__GNUC__) && (__GNUC__ >= 3)
+            return (__builtin_ctz((U32)val) >> 3);
+#       else
+            static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0,
+                                                     3, 2, 2, 1, 3, 2, 0, 1,
+                                                     3, 3, 1, 2, 2, 2, 2, 0,
+                                                     3, 1, 2, 0, 1, 0, 1, 1 };
+            return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
+#       endif
+        }
+    } else {  /* Big Endian CPU */
+        if (MEM_64bits()) {
+#       if defined(_MSC_VER) && defined(_WIN64)
+            unsigned long r = 0;
+            _BitScanReverse64( &r, val );
+            return (unsigned)(r>>3);
+#       elif defined(__GNUC__) && (__GNUC__ >= 4)
+            return (__builtin_clzll(val) >> 3);
+#       else
+            unsigned r;
+            const unsigned n32 = sizeof(size_t)*4;   /* calculate this way due to compiler complaining in 32-bits mode */
+            if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; }
+            if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
+            r += (!val);
+            return r;
+#       endif
+        } else { /* 32 bits */
+#       if defined(_MSC_VER)
+            unsigned long r = 0;
+            _BitScanReverse( &r, (unsigned long)val );
+            return (unsigned)(r>>3);
+#       elif defined(__GNUC__) && (__GNUC__ >= 3)
+            return (__builtin_clz((U32)val) >> 3);
+#       else
+            unsigned r;
+            if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
+            r += (!val);
+            return r;
+#       endif
+    }   }
+}
+
+
+MEM_STATIC size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit)
+{
+    const BYTE* const pStart = pIn;
+    const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1);
+
+    while (pIn < pInLoopLimit) {
+        size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);
+        if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; }
+        pIn += ZSTD_NbCommonBytes(diff);
+        return (size_t)(pIn - pStart);
+    }
+    if (MEM_64bits()) if ((pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; }
+    if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; }
+    if ((pIn<pInLimit) && (*pMatch == *pIn)) pIn++;
+    return (size_t)(pIn - pStart);
+}
+
+/** ZSTD_count_2segments() :
+*   can count match length with `ip` & `match` in 2 different segments.
+*   convention : on reaching mEnd, match count continue starting from iStart
+*/
+MEM_STATIC size_t ZSTD_count_2segments(const BYTE* ip, const BYTE* match, const BYTE* iEnd, const BYTE* mEnd, const BYTE* iStart)
+{
+    const BYTE* const vEnd = MIN( ip + (mEnd - match), iEnd);
+    size_t const matchLength = ZSTD_count(ip, match, vEnd);
+    if (match + matchLength != mEnd) return matchLength;
+    return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd);
+}
+
+
+/*-*************************************
+*  Hashes
+***************************************/
+static const U32 prime3bytes = 506832829U;
+static U32    ZSTD_hash3(U32 u, U32 h) { return ((u << (32-24)) * prime3bytes)  >> (32-h) ; }
+MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h); } /* only in zstd_opt.h */
+
+static const U32 prime4bytes = 2654435761U;
+static U32    ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32-h) ; }
+static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_read32(ptr), h); }
+
+static const U64 prime5bytes = 889523592379ULL;
+static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u  << (64-40)) * prime5bytes) >> (64-h)) ; }
+static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h); }
+
+static const U64 prime6bytes = 227718039650203ULL;
+static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u  << (64-48)) * prime6bytes) >> (64-h)) ; }
+static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }
+
+static const U64 prime7bytes = 58295818150454627ULL;
+static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u  << (64-56)) * prime7bytes) >> (64-h)) ; }
+static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h); }
+
+static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
+static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }
+static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }
+
+MEM_STATIC size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)
+{
+    switch(mls)
+    {
+    default:
+    case 4: return ZSTD_hash4Ptr(p, hBits);
+    case 5: return ZSTD_hash5Ptr(p, hBits);
+    case 6: return ZSTD_hash6Ptr(p, hBits);
+    case 7: return ZSTD_hash7Ptr(p, hBits);
+    case 8: return ZSTD_hash8Ptr(p, hBits);
+    }
+}
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_COMPRESS_H */
diff --git a/thirdparty/zstd/compress/zstd_double_fast.c b/thirdparty/zstd/compress/zstd_double_fast.c
new file mode 100644
index 0000000000..876a36042c
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_double_fast.c
@@ -0,0 +1,308 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_double_fast.h"
+
+
+void ZSTD_fillDoubleHashTable(ZSTD_CCtx* cctx, const void* end, const U32 mls)
+{
+    U32* const hashLarge = cctx->hashTable;
+    U32  const hBitsL = cctx->appliedParams.cParams.hashLog;
+    U32* const hashSmall = cctx->chainTable;
+    U32  const hBitsS = cctx->appliedParams.cParams.chainLog;
+    const BYTE* const base = cctx->base;
+    const BYTE* ip = base + cctx->nextToUpdate;
+    const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
+    const size_t fastHashFillStep = 3;
+
+    while(ip <= iend) {
+        hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
+        hashLarge[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
+        ip += fastHashFillStep;
+    }
+}
+
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx* cctx,
+                                 const void* src, size_t srcSize,
+                                 const U32 mls)
+{
+    U32* const hashLong = cctx->hashTable;
+    const U32 hBitsL = cctx->appliedParams.cParams.hashLog;
+    U32* const hashSmall = cctx->chainTable;
+    const U32 hBitsS = cctx->appliedParams.cParams.chainLog;
+    seqStore_t* seqStorePtr = &(cctx->seqStore);
+    const BYTE* const base = cctx->base;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32 lowestIndex = cctx->dictLimit;
+    const BYTE* const lowest = base + lowestIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - HASH_READ_SIZE;
+    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
+    U32 offsetSaved = 0;
+
+    /* init */
+    ip += (ip==lowest);
+    {   U32 const maxRep = (U32)(ip-lowest);
+        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
+    }
+
+    /* Main Search Loop */
+    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
+        size_t mLength;
+        size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);
+        size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);
+        U32 const current = (U32)(ip-base);
+        U32 const matchIndexL = hashLong[h2];
+        U32 const matchIndexS = hashSmall[h];
+        const BYTE* matchLong = base + matchIndexL;
+        const BYTE* match = base + matchIndexS;
+        hashLong[h2] = hashSmall[h] = current;   /* update hash tables */
+
+        assert(offset_1 <= current);   /* supposed guaranteed by construction */
+        if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {
+            /* favor repcode */
+            mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
+        } else {
+            U32 offset;
+            if ( (matchIndexL > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip)) ) {
+                mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
+                offset = (U32)(ip-matchLong);
+                while (((ip>anchor) & (matchLong>lowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
+            } else if ( (matchIndexS > lowestIndex) && (MEM_read32(match) == MEM_read32(ip)) ) {
+                size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
+                U32 const matchIndexL3 = hashLong[hl3];
+                const BYTE* matchL3 = base + matchIndexL3;
+                hashLong[hl3] = current + 1;
+                if ( (matchIndexL3 > lowestIndex) && (MEM_read64(matchL3) == MEM_read64(ip+1)) ) {
+                    mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;
+                    ip++;
+                    offset = (U32)(ip-matchL3);
+                    while (((ip>anchor) & (matchL3>lowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */
+                } else {
+                    mLength = ZSTD_count(ip+4, match+4, iend) + 4;
+                    offset = (U32)(ip-match);
+                    while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
+                }
+            } else {
+                ip += ((ip-anchor) >> g_searchStrength) + 1;
+                continue;
+            }
+
+            offset_2 = offset_1;
+            offset_1 = offset;
+
+            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+        }
+
+        /* match found */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Fill Table */
+            hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] =
+                hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;  /* here because current+2 could be > iend-8 */
+            hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] =
+                hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
+
+            /* check immediate repcode */
+            while ( (ip <= ilimit)
+                 && ( (offset_2>0)
+                 & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
+                /* store sequence */
+                size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
+                { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */
+                hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);
+                hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);
+                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);
+                ip += rLength;
+                anchor = ip;
+                continue;   /* faster when present ... (?) */
+    }   }   }
+
+    /* save reps for next block */
+    seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
+    seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_doubleFast(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    const U32 mls = ctx->appliedParams.cParams.searchLength;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 7);
+    }
+}
+
+
+static size_t ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx* ctx,
+                                 const void* src, size_t srcSize,
+                                 const U32 mls)
+{
+    U32* const hashLong = ctx->hashTable;
+    U32  const hBitsL = ctx->appliedParams.cParams.hashLog;
+    U32* const hashSmall = ctx->chainTable;
+    U32  const hBitsS = ctx->appliedParams.cParams.chainLog;
+    seqStore_t* seqStorePtr = &(ctx->seqStore);
+    const BYTE* const base = ctx->base;
+    const BYTE* const dictBase = ctx->dictBase;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   lowestIndex = ctx->lowLimit;
+    const BYTE* const dictStart = dictBase + lowestIndex;
+    const U32   dictLimit = ctx->dictLimit;
+    const BYTE* const lowPrefixPtr = base + dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
+
+    /* Search Loop */
+    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
+        const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
+        const U32 matchIndex = hashSmall[hSmall];
+        const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;
+        const BYTE* match = matchBase + matchIndex;
+
+        const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
+        const U32 matchLongIndex = hashLong[hLong];
+        const BYTE* matchLongBase = matchLongIndex < dictLimit ? dictBase : base;
+        const BYTE* matchLong = matchLongBase + matchLongIndex;
+
+        const U32 current = (U32)(ip-base);
+        const U32 repIndex = current + 1 - offset_1;   /* offset_1 expected <= current +1 */
+        const BYTE* repBase = repIndex < dictLimit ? dictBase : base;
+        const BYTE* repMatch = repBase + repIndex;
+        size_t mLength;
+        hashSmall[hSmall] = hashLong[hLong] = current;   /* update hash table */
+
+        if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))
+           && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+            const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
+            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, lowPrefixPtr) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
+        } else {
+            if ((matchLongIndex > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
+                const BYTE* matchEnd = matchLongIndex < dictLimit ? dictEnd : iend;
+                const BYTE* lowMatchPtr = matchLongIndex < dictLimit ? dictStart : lowPrefixPtr;
+                U32 offset;
+                mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, lowPrefixPtr) + 8;
+                offset = current - matchLongIndex;
+                while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; }   /* catch up */
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+
+            } else if ((matchIndex > lowestIndex) && (MEM_read32(match) == MEM_read32(ip))) {
+                size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
+                U32 const matchIndex3 = hashLong[h3];
+                const BYTE* const match3Base = matchIndex3 < dictLimit ? dictBase : base;
+                const BYTE* match3 = match3Base + matchIndex3;
+                U32 offset;
+                hashLong[h3] = current + 1;
+                if ( (matchIndex3 > lowestIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {
+                    const BYTE* matchEnd = matchIndex3 < dictLimit ? dictEnd : iend;
+                    const BYTE* lowMatchPtr = matchIndex3 < dictLimit ? dictStart : lowPrefixPtr;
+                    mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, lowPrefixPtr) + 8;
+                    ip++;
+                    offset = current+1 - matchIndex3;
+                    while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */
+                } else {
+                    const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;
+                    const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
+                    mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, lowPrefixPtr) + 4;
+                    offset = current - matchIndex;
+                    while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
+                }
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+
+            } else {
+                ip += ((ip-anchor) >> g_searchStrength) + 1;
+                continue;
+        }   }
+
+        /* found a match : store it */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Fill Table */
+            hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;
+            hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2;
+            hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
+            hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
+            /* check immediate repcode */
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex2 = current2 - offset_2;
+                const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
+                if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex))  /* intentional overflow */
+                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
+                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, lowPrefixPtr) + 4;
+                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);
+                    hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
+                    hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
+                    ip += repLength2;
+                    anchor = ip;
+                    continue;
+                }
+                break;
+    }   }   }
+
+    /* save reps for next block */
+    seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_doubleFast_extDict(ZSTD_CCtx* ctx,
+                         const void* src, size_t srcSize)
+{
+    U32 const mls = ctx->appliedParams.cParams.searchLength;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 7);
+    }
+}
diff --git a/thirdparty/zstd/compress/zstd_double_fast.h b/thirdparty/zstd/compress/zstd_double_fast.h
new file mode 100644
index 0000000000..3dba6c7108
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_double_fast.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_DOUBLE_FAST_H
+#define ZSTD_DOUBLE_FAST_H
+
+#include "zstd_compress.h"
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+void ZSTD_fillDoubleHashTable(ZSTD_CCtx* cctx, const void* end, const U32 mls);
+size_t ZSTD_compressBlock_doubleFast(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_doubleFast_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_DOUBLE_FAST_H */
diff --git a/thirdparty/zstd/compress/zstd_fast.c b/thirdparty/zstd/compress/zstd_fast.c
new file mode 100644
index 0000000000..2e057017b9
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_fast.c
@@ -0,0 +1,242 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_fast.h"
+
+
+void ZSTD_fillHashTable (ZSTD_CCtx* zc, const void* end, const U32 mls)
+{
+    U32* const hashTable = zc->hashTable;
+    U32  const hBits = zc->appliedParams.cParams.hashLog;
+    const BYTE* const base = zc->base;
+    const BYTE* ip = base + zc->nextToUpdate;
+    const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
+    const size_t fastHashFillStep = 3;
+
+    while(ip <= iend) {
+        hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);
+        ip += fastHashFillStep;
+    }
+}
+
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_fast_generic(ZSTD_CCtx* cctx,
+                                 const void* src, size_t srcSize,
+                                 const U32 mls)
+{
+    U32* const hashTable = cctx->hashTable;
+    U32  const hBits = cctx->appliedParams.cParams.hashLog;
+    seqStore_t* seqStorePtr = &(cctx->seqStore);
+    const BYTE* const base = cctx->base;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   lowestIndex = cctx->dictLimit;
+    const BYTE* const lowest = base + lowestIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - HASH_READ_SIZE;
+    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
+    U32 offsetSaved = 0;
+
+    /* init */
+    ip += (ip==lowest);
+    {   U32 const maxRep = (U32)(ip-lowest);
+        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
+    }
+
+    /* Main Search Loop */
+    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
+        size_t mLength;
+        size_t const h = ZSTD_hashPtr(ip, hBits, mls);
+        U32 const current = (U32)(ip-base);
+        U32 const matchIndex = hashTable[h];
+        const BYTE* match = base + matchIndex;
+        hashTable[h] = current;   /* update hash table */
+
+        if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {
+            mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
+        } else {
+            U32 offset;
+            if ( (matchIndex <= lowestIndex) || (MEM_read32(match) != MEM_read32(ip)) ) {
+                ip += ((ip-anchor) >> g_searchStrength) + 1;
+                continue;
+            }
+            mLength = ZSTD_count(ip+4, match+4, iend) + 4;
+            offset = (U32)(ip-match);
+            while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
+            offset_2 = offset_1;
+            offset_1 = offset;
+
+            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+        }
+
+        /* match found */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Fill Table */
+            hashTable[ZSTD_hashPtr(base+current+2, hBits, mls)] = current+2;  /* here because current+2 could be > iend-8 */
+            hashTable[ZSTD_hashPtr(ip-2, hBits, mls)] = (U32)(ip-2-base);
+            /* check immediate repcode */
+            while ( (ip <= ilimit)
+                 && ( (offset_2>0)
+                 & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
+                /* store sequence */
+                size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
+                { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; }  /* swap offset_2 <=> offset_1 */
+                hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip-base);
+                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);
+                ip += rLength;
+                anchor = ip;
+                continue;   /* faster when present ... (?) */
+    }   }   }
+
+    /* save reps for next block */
+    seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
+    seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_fast(ZSTD_CCtx* ctx,
+                       const void* src, size_t srcSize)
+{
+    const U32 mls = ctx->appliedParams.cParams.searchLength;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 7);
+    }
+}
+
+
+static size_t ZSTD_compressBlock_fast_extDict_generic(ZSTD_CCtx* ctx,
+                                 const void* src, size_t srcSize,
+                                 const U32 mls)
+{
+    U32* hashTable = ctx->hashTable;
+    const U32 hBits = ctx->appliedParams.cParams.hashLog;
+    seqStore_t* seqStorePtr = &(ctx->seqStore);
+    const BYTE* const base = ctx->base;
+    const BYTE* const dictBase = ctx->dictBase;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   lowestIndex = ctx->lowLimit;
+    const BYTE* const dictStart = dictBase + lowestIndex;
+    const U32   dictLimit = ctx->dictLimit;
+    const BYTE* const lowPrefixPtr = base + dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
+
+    /* Search Loop */
+    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
+        const size_t h = ZSTD_hashPtr(ip, hBits, mls);
+        const U32 matchIndex = hashTable[h];
+        const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;
+        const BYTE* match = matchBase + matchIndex;
+        const U32 current = (U32)(ip-base);
+        const U32 repIndex = current + 1 - offset_1;   /* offset_1 expected <= current +1 */
+        const BYTE* repBase = repIndex < dictLimit ? dictBase : base;
+        const BYTE* repMatch = repBase + repIndex;
+        size_t mLength;
+        hashTable[h] = current;   /* update hash table */
+
+        if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))
+           && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
+            const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
+            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, lowPrefixPtr) + 4;
+            ip++;
+            ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
+        } else {
+            if ( (matchIndex < lowestIndex) ||
+                 (MEM_read32(match) != MEM_read32(ip)) ) {
+                ip += ((ip-anchor) >> g_searchStrength) + 1;
+                continue;
+            }
+            {   const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;
+                const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
+                U32 offset;
+                mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, lowPrefixPtr) + 4;
+                while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
+                offset = current - matchIndex;
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+        }   }
+
+        /* found a match : store it */
+        ip += mLength;
+        anchor = ip;
+
+        if (ip <= ilimit) {
+            /* Fill Table */
+            hashTable[ZSTD_hashPtr(base+current+2, hBits, mls)] = current+2;
+            hashTable[ZSTD_hashPtr(ip-2, hBits, mls)] = (U32)(ip-2-base);
+            /* check immediate repcode */
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex2 = current2 - offset_2;
+                const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
+                if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex))  /* intentional overflow */
+                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
+                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, lowPrefixPtr) + 4;
+                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);
+                    hashTable[ZSTD_hashPtr(ip, hBits, mls)] = current2;
+                    ip += repLength2;
+                    anchor = ip;
+                    continue;
+                }
+                break;
+    }   }   }
+
+    /* save reps for next block */
+    seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_fast_extDict(ZSTD_CCtx* ctx,
+                         const void* src, size_t srcSize)
+{
+    U32 const mls = ctx->appliedParams.cParams.searchLength;
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 7);
+    }
+}
diff --git a/thirdparty/zstd/compress/zstd_fast.h b/thirdparty/zstd/compress/zstd_fast.h
new file mode 100644
index 0000000000..4205141a9a
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_fast.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_FAST_H
+#define ZSTD_FAST_H
+
+#include "zstd_compress.h"
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+void ZSTD_fillHashTable(ZSTD_CCtx* zc, const void* end, const U32 mls);
+size_t ZSTD_compressBlock_fast(ZSTD_CCtx* ctx,
+                         const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_fast_extDict(ZSTD_CCtx* ctx,
+                         const void* src, size_t srcSize);
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_FAST_H */
diff --git a/thirdparty/zstd/compress/zstd_lazy.c b/thirdparty/zstd/compress/zstd_lazy.c
new file mode 100644
index 0000000000..2a7f6a0fe2
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_lazy.c
@@ -0,0 +1,749 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_lazy.h"
+
+
+/*-*************************************
+*  Binary Tree search
+***************************************/
+/** ZSTD_insertBt1() : add one or multiple positions to tree.
+*   ip : assumed <= iend-8 .
+*   @return : nb of positions added */
+static U32 ZSTD_insertBt1(ZSTD_CCtx* zc, const BYTE* const ip, const U32 mls, const BYTE* const iend, U32 nbCompares,
+                          U32 extDict)
+{
+    U32*   const hashTable = zc->hashTable;
+    U32    const hashLog = zc->appliedParams.cParams.hashLog;
+    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32*   const bt = zc->chainTable;
+    U32    const btLog  = zc->appliedParams.cParams.chainLog - 1;
+    U32    const btMask = (1 << btLog) - 1;
+    U32 matchIndex = hashTable[h];
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+    const BYTE* const base = zc->base;
+    const BYTE* const dictBase = zc->dictBase;
+    const U32 dictLimit = zc->dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* match;
+    const U32 current = (U32)(ip-base);
+    const U32 btLow = btMask >= current ? 0 : current - btMask;
+    U32* smallerPtr = bt + 2*(current&btMask);
+    U32* largerPtr  = smallerPtr + 1;
+    U32 dummy32;   /* to be nullified at the end */
+    U32 const windowLow = zc->lowLimit;
+    U32 matchEndIdx = current+8;
+    size_t bestLength = 8;
+#ifdef ZSTD_C_PREDICT
+    U32 predictedSmall = *(bt + 2*((current-1)&btMask) + 0);
+    U32 predictedLarge = *(bt + 2*((current-1)&btMask) + 1);
+    predictedSmall += (predictedSmall>0);
+    predictedLarge += (predictedLarge>0);
+#endif /* ZSTD_C_PREDICT */
+
+    assert(ip <= iend-8);   /* required for h calculation */
+    hashTable[h] = current;   /* Update Hash Table */
+
+    while (nbCompares-- && (matchIndex > windowLow)) {
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+
+#ifdef ZSTD_C_PREDICT   /* note : can create issues when hlog small <= 11 */
+        const U32* predictPtr = bt + 2*((matchIndex-1) & btMask);   /* written this way, as bt is a roll buffer */
+        if (matchIndex == predictedSmall) {
+            /* no need to check length, result known */
+            *smallerPtr = matchIndex;
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
+            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+            predictedSmall = predictPtr[1] + (predictPtr[1]>0);
+            continue;
+        }
+        if (matchIndex == predictedLarge) {
+            *largerPtr = matchIndex;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+            predictedLarge = predictPtr[0] + (predictPtr[0]>0);
+            continue;
+        }
+#endif
+        if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
+            match = base + matchIndex;
+            if (match[matchLength] == ip[matchLength])
+                matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
+        } else {
+            match = dictBase + matchIndex;
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
+            if (matchIndex+matchLength >= dictLimit)
+                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
+        }
+
+        if (matchLength > bestLength) {
+            bestLength = matchLength;
+            if (matchLength > matchEndIdx - matchIndex)
+                matchEndIdx = matchIndex + (U32)matchLength;
+        }
+
+        if (ip+matchLength == iend)   /* equal : no way to know if inf or sup */
+            break;   /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */
+
+        if (match[matchLength] < ip[matchLength]) {  /* necessarily within buffer */
+            /* match+1 is smaller than current */
+            *smallerPtr = matchIndex;             /* update smaller idx */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop searching */
+            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
+            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+        } else {
+            /* match is larger than current */
+            *largerPtr = matchIndex;
+            commonLengthLarger = matchLength;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop searching */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+    }   }
+
+    *smallerPtr = *largerPtr = 0;
+    if (bestLength > 384) return MIN(192, (U32)(bestLength - 384));   /* speed optimization */
+    if (matchEndIdx > current + 8) return matchEndIdx - (current + 8);
+    return 1;
+}
+
+
+static size_t ZSTD_insertBtAndFindBestMatch (
+                        ZSTD_CCtx* zc,
+                        const BYTE* const ip, const BYTE* const iend,
+                        size_t* offsetPtr,
+                        U32 nbCompares, const U32 mls,
+                        U32 extDict)
+{
+    U32*   const hashTable = zc->hashTable;
+    U32    const hashLog = zc->appliedParams.cParams.hashLog;
+    size_t const h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32*   const bt = zc->chainTable;
+    U32    const btLog  = zc->appliedParams.cParams.chainLog - 1;
+    U32    const btMask = (1 << btLog) - 1;
+    U32 matchIndex  = hashTable[h];
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+    const BYTE* const base = zc->base;
+    const BYTE* const dictBase = zc->dictBase;
+    const U32 dictLimit = zc->dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const U32 current = (U32)(ip-base);
+    const U32 btLow = btMask >= current ? 0 : current - btMask;
+    const U32 windowLow = zc->lowLimit;
+    U32* smallerPtr = bt + 2*(current&btMask);
+    U32* largerPtr  = bt + 2*(current&btMask) + 1;
+    U32 matchEndIdx = current+8;
+    U32 dummy32;   /* to be nullified at the end */
+    size_t bestLength = 0;
+
+    assert(ip <= iend-8);   /* required for h calculation */
+    hashTable[h] = current;   /* Update Hash Table */
+
+    while (nbCompares-- && (matchIndex > windowLow)) {
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+        const BYTE* match;
+
+        if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
+            match = base + matchIndex;
+            if (match[matchLength] == ip[matchLength])
+                matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
+        } else {
+            match = dictBase + matchIndex;
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
+            if (matchIndex+matchLength >= dictLimit)
+                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
+        }
+
+        if (matchLength > bestLength) {
+            if (matchLength > matchEndIdx - matchIndex)
+                matchEndIdx = matchIndex + (U32)matchLength;
+            if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )
+                bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;
+            if (ip+matchLength == iend)   /* equal : no way to know if inf or sup */
+                break;   /* drop, to guarantee consistency (miss a little bit of compression) */
+        }
+
+        if (match[matchLength] < ip[matchLength]) {
+            /* match is smaller than current */
+            *smallerPtr = matchIndex;             /* update smaller idx */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
+            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+        } else {
+            /* match is larger than current */
+            *largerPtr = matchIndex;
+            commonLengthLarger = matchLength;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+    }   }
+
+    *smallerPtr = *largerPtr = 0;
+
+    zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;
+    return bestLength;
+}
+
+
+void ZSTD_updateTree(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls)
+{
+    const BYTE* const base = zc->base;
+    const U32 target = (U32)(ip - base);
+    U32 idx = zc->nextToUpdate;
+
+    while(idx < target)
+        idx += ZSTD_insertBt1(zc, base+idx, mls, iend, nbCompares, 0);
+}
+
+/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */
+static size_t ZSTD_BtFindBestMatch (
+                        ZSTD_CCtx* zc,
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 mls)
+{
+    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
+    ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
+    return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 0);
+}
+
+
+static size_t ZSTD_BtFindBestMatch_selectMLS (
+                        ZSTD_CCtx* zc,   /* Index table will be updated */
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 matchLengthSearch)
+{
+    switch(matchLengthSearch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
+    case 5 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
+    case 7 :
+    case 6 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
+    }
+}
+
+
+void ZSTD_updateTree_extDict(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls)
+{
+    const BYTE* const base = zc->base;
+    const U32 target = (U32)(ip - base);
+    U32 idx = zc->nextToUpdate;
+
+    while (idx < target) idx += ZSTD_insertBt1(zc, base+idx, mls, iend, nbCompares, 1);
+}
+
+
+/** Tree updater, providing best match */
+static size_t ZSTD_BtFindBestMatch_extDict (
+                        ZSTD_CCtx* zc,
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 mls)
+{
+    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
+    ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
+    return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 1);
+}
+
+
+static size_t ZSTD_BtFindBestMatch_selectMLS_extDict (
+                        ZSTD_CCtx* zc,   /* Index table will be updated */
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 matchLengthSearch)
+{
+    switch(matchLengthSearch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
+    case 5 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
+    case 7 :
+    case 6 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
+    }
+}
+
+
+
+/* *********************************
+*  Hash Chain
+***********************************/
+#define NEXT_IN_CHAIN(d, mask)   chainTable[(d) & mask]
+
+/* Update chains up to ip (excluded)
+   Assumption : always within prefix (i.e. not within extDict) */
+U32 ZSTD_insertAndFindFirstIndex (ZSTD_CCtx* zc, const BYTE* ip, U32 mls)
+{
+    U32* const hashTable  = zc->hashTable;
+    const U32 hashLog = zc->appliedParams.cParams.hashLog;
+    U32* const chainTable = zc->chainTable;
+    const U32 chainMask = (1 << zc->appliedParams.cParams.chainLog) - 1;
+    const BYTE* const base = zc->base;
+    const U32 target = (U32)(ip - base);
+    U32 idx = zc->nextToUpdate;
+
+    while(idx < target) { /* catch up */
+        size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);
+        NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
+        hashTable[h] = idx;
+        idx++;
+    }
+
+    zc->nextToUpdate = target;
+    return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
+}
+
+
+/* inlining is important to hardwire a hot branch (template emulation) */
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_HcFindBestMatch_generic (
+                        ZSTD_CCtx* zc,   /* Index table will be updated */
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 mls, const U32 extDict)
+{
+    U32* const chainTable = zc->chainTable;
+    const U32 chainSize = (1 << zc->appliedParams.cParams.chainLog);
+    const U32 chainMask = chainSize-1;
+    const BYTE* const base = zc->base;
+    const BYTE* const dictBase = zc->dictBase;
+    const U32 dictLimit = zc->dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const U32 lowLimit = zc->lowLimit;
+    const U32 current = (U32)(ip-base);
+    const U32 minChain = current > chainSize ? current - chainSize : 0;
+    int nbAttempts=maxNbAttempts;
+    size_t ml=4-1;
+
+    /* HC4 match finder */
+    U32 matchIndex = ZSTD_insertAndFindFirstIndex (zc, ip, mls);
+
+    for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) {
+        const BYTE* match;
+        size_t currentMl=0;
+        if ((!extDict) || matchIndex >= dictLimit) {
+            match = base + matchIndex;
+            if (match[ml] == ip[ml])   /* potentially better */
+                currentMl = ZSTD_count(ip, match, iLimit);
+        } else {
+            match = dictBase + matchIndex;
+            if (MEM_read32(match) == MEM_read32(ip))   /* assumption : matchIndex <= dictLimit-4 (by table construction) */
+                currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4;
+        }
+
+        /* save best solution */
+        if (currentMl > ml) {
+            ml = currentMl;
+            *offsetPtr = current - matchIndex + ZSTD_REP_MOVE;
+            if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
+        }
+
+        if (matchIndex <= minChain) break;
+        matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
+    }
+
+    return ml;
+}
+
+
+FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_selectMLS (
+                        ZSTD_CCtx* zc,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 matchLengthSearch)
+{
+    switch(matchLengthSearch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 0);
+    case 5 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 0);
+    case 7 :
+    case 6 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 0);
+    }
+}
+
+
+FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_extDict_selectMLS (
+                        ZSTD_CCtx* zc,
+                        const BYTE* ip, const BYTE* const iLimit,
+                        size_t* offsetPtr,
+                        const U32 maxNbAttempts, const U32 matchLengthSearch)
+{
+    switch(matchLengthSearch)
+    {
+    default : /* includes case 3 */
+    case 4 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 1);
+    case 5 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 1);
+    case 7 :
+    case 6 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 1);
+    }
+}
+
+
+/* *******************************
+*  Common parser - lazy strategy
+*********************************/
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_lazy_generic(ZSTD_CCtx* ctx,
+                                       const void* src, size_t srcSize,
+                                       const U32 searchMethod, const U32 depth)
+{
+    seqStore_t* seqStorePtr = &(ctx->seqStore);
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ctx->base + ctx->dictLimit;
+
+    U32 const maxSearches = 1 << ctx->appliedParams.cParams.searchLog;
+    U32 const mls = ctx->appliedParams.cParams.searchLength;
+
+    typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
+                        size_t* offsetPtr,
+                        U32 maxNbAttempts, U32 matchLengthSearch);
+    searchMax_f const searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS;
+    U32 offset_1 = seqStorePtr->rep[0], offset_2 = seqStorePtr->rep[1], savedOffset=0;
+
+    /* init */
+    ip += (ip==base);
+    ctx->nextToUpdate3 = ctx->nextToUpdate;
+    {   U32 const maxRep = (U32)(ip-base);
+        if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0;
+    }
+
+    /* Match Loop */
+    while (ip < ilimit) {
+        size_t matchLength=0;
+        size_t offset=0;
+        const BYTE* start=ip+1;
+
+        /* check repCode */
+        if ((offset_1>0) & (MEM_read32(ip+1) == MEM_read32(ip+1 - offset_1))) {
+            /* repcode : we take it */
+            matchLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
+            if (depth==0) goto _storeSequence;
+        }
+
+        /* first search (depth 0) */
+        {   size_t offsetFound = 99999999;
+            size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
+            if (ml2 > matchLength)
+                matchLength = ml2, start = ip, offset=offsetFound;
+        }
+
+        if (matchLength < 4) {
+            ip += ((ip-anchor) >> g_searchStrength) + 1;   /* jump faster over incompressible sections */
+            continue;
+        }
+
+        /* let's try to find a better solution */
+        if (depth>=1)
+        while (ip<ilimit) {
+            ip ++;
+            if ((offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {
+                size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;
+                int const gain2 = (int)(mlRep * 3);
+                int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
+                if ((mlRep >= 4) && (gain2 > gain1))
+                    matchLength = mlRep, offset = 0, start = ip;
+            }
+            {   size_t offset2=99999999;
+                size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
+                int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);
+                if ((ml2 >= 4) && (gain2 > gain1)) {
+                    matchLength = ml2, offset = offset2, start = ip;
+                    continue;   /* search a better one */
+            }   }
+
+            /* let's find an even better one */
+            if ((depth==2) && (ip<ilimit)) {
+                ip ++;
+                if ((offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {
+                    size_t const ml2 = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;
+                    int const gain2 = (int)(ml2 * 4);
+                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
+                    if ((ml2 >= 4) && (gain2 > gain1))
+                        matchLength = ml2, offset = 0, start = ip;
+                }
+                {   size_t offset2=99999999;
+                    size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
+                    int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);
+                    if ((ml2 >= 4) && (gain2 > gain1)) {
+                        matchLength = ml2, offset = offset2, start = ip;
+                        continue;
+            }   }   }
+            break;  /* nothing found : store previous solution */
+        }
+
+        /* NOTE:
+         * start[-offset+ZSTD_REP_MOVE-1] is undefined behavior.
+         * (-offset+ZSTD_REP_MOVE-1) is unsigned, and is added to start, which
+         * overflows the pointer, which is undefined behavior.
+         */
+        /* catch up */
+        if (offset) {
+            while ( (start > anchor)
+                 && (start > base+offset-ZSTD_REP_MOVE)
+                 && (start[-1] == (start-offset+ZSTD_REP_MOVE)[-1]) )  /* only search for offset within prefix */
+                { start--; matchLength++; }
+            offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);
+        }
+        /* store sequence */
+_storeSequence:
+        {   size_t const litLength = start - anchor;
+            ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength-MINMATCH);
+            anchor = ip = start + matchLength;
+        }
+
+        /* check immediate repcode */
+        while ( (ip <= ilimit)
+             && ((offset_2>0)
+             & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
+            /* store sequence */
+            matchLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
+            offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap repcodes */
+            ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength-MINMATCH);
+            ip += matchLength;
+            anchor = ip;
+            continue;   /* faster when present ... (?) */
+    }   }
+
+    /* Save reps for next block */
+    seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : savedOffset;
+    seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : savedOffset;
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_btlazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 1, 2);
+}
+
+size_t ZSTD_compressBlock_lazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 2);
+}
+
+size_t ZSTD_compressBlock_lazy(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 1);
+}
+
+size_t ZSTD_compressBlock_greedy(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 0);
+}
+
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_lazy_extDict_generic(ZSTD_CCtx* ctx,
+                                     const void* src, size_t srcSize,
+                                     const U32 searchMethod, const U32 depth)
+{
+    seqStore_t* seqStorePtr = &(ctx->seqStore);
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ctx->base;
+    const U32 dictLimit = ctx->dictLimit;
+    const U32 lowestIndex = ctx->lowLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* const dictBase = ctx->dictBase;
+    const BYTE* const dictEnd  = dictBase + dictLimit;
+    const BYTE* const dictStart  = dictBase + ctx->lowLimit;
+
+    const U32 maxSearches = 1 << ctx->appliedParams.cParams.searchLog;
+    const U32 mls = ctx->appliedParams.cParams.searchLength;
+
+    typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
+                        size_t* offsetPtr,
+                        U32 maxNbAttempts, U32 matchLengthSearch);
+    searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS_extDict : ZSTD_HcFindBestMatch_extDict_selectMLS;
+
+    U32 offset_1 = seqStorePtr->rep[0], offset_2 = seqStorePtr->rep[1];
+
+    /* init */
+    ctx->nextToUpdate3 = ctx->nextToUpdate;
+    ip += (ip == prefixStart);
+
+    /* Match Loop */
+    while (ip < ilimit) {
+        size_t matchLength=0;
+        size_t offset=0;
+        const BYTE* start=ip+1;
+        U32 current = (U32)(ip-base);
+
+        /* check repCode */
+        {   const U32 repIndex = (U32)(current+1 - offset_1);
+            const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+            const BYTE* const repMatch = repBase + repIndex;
+            if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))   /* intentional overflow */
+            if (MEM_read32(ip+1) == MEM_read32(repMatch)) {
+                /* repcode detected we should take it */
+                const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                if (depth==0) goto _storeSequence;
+        }   }
+
+        /* first search (depth 0) */
+        {   size_t offsetFound = 99999999;
+            size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
+            if (ml2 > matchLength)
+                matchLength = ml2, start = ip, offset=offsetFound;
+        }
+
+         if (matchLength < 4) {
+            ip += ((ip-anchor) >> g_searchStrength) + 1;   /* jump faster over incompressible sections */
+            continue;
+        }
+
+        /* let's try to find a better solution */
+        if (depth>=1)
+        while (ip<ilimit) {
+            ip ++;
+            current++;
+            /* check repCode */
+            if (offset) {
+                const U32 repIndex = (U32)(current - offset_1);
+                const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+                const BYTE* const repMatch = repBase + repIndex;
+                if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))  /* intentional overflow */
+                if (MEM_read32(ip) == MEM_read32(repMatch)) {
+                    /* repcode detected */
+                    const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                    size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                    int const gain2 = (int)(repLength * 3);
+                    int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);
+                    if ((repLength >= 4) && (gain2 > gain1))
+                        matchLength = repLength, offset = 0, start = ip;
+            }   }
+
+            /* search match, depth 1 */
+            {   size_t offset2=99999999;
+                size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
+                int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);
+                if ((ml2 >= 4) && (gain2 > gain1)) {
+                    matchLength = ml2, offset = offset2, start = ip;
+                    continue;   /* search a better one */
+            }   }
+
+            /* let's find an even better one */
+            if ((depth==2) && (ip<ilimit)) {
+                ip ++;
+                current++;
+                /* check repCode */
+                if (offset) {
+                    const U32 repIndex = (U32)(current - offset_1);
+                    const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+                    const BYTE* const repMatch = repBase + repIndex;
+                    if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))  /* intentional overflow */
+                    if (MEM_read32(ip) == MEM_read32(repMatch)) {
+                        /* repcode detected */
+                        const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                        size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                        int const gain2 = (int)(repLength * 4);
+                        int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);
+                        if ((repLength >= 4) && (gain2 > gain1))
+                            matchLength = repLength, offset = 0, start = ip;
+                }   }
+
+                /* search match, depth 2 */
+                {   size_t offset2=99999999;
+                    size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
+                    int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1));   /* raw approx */
+                    int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);
+                    if ((ml2 >= 4) && (gain2 > gain1)) {
+                        matchLength = ml2, offset = offset2, start = ip;
+                        continue;
+            }   }   }
+            break;  /* nothing found : store previous solution */
+        }
+
+        /* catch up */
+        if (offset) {
+            U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));
+            const BYTE* match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;
+            const BYTE* const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;
+            while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; }  /* catch up */
+            offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);
+        }
+
+        /* store sequence */
+_storeSequence:
+        {   size_t const litLength = start - anchor;
+            ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength-MINMATCH);
+            anchor = ip = start + matchLength;
+        }
+
+        /* check immediate repcode */
+        while (ip <= ilimit) {
+            const U32 repIndex = (U32)((ip-base) - offset_2);
+            const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+            const BYTE* const repMatch = repBase + repIndex;
+            if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex))  /* intentional overflow */
+            if (MEM_read32(ip) == MEM_read32(repMatch)) {
+                /* repcode detected we should take it */
+                const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
+                offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset;   /* swap offset history */
+                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength-MINMATCH);
+                ip += matchLength;
+                anchor = ip;
+                continue;   /* faster when present ... (?) */
+            }
+            break;
+    }   }
+
+    /* Save reps for next block */
+    seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_greedy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 0);
+}
+
+size_t ZSTD_compressBlock_lazy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 1);
+}
+
+size_t ZSTD_compressBlock_lazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 2);
+}
+
+size_t ZSTD_compressBlock_btlazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 1, 2);
+}
diff --git a/thirdparty/zstd/compress/zstd_lazy.h b/thirdparty/zstd/compress/zstd_lazy.h
new file mode 100644
index 0000000000..a9c4daed25
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_lazy.h
@@ -0,0 +1,38 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_LAZY_H
+#define ZSTD_LAZY_H
+
+#include "zstd_compress.h"
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+U32 ZSTD_insertAndFindFirstIndex (ZSTD_CCtx* zc, const BYTE* ip, U32 mls);
+void ZSTD_updateTree(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls);
+void ZSTD_updateTree_extDict(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls);
+
+size_t ZSTD_compressBlock_btlazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_greedy(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+
+size_t ZSTD_compressBlock_greedy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_btlazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_LAZY_H */
diff --git a/thirdparty/zstd/compress/zstd_ldm.c b/thirdparty/zstd/compress/zstd_ldm.c
new file mode 100644
index 0000000000..be50872cf7
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_ldm.c
@@ -0,0 +1,707 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ */
+
+#include "zstd_ldm.h"
+
+#include "zstd_fast.h"          /* ZSTD_fillHashTable() */
+#include "zstd_double_fast.h"   /* ZSTD_fillDoubleHashTable() */
+
+#define LDM_BUCKET_SIZE_LOG 3
+#define LDM_MIN_MATCH_LENGTH 64
+#define LDM_HASH_RLOG 7
+#define LDM_HASH_CHAR_OFFSET 10
+
+size_t ZSTD_ldm_initializeParameters(ldmParams_t* params, U32 enableLdm)
+{
+    ZSTD_STATIC_ASSERT(LDM_BUCKET_SIZE_LOG <= ZSTD_LDM_BUCKETSIZELOG_MAX);
+    params->enableLdm = enableLdm>0;
+    params->hashLog = 0;
+    params->bucketSizeLog = LDM_BUCKET_SIZE_LOG;
+    params->minMatchLength = LDM_MIN_MATCH_LENGTH;
+    params->hashEveryLog = ZSTD_LDM_HASHEVERYLOG_NOTSET;
+    return 0;
+}
+
+void ZSTD_ldm_adjustParameters(ldmParams_t* params, U32 windowLog)
+{
+    if (params->hashLog == 0) {
+        params->hashLog = MAX(ZSTD_HASHLOG_MIN, windowLog - LDM_HASH_RLOG);
+        assert(params->hashLog <= ZSTD_HASHLOG_MAX);
+    }
+    if (params->hashEveryLog == ZSTD_LDM_HASHEVERYLOG_NOTSET) {
+        params->hashEveryLog =
+                windowLog < params->hashLog ? 0 : windowLog - params->hashLog;
+    }
+    params->bucketSizeLog = MIN(params->bucketSizeLog, params->hashLog);
+}
+
+size_t ZSTD_ldm_getTableSize(U32 hashLog, U32 bucketSizeLog) {
+    size_t const ldmHSize = ((size_t)1) << hashLog;
+    size_t const ldmBucketSizeLog = MIN(bucketSizeLog, hashLog);
+    size_t const ldmBucketSize =
+        ((size_t)1) << (hashLog - ldmBucketSizeLog);
+    return ldmBucketSize + (ldmHSize * (sizeof(ldmEntry_t)));
+}
+
+/** ZSTD_ldm_getSmallHash() :
+ *  numBits should be <= 32
+ *  If numBits==0, returns 0.
+ *  @return : the most significant numBits of value. */
+static U32 ZSTD_ldm_getSmallHash(U64 value, U32 numBits)
+{
+    assert(numBits <= 32);
+    return numBits == 0 ? 0 : (U32)(value >> (64 - numBits));
+}
+
+/** ZSTD_ldm_getChecksum() :
+ *  numBitsToDiscard should be <= 32
+ *  @return : the next most significant 32 bits after numBitsToDiscard */
+static U32 ZSTD_ldm_getChecksum(U64 hash, U32 numBitsToDiscard)
+{
+    assert(numBitsToDiscard <= 32);
+    return (hash >> (64 - 32 - numBitsToDiscard)) & 0xFFFFFFFF;
+}
+
+/** ZSTD_ldm_getTag() ;
+ *  Given the hash, returns the most significant numTagBits bits
+ *  after (32 + hbits) bits.
+ *
+ *  If there are not enough bits remaining, return the last
+ *  numTagBits bits. */
+static U32 ZSTD_ldm_getTag(U64 hash, U32 hbits, U32 numTagBits)
+{
+    assert(numTagBits < 32 && hbits <= 32);
+    if (32 - hbits < numTagBits) {
+        return hash & (((U32)1 << numTagBits) - 1);
+    } else {
+        return (hash >> (32 - hbits - numTagBits)) & (((U32)1 << numTagBits) - 1);
+    }
+}
+
+/** ZSTD_ldm_getBucket() :
+ *  Returns a pointer to the start of the bucket associated with hash. */
+static ldmEntry_t* ZSTD_ldm_getBucket(
+        ldmState_t* ldmState, size_t hash, ldmParams_t const ldmParams)
+{
+    return ldmState->hashTable + (hash << ldmParams.bucketSizeLog);
+}
+
+/** ZSTD_ldm_insertEntry() :
+ *  Insert the entry with corresponding hash into the hash table */
+static void ZSTD_ldm_insertEntry(ldmState_t* ldmState,
+                                 size_t const hash, const ldmEntry_t entry,
+                                 ldmParams_t const ldmParams)
+{
+    BYTE* const bucketOffsets = ldmState->bucketOffsets;
+    *(ZSTD_ldm_getBucket(ldmState, hash, ldmParams) + bucketOffsets[hash]) = entry;
+    bucketOffsets[hash]++;
+    bucketOffsets[hash] &= ((U32)1 << ldmParams.bucketSizeLog) - 1;
+}
+
+/** ZSTD_ldm_makeEntryAndInsertByTag() :
+ *
+ *  Gets the small hash, checksum, and tag from the rollingHash.
+ *
+ *  If the tag matches (1 << ldmParams.hashEveryLog)-1, then
+ *  creates an ldmEntry from the offset, and inserts it into the hash table.
+ *
+ *  hBits is the length of the small hash, which is the most significant hBits
+ *  of rollingHash. The checksum is the next 32 most significant bits, followed
+ *  by ldmParams.hashEveryLog bits that make up the tag. */
+static void ZSTD_ldm_makeEntryAndInsertByTag(ldmState_t* ldmState,
+                                             U64 const rollingHash,
+                                             U32 const hBits,
+                                             U32 const offset,
+                                             ldmParams_t const ldmParams)
+{
+    U32 const tag = ZSTD_ldm_getTag(rollingHash, hBits, ldmParams.hashEveryLog);
+    U32 const tagMask = ((U32)1 << ldmParams.hashEveryLog) - 1;
+    if (tag == tagMask) {
+        U32 const hash = ZSTD_ldm_getSmallHash(rollingHash, hBits);
+        U32 const checksum = ZSTD_ldm_getChecksum(rollingHash, hBits);
+        ldmEntry_t entry;
+        entry.offset = offset;
+        entry.checksum = checksum;
+        ZSTD_ldm_insertEntry(ldmState, hash, entry, ldmParams);
+    }
+}
+
+/** ZSTD_ldm_getRollingHash() :
+ *  Get a 64-bit hash using the first len bytes from buf.
+ *
+ *  Giving bytes s = s_1, s_2, ... s_k, the hash is defined to be
+ *  H(s) = s_1*(a^(k-1)) + s_2*(a^(k-2)) + ... + s_k*(a^0)
+ *
+ *  where the constant a is defined to be prime8bytes.
+ *
+ *  The implementation adds an offset to each byte, so
+ *  H(s) = (s_1 + HASH_CHAR_OFFSET)*(a^(k-1)) + ... */
+static U64 ZSTD_ldm_getRollingHash(const BYTE* buf, U32 len)
+{
+    U64 ret = 0;
+    U32 i;
+    for (i = 0; i < len; i++) {
+        ret *= prime8bytes;
+        ret += buf[i] + LDM_HASH_CHAR_OFFSET;
+    }
+    return ret;
+}
+
+/** ZSTD_ldm_ipow() :
+ *  Return base^exp. */
+static U64 ZSTD_ldm_ipow(U64 base, U64 exp)
+{
+    U64 ret = 1;
+    while (exp) {
+        if (exp & 1) { ret *= base; }
+        exp >>= 1;
+        base *= base;
+    }
+    return ret;
+}
+
+U64 ZSTD_ldm_getHashPower(U32 minMatchLength) {
+    assert(minMatchLength >= ZSTD_LDM_MINMATCH_MIN);
+    return ZSTD_ldm_ipow(prime8bytes, minMatchLength - 1);
+}
+
+/** ZSTD_ldm_updateHash() :
+ *  Updates hash by removing toRemove and adding toAdd. */
+static U64 ZSTD_ldm_updateHash(U64 hash, BYTE toRemove, BYTE toAdd, U64 hashPower)
+{
+    hash -= ((toRemove + LDM_HASH_CHAR_OFFSET) * hashPower);
+    hash *= prime8bytes;
+    hash += toAdd + LDM_HASH_CHAR_OFFSET;
+    return hash;
+}
+
+/** ZSTD_ldm_countBackwardsMatch() :
+ *  Returns the number of bytes that match backwards before pIn and pMatch.
+ *
+ *  We count only bytes where pMatch >= pBase and pIn >= pAnchor. */
+static size_t ZSTD_ldm_countBackwardsMatch(
+            const BYTE* pIn, const BYTE* pAnchor,
+            const BYTE* pMatch, const BYTE* pBase)
+{
+    size_t matchLength = 0;
+    while (pIn > pAnchor && pMatch > pBase && pIn[-1] == pMatch[-1]) {
+        pIn--;
+        pMatch--;
+        matchLength++;
+    }
+    return matchLength;
+}
+
+/** ZSTD_ldm_fillFastTables() :
+ *
+ *  Fills the relevant tables for the ZSTD_fast and ZSTD_dfast strategies.
+ *  This is similar to ZSTD_loadDictionaryContent.
+ *
+ *  The tables for the other strategies are filled within their
+ *  block compressors. */
+static size_t ZSTD_ldm_fillFastTables(ZSTD_CCtx* zc, const void* end)
+{
+    const BYTE* const iend = (const BYTE*)end;
+    const U32 mls = zc->appliedParams.cParams.searchLength;
+
+    switch(zc->appliedParams.cParams.strategy)
+    {
+    case ZSTD_fast:
+        ZSTD_fillHashTable(zc, iend, mls);
+        zc->nextToUpdate = (U32)(iend - zc->base);
+        break;
+
+    case ZSTD_dfast:
+        ZSTD_fillDoubleHashTable(zc, iend, mls);
+        zc->nextToUpdate = (U32)(iend - zc->base);
+        break;
+
+    case ZSTD_greedy:
+    case ZSTD_lazy:
+    case ZSTD_lazy2:
+    case ZSTD_btlazy2:
+    case ZSTD_btopt:
+    case ZSTD_btultra:
+        break;
+    default:
+        assert(0);  /* not possible : not a valid strategy id */
+    }
+
+    return 0;
+}
+
+/** ZSTD_ldm_fillLdmHashTable() :
+ *
+ *  Fills hashTable from (lastHashed + 1) to iend (non-inclusive).
+ *  lastHash is the rolling hash that corresponds to lastHashed.
+ *
+ *  Returns the rolling hash corresponding to position iend-1. */
+static U64 ZSTD_ldm_fillLdmHashTable(ldmState_t* state,
+                                     U64 lastHash, const BYTE* lastHashed,
+                                     const BYTE* iend, const BYTE* base,
+                                     U32 hBits, ldmParams_t const ldmParams)
+{
+    U64 rollingHash = lastHash;
+    const BYTE* cur = lastHashed + 1;
+
+    while (cur < iend) {
+        rollingHash = ZSTD_ldm_updateHash(rollingHash, cur[-1],
+                                          cur[ldmParams.minMatchLength-1],
+                                          state->hashPower);
+        ZSTD_ldm_makeEntryAndInsertByTag(state,
+                                         rollingHash, hBits,
+                                         (U32)(cur - base), ldmParams);
+        ++cur;
+    }
+    return rollingHash;
+}
+
+
+/** ZSTD_ldm_limitTableUpdate() :
+ *
+ *  Sets cctx->nextToUpdate to a position corresponding closer to anchor
+ *  if it is far way
+ *  (after a long match, only update tables a limited amount). */
+static void ZSTD_ldm_limitTableUpdate(ZSTD_CCtx* cctx, const BYTE* anchor)
+{
+    U32 const current = (U32)(anchor - cctx->base);
+    if (current > cctx->nextToUpdate + 1024) {
+        cctx->nextToUpdate =
+            current - MIN(512, current - cctx->nextToUpdate - 1024);
+    }
+}
+
+typedef size_t (*ZSTD_blockCompressor) (ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+/* defined in zstd_compress.c */
+ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict);
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_ldm_generic(ZSTD_CCtx* cctx,
+                                      const void* src, size_t srcSize)
+{
+    ldmState_t* const ldmState = &(cctx->ldmState);
+    const ldmParams_t ldmParams = cctx->appliedParams.ldmParams;
+    const U64 hashPower = ldmState->hashPower;
+    const U32 hBits = ldmParams.hashLog - ldmParams.bucketSizeLog;
+    const U32 ldmBucketSize = ((U32)1 << ldmParams.bucketSizeLog);
+    const U32 ldmTagMask = ((U32)1 << ldmParams.hashEveryLog) - 1;
+    seqStore_t* const seqStorePtr = &(cctx->seqStore);
+    const BYTE* const base = cctx->base;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   lowestIndex = cctx->dictLimit;
+    const BYTE* const lowest = base + lowestIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - MAX(ldmParams.minMatchLength, HASH_READ_SIZE);
+
+    const ZSTD_blockCompressor blockCompressor =
+        ZSTD_selectBlockCompressor(cctx->appliedParams.cParams.strategy, 0);
+    U32* const repToConfirm = seqStorePtr->repToConfirm;
+    U32 savedRep[ZSTD_REP_NUM];
+    U64 rollingHash = 0;
+    const BYTE* lastHashed = NULL;
+    size_t i, lastLiterals;
+
+    /* Save seqStorePtr->rep and copy repToConfirm */
+    for (i = 0; i < ZSTD_REP_NUM; i++)
+        savedRep[i] = repToConfirm[i] = seqStorePtr->rep[i];
+
+    /* Main Search Loop */
+    while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
+        size_t mLength;
+        U32 const current = (U32)(ip - base);
+        size_t forwardMatchLength = 0, backwardMatchLength = 0;
+        ldmEntry_t* bestEntry = NULL;
+        if (ip != istart) {
+            rollingHash = ZSTD_ldm_updateHash(rollingHash, lastHashed[0],
+                                              lastHashed[ldmParams.minMatchLength],
+                                              hashPower);
+        } else {
+            rollingHash = ZSTD_ldm_getRollingHash(ip, ldmParams.minMatchLength);
+        }
+        lastHashed = ip;
+
+        /* Do not insert and do not look for a match */
+        if (ZSTD_ldm_getTag(rollingHash, hBits, ldmParams.hashEveryLog) !=
+                ldmTagMask) {
+           ip++;
+           continue;
+        }
+
+        /* Get the best entry and compute the match lengths */
+        {
+            ldmEntry_t* const bucket =
+                ZSTD_ldm_getBucket(ldmState,
+                                   ZSTD_ldm_getSmallHash(rollingHash, hBits),
+                                   ldmParams);
+            ldmEntry_t* cur;
+            size_t bestMatchLength = 0;
+            U32 const checksum = ZSTD_ldm_getChecksum(rollingHash, hBits);
+
+            for (cur = bucket; cur < bucket + ldmBucketSize; ++cur) {
+                const BYTE* const pMatch = cur->offset + base;
+                size_t curForwardMatchLength, curBackwardMatchLength,
+                       curTotalMatchLength;
+                if (cur->checksum != checksum || cur->offset <= lowestIndex) {
+                    continue;
+                }
+
+                curForwardMatchLength = ZSTD_count(ip, pMatch, iend);
+                if (curForwardMatchLength < ldmParams.minMatchLength) {
+                    continue;
+                }
+                curBackwardMatchLength = ZSTD_ldm_countBackwardsMatch(
+                                             ip, anchor, pMatch, lowest);
+                curTotalMatchLength = curForwardMatchLength +
+                                      curBackwardMatchLength;
+
+                if (curTotalMatchLength > bestMatchLength) {
+                    bestMatchLength = curTotalMatchLength;
+                    forwardMatchLength = curForwardMatchLength;
+                    backwardMatchLength = curBackwardMatchLength;
+                    bestEntry = cur;
+                }
+            }
+        }
+
+        /* No match found -- continue searching */
+        if (bestEntry == NULL) {
+            ZSTD_ldm_makeEntryAndInsertByTag(ldmState, rollingHash,
+                                             hBits, current,
+                                             ldmParams);
+            ip++;
+            continue;
+        }
+
+        /* Match found */
+        mLength = forwardMatchLength + backwardMatchLength;
+        ip -= backwardMatchLength;
+
+        /* Call the block compressor on the remaining literals */
+        {
+            U32 const matchIndex = bestEntry->offset;
+            const BYTE* const match = base + matchIndex - backwardMatchLength;
+            U32 const offset = (U32)(ip - match);
+
+            /* Overwrite rep codes */
+            for (i = 0; i < ZSTD_REP_NUM; i++)
+                seqStorePtr->rep[i] = repToConfirm[i];
+
+            /* Fill tables for block compressor */
+            ZSTD_ldm_limitTableUpdate(cctx, anchor);
+            ZSTD_ldm_fillFastTables(cctx, anchor);
+
+            /* Call block compressor and get remaining literals */
+            lastLiterals = blockCompressor(cctx, anchor, ip - anchor);
+            cctx->nextToUpdate = (U32)(ip - base);
+
+            /* Update repToConfirm with the new offset */
+            for (i = ZSTD_REP_NUM - 1; i > 0; i--)
+                repToConfirm[i] = repToConfirm[i-1];
+            repToConfirm[0] = offset;
+
+            /* Store the sequence with the leftover literals */
+            ZSTD_storeSeq(seqStorePtr, lastLiterals, ip - lastLiterals,
+                          offset + ZSTD_REP_MOVE, mLength - MINMATCH);
+        }
+
+        /* Insert the current entry into the hash table */
+        ZSTD_ldm_makeEntryAndInsertByTag(ldmState, rollingHash, hBits,
+                                         (U32)(lastHashed - base),
+                                         ldmParams);
+
+        assert(ip + backwardMatchLength == lastHashed);
+
+        /* Fill the hash table from lastHashed+1 to ip+mLength*/
+        /* Heuristic: don't need to fill the entire table at end of block */
+        if (ip + mLength < ilimit) {
+            rollingHash = ZSTD_ldm_fillLdmHashTable(
+                              ldmState, rollingHash, lastHashed,
+                              ip + mLength, base, hBits, ldmParams);
+            lastHashed = ip + mLength - 1;
+        }
+        ip += mLength;
+        anchor = ip;
+        /* Check immediate repcode */
+        while ( (ip < ilimit)
+             && ( (repToConfirm[1] > 0) && (repToConfirm[1] <= (U32)(ip-lowest))
+             && (MEM_read32(ip) == MEM_read32(ip - repToConfirm[1])) )) {
+
+            size_t const rLength = ZSTD_count(ip+4, ip+4-repToConfirm[1],
+                                              iend) + 4;
+            /* Swap repToConfirm[1] <=> repToConfirm[0] */
+            {
+                U32 const tmpOff = repToConfirm[1];
+                repToConfirm[1] = repToConfirm[0];
+                repToConfirm[0] = tmpOff;
+            }
+
+            ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);
+
+            /* Fill the  hash table from lastHashed+1 to ip+rLength*/
+            if (ip + rLength < ilimit) {
+                rollingHash = ZSTD_ldm_fillLdmHashTable(
+                                ldmState, rollingHash, lastHashed,
+                                ip + rLength, base, hBits, ldmParams);
+                lastHashed = ip + rLength - 1;
+            }
+            ip += rLength;
+            anchor = ip;
+        }
+    }
+
+    /* Overwrite rep */
+    for (i = 0; i < ZSTD_REP_NUM; i++)
+        seqStorePtr->rep[i] = repToConfirm[i];
+
+    ZSTD_ldm_limitTableUpdate(cctx, anchor);
+    ZSTD_ldm_fillFastTables(cctx, anchor);
+
+    lastLiterals = blockCompressor(cctx, anchor, iend - anchor);
+    cctx->nextToUpdate = (U32)(iend - base);
+
+    /* Restore seqStorePtr->rep */
+    for (i = 0; i < ZSTD_REP_NUM; i++)
+        seqStorePtr->rep[i] = savedRep[i];
+
+    /* Return the last literals size */
+    return lastLiterals;
+}
+
+size_t ZSTD_compressBlock_ldm(ZSTD_CCtx* ctx,
+                              const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_ldm_generic(ctx, src, srcSize);
+}
+
+static size_t ZSTD_compressBlock_ldm_extDict_generic(
+                                 ZSTD_CCtx* ctx,
+                                 const void* src, size_t srcSize)
+{
+    ldmState_t* const ldmState = &(ctx->ldmState);
+    const ldmParams_t ldmParams = ctx->appliedParams.ldmParams;
+    const U64 hashPower = ldmState->hashPower;
+    const U32 hBits = ldmParams.hashLog - ldmParams.bucketSizeLog;
+    const U32 ldmBucketSize = ((U32)1 << ldmParams.bucketSizeLog);
+    const U32 ldmTagMask = ((U32)1 << ldmParams.hashEveryLog) - 1;
+    seqStore_t* const seqStorePtr = &(ctx->seqStore);
+    const BYTE* const base = ctx->base;
+    const BYTE* const dictBase = ctx->dictBase;
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const U32   lowestIndex = ctx->lowLimit;
+    const BYTE* const dictStart = dictBase + lowestIndex;
+    const U32   dictLimit = ctx->dictLimit;
+    const BYTE* const lowPrefixPtr = base + dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - MAX(ldmParams.minMatchLength, HASH_READ_SIZE);
+
+    const ZSTD_blockCompressor blockCompressor =
+        ZSTD_selectBlockCompressor(ctx->appliedParams.cParams.strategy, 1);
+    U32* const repToConfirm = seqStorePtr->repToConfirm;
+    U32 savedRep[ZSTD_REP_NUM];
+    U64 rollingHash = 0;
+    const BYTE* lastHashed = NULL;
+    size_t i, lastLiterals;
+
+    /* Save seqStorePtr->rep and copy repToConfirm */
+    for (i = 0; i < ZSTD_REP_NUM; i++) {
+        savedRep[i] = repToConfirm[i] = seqStorePtr->rep[i];
+    }
+
+    /* Search Loop */
+    while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
+        size_t mLength;
+        const U32 current = (U32)(ip-base);
+        size_t forwardMatchLength = 0, backwardMatchLength = 0;
+        ldmEntry_t* bestEntry = NULL;
+        if (ip != istart) {
+          rollingHash = ZSTD_ldm_updateHash(rollingHash, lastHashed[0],
+                                       lastHashed[ldmParams.minMatchLength],
+                                       hashPower);
+        } else {
+            rollingHash = ZSTD_ldm_getRollingHash(ip, ldmParams.minMatchLength);
+        }
+        lastHashed = ip;
+
+        if (ZSTD_ldm_getTag(rollingHash, hBits, ldmParams.hashEveryLog) !=
+                ldmTagMask) {
+            /* Don't insert and don't look for a match */
+           ip++;
+           continue;
+        }
+
+        /* Get the best entry and compute the match lengths */
+        {
+            ldmEntry_t* const bucket =
+                ZSTD_ldm_getBucket(ldmState,
+                                   ZSTD_ldm_getSmallHash(rollingHash, hBits),
+                                   ldmParams);
+            ldmEntry_t* cur;
+            size_t bestMatchLength = 0;
+            U32 const checksum = ZSTD_ldm_getChecksum(rollingHash, hBits);
+
+            for (cur = bucket; cur < bucket + ldmBucketSize; ++cur) {
+                const BYTE* const curMatchBase =
+                    cur->offset < dictLimit ? dictBase : base;
+                const BYTE* const pMatch = curMatchBase + cur->offset;
+                const BYTE* const matchEnd =
+                    cur->offset < dictLimit ? dictEnd : iend;
+                const BYTE* const lowMatchPtr =
+                    cur->offset < dictLimit ? dictStart : lowPrefixPtr;
+                size_t curForwardMatchLength, curBackwardMatchLength,
+                       curTotalMatchLength;
+
+                if (cur->checksum != checksum || cur->offset <= lowestIndex) {
+                    continue;
+                }
+
+                curForwardMatchLength = ZSTD_count_2segments(
+                                            ip, pMatch, iend,
+                                            matchEnd, lowPrefixPtr);
+                if (curForwardMatchLength < ldmParams.minMatchLength) {
+                    continue;
+                }
+                curBackwardMatchLength = ZSTD_ldm_countBackwardsMatch(
+                                             ip, anchor, pMatch, lowMatchPtr);
+                curTotalMatchLength = curForwardMatchLength +
+                                      curBackwardMatchLength;
+
+                if (curTotalMatchLength > bestMatchLength) {
+                    bestMatchLength = curTotalMatchLength;
+                    forwardMatchLength = curForwardMatchLength;
+                    backwardMatchLength = curBackwardMatchLength;
+                    bestEntry = cur;
+                }
+            }
+        }
+
+        /* No match found -- continue searching */
+        if (bestEntry == NULL) {
+            ZSTD_ldm_makeEntryAndInsertByTag(ldmState, rollingHash, hBits,
+                                             (U32)(lastHashed - base),
+                                             ldmParams);
+            ip++;
+            continue;
+        }
+
+        /* Match found */
+        mLength = forwardMatchLength + backwardMatchLength;
+        ip -= backwardMatchLength;
+
+        /* Call the block compressor on the remaining literals */
+        {
+            /* ip = current - backwardMatchLength
+             * The match is at (bestEntry->offset - backwardMatchLength) */
+            U32 const matchIndex = bestEntry->offset;
+            U32 const offset = current - matchIndex;
+
+            /* Overwrite rep codes */
+            for (i = 0; i < ZSTD_REP_NUM; i++)
+                seqStorePtr->rep[i] = repToConfirm[i];
+
+            /* Fill the hash table for the block compressor */
+            ZSTD_ldm_limitTableUpdate(ctx, anchor);
+            ZSTD_ldm_fillFastTables(ctx, anchor);
+
+            /* Call block compressor and get remaining literals  */
+            lastLiterals = blockCompressor(ctx, anchor, ip - anchor);
+            ctx->nextToUpdate = (U32)(ip - base);
+
+            /* Update repToConfirm with the new offset */
+            for (i = ZSTD_REP_NUM - 1; i > 0; i--)
+                repToConfirm[i] = repToConfirm[i-1];
+            repToConfirm[0] = offset;
+
+            /* Store the sequence with the leftover literals */
+            ZSTD_storeSeq(seqStorePtr, lastLiterals, ip - lastLiterals,
+                          offset + ZSTD_REP_MOVE, mLength - MINMATCH);
+        }
+
+        /* Insert the current entry into the hash table */
+        ZSTD_ldm_makeEntryAndInsertByTag(ldmState, rollingHash, hBits,
+                                         (U32)(lastHashed - base),
+                                         ldmParams);
+
+        /* Fill the hash table from lastHashed+1 to ip+mLength */
+        assert(ip + backwardMatchLength == lastHashed);
+        if (ip + mLength < ilimit) {
+            rollingHash = ZSTD_ldm_fillLdmHashTable(
+                              ldmState, rollingHash, lastHashed,
+                              ip + mLength, base, hBits,
+                              ldmParams);
+            lastHashed = ip + mLength - 1;
+        }
+        ip += mLength;
+        anchor = ip;
+
+        /* check immediate repcode */
+        while (ip < ilimit) {
+            U32 const current2 = (U32)(ip-base);
+            U32 const repIndex2 = current2 - repToConfirm[1];
+            const BYTE* repMatch2 = repIndex2 < dictLimit ?
+                                    dictBase + repIndex2 : base + repIndex2;
+            if ( (((U32)((dictLimit-1) - repIndex2) >= 3) &
+                        (repIndex2 > lowestIndex))  /* intentional overflow */
+               && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                const BYTE* const repEnd2 = repIndex2 < dictLimit ?
+                                            dictEnd : iend;
+                size_t const repLength2 =
+                        ZSTD_count_2segments(ip+4, repMatch2+4, iend,
+                                             repEnd2, lowPrefixPtr) + 4;
+
+                U32 tmpOffset = repToConfirm[1];
+                repToConfirm[1] = repToConfirm[0];
+                repToConfirm[0] = tmpOffset;
+
+                ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);
+
+                /* Fill the  hash table from lastHashed+1 to ip+repLength2*/
+                if (ip + repLength2 < ilimit) {
+                    rollingHash = ZSTD_ldm_fillLdmHashTable(
+                                      ldmState, rollingHash, lastHashed,
+                                      ip + repLength2, base, hBits,
+                                      ldmParams);
+                    lastHashed = ip + repLength2 - 1;
+                }
+                ip += repLength2;
+                anchor = ip;
+                continue;
+            }
+            break;
+        }
+    }
+
+    /* Overwrite rep */
+    for (i = 0; i < ZSTD_REP_NUM; i++)
+        seqStorePtr->rep[i] = repToConfirm[i];
+
+    ZSTD_ldm_limitTableUpdate(ctx, anchor);
+    ZSTD_ldm_fillFastTables(ctx, anchor);
+
+    /* Call the block compressor one last time on the last literals */
+    lastLiterals = blockCompressor(ctx, anchor, iend - anchor);
+    ctx->nextToUpdate = (U32)(iend - base);
+
+    /* Restore seqStorePtr->rep */
+    for (i = 0; i < ZSTD_REP_NUM; i++)
+        seqStorePtr->rep[i] = savedRep[i];
+
+    /* Return the last literals size */
+    return lastLiterals;
+}
+
+size_t ZSTD_compressBlock_ldm_extDict(ZSTD_CCtx* ctx,
+                                      const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_ldm_extDict_generic(ctx, src, srcSize);
+}
diff --git a/thirdparty/zstd/compress/zstd_ldm.h b/thirdparty/zstd/compress/zstd_ldm.h
new file mode 100644
index 0000000000..d6d3d42c33
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_ldm.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ */
+
+#ifndef ZSTD_LDM_H
+#define ZSTD_LDM_H
+
+#include "zstd_compress.h"
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+/*-*************************************
+*  Long distance matching
+***************************************/
+
+#define ZSTD_LDM_DEFAULT_WINDOW_LOG ZSTD_WINDOWLOG_DEFAULTMAX
+#define ZSTD_LDM_HASHEVERYLOG_NOTSET 9999
+
+/** ZSTD_compressBlock_ldm_generic() :
+ *
+ *  This is a block compressor intended for long distance matching.
+ *
+ *  The function searches for matches of length at least
+ *  ldmParams.minMatchLength using a hash table in cctx->ldmState.
+ *  Matches can be at a distance of up to cParams.windowLog.
+ *
+ *  Upon finding a match, the unmatched literals are compressed using a
+ *  ZSTD_blockCompressor (depending on the strategy in the compression
+ *  parameters), which stores the matched sequences. The "long distance"
+ *  match is then stored with the remaining literals from the
+ *  ZSTD_blockCompressor. */
+size_t ZSTD_compressBlock_ldm(ZSTD_CCtx* cctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_ldm_extDict(ZSTD_CCtx* ctx,
+                                      const void* src, size_t srcSize);
+
+/** ZSTD_ldm_initializeParameters() :
+ *  Initialize the long distance matching parameters to their default values. */
+size_t ZSTD_ldm_initializeParameters(ldmParams_t* params, U32 enableLdm);
+
+/** ZSTD_ldm_getTableSize() :
+ *  Estimate the space needed for long distance matching tables. */
+size_t ZSTD_ldm_getTableSize(U32 hashLog, U32 bucketSizeLog);
+
+/** ZSTD_ldm_getTableSize() :
+ *  Return prime8bytes^(minMatchLength-1) */
+U64 ZSTD_ldm_getHashPower(U32 minMatchLength);
+
+/** ZSTD_ldm_adjustParameters() :
+ *  If the params->hashEveryLog is not set, set it to its default value based on
+ *  windowLog and params->hashLog.
+ *
+ *  Ensures that params->bucketSizeLog is <= params->hashLog (setting it to
+ *  params->hashLog if it is not). */
+void ZSTD_ldm_adjustParameters(ldmParams_t* params, U32 windowLog);
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_FAST_H */
diff --git a/thirdparty/zstd/compress/zstd_opt.c b/thirdparty/zstd/compress/zstd_opt.c
new file mode 100644
index 0000000000..c47ce23ad5
--- /dev/null
+++ b/thirdparty/zstd/compress/zstd_opt.c
@@ -0,0 +1,957 @@
+/*
+ * Copyright (c) 2016-present, Przemyslaw Skibinski, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "zstd_opt.h"
+#include "zstd_lazy.h"
+
+
+#define ZSTD_LITFREQ_ADD    2
+#define ZSTD_FREQ_DIV       4
+#define ZSTD_MAX_PRICE      (1<<30)
+
+/*-*************************************
+*  Price functions for optimal parser
+***************************************/
+static void ZSTD_setLog2Prices(optState_t* optPtr)
+{
+    optPtr->log2matchLengthSum = ZSTD_highbit32(optPtr->matchLengthSum+1);
+    optPtr->log2litLengthSum = ZSTD_highbit32(optPtr->litLengthSum+1);
+    optPtr->log2litSum = ZSTD_highbit32(optPtr->litSum+1);
+    optPtr->log2offCodeSum = ZSTD_highbit32(optPtr->offCodeSum+1);
+    optPtr->factor = 1 + ((optPtr->litSum>>5) / optPtr->litLengthSum) + ((optPtr->litSum<<1) / (optPtr->litSum + optPtr->matchSum));
+}
+
+
+static void ZSTD_rescaleFreqs(optState_t* optPtr, const BYTE* src, size_t srcSize)
+{
+    unsigned u;
+
+    optPtr->cachedLiterals = NULL;
+    optPtr->cachedPrice = optPtr->cachedLitLength = 0;
+    optPtr->staticPrices = 0;
+
+    if (optPtr->litLengthSum == 0) {
+        if (srcSize <= 1024) optPtr->staticPrices = 1;
+
+        assert(optPtr->litFreq!=NULL);
+        for (u=0; u<=MaxLit; u++)
+            optPtr->litFreq[u] = 0;
+        for (u=0; u<srcSize; u++)
+            optPtr->litFreq[src[u]]++;
+
+        optPtr->litSum = 0;
+        optPtr->litLengthSum = MaxLL+1;
+        optPtr->matchLengthSum = MaxML+1;
+        optPtr->offCodeSum = (MaxOff+1);
+        optPtr->matchSum = (ZSTD_LITFREQ_ADD<<Litbits);
+
+        for (u=0; u<=MaxLit; u++) {
+            optPtr->litFreq[u] = 1 + (optPtr->litFreq[u]>>ZSTD_FREQ_DIV);
+            optPtr->litSum += optPtr->litFreq[u];
+        }
+        for (u=0; u<=MaxLL; u++)
+            optPtr->litLengthFreq[u] = 1;
+        for (u=0; u<=MaxML; u++)
+            optPtr->matchLengthFreq[u] = 1;
+        for (u=0; u<=MaxOff; u++)
+            optPtr->offCodeFreq[u] = 1;
+    } else {
+        optPtr->matchLengthSum = 0;
+        optPtr->litLengthSum = 0;
+        optPtr->offCodeSum = 0;
+        optPtr->matchSum = 0;
+        optPtr->litSum = 0;
+
+        for (u=0; u<=MaxLit; u++) {
+            optPtr->litFreq[u] = 1 + (optPtr->litFreq[u]>>(ZSTD_FREQ_DIV+1));
+            optPtr->litSum += optPtr->litFreq[u];
+        }
+        for (u=0; u<=MaxLL; u++) {
+            optPtr->litLengthFreq[u] = 1 + (optPtr->litLengthFreq[u]>>(ZSTD_FREQ_DIV+1));
+            optPtr->litLengthSum += optPtr->litLengthFreq[u];
+        }
+        for (u=0; u<=MaxML; u++) {
+            optPtr->matchLengthFreq[u] = 1 + (optPtr->matchLengthFreq[u]>>ZSTD_FREQ_DIV);
+            optPtr->matchLengthSum += optPtr->matchLengthFreq[u];
+            optPtr->matchSum += optPtr->matchLengthFreq[u] * (u + 3);
+        }
+        optPtr->matchSum *= ZSTD_LITFREQ_ADD;
+        for (u=0; u<=MaxOff; u++) {
+            optPtr->offCodeFreq[u] = 1 + (optPtr->offCodeFreq[u]>>ZSTD_FREQ_DIV);
+            optPtr->offCodeSum += optPtr->offCodeFreq[u];
+        }
+    }
+
+    ZSTD_setLog2Prices(optPtr);
+}
+
+
+static U32 ZSTD_getLiteralPrice(optState_t* optPtr, U32 litLength, const BYTE* literals)
+{
+    U32 price, u;
+
+    if (optPtr->staticPrices)
+        return ZSTD_highbit32((U32)litLength+1) + (litLength*6);
+
+    if (litLength == 0)
+        return optPtr->log2litLengthSum - ZSTD_highbit32(optPtr->litLengthFreq[0]+1);
+
+    /* literals */
+    if (optPtr->cachedLiterals == literals) {
+        U32 const additional = litLength - optPtr->cachedLitLength;
+        const BYTE* literals2 = optPtr->cachedLiterals + optPtr->cachedLitLength;
+        price = optPtr->cachedPrice + additional * optPtr->log2litSum;
+        for (u=0; u < additional; u++)
+            price -= ZSTD_highbit32(optPtr->litFreq[literals2[u]]+1);
+        optPtr->cachedPrice = price;
+        optPtr->cachedLitLength = litLength;
+    } else {
+        price = litLength * optPtr->log2litSum;
+        for (u=0; u < litLength; u++)
+            price -= ZSTD_highbit32(optPtr->litFreq[literals[u]]+1);
+
+        if (litLength >= 12) {
+            optPtr->cachedLiterals = literals;
+            optPtr->cachedPrice = price;
+            optPtr->cachedLitLength = litLength;
+        }
+    }
+
+    /* literal Length */
+    {   const BYTE LL_deltaCode = 19;
+        const BYTE llCode = (litLength>63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
+        price += LL_bits[llCode] + optPtr->log2litLengthSum - ZSTD_highbit32(optPtr->litLengthFreq[llCode]+1);
+    }
+
+    return price;
+}
+
+
+FORCE_INLINE_TEMPLATE U32 ZSTD_getPrice(optState_t* optPtr, U32 litLength, const BYTE* literals, U32 offset, U32 matchLength, const int ultra)
+{
+    /* offset */
+    U32 price;
+    BYTE const offCode = (BYTE)ZSTD_highbit32(offset+1);
+
+    if (optPtr->staticPrices)
+        return ZSTD_getLiteralPrice(optPtr, litLength, literals) + ZSTD_highbit32((U32)matchLength+1) + 16 + offCode;
+
+    price = offCode + optPtr->log2offCodeSum - ZSTD_highbit32(optPtr->offCodeFreq[offCode]+1);
+    if (!ultra && offCode >= 20) price += (offCode-19)*2;
+
+    /* match Length */
+    {   const BYTE ML_deltaCode = 36;
+        const BYTE mlCode = (matchLength>127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];
+        price += ML_bits[mlCode] + optPtr->log2matchLengthSum - ZSTD_highbit32(optPtr->matchLengthFreq[mlCode]+1);
+    }
+
+    return price + ZSTD_getLiteralPrice(optPtr, litLength, literals) + optPtr->factor;
+}
+
+
+static void ZSTD_updatePrice(optState_t* optPtr, U32 litLength, const BYTE* literals, U32 offset, U32 matchLength)
+{
+    U32 u;
+
+    /* literals */
+    optPtr->litSum += litLength*ZSTD_LITFREQ_ADD;
+    for (u=0; u < litLength; u++)
+        optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;
+
+    /* literal Length */
+    {   const BYTE LL_deltaCode = 19;
+        const BYTE llCode = (litLength>63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
+        optPtr->litLengthFreq[llCode]++;
+        optPtr->litLengthSum++;
+    }
+
+    /* match offset */
+    {   BYTE const offCode = (BYTE)ZSTD_highbit32(offset+1);
+        optPtr->offCodeSum++;
+        optPtr->offCodeFreq[offCode]++;
+    }
+
+    /* match Length */
+    {   const BYTE ML_deltaCode = 36;
+        const BYTE mlCode = (matchLength>127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];
+        optPtr->matchLengthFreq[mlCode]++;
+        optPtr->matchLengthSum++;
+    }
+
+    ZSTD_setLog2Prices(optPtr);
+}
+
+
+#define SET_PRICE(pos, mlen_, offset_, litlen_, price_)   \
+    {                                                 \
+        while (last_pos < pos)  { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; } \
+        opt[pos].mlen = mlen_;                         \
+        opt[pos].off = offset_;                        \
+        opt[pos].litlen = litlen_;                     \
+        opt[pos].price = price_;                       \
+    }
+
+
+/* function safe only for comparisons */
+static U32 ZSTD_readMINMATCH(const void* memPtr, U32 length)
+{
+    switch (length)
+    {
+    default :
+    case 4 : return MEM_read32(memPtr);
+    case 3 : if (MEM_isLittleEndian())
+                return MEM_read32(memPtr)<<8;
+             else
+                return MEM_read32(memPtr)>>8;
+    }
+}
+
+
+/* Update hashTable3 up to ip (excluded)
+   Assumption : always within prefix (i.e. not within extDict) */
+static
+U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_CCtx* zc, const BYTE* ip)
+{
+    U32* const hashTable3  = zc->hashTable3;
+    U32 const hashLog3  = zc->hashLog3;
+    const BYTE* const base = zc->base;
+    U32 idx = zc->nextToUpdate3;
+    const U32 target = zc->nextToUpdate3 = (U32)(ip - base);
+    const size_t hash3 = ZSTD_hash3Ptr(ip, hashLog3);
+
+    while(idx < target) {
+        hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx;
+        idx++;
+    }
+
+    return hashTable3[hash3];
+}
+
+
+/*-*************************************
+*  Binary Tree search
+***************************************/
+static U32 ZSTD_insertBtAndGetAllMatches (
+                        ZSTD_CCtx* zc,
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        U32 nbCompares, const U32 mls,
+                        U32 extDict, ZSTD_match_t* matches, const U32 minMatchLen)
+{
+    const BYTE* const base = zc->base;
+    const U32 current = (U32)(ip-base);
+    const U32 hashLog = zc->appliedParams.cParams.hashLog;
+    const size_t h  = ZSTD_hashPtr(ip, hashLog, mls);
+    U32* const hashTable = zc->hashTable;
+    U32 matchIndex  = hashTable[h];
+    U32* const bt   = zc->chainTable;
+    const U32 btLog = zc->appliedParams.cParams.chainLog - 1;
+    const U32 btMask= (1U << btLog) - 1;
+    size_t commonLengthSmaller=0, commonLengthLarger=0;
+    const BYTE* const dictBase = zc->dictBase;
+    const U32 dictLimit = zc->dictLimit;
+    const BYTE* const dictEnd = dictBase + dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const U32 btLow = btMask >= current ? 0 : current - btMask;
+    const U32 windowLow = zc->lowLimit;
+    U32* smallerPtr = bt + 2*(current&btMask);
+    U32* largerPtr  = bt + 2*(current&btMask) + 1;
+    U32 matchEndIdx = current+8;
+    U32 dummy32;   /* to be nullified at the end */
+    U32 mnum = 0;
+
+    const U32 minMatch = (mls == 3) ? 3 : 4;
+    size_t bestLength = minMatchLen-1;
+
+    if (minMatch == 3) { /* HC3 match finder */
+        U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3 (zc, ip);
+        if (matchIndex3>windowLow && (current - matchIndex3 < (1<<18))) {
+            const BYTE* match;
+            size_t currentMl=0;
+            if ((!extDict) || matchIndex3 >= dictLimit) {
+                match = base + matchIndex3;
+                if (match[bestLength] == ip[bestLength]) currentMl = ZSTD_count(ip, match, iLimit);
+            } else {
+                match = dictBase + matchIndex3;
+                if (ZSTD_readMINMATCH(match, MINMATCH) == ZSTD_readMINMATCH(ip, MINMATCH))    /* assumption : matchIndex3 <= dictLimit-4 (by table construction) */
+                    currentMl = ZSTD_count_2segments(ip+MINMATCH, match+MINMATCH, iLimit, dictEnd, prefixStart) + MINMATCH;
+            }
+
+            /* save best solution */
+            if (currentMl > bestLength) {
+                bestLength = currentMl;
+                matches[mnum].off = ZSTD_REP_MOVE_OPT + current - matchIndex3;
+                matches[mnum].len = (U32)currentMl;
+                mnum++;
+                if (currentMl > ZSTD_OPT_NUM) goto update;
+                if (ip+currentMl == iLimit) goto update; /* best possible, and avoid read overflow*/
+            }
+        }
+    }
+
+    hashTable[h] = current;   /* Update Hash Table */
+
+    while (nbCompares-- && (matchIndex > windowLow)) {
+        U32* nextPtr = bt + 2*(matchIndex & btMask);
+        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+        const BYTE* match;
+
+        if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
+            match = base + matchIndex;
+            if (match[matchLength] == ip[matchLength]) {
+                matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iLimit) +1;
+            }
+        } else {
+            match = dictBase + matchIndex;
+            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);
+            if (matchIndex+matchLength >= dictLimit)
+                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
+        }
+
+        if (matchLength > bestLength) {
+            if (matchLength > matchEndIdx - matchIndex) matchEndIdx = matchIndex + (U32)matchLength;
+            bestLength = matchLength;
+            matches[mnum].off = ZSTD_REP_MOVE_OPT + current - matchIndex;
+            matches[mnum].len = (U32)matchLength;
+            mnum++;
+            if (matchLength > ZSTD_OPT_NUM) break;
+            if (ip+matchLength == iLimit)   /* equal : no way to know if inf or sup */
+                break;   /* drop, to guarantee consistency (miss a little bit of compression) */
+        }
+
+        if (match[matchLength] < ip[matchLength]) {
+            /* match is smaller than current */
+            *smallerPtr = matchIndex;             /* update smaller idx */
+            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
+            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
+            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
+        } else {
+            /* match is larger than current */
+            *largerPtr = matchIndex;
+            commonLengthLarger = matchLength;
+            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
+            largerPtr = nextPtr;
+            matchIndex = nextPtr[0];
+    }   }
+
+    *smallerPtr = *largerPtr = 0;
+
+update:
+    zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;
+    return mnum;
+}
+
+
+/** Tree updater, providing best match */
+static U32 ZSTD_BtGetAllMatches (
+                        ZSTD_CCtx* zc,
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        const U32 maxNbAttempts, const U32 mls, ZSTD_match_t* matches, const U32 minMatchLen)
+{
+    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
+    ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
+    return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 0, matches, minMatchLen);
+}
+
+
+static U32 ZSTD_BtGetAllMatches_selectMLS (
+                        ZSTD_CCtx* zc,   /* Index table will be updated */
+                        const BYTE* ip, const BYTE* const iHighLimit,
+                        const U32 maxNbAttempts, const U32 matchLengthSearch, ZSTD_match_t* matches, const U32 minMatchLen)
+{
+    switch(matchLengthSearch)
+    {
+    case 3 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);
+    default :
+    case 4 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);
+    case 5 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);
+    case 7 :
+    case 6 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);
+    }
+}
+
+/** Tree updater, providing best match */
+static U32 ZSTD_BtGetAllMatches_extDict (
+                        ZSTD_CCtx* zc,
+                        const BYTE* const ip, const BYTE* const iLimit,
+                        const U32 maxNbAttempts, const U32 mls, ZSTD_match_t* matches, const U32 minMatchLen)
+{
+    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
+    ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
+    return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 1, matches, minMatchLen);
+}
+
+
+static U32 ZSTD_BtGetAllMatches_selectMLS_extDict (
+                        ZSTD_CCtx* zc,   /* Index table will be updated */
+                        const BYTE* ip, const BYTE* const iHighLimit,
+                        const U32 maxNbAttempts, const U32 matchLengthSearch, ZSTD_match_t* matches, const U32 minMatchLen)
+{
+    switch(matchLengthSearch)
+    {
+    case 3 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);
+    default :
+    case 4 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);
+    case 5 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);
+    case 7 :
+    case 6 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);
+    }
+}
+
+
+/*-*******************************
+*  Optimal parser
+*********************************/
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_opt_generic(ZSTD_CCtx* ctx,
+                                      const void* src, size_t srcSize, const int ultra)
+{
+    seqStore_t* seqStorePtr = &(ctx->seqStore);
+    optState_t* optStatePtr = &(ctx->optState);
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ctx->base;
+    const BYTE* const prefixStart = base + ctx->dictLimit;
+
+    const U32 maxSearches = 1U << ctx->appliedParams.cParams.searchLog;
+    const U32 sufficient_len = ctx->appliedParams.cParams.targetLength;
+    const U32 mls = ctx->appliedParams.cParams.searchLength;
+    const U32 minMatch = (ctx->appliedParams.cParams.searchLength == 3) ? 3 : 4;
+
+    ZSTD_optimal_t* opt = optStatePtr->priceTable;
+    ZSTD_match_t* matches = optStatePtr->matchTable;
+    const BYTE* inr;
+    U32 offset, rep[ZSTD_REP_NUM];
+
+    /* init */
+    ctx->nextToUpdate3 = ctx->nextToUpdate;
+    ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize);
+    ip += (ip==prefixStart);
+    { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) rep[i]=seqStorePtr->rep[i]; }
+
+    /* Match Loop */
+    while (ip < ilimit) {
+        U32 cur, match_num, last_pos, litlen, price;
+        U32 u, mlen, best_mlen, best_off, litLength;
+        memset(opt, 0, sizeof(ZSTD_optimal_t));
+        last_pos = 0;
+        litlen = (U32)(ip - anchor);
+
+        /* check repCode */
+        {   U32 i, last_i = ZSTD_REP_CHECK + (ip==anchor);
+            for (i=(ip == anchor); i<last_i; i++) {
+                const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];
+                if ( (repCur > 0) && (repCur < (S32)(ip-prefixStart))
+                    && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(ip - repCur, minMatch))) {
+                    mlen = (U32)ZSTD_count(ip+minMatch, ip+minMatch-repCur, iend) + minMatch;
+                    if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {
+                        best_mlen = mlen; best_off = i; cur = 0; last_pos = 1;
+                        goto _storeSequence;
+                    }
+                    best_off = i - (ip == anchor);
+                    do {
+                        price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
+                        if (mlen > last_pos || price < opt[mlen].price)
+                            SET_PRICE(mlen, mlen, i, litlen, price);   /* note : macro modifies last_pos */
+                        mlen--;
+                    } while (mlen >= minMatch);
+        }   }   }
+
+        match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, ip, iend, maxSearches, mls, matches, minMatch);
+
+        if (!last_pos && !match_num) { ip++; continue; }
+
+        if (match_num && (matches[match_num-1].len > sufficient_len || matches[match_num-1].len >= ZSTD_OPT_NUM)) {
+            best_mlen = matches[match_num-1].len;
+            best_off = matches[match_num-1].off;
+            cur = 0;
+            last_pos = 1;
+            goto _storeSequence;
+        }
+
+        /* set prices using matches at position = 0 */
+        best_mlen = (last_pos) ? last_pos : minMatch;
+        for (u = 0; u < match_num; u++) {
+            mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
+            best_mlen = matches[u].len;
+            while (mlen <= best_mlen) {
+                price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
+                if (mlen > last_pos || price < opt[mlen].price)
+                    SET_PRICE(mlen, mlen, matches[u].off, litlen, price);   /* note : macro modifies last_pos */
+                mlen++;
+        }   }
+
+        if (last_pos < minMatch) { ip++; continue; }
+
+        /* initialize opt[0] */
+        { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
+        opt[0].mlen = 1;
+        opt[0].litlen = litlen;
+
+         /* check further positions */
+        for (cur = 1; cur <= last_pos; cur++) {
+           inr = ip + cur;
+
+           if (opt[cur-1].mlen == 1) {
+                litlen = opt[cur-1].litlen + 1;
+                if (cur > litlen) {
+                    price = opt[cur - litlen].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-litlen);
+                } else
+                    price = ZSTD_getLiteralPrice(optStatePtr, litlen, anchor);
+           } else {
+                litlen = 1;
+                price = opt[cur - 1].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-1);
+           }
+
+           if (cur > last_pos || price <= opt[cur].price)
+                SET_PRICE(cur, 1, 0, litlen, price);
+
+           if (cur == last_pos) break;
+
+           if (inr > ilimit)  /* last match must start at a minimum distance of 8 from oend */
+               continue;
+
+           mlen = opt[cur].mlen;
+           if (opt[cur].off > ZSTD_REP_MOVE_OPT) {
+                opt[cur].rep[2] = opt[cur-mlen].rep[1];
+                opt[cur].rep[1] = opt[cur-mlen].rep[0];
+                opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;
+           } else {
+                opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur-mlen].rep[1] : opt[cur-mlen].rep[2];
+                opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur-mlen].rep[0] : opt[cur-mlen].rep[1];
+                /* If opt[cur].off == ZSTD_REP_MOVE_OPT, then mlen != 1.
+                 * offset ZSTD_REP_MOVE_OPT is used for the special case
+                 * litLength == 0, where offset 0 means something special.
+                 * mlen == 1 means the previous byte was stored as a literal,
+                 * so they are mutually exclusive.
+                 */
+                assert(!(opt[cur].off == ZSTD_REP_MOVE_OPT && mlen == 1));
+                opt[cur].rep[0] = (opt[cur].off == ZSTD_REP_MOVE_OPT) ? (opt[cur-mlen].rep[0] - 1) : (opt[cur-mlen].rep[opt[cur].off]);
+           }
+
+            best_mlen = minMatch;
+            {   U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);
+                for (i=(opt[cur].mlen != 1); i<last_i; i++) {  /* check rep */
+                    const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];
+                    if ( (repCur > 0) && (repCur < (S32)(inr-prefixStart))
+                       && (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(inr - repCur, minMatch))) {
+                       mlen = (U32)ZSTD_count(inr+minMatch, inr+minMatch - repCur, iend) + minMatch;
+
+                       if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {
+                            best_mlen = mlen; best_off = i; last_pos = cur + 1;
+                            goto _storeSequence;
+                       }
+
+                       best_off = i - (opt[cur].mlen != 1);
+                       if (mlen > best_mlen) best_mlen = mlen;
+
+                       do {
+                           if (opt[cur].mlen == 1) {
+                                litlen = opt[cur].litlen;
+                                if (cur > litlen) {
+                                    price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, inr-litlen, best_off, mlen - MINMATCH, ultra);
+                                } else
+                                    price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
+                            } else {
+                                litlen = 0;
+                                price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);
+                            }
+
+                            if (cur + mlen > last_pos || price <= opt[cur + mlen].price)
+                                SET_PRICE(cur + mlen, mlen, i, litlen, price);
+                            mlen--;
+                        } while (mlen >= minMatch);
+            }   }   }
+
+            match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, inr, iend, maxSearches, mls, matches, best_mlen);
+
+            if (match_num > 0 && (matches[match_num-1].len > sufficient_len || cur + matches[match_num-1].len >= ZSTD_OPT_NUM)) {
+                best_mlen = matches[match_num-1].len;
+                best_off = matches[match_num-1].off;
+                last_pos = cur + 1;
+                goto _storeSequence;
+            }
+
+            /* set prices using matches at position = cur */
+            for (u = 0; u < match_num; u++) {
+                mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
+                best_mlen = matches[u].len;
+
+                while (mlen <= best_mlen) {
+                    if (opt[cur].mlen == 1) {
+                        litlen = opt[cur].litlen;
+                        if (cur > litlen)
+                            price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, ip+cur-litlen, matches[u].off-1, mlen - MINMATCH, ultra);
+                        else
+                            price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
+                    } else {
+                        litlen = 0;
+                        price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, matches[u].off-1, mlen - MINMATCH, ultra);
+                    }
+
+                    if (cur + mlen > last_pos || (price < opt[cur + mlen].price))
+                        SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);
+
+                    mlen++;
+        }   }   }
+
+        best_mlen = opt[last_pos].mlen;
+        best_off = opt[last_pos].off;
+        cur = last_pos - best_mlen;
+
+        /* store sequence */
+_storeSequence:   /* cur, last_pos, best_mlen, best_off have to be set */
+        opt[0].mlen = 1;
+
+        while (1) {
+            mlen = opt[cur].mlen;
+            offset = opt[cur].off;
+            opt[cur].mlen = best_mlen;
+            opt[cur].off = best_off;
+            best_mlen = mlen;
+            best_off = offset;
+            if (mlen > cur) break;
+            cur -= mlen;
+        }
+
+        for (u = 0; u <= last_pos;) {
+            u += opt[u].mlen;
+        }
+
+        for (cur=0; cur < last_pos; ) {
+            mlen = opt[cur].mlen;
+            if (mlen == 1) { ip++; cur++; continue; }
+            offset = opt[cur].off;
+            cur += mlen;
+            litLength = (U32)(ip - anchor);
+
+            if (offset > ZSTD_REP_MOVE_OPT) {
+                rep[2] = rep[1];
+                rep[1] = rep[0];
+                rep[0] = offset - ZSTD_REP_MOVE_OPT;
+                offset--;
+            } else {
+                if (offset != 0) {
+                    best_off = (offset==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);
+                    if (offset != 1) rep[2] = rep[1];
+                    rep[1] = rep[0];
+                    rep[0] = best_off;
+                }
+                if (litLength==0) offset--;
+            }
+
+            ZSTD_updatePrice(optStatePtr, litLength, anchor, offset, mlen-MINMATCH);
+            ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);
+            anchor = ip = ip + mlen;
+    }    }   /* for (cur=0; cur < last_pos; ) */
+
+    /* Save reps for next block */
+    { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqStorePtr->repToConfirm[i] = rep[i]; }
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_btopt(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 0);
+}
+
+size_t ZSTD_compressBlock_btultra(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 1);
+}
+
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_opt_extDict_generic(ZSTD_CCtx* ctx,
+                                     const void* src, size_t srcSize, const int ultra)
+{
+    seqStore_t* seqStorePtr = &(ctx->seqStore);
+    optState_t* optStatePtr = &(ctx->optState);
+    const BYTE* const istart = (const BYTE*)src;
+    const BYTE* ip = istart;
+    const BYTE* anchor = istart;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - 8;
+    const BYTE* const base = ctx->base;
+    const U32 lowestIndex = ctx->lowLimit;
+    const U32 dictLimit = ctx->dictLimit;
+    const BYTE* const prefixStart = base + dictLimit;
+    const BYTE* const dictBase = ctx->dictBase;
+    const BYTE* const dictEnd  = dictBase + dictLimit;
+
+    const U32 maxSearches = 1U << ctx->appliedParams.cParams.searchLog;
+    const U32 sufficient_len = ctx->appliedParams.cParams.targetLength;
+    const U32 mls = ctx->appliedParams.cParams.searchLength;
+    const U32 minMatch = (ctx->appliedParams.cParams.searchLength == 3) ? 3 : 4;
+
+    ZSTD_optimal_t* opt = optStatePtr->priceTable;
+    ZSTD_match_t* matches = optStatePtr->matchTable;
+    const BYTE* inr;
+
+    /* init */
+    U32 offset, rep[ZSTD_REP_NUM];
+    { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) rep[i]=seqStorePtr->rep[i]; }
+
+    ctx->nextToUpdate3 = ctx->nextToUpdate;
+    ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize);
+    ip += (ip==prefixStart);
+
+    /* Match Loop */
+    while (ip < ilimit) {
+        U32 cur, match_num, last_pos, litlen, price;
+        U32 u, mlen, best_mlen, best_off, litLength;
+        U32 current = (U32)(ip-base);
+        memset(opt, 0, sizeof(ZSTD_optimal_t));
+        last_pos = 0;
+        opt[0].litlen = (U32)(ip - anchor);
+
+        /* check repCode */
+        {   U32 i, last_i = ZSTD_REP_CHECK + (ip==anchor);
+            for (i = (ip==anchor); i<last_i; i++) {
+                const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];
+                const U32 repIndex = (U32)(current - repCur);
+                const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+                const BYTE* const repMatch = repBase + repIndex;
+                if ( (repCur > 0 && repCur <= (S32)current)
+                   && (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex>lowestIndex))  /* intentional overflow */
+                   && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
+                    /* repcode detected we should take it */
+                    const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                    mlen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iend, repEnd, prefixStart) + minMatch;
+
+                    if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {
+                        best_mlen = mlen; best_off = i; cur = 0; last_pos = 1;
+                        goto _storeSequence;
+                    }
+
+                    best_off = i - (ip==anchor);
+                    litlen = opt[0].litlen;
+                    do {
+                        price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
+                        if (mlen > last_pos || price < opt[mlen].price)
+                            SET_PRICE(mlen, mlen, i, litlen, price);   /* note : macro modifies last_pos */
+                        mlen--;
+                    } while (mlen >= minMatch);
+        }   }   }
+
+        match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, ip, iend, maxSearches, mls, matches, minMatch);  /* first search (depth 0) */
+
+        if (!last_pos && !match_num) { ip++; continue; }
+
+        { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
+        opt[0].mlen = 1;
+
+        if (match_num && (matches[match_num-1].len > sufficient_len || matches[match_num-1].len >= ZSTD_OPT_NUM)) {
+            best_mlen = matches[match_num-1].len;
+            best_off = matches[match_num-1].off;
+            cur = 0;
+            last_pos = 1;
+            goto _storeSequence;
+        }
+
+        best_mlen = (last_pos) ? last_pos : minMatch;
+
+        /* set prices using matches at position = 0 */
+        for (u = 0; u < match_num; u++) {
+            mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
+            best_mlen = matches[u].len;
+            litlen = opt[0].litlen;
+            while (mlen <= best_mlen) {
+                price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
+                if (mlen > last_pos || price < opt[mlen].price)
+                    SET_PRICE(mlen, mlen, matches[u].off, litlen, price);
+                mlen++;
+        }   }
+
+        if (last_pos < minMatch) {
+            ip++; continue;
+        }
+
+        /* check further positions */
+        for (cur = 1; cur <= last_pos; cur++) {
+            inr = ip + cur;
+
+            if (opt[cur-1].mlen == 1) {
+                litlen = opt[cur-1].litlen + 1;
+                if (cur > litlen) {
+                    price = opt[cur - litlen].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-litlen);
+                } else
+                    price = ZSTD_getLiteralPrice(optStatePtr, litlen, anchor);
+            } else {
+                litlen = 1;
+                price = opt[cur - 1].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-1);
+            }
+
+            if (cur > last_pos || price <= opt[cur].price)
+                SET_PRICE(cur, 1, 0, litlen, price);
+
+            if (cur == last_pos) break;
+
+            if (inr > ilimit)  /* last match must start at a minimum distance of 8 from oend */
+                continue;
+
+            mlen = opt[cur].mlen;
+            if (opt[cur].off > ZSTD_REP_MOVE_OPT) {
+                opt[cur].rep[2] = opt[cur-mlen].rep[1];
+                opt[cur].rep[1] = opt[cur-mlen].rep[0];
+                opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;
+            } else {
+                opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur-mlen].rep[1] : opt[cur-mlen].rep[2];
+                opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur-mlen].rep[0] : opt[cur-mlen].rep[1];
+                assert(!(opt[cur].off == ZSTD_REP_MOVE_OPT && mlen == 1));
+                opt[cur].rep[0] = (opt[cur].off == ZSTD_REP_MOVE_OPT) ? (opt[cur-mlen].rep[0] - 1) : (opt[cur-mlen].rep[opt[cur].off]);
+            }
+
+            best_mlen = minMatch;
+            {   U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);
+                for (i = (mlen != 1); i<last_i; i++) {
+                    const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];
+                    const U32 repIndex = (U32)(current+cur - repCur);
+                    const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
+                    const BYTE* const repMatch = repBase + repIndex;
+                    if ( (repCur > 0 && repCur <= (S32)(current+cur))
+                      && (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex>lowestIndex))  /* intentional overflow */
+                      && (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
+                        /* repcode detected */
+                        const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
+                        mlen = (U32)ZSTD_count_2segments(inr+minMatch, repMatch+minMatch, iend, repEnd, prefixStart) + minMatch;
+
+                        if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {
+                            best_mlen = mlen; best_off = i; last_pos = cur + 1;
+                            goto _storeSequence;
+                        }
+
+                        best_off = i - (opt[cur].mlen != 1);
+                        if (mlen > best_mlen) best_mlen = mlen;
+
+                        do {
+                            if (opt[cur].mlen == 1) {
+                                litlen = opt[cur].litlen;
+                                if (cur > litlen) {
+                                    price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, inr-litlen, best_off, mlen - MINMATCH, ultra);
+                                } else
+                                    price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
+                            } else {
+                                litlen = 0;
+                                price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);
+                            }
+
+                            if (cur + mlen > last_pos || price <= opt[cur + mlen].price)
+                                SET_PRICE(cur + mlen, mlen, i, litlen, price);
+                            mlen--;
+                        } while (mlen >= minMatch);
+            }   }   }
+
+            match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, inr, iend, maxSearches, mls, matches, minMatch);
+
+            if (match_num > 0 && (matches[match_num-1].len > sufficient_len || cur + matches[match_num-1].len >= ZSTD_OPT_NUM)) {
+                best_mlen = matches[match_num-1].len;
+                best_off = matches[match_num-1].off;
+                last_pos = cur + 1;
+                goto _storeSequence;
+            }
+
+            /* set prices using matches at position = cur */
+            for (u = 0; u < match_num; u++) {
+                mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
+                best_mlen = matches[u].len;
+
+                while (mlen <= best_mlen) {
+                    if (opt[cur].mlen == 1) {
+                        litlen = opt[cur].litlen;
+                        if (cur > litlen)
+                            price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, ip+cur-litlen, matches[u].off-1, mlen - MINMATCH, ultra);
+                        else
+                            price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
+                    } else {
+                        litlen = 0;
+                        price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, matches[u].off-1, mlen - MINMATCH, ultra);
+                    }
+
+                    if (cur + mlen > last_pos || (price < opt[cur + mlen].price))
+                        SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);
+
+                    mlen++;
+        }   }   }   /* for (cur = 1; cur <= last_pos; cur++) */
+
+        best_mlen = opt[last_pos].mlen;
+        best_off = opt[last_pos].off;
+        cur = last_pos - best_mlen;
+
+        /* store sequence */
+_storeSequence:   /* cur, last_pos, best_mlen, best_off have to be set */
+        opt[0].mlen = 1;
+
+        while (1) {
+            mlen = opt[cur].mlen;
+            offset = opt[cur].off;
+            opt[cur].mlen = best_mlen;
+            opt[cur].off = best_off;
+            best_mlen = mlen;
+            best_off = offset;
+            if (mlen > cur) break;
+            cur -= mlen;
+        }
+
+        for (u = 0; u <= last_pos; ) {
+            u += opt[u].mlen;
+        }
+
+        for (cur=0; cur < last_pos; ) {
+            mlen = opt[cur].mlen;
+            if (mlen == 1) { ip++; cur++; continue; }
+            offset = opt[cur].off;
+            cur += mlen;
+            litLength = (U32)(ip - anchor);
+
+            if (offset > ZSTD_REP_MOVE_OPT) {
+                rep[2] = rep[1];
+                rep[1] = rep[0];
+                rep[0] = offset - ZSTD_REP_MOVE_OPT;
+                offset--;
+            } else {
+                if (offset != 0) {
+                    best_off = (offset==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);
+                    if (offset != 1) rep[2] = rep[1];
+                    rep[1] = rep[0];
+                    rep[0] = best_off;
+                }
+
+                if (litLength==0) offset--;
+            }
+
+            ZSTD_updatePrice(optStatePtr, litLength, anchor, offset, mlen-MINMATCH);
+            ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);
+            anchor = ip = ip + mlen;
+    }    }   /* for (cur=0; cur < last_pos; ) */
+
+    /* Save reps for next block */
+    { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqStorePtr->repToConfirm[i] = rep[i]; }
+
+    /* Return the last literals size */
+    return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_btopt_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 0);
+}
+
+size_t ZSTD_compressBlock_btultra_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 1);
+}
diff --git a/thirdparty/zstd/compress/zstd_opt.h b/thirdparty/zstd/compress/zstd_opt.h
index ae24732c7d..816a1fabbf 100644
--- a/thirdparty/zstd/compress/zstd_opt.h
+++ b/thirdparty/zstd/compress/zstd_opt.h
@@ -5,934 +5,26 @@
  * 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_OPT_H
+#define ZSTD_OPT_H
 
-/* Note : this file is intended to be included within zstd_compress.c */
+#include "zstd_compress.h"
 
+#if defined (__cplusplus)
+extern "C" {
+#endif
 
-#ifndef ZSTD_OPT_H_91842398743
-#define ZSTD_OPT_H_91842398743
+size_t ZSTD_compressBlock_btopt(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
 
+size_t ZSTD_compressBlock_btopt_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize);
 
-#define ZSTD_LITFREQ_ADD    2
-#define ZSTD_FREQ_DIV       4
-#define ZSTD_MAX_PRICE      (1<<30)
-
-/*-*************************************
-*  Price functions for optimal parser
-***************************************/
-static void ZSTD_setLog2Prices(optState_t* optPtr)
-{
-    optPtr->log2matchLengthSum = ZSTD_highbit32(optPtr->matchLengthSum+1);
-    optPtr->log2litLengthSum = ZSTD_highbit32(optPtr->litLengthSum+1);
-    optPtr->log2litSum = ZSTD_highbit32(optPtr->litSum+1);
-    optPtr->log2offCodeSum = ZSTD_highbit32(optPtr->offCodeSum+1);
-    optPtr->factor = 1 + ((optPtr->litSum>>5) / optPtr->litLengthSum) + ((optPtr->litSum<<1) / (optPtr->litSum + optPtr->matchSum));
-}
-
-
-static void ZSTD_rescaleFreqs(optState_t* optPtr, const BYTE* src, size_t srcSize)
-{
-    unsigned u;
-
-    optPtr->cachedLiterals = NULL;
-    optPtr->cachedPrice = optPtr->cachedLitLength = 0;
-    optPtr->staticPrices = 0;
-
-    if (optPtr->litLengthSum == 0) {
-        if (srcSize <= 1024) optPtr->staticPrices = 1;
-
-        assert(optPtr->litFreq!=NULL);
-        for (u=0; u<=MaxLit; u++)
-            optPtr->litFreq[u] = 0;
-        for (u=0; u<srcSize; u++)
-            optPtr->litFreq[src[u]]++;
-
-        optPtr->litSum = 0;
-        optPtr->litLengthSum = MaxLL+1;
-        optPtr->matchLengthSum = MaxML+1;
-        optPtr->offCodeSum = (MaxOff+1);
-        optPtr->matchSum = (ZSTD_LITFREQ_ADD<<Litbits);
-
-        for (u=0; u<=MaxLit; u++) {
-            optPtr->litFreq[u] = 1 + (optPtr->litFreq[u]>>ZSTD_FREQ_DIV);
-            optPtr->litSum += optPtr->litFreq[u];
-        }
-        for (u=0; u<=MaxLL; u++)
-            optPtr->litLengthFreq[u] = 1;
-        for (u=0; u<=MaxML; u++)
-            optPtr->matchLengthFreq[u] = 1;
-        for (u=0; u<=MaxOff; u++)
-            optPtr->offCodeFreq[u] = 1;
-    } else {
-        optPtr->matchLengthSum = 0;
-        optPtr->litLengthSum = 0;
-        optPtr->offCodeSum = 0;
-        optPtr->matchSum = 0;
-        optPtr->litSum = 0;
-
-        for (u=0; u<=MaxLit; u++) {
-            optPtr->litFreq[u] = 1 + (optPtr->litFreq[u]>>(ZSTD_FREQ_DIV+1));
-            optPtr->litSum += optPtr->litFreq[u];
-        }
-        for (u=0; u<=MaxLL; u++) {
-            optPtr->litLengthFreq[u] = 1 + (optPtr->litLengthFreq[u]>>(ZSTD_FREQ_DIV+1));
-            optPtr->litLengthSum += optPtr->litLengthFreq[u];
-        }
-        for (u=0; u<=MaxML; u++) {
-            optPtr->matchLengthFreq[u] = 1 + (optPtr->matchLengthFreq[u]>>ZSTD_FREQ_DIV);
-            optPtr->matchLengthSum += optPtr->matchLengthFreq[u];
-            optPtr->matchSum += optPtr->matchLengthFreq[u] * (u + 3);
-        }
-        optPtr->matchSum *= ZSTD_LITFREQ_ADD;
-        for (u=0; u<=MaxOff; u++) {
-            optPtr->offCodeFreq[u] = 1 + (optPtr->offCodeFreq[u]>>ZSTD_FREQ_DIV);
-            optPtr->offCodeSum += optPtr->offCodeFreq[u];
-        }
-    }
-
-    ZSTD_setLog2Prices(optPtr);
-}
-
-
-static U32 ZSTD_getLiteralPrice(optState_t* optPtr, U32 litLength, const BYTE* literals)
-{
-    U32 price, u;
-
-    if (optPtr->staticPrices)
-        return ZSTD_highbit32((U32)litLength+1) + (litLength*6);
-
-    if (litLength == 0)
-        return optPtr->log2litLengthSum - ZSTD_highbit32(optPtr->litLengthFreq[0]+1);
-
-    /* literals */
-    if (optPtr->cachedLiterals == literals) {
-        U32 const additional = litLength - optPtr->cachedLitLength;
-        const BYTE* literals2 = optPtr->cachedLiterals + optPtr->cachedLitLength;
-        price = optPtr->cachedPrice + additional * optPtr->log2litSum;
-        for (u=0; u < additional; u++)
-            price -= ZSTD_highbit32(optPtr->litFreq[literals2[u]]+1);
-        optPtr->cachedPrice = price;
-        optPtr->cachedLitLength = litLength;
-    } else {
-        price = litLength * optPtr->log2litSum;
-        for (u=0; u < litLength; u++)
-            price -= ZSTD_highbit32(optPtr->litFreq[literals[u]]+1);
-
-        if (litLength >= 12) {
-            optPtr->cachedLiterals = literals;
-            optPtr->cachedPrice = price;
-            optPtr->cachedLitLength = litLength;
-        }
-    }
-
-    /* literal Length */
-    {   const BYTE LL_deltaCode = 19;
-        const BYTE llCode = (litLength>63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
-        price += LL_bits[llCode] + optPtr->log2litLengthSum - ZSTD_highbit32(optPtr->litLengthFreq[llCode]+1);
-    }
-
-    return price;
-}
-
-
-FORCE_INLINE_TEMPLATE U32 ZSTD_getPrice(optState_t* optPtr, U32 litLength, const BYTE* literals, U32 offset, U32 matchLength, const int ultra)
-{
-    /* offset */
-    U32 price;
-    BYTE const offCode = (BYTE)ZSTD_highbit32(offset+1);
-
-    if (optPtr->staticPrices)
-        return ZSTD_getLiteralPrice(optPtr, litLength, literals) + ZSTD_highbit32((U32)matchLength+1) + 16 + offCode;
-
-    price = offCode + optPtr->log2offCodeSum - ZSTD_highbit32(optPtr->offCodeFreq[offCode]+1);
-    if (!ultra && offCode >= 20) price += (offCode-19)*2;
-
-    /* match Length */
-    {   const BYTE ML_deltaCode = 36;
-        const BYTE mlCode = (matchLength>127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];
-        price += ML_bits[mlCode] + optPtr->log2matchLengthSum - ZSTD_highbit32(optPtr->matchLengthFreq[mlCode]+1);
-    }
-
-    return price + ZSTD_getLiteralPrice(optPtr, litLength, literals) + optPtr->factor;
-}
-
-
-static void ZSTD_updatePrice(optState_t* optPtr, U32 litLength, const BYTE* literals, U32 offset, U32 matchLength)
-{
-    U32 u;
-
-    /* literals */
-    optPtr->litSum += litLength*ZSTD_LITFREQ_ADD;
-    for (u=0; u < litLength; u++)
-        optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;
-
-    /* literal Length */
-    {   const BYTE LL_deltaCode = 19;
-        const BYTE llCode = (litLength>63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
-        optPtr->litLengthFreq[llCode]++;
-        optPtr->litLengthSum++;
-    }
-
-    /* match offset */
-    {   BYTE const offCode = (BYTE)ZSTD_highbit32(offset+1);
-        optPtr->offCodeSum++;
-        optPtr->offCodeFreq[offCode]++;
-    }
-
-    /* match Length */
-    {   const BYTE ML_deltaCode = 36;
-        const BYTE mlCode = (matchLength>127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];
-        optPtr->matchLengthFreq[mlCode]++;
-        optPtr->matchLengthSum++;
-    }
-
-    ZSTD_setLog2Prices(optPtr);
-}
-
-
-#define SET_PRICE(pos, mlen_, offset_, litlen_, price_)   \
-    {                                                 \
-        while (last_pos < pos)  { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; } \
-        opt[pos].mlen = mlen_;                         \
-        opt[pos].off = offset_;                        \
-        opt[pos].litlen = litlen_;                     \
-        opt[pos].price = price_;                       \
-    }
-
-
-/* function safe only for comparisons */
-static U32 ZSTD_readMINMATCH(const void* memPtr, U32 length)
-{
-    switch (length)
-    {
-    default :
-    case 4 : return MEM_read32(memPtr);
-    case 3 : if (MEM_isLittleEndian())
-                return MEM_read32(memPtr)<<8;
-             else
-                return MEM_read32(memPtr)>>8;
-    }
-}
-
-
-/* Update hashTable3 up to ip (excluded)
-   Assumption : always within prefix (i.e. not within extDict) */
-static
-U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_CCtx* zc, const BYTE* ip)
-{
-    U32* const hashTable3  = zc->hashTable3;
-    U32 const hashLog3  = zc->hashLog3;
-    const BYTE* const base = zc->base;
-    U32 idx = zc->nextToUpdate3;
-    const U32 target = zc->nextToUpdate3 = (U32)(ip - base);
-    const size_t hash3 = ZSTD_hash3Ptr(ip, hashLog3);
-
-    while(idx < target) {
-        hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx;
-        idx++;
-    }
-
-    return hashTable3[hash3];
-}
-
-
-/*-*************************************
-*  Binary Tree search
-***************************************/
-static U32 ZSTD_insertBtAndGetAllMatches (
-                        ZSTD_CCtx* zc,
-                        const BYTE* const ip, const BYTE* const iLimit,
-                        U32 nbCompares, const U32 mls,
-                        U32 extDict, ZSTD_match_t* matches, const U32 minMatchLen)
-{
-    const BYTE* const base = zc->base;
-    const U32 current = (U32)(ip-base);
-    const U32 hashLog = zc->appliedParams.cParams.hashLog;
-    const size_t h  = ZSTD_hashPtr(ip, hashLog, mls);
-    U32* const hashTable = zc->hashTable;
-    U32 matchIndex  = hashTable[h];
-    U32* const bt   = zc->chainTable;
-    const U32 btLog = zc->appliedParams.cParams.chainLog - 1;
-    const U32 btMask= (1U << btLog) - 1;
-    size_t commonLengthSmaller=0, commonLengthLarger=0;
-    const BYTE* const dictBase = zc->dictBase;
-    const U32 dictLimit = zc->dictLimit;
-    const BYTE* const dictEnd = dictBase + dictLimit;
-    const BYTE* const prefixStart = base + dictLimit;
-    const U32 btLow = btMask >= current ? 0 : current - btMask;
-    const U32 windowLow = zc->lowLimit;
-    U32* smallerPtr = bt + 2*(current&btMask);
-    U32* largerPtr  = bt + 2*(current&btMask) + 1;
-    U32 matchEndIdx = current+8;
-    U32 dummy32;   /* to be nullified at the end */
-    U32 mnum = 0;
-
-    const U32 minMatch = (mls == 3) ? 3 : 4;
-    size_t bestLength = minMatchLen-1;
-
-    if (minMatch == 3) { /* HC3 match finder */
-        U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3 (zc, ip);
-        if (matchIndex3>windowLow && (current - matchIndex3 < (1<<18))) {
-            const BYTE* match;
-            size_t currentMl=0;
-            if ((!extDict) || matchIndex3 >= dictLimit) {
-                match = base + matchIndex3;
-                if (match[bestLength] == ip[bestLength]) currentMl = ZSTD_count(ip, match, iLimit);
-            } else {
-                match = dictBase + matchIndex3;
-                if (ZSTD_readMINMATCH(match, MINMATCH) == ZSTD_readMINMATCH(ip, MINMATCH))    /* assumption : matchIndex3 <= dictLimit-4 (by table construction) */
-                    currentMl = ZSTD_count_2segments(ip+MINMATCH, match+MINMATCH, iLimit, dictEnd, prefixStart) + MINMATCH;
-            }
-
-            /* save best solution */
-            if (currentMl > bestLength) {
-                bestLength = currentMl;
-                matches[mnum].off = ZSTD_REP_MOVE_OPT + current - matchIndex3;
-                matches[mnum].len = (U32)currentMl;
-                mnum++;
-                if (currentMl > ZSTD_OPT_NUM) goto update;
-                if (ip+currentMl == iLimit) goto update; /* best possible, and avoid read overflow*/
-            }
-        }
-    }
-
-    hashTable[h] = current;   /* Update Hash Table */
-
-    while (nbCompares-- && (matchIndex > windowLow)) {
-        U32* nextPtr = bt + 2*(matchIndex & btMask);
-        size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
-        const BYTE* match;
-
-        if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
-            match = base + matchIndex;
-            if (match[matchLength] == ip[matchLength]) {
-                matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iLimit) +1;
-            }
-        } else {
-            match = dictBase + matchIndex;
-            matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);
-            if (matchIndex+matchLength >= dictLimit)
-                match = base + matchIndex;   /* to prepare for next usage of match[matchLength] */
-        }
-
-        if (matchLength > bestLength) {
-            if (matchLength > matchEndIdx - matchIndex) matchEndIdx = matchIndex + (U32)matchLength;
-            bestLength = matchLength;
-            matches[mnum].off = ZSTD_REP_MOVE_OPT + current - matchIndex;
-            matches[mnum].len = (U32)matchLength;
-            mnum++;
-            if (matchLength > ZSTD_OPT_NUM) break;
-            if (ip+matchLength == iLimit)   /* equal : no way to know if inf or sup */
-                break;   /* drop, to guarantee consistency (miss a little bit of compression) */
-        }
-
-        if (match[matchLength] < ip[matchLength]) {
-            /* match is smaller than current */
-            *smallerPtr = matchIndex;             /* update smaller idx */
-            commonLengthSmaller = matchLength;    /* all smaller will now have at least this guaranteed common length */
-            if (matchIndex <= btLow) { smallerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            smallerPtr = nextPtr+1;               /* new "smaller" => larger of match */
-            matchIndex = nextPtr[1];              /* new matchIndex larger than previous (closer to current) */
-        } else {
-            /* match is larger than current */
-            *largerPtr = matchIndex;
-            commonLengthLarger = matchLength;
-            if (matchIndex <= btLow) { largerPtr=&dummy32; break; }   /* beyond tree size, stop the search */
-            largerPtr = nextPtr;
-            matchIndex = nextPtr[0];
-    }   }
-
-    *smallerPtr = *largerPtr = 0;
-
-update:
-    zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;
-    return mnum;
-}
-
-
-/** Tree updater, providing best match */
-static U32 ZSTD_BtGetAllMatches (
-                        ZSTD_CCtx* zc,
-                        const BYTE* const ip, const BYTE* const iLimit,
-                        const U32 maxNbAttempts, const U32 mls, ZSTD_match_t* matches, const U32 minMatchLen)
-{
-    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
-    ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
-    return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 0, matches, minMatchLen);
-}
-
-
-static U32 ZSTD_BtGetAllMatches_selectMLS (
-                        ZSTD_CCtx* zc,   /* Index table will be updated */
-                        const BYTE* ip, const BYTE* const iHighLimit,
-                        const U32 maxNbAttempts, const U32 matchLengthSearch, ZSTD_match_t* matches, const U32 minMatchLen)
-{
-    switch(matchLengthSearch)
-    {
-    case 3 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);
-    default :
-    case 4 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);
-    case 5 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);
-    case 7 :
-    case 6 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);
-    }
-}
-
-/** Tree updater, providing best match */
-static U32 ZSTD_BtGetAllMatches_extDict (
-                        ZSTD_CCtx* zc,
-                        const BYTE* const ip, const BYTE* const iLimit,
-                        const U32 maxNbAttempts, const U32 mls, ZSTD_match_t* matches, const U32 minMatchLen)
-{
-    if (ip < zc->base + zc->nextToUpdate) return 0;   /* skipped area */
-    ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
-    return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 1, matches, minMatchLen);
-}
-
-
-static U32 ZSTD_BtGetAllMatches_selectMLS_extDict (
-                        ZSTD_CCtx* zc,   /* Index table will be updated */
-                        const BYTE* ip, const BYTE* const iHighLimit,
-                        const U32 maxNbAttempts, const U32 matchLengthSearch, ZSTD_match_t* matches, const U32 minMatchLen)
-{
-    switch(matchLengthSearch)
-    {
-    case 3 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);
-    default :
-    case 4 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);
-    case 5 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);
-    case 7 :
-    case 6 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);
-    }
-}
-
-
-/*-*******************************
-*  Optimal parser
-*********************************/
-FORCE_INLINE_TEMPLATE
-void ZSTD_compressBlock_opt_generic(ZSTD_CCtx* ctx,
-                                    const void* src, size_t srcSize, const int ultra)
-{
-    seqStore_t* seqStorePtr = &(ctx->seqStore);
-    optState_t* optStatePtr = &(ctx->optState);
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - 8;
-    const BYTE* const base = ctx->base;
-    const BYTE* const prefixStart = base + ctx->dictLimit;
-
-    const U32 maxSearches = 1U << ctx->appliedParams.cParams.searchLog;
-    const U32 sufficient_len = ctx->appliedParams.cParams.targetLength;
-    const U32 mls = ctx->appliedParams.cParams.searchLength;
-    const U32 minMatch = (ctx->appliedParams.cParams.searchLength == 3) ? 3 : 4;
-
-    ZSTD_optimal_t* opt = optStatePtr->priceTable;
-    ZSTD_match_t* matches = optStatePtr->matchTable;
-    const BYTE* inr;
-    U32 offset, rep[ZSTD_REP_NUM];
-
-    /* init */
-    ctx->nextToUpdate3 = ctx->nextToUpdate;
-    ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize);
-    ip += (ip==prefixStart);
-    { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) rep[i]=seqStorePtr->rep[i]; }
-
-    /* Match Loop */
-    while (ip < ilimit) {
-        U32 cur, match_num, last_pos, litlen, price;
-        U32 u, mlen, best_mlen, best_off, litLength;
-        memset(opt, 0, sizeof(ZSTD_optimal_t));
-        last_pos = 0;
-        litlen = (U32)(ip - anchor);
-
-        /* check repCode */
-        {   U32 i, last_i = ZSTD_REP_CHECK + (ip==anchor);
-            for (i=(ip == anchor); i<last_i; i++) {
-                const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];
-                if ( (repCur > 0) && (repCur < (S32)(ip-prefixStart))
-                    && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(ip - repCur, minMatch))) {
-                    mlen = (U32)ZSTD_count(ip+minMatch, ip+minMatch-repCur, iend) + minMatch;
-                    if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {
-                        best_mlen = mlen; best_off = i; cur = 0; last_pos = 1;
-                        goto _storeSequence;
-                    }
-                    best_off = i - (ip == anchor);
-                    do {
-                        price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-                        if (mlen > last_pos || price < opt[mlen].price)
-                            SET_PRICE(mlen, mlen, i, litlen, price);   /* note : macro modifies last_pos */
-                        mlen--;
-                    } while (mlen >= minMatch);
-        }   }   }
-
-        match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, ip, iend, maxSearches, mls, matches, minMatch);
-
-        if (!last_pos && !match_num) { ip++; continue; }
-
-        if (match_num && (matches[match_num-1].len > sufficient_len || matches[match_num-1].len >= ZSTD_OPT_NUM)) {
-            best_mlen = matches[match_num-1].len;
-            best_off = matches[match_num-1].off;
-            cur = 0;
-            last_pos = 1;
-            goto _storeSequence;
-        }
-
-        /* set prices using matches at position = 0 */
-        best_mlen = (last_pos) ? last_pos : minMatch;
-        for (u = 0; u < match_num; u++) {
-            mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
-            best_mlen = matches[u].len;
-            while (mlen <= best_mlen) {
-                price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
-                if (mlen > last_pos || price < opt[mlen].price)
-                    SET_PRICE(mlen, mlen, matches[u].off, litlen, price);   /* note : macro modifies last_pos */
-                mlen++;
-        }   }
-
-        if (last_pos < minMatch) { ip++; continue; }
-
-        /* initialize opt[0] */
-        { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
-        opt[0].mlen = 1;
-        opt[0].litlen = litlen;
-
-         /* check further positions */
-        for (cur = 1; cur <= last_pos; cur++) {
-           inr = ip + cur;
-
-           if (opt[cur-1].mlen == 1) {
-                litlen = opt[cur-1].litlen + 1;
-                if (cur > litlen) {
-                    price = opt[cur - litlen].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-litlen);
-                } else
-                    price = ZSTD_getLiteralPrice(optStatePtr, litlen, anchor);
-           } else {
-                litlen = 1;
-                price = opt[cur - 1].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-1);
-           }
-
-           if (cur > last_pos || price <= opt[cur].price)
-                SET_PRICE(cur, 1, 0, litlen, price);
-
-           if (cur == last_pos) break;
-
-           if (inr > ilimit)  /* last match must start at a minimum distance of 8 from oend */
-               continue;
-
-           mlen = opt[cur].mlen;
-           if (opt[cur].off > ZSTD_REP_MOVE_OPT) {
-                opt[cur].rep[2] = opt[cur-mlen].rep[1];
-                opt[cur].rep[1] = opt[cur-mlen].rep[0];
-                opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;
-           } else {
-                opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur-mlen].rep[1] : opt[cur-mlen].rep[2];
-                opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur-mlen].rep[0] : opt[cur-mlen].rep[1];
-                opt[cur].rep[0] = ((opt[cur].off==ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur-mlen].rep[0] - 1) : (opt[cur-mlen].rep[opt[cur].off]);
-           }
-
-            best_mlen = minMatch;
-            {   U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);
-                for (i=(opt[cur].mlen != 1); i<last_i; i++) {  /* check rep */
-                    const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];
-                    if ( (repCur > 0) && (repCur < (S32)(inr-prefixStart))
-                       && (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(inr - repCur, minMatch))) {
-                       mlen = (U32)ZSTD_count(inr+minMatch, inr+minMatch - repCur, iend) + minMatch;
-
-                       if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {
-                            best_mlen = mlen; best_off = i; last_pos = cur + 1;
-                            goto _storeSequence;
-                       }
-
-                       best_off = i - (opt[cur].mlen != 1);
-                       if (mlen > best_mlen) best_mlen = mlen;
-
-                       do {
-                           if (opt[cur].mlen == 1) {
-                                litlen = opt[cur].litlen;
-                                if (cur > litlen) {
-                                    price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, inr-litlen, best_off, mlen - MINMATCH, ultra);
-                                } else
-                                    price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-                            } else {
-                                litlen = 0;
-                                price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);
-                            }
-
-                            if (cur + mlen > last_pos || price <= opt[cur + mlen].price)
-                                SET_PRICE(cur + mlen, mlen, i, litlen, price);
-                            mlen--;
-                        } while (mlen >= minMatch);
-            }   }   }
-
-            match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, inr, iend, maxSearches, mls, matches, best_mlen);
-
-            if (match_num > 0 && (matches[match_num-1].len > sufficient_len || cur + matches[match_num-1].len >= ZSTD_OPT_NUM)) {
-                best_mlen = matches[match_num-1].len;
-                best_off = matches[match_num-1].off;
-                last_pos = cur + 1;
-                goto _storeSequence;
-            }
-
-            /* set prices using matches at position = cur */
-            for (u = 0; u < match_num; u++) {
-                mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
-                best_mlen = matches[u].len;
-
-                while (mlen <= best_mlen) {
-                    if (opt[cur].mlen == 1) {
-                        litlen = opt[cur].litlen;
-                        if (cur > litlen)
-                            price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, ip+cur-litlen, matches[u].off-1, mlen - MINMATCH, ultra);
-                        else
-                            price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
-                    } else {
-                        litlen = 0;
-                        price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, matches[u].off-1, mlen - MINMATCH, ultra);
-                    }
-
-                    if (cur + mlen > last_pos || (price < opt[cur + mlen].price))
-                        SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);
-
-                    mlen++;
-        }   }   }
-
-        best_mlen = opt[last_pos].mlen;
-        best_off = opt[last_pos].off;
-        cur = last_pos - best_mlen;
-
-        /* store sequence */
-_storeSequence:   /* cur, last_pos, best_mlen, best_off have to be set */
-        opt[0].mlen = 1;
-
-        while (1) {
-            mlen = opt[cur].mlen;
-            offset = opt[cur].off;
-            opt[cur].mlen = best_mlen;
-            opt[cur].off = best_off;
-            best_mlen = mlen;
-            best_off = offset;
-            if (mlen > cur) break;
-            cur -= mlen;
-        }
-
-        for (u = 0; u <= last_pos;) {
-            u += opt[u].mlen;
-        }
-
-        for (cur=0; cur < last_pos; ) {
-            mlen = opt[cur].mlen;
-            if (mlen == 1) { ip++; cur++; continue; }
-            offset = opt[cur].off;
-            cur += mlen;
-            litLength = (U32)(ip - anchor);
-
-            if (offset > ZSTD_REP_MOVE_OPT) {
-                rep[2] = rep[1];
-                rep[1] = rep[0];
-                rep[0] = offset - ZSTD_REP_MOVE_OPT;
-                offset--;
-            } else {
-                if (offset != 0) {
-                    best_off = (offset==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);
-                    if (offset != 1) rep[2] = rep[1];
-                    rep[1] = rep[0];
-                    rep[0] = best_off;
-                }
-                if (litLength==0) offset--;
-            }
-
-            ZSTD_updatePrice(optStatePtr, litLength, anchor, offset, mlen-MINMATCH);
-            ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);
-            anchor = ip = ip + mlen;
-    }    }   /* for (cur=0; cur < last_pos; ) */
-
-    /* Save reps for next block */
-    { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqStorePtr->repToConfirm[i] = rep[i]; }
-
-    /* Last Literals */
-    {   size_t const lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
-}
-
-
-FORCE_INLINE_TEMPLATE
-void ZSTD_compressBlock_opt_extDict_generic(ZSTD_CCtx* ctx,
-                                     const void* src, size_t srcSize, const int ultra)
-{
-    seqStore_t* seqStorePtr = &(ctx->seqStore);
-    optState_t* optStatePtr = &(ctx->optState);
-    const BYTE* const istart = (const BYTE*)src;
-    const BYTE* ip = istart;
-    const BYTE* anchor = istart;
-    const BYTE* const iend = istart + srcSize;
-    const BYTE* const ilimit = iend - 8;
-    const BYTE* const base = ctx->base;
-    const U32 lowestIndex = ctx->lowLimit;
-    const U32 dictLimit = ctx->dictLimit;
-    const BYTE* const prefixStart = base + dictLimit;
-    const BYTE* const dictBase = ctx->dictBase;
-    const BYTE* const dictEnd  = dictBase + dictLimit;
-
-    const U32 maxSearches = 1U << ctx->appliedParams.cParams.searchLog;
-    const U32 sufficient_len = ctx->appliedParams.cParams.targetLength;
-    const U32 mls = ctx->appliedParams.cParams.searchLength;
-    const U32 minMatch = (ctx->appliedParams.cParams.searchLength == 3) ? 3 : 4;
-
-    ZSTD_optimal_t* opt = optStatePtr->priceTable;
-    ZSTD_match_t* matches = optStatePtr->matchTable;
-    const BYTE* inr;
-
-    /* init */
-    U32 offset, rep[ZSTD_REP_NUM];
-    { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) rep[i]=seqStorePtr->rep[i]; }
-
-    ctx->nextToUpdate3 = ctx->nextToUpdate;
-    ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize);
-    ip += (ip==prefixStart);
-
-    /* Match Loop */
-    while (ip < ilimit) {
-        U32 cur, match_num, last_pos, litlen, price;
-        U32 u, mlen, best_mlen, best_off, litLength;
-        U32 current = (U32)(ip-base);
-        memset(opt, 0, sizeof(ZSTD_optimal_t));
-        last_pos = 0;
-        opt[0].litlen = (U32)(ip - anchor);
-
-        /* check repCode */
-        {   U32 i, last_i = ZSTD_REP_CHECK + (ip==anchor);
-            for (i = (ip==anchor); i<last_i; i++) {
-                const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];
-                const U32 repIndex = (U32)(current - repCur);
-                const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
-                const BYTE* const repMatch = repBase + repIndex;
-                if ( (repCur > 0 && repCur <= (S32)current)
-                   && (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex>lowestIndex))  /* intentional overflow */
-                   && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
-                    /* repcode detected we should take it */
-                    const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
-                    mlen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iend, repEnd, prefixStart) + minMatch;
-
-                    if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {
-                        best_mlen = mlen; best_off = i; cur = 0; last_pos = 1;
-                        goto _storeSequence;
-                    }
-
-                    best_off = i - (ip==anchor);
-                    litlen = opt[0].litlen;
-                    do {
-                        price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-                        if (mlen > last_pos || price < opt[mlen].price)
-                            SET_PRICE(mlen, mlen, i, litlen, price);   /* note : macro modifies last_pos */
-                        mlen--;
-                    } while (mlen >= minMatch);
-        }   }   }
-
-        match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, ip, iend, maxSearches, mls, matches, minMatch);  /* first search (depth 0) */
-
-        if (!last_pos && !match_num) { ip++; continue; }
-
-        { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
-        opt[0].mlen = 1;
-
-        if (match_num && (matches[match_num-1].len > sufficient_len || matches[match_num-1].len >= ZSTD_OPT_NUM)) {
-            best_mlen = matches[match_num-1].len;
-            best_off = matches[match_num-1].off;
-            cur = 0;
-            last_pos = 1;
-            goto _storeSequence;
-        }
-
-        best_mlen = (last_pos) ? last_pos : minMatch;
-
-        /* set prices using matches at position = 0 */
-        for (u = 0; u < match_num; u++) {
-            mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
-            best_mlen = matches[u].len;
-            litlen = opt[0].litlen;
-            while (mlen <= best_mlen) {
-                price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
-                if (mlen > last_pos || price < opt[mlen].price)
-                    SET_PRICE(mlen, mlen, matches[u].off, litlen, price);
-                mlen++;
-        }   }
-
-        if (last_pos < minMatch) {
-            ip++; continue;
-        }
-
-        /* check further positions */
-        for (cur = 1; cur <= last_pos; cur++) {
-            inr = ip + cur;
-
-            if (opt[cur-1].mlen == 1) {
-                litlen = opt[cur-1].litlen + 1;
-                if (cur > litlen) {
-                    price = opt[cur - litlen].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-litlen);
-                } else
-                    price = ZSTD_getLiteralPrice(optStatePtr, litlen, anchor);
-            } else {
-                litlen = 1;
-                price = opt[cur - 1].price + ZSTD_getLiteralPrice(optStatePtr, litlen, inr-1);
-            }
-
-            if (cur > last_pos || price <= opt[cur].price)
-                SET_PRICE(cur, 1, 0, litlen, price);
-
-            if (cur == last_pos) break;
-
-            if (inr > ilimit)  /* last match must start at a minimum distance of 8 from oend */
-                continue;
-
-            mlen = opt[cur].mlen;
-            if (opt[cur].off > ZSTD_REP_MOVE_OPT) {
-                opt[cur].rep[2] = opt[cur-mlen].rep[1];
-                opt[cur].rep[1] = opt[cur-mlen].rep[0];
-                opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;
-            } else {
-                opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur-mlen].rep[1] : opt[cur-mlen].rep[2];
-                opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur-mlen].rep[0] : opt[cur-mlen].rep[1];
-                opt[cur].rep[0] = ((opt[cur].off==ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur-mlen].rep[0] - 1) : (opt[cur-mlen].rep[opt[cur].off]);
-            }
-
-            best_mlen = minMatch;
-            {   U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);
-                for (i = (mlen != 1); i<last_i; i++) {
-                    const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];
-                    const U32 repIndex = (U32)(current+cur - repCur);
-                    const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
-                    const BYTE* const repMatch = repBase + repIndex;
-                    if ( (repCur > 0 && repCur <= (S32)(current+cur))
-                      && (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex>lowestIndex))  /* intentional overflow */
-                      && (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
-                        /* repcode detected */
-                        const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
-                        mlen = (U32)ZSTD_count_2segments(inr+minMatch, repMatch+minMatch, iend, repEnd, prefixStart) + minMatch;
-
-                        if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {
-                            best_mlen = mlen; best_off = i; last_pos = cur + 1;
-                            goto _storeSequence;
-                        }
-
-                        best_off = i - (opt[cur].mlen != 1);
-                        if (mlen > best_mlen) best_mlen = mlen;
-
-                        do {
-                            if (opt[cur].mlen == 1) {
-                                litlen = opt[cur].litlen;
-                                if (cur > litlen) {
-                                    price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, inr-litlen, best_off, mlen - MINMATCH, ultra);
-                                } else
-                                    price = ZSTD_getPrice(optStatePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-                            } else {
-                                litlen = 0;
-                                price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);
-                            }
-
-                            if (cur + mlen > last_pos || price <= opt[cur + mlen].price)
-                                SET_PRICE(cur + mlen, mlen, i, litlen, price);
-                            mlen--;
-                        } while (mlen >= minMatch);
-            }   }   }
-
-            match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, inr, iend, maxSearches, mls, matches, minMatch);
-
-            if (match_num > 0 && (matches[match_num-1].len > sufficient_len || cur + matches[match_num-1].len >= ZSTD_OPT_NUM)) {
-                best_mlen = matches[match_num-1].len;
-                best_off = matches[match_num-1].off;
-                last_pos = cur + 1;
-                goto _storeSequence;
-            }
-
-            /* set prices using matches at position = cur */
-            for (u = 0; u < match_num; u++) {
-                mlen = (u>0) ? matches[u-1].len+1 : best_mlen;
-                best_mlen = matches[u].len;
-
-                while (mlen <= best_mlen) {
-                    if (opt[cur].mlen == 1) {
-                        litlen = opt[cur].litlen;
-                        if (cur > litlen)
-                            price = opt[cur - litlen].price + ZSTD_getPrice(optStatePtr, litlen, ip+cur-litlen, matches[u].off-1, mlen - MINMATCH, ultra);
-                        else
-                            price = ZSTD_getPrice(optStatePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);
-                    } else {
-                        litlen = 0;
-                        price = opt[cur].price + ZSTD_getPrice(optStatePtr, 0, NULL, matches[u].off-1, mlen - MINMATCH, ultra);
-                    }
-
-                    if (cur + mlen > last_pos || (price < opt[cur + mlen].price))
-                        SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);
-
-                    mlen++;
-        }   }   }   /* for (cur = 1; cur <= last_pos; cur++) */
-
-        best_mlen = opt[last_pos].mlen;
-        best_off = opt[last_pos].off;
-        cur = last_pos - best_mlen;
-
-        /* store sequence */
-_storeSequence:   /* cur, last_pos, best_mlen, best_off have to be set */
-        opt[0].mlen = 1;
-
-        while (1) {
-            mlen = opt[cur].mlen;
-            offset = opt[cur].off;
-            opt[cur].mlen = best_mlen;
-            opt[cur].off = best_off;
-            best_mlen = mlen;
-            best_off = offset;
-            if (mlen > cur) break;
-            cur -= mlen;
-        }
-
-        for (u = 0; u <= last_pos; ) {
-            u += opt[u].mlen;
-        }
-
-        for (cur=0; cur < last_pos; ) {
-            mlen = opt[cur].mlen;
-            if (mlen == 1) { ip++; cur++; continue; }
-            offset = opt[cur].off;
-            cur += mlen;
-            litLength = (U32)(ip - anchor);
-
-            if (offset > ZSTD_REP_MOVE_OPT) {
-                rep[2] = rep[1];
-                rep[1] = rep[0];
-                rep[0] = offset - ZSTD_REP_MOVE_OPT;
-                offset--;
-            } else {
-                if (offset != 0) {
-                    best_off = (offset==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);
-                    if (offset != 1) rep[2] = rep[1];
-                    rep[1] = rep[0];
-                    rep[0] = best_off;
-                }
-
-                if (litLength==0) offset--;
-            }
-
-            ZSTD_updatePrice(optStatePtr, litLength, anchor, offset, mlen-MINMATCH);
-            ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);
-            anchor = ip = ip + mlen;
-    }    }   /* for (cur=0; cur < last_pos; ) */
-
-    /* Save reps for next block */
-    { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqStorePtr->repToConfirm[i] = rep[i]; }
-
-    /* Last Literals */
-    {   size_t lastLLSize = iend - anchor;
-        memcpy(seqStorePtr->lit, anchor, lastLLSize);
-        seqStorePtr->lit += lastLLSize;
-    }
+#if defined (__cplusplus)
 }
+#endif
 
-#endif  /* ZSTD_OPT_H_91842398743 */
+#endif /* ZSTD_OPT_H */
diff --git a/thirdparty/zstd/compress/zstdmt_compress.c b/thirdparty/zstd/compress/zstdmt_compress.c
index 8564bc4392..7831cd3bd8 100644
--- a/thirdparty/zstd/compress/zstdmt_compress.c
+++ b/thirdparty/zstd/compress/zstdmt_compress.c
@@ -5,11 +5,12 @@
  * 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.
  */
 
 
 /* ======   Tuning parameters   ====== */
-#define ZSTDMT_NBTHREADS_MAX 256
+#define ZSTDMT_NBTHREADS_MAX 200
 #define ZSTDMT_OVERLAPLOG_DEFAULT 6
 
 
@@ -52,22 +53,24 @@ static unsigned long long GetCurrentClockTimeMicroseconds(void)
 }
 
 #define MUTEX_WAIT_TIME_DLEVEL 6
-#define PTHREAD_MUTEX_LOCK(mutex) {               \
-    if (ZSTD_DEBUG>=MUTEX_WAIT_TIME_DLEVEL) {   \
+#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) {          \
+    if (ZSTD_DEBUG >= MUTEX_WAIT_TIME_DLEVEL) {   \
         unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds(); \
-        pthread_mutex_lock(mutex);                \
+        ZSTD_pthread_mutex_lock(mutex);           \
         {   unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \
             unsigned long long const elapsedTime = (afterTime-beforeTime); \
             if (elapsedTime > 1000) {  /* or whatever threshold you like; I'm using 1 millisecond here */ \
                 DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, "Thread took %llu microseconds to acquire mutex %s \n", \
                    elapsedTime, #mutex);          \
         }   }                                     \
-    } else pthread_mutex_lock(mutex);             \
+    } else {                                      \
+        ZSTD_pthread_mutex_lock(mutex);           \
+    }                                             \
 }
 
 #else
 
-#  define PTHREAD_MUTEX_LOCK(m) pthread_mutex_lock(m)
+#  define ZSTD_PTHREAD_MUTEX_LOCK(m) ZSTD_pthread_mutex_lock(m)
 #  define DEBUG_PRINTHEX(l,p,n) {}
 
 #endif
@@ -84,7 +87,7 @@ typedef struct buffer_s {
 static const buffer_t g_nullBuffer = { NULL, 0 };
 
 typedef struct ZSTDMT_bufferPool_s {
-    pthread_mutex_t poolMutex;
+    ZSTD_pthread_mutex_t poolMutex;
     size_t bufferSize;
     unsigned totalBuffers;
     unsigned nbBuffers;
@@ -98,7 +101,7 @@ static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned nbThreads, ZSTD_custo
     ZSTDMT_bufferPool* const bufPool = (ZSTDMT_bufferPool*)ZSTD_calloc(
         sizeof(ZSTDMT_bufferPool) + (maxNbBuffers-1) * sizeof(buffer_t), cMem);
     if (bufPool==NULL) return NULL;
-    if (pthread_mutex_init(&bufPool->poolMutex, NULL)) {
+    if (ZSTD_pthread_mutex_init(&bufPool->poolMutex, NULL)) {
         ZSTD_free(bufPool, cMem);
         return NULL;
     }
@@ -112,10 +115,13 @@ static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned nbThreads, ZSTD_custo
 static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool)
 {
     unsigned u;
+    DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool);
     if (!bufPool) return;   /* compatibility with free on NULL */
-    for (u=0; u<bufPool->totalBuffers; u++)
+    for (u=0; u<bufPool->totalBuffers; u++) {
+        DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->bTable[u].start);
         ZSTD_free(bufPool->bTable[u].start, bufPool->cMem);
-    pthread_mutex_destroy(&bufPool->poolMutex);
+    }
+    ZSTD_pthread_mutex_destroy(&bufPool->poolMutex);
     ZSTD_free(bufPool, bufPool->cMem);
 }
 
@@ -126,10 +132,10 @@ static size_t ZSTDMT_sizeof_bufferPool(ZSTDMT_bufferPool* bufPool)
                             + (bufPool->totalBuffers - 1) * sizeof(buffer_t);
     unsigned u;
     size_t totalBufferSize = 0;
-    pthread_mutex_lock(&bufPool->poolMutex);
+    ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
     for (u=0; u<bufPool->totalBuffers; u++)
         totalBufferSize += bufPool->bTable[u].size;
-    pthread_mutex_unlock(&bufPool->poolMutex);
+    ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
 
     return poolSize + totalBufferSize;
 }
@@ -145,20 +151,21 @@ static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)
 {
     size_t const bSize = bufPool->bufferSize;
     DEBUGLOG(5, "ZSTDMT_getBuffer");
-    pthread_mutex_lock(&bufPool->poolMutex);
+    ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
     if (bufPool->nbBuffers) {   /* try to use an existing buffer */
         buffer_t const buf = bufPool->bTable[--(bufPool->nbBuffers)];
         size_t const availBufferSize = buf.size;
+        bufPool->bTable[bufPool->nbBuffers] = g_nullBuffer;
         if ((availBufferSize >= bSize) & (availBufferSize <= 10*bSize)) {
             /* large enough, but not too much */
-            pthread_mutex_unlock(&bufPool->poolMutex);
+            ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
             return buf;
         }
         /* size conditions not respected : scratch this buffer, create new one */
         DEBUGLOG(5, "existing buffer does not meet size conditions => freeing");
         ZSTD_free(buf.start, bufPool->cMem);
     }
-    pthread_mutex_unlock(&bufPool->poolMutex);
+    ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
     /* create new buffer */
     DEBUGLOG(5, "create a new buffer");
     {   buffer_t buffer;
@@ -174,24 +181,38 @@ static void ZSTDMT_releaseBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buf)
 {
     if (buf.start == NULL) return;   /* compatible with release on NULL */
     DEBUGLOG(5, "ZSTDMT_releaseBuffer");
-    pthread_mutex_lock(&bufPool->poolMutex);
+    ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
     if (bufPool->nbBuffers < bufPool->totalBuffers) {
         bufPool->bTable[bufPool->nbBuffers++] = buf;  /* stored for later use */
-        pthread_mutex_unlock(&bufPool->poolMutex);
+        ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
         return;
     }
-    pthread_mutex_unlock(&bufPool->poolMutex);
+    ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
     /* Reached bufferPool capacity (should not happen) */
     DEBUGLOG(5, "buffer pool capacity reached => freeing ");
     ZSTD_free(buf.start, bufPool->cMem);
 }
 
+/* Sets parameters relevant to the compression job, initializing others to
+ * default values. Notably, nbThreads should probably be zero. */
+static ZSTD_CCtx_params ZSTDMT_makeJobCCtxParams(ZSTD_CCtx_params const params)
+{
+    ZSTD_CCtx_params jobParams;
+    memset(&jobParams, 0, sizeof(jobParams));
+
+    jobParams.cParams = params.cParams;
+    jobParams.fParams = params.fParams;
+    jobParams.compressionLevel = params.compressionLevel;
+
+    jobParams.ldmParams = params.ldmParams;
+    return jobParams;
+}
 
 /* =====   CCtx Pool   ===== */
 /* a single CCtx Pool can be invoked from multiple threads in parallel */
 
 typedef struct {
-    pthread_mutex_t poolMutex;
+    ZSTD_pthread_mutex_t poolMutex;
     unsigned totalCCtx;
     unsigned availCCtx;
     ZSTD_customMem cMem;
@@ -204,7 +225,7 @@ static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool)
     unsigned u;
     for (u=0; u<pool->totalCCtx; u++)
         ZSTD_freeCCtx(pool->cctx[u]);  /* note : compatible with free on NULL */
-    pthread_mutex_destroy(&pool->poolMutex);
+    ZSTD_pthread_mutex_destroy(&pool->poolMutex);
     ZSTD_free(pool, pool->cMem);
 }
 
@@ -216,7 +237,7 @@ static ZSTDMT_CCtxPool* ZSTDMT_createCCtxPool(unsigned nbThreads,
     ZSTDMT_CCtxPool* const cctxPool = (ZSTDMT_CCtxPool*) ZSTD_calloc(
         sizeof(ZSTDMT_CCtxPool) + (nbThreads-1)*sizeof(ZSTD_CCtx*), cMem);
     if (!cctxPool) return NULL;
-    if (pthread_mutex_init(&cctxPool->poolMutex, NULL)) {
+    if (ZSTD_pthread_mutex_init(&cctxPool->poolMutex, NULL)) {
         ZSTD_free(cctxPool, cMem);
         return NULL;
     }
@@ -232,7 +253,7 @@ static ZSTDMT_CCtxPool* ZSTDMT_createCCtxPool(unsigned nbThreads,
 /* only works during initialization phase, not during compression */
 static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool)
 {
-    pthread_mutex_lock(&cctxPool->poolMutex);
+    ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);
     {   unsigned const nbThreads = cctxPool->totalCCtx;
         size_t const poolSize = sizeof(*cctxPool)
                                 + (nbThreads-1)*sizeof(ZSTD_CCtx*);
@@ -241,7 +262,7 @@ static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool)
         for (u=0; u<nbThreads; u++) {
             totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctx[u]);
         }
-        pthread_mutex_unlock(&cctxPool->poolMutex);
+        ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
         return poolSize + totalCCtxSize;
     }
 }
@@ -249,14 +270,14 @@ static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool)
 static ZSTD_CCtx* ZSTDMT_getCCtx(ZSTDMT_CCtxPool* cctxPool)
 {
     DEBUGLOG(5, "ZSTDMT_getCCtx");
-    pthread_mutex_lock(&cctxPool->poolMutex);
+    ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);
     if (cctxPool->availCCtx) {
         cctxPool->availCCtx--;
         {   ZSTD_CCtx* const cctx = cctxPool->cctx[cctxPool->availCCtx];
-            pthread_mutex_unlock(&cctxPool->poolMutex);
+            ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
             return cctx;
     }   }
-    pthread_mutex_unlock(&cctxPool->poolMutex);
+    ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
     DEBUGLOG(5, "create one more CCtx");
     return ZSTD_createCCtx_advanced(cctxPool->cMem);   /* note : can be NULL, when creation fails ! */
 }
@@ -264,7 +285,7 @@ static ZSTD_CCtx* ZSTDMT_getCCtx(ZSTDMT_CCtxPool* cctxPool)
 static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx)
 {
     if (cctx==NULL) return;   /* compatibility with release on NULL */
-    pthread_mutex_lock(&pool->poolMutex);
+    ZSTD_pthread_mutex_lock(&pool->poolMutex);
     if (pool->availCCtx < pool->totalCCtx)
         pool->cctx[pool->availCCtx++] = cctx;
     else {
@@ -272,7 +293,7 @@ static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx)
         DEBUGLOG(5, "CCtx pool overflow : free cctx");
         ZSTD_freeCCtx(cctx);
     }
-    pthread_mutex_unlock(&pool->poolMutex);
+    ZSTD_pthread_mutex_unlock(&pool->poolMutex);
 }
 
 
@@ -290,9 +311,9 @@ typedef struct {
     unsigned lastChunk;
     unsigned jobCompleted;
     unsigned jobScanned;
-    pthread_mutex_t* jobCompleted_mutex;
-    pthread_cond_t* jobCompleted_cond;
-    ZSTD_parameters params;
+    ZSTD_pthread_mutex_t* jobCompleted_mutex;
+    ZSTD_pthread_cond_t* jobCompleted_cond;
+    ZSTD_CCtx_params params;
     const ZSTD_CDict* cdict;
     ZSTDMT_CCtxPool* cctxPool;
     ZSTDMT_bufferPool* bufPool;
@@ -329,10 +350,15 @@ void ZSTDMT_compressChunk(void* jobDescription)
         if (ZSTD_isError(initError)) { job->cSize = initError; goto _endJob; }
     } else {  /* srcStart points at reloaded section */
         if (!job->firstChunk) job->params.fParams.contentSizeFlag = 0;  /* ensure no srcSize control */
-        {   size_t const dictModeError = ZSTD_setCCtxParameter(cctx, ZSTD_p_forceRawDict, 1);  /* Force loading dictionary in "content-only" mode (no header analysis) */
-            size_t const initError = ZSTD_compressBegin_advanced(cctx, job->srcStart, job->dictSize, job->params, job->fullFrameSize);
-            if (ZSTD_isError(initError) || ZSTD_isError(dictModeError)) { job->cSize = initError; goto _endJob; }
-            ZSTD_setCCtxParameter(cctx, ZSTD_p_forceWindow, 1);
+        { ZSTD_CCtx_params jobParams = job->params;
+          size_t const forceWindowError =
+              ZSTD_CCtxParam_setParameter(&jobParams, ZSTD_p_forceMaxWindow, !job->firstChunk);
+          /* Force loading dictionary in "content-only" mode (no header analysis) */
+          size_t const initError = ZSTD_compressBegin_advanced_internal(cctx, job->srcStart, job->dictSize, ZSTD_dm_rawContent, jobParams, job->fullFrameSize);
+            if (ZSTD_isError(initError) || ZSTD_isError(forceWindowError)) {
+                job->cSize = initError;
+                goto _endJob;
+            }
     }   }
     if (!job->firstChunk) {  /* flush and overwrite frame header when it's not first segment */
         size_t const hSize = ZSTD_compressContinue(cctx, dstBuff.start, dstBuff.size, src, 0);
@@ -353,11 +379,11 @@ _endJob:
     ZSTDMT_releaseCCtx(job->cctxPool, cctx);
     ZSTDMT_releaseBuffer(job->bufPool, job->src);
     job->src = g_nullBuffer; job->srcStart = NULL;
-    PTHREAD_MUTEX_LOCK(job->jobCompleted_mutex);
+    ZSTD_PTHREAD_MUTEX_LOCK(job->jobCompleted_mutex);
     job->jobCompleted = 1;
     job->jobScanned = 0;
-    pthread_cond_signal(job->jobCompleted_cond);
-    pthread_mutex_unlock(job->jobCompleted_mutex);
+    ZSTD_pthread_cond_signal(job->jobCompleted_cond);
+    ZSTD_pthread_mutex_unlock(job->jobCompleted_mutex);
 }
 
 
@@ -375,24 +401,21 @@ struct ZSTDMT_CCtx_s {
     ZSTDMT_jobDescription* jobs;
     ZSTDMT_bufferPool* bufPool;
     ZSTDMT_CCtxPool* cctxPool;
-    pthread_mutex_t jobCompleted_mutex;
-    pthread_cond_t jobCompleted_cond;
+    ZSTD_pthread_mutex_t jobCompleted_mutex;
+    ZSTD_pthread_cond_t jobCompleted_cond;
     size_t targetSectionSize;
     size_t inBuffSize;
     size_t dictSize;
     size_t targetDictSize;
     inBuff_t inBuff;
-    ZSTD_parameters params;
+    ZSTD_CCtx_params params;
     XXH64_state_t xxhState;
-    unsigned nbThreads;
     unsigned jobIDMask;
     unsigned doneJobID;
     unsigned nextJobID;
     unsigned frameEnded;
     unsigned allJobsCompleted;
-    unsigned overlapLog;
     unsigned long long frameContentSize;
-    size_t sectionSize;
     ZSTD_customMem cMem;
     ZSTD_CDict* cdictLocal;
     const ZSTD_CDict* cdict;
@@ -407,6 +430,15 @@ static ZSTDMT_jobDescription* ZSTDMT_allocJobsTable(U32* nbJobsPtr, ZSTD_customM
                             nbJobs * sizeof(ZSTDMT_jobDescription), cMem);
 }
 
+/* Internal only */
+size_t ZSTDMT_initializeCCtxParameters(ZSTD_CCtx_params* params, unsigned nbThreads)
+{
+    params->nbThreads = nbThreads;
+    params->overlapSizeLog = ZSTDMT_OVERLAPLOG_DEFAULT;
+    params->jobSize = 0;
+    return 0;
+}
+
 ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbThreads, ZSTD_customMem cMem)
 {
     ZSTDMT_CCtx* mtctx;
@@ -421,12 +453,10 @@ ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbThreads, ZSTD_customMem cMem)
 
     mtctx = (ZSTDMT_CCtx*) ZSTD_calloc(sizeof(ZSTDMT_CCtx), cMem);
     if (!mtctx) return NULL;
+    ZSTDMT_initializeCCtxParameters(&mtctx->params, nbThreads);
     mtctx->cMem = cMem;
-    mtctx->nbThreads = nbThreads;
     mtctx->allJobsCompleted = 1;
-    mtctx->sectionSize = 0;
-    mtctx->overlapLog = ZSTDMT_OVERLAPLOG_DEFAULT;
-    mtctx->factory = POOL_create(nbThreads, 0);
+    mtctx->factory = POOL_create_advanced(nbThreads, 0, cMem);
     mtctx->jobs = ZSTDMT_allocJobsTable(&nbJobs, cMem);
     mtctx->jobIDMask = nbJobs - 1;
     mtctx->bufPool = ZSTDMT_createBufferPool(nbThreads, cMem);
@@ -435,11 +465,11 @@ ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbThreads, ZSTD_customMem cMem)
         ZSTDMT_freeCCtx(mtctx);
         return NULL;
     }
-    if (pthread_mutex_init(&mtctx->jobCompleted_mutex, NULL)) {
+    if (ZSTD_pthread_mutex_init(&mtctx->jobCompleted_mutex, NULL)) {
         ZSTDMT_freeCCtx(mtctx);
         return NULL;
     }
-    if (pthread_cond_init(&mtctx->jobCompleted_cond, NULL)) {
+    if (ZSTD_pthread_cond_init(&mtctx->jobCompleted_cond, NULL)) {
         ZSTDMT_freeCCtx(mtctx);
         return NULL;
     }
@@ -459,28 +489,46 @@ static void ZSTDMT_releaseAllJobResources(ZSTDMT_CCtx* mtctx)
     unsigned jobID;
     DEBUGLOG(3, "ZSTDMT_releaseAllJobResources");
     for (jobID=0; jobID <= mtctx->jobIDMask; jobID++) {
+        DEBUGLOG(4, "job%02u: release dst address %08X", jobID, (U32)(size_t)mtctx->jobs[jobID].dstBuff.start);
         ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[jobID].dstBuff);
         mtctx->jobs[jobID].dstBuff = g_nullBuffer;
+        DEBUGLOG(4, "job%02u: release src address %08X", jobID, (U32)(size_t)mtctx->jobs[jobID].src.start);
         ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[jobID].src);
         mtctx->jobs[jobID].src = g_nullBuffer;
     }
     memset(mtctx->jobs, 0, (mtctx->jobIDMask+1)*sizeof(ZSTDMT_jobDescription));
+    DEBUGLOG(4, "input: release address %08X", (U32)(size_t)mtctx->inBuff.buffer.start);
     ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->inBuff.buffer);
     mtctx->inBuff.buffer = g_nullBuffer;
     mtctx->allJobsCompleted = 1;
 }
 
+static void ZSTDMT_waitForAllJobsCompleted(ZSTDMT_CCtx* zcs)
+{
+    DEBUGLOG(4, "ZSTDMT_waitForAllJobsCompleted");
+    while (zcs->doneJobID < zcs->nextJobID) {
+        unsigned const jobID = zcs->doneJobID & zcs->jobIDMask;
+        ZSTD_PTHREAD_MUTEX_LOCK(&zcs->jobCompleted_mutex);
+        while (zcs->jobs[jobID].jobCompleted==0) {
+            DEBUGLOG(5, "waiting for jobCompleted signal from chunk %u", zcs->doneJobID);   /* we want to block when waiting for data to flush */
+            ZSTD_pthread_cond_wait(&zcs->jobCompleted_cond, &zcs->jobCompleted_mutex);
+        }
+        ZSTD_pthread_mutex_unlock(&zcs->jobCompleted_mutex);
+        zcs->doneJobID++;
+    }
+}
+
 size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx)
 {
     if (mtctx==NULL) return 0;   /* compatible with free on NULL */
-    POOL_free(mtctx->factory);
-    if (!mtctx->allJobsCompleted) ZSTDMT_releaseAllJobResources(mtctx); /* stop workers first */
-    ZSTDMT_freeBufferPool(mtctx->bufPool);  /* release job resources into pools first */
+    POOL_free(mtctx->factory);   /* stop and free worker threads */
+    ZSTDMT_releaseAllJobResources(mtctx);  /* release job resources into pools first */
     ZSTD_free(mtctx->jobs, mtctx->cMem);
+    ZSTDMT_freeBufferPool(mtctx->bufPool);
     ZSTDMT_freeCCtxPool(mtctx->cctxPool);
     ZSTD_freeCDict(mtctx->cdictLocal);
-    pthread_mutex_destroy(&mtctx->jobCompleted_mutex);
-    pthread_cond_destroy(&mtctx->jobCompleted_cond);
+    ZSTD_pthread_mutex_destroy(&mtctx->jobCompleted_mutex);
+    ZSTD_pthread_cond_destroy(&mtctx->jobCompleted_cond);
     ZSTD_free(mtctx, mtctx->cMem);
     return 0;
 }
@@ -496,22 +544,35 @@ size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx)
             + ZSTD_sizeof_CDict(mtctx->cdictLocal);
 }
 
-size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSDTMT_parameter parameter, unsigned value)
-{
+/* Internal only */
+size_t ZSTDMT_CCtxParam_setMTCtxParameter(
+    ZSTD_CCtx_params* params, ZSTDMT_parameter parameter, unsigned value) {
     switch(parameter)
     {
     case ZSTDMT_p_sectionSize :
-        mtctx->sectionSize = value;
+        params->jobSize = value;
         return 0;
     case ZSTDMT_p_overlapSectionLog :
-        DEBUGLOG(5, "ZSTDMT_p_overlapSectionLog : %u", value);
-        mtctx->overlapLog = (value >= 9) ? 9 : value;
+        DEBUGLOG(4, "ZSTDMT_p_overlapSectionLog : %u", value);
+        params->overlapSizeLog = (value >= 9) ? 9 : value;
         return 0;
     default :
         return ERROR(parameter_unsupported);
     }
 }
 
+size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, unsigned value)
+{
+    switch(parameter)
+    {
+    case ZSTDMT_p_sectionSize :
+        return ZSTDMT_CCtxParam_setMTCtxParameter(&mtctx->params, parameter, value);
+    case ZSTDMT_p_overlapSectionLog :
+        return ZSTDMT_CCtxParam_setMTCtxParameter(&mtctx->params, parameter, value);
+    default :
+        return ERROR(parameter_unsupported);
+    }
+}
 
 /* ------------------------------------------ */
 /* =====   Multi-threaded compression   ===== */
@@ -528,17 +589,17 @@ static unsigned computeNbChunks(size_t srcSize, unsigned windowLog, unsigned nbT
     return (multiplier>1) ? nbChunksLarge : nbChunksSmall;
 }
 
-
-size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
-                               void* dst, size_t dstCapacity,
-                         const void* src, size_t srcSize,
-                         const ZSTD_CDict* cdict,
-                               ZSTD_parameters const params,
-                               unsigned overlapLog)
+static size_t ZSTDMT_compress_advanced_internal(
+                ZSTDMT_CCtx* mtctx,
+                void* dst, size_t dstCapacity,
+          const void* src, size_t srcSize,
+          const ZSTD_CDict* cdict,
+                ZSTD_CCtx_params const params)
 {
-    unsigned const overlapRLog = (overlapLog>9) ? 0 : 9-overlapLog;
+    ZSTD_CCtx_params const jobParams = ZSTDMT_makeJobCCtxParams(params);
+    unsigned const overlapRLog = (params.overlapSizeLog>9) ? 0 : 9-params.overlapSizeLog;
     size_t const overlapSize = (overlapRLog>=9) ? 0 : (size_t)1 << (params.cParams.windowLog - overlapRLog);
-    unsigned nbChunks = computeNbChunks(srcSize, params.cParams.windowLog, mtctx->nbThreads);
+    unsigned nbChunks = computeNbChunks(srcSize, params.cParams.windowLog, params.nbThreads);
     size_t const proposedChunkSize = (srcSize + (nbChunks-1)) / nbChunks;
     size_t const avgChunkSize = ((proposedChunkSize & 0x1FFFF) < 0x7FFF) ? proposedChunkSize + 0xFFFF : proposedChunkSize;   /* avoid too small last block */
     const char* const srcStart = (const char*)src;
@@ -546,12 +607,14 @@ size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
     unsigned const compressWithinDst = (dstCapacity >= ZSTD_compressBound(srcSize)) ? nbChunks : (unsigned)(dstCapacity / ZSTD_compressBound(avgChunkSize));  /* presumes avgChunkSize >= 256 KB, which should be the case */
     size_t frameStartPos = 0, dstBufferPos = 0;
     XXH64_state_t xxh64;
+    assert(jobParams.nbThreads == 0);
+    assert(mtctx->cctxPool->totalCCtx == params.nbThreads);
 
     DEBUGLOG(4, "nbChunks  : %2u   (chunkSize : %u bytes)   ", nbChunks, (U32)avgChunkSize);
     if (nbChunks==1) {   /* fallback to single-thread mode */
         ZSTD_CCtx* const cctx = mtctx->cctxPool->cctx[0];
-        if (cdict) return ZSTD_compress_usingCDict_advanced(cctx, dst, dstCapacity, src, srcSize, cdict, params.fParams);
-        return ZSTD_compress_advanced(cctx, dst, dstCapacity, src, srcSize, NULL, 0, params);
+        if (cdict) return ZSTD_compress_usingCDict_advanced(cctx, dst, dstCapacity, src, srcSize, cdict, jobParams.fParams);
+        return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, NULL, 0, jobParams);
     }
     assert(avgChunkSize >= 256 KB);  /* condition for ZSTD_compressBound(A) + ZSTD_compressBound(B) <= ZSTD_compressBound(A+B), which is required for compressWithinDst */
     ZSTDMT_setBufferSize(mtctx->bufPool, ZSTD_compressBound(avgChunkSize) );
@@ -580,7 +643,7 @@ size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
             mtctx->jobs[u].srcSize = chunkSize;
             mtctx->jobs[u].cdict = mtctx->nextJobID==0 ? cdict : NULL;
             mtctx->jobs[u].fullFrameSize = srcSize;
-            mtctx->jobs[u].params = params;
+            mtctx->jobs[u].params = jobParams;
             /* do not calculate checksum within sections, but write it in header for first section */
             if (u!=0) mtctx->jobs[u].params.fParams.checksumFlag = 0;
             mtctx->jobs[u].dstBuff = dstBuffer;
@@ -610,12 +673,12 @@ size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
         unsigned chunkID;
         for (chunkID=0; chunkID<nbChunks; chunkID++) {
             DEBUGLOG(5, "waiting for chunk %u ", chunkID);
-            PTHREAD_MUTEX_LOCK(&mtctx->jobCompleted_mutex);
+            ZSTD_PTHREAD_MUTEX_LOCK(&mtctx->jobCompleted_mutex);
             while (mtctx->jobs[chunkID].jobCompleted==0) {
                 DEBUGLOG(5, "waiting for jobCompleted signal from chunk %u", chunkID);
-                pthread_cond_wait(&mtctx->jobCompleted_cond, &mtctx->jobCompleted_mutex);
+                ZSTD_pthread_cond_wait(&mtctx->jobCompleted_cond, &mtctx->jobCompleted_mutex);
             }
-            pthread_mutex_unlock(&mtctx->jobCompleted_mutex);
+            ZSTD_pthread_mutex_unlock(&mtctx->jobCompleted_mutex);
             DEBUGLOG(5, "ready to write chunk %u ", chunkID);
 
             mtctx->jobs[chunkID].srcStart = NULL;
@@ -628,9 +691,8 @@ size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
                     if (chunkID >= compressWithinDst) {  /* chunk compressed into its own buffer, which must be released */
                         DEBUGLOG(5, "releasing buffer %u>=%u", chunkID, compressWithinDst);
                         ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[chunkID].dstBuff);
-                    }
-                    mtctx->jobs[chunkID].dstBuff = g_nullBuffer;
-                }
+                }   }
+                mtctx->jobs[chunkID].dstBuff = g_nullBuffer;
                 dstPos += cSize ;
             }
         }  /* for (chunkID=0; chunkID<nbChunks; chunkID++) */
@@ -651,6 +713,23 @@ size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
     }
 }
 
+size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize,
+                         const ZSTD_CDict* cdict,
+                               ZSTD_parameters const params,
+                               unsigned overlapLog)
+{
+    ZSTD_CCtx_params cctxParams = mtctx->params;
+    cctxParams.cParams = params.cParams;
+    cctxParams.fParams = params.fParams;
+    cctxParams.overlapSizeLog = overlapLog;
+    return ZSTDMT_compress_advanced_internal(mtctx,
+                                             dst, dstCapacity,
+                                             src, srcSize,
+                                             cdict, cctxParams);
+}
+
 
 size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,
                            void* dst, size_t dstCapacity,
@@ -668,38 +747,25 @@ size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,
 /* =======      Streaming API     ======= */
 /* ====================================== */
 
-static void ZSTDMT_waitForAllJobsCompleted(ZSTDMT_CCtx* zcs)
-{
-    DEBUGLOG(4, "ZSTDMT_waitForAllJobsCompleted");
-    while (zcs->doneJobID < zcs->nextJobID) {
-        unsigned const jobID = zcs->doneJobID & zcs->jobIDMask;
-        PTHREAD_MUTEX_LOCK(&zcs->jobCompleted_mutex);
-        while (zcs->jobs[jobID].jobCompleted==0) {
-            DEBUGLOG(5, "waiting for jobCompleted signal from chunk %u", zcs->doneJobID);   /* we want to block when waiting for data to flush */
-            pthread_cond_wait(&zcs->jobCompleted_cond, &zcs->jobCompleted_mutex);
-        }
-        pthread_mutex_unlock(&zcs->jobCompleted_mutex);
-        zcs->doneJobID++;
-    }
-}
-
-
-/** ZSTDMT_initCStream_internal() :
- *  internal usage only */
-size_t ZSTDMT_initCStream_internal(ZSTDMT_CCtx* zcs,
-                    const void* dict, size_t dictSize, const ZSTD_CDict* cdict,
-                    ZSTD_parameters params, unsigned long long pledgedSrcSize)
+size_t ZSTDMT_initCStream_internal(
+        ZSTDMT_CCtx* zcs,
+        const void* dict, size_t dictSize, ZSTD_dictMode_e dictMode,
+        const ZSTD_CDict* cdict, ZSTD_CCtx_params params,
+        unsigned long long pledgedSrcSize)
 {
     DEBUGLOG(4, "ZSTDMT_initCStream_internal");
     /* params are supposed to be fully validated at this point */
     assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
     assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
+    assert(zcs->cctxPool->totalCCtx == params.nbThreads);
 
-    if (zcs->nbThreads==1) {
+    if (params.nbThreads==1) {
+        ZSTD_CCtx_params const singleThreadParams = ZSTDMT_makeJobCCtxParams(params);
         DEBUGLOG(4, "single thread mode");
+        assert(singleThreadParams.nbThreads == 0);
         return ZSTD_initCStream_internal(zcs->cctxPool->cctx[0],
-                                        dict, dictSize, cdict,
-                                        params, pledgedSrcSize);
+                                         dict, dictSize, cdict,
+                                         singleThreadParams, pledgedSrcSize);
     }
 
     if (zcs->allJobsCompleted == 0) {   /* previous compression not correctly finished */
@@ -714,7 +780,7 @@ size_t ZSTDMT_initCStream_internal(ZSTDMT_CCtx* zcs,
         DEBUGLOG(4,"cdictLocal: %08X", (U32)(size_t)zcs->cdictLocal);
         ZSTD_freeCDict(zcs->cdictLocal);
         zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,
-                                                    0 /* byRef */, ZSTD_dm_auto,   /* note : a loadPrefix becomes an internal CDict */
+                                                    ZSTD_dlm_byCopy, dictMode, /* note : a loadPrefix becomes an internal CDict */
                                                     params.cParams, zcs->cMem);
         zcs->cdict = zcs->cdictLocal;
         if (zcs->cdictLocal == NULL) return ERROR(memory_allocation);
@@ -725,10 +791,10 @@ size_t ZSTDMT_initCStream_internal(ZSTDMT_CCtx* zcs,
         zcs->cdict = cdict;
     }
 
-    zcs->targetDictSize = (zcs->overlapLog==0) ? 0 : (size_t)1 << (zcs->params.cParams.windowLog - (9 - zcs->overlapLog));
-    DEBUGLOG(4, "overlapLog : %u ", zcs->overlapLog);
+    zcs->targetDictSize = (params.overlapSizeLog==0) ? 0 : (size_t)1 << (params.cParams.windowLog - (9 - params.overlapSizeLog));
+    DEBUGLOG(4, "overlapLog : %u ", params.overlapSizeLog);
     DEBUGLOG(4, "overlap Size : %u KB", (U32)(zcs->targetDictSize>>10));
-    zcs->targetSectionSize = zcs->sectionSize ? zcs->sectionSize : (size_t)1 << (zcs->params.cParams.windowLog + 2);
+    zcs->targetSectionSize = params.jobSize ? params.jobSize : (size_t)1 << (params.cParams.windowLog + 2);
     zcs->targetSectionSize = MAX(ZSTDMT_SECTION_SIZE_MIN, zcs->targetSectionSize);
     zcs->targetSectionSize = MAX(zcs->targetDictSize, zcs->targetSectionSize);
     DEBUGLOG(4, "Section Size : %u KB", (U32)(zcs->targetSectionSize>>10));
@@ -749,8 +815,12 @@ size_t ZSTDMT_initCStream_advanced(ZSTDMT_CCtx* mtctx,
                                    ZSTD_parameters params,
                                    unsigned long long pledgedSrcSize)
 {
+    ZSTD_CCtx_params cctxParams = mtctx->params;
     DEBUGLOG(5, "ZSTDMT_initCStream_advanced");
-    return ZSTDMT_initCStream_internal(mtctx, dict, dictSize, NULL, params, pledgedSrcSize);
+    cctxParams.cParams = params.cParams;
+    cctxParams.fParams = params.fParams;
+    return ZSTDMT_initCStream_internal(mtctx, dict, dictSize, ZSTD_dm_auto, NULL,
+                                       cctxParams, pledgedSrcSize);
 }
 
 size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
@@ -758,11 +828,12 @@ size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
                                      ZSTD_frameParameters fParams,
                                      unsigned long long pledgedSrcSize)
 {
-    ZSTD_parameters params = ZSTD_getParamsFromCDict(cdict);
+    ZSTD_CCtx_params cctxParams = mtctx->params;
+    cctxParams.cParams = ZSTD_getCParamsFromCDict(cdict);
+    cctxParams.fParams = fParams;
     if (cdict==NULL) return ERROR(dictionary_wrong);   /* method incompatible with NULL cdict */
-    params.fParams = fParams;
-    return ZSTDMT_initCStream_internal(mtctx, NULL, 0 /*dictSize*/, cdict,
-                                        params, pledgedSrcSize);
+    return ZSTDMT_initCStream_internal(mtctx, NULL, 0 /*dictSize*/, ZSTD_dm_auto, cdict,
+                                       cctxParams, pledgedSrcSize);
 }
 
 
@@ -770,14 +841,18 @@ size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
  * pledgedSrcSize is optional and can be zero == unknown */
 size_t ZSTDMT_resetCStream(ZSTDMT_CCtx* zcs, unsigned long long pledgedSrcSize)
 {
-    if (zcs->nbThreads==1)
+    if (zcs->params.nbThreads==1)
         return ZSTD_resetCStream(zcs->cctxPool->cctx[0], pledgedSrcSize);
-    return ZSTDMT_initCStream_internal(zcs, NULL, 0, 0, zcs->params, pledgedSrcSize);
+    return ZSTDMT_initCStream_internal(zcs, NULL, 0, ZSTD_dm_auto, 0, zcs->params,
+                                       pledgedSrcSize);
 }
 
 size_t ZSTDMT_initCStream(ZSTDMT_CCtx* zcs, int compressionLevel) {
     ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, 0);
-    return ZSTDMT_initCStream_internal(zcs, NULL, 0, NULL, params, 0);
+    ZSTD_CCtx_params cctxParams = zcs->params;
+    cctxParams.cParams = params.cParams;
+    cctxParams.fParams = params.fParams;
+    return ZSTDMT_initCStream_internal(zcs, NULL, 0, ZSTD_dm_auto, NULL, cctxParams, 0);
 }
 
 
@@ -856,13 +931,13 @@ static size_t ZSTDMT_flushNextJob(ZSTDMT_CCtx* zcs, ZSTD_outBuffer* output, unsi
 {
     unsigned const wJobID = zcs->doneJobID & zcs->jobIDMask;
     if (zcs->doneJobID == zcs->nextJobID) return 0;   /* all flushed ! */
-    PTHREAD_MUTEX_LOCK(&zcs->jobCompleted_mutex);
+    ZSTD_PTHREAD_MUTEX_LOCK(&zcs->jobCompleted_mutex);
     while (zcs->jobs[wJobID].jobCompleted==0) {
         DEBUGLOG(5, "waiting for jobCompleted signal from job %u", zcs->doneJobID);
-        if (!blockToFlush) { pthread_mutex_unlock(&zcs->jobCompleted_mutex); return 0; }  /* nothing ready to be flushed => skip */
-        pthread_cond_wait(&zcs->jobCompleted_cond, &zcs->jobCompleted_mutex);  /* block when nothing available to flush */
+        if (!blockToFlush) { ZSTD_pthread_mutex_unlock(&zcs->jobCompleted_mutex); return 0; }  /* nothing ready to be flushed => skip */
+        ZSTD_pthread_cond_wait(&zcs->jobCompleted_cond, &zcs->jobCompleted_mutex);  /* block when nothing available to flush */
     }
-    pthread_mutex_unlock(&zcs->jobCompleted_mutex);
+    ZSTD_pthread_mutex_unlock(&zcs->jobCompleted_mutex);
     /* compression job completed : output can be flushed */
     {   ZSTDMT_jobDescription job = zcs->jobs[wJobID];
         if (!job.jobScanned) {
@@ -906,7 +981,7 @@ static size_t ZSTDMT_flushNextJob(ZSTDMT_CCtx* zcs, ZSTD_outBuffer* output, unsi
 
 
 /** ZSTDMT_compressStream_generic() :
- *  internal use only
+ *  internal use only - exposed to be invoked from zstd_compress.c
  *  assumption : output and input are valid (pos <= size)
  * @return : minimum amount of data remaining to flush, 0 if none */
 size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
@@ -915,25 +990,26 @@ size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
                                      ZSTD_EndDirective endOp)
 {
     size_t const newJobThreshold = mtctx->dictSize + mtctx->targetSectionSize;
+    unsigned forwardInputProgress = 0;
     assert(output->pos <= output->size);
     assert(input->pos  <= input->size);
     if ((mtctx->frameEnded) && (endOp==ZSTD_e_continue)) {
-        /* current frame being ended. Only flush/end are allowed. Or start new frame with init */
+        /* current frame being ended. Only flush/end are allowed */
         return ERROR(stage_wrong);
     }
-    if (mtctx->nbThreads==1) {  /* delegate to single-thread (synchronous) */
+    if (mtctx->params.nbThreads==1) {  /* delegate to single-thread (synchronous) */
         return ZSTD_compressStream_generic(mtctx->cctxPool->cctx[0], output, input, endOp);
     }
 
-    /* single-pass shortcut (note : this is synchronous-mode) */
-    if ( (mtctx->nextJobID==0)      /* just started */
-      && (mtctx->inBuff.filled==0)  /* nothing buffered */
-      && (endOp==ZSTD_e_end)        /* end order */
+    /* single-pass shortcut (note : synchronous-mode) */
+    if ( (mtctx->nextJobID == 0)     /* just started */
+      && (mtctx->inBuff.filled == 0) /* nothing buffered */
+      && (endOp == ZSTD_e_end)       /* end order */
       && (output->size - output->pos >= ZSTD_compressBound(input->size - input->pos)) ) { /* enough room */
-        size_t const cSize = ZSTDMT_compress_advanced(mtctx,
+        size_t const cSize = ZSTDMT_compress_advanced_internal(mtctx,
                 (char*)output->dst + output->pos, output->size - output->pos,
                 (const char*)input->src + input->pos, input->size - input->pos,
-                mtctx->cdict, mtctx->params, mtctx->overlapLog);
+                mtctx->cdict, mtctx->params);
         if (ZSTD_isError(cSize)) return cSize;
         input->pos = input->size;
         output->pos += cSize;
@@ -946,15 +1022,16 @@ size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
     /* fill input buffer */
     if (input->size > input->pos) {   /* support NULL input */
         if (mtctx->inBuff.buffer.start == NULL) {
-            mtctx->inBuff.buffer = ZSTDMT_getBuffer(mtctx->bufPool);
-            if (mtctx->inBuff.buffer.start == NULL) return ERROR(memory_allocation);
+            mtctx->inBuff.buffer = ZSTDMT_getBuffer(mtctx->bufPool);  /* note : may fail, in which case, no forward input progress */
             mtctx->inBuff.filled = 0;
         }
-        {   size_t const toLoad = MIN(input->size - input->pos, mtctx->inBuffSize - mtctx->inBuff.filled);
+        if (mtctx->inBuff.buffer.start) {
+            size_t const toLoad = MIN(input->size - input->pos, mtctx->inBuffSize - mtctx->inBuff.filled);
             DEBUGLOG(5, "inBuff:%08X;  inBuffSize=%u;  ToCopy=%u", (U32)(size_t)mtctx->inBuff.buffer.start, (U32)mtctx->inBuffSize, (U32)toLoad);
             memcpy((char*)mtctx->inBuff.buffer.start + mtctx->inBuff.filled, (const char*)input->src + input->pos, toLoad);
             input->pos += toLoad;
             mtctx->inBuff.filled += toLoad;
+            forwardInputProgress = toLoad>0;
     }   }
 
     if ( (mtctx->inBuff.filled >= newJobThreshold)  /* filled enough : let's compress */
@@ -963,7 +1040,7 @@ size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
     }
 
     /* check for potential compressed data ready to be flushed */
-    CHECK_F( ZSTDMT_flushNextJob(mtctx, output, (mtctx->inBuff.filled == mtctx->inBuffSize) /* blockToFlush */) ); /* block if it wasn't possible to create new job due to saturation */
+    CHECK_F( ZSTDMT_flushNextJob(mtctx, output, !forwardInputProgress /* blockToFlush */) ); /* block if there was no forward input progress */
 
     if (input->pos < input->size)  /* input not consumed : do not flush yet */
         endOp = ZSTD_e_continue;
@@ -1008,7 +1085,7 @@ static size_t ZSTDMT_flushStream_internal(ZSTDMT_CCtx* zcs, ZSTD_outBuffer* outp
 size_t ZSTDMT_flushStream(ZSTDMT_CCtx* zcs, ZSTD_outBuffer* output)
 {
     DEBUGLOG(5, "ZSTDMT_flushStream");
-    if (zcs->nbThreads==1)
+    if (zcs->params.nbThreads==1)
         return ZSTD_flushStream(zcs->cctxPool->cctx[0], output);
     return ZSTDMT_flushStream_internal(zcs, output, 0 /* endFrame */);
 }
@@ -1016,7 +1093,7 @@ size_t ZSTDMT_flushStream(ZSTDMT_CCtx* zcs, ZSTD_outBuffer* output)
 size_t ZSTDMT_endStream(ZSTDMT_CCtx* zcs, ZSTD_outBuffer* output)
 {
     DEBUGLOG(4, "ZSTDMT_endStream");
-    if (zcs->nbThreads==1)
+    if (zcs->params.nbThreads==1)
         return ZSTD_endStream(zcs->cctxPool->cctx[0], output);
     return ZSTDMT_flushStream_internal(zcs, output, 1 /* endFrame */);
 }
diff --git a/thirdparty/zstd/compress/zstdmt_compress.h b/thirdparty/zstd/compress/zstdmt_compress.h
index 0f0fc2b03f..8c59c684f1 100644
--- a/thirdparty/zstd/compress/zstdmt_compress.h
+++ b/thirdparty/zstd/compress/zstdmt_compress.h
@@ -5,6 +5,7 @@
  * 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 ZSTDMT_COMPRESS_H
@@ -80,19 +81,19 @@ ZSTDLIB_API size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
                                         ZSTD_frameParameters fparams,
                                         unsigned long long pledgedSrcSize);  /* note : zero means empty */
 
-/* ZSDTMT_parameter :
+/* ZSTDMT_parameter :
  * List of parameters that can be set using ZSTDMT_setMTCtxParameter() */
 typedef enum {
     ZSTDMT_p_sectionSize,        /* size of input "section". Each section is compressed in parallel. 0 means default, which is dynamically determined within compression functions */
     ZSTDMT_p_overlapSectionLog   /* Log of overlapped section; 0 == no overlap, 6(default) == use 1/8th of window, >=9 == use full window */
-} ZSDTMT_parameter;
+} ZSTDMT_parameter;
 
 /* ZSTDMT_setMTCtxParameter() :
  * allow setting individual parameters, one at a time, among a list of enums defined in ZSTDMT_parameter.
  * The function must be called typically after ZSTD_createCCtx().
  * Parameters not explicitly reset by ZSTDMT_init*() remain the same in consecutive compression sessions.
  * @return : 0, or an error code (which can be tested using ZSTD_isError()) */
-ZSTDLIB_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSDTMT_parameter parameter, unsigned value);
+ZSTDLIB_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, unsigned value);
 
 
 /*! ZSTDMT_compressStream_generic() :
@@ -107,6 +108,22 @@ ZSTDLIB_API size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
                                                 ZSTD_EndDirective endOp);
 
 
+/* ===   Private definitions; never ever use directly  === */
+
+size_t ZSTDMT_CCtxParam_setMTCtxParameter(ZSTD_CCtx_params* params, ZSTDMT_parameter parameter, unsigned value);
+
+size_t ZSTDMT_initializeCCtxParameters(ZSTD_CCtx_params* params, unsigned nbThreads);
+
+/*! ZSTDMT_initCStream_internal() :
+ *  Private use only. Init streaming operation.
+ *  expects params to be valid.
+ *  must receive dict, or cdict, or none, but not both.
+ *  @return : 0, or an error code */
+size_t ZSTDMT_initCStream_internal(ZSTDMT_CCtx* zcs,
+                    const void* dict, size_t dictSize, ZSTD_dictMode_e dictMode,
+                    const ZSTD_CDict* cdict,
+                    ZSTD_CCtx_params params, unsigned long long pledgedSrcSize);
+
 
 #if defined (__cplusplus)
 }
diff --git a/thirdparty/zstd/decompress/zstd_decompress.c b/thirdparty/zstd/decompress/zstd_decompress.c
index d2bc545e52..96fc609089 100644
--- a/thirdparty/zstd/decompress/zstd_decompress.c
+++ b/thirdparty/zstd/decompress/zstd_decompress.c
@@ -5,6 +5,7 @@
  * 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.
  */
 
 
@@ -34,7 +35,7 @@
 *  Frames requiring more memory will be rejected.
 */
 #ifndef ZSTD_MAXWINDOWSIZE_DEFAULT
-#  define ZSTD_MAXWINDOWSIZE_DEFAULT ((1 << ZSTD_WINDOWLOG_MAX) + 1)   /* defined within zstd.h */
+#  define ZSTD_MAXWINDOWSIZE_DEFAULT (((U32)1 << ZSTD_WINDOWLOG_DEFAULTMAX) + 1)
 #endif
 
 
@@ -101,13 +102,15 @@ struct ZSTD_DCtx_s
     const void* dictEnd;          /* end of previous segment */
     size_t expected;
     ZSTD_frameHeader fParams;
-    blockType_e bType;   /* used in ZSTD_decompressContinue(), to transfer blockType between header decoding and block decoding stages */
+    U64 decodedSize;
+    blockType_e bType;            /* used in ZSTD_decompressContinue(), store blockType between block header decoding and block decompression stages */
     ZSTD_dStage stage;
     U32 litEntropy;
     U32 fseEntropy;
     XXH64_state_t xxhState;
     size_t headerSize;
     U32 dictID;
+    ZSTD_format_e format;
     const BYTE* litPtr;
     ZSTD_customMem customMem;
     size_t litSize;
@@ -126,7 +129,6 @@ struct ZSTD_DCtx_s
     size_t outBuffSize;
     size_t outStart;
     size_t outEnd;
-    size_t blockSize;
     size_t lhSize;
     void* legacyContext;
     U32 previousLegacyVersion;
@@ -148,39 +150,44 @@ size_t ZSTD_sizeof_DCtx (const ZSTD_DCtx* dctx)
 
 size_t ZSTD_estimateDCtxSize(void) { return sizeof(ZSTD_DCtx); }
 
-size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)
+
+static size_t ZSTD_startingInputLength(ZSTD_format_e format)
 {
-    dctx->expected = ZSTD_frameHeaderSize_prefix;
-    dctx->stage = ZSTDds_getFrameHeaderSize;
-    dctx->previousDstEnd = NULL;
-    dctx->base = NULL;
-    dctx->vBase = NULL;
-    dctx->dictEnd = NULL;
-    dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001);  /* cover both little and big endian */
-    dctx->litEntropy = dctx->fseEntropy = 0;
-    dctx->dictID = 0;
-    MEM_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
-    memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue));  /* initial repcodes */
-    dctx->LLTptr = dctx->entropy.LLTable;
-    dctx->MLTptr = dctx->entropy.MLTable;
-    dctx->OFTptr = dctx->entropy.OFTable;
-    dctx->HUFptr = dctx->entropy.hufTable;
-    return 0;
+    size_t const startingInputLength = (format==ZSTD_f_zstd1_magicless) ?
+                    ZSTD_frameHeaderSize_prefix - ZSTD_frameIdSize :
+                    ZSTD_frameHeaderSize_prefix;
+    ZSTD_STATIC_ASSERT(ZSTD_FRAMEHEADERSIZE_PREFIX >= ZSTD_FRAMEIDSIZE);
+    /* only supports formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless */
+    assert( (format == ZSTD_f_zstd1) || (format == ZSTD_f_zstd1_magicless) );
+    return startingInputLength;
 }
 
 static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)
 {
-    ZSTD_decompressBegin(dctx);   /* cannot fail */
-    dctx->staticSize = 0;
+    dctx->format = ZSTD_f_zstd1;  /* ZSTD_decompressBegin() invokes ZSTD_startingInputLength() with argument dctx->format */
+    dctx->staticSize  = 0;
     dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
-    dctx->ddict   = NULL;
-    dctx->ddictLocal = NULL;
-    dctx->inBuff  = NULL;
-    dctx->inBuffSize = 0;
-    dctx->outBuffSize= 0;
+    dctx->ddict       = NULL;
+    dctx->ddictLocal  = NULL;
+    dctx->inBuff      = NULL;
+    dctx->inBuffSize  = 0;
+    dctx->outBuffSize = 0;
     dctx->streamStage = zdss_init;
 }
 
+ZSTD_DCtx* ZSTD_initStaticDCtx(void *workspace, size_t workspaceSize)
+{
+    ZSTD_DCtx* const dctx = (ZSTD_DCtx*) workspace;
+
+    if ((size_t)workspace & 7) return NULL;  /* 8-aligned */
+    if (workspaceSize < sizeof(ZSTD_DCtx)) return NULL;  /* minimum size */
+
+    ZSTD_initDCtx_internal(dctx);
+    dctx->staticSize = workspaceSize;
+    dctx->inBuff = (char*)(dctx+1);
+    return dctx;
+}
+
 ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem)
 {
     if (!customMem.customAlloc ^ !customMem.customFree) return NULL;
@@ -195,19 +202,6 @@ ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem)
     }
 }
 
-ZSTD_DCtx* ZSTD_initStaticDCtx(void *workspace, size_t workspaceSize)
-{
-    ZSTD_DCtx* dctx = (ZSTD_DCtx*) workspace;
-
-    if ((size_t)workspace & 7) return NULL;  /* 8-aligned */
-    if (workspaceSize < sizeof(ZSTD_DCtx)) return NULL;  /* minimum size */
-
-    ZSTD_initDCtx_internal(dctx);
-    dctx->staticSize = workspaceSize;
-    dctx->inBuff = (char*)(dctx+1);
-    return dctx;
-}
-
 ZSTD_DCtx* ZSTD_createDCtx(void)
 {
     return ZSTD_createDCtx_advanced(ZSTD_defaultCMem);
@@ -250,7 +244,7 @@ void ZSTD_copyDCtx(ZSTD_DCtx* dstDCtx, const ZSTD_DCtx* srcDCtx)
  *  Note 3 : Skippable Frame Identifiers are considered valid. */
 unsigned ZSTD_isFrame(const void* buffer, size_t size)
 {
-    if (size < 4) return 0;
+    if (size < ZSTD_frameIdSize) return 0;
     {   U32 const magic = MEM_readLE32(buffer);
         if (magic == ZSTD_MAGICNUMBER) return 1;
         if ((magic & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) return 1;
@@ -261,55 +255,70 @@ unsigned ZSTD_isFrame(const void* buffer, size_t size)
     return 0;
 }
 
-
-/** ZSTD_frameHeaderSize() :
-*   srcSize must be >= ZSTD_frameHeaderSize_prefix.
-*   @return : size of the Frame Header */
-size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize)
+/** ZSTD_frameHeaderSize_internal() :
+ *  srcSize must be large enough to reach header size fields.
+ *  note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless
+ * @return : size of the Frame Header
+ *           or an error code, which can be tested with ZSTD_isError() */
+static size_t ZSTD_frameHeaderSize_internal(const void* src, size_t srcSize, ZSTD_format_e format)
 {
-    if (srcSize < ZSTD_frameHeaderSize_prefix) return ERROR(srcSize_wrong);
-    {   BYTE const fhd = ((const BYTE*)src)[4];
+    size_t const minInputSize = ZSTD_startingInputLength(format);
+    if (srcSize < minInputSize) return ERROR(srcSize_wrong);
+
+    {   BYTE const fhd = ((const BYTE*)src)[minInputSize-1];
         U32 const dictID= fhd & 3;
         U32 const singleSegment = (fhd >> 5) & 1;
         U32 const fcsId = fhd >> 6;
-        return ZSTD_frameHeaderSize_prefix + !singleSegment + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId]
-                + (singleSegment && !fcsId);
+        return minInputSize + !singleSegment
+             + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId]
+             + (singleSegment && !fcsId);
     }
 }
 
+/** ZSTD_frameHeaderSize() :
+ *  srcSize must be >= ZSTD_frameHeaderSize_prefix.
+ * @return : size of the Frame Header */
+size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize)
+{
+    return ZSTD_frameHeaderSize_internal(src, srcSize, ZSTD_f_zstd1);
+}
 
-/** ZSTD_getFrameHeader() :
-*   decode Frame Header, or require larger `srcSize`.
-*   @return : 0, `zfhPtr` is correctly filled,
-*            >0, `srcSize` is too small, result is expected `srcSize`,
-*             or an error code, which can be tested using ZSTD_isError() */
-size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize)
+
+/** ZSTD_getFrameHeader_internal() :
+ *  decode Frame Header, or require larger `srcSize`.
+ *  note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless
+ * @return : 0, `zfhPtr` is correctly filled,
+ *          >0, `srcSize` is too small, value is wanted `srcSize` amount,
+ *           or an error code, which can be tested using ZSTD_isError() */
+static size_t ZSTD_getFrameHeader_internal(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format)
 {
     const BYTE* ip = (const BYTE*)src;
-    if (srcSize < ZSTD_frameHeaderSize_prefix) return ZSTD_frameHeaderSize_prefix;
+    size_t const minInputSize = ZSTD_startingInputLength(format);
 
-    if (MEM_readLE32(src) != ZSTD_MAGICNUMBER) {
+    if (srcSize < minInputSize) return minInputSize;
+
+    if ( (format != ZSTD_f_zstd1_magicless)
+      && (MEM_readLE32(src) != ZSTD_MAGICNUMBER) ) {
         if ((MEM_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
             /* skippable frame */
             if (srcSize < ZSTD_skippableHeaderSize)
                 return ZSTD_skippableHeaderSize; /* magic number + frame length */
             memset(zfhPtr, 0, sizeof(*zfhPtr));
-            zfhPtr->frameContentSize = MEM_readLE32((const char *)src + 4);
+            zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_frameIdSize);
             zfhPtr->frameType = ZSTD_skippableFrame;
-            zfhPtr->windowSize = 0;
             return 0;
         }
         return ERROR(prefix_unknown);
     }
 
     /* ensure there is enough `srcSize` to fully read/decode frame header */
-    {   size_t const fhsize = ZSTD_frameHeaderSize(src, srcSize);
+    {   size_t const fhsize = ZSTD_frameHeaderSize_internal(src, srcSize, format);
         if (srcSize < fhsize) return fhsize;
         zfhPtr->headerSize = (U32)fhsize;
     }
 
-    {   BYTE const fhdByte = ip[4];
-        size_t pos = 5;
+    {   BYTE const fhdByte = ip[minInputSize-1];
+        size_t pos = minInputSize;
         U32 const dictIDSizeCode = fhdByte&3;
         U32 const checksumFlag = (fhdByte>>2)&1;
         U32 const singleSegment = (fhdByte>>5)&1;
@@ -349,12 +358,25 @@ size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t src
         zfhPtr->frameType = ZSTD_frame;
         zfhPtr->frameContentSize = frameContentSize;
         zfhPtr->windowSize = windowSize;
+        zfhPtr->blockSizeMax = (unsigned) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
         zfhPtr->dictID = dictID;
         zfhPtr->checksumFlag = checksumFlag;
     }
     return 0;
 }
 
+/** ZSTD_getFrameHeader() :
+ *  decode Frame Header, or require larger `srcSize`.
+ *  note : this function does not consume input, it only reads it.
+ * @return : 0, `zfhPtr` is correctly filled,
+ *          >0, `srcSize` is too small, value is wanted `srcSize` amount,
+ *           or an error code, which can be tested using ZSTD_isError() */
+size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize)
+{
+    return ZSTD_getFrameHeader_internal(zfhPtr, src, srcSize, ZSTD_f_zstd1);
+}
+
+
 /** ZSTD_getFrameContentSize() :
  *  compatible with legacy mode
  * @return : decompressed size of the single frame pointed to be `src` if known, otherwise
@@ -388,14 +410,14 @@ unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
     unsigned long long totalDstSize = 0;
 
     while (srcSize >= ZSTD_frameHeaderSize_prefix) {
-        const U32 magicNumber = MEM_readLE32(src);
+        U32 const magicNumber = MEM_readLE32(src);
 
         if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
             size_t skippableSize;
             if (srcSize < ZSTD_skippableHeaderSize)
                 return ERROR(srcSize_wrong);
-            skippableSize = MEM_readLE32((const BYTE *)src + 4) +
-                            ZSTD_skippableHeaderSize;
+            skippableSize = MEM_readLE32((const BYTE *)src + ZSTD_frameIdSize)
+                          + ZSTD_skippableHeaderSize;
             if (srcSize < skippableSize) {
                 return ZSTD_CONTENTSIZE_ERROR;
             }
@@ -420,11 +442,9 @@ unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
             src = (const BYTE *)src + frameSrcSize;
             srcSize -= frameSrcSize;
         }
-    }
+    }  /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */
 
-    if (srcSize) {
-        return ZSTD_CONTENTSIZE_ERROR;
-    }
+    if (srcSize) return ZSTD_CONTENTSIZE_ERROR;
 
     return totalDstSize;
 }
@@ -440,7 +460,8 @@ unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
 unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize)
 {
     unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
-    return ret >= ZSTD_CONTENTSIZE_ERROR ? 0 : ret;
+    ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_ERROR < ZSTD_CONTENTSIZE_UNKNOWN);
+    return (ret >= ZSTD_CONTENTSIZE_ERROR) ? 0 : ret;
 }
 
 
@@ -449,9 +470,9 @@ unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize)
 *   @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */
 static size_t ZSTD_decodeFrameHeader(ZSTD_DCtx* dctx, const void* src, size_t headerSize)
 {
-    size_t const result = ZSTD_getFrameHeader(&(dctx->fParams), src, headerSize);
-    if (ZSTD_isError(result)) return result;  /* invalid header */
-    if (result>0) return ERROR(srcSize_wrong);   /* headerSize too small */
+    size_t const result = ZSTD_getFrameHeader_internal(&(dctx->fParams), src, headerSize, dctx->format);
+    if (ZSTD_isError(result)) return result;    /* invalid header */
+    if (result>0) return ERROR(srcSize_wrong);  /* headerSize too small */
     if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID))
         return ERROR(dictionary_wrong);
     if (dctx->fParams.checksumFlag) XXH64_reset(&dctx->xxhState, 0);
@@ -497,7 +518,8 @@ static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity,
 }
 
 /*! ZSTD_decodeLiteralsBlock() :
-    @return : nb of bytes read from src (< srcSize ) */
+ * @return : nb of bytes read from src (< srcSize )
+ *  note : symbol not declared but exposed for fullbench */
 size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                           const void* src, size_t srcSize)   /* note : srcSize < BLOCKSIZE */
 {
@@ -698,9 +720,9 @@ static const FSE_decode_t4 OF_defaultDTable[(1<<OF_DEFAULTNORMLOG)+1] = {
 };   /* OF_defaultDTable */
 
 /*! ZSTD_buildSeqTable() :
-    @return : nb bytes read from src,
-              or an error code if it fails, testable with ZSTD_isError()
-*/
+ * @return : nb bytes read from src,
+ *           or an error code if it fails, testable with ZSTD_isError()
+ */
 static size_t ZSTD_buildSeqTable(FSE_DTable* DTableSpace, const FSE_DTable** DTablePtr,
                                  symbolEncodingType_e type, U32 max, U32 maxLog,
                                  const void* src, size_t srcSize,
@@ -858,13 +880,25 @@ size_t ZSTD_execSequenceLast7(BYTE* op,
 }
 
 
-static seq_t ZSTD_decodeSequence(seqState_t* seqState)
+typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
+
+/* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum
+ * offset bits. But we can only read at most (STREAM_ACCUMULATOR_MIN_32 - 1)
+ * bits before reloading. This value is the maximum number of bytes we read
+ * after reloading when we are decoding long offets.
+ */
+#define LONG_OFFSETS_MAX_EXTRA_BITS_32                                         \
+    (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32                         \
+        ? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32                    \
+        : 0)
+
+static seq_t ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
 {
     seq_t seq;
 
     U32 const llCode = FSE_peekSymbol(&seqState->stateLL);
     U32 const mlCode = FSE_peekSymbol(&seqState->stateML);
-    U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb);   /* <= maxOff, by table construction */
+    U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb);   /* <= MaxOff, by table construction */
 
     U32 const llBits = LL_bits[llCode];
     U32 const mlBits = ML_bits[mlCode];
@@ -891,15 +925,25 @@ static seq_t ZSTD_decodeSequence(seqState_t* seqState)
                      0,        1,       1,       5,     0xD,     0x1D,     0x3D,     0x7D,
                      0xFD,   0x1FD,   0x3FD,   0x7FD,   0xFFD,   0x1FFD,   0x3FFD,   0x7FFD,
                      0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD,
-                     0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD };
+                     0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD, 0x1FFFFFFD, 0x3FFFFFFD, 0x7FFFFFFD };
 
     /* sequence */
     {   size_t offset;
         if (!ofCode)
             offset = 0;
         else {
-            offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits);   /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
-            if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
+            ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
+            ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
+            assert(ofBits <= MaxOff);
+            if (MEM_32bits() && longOffsets) {
+                U32 const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN_32-1);
+                offset = OF_base[ofCode] + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
+                if (MEM_32bits() || extraBits) BIT_reloadDStream(&seqState->DStream);
+                if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
+            } else {
+                offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits);   /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
+                if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
+            }
         }
 
         if (ofCode <= 1) {
@@ -923,13 +967,17 @@ static seq_t ZSTD_decodeSequence(seqState_t* seqState)
 
     seq.matchLength = ML_base[mlCode]
                     + ((mlCode>31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0);  /* <=  16 bits */
-    if (MEM_32bits() && (mlBits+llBits>24)) BIT_reloadDStream(&seqState->DStream);
+    if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
+        BIT_reloadDStream(&seqState->DStream);
+    if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
+        BIT_reloadDStream(&seqState->DStream);
+    /* Verify that there is enough bits to read the rest of the data in 64-bit mode. */
+    ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
 
     seq.litLength = LL_base[llCode]
                   + ((llCode>15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0);    /* <=  16 bits */
-    if (  MEM_32bits()
-      || (totalBits > 64 - 7 - (LLFSELog+MLFSELog+OffFSELog)) )
-       BIT_reloadDStream(&seqState->DStream);
+    if (MEM_32bits())
+        BIT_reloadDStream(&seqState->DStream);
 
     DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",
                 (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
@@ -1028,7 +1076,8 @@ size_t ZSTD_execSequence(BYTE* op,
 static size_t ZSTD_decompressSequences(
                                ZSTD_DCtx* dctx,
                                void* dst, size_t maxDstSize,
-                         const void* seqStart, size_t seqSize)
+                         const void* seqStart, size_t seqSize,
+                         const ZSTD_longOffset_e isLongOffset)
 {
     const BYTE* ip = (const BYTE*)seqStart;
     const BYTE* const iend = ip + seqSize;
@@ -1063,7 +1112,7 @@ static size_t ZSTD_decompressSequences(
 
         for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {
             nbSeq--;
-            {   seq_t const sequence = ZSTD_decodeSequence(&seqState);
+            {   seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
                 size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);
                 DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
                 if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
@@ -1088,13 +1137,14 @@ static size_t ZSTD_decompressSequences(
 }
 
 
-FORCE_INLINE_TEMPLATE seq_t ZSTD_decodeSequenceLong_generic(seqState_t* seqState, int const longOffsets)
+HINT_INLINE
+seq_t ZSTD_decodeSequenceLong(seqState_t* seqState, ZSTD_longOffset_e const longOffsets)
 {
     seq_t seq;
 
     U32 const llCode = FSE_peekSymbol(&seqState->stateLL);
     U32 const mlCode = FSE_peekSymbol(&seqState->stateML);
-    U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb);   /* <= maxOff, by table construction */
+    U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb);   /* <= MaxOff, by table construction */
 
     U32 const llBits = LL_bits[llCode];
     U32 const mlBits = ML_bits[mlCode];
@@ -1121,15 +1171,18 @@ FORCE_INLINE_TEMPLATE seq_t ZSTD_decodeSequenceLong_generic(seqState_t* seqState
                      0,        1,       1,       5,     0xD,     0x1D,     0x3D,     0x7D,
                      0xFD,   0x1FD,   0x3FD,   0x7FD,   0xFFD,   0x1FFD,   0x3FFD,   0x7FFD,
                      0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD,
-                     0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD };
+                     0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD, 0x1FFFFFFD, 0x3FFFFFFD, 0x7FFFFFFD };
 
     /* sequence */
     {   size_t offset;
         if (!ofCode)
             offset = 0;
         else {
-            if (longOffsets) {
-                int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN);
+            ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
+            ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
+            assert(ofBits <= MaxOff);
+            if (MEM_32bits() && longOffsets) {
+                U32 const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN_32-1);
                 offset = OF_base[ofCode] + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
                 if (MEM_32bits() || extraBits) BIT_reloadDStream(&seqState->DStream);
                 if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
@@ -1159,11 +1212,16 @@ FORCE_INLINE_TEMPLATE seq_t ZSTD_decodeSequenceLong_generic(seqState_t* seqState
     }
 
     seq.matchLength = ML_base[mlCode] + ((mlCode>31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0);  /* <=  16 bits */
-    if (MEM_32bits() && (mlBits+llBits>24)) BIT_reloadDStream(&seqState->DStream);
+    if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
+        BIT_reloadDStream(&seqState->DStream);
+    if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
+        BIT_reloadDStream(&seqState->DStream);
+    /* Verify that there is enough bits to read the rest of the data in 64-bit mode. */
+    ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
 
     seq.litLength = LL_base[llCode] + ((llCode>15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0);    /* <=  16 bits */
-    if (MEM_32bits() ||
-       (totalBits > 64 - 7 - (LLFSELog+MLFSELog+OffFSELog)) ) BIT_reloadDStream(&seqState->DStream);
+    if (MEM_32bits())
+        BIT_reloadDStream(&seqState->DStream);
 
     {   size_t const pos = seqState->pos + seq.litLength;
         seq.match = seqState->base + pos - seq.offset;    /* single memory segment */
@@ -1180,19 +1238,12 @@ FORCE_INLINE_TEMPLATE seq_t ZSTD_decodeSequenceLong_generic(seqState_t* seqState
     return seq;
 }
 
-static seq_t ZSTD_decodeSequenceLong(seqState_t* seqState, unsigned const windowSize) {
-    if (ZSTD_highbit32(windowSize) > STREAM_ACCUMULATOR_MIN) {
-        return ZSTD_decodeSequenceLong_generic(seqState, 1);
-    } else {
-        return ZSTD_decodeSequenceLong_generic(seqState, 0);
-    }
-}
 
 HINT_INLINE
 size_t ZSTD_execSequenceLong(BYTE* op,
-                                BYTE* const oend, seq_t sequence,
-                                const BYTE** litPtr, const BYTE* const litLimit,
-                                const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)
+                             BYTE* const oend, seq_t sequence,
+                             const BYTE** litPtr, const BYTE* const litLimit,
+                             const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)
 {
     BYTE* const oLitEnd = op + sequence.litLength;
     size_t const sequenceLength = sequence.litLength + sequence.matchLength;
@@ -1202,11 +1253,9 @@ size_t ZSTD_execSequenceLong(BYTE* op,
     const BYTE* match = sequence.match;
 
     /* check */
-#if 1
     if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
     if (iLitEnd > litLimit) return ERROR(corruption_detected);   /* over-read beyond lit buffer */
     if (oLitEnd>oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);
-#endif
 
     /* copy Literals */
     ZSTD_copy8(op, *litPtr);
@@ -1216,7 +1265,6 @@ size_t ZSTD_execSequenceLong(BYTE* op,
     *litPtr = iLitEnd;   /* update for next sequence */
 
     /* copy Match */
-#if 1
     if (sequence.offset > (size_t)(oLitEnd - base)) {
         /* offset beyond prefix */
         if (sequence.offset > (size_t)(oLitEnd - vBase)) return ERROR(corruption_detected);
@@ -1236,8 +1284,8 @@ size_t ZSTD_execSequenceLong(BYTE* op,
               return sequenceLength;
             }
     }   }
-    /* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */
-#endif
+    assert(op <= oend_w);
+    assert(sequence.matchLength >= MINMATCH);
 
     /* match within prefix */
     if (sequence.offset < 8) {
@@ -1273,7 +1321,8 @@ size_t ZSTD_execSequenceLong(BYTE* op,
 static size_t ZSTD_decompressSequencesLong(
                                ZSTD_DCtx* dctx,
                                void* dst, size_t maxDstSize,
-                         const void* seqStart, size_t seqSize)
+                         const void* seqStart, size_t seqSize,
+                         const ZSTD_longOffset_e isLongOffset)
 {
     const BYTE* ip = (const BYTE*)seqStart;
     const BYTE* const iend = ip + seqSize;
@@ -1285,7 +1334,6 @@ static size_t ZSTD_decompressSequencesLong(
     const BYTE* const base = (const BYTE*) (dctx->base);
     const BYTE* const vBase = (const BYTE*) (dctx->vBase);
     const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
-    unsigned const windowSize32 = (unsigned)dctx->fParams.windowSize;
     int nbSeq;
 
     /* Build Decoding Tables */
@@ -1315,13 +1363,13 @@ static size_t ZSTD_decompressSequencesLong(
 
         /* prepare in advance */
         for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && seqNb<seqAdvance; seqNb++) {
-            sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, windowSize32);
+            sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, isLongOffset);
         }
         if (seqNb<seqAdvance) return ERROR(corruption_detected);
 
         /* decode and decompress */
         for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && seqNb<nbSeq ; seqNb++) {
-            seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, windowSize32);
+            seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, isLongOffset);
             size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[(seqNb-ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);
             if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
             PREFETCH(sequence.match);
@@ -1355,10 +1403,20 @@ static size_t ZSTD_decompressSequencesLong(
 
 static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
                             void* dst, size_t dstCapacity,
-                      const void* src, size_t srcSize)
+                      const void* src, size_t srcSize, const int frame)
 {   /* blockType == blockCompressed */
     const BYTE* ip = (const BYTE*)src;
-    DEBUGLOG(5, "ZSTD_decompressBlock_internal");
+    /* isLongOffset must be true if there are long offsets.
+     * Offsets are long if they are larger than 2^STREAM_ACCUMULATOR_MIN.
+     * We don't expect that to be the case in 64-bit mode.
+     * If we are in block mode we don't know the window size, so we have to be
+     * conservative.
+     */
+    ZSTD_longOffset_e const isLongOffset = (ZSTD_longOffset_e)(MEM_32bits() && (!frame || dctx->fParams.windowSize > (1ULL << STREAM_ACCUMULATOR_MIN)));
+    /* windowSize could be any value at this point, since it is only validated
+     * in the streaming API.
+     */
+    DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);
 
     if (srcSize >= ZSTD_BLOCKSIZE_MAX) return ERROR(srcSize_wrong);
 
@@ -1369,13 +1427,9 @@ static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
         ip += litCSize;
         srcSize -= litCSize;
     }
-    if (sizeof(size_t) > 4)  /* do not enable prefetching on 32-bits x86, as it's performance detrimental */
-                             /* likely because of register pressure */
-                             /* if that's the correct cause, then 32-bits ARM should be affected differently */
-                             /* it would be good to test this on ARM real hardware, to see if prefetch version improves speed */
-        if (dctx->fParams.windowSize > (1<<23))
-            return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize);
-    return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize);
+    if (frame && dctx->fParams.windowSize > (1<<23))
+        return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, isLongOffset);
+    return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, isLongOffset);
 }
 
 
@@ -1395,7 +1449,7 @@ size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,
 {
     size_t dSize;
     ZSTD_checkContinuity(dctx, dst);
-    dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize);
+    dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 0);
     dctx->previousDstEnd = (char*)dst + dSize;
     return dSize;
 }
@@ -1411,7 +1465,7 @@ ZSTDLIB_API size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, siz
 }
 
 
-size_t ZSTD_generateNxBytes(void* dst, size_t dstCapacity, BYTE byte, size_t length)
+static size_t ZSTD_generateNxBytes(void* dst, size_t dstCapacity, BYTE byte, size_t length)
 {
     if (length > dstCapacity) return ERROR(dstSize_tooSmall);
     memset(dst, byte, length);
@@ -1431,7 +1485,7 @@ size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
 #endif
     if ( (srcSize >= ZSTD_skippableHeaderSize)
       && (MEM_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START ) {
-        return ZSTD_skippableHeaderSize + MEM_readLE32((const BYTE*)src + 4);
+        return ZSTD_skippableHeaderSize + MEM_readLE32((const BYTE*)src + ZSTD_frameIdSize);
     } else {
         const BYTE* ip = (const BYTE*)src;
         const BYTE* const ipstart = ip;
@@ -1511,7 +1565,7 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
         switch(blockProperties.blockType)
         {
         case bt_compressed:
-            decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend-op, ip, cBlockSize);
+            decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend-op, ip, cBlockSize, /* frame */ 1);
             break;
         case bt_raw :
             decodedSize = ZSTD_copyRawBlock(op, oend-op, ip, cBlockSize);
@@ -1533,6 +1587,10 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
         if (blockProperties.lastBlock) break;
     }
 
+    if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {
+        if ((U64)(op-ostart) != dctx->fParams.frameContentSize) {
+            return ERROR(corruption_detected);
+    }   }
     if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */
         U32 const checkCalc = (U32)XXH64_digest(&dctx->xxhState);
         U32 checkRead;
@@ -1590,13 +1648,15 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
 #endif
 
         magicNumber = MEM_readLE32(src);
+        DEBUGLOG(4, "reading magic number %08X (expecting %08X)",
+                    (U32)magicNumber, (U32)ZSTD_MAGICNUMBER);
         if (magicNumber != ZSTD_MAGICNUMBER) {
             if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
                 size_t skippableSize;
                 if (srcSize < ZSTD_skippableHeaderSize)
                     return ERROR(srcSize_wrong);
-                skippableSize = MEM_readLE32((const BYTE *)src + 4) +
-                                ZSTD_skippableHeaderSize;
+                skippableSize = MEM_readLE32((const BYTE*)src + ZSTD_frameIdSize)
+                              + ZSTD_skippableHeaderSize;
                 if (srcSize < skippableSize) return ERROR(srcSize_wrong);
 
                 src = (const BYTE *)src + skippableSize;
@@ -1699,33 +1759,31 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
 {
     DEBUGLOG(5, "ZSTD_decompressContinue");
     /* Sanity check */
-    if (srcSize != dctx->expected) return ERROR(srcSize_wrong);   /* unauthorized */
+    if (srcSize != dctx->expected) return ERROR(srcSize_wrong);  /* not allowed */
     if (dstCapacity) ZSTD_checkContinuity(dctx, dst);
 
     switch (dctx->stage)
     {
     case ZSTDds_getFrameHeaderSize :
-        if (srcSize != ZSTD_frameHeaderSize_prefix) return ERROR(srcSize_wrong);      /* unauthorized */
         assert(src != NULL);
-        if ((MEM_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {        /* skippable frame */
-            memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);
-            dctx->expected = ZSTD_skippableHeaderSize - ZSTD_frameHeaderSize_prefix;  /* magic number + skippable frame length */
-            dctx->stage = ZSTDds_decodeSkippableHeader;
-            return 0;
-        }
-        dctx->headerSize = ZSTD_frameHeaderSize(src, ZSTD_frameHeaderSize_prefix);
+        if (dctx->format == ZSTD_f_zstd1) {  /* allows header */
+            assert(srcSize >= ZSTD_frameIdSize);  /* to read skippable magic number */
+            if ((MEM_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {        /* skippable frame */
+                memcpy(dctx->headerBuffer, src, srcSize);
+                dctx->expected = ZSTD_skippableHeaderSize - srcSize;  /* remaining to load to get full skippable frame header */
+                dctx->stage = ZSTDds_decodeSkippableHeader;
+                return 0;
+        }   }
+        dctx->headerSize = ZSTD_frameHeaderSize_internal(src, srcSize, dctx->format);
         if (ZSTD_isError(dctx->headerSize)) return dctx->headerSize;
-        memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);
-        if (dctx->headerSize > ZSTD_frameHeaderSize_prefix) {
-            dctx->expected = dctx->headerSize - ZSTD_frameHeaderSize_prefix;
-            dctx->stage = ZSTDds_decodeFrameHeader;
-            return 0;
-        }
-        dctx->expected = 0;   /* not necessary to copy more */
-        /* fall-through */
+        memcpy(dctx->headerBuffer, src, srcSize);
+        dctx->expected = dctx->headerSize - srcSize;
+        dctx->stage = ZSTDds_decodeFrameHeader;
+        return 0;
+
     case ZSTDds_decodeFrameHeader:
         assert(src != NULL);
-        memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);
+        memcpy(dctx->headerBuffer + (dctx->headerSize - srcSize), src, srcSize);
         CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));
         dctx->expected = ZSTD_blockHeaderSize;
         dctx->stage = ZSTDds_decodeBlockHeader;
@@ -1757,6 +1815,7 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
             }
             return 0;
         }
+
     case ZSTDds_decompressLastBlock:
     case ZSTDds_decompressBlock:
         DEBUGLOG(5, "case ZSTDds_decompressBlock");
@@ -1765,7 +1824,7 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
             {
             case bt_compressed:
                 DEBUGLOG(5, "case bt_compressed");
-                rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize);
+                rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 1);
                 break;
             case bt_raw :
                 rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize);
@@ -1778,9 +1837,16 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
                 return ERROR(corruption_detected);
             }
             if (ZSTD_isError(rSize)) return rSize;
+            DEBUGLOG(5, "decoded size from block : %u", (U32)rSize);
+            dctx->decodedSize += rSize;
             if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, dst, rSize);
 
             if (dctx->stage == ZSTDds_decompressLastBlock) {   /* end of frame */
+                DEBUGLOG(4, "decoded size from frame : %u", (U32)dctx->decodedSize);
+                if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {
+                    if (dctx->decodedSize != dctx->fParams.frameContentSize) {
+                        return ERROR(corruption_detected);
+                }   }
                 if (dctx->fParams.checksumFlag) {  /* another round for frame checksum */
                     dctx->expected = 4;
                     dctx->stage = ZSTDds_checkChecksum;
@@ -1795,26 +1861,31 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
             }
             return rSize;
         }
+
     case ZSTDds_checkChecksum:
+        assert(srcSize == 4);  /* guaranteed by dctx->expected */
         {   U32 const h32 = (U32)XXH64_digest(&dctx->xxhState);
-            U32 const check32 = MEM_readLE32(src);   /* srcSize == 4, guaranteed by dctx->expected */
+            U32 const check32 = MEM_readLE32(src);
+            DEBUGLOG(4, "checksum : calculated %08X :: %08X read", h32, check32);
             if (check32 != h32) return ERROR(checksum_wrong);
             dctx->expected = 0;
             dctx->stage = ZSTDds_getFrameHeaderSize;
             return 0;
         }
+
     case ZSTDds_decodeSkippableHeader:
-        {   assert(src != NULL);
-            memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);
-            dctx->expected = MEM_readLE32(dctx->headerBuffer + 4);
-            dctx->stage = ZSTDds_skipFrame;
-            return 0;
-        }
+        assert(src != NULL);
+        assert(srcSize <= ZSTD_skippableHeaderSize);
+        memcpy(dctx->headerBuffer + (ZSTD_skippableHeaderSize - srcSize), src, srcSize);   /* complete skippable header */
+        dctx->expected = MEM_readLE32(dctx->headerBuffer + ZSTD_frameIdSize);   /* note : dctx->expected can grow seriously large, beyond local buffer size */
+        dctx->stage = ZSTDds_skipFrame;
+        return 0;
+
     case ZSTDds_skipFrame:
-        {   dctx->expected = 0;
-            dctx->stage = ZSTDds_getFrameHeaderSize;
-            return 0;
-        }
+        dctx->expected = 0;
+        dctx->stage = ZSTDds_getFrameHeaderSize;
+        return 0;
+
     default:
         return ERROR(GENERIC);   /* impossible */
     }
@@ -1895,7 +1966,7 @@ static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict
         if (magic != ZSTD_MAGIC_DICTIONARY) {
             return ZSTD_refDictContent(dctx, dict, dictSize);   /* pure content mode */
     }   }
-    dctx->dictID = MEM_readLE32((const char*)dict + 4);
+    dctx->dictID = MEM_readLE32((const char*)dict + ZSTD_frameIdSize);
 
     /* load entropy tables */
     {   size_t const eSize = ZSTD_loadEntropy(&dctx->entropy, dict, dictSize);
@@ -1909,6 +1980,29 @@ static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict
     return ZSTD_refDictContent(dctx, dict, dictSize);
 }
 
+/* Note : this function cannot fail */
+size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)
+{
+    assert(dctx != NULL);
+    dctx->expected = ZSTD_startingInputLength(dctx->format);  /* dctx->format must be properly set */
+    dctx->stage = ZSTDds_getFrameHeaderSize;
+    dctx->decodedSize = 0;
+    dctx->previousDstEnd = NULL;
+    dctx->base = NULL;
+    dctx->vBase = NULL;
+    dctx->dictEnd = NULL;
+    dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001);  /* cover both little and big endian */
+    dctx->litEntropy = dctx->fseEntropy = 0;
+    dctx->dictID = 0;
+    ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
+    memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue));  /* initial repcodes */
+    dctx->LLTptr = dctx->entropy.LLTable;
+    dctx->MLTptr = dctx->entropy.MLTable;
+    dctx->OFTptr = dctx->entropy.OFTable;
+    dctx->HUFptr = dctx->entropy.hufTable;
+    return 0;
+}
+
 size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
 {
     CHECK_F( ZSTD_decompressBegin(dctx) );
@@ -1975,7 +2069,7 @@ static size_t ZSTD_loadEntropy_inDDict(ZSTD_DDict* ddict)
     {   U32 const magic = MEM_readLE32(ddict->dictContent);
         if (magic != ZSTD_MAGIC_DICTIONARY) return 0;   /* pure content mode */
     }
-    ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + 4);
+    ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_frameIdSize);
 
     /* load entropy tables */
     CHECK_E( ZSTD_loadEntropy(&ddict->entropy, ddict->dictContent, ddict->dictSize), dictionary_corrupted );
@@ -1984,9 +2078,9 @@ static size_t ZSTD_loadEntropy_inDDict(ZSTD_DDict* ddict)
 }
 
 
-static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict, const void* dict, size_t dictSize, unsigned byReference)
+static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod)
 {
-    if ((byReference) || (!dict) || (!dictSize)) {
+    if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dict) || (!dictSize)) {
         ddict->dictBuffer = NULL;
         ddict->dictContent = dict;
     } else {
@@ -2005,7 +2099,7 @@ static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict, const void* dict, size_
     return 0;
 }
 
-ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize, unsigned byReference, ZSTD_customMem customMem)
+ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_customMem customMem)
 {
     if (!customMem.customAlloc ^ !customMem.customFree) return NULL;
 
@@ -2013,7 +2107,7 @@ ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize, unsigne
         if (!ddict) return NULL;
         ddict->cMem = customMem;
 
-        if (ZSTD_isError( ZSTD_initDDict_internal(ddict, dict, dictSize, byReference) )) {
+        if (ZSTD_isError( ZSTD_initDDict_internal(ddict, dict, dictSize, dictLoadMethod) )) {
             ZSTD_freeDDict(ddict);
             return NULL;
         }
@@ -2029,7 +2123,7 @@ ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize, unsigne
 ZSTD_DDict* ZSTD_createDDict(const void* dict, size_t dictSize)
 {
     ZSTD_customMem const allocator = { NULL, NULL, NULL };
-    return ZSTD_createDDict_advanced(dict, dictSize, 0, allocator);
+    return ZSTD_createDDict_advanced(dict, dictSize, ZSTD_dlm_byCopy, allocator);
 }
 
 /*! ZSTD_createDDict_byReference() :
@@ -2039,25 +2133,26 @@ ZSTD_DDict* ZSTD_createDDict(const void* dict, size_t dictSize)
 ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize)
 {
     ZSTD_customMem const allocator = { NULL, NULL, NULL };
-    return ZSTD_createDDict_advanced(dictBuffer, dictSize, 1, allocator);
+    return ZSTD_createDDict_advanced(dictBuffer, dictSize, ZSTD_dlm_byRef, allocator);
 }
 
 
 ZSTD_DDict* ZSTD_initStaticDDict(void* workspace, size_t workspaceSize,
                                  const void* dict, size_t dictSize,
-                                 unsigned byReference)
+                                 ZSTD_dictLoadMethod_e dictLoadMethod)
 {
-    size_t const neededSpace = sizeof(ZSTD_DDict) + (byReference ? 0 : dictSize);
+    size_t const neededSpace =
+            sizeof(ZSTD_DDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
     ZSTD_DDict* const ddict = (ZSTD_DDict*)workspace;
     assert(workspace != NULL);
     assert(dict != NULL);
     if ((size_t)workspace & 7) return NULL;  /* 8-aligned */
     if (workspaceSize < neededSpace) return NULL;
-    if (!byReference) {
+    if (dictLoadMethod == ZSTD_dlm_byCopy) {
         memcpy(ddict+1, dict, dictSize);  /* local copy */
         dict = ddict+1;
     }
-    if (ZSTD_isError( ZSTD_initDDict_internal(ddict, dict, dictSize, 1 /* byRef */) ))
+    if (ZSTD_isError( ZSTD_initDDict_internal(ddict, dict, dictSize, ZSTD_dlm_byRef) ))
         return NULL;
     return ddict;
 }
@@ -2075,10 +2170,10 @@ size_t ZSTD_freeDDict(ZSTD_DDict* ddict)
 
 /*! ZSTD_estimateDDictSize() :
  *  Estimate amount of memory that will be needed to create a dictionary for decompression.
- *  Note : dictionary created "byReference" are smaller */
-size_t ZSTD_estimateDDictSize(size_t dictSize, unsigned byReference)
+ *  Note : dictionary created by reference using ZSTD_dlm_byRef are smaller */
+size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod)
 {
-    return sizeof(ZSTD_DDict) + (byReference ? 0 : dictSize);
+    return sizeof(ZSTD_DDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
 }
 
 size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict)
@@ -2095,7 +2190,7 @@ unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize)
 {
     if (dictSize < 8) return 0;
     if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) return 0;
-    return MEM_readLE32((const char*)dict + 4);
+    return MEM_readLE32((const char*)dict + ZSTD_frameIdSize);
 }
 
 /*! ZSTD_getDictID_fromDDict() :
@@ -2123,7 +2218,7 @@ unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict)
  *  ZSTD_getFrameHeader(), which will provide a more precise error code. */
 unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize)
 {
-    ZSTD_frameHeader zfp = { 0, 0, ZSTD_frame, 0, 0, 0 };
+    ZSTD_frameHeader zfp = { 0, 0, 0, ZSTD_frame, 0, 0, 0 };
     size_t const hError = ZSTD_getFrameHeader(&zfp, src, srcSize);
     if (ZSTD_isError(hError)) return 0;
     return zfp.dictID;
@@ -2190,13 +2285,15 @@ size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t di
     return ZSTD_frameHeaderSize_prefix;
 }
 
+/* note : this variant can't fail */
 size_t ZSTD_initDStream(ZSTD_DStream* zds)
 {
     return ZSTD_initDStream_usingDict(zds, NULL, 0);
 }
 
 /* ZSTD_initDStream_usingDDict() :
- * ddict will just be referenced, and must outlive decompression session */
+ * ddict will just be referenced, and must outlive decompression session
+ * this function cannot fail */
 size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict)
 {
     size_t const initResult = ZSTD_initDStream(zds);
@@ -2216,31 +2313,66 @@ size_t ZSTD_resetDStream(ZSTD_DStream* zds)
 size_t ZSTD_setDStreamParameter(ZSTD_DStream* zds,
                                 ZSTD_DStreamParameter_e paramType, unsigned paramValue)
 {
+    ZSTD_STATIC_ASSERT((unsigned)zdss_loadHeader >= (unsigned)zdss_init);
+    if ((unsigned)zds->streamStage > (unsigned)zdss_loadHeader)
+        return ERROR(stage_wrong);
     switch(paramType)
     {
         default : return ERROR(parameter_unsupported);
-        case DStream_p_maxWindowSize : zds->maxWindowSize = paramValue ? paramValue : (U32)(-1); break;
+        case DStream_p_maxWindowSize :
+            DEBUGLOG(4, "setting maxWindowSize = %u KB", paramValue >> 10);
+            zds->maxWindowSize = paramValue ? paramValue : (U32)(-1);
+            break;
     }
     return 0;
 }
 
+size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize)
+{
+    ZSTD_STATIC_ASSERT((unsigned)zdss_loadHeader >= (unsigned)zdss_init);
+    if ((unsigned)dctx->streamStage > (unsigned)zdss_loadHeader)
+        return ERROR(stage_wrong);
+    dctx->maxWindowSize = maxWindowSize;
+    return 0;
+}
+
+size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format)
+{
+    DEBUGLOG(4, "ZSTD_DCtx_setFormat : %u", (unsigned)format);
+    ZSTD_STATIC_ASSERT((unsigned)zdss_loadHeader >= (unsigned)zdss_init);
+    if ((unsigned)dctx->streamStage > (unsigned)zdss_loadHeader)
+        return ERROR(stage_wrong);
+    dctx->format = format;
+    return 0;
+}
+
 
 size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds)
 {
     return ZSTD_sizeof_DCtx(zds);
 }
 
+size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
+{
+    size_t const blockSize = (size_t) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
+    unsigned long long const neededRBSize = windowSize + blockSize + (WILDCOPY_OVERLENGTH * 2);
+    unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);
+    size_t const minRBSize = (size_t) neededSize;
+    if ((unsigned long long)minRBSize != neededSize) return ERROR(frameParameter_windowTooLarge);
+    return minRBSize;
+}
+
 size_t ZSTD_estimateDStreamSize(size_t windowSize)
 {
     size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
     size_t const inBuffSize = blockSize;  /* no block can be larger */
-    size_t const outBuffSize = windowSize + blockSize + (WILDCOPY_OVERLENGTH * 2);
+    size_t const outBuffSize = ZSTD_decodingBufferSize_min(windowSize, ZSTD_CONTENTSIZE_UNKNOWN);
     return ZSTD_estimateDCtxSize() + inBuffSize + outBuffSize;
 }
 
-ZSTDLIB_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize)
+size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize)
 {
-    U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX;
+    U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX;   /* note : should be user-selectable */
     ZSTD_frameHeader zfh;
     size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize);
     if (ZSTD_isError(err)) return err;
@@ -2272,7 +2404,18 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
     U32 someMoreWork = 1;
 
     DEBUGLOG(5, "ZSTD_decompressStream");
+    if (input->pos > input->size) {  /* forbidden */
+        DEBUGLOG(5, "in: pos: %u   vs size: %u",
+                    (U32)input->pos, (U32)input->size);
+        return ERROR(srcSize_wrong);
+    }
+    if (output->pos > output->size) {  /* forbidden */
+        DEBUGLOG(5, "out: pos: %u   vs size: %u",
+                    (U32)output->pos, (U32)output->size);
+        return ERROR(dstSize_tooSmall);
+    }
     DEBUGLOG(5, "input size : %u", (U32)(input->size - input->pos));
+
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
     if (zds->legacyVersion) {
         /* legacy support is incompatible with static dctx */
@@ -2289,7 +2432,9 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
             /* fall-through */
 
         case zdss_loadHeader :
-            {   size_t const hSize = ZSTD_getFrameHeader(&zds->fParams, zds->headerBuffer, zds->lhSize);
+            DEBUGLOG(5, "stage zdss_loadHeader (srcSize : %u)", (U32)(iend - ip));
+            {   size_t const hSize = ZSTD_getFrameHeader_internal(&zds->fParams, zds->headerBuffer, zds->lhSize, zds->format);
+                DEBUGLOG(5, "header size : %u", (U32)hSize);
                 if (ZSTD_isError(hSize)) {
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
                     U32 const legacyVersion = ZSTD_isLegacy(istart, iend-istart);
@@ -2342,7 +2487,7 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
             CHECK_F(ZSTD_decompressBegin_usingDDict(zds, zds->ddict));
 
             if ((MEM_readLE32(zds->headerBuffer) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {  /* skippable frame */
-                zds->expected = MEM_readLE32(zds->headerBuffer + 4);
+                zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_frameIdSize);
                 zds->stage = ZSTDds_skipFrame;
             } else {
                 CHECK_F(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize));
@@ -2351,20 +2496,20 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
             }
 
             /* control buffer memory usage */
-            DEBUGLOG(4, "Control max buffer memory usage");
+            DEBUGLOG(4, "Control max buffer memory usage (max %u KB)",
+                        (U32)(zds->maxWindowSize >> 10));
             zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
             if (zds->fParams.windowSize > zds->maxWindowSize) return ERROR(frameParameter_windowTooLarge);
 
             /* Adapt buffer sizes to frame header instructions */
-            {   size_t const blockSize = (size_t)(MIN(zds->fParams.windowSize, ZSTD_BLOCKSIZE_MAX));
-                size_t const neededOutSize = (size_t)(zds->fParams.windowSize + blockSize + WILDCOPY_OVERLENGTH * 2);
-                zds->blockSize = blockSize;
-                if ((zds->inBuffSize < blockSize) || (zds->outBuffSize < neededOutSize)) {
-                    size_t const bufferSize = blockSize + neededOutSize;
+            {   size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);
+                size_t const neededOutBuffSize = ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize);
+                if ((zds->inBuffSize < neededInBuffSize) || (zds->outBuffSize < neededOutBuffSize)) {
+                    size_t const bufferSize = neededInBuffSize + neededOutBuffSize;
                     DEBUGLOG(4, "inBuff  : from %u to %u",
-                                (U32)zds->inBuffSize, (U32)blockSize);
+                                (U32)zds->inBuffSize, (U32)neededInBuffSize);
                     DEBUGLOG(4, "outBuff : from %u to %u",
-                                (U32)zds->outBuffSize, (U32)neededOutSize);
+                                (U32)zds->outBuffSize, (U32)neededOutBuffSize);
                     if (zds->staticSize) {  /* static DCtx */
                         DEBUGLOG(4, "staticSize : %u", (U32)zds->staticSize);
                         assert(zds->staticSize >= sizeof(ZSTD_DCtx));  /* controlled at init */
@@ -2377,9 +2522,9 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
                         zds->inBuff = (char*)ZSTD_malloc(bufferSize, zds->customMem);
                         if (zds->inBuff == NULL) return ERROR(memory_allocation);
                     }
-                    zds->inBuffSize = blockSize;
+                    zds->inBuffSize = neededInBuffSize;
                     zds->outBuff = zds->inBuff + zds->inBuffSize;
-                    zds->outBuffSize = neededOutSize;
+                    zds->outBuffSize = neededOutBuffSize;
             }   }
             zds->streamStage = zdss_read;
             /* fall-through */
@@ -2437,8 +2582,13 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
                 zds->outStart += flushedSize;
                 if (flushedSize == toFlushSize) {  /* flush completed */
                     zds->streamStage = zdss_read;
-                    if (zds->outStart + zds->blockSize > zds->outBuffSize)
+                    if ( (zds->outBuffSize < zds->fParams.frameContentSize)
+                      && (zds->outStart + zds->fParams.blockSizeMax > zds->outBuffSize) ) {
+                        DEBUGLOG(5, "restart filling outBuff from beginning (left:%i, needed:%u)",
+                                (int)(zds->outBuffSize - zds->outStart),
+                                (U32)zds->fParams.blockSizeMax);
                         zds->outStart = zds->outEnd = 0;
+                    }
                     break;
             }   }
             /* cannot complete flush */
@@ -2476,3 +2626,30 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
         return nextSrcSizeHint;
     }
 }
+
+
+size_t ZSTD_decompress_generic(ZSTD_DCtx* dctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
+{
+    return ZSTD_decompressStream(dctx, output, input);
+}
+
+size_t ZSTD_decompress_generic_simpleArgs (
+                            ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity, size_t* dstPos,
+                      const void* src, size_t srcSize, size_t* srcPos)
+{
+    ZSTD_outBuffer output = { dst, dstCapacity, *dstPos };
+    ZSTD_inBuffer  input  = { src, srcSize, *srcPos };
+    /* ZSTD_compress_generic() will check validity of dstPos and srcPos */
+    size_t const cErr = ZSTD_decompress_generic(dctx, &output, &input);
+    *dstPos = output.pos;
+    *srcPos = input.pos;
+    return cErr;
+}
+
+void ZSTD_DCtx_reset(ZSTD_DCtx* dctx)
+{
+    (void)ZSTD_initDStream(dctx);
+    dctx->format = ZSTD_f_zstd1;
+    dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
+}
diff --git a/thirdparty/zstd/zstd.h b/thirdparty/zstd/zstd.h
index 13b4563fd6..2194a3b23d 100644
--- a/thirdparty/zstd/zstd.h
+++ b/thirdparty/zstd/zstd.h
@@ -5,6 +5,7 @@
  * 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.
  */
 #if defined (__cplusplus)
 extern "C" {
@@ -58,7 +59,7 @@ extern "C" {
 /*------   Version   ------*/
 #define ZSTD_VERSION_MAJOR    1
 #define ZSTD_VERSION_MINOR    3
-#define ZSTD_VERSION_RELEASE  1
+#define ZSTD_VERSION_RELEASE  2
 
 #define ZSTD_VERSION_NUMBER  (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)
 ZSTDLIB_API unsigned ZSTD_versionNumber(void);   /**< useful to check dll version */
@@ -130,10 +131,11 @@ ZSTDLIB_API unsigned long long ZSTD_getDecompressedSize(const void* src, size_t
 
 
 /*======  Helper functions  ======*/
-ZSTDLIB_API int         ZSTD_maxCLevel(void);               /*!< maximum compression level available */
+#define ZSTD_COMPRESSBOUND(srcSize)   ((srcSize) + ((srcSize)>>8) + (((srcSize) < 128 KB) ? ((128 KB - (srcSize)) >> 11) /* margin, from 64 to 0 */ : 0))  /* this formula ensures that bound(A) + bound(B) <= bound(A+B) as long as A and B >= 128 KB */
 ZSTDLIB_API size_t      ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case scenario */
 ZSTDLIB_API unsigned    ZSTD_isError(size_t code);          /*!< tells if a `size_t` function result is an error code */
 ZSTDLIB_API const char* ZSTD_getErrorName(size_t code);     /*!< provides readable string from an error code */
+ZSTDLIB_API int         ZSTD_maxCLevel(void);               /*!< maximum compression level available */
 
 
 /***************************************
@@ -375,27 +377,31 @@ ZSTDLIB_API size_t ZSTD_DStreamOutSize(void);   /*!< recommended size for output
 #define ZSTD_MAGIC_SKIPPABLE_START  0x184D2A50U
 #define ZSTD_MAGIC_DICTIONARY       0xEC30A437   /* v0.7+ */
 
-#define ZSTD_WINDOWLOG_MAX_32  27
-#define ZSTD_WINDOWLOG_MAX_64  27
+#define ZSTD_WINDOWLOG_MAX_32   30
+#define ZSTD_WINDOWLOG_MAX_64   31
 #define ZSTD_WINDOWLOG_MAX    ((unsigned)(sizeof(size_t) == 4 ? ZSTD_WINDOWLOG_MAX_32 : ZSTD_WINDOWLOG_MAX_64))
-#define ZSTD_WINDOWLOG_MIN     10
-#define ZSTD_HASHLOG_MAX       ZSTD_WINDOWLOG_MAX
-#define ZSTD_HASHLOG_MIN        6
-#define ZSTD_CHAINLOG_MAX     (ZSTD_WINDOWLOG_MAX+1)
-#define ZSTD_CHAINLOG_MIN      ZSTD_HASHLOG_MIN
-#define ZSTD_HASHLOG3_MAX      17
-#define ZSTD_SEARCHLOG_MAX    (ZSTD_WINDOWLOG_MAX-1)
-#define ZSTD_SEARCHLOG_MIN      1
-#define ZSTD_SEARCHLENGTH_MAX   7   /* only for ZSTD_fast, other strategies are limited to 6 */
-#define ZSTD_SEARCHLENGTH_MIN   3   /* only for ZSTD_btopt, other strategies are limited to 4 */
-#define ZSTD_TARGETLENGTH_MIN   4
-#define ZSTD_TARGETLENGTH_MAX 999
-
-#define ZSTD_FRAMEHEADERSIZE_MAX 18    /* for static allocation */
-#define ZSTD_FRAMEHEADERSIZE_MIN  6
-static const size_t ZSTD_frameHeaderSize_prefix = 5;  /* minimum input size to know frame header size */
-static const size_t ZSTD_frameHeaderSize_max = ZSTD_FRAMEHEADERSIZE_MAX;
+#define ZSTD_WINDOWLOG_MIN      10
+#define ZSTD_HASHLOG_MAX        MIN(ZSTD_WINDOWLOG_MAX, 30)
+#define ZSTD_HASHLOG_MIN         6
+#define ZSTD_CHAINLOG_MAX       MIN(ZSTD_WINDOWLOG_MAX+1, 30)
+#define ZSTD_CHAINLOG_MIN       ZSTD_HASHLOG_MIN
+#define ZSTD_HASHLOG3_MAX       17
+#define ZSTD_SEARCHLOG_MAX     (ZSTD_WINDOWLOG_MAX-1)
+#define ZSTD_SEARCHLOG_MIN       1
+#define ZSTD_SEARCHLENGTH_MAX    7   /* only for ZSTD_fast, other strategies are limited to 6 */
+#define ZSTD_SEARCHLENGTH_MIN    3   /* only for ZSTD_btopt, other strategies are limited to 4 */
+#define ZSTD_TARGETLENGTH_MIN    4   /* only useful for btopt */
+#define ZSTD_TARGETLENGTH_MAX  999   /* only useful for btopt */
+#define ZSTD_LDM_MINMATCH_MIN    4
+#define ZSTD_LDM_MINMATCH_MAX 4096
+#define ZSTD_LDM_BUCKETSIZELOG_MAX 8
+
+#define ZSTD_FRAMEHEADERSIZE_PREFIX 5   /* minimum input size to know frame header size */
+#define ZSTD_FRAMEHEADERSIZE_MIN    6
+#define ZSTD_FRAMEHEADERSIZE_MAX   18   /* for static allocation */
+static const size_t ZSTD_frameHeaderSize_prefix = ZSTD_FRAMEHEADERSIZE_PREFIX;
 static const size_t ZSTD_frameHeaderSize_min = ZSTD_FRAMEHEADERSIZE_MIN;
+static const size_t ZSTD_frameHeaderSize_max = ZSTD_FRAMEHEADERSIZE_MAX;
 static const size_t ZSTD_skippableHeaderSize = 8;  /* magic number + skippable frame length */
 
 
@@ -424,6 +430,8 @@ typedef struct {
     ZSTD_frameParameters fParams;
 } ZSTD_parameters;
 
+typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params;
+
 /*= Custom memory allocation functions */
 typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size);
 typedef void  (*ZSTD_freeFunction) (void* opaque, void* address);
@@ -480,7 +488,7 @@ ZSTDLIB_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
 
 /*! ZSTD_sizeof_*() :
  *  These functions give the current memory usage of selected object.
- *  Object memory usage can evolve if it's re-used multiple times. */
+ *  Object memory usage can evolve when re-used multiple times. */
 ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
 ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
 ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
@@ -493,18 +501,21 @@ ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
  *  of a future {D,C}Ctx, before its creation.
  *  ZSTD_estimateCCtxSize() will provide a budget large enough for any compression level up to selected one.
  *  It will also consider src size to be arbitrarily "large", which is worst case.
- *  If srcSize is known to always be small, ZSTD_estimateCCtxSize_advanced() can provide a tighter estimation.
- *  ZSTD_estimateCCtxSize_advanced() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
+ *  If srcSize is known to always be small, ZSTD_estimateCCtxSize_usingCParams() can provide a tighter estimation.
+ *  ZSTD_estimateCCtxSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
+ *  ZSTD_estimateCCtxSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_p_nbThreads is > 1.
  *  Note : CCtx estimation is only correct for single-threaded compression */
 ZSTDLIB_API size_t ZSTD_estimateCCtxSize(int compressionLevel);
-ZSTDLIB_API size_t ZSTD_estimateCCtxSize_advanced(ZSTD_compressionParameters cParams);
+ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
+ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
 ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);
 
-/*! ZSTD_estimate?StreamSize() :
+/*! ZSTD_estimateCStreamSize() :
  *  ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one.
  *  It will also consider src size to be arbitrarily "large", which is worst case.
- *  If srcSize is known to always be small, ZSTD_estimateCStreamSize_advanced() can provide a tighter estimation.
- *  ZSTD_estimateCStreamSize_advanced() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
+ *  If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.
+ *  ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
+ *  ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_p_nbThreads is set to a value > 1.
  *  Note : CStream estimation is only correct for single-threaded compression.
  *  ZSTD_DStream memory budget depends on window Size.
  *  This information can be passed manually, using ZSTD_estimateDStreamSize,
@@ -513,17 +524,24 @@ ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);
  *         an internal ?Dict will be created, which additional size is not estimated here.
  *         In this case, get total size by adding ZSTD_estimate?DictSize */
 ZSTDLIB_API size_t ZSTD_estimateCStreamSize(int compressionLevel);
-ZSTDLIB_API size_t ZSTD_estimateCStreamSize_advanced(ZSTD_compressionParameters cParams);
+ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);
+ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);
 ZSTDLIB_API size_t ZSTD_estimateDStreamSize(size_t windowSize);
 ZSTDLIB_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);
 
+typedef enum {
+    ZSTD_dlm_byCopy = 0,     /**< Copy dictionary content internally */
+    ZSTD_dlm_byRef,          /**< Reference dictionary content -- the dictionary buffer must outlive its users. */
+} ZSTD_dictLoadMethod_e;
+
 /*! ZSTD_estimate?DictSize() :
  *  ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict().
- *  ZSTD_estimateCStreamSize_advanced() makes it possible to control precisely compression parameters, like ZSTD_createCDict_advanced().
- *  Note : dictionary created "byReference" are smaller */
+ *  ZSTD_estimateCStreamSize_advanced_usingCParams() makes it possible to control precisely compression parameters, like ZSTD_createCDict_advanced().
+ *  Note : dictionary created by reference using ZSTD_dlm_byRef are smaller
+ */
 ZSTDLIB_API size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel);
-ZSTDLIB_API size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, unsigned byReference);
-ZSTDLIB_API size_t ZSTD_estimateDDictSize(size_t dictSize, unsigned byReference);
+ZSTDLIB_API size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod);
+ZSTDLIB_API size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod);
 
 
 /***************************************
@@ -551,24 +569,12 @@ ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem);
 ZSTDLIB_API ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize);
 
 
-/* !!! To be deprecated !!! */
-typedef enum {
-    ZSTD_p_forceWindow,   /* Force back-references to remain < windowSize, even when referencing Dictionary content (default:0) */
-    ZSTD_p_forceRawDict   /* Force loading dictionary in "content-only" mode (no header analysis) */
-} ZSTD_CCtxParameter;
-/*! ZSTD_setCCtxParameter() :
- *  Set advanced parameters, selected through enum ZSTD_CCtxParameter
- *  @result : 0, or an error code (which can be tested with ZSTD_isError()) */
-ZSTDLIB_API size_t ZSTD_setCCtxParameter(ZSTD_CCtx* cctx, ZSTD_CCtxParameter param, unsigned value);
-
-
 /*! ZSTD_createCDict_byReference() :
  *  Create a digested dictionary for compression
  *  Dictionary content is simply referenced, and therefore stays in dictBuffer.
  *  It is important that dictBuffer outlives CDict, it must remain read accessible throughout the lifetime of CDict */
 ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
 
-
 typedef enum { ZSTD_dm_auto=0,        /* dictionary is "full" if it starts with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */
                ZSTD_dm_rawContent,    /* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */
                ZSTD_dm_fullDict       /* refuses to load a dictionary if it does not respect Zstandard's specification */
@@ -576,7 +582,8 @@ typedef enum { ZSTD_dm_auto=0,        /* dictionary is "full" if it starts with
 /*! ZSTD_createCDict_advanced() :
  *  Create a ZSTD_CDict using external alloc and free, and customized compression parameters */
 ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize,
-                                                  unsigned byReference, ZSTD_dictMode_e dictMode,
+                                                  ZSTD_dictLoadMethod_e dictLoadMethod,
+                                                  ZSTD_dictMode_e dictMode,
                                                   ZSTD_compressionParameters cParams,
                                                   ZSTD_customMem customMem);
 
@@ -596,7 +603,7 @@ ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictS
 ZSTDLIB_API ZSTD_CDict* ZSTD_initStaticCDict(
                             void* workspace, size_t workspaceSize,
                       const void* dict, size_t dictSize,
-                            unsigned byReference, ZSTD_dictMode_e dictMode,
+                            ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode,
                             ZSTD_compressionParameters cParams);
 
 /*! ZSTD_getCParams() :
@@ -674,7 +681,8 @@ ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, siz
 /*! ZSTD_createDDict_advanced() :
  *  Create a ZSTD_DDict using external alloc and free, optionally by reference */
 ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,
-                                                  unsigned byReference, ZSTD_customMem customMem);
+                                                  ZSTD_dictLoadMethod_e dictLoadMethod,
+                                                  ZSTD_customMem customMem);
 
 /*! ZSTD_initStaticDDict() :
  *  Generate a digested dictionary in provided memory area.
@@ -689,7 +697,7 @@ ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictS
  */
 ZSTDLIB_API ZSTD_DDict* ZSTD_initStaticDDict(void* workspace, size_t workspaceSize,
                                              const void* dict, size_t dictSize,
-                                             unsigned byReference);
+                                             ZSTD_dictLoadMethod_e dictLoadMethod);
 
 /*! ZSTD_getDictID_fromDict() :
  *  Provides the dictID stored within dictionary.
@@ -724,9 +732,9 @@ ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
 ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem);
 ZSTDLIB_API ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize);    /**< same as ZSTD_initStaticCCtx() */
 ZSTDLIB_API size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize);   /**< pledgedSrcSize must be correct, a size of 0 means unknown.  for a frame size of 0 use initCStream_advanced */
-ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel); /**< creates of an internal CDict (incompatible with static CCtx), except if dict == NULL or dictSize < 8, in which case no dict is used. */
+ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel); /**< creates of an internal CDict (incompatible with static CCtx), except if dict == NULL or dictSize < 8, in which case no dict is used. Note: dict is loaded with ZSTD_dm_auto (treated as a full zstd dictionary if it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.*/
 ZSTDLIB_API size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize,
-                                             ZSTD_parameters params, unsigned long long pledgedSrcSize);  /**< pledgedSrcSize is optional and can be 0 (meaning unknown). note: if the contentSizeFlag is set, pledgedSrcSize == 0 means the source size is actually 0 */
+                                             ZSTD_parameters params, unsigned long long pledgedSrcSize);  /**< pledgedSrcSize is optional and can be 0 (meaning unknown). note: if the contentSizeFlag is set, pledgedSrcSize == 0 means the source size is actually 0. dict is loaded with ZSTD_dm_auto and ZSTD_dlm_byCopy. */
 ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);  /**< note : cdict will just be referenced, and must outlive compression session */
 ZSTDLIB_API size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize);  /**< same as ZSTD_initCStream_usingCDict(), with control over frame parameters */
 
@@ -741,12 +749,12 @@ ZSTDLIB_API size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledg
 
 
 /*=====   Advanced Streaming decompression functions  =====*/
-typedef enum { DStream_p_maxWindowSize } ZSTD_DStreamParameter_e;
 ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem);
 ZSTDLIB_API ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize);    /**< same as ZSTD_initStaticDCtx() */
-ZSTDLIB_API size_t ZSTD_setDStreamParameter(ZSTD_DStream* zds, ZSTD_DStreamParameter_e paramType, unsigned paramValue);
-ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize); /**< note: a dict will not be used if dict == NULL or dictSize < 8 */
-ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);  /**< note : ddict will just be referenced, and must outlive decompression session */
+typedef enum { DStream_p_maxWindowSize } ZSTD_DStreamParameter_e;
+ZSTDLIB_API size_t ZSTD_setDStreamParameter(ZSTD_DStream* zds, ZSTD_DStreamParameter_e paramType, unsigned paramValue);   /* obsolete : this API will be removed in a future version */
+ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize); /**< note: no dictionary will be used if dict == NULL or dictSize < 8 */
+ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);  /**< note : ddict is referenced, it must outlive decompression session */
 ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);  /**< re-use decompression parameters from previous init; saves dictionary loading */
 
 
@@ -754,8 +762,8 @@ ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);  /**< re-use decompress
 *  Buffer-less and synchronous inner streaming functions
 *
 *  This is an advanced API, giving full control over buffer management, for users which need direct control over memory.
-*  But it's also a complex one, with many restrictions (documented below).
-*  Prefer using normal streaming API for an easier experience
+*  But it's also a complex one, with several restrictions, documented below.
+*  Prefer normal streaming API for an easier experience.
 ********************************************************************* */
 
 /**
@@ -772,8 +780,8 @@ ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);  /**< re-use decompress
 
   Then, consume your input using ZSTD_compressContinue().
   There are some important considerations to keep in mind when using this advanced function :
-  - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffer only.
-  - Interface is synchronous : input is consumed entirely and produce 1+ (or more) compressed blocks.
+  - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffers only.
+  - Interface is synchronous : input is consumed entirely and produces 1+ compressed blocks.
   - Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario.
     Worst case evaluation is provided by ZSTD_compressBound().
     ZSTD_compressContinue() doesn't guarantee recover after a failed compression.
@@ -784,9 +792,9 @@ ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);  /**< re-use decompress
 
   Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum.
   It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame.
-  Without last block mark, frames will be considered unfinished (corrupted) by decoders.
+  Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders.
 
-  `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress some new frame.
+  `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress again.
 */
 
 /*=====   Buffer-less streaming compression functions  =====*/
@@ -809,40 +817,53 @@ ZSTDLIB_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapaci
   A ZSTD_DCtx object can be re-used multiple times.
 
   First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader().
-  It fills a ZSTD_frameHeader structure with important information to correctly decode the frame,
-  such as minimum rolling buffer size to allocate to decompress data (`windowSize`),
-  and the dictionary ID in use.
-  (Note : content size is optional, it may not be present. 0 means : content size unknown).
-  Note that these values could be wrong, either because of data malformation, or because an attacker is spoofing deliberate false information.
-  As a consequence, check that values remain within valid application range, especially `windowSize`, before allocation.
-  Each application can set its own limit, depending on local restrictions.
-  For extended interoperability, it is recommended to support windowSize of at least 8 MB.
   Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough.
   Data fragment must be large enough to ensure successful decoding.
-  `ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough.
+ `ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough.
   @result : 0 : successful decoding, the `ZSTD_frameHeader` structure is correctly filled.
            >0 : `srcSize` is too small, please provide at least @result bytes on next attempt.
            errorCode, which can be tested using ZSTD_isError().
 
-  Start decompression, with ZSTD_decompressBegin().
+  It fills a ZSTD_frameHeader structure with important information to correctly decode the frame,
+  such as the dictionary ID, content size, or maximum back-reference distance (`windowSize`).
+  Note that these values could be wrong, either because of data corruption, or because a 3rd party deliberately spoofs false information.
+  As a consequence, check that values remain within valid application range.
+  For example, do not allocate memory blindly, check that `windowSize` is within expectation.
+  Each application can set its own limits, depending on local restrictions.
+  For extended interoperability, it is recommended to support `windowSize` of at least 8 MB.
+
+  ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize` bytes.
+  ZSTD_decompressContinue() is very sensitive to contiguity,
+  if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,
+  or that previous contiguous segment is large enough to properly handle maximum back-reference distance.
+  There are multiple ways to guarantee this condition.
+
+  The most memory efficient way is to use a round buffer of sufficient size.
+  Sufficient size is determined by invoking ZSTD_decodingBufferSize_min(),
+  which can @return an error code if required value is too large for current system (in 32-bits mode).
+  In a round buffer methodology, ZSTD_decompressContinue() decompresses each block next to previous one,
+  up to the moment there is not enough room left in the buffer to guarantee decoding another full block,
+  which maximum size is provided in `ZSTD_frameHeader` structure, field `blockSizeMax`.
+  At which point, decoding can resume from the beginning of the buffer.
+  Note that already decoded data stored in the buffer should be flushed before being overwritten.
+
+  There are alternatives possible, for example using two or more buffers of size `windowSize` each, though they consume more memory.
+
+  Finally, if you control the compression process, you can also ignore all buffer size rules,
+  as long as the encoder and decoder progress in "lock-step",
+  aka use exactly the same buffer sizes, break contiguity at the same place, etc.
+
+  Once buffers are setup, start decompression, with ZSTD_decompressBegin().
   If decompression requires a dictionary, use ZSTD_decompressBegin_usingDict() or ZSTD_decompressBegin_usingDDict().
-  Alternatively, you can copy a prepared context, using ZSTD_copyDCtx().
 
   Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.
   ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue().
   ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail.
 
-  @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).
-  It can be zero, which is not an error; it just means ZSTD_decompressContinue() has decoded some metadata item.
+ @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).
+  It can be zero : it just means ZSTD_decompressContinue() has decoded some metadata item.
   It can also be an error code, which can be tested with ZSTD_isError().
 
-  ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize`.
-  They should preferably be located contiguously, prior to current block.
-  Alternatively, a round buffer of sufficient size is also possible. Sufficient size is determined by frame parameters.
-  ZSTD_decompressContinue() is very sensitive to contiguity,
-  if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,
-  or that previous contiguous segment is large enough to properly handle maximum back-reference.
-
   A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
   Context can then be reset to start a new decompression.
 
@@ -852,75 +873,101 @@ ZSTDLIB_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapaci
   == Special case : skippable frames ==
 
   Skippable frames allow integration of user-defined data into a flow of concatenated frames.
-  Skippable frames will be ignored (skipped) by a decompressor. The format of skippable frames is as follows :
+  Skippable frames will be ignored (skipped) by decompressor.
+  The format of skippable frames is as follows :
   a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F
   b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits
   c) Frame Content - any content (User Data) of length equal to Frame Size
-  For skippable frames ZSTD_decompressContinue() always returns 0.
-  For skippable frames ZSTD_getFrameHeader() returns fparamsPtr->windowLog==0 what means that a frame is skippable.
-    Note : If fparamsPtr->frameContentSize==0, it is ambiguous: the frame might actually be a Zstd encoded frame with no content.
-           For purposes of decompression, it is valid in both cases to skip the frame using
-           ZSTD_findFrameCompressedSize to find its size in bytes.
-  It also returns Frame Size as fparamsPtr->frameContentSize.
+  For skippable frames ZSTD_getFrameHeader() returns zfhPtr->frameType==ZSTD_skippableFrame.
+  For skippable frames ZSTD_decompressContinue() always returns 0 : it only skips the content.
 */
 
 /*=====   Buffer-less streaming decompression functions  =====*/
 typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
 typedef struct {
-    unsigned long long frameContentSize; /* ZSTD_CONTENTSIZE_UNKNOWN means this field is not available. 0 means "empty" */
+    unsigned long long frameContentSize; /* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */
     unsigned long long windowSize;       /* can be very large, up to <= frameContentSize */
+    unsigned blockSizeMax;
     ZSTD_frameType_e frameType;          /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */
     unsigned headerSize;
     unsigned dictID;
     unsigned checksumFlag;
 } ZSTD_frameHeader;
 ZSTDLIB_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize);   /**< doesn't consume input */
+ZSTDLIB_API size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize);  /**< when frame content size is not known, pass in frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN */
+
 ZSTDLIB_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx);
 ZSTDLIB_API size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
 ZSTDLIB_API size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
-ZSTDLIB_API void   ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx);
 
 ZSTDLIB_API size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx);
 ZSTDLIB_API size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+/* misc */
+ZSTDLIB_API void   ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx);
 typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;
 ZSTDLIB_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);
 
 
 
-/*===   New advanced API (experimental, and compression only)  ===*/
+/* ============================================ */
+/**       New advanced API (experimental)       */
+/* ============================================ */
 
 /* notes on API design :
- *   In this proposal, parameters are pushed one by one into an existing CCtx,
+ *   In this proposal, parameters are pushed one by one into an existing context,
  *   and then applied on all subsequent compression jobs.
  *   When no parameter is ever provided, CCtx is created with compression level ZSTD_CLEVEL_DEFAULT.
  *
  *   This API is intended to replace all others experimental API.
  *   It can basically do all other use cases, and even new ones.
- *   It stands a good chance to become "stable",
- *   after a reasonable testing period.
+ *   In constrast with _advanced() variants, it stands a reasonable chance to become "stable",
+ *   after a good testing period.
  */
 
 /* note on naming convention :
  *   Initially, the API favored names like ZSTD_setCCtxParameter() .
  *   In this proposal, convention is changed towards ZSTD_CCtx_setParameter() .
  *   The main driver is that it identifies more clearly the target object type.
- *   It feels clearer in light of potential variants :
+ *   It feels clearer when considering multiple targets :
  *   ZSTD_CDict_setParameter() (rather than ZSTD_setCDictParameter())
- *   ZSTD_DCtx_setParameter()  (rather than ZSTD_setDCtxParameter() )
- *   Left variant feels easier to distinguish.
+ *   ZSTD_CCtxParams_setParameter()  (rather than ZSTD_setCCtxParamsParameter() )
+ *   etc...
  */
 
 /* note on enum design :
- * All enum will be manually set to explicit values before reaching "stable API" status */
+ * All enum will be pinned to explicit values before reaching "stable API" status */
 
 typedef enum {
+    /* Question : should we have a format ZSTD_f_auto ?
+     * For the time being, it would mean exactly the same as ZSTD_f_zstd1.
+     * But, in the future, should several formats be supported,
+     * on the compression side, it would mean "default format".
+     * On the decompression side, it would mean "multi format",
+     * and ZSTD_f_zstd1 could be reserved to mean "accept *only* zstd frames".
+     * Since meaning is a little different, another option could be to define different enums for compression and decompression.
+     * This question could be kept for later, when there are actually multiple formats to support,
+     * but there is also the question of pinning enum values, and pinning value `0` is especially important */
+    ZSTD_f_zstd1 = 0,        /* zstd frame format, specified in zstd_compression_format.md (default) */
+    ZSTD_f_zstd1_magicless,  /* Variant of zstd frame format, without initial 4-bytes magic number.
+                              * Useful to save 4 bytes per generated frame.
+                              * Decoder cannot recognise automatically this format, requiring instructions. */
+} ZSTD_format_e;
+
+typedef enum {
+    /* compression format */
+    ZSTD_p_format = 10,      /* See ZSTD_format_e enum definition.
+                              * Cast selected format as unsigned for ZSTD_CCtx_setParameter() compatibility. */
+
     /* compression parameters */
     ZSTD_p_compressionLevel=100, /* Update all compression parameters according to pre-defined cLevel table
                               * Default level is ZSTD_CLEVEL_DEFAULT==3.
                               * Special: value 0 means "do not change cLevel". */
     ZSTD_p_windowLog,        /* Maximum allowed back-reference distance, expressed as power of 2.
                               * Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX.
-                              * Special: value 0 means "do not change windowLog". */
+                              * Special: value 0 means "do not change windowLog".
+                              * Note: Using a window size greater than ZSTD_MAXWINDOWSIZE_DEFAULT (default: 2^27)
+                              * requires setting the maximum window size at least as large during decompression. */
     ZSTD_p_hashLog,          /* Size of the probe table, as a power of 2.
                               * Resulting table size is (1 << (hashLog+2)).
                               * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX.
@@ -959,12 +1006,6 @@ typedef enum {
     ZSTD_p_checksumFlag,     /* A 32-bits checksum of content is written at end of frame (default:0) */
     ZSTD_p_dictIDFlag,       /* When applicable, dictID of dictionary is provided in frame header (default:1) */
 
-    /* dictionary parameters (must be set before ZSTD_CCtx_loadDictionary) */
-    ZSTD_p_dictMode=300,     /* Select how dictionary content must be interpreted. Value must be from type ZSTD_dictMode_e.
-                              * default : 0==auto : dictionary will be "full" if it respects specification, otherwise it will be "rawContent" */
-    ZSTD_p_refDictContent,   /* Dictionary content will be referenced, instead of copied (default:0==byCopy).
-                              * It requires that dictionary buffer outlives its users */
-
     /* multi-threading parameters */
     ZSTD_p_nbThreads=400,    /* Select how many threads a compression job can spawn (default:1)
                               * More threads improve speed, but also increase memory usage.
@@ -980,6 +1021,35 @@ typedef enum {
     /* advanced parameters - may not remain available after API update */
     ZSTD_p_forceMaxWindow=1100, /* Force back-reference distances to remain < windowSize,
                               * even when referencing into Dictionary content (default:0) */
+    ZSTD_p_enableLongDistanceMatching=1200,  /* Enable long distance matching.
+                                         * This parameter is designed to improve the compression
+                                         * ratio for large inputs with long distance matches.
+                                         * This increases the memory usage as well as window size.
+                                         * Note: setting this parameter sets all the LDM parameters
+                                         * as well as ZSTD_p_windowLog. It should be set after
+                                         * ZSTD_p_compressionLevel and before ZSTD_p_windowLog and
+                                         * other LDM parameters. Setting the compression level
+                                         * after this parameter overrides the window log, though LDM
+                                         * will remain enabled until explicitly disabled. */
+    ZSTD_p_ldmHashLog,   /* Size of the table for long distance matching, as a power of 2.
+                          * Larger values increase memory usage and compression ratio, but decrease
+                          * compression speed.
+                          * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX
+                          * (default: windowlog - 7). */
+    ZSTD_p_ldmMinMatch,  /* Minimum size of searched matches for long distance matcher.
+                          * Larger/too small values usually decrease compression ratio.
+                          * Must be clamped between ZSTD_LDM_MINMATCH_MIN
+                          * and ZSTD_LDM_MINMATCH_MAX (default: 64). */
+    ZSTD_p_ldmBucketSizeLog,  /* Log size of each bucket in the LDM hash table for collision resolution.
+                               * Larger values usually improve collision resolution but may decrease
+                               * compression speed.
+                               * The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX (default: 3). */
+    ZSTD_p_ldmHashEveryLog,  /* Frequency of inserting/looking up entries in the LDM hash table.
+                              * The default is MAX(0, (windowLog - ldmHashLog)) to
+                              * optimize hash table usage.
+                              * Larger values improve compression speed. Deviating far from the
+                              * default value will likely result in a decrease in compression ratio.
+                              * Must be clamped between 0 and ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN. */
 
 } ZSTD_cParameter;
 
@@ -1007,14 +1077,22 @@ ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long lo
  * @result : 0, or an error code (which can be tested with ZSTD_isError()).
  *  Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,
  *            meaning "return to no-dictionary mode".
- *  Note 1 : `dict` content will be copied internally,
- *           except if ZSTD_p_refDictContent is set before loading.
+ *  Note 1 : `dict` content will be copied internally. Use
+ *            ZSTD_CCtx_loadDictionary_byReference() to reference dictionary
+ *            content instead. The dictionary buffer must then outlive its
+ *            users.
  *  Note 2 : Loading a dictionary involves building tables, which are dependent on compression parameters.
  *           For this reason, compression parameters cannot be changed anymore after loading a dictionary.
  *           It's also a CPU-heavy operation, with non-negligible impact on latency.
  *  Note 3 : Dictionary will be used for all future compression jobs.
- *           To return to "no-dictionary" situation, load a NULL dictionary */
+ *           To return to "no-dictionary" situation, load a NULL dictionary
+ *  Note 5 : Use ZSTD_CCtx_loadDictionary_advanced() to select how dictionary
+ *           content will be interpreted.
+ */
 ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode);
+
 
 /*! ZSTD_CCtx_refCDict() :
  *  Reference a prepared dictionary, to be used for all next compression jobs.
@@ -1040,23 +1118,26 @@ ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
  *  Note 1 : Prefix buffer is referenced. It must outlive compression job.
  *  Note 2 : Referencing a prefix involves building tables, which are dependent on compression parameters.
  *           It's a CPU-heavy operation, with non-negligible impact on latency.
- *  Note 3 : it's possible to alter ZSTD_p_dictMode using ZSTD_CCtx_setParameter() */
+ *  Note 3 : By default, the prefix is treated as raw content
+ *           (ZSTD_dm_rawContent). Use ZSTD_CCtx_refPrefix_advanced() to alter
+ *           dictMode. */
 ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize);
+ZSTDLIB_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictMode_e dictMode);
 
 
 
 typedef enum {
     ZSTD_e_continue=0, /* collect more data, encoder transparently decides when to output result, for optimal conditions */
     ZSTD_e_flush,      /* flush any data provided so far - frame will continue, future data can still reference previous data for better compression */
-    ZSTD_e_end         /* flush any remaining data and ends current frame. Any future compression starts a new frame. */
+    ZSTD_e_end         /* flush any remaining data and close current frame. Any additional data starts a new frame. */
 } ZSTD_EndDirective;
 
 /*! ZSTD_compress_generic() :
  *  Behave about the same as ZSTD_compressStream. To note :
  *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_setParameter()
  *  - Compression parameters cannot be changed once compression is started.
- *  - *dstPos must be <= dstCapacity, *srcPos must be <= srcSize
- *  - *dspPos and *srcPos will be updated. They are guaranteed to remain below their respective limit.
+ *  - outpot->pos must be <= dstCapacity, input->pos must be <= srcSize
+ *  - outpot->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
  *  - @return provides the minimum amount of data still to flush from internal buffers
  *            or an error code, which can be tested using ZSTD_isError().
  *            if @return != 0, flush is not fully completed, there is some data left within internal buffers.
@@ -1075,6 +1156,7 @@ ZSTDLIB_API size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,
  *  Useful after an error, or to interrupt an ongoing compression job and start a new one.
  *  Any internal data not yet flushed is cancelled.
  *  Dictionary (if any) is dropped.
+ *  All parameters are back to default values.
  *  It's possible to modify compression parameters after a reset.
  */
 ZSTDLIB_API void ZSTD_CCtx_reset(ZSTD_CCtx* cctx);   /* Not ready yet ! */
@@ -1083,21 +1165,187 @@ ZSTDLIB_API void ZSTD_CCtx_reset(ZSTD_CCtx* cctx);   /* Not ready yet ! */
 /*! ZSTD_compress_generic_simpleArgs() :
  *  Same as ZSTD_compress_generic(),
  *  but using only integral types as arguments.
- *  Argument list is larger and less expressive than ZSTD_{in,out}Buffer,
+ *  Argument list is larger than ZSTD_{in,out}Buffer,
  *  but can be helpful for binders from dynamic languages
  *  which have troubles handling structures containing memory pointers.
  */
-size_t ZSTD_compress_generic_simpleArgs (
+ZSTDLIB_API size_t ZSTD_compress_generic_simpleArgs (
                             ZSTD_CCtx* cctx,
                             void* dst, size_t dstCapacity, size_t* dstPos,
                       const void* src, size_t srcSize, size_t* srcPos,
                             ZSTD_EndDirective endOp);
 
 
+/*! ZSTD_CCtx_params :
+ *  Quick howto :
+ *  - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
+ *  - ZSTD_CCtxParam_setParameter() : Push parameters one by one into
+ *                                    an existing ZSTD_CCtx_params structure.
+ *                                    This is similar to
+ *                                    ZSTD_CCtx_setParameter().
+ *  - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to
+ *                                    an existing CCtx.
+ *                                    These parameters will be applied to
+ *                                    all subsequent compression jobs.
+ *  - ZSTD_compress_generic() : Do compression using the CCtx.
+ *  - ZSTD_freeCCtxParams() : Free the memory.
+ *
+ *  This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()
+ *  for static allocation for single-threaded compression.
+ */
+ZSTDLIB_API ZSTD_CCtx_params* ZSTD_createCCtxParams(void);
 
-/**
-    Block functions
+/*! ZSTD_resetCCtxParams() :
+ *  Reset params to default, with the default compression level.
+ */
+ZSTDLIB_API size_t ZSTD_resetCCtxParams(ZSTD_CCtx_params* params);
+
+/*! ZSTD_initCCtxParams() :
+ *  Initializes the compression parameters of cctxParams according to
+ *  compression level. All other parameters are reset to their default values.
+ */
+ZSTDLIB_API size_t ZSTD_initCCtxParams(ZSTD_CCtx_params* cctxParams, int compressionLevel);
+
+/*! ZSTD_initCCtxParams_advanced() :
+ *  Initializes the compression and frame parameters of cctxParams according to
+ *  params. All other parameters are reset to their default values.
+ */
+ZSTDLIB_API size_t ZSTD_initCCtxParams_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
+
+ZSTDLIB_API size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params);
+
+/*! ZSTD_CCtxParam_setParameter() :
+ *  Similar to ZSTD_CCtx_setParameter.
+ *  Set one compression parameter, selected by enum ZSTD_cParameter.
+ *  Parameters must be applied to a ZSTD_CCtx using ZSTD_CCtx_setParametersUsingCCtxParams().
+ *  Note : when `value` is an enum, cast it to unsigned for proper type checking.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_CCtxParam_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, unsigned value);
+
+/*! ZSTD_CCtx_setParametersUsingCCtxParams() :
+ *  Apply a set of ZSTD_CCtx_params to the compression context.
+ *  This must be done before the dictionary is loaded.
+ *  The pledgedSrcSize is treated as unknown.
+ *  Multithreading parameters are applied only if nbThreads > 1.
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_setParametersUsingCCtxParams(
+        ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params);
+
+
+/*===   Advanced parameters for decompression API  ===*/
+
+/* The following parameters must be set after creating a ZSTD_DCtx* (or ZSTD_DStream*) object,
+ * but before starting decompression of a frame.
+ */
+
+/*! ZSTD_DCtx_loadDictionary() :
+ *  Create an internal DDict from dict buffer,
+ *  to be used to decompress next frames.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,
+ *            meaning "return to no-dictionary mode".
+ *  Note 1 : `dict` content will be copied internally.
+ *            Use ZSTD_DCtx_loadDictionary_byReference()
+ *            to reference dictionary content instead.
+ *            In which case, the dictionary buffer must outlive its users.
+ *  Note 2 : Loading a dictionary involves building tables,
+ *           which has a non-negligible impact on CPU usage and latency.
+ *  Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to select
+ *           how dictionary content will be interpreted and loaded.
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);   /* not implemented */
+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);   /* not implemented */
+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode);   /* not implemented */
+
+
+/*! ZSTD_DCtx_refDDict() :
+ *  Reference a prepared dictionary, to be used to decompress next frames.
+ *  The dictionary remains active for decompression of future frames using same DCtx.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Note 1 : Currently, only one dictionary can be managed.
+ *           Referencing a new dictionary effectively "discards" any previous one.
+ *  Special : adding a NULL DDict means "return to no-dictionary mode".
+ *  Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx.
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);   /* not implemented */
+
+
+/*! ZSTD_DCtx_refPrefix() :
+ *  Reference a prefix (single-usage dictionary) for next compression job.
+ *  Prefix is **only used once**. It must be explicitly referenced before each frame.
+ *  If there is a need to use same prefix multiple times, consider embedding it into a ZSTD_DDict instead.
+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).
+ *  Note 1 : Adding any prefix (including NULL) invalidates any previously set prefix or dictionary
+ *  Note 2 : Prefix buffer is referenced. It must outlive compression job.
+ *  Note 3 : By default, the prefix is treated as raw content (ZSTD_dm_rawContent).
+ *           Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode.
+ *  Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost.
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize);   /* not implemented */
+ZSTDLIB_API size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictMode_e dictMode);   /* not implemented */
+
+
+/*! ZSTD_DCtx_setMaxWindowSize() :
+ *  Refuses allocating internal buffers for frames requiring a window size larger than provided limit.
+ *  This is useful to prevent a decoder context from reserving too much memory for itself (potential attack scenario).
+ *  This parameter is only useful in streaming mode, since no internal buffer is allocated in direct mode.
+ *  By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_MAX)
+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize);
+
+
+/*! ZSTD_DCtx_setFormat() :
+ *  Instruct the decoder context about what kind of data to decode next.
+ *  This instruction is mandatory to decode data without a fully-formed header,
+ *  such ZSTD_f_zstd1_magicless for example.
+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format);
+
+
+/*! ZSTD_decompress_generic() :
+ *  Behave the same as ZSTD_decompressStream.
+ *  Decompression parameters cannot be changed once decompression is started.
+ * @return : an error code, which can be tested using ZSTD_isError()
+ *           if >0, a hint, nb of expected input bytes for next invocation.
+ *           `0` means : a frame has just been fully decoded and flushed.
+ */
+ZSTDLIB_API size_t ZSTD_decompress_generic(ZSTD_DCtx* dctx,
+                                           ZSTD_outBuffer* output,
+                                           ZSTD_inBuffer* input);
+
+
+/*! ZSTD_decompress_generic_simpleArgs() :
+ *  Same as ZSTD_decompress_generic(),
+ *  but using only integral types as arguments.
+ *  Argument list is larger than ZSTD_{in,out}Buffer,
+ *  but can be helpful for binders from dynamic languages
+ *  which have troubles handling structures containing memory pointers.
+ */
+ZSTDLIB_API size_t ZSTD_decompress_generic_simpleArgs (
+                            ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity, size_t* dstPos,
+                      const void* src, size_t srcSize, size_t* srcPos);
+
+
+/*! ZSTD_DCtx_reset() :
+ *  Return a DCtx to clean state.
+ *  If a decompression was ongoing, any internal data not yet flushed is cancelled.
+ *  All parameters are back to default values, including sticky ones.
+ *  Dictionary (if any) is dropped.
+ *  Parameters can be modified again after a reset.
+ */
+ZSTDLIB_API void ZSTD_DCtx_reset(ZSTD_DCtx* dctx);
+
+
+
+/* ============================ */
+/**       Block level API       */
+/* ============================ */
 
+/*!
     Block functions produce and decode raw zstd blocks, without frame metadata.
     Frame metadata cost is typically ~18 bytes, which can be non-negligible for very small blocks (< 100 bytes).
     User will have to take in charge required information to regenerate data, such as compressed and content sizes.
@@ -1109,7 +1357,7 @@ size_t ZSTD_compress_generic_simpleArgs (
       + compression : any ZSTD_compressBegin*() variant, including with dictionary
       + decompression : any ZSTD_decompressBegin*() variant, including with dictionary
       + copyCCtx() and copyDCtx() can be used too
-    - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX
+    - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB
       + If input is larger than a block size, it's necessary to split input data into multiple blocks
       + For inputs larger than a single block size, consider using the regular ZSTD_compress() instead.
         Frame metadata is not that costly, and quickly becomes negligible as source size grows larger.
@@ -1128,7 +1376,7 @@ size_t ZSTD_compress_generic_simpleArgs (
 ZSTDLIB_API size_t ZSTD_getBlockSize   (const ZSTD_CCtx* cctx);
 ZSTDLIB_API size_t ZSTD_compressBlock  (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
 ZSTDLIB_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
-ZSTDLIB_API size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize);  /**< insert block into `dctx` history. Useful for uncompressed blocks */
+ZSTDLIB_API size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize);  /**< insert uncompressed block into `dctx` history. Useful for multi-blocks decompression */
 
 
 #endif   /* ZSTD_H_ZSTD_STATIC_LINKING_ONLY */