Update contrib/libs/zstd to 1.5.5

32 files changed, 1353 insertions, 696 deletions

diff --git a/contrib/libs/zstd/CHANGELOG b/contrib/libs/zstd/CHANGELOG
index 4010c1ff5d..c7a7506ee3 100644
--- a/contrib/libs/zstd/CHANGELOG
+++ b/contrib/libs/zstd/CHANGELOG
@@ -1,3 +1,22 @@
+v1.5.5 (Apr 2023)
+fix: fix rare corruption bug affecting the high compression mode, reported by @danlark1 (#3517, @terrelln)
+perf: improve mid-level compression speed (#3529, #3533, #3543, @yoniko and #3552, @terrelln)
+lib: deprecated bufferless block-level API (#3534) by @terrelln
+cli: mmap large dictionaries to save memory, by @daniellerozenblit
+cli: improve speed of --patch-from mode (~+50%) (#3545) by @daniellerozenblit
+cli: improve i/o speed (~+10%) when processing lots of small files (#3479) by @felixhandte
+cli: zstd no longer crashes when requested to write into write-protected directory (#3541) by @felixhandte
+cli: fix decompression into block device using -o, reported by @georgmu (#3583)
+build: fix zstd CLI compiled with lzma support but not zlib support (#3494) by @Hello71
+build: fix cmake does no longer require 3.18 as minimum version (#3510) by @kou
+build: fix MSVC+ClangCL linking issue (#3569) by @tru
+build: fix zstd-dll, version of zstd CLI that links to the dynamic library (#3496) by @yoniko
+build: fix MSVC warnings (#3495) by @embg
+doc: updated zstd specification to clarify corner cases, by @Cyan4973
+doc: document how to create fat binaries for macos (#3568) by @rickmark
+misc: improve seekable format ingestion speed (~+100%) for very small chunk sizes (#3544) by @Cyan4973
+misc: tests/fullbench can benchmark multiple files (#3516) by @dloidolt
+
 v1.5.4 (Feb 2023)
 perf: +20% faster huffman decompression for targets that can't compile x64 assembly (#3449, @terrelln)
 perf: up to +10% faster streaming compression at levels 1-2 (#3114, @embg)
diff --git a/contrib/libs/zstd/README.md b/contrib/libs/zstd/README.md
index 6bcf757d86..f91e68fdb1 100644
--- a/contrib/libs/zstd/README.md
+++ b/contrib/libs/zstd/README.md
@@ -13,15 +13,12 @@ a list of known ports and bindings is provided on [Zstandard homepage](https://f
 **Development branch status:**
 
 [![Build Status][travisDevBadge]][travisLink]
-[![Build status][AppveyorDevBadge]][AppveyorLink]
 [![Build status][CircleDevBadge]][CircleLink]
 [![Build status][CirrusDevBadge]][CirrusLink]
 [![Fuzzing Status][OSSFuzzBadge]][OSSFuzzLink]
 
 [travisDevBadge]: https://api.travis-ci.com/facebook/zstd.svg?branch=dev "Continuous Integration test suite"
 [travisLink]: https://travis-ci.com/facebook/zstd
-[AppveyorDevBadge]: https://ci.appveyor.com/api/projects/status/xt38wbdxjk5mrbem/branch/dev?svg=true "Windows test suite"
-[AppveyorLink]: https://ci.appveyor.com/project/YannCollet/zstd-p0yf0
 [CircleDevBadge]: https://circleci.com/gh/facebook/zstd/tree/dev.svg?style=shield "Short test suite"
 [CircleLink]: https://circleci.com/gh/facebook/zstd
 [CirrusDevBadge]: https://api.cirrus-ci.com/github/facebook/zstd.svg?branch=dev
@@ -154,6 +151,18 @@ to create `zstd` binary, and `libzstd` dynamic and static libraries.
 
 By default, `CMAKE_BUILD_TYPE` is set to `Release`.
 
+#### Support for Fat (Universal2) Output
+
+`zstd` can be built and installed with support for both Apple Silicon (M1/M2) as well as Intel by using CMake's Universal2 support.
+To perform a Fat/Universal2 build and install use the following commands:
+
+```bash
+cmake -B build-cmake-debug -S build/cmake -G Ninja -DCMAKE_OSX_ARCHITECTURES="x86_64;x86_64h;arm64"
+cd build-cmake-debug
+ninja
+sudo ninja install
+```
+
 ### Meson
 
 A Meson project is provided within [`build/meson`](build/meson). Follow
diff --git a/contrib/libs/zstd/lib/common/allocations.h b/contrib/libs/zstd/lib/common/allocations.h
new file mode 100644
index 0000000000..a3153c4bac
--- /dev/null
+++ b/contrib/libs/zstd/lib/common/allocations.h
@@ -0,0 +1,55 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* This file provides custom allocation primitives
+ */
+
+#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"   /* ZSTD_malloc, ZSTD_calloc, ZSTD_free, ZSTD_memset */
+
+#include "mem.h" /* MEM_STATIC */
+#define ZSTD_STATIC_LINKING_ONLY
+#include "../zstd.h" /* ZSTD_customMem */
+
+#ifndef ZSTD_ALLOCATIONS_H
+#define ZSTD_ALLOCATIONS_H
+
+/* custom memory allocation functions */
+
+MEM_STATIC void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem)
+{
+    if (customMem.customAlloc)
+        return customMem.customAlloc(customMem.opaque, size);
+    return ZSTD_malloc(size);
+}
+
+MEM_STATIC void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem)
+{
+    if (customMem.customAlloc) {
+        /* calloc implemented as malloc+memset;
+         * not as efficient as calloc, but next best guess for custom malloc */
+        void* const ptr = customMem.customAlloc(customMem.opaque, size);
+        ZSTD_memset(ptr, 0, size);
+        return ptr;
+    }
+    return ZSTD_calloc(1, size);
+}
+
+MEM_STATIC void ZSTD_customFree(void* ptr, ZSTD_customMem customMem)
+{
+    if (ptr!=NULL) {
+        if (customMem.customFree)
+            customMem.customFree(customMem.opaque, ptr);
+        else
+            ZSTD_free(ptr);
+    }
+}
+
+#endif /* ZSTD_ALLOCATIONS_H */
diff --git a/contrib/libs/zstd/lib/common/bits.h b/contrib/libs/zstd/lib/common/bits.h
index 7939f3d0f3..def56c474c 100644
--- a/contrib/libs/zstd/lib/common/bits.h
+++ b/contrib/libs/zstd/lib/common/bits.h
@@ -17,7 +17,7 @@ MEM_STATIC unsigned ZSTD_countTrailingZeros32_fallback(U32 val)
 {
     assert(val != 0);
     {
-        static const int DeBruijnBytePos[32] = {0, 1, 28, 2, 29, 14, 24, 3,
+        static const U32 DeBruijnBytePos[32] = {0, 1, 28, 2, 29, 14, 24, 3,
                                                 30, 22, 20, 15, 25, 17, 4, 8,
                                                 31, 27, 13, 23, 21, 19, 16, 7,
                                                 26, 12, 18, 6, 11, 5, 10, 9};
@@ -30,7 +30,7 @@ MEM_STATIC unsigned ZSTD_countTrailingZeros32(U32 val)
     assert(val != 0);
 #   if defined(_MSC_VER)
 #       if STATIC_BMI2 == 1
-            return _tzcnt_u32(val);
+            return (unsigned)_tzcnt_u32(val);
 #       else
             if (val != 0) {
                 unsigned long r;
@@ -69,7 +69,7 @@ MEM_STATIC unsigned ZSTD_countLeadingZeros32(U32 val)
     assert(val != 0);
 #   if defined(_MSC_VER)
 #       if STATIC_BMI2 == 1
-            return _lzcnt_u32(val);
+            return (unsigned)_lzcnt_u32(val);
 #       else
             if (val != 0) {
                 unsigned long r;
@@ -92,7 +92,7 @@ MEM_STATIC unsigned ZSTD_countTrailingZeros64(U64 val)
     assert(val != 0);
 #   if defined(_MSC_VER) && defined(_WIN64)
 #       if STATIC_BMI2 == 1
-            return _tzcnt_u64(val);
+            return (unsigned)_tzcnt_u64(val);
 #       else
             if (val != 0) {
                 unsigned long r;
@@ -123,7 +123,7 @@ MEM_STATIC unsigned ZSTD_countLeadingZeros64(U64 val)
     assert(val != 0);
 #   if defined(_MSC_VER) && defined(_WIN64)
 #       if STATIC_BMI2 == 1
-            return _lzcnt_u64(val);
+            return (unsigned)_lzcnt_u64(val);
 #       else
             if (val != 0) {
                 unsigned long r;
@@ -172,4 +172,29 @@ MEM_STATIC unsigned ZSTD_highbit32(U32 val)   /* compress, dictBuilder, decodeCo
     return 31 - ZSTD_countLeadingZeros32(val);
 }
 
+/* ZSTD_rotateRight_*():
+ * Rotates a bitfield to the right by "count" bits.
+ * https://en.wikipedia.org/w/index.php?title=Circular_shift&oldid=991635599#Implementing_circular_shifts
+ */
+MEM_STATIC
+U64 ZSTD_rotateRight_U64(U64 const value, U32 count) {
+    assert(count < 64);
+    count &= 0x3F; /* for fickle pattern recognition */
+    return (value >> count) | (U64)(value << ((0U - count) & 0x3F));
+}
+
+MEM_STATIC
+U32 ZSTD_rotateRight_U32(U32 const value, U32 count) {
+    assert(count < 32);
+    count &= 0x1F; /* for fickle pattern recognition */
+    return (value >> count) | (U32)(value << ((0U - count) & 0x1F));
+}
+
+MEM_STATIC
+U16 ZSTD_rotateRight_U16(U16 const value, U32 count) {
+    assert(count < 16);
+    count &= 0x0F; /* for fickle pattern recognition */
+    return (value >> count) | (U16)(value << ((0U - count) & 0x0F));
+}
+
 #endif /* ZSTD_BITS_H */
diff --git a/contrib/libs/zstd/lib/common/bitstream.h b/contrib/libs/zstd/lib/common/bitstream.h
index db1b4cf136..72b0b3df22 100644
--- a/contrib/libs/zstd/lib/common/bitstream.h
+++ b/contrib/libs/zstd/lib/common/bitstream.h
@@ -396,7 +396,7 @@ MEM_STATIC BIT_DStream_status BIT_reloadDStreamFast(BIT_DStream_t* bitD)
  *  This function is safe, it guarantees it will not read beyond src buffer.
  * @return : status of `BIT_DStream_t` internal register.
  *           when status == BIT_DStream_unfinished, internal register is filled with at least 25 or 57 bits */
-MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
+MEM_STATIC FORCE_INLINE_ATTR BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
 {
     if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))  /* overflow detected, like end of stream */
         return BIT_DStream_overflow;
diff --git a/contrib/libs/zstd/lib/common/compiler.h b/contrib/libs/zstd/lib/common/compiler.h
index d4f2f285d7..73f8d01998 100644
--- a/contrib/libs/zstd/lib/common/compiler.h
+++ b/contrib/libs/zstd/lib/common/compiler.h
@@ -311,6 +311,10 @@ void __msan_poison(const volatile void *a, size_t size);
 /* Returns the offset of the first (at least partially) poisoned byte in the
    memory range, or -1 if the whole range is good. */
 intptr_t __msan_test_shadow(const volatile void *x, size_t size);
+
+/* Print shadow and origin for the memory range to stderr in a human-readable
+   format. */
+void __msan_print_shadow(const volatile void *x, size_t size);
 #endif
 
 #if ZSTD_ADDRESS_SANITIZER && !defined(ZSTD_ASAN_DONT_POISON_WORKSPACE)
diff --git a/contrib/libs/zstd/lib/common/pool.c b/contrib/libs/zstd/lib/common/pool.c
index f3d9d08547..d5ca5a7808 100644
--- a/contrib/libs/zstd/lib/common/pool.c
+++ b/contrib/libs/zstd/lib/common/pool.c
@@ -10,9 +10,9 @@
 
 
 /* ======   Dependencies   ======= */
+#include "../common/allocations.h"  /* ZSTD_customCalloc, ZSTD_customFree */
 #include "zstd_deps.h" /* size_t */
 #include "debug.h"     /* assert */
-#include "zstd_internal.h"  /* ZSTD_customCalloc, ZSTD_customFree */
 #include "pool.h"
 
 /* ======   Compiler specifics   ====== */
diff --git a/contrib/libs/zstd/lib/common/threading.c b/contrib/libs/zstd/lib/common/threading.c
index f2341105a1..ca155b9b9d 100644
--- a/contrib/libs/zstd/lib/common/threading.c
+++ b/contrib/libs/zstd/lib/common/threading.c
@@ -47,7 +47,7 @@ static unsigned __stdcall worker(void *arg)
     void* (*start_routine)(void*);
     void* thread_arg;
 
-    /* Inialized thread_arg and start_routine and signal main thread that we don't need it
+    /* Initialized thread_arg and start_routine and signal main thread that we don't need it
      * to wait any longer.
      */
     {
diff --git a/contrib/libs/zstd/lib/common/zstd_common.c b/contrib/libs/zstd/lib/common/zstd_common.c
index 3208552475..3f04c22abf 100644
--- a/contrib/libs/zstd/lib/common/zstd_common.c
+++ b/contrib/libs/zstd/lib/common/zstd_common.c
@@ -14,7 +14,6 @@
 *  Dependencies
 ***************************************/
 #define ZSTD_DEPS_NEED_MALLOC
-#include "zstd_deps.h"   /* ZSTD_malloc, ZSTD_calloc, ZSTD_free, ZSTD_memset */
 #include "error_private.h"
 #include "zstd_internal.h"
 
@@ -47,37 +46,3 @@ ZSTD_ErrorCode ZSTD_getErrorCode(size_t code) { return ERR_getErrorCode(code); }
 /*! ZSTD_getErrorString() :
  *  provides error code string from enum */
 const char* ZSTD_getErrorString(ZSTD_ErrorCode code) { return ERR_getErrorString(code); }
-
-
-
-/*=**************************************************************
-*  Custom allocator
-****************************************************************/
-void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem)
-{
-    if (customMem.customAlloc)
-        return customMem.customAlloc(customMem.opaque, size);
-    return ZSTD_malloc(size);
-}
-
-void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem)
-{
-    if (customMem.customAlloc) {
-        /* calloc implemented as malloc+memset;
-         * not as efficient as calloc, but next best guess for custom malloc */
-        void* const ptr = customMem.customAlloc(customMem.opaque, size);
-        ZSTD_memset(ptr, 0, size);
-        return ptr;
-    }
-    return ZSTD_calloc(1, size);
-}
-
-void ZSTD_customFree(void* ptr, ZSTD_customMem customMem)
-{
-    if (ptr!=NULL) {
-        if (customMem.customFree)
-            customMem.customFree(customMem.opaque, ptr);
-        else
-            ZSTD_free(ptr);
-    }
-}
diff --git a/contrib/libs/zstd/lib/common/zstd_internal.h b/contrib/libs/zstd/lib/common/zstd_internal.h
index 54792712fc..d1bcada74b 100644
--- a/contrib/libs/zstd/lib/common/zstd_internal.h
+++ b/contrib/libs/zstd/lib/common/zstd_internal.h
@@ -350,11 +350,6 @@ typedef struct {
 const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx);   /* compress & dictBuilder */
 int ZSTD_seqToCodes(const seqStore_t* seqStorePtr);   /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
 
-/* custom memory allocation functions */
-void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem);
-void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem);
-void ZSTD_customFree(void* ptr, ZSTD_customMem customMem);
-
 
 /* ZSTD_invalidateRepCodes() :
  * ensures next compression will not use repcodes from previous block.
diff --git a/contrib/libs/zstd/lib/compress/zstd_compress.c b/contrib/libs/zstd/lib/compress/zstd_compress.c
index b55f684cd7..d6133e70b4 100644
--- a/contrib/libs/zstd/lib/compress/zstd_compress.c
+++ b/contrib/libs/zstd/lib/compress/zstd_compress.c
@@ -11,6 +11,7 @@
 /*-*************************************
 *  Dependencies
 ***************************************/
+#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
 #include "../common/zstd_deps.h"  /* INT_MAX, ZSTD_memset, ZSTD_memcpy */
 #include "../common/mem.h"
 #include "hist.h"           /* HIST_countFast_wksp */
@@ -26,7 +27,7 @@
 #include "zstd_opt.h"
 #include "zstd_ldm.h"
 #include "zstd_compress_superblock.h"
-#include  "../common/bits.h"      /* ZSTD_highbit32 */
+#include  "../common/bits.h"      /* ZSTD_highbit32, ZSTD_rotateRight_U64 */
 
 /* ***************************************************************
 *  Tuning parameters
@@ -1177,16 +1178,39 @@ size_t ZSTD_CCtx_setParametersUsingCCtxParams(
 
 size_t ZSTD_CCtx_setCParams(ZSTD_CCtx* cctx, ZSTD_compressionParameters cparams)
 {
+    ZSTD_STATIC_ASSERT(sizeof(cparams) == 7 * 4 /* all params are listed below */);
     DEBUGLOG(4, "ZSTD_CCtx_setCParams");
-    assert(cctx != NULL);
-    if (cctx->streamStage != zcss_init) {
-        /* All parameters in @cparams are allowed to be updated during MT compression.
-         * This must be signaled, so that MT compression picks up the changes */
-        cctx->cParamsChanged = 1;
-    }
-    /* only update if parameters are valid */
+    /* only update if all parameters are valid */
     FORWARD_IF_ERROR(ZSTD_checkCParams(cparams), "");
-    cctx->requestedParams.cParams = cparams;
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_windowLog, cparams.windowLog), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_chainLog, cparams.chainLog), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_hashLog, cparams.hashLog), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_searchLog, cparams.searchLog), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_minMatch, cparams.minMatch), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_targetLength, cparams.targetLength), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_strategy, cparams.strategy), "");
+    return 0;
+}
+
+size_t ZSTD_CCtx_setFParams(ZSTD_CCtx* cctx, ZSTD_frameParameters fparams)
+{
+    ZSTD_STATIC_ASSERT(sizeof(fparams) == 3 * 4 /* all params are listed below */);
+    DEBUGLOG(4, "ZSTD_CCtx_setFParams");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_contentSizeFlag, fparams.contentSizeFlag != 0), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_checksumFlag, fparams.checksumFlag != 0), "");
+    FORWARD_IF_ERROR(ZSTD_CCtx_setParameter(cctx, ZSTD_c_dictIDFlag, fparams.noDictIDFlag == 0), "");
+    return 0;
+}
+
+size_t ZSTD_CCtx_setParams(ZSTD_CCtx* cctx, ZSTD_parameters params)
+{
+    DEBUGLOG(4, "ZSTD_CCtx_setParams");
+    /* First check cParams, because we want to update all or none. */
+    FORWARD_IF_ERROR(ZSTD_checkCParams(params.cParams), "");
+    /* Next set fParams, because this could fail if the cctx isn't in init stage. */
+    FORWARD_IF_ERROR(ZSTD_CCtx_setFParams(cctx, params.fParams), "");
+    /* Finally set cParams, which should succeed. */
+    FORWARD_IF_ERROR(ZSTD_CCtx_setCParams(cctx, params.cParams), "");
     return 0;
 }
 
@@ -1208,9 +1232,9 @@ static void ZSTD_dedicatedDictSearch_revertCParams(
         ZSTD_compressionParameters* cParams);
 
 /**
- * Initializes the local dict using the requested parameters.
- * NOTE: This does not use the pledged src size, because it may be used for more
- * than one compression.
+ * Initializes the local dictionary using requested parameters.
+ * NOTE: Initialization does not employ the pledged src size,
+ * because the dictionary may be used for multiple compressions.
  */
 static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx)
 {
@@ -1223,8 +1247,8 @@ static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx)
         return 0;
     }
     if (dl->cdict != NULL) {
-        assert(cctx->cdict == dl->cdict);
         /* Local dictionary already initialized. */
+        assert(cctx->cdict == dl->cdict);
         return 0;
     }
     assert(dl->dictSize > 0);
@@ -1244,26 +1268,30 @@ static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx)
 }
 
 size_t ZSTD_CCtx_loadDictionary_advanced(
-        ZSTD_CCtx* cctx, const void* dict, size_t dictSize,
-        ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType)
+        ZSTD_CCtx* cctx,
+        const void* dict, size_t dictSize,
+        ZSTD_dictLoadMethod_e dictLoadMethod,
+        ZSTD_dictContentType_e dictContentType)
 {
-    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
-                    "Can't load a dictionary when ctx is not in init stage.");
     DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize);
-    ZSTD_clearAllDicts(cctx);  /* in case one already exists */
-    if (dict == NULL || dictSize == 0)  /* no dictionary mode */
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
+                    "Can't load a dictionary when cctx is not in init stage.");
+    ZSTD_clearAllDicts(cctx);  /* erase any previously set dictionary */
+    if (dict == NULL || dictSize == 0)  /* no dictionary */
         return 0;
     if (dictLoadMethod == ZSTD_dlm_byRef) {
         cctx->localDict.dict = dict;
     } else {
+        /* copy dictionary content inside CCtx to own its lifetime */
         void* dictBuffer;
         RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
-                        "no malloc for static CCtx");
+                        "static CCtx can't allocate for an internal copy of dictionary");
         dictBuffer = ZSTD_customMalloc(dictSize, cctx->customMem);
-        RETURN_ERROR_IF(!dictBuffer, memory_allocation, "NULL pointer!");
+        RETURN_ERROR_IF(dictBuffer==NULL, memory_allocation,
+                        "allocation failed for dictionary content");
         ZSTD_memcpy(dictBuffer, dict, dictSize);
-        cctx->localDict.dictBuffer = dictBuffer;
-        cctx->localDict.dict = dictBuffer;
+        cctx->localDict.dictBuffer = dictBuffer;  /* owned ptr to free */
+        cctx->localDict.dict = dictBuffer;        /* read-only reference */
     }
     cctx->localDict.dictSize = dictSize;
     cctx->localDict.dictContentType = dictContentType;
@@ -1333,7 +1361,7 @@ size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset)
     if ( (reset == ZSTD_reset_parameters)
       || (reset == ZSTD_reset_session_and_parameters) ) {
         RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
-                        "Can't reset parameters only when not in init stage.");
+                        "Reset parameters is only possible during init stage.");
         ZSTD_clearAllDicts(cctx);
         ZSTD_memset(&cctx->externalMatchCtx, 0, sizeof(cctx->externalMatchCtx));
         return ZSTD_CCtxParams_reset(&cctx->requestedParams);
@@ -1592,7 +1620,7 @@ ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams,
       + ZSTD_cwksp_aligned_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t))
       + ZSTD_cwksp_aligned_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t));
     size_t const lazyAdditionalSpace = ZSTD_rowMatchFinderUsed(cParams->strategy, useRowMatchFinder)
-                                            ? ZSTD_cwksp_aligned_alloc_size(hSize*sizeof(U16))
+                                            ? ZSTD_cwksp_aligned_alloc_size(hSize)
                                             : 0;
     size_t const optSpace = (forCCtx && (cParams->strategy >= ZSTD_btopt))
                                 ? optPotentialSpace
@@ -1883,6 +1911,19 @@ typedef enum {
     ZSTD_resetTarget_CCtx
 } ZSTD_resetTarget_e;
 
+/* Mixes bits in a 64 bits in a value, based on XXH3_rrmxmx */
+static U64 ZSTD_bitmix(U64 val, U64 len) {
+    val ^= ZSTD_rotateRight_U64(val, 49) ^ ZSTD_rotateRight_U64(val, 24);
+    val *= 0x9FB21C651E98DF25ULL;
+    val ^= (val >> 35) + len ;
+    val *= 0x9FB21C651E98DF25ULL;
+    return val ^ (val >> 28);
+}
+
+/* Mixes in the hashSalt and hashSaltEntropy to create a new hashSalt */
+static void ZSTD_advanceHashSalt(ZSTD_matchState_t* ms) {
+    ms->hashSalt = ZSTD_bitmix(ms->hashSalt, 8) ^ ZSTD_bitmix((U64) ms->hashSaltEntropy, 4);
+}
 
 static size_t
 ZSTD_reset_matchState(ZSTD_matchState_t* ms,
@@ -1910,6 +1951,7 @@ ZSTD_reset_matchState(ZSTD_matchState_t* ms,
     }
 
     ms->hashLog3 = hashLog3;
+    ms->lazySkipping = 0;
 
     ZSTD_invalidateMatchState(ms);
 
@@ -1931,6 +1973,27 @@ ZSTD_reset_matchState(ZSTD_matchState_t* ms,
         ZSTD_cwksp_clean_tables(ws);
     }
 
+    if (ZSTD_rowMatchFinderUsed(cParams->strategy, useRowMatchFinder)) {
+        /* Row match finder needs an additional table of hashes ("tags") */
+        size_t const tagTableSize = hSize;
+        /* We want to generate a new salt in case we reset a Cctx, but we always want to use
+         * 0 when we reset a Cdict */
+        if(forWho == ZSTD_resetTarget_CCtx) {
+            ms->tagTable = (BYTE*) ZSTD_cwksp_reserve_aligned_init_once(ws, tagTableSize);
+            ZSTD_advanceHashSalt(ms);
+        } else {
+            /* When we are not salting we want to always memset the memory */
+            ms->tagTable = (BYTE*) ZSTD_cwksp_reserve_aligned(ws, tagTableSize);
+            ZSTD_memset(ms->tagTable, 0, tagTableSize);
+            ms->hashSalt = 0;
+        }
+        {   /* Switch to 32-entry rows if searchLog is 5 (or more) */
+            U32 const rowLog = BOUNDED(4, cParams->searchLog, 6);
+            assert(cParams->hashLog >= rowLog);
+            ms->rowHashLog = cParams->hashLog - rowLog;
+        }
+    }
+
     /* opt parser space */
     if ((forWho == ZSTD_resetTarget_CCtx) && (cParams->strategy >= ZSTD_btopt)) {
         DEBUGLOG(4, "reserving optimal parser space");
@@ -1942,19 +2005,6 @@ ZSTD_reset_matchState(ZSTD_matchState_t* ms,
         ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t));
     }
 
-    if (ZSTD_rowMatchFinderUsed(cParams->strategy, useRowMatchFinder)) {
-        {   /* Row match finder needs an additional table of hashes ("tags") */
-            size_t const tagTableSize = hSize*sizeof(U16);
-            ms->tagTable = (U16*)ZSTD_cwksp_reserve_aligned(ws, tagTableSize);
-            if (ms->tagTable) ZSTD_memset(ms->tagTable, 0, tagTableSize);
-        }
-        {   /* Switch to 32-entry rows if searchLog is 5 (or more) */
-            U32 const rowLog = BOUNDED(4, cParams->searchLog, 6);
-            assert(cParams->hashLog >= rowLog);
-            ms->rowHashLog = cParams->hashLog - rowLog;
-        }
-    }
-
     ms->cParams = *cParams;
 
     RETURN_ERROR_IF(ZSTD_cwksp_reserve_failed(ws), memory_allocation,
@@ -2101,13 +2151,46 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
 
         ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock);
 
+        FORWARD_IF_ERROR(ZSTD_reset_matchState(
+                &zc->blockState.matchState,
+                ws,
+                &params->cParams,
+                params->useRowMatchFinder,
+                crp,
+                needsIndexReset,
+                ZSTD_resetTarget_CCtx), "");
+
+        zc->seqStore.sequencesStart = (seqDef*)ZSTD_cwksp_reserve_aligned(ws, maxNbSeq * sizeof(seqDef));
+
+        /* ldm hash table */
+        if (params->ldmParams.enableLdm == ZSTD_ps_enable) {
+            /* TODO: avoid memset? */
+            size_t const ldmHSize = ((size_t)1) << params->ldmParams.hashLog;
+            zc->ldmState.hashTable = (ldmEntry_t*)ZSTD_cwksp_reserve_aligned(ws, ldmHSize * sizeof(ldmEntry_t));
+            ZSTD_memset(zc->ldmState.hashTable, 0, ldmHSize * sizeof(ldmEntry_t));
+            zc->ldmSequences = (rawSeq*)ZSTD_cwksp_reserve_aligned(ws, maxNbLdmSeq * sizeof(rawSeq));
+            zc->maxNbLdmSequences = maxNbLdmSeq;
+
+            ZSTD_window_init(&zc->ldmState.window);
+            zc->ldmState.loadedDictEnd = 0;
+        }
+
+        /* reserve space for block-level external sequences */
+        if (params->useSequenceProducer) {
+            size_t const maxNbExternalSeq = ZSTD_sequenceBound(blockSize);
+            zc->externalMatchCtx.seqBufferCapacity = maxNbExternalSeq;
+            zc->externalMatchCtx.seqBuffer =
+                (ZSTD_Sequence*)ZSTD_cwksp_reserve_aligned(ws, maxNbExternalSeq * sizeof(ZSTD_Sequence));
+        }
+
+        /* buffers */
+
         /* ZSTD_wildcopy() is used to copy into the literals buffer,
          * so we have to oversize the buffer by WILDCOPY_OVERLENGTH bytes.
          */
         zc->seqStore.litStart = ZSTD_cwksp_reserve_buffer(ws, blockSize + WILDCOPY_OVERLENGTH);
         zc->seqStore.maxNbLit = blockSize;
 
-        /* buffers */
         zc->bufferedPolicy = zbuff;
         zc->inBuffSize = buffInSize;
         zc->inBuff = (char*)ZSTD_cwksp_reserve_buffer(ws, buffInSize);
@@ -2130,40 +2213,9 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
         zc->seqStore.llCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE));
         zc->seqStore.mlCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE));
         zc->seqStore.ofCode = ZSTD_cwksp_reserve_buffer(ws, maxNbSeq * sizeof(BYTE));
-        zc->seqStore.sequencesStart = (seqDef*)ZSTD_cwksp_reserve_aligned(ws, maxNbSeq * sizeof(seqDef));
-
-        FORWARD_IF_ERROR(ZSTD_reset_matchState(
-            &zc->blockState.matchState,
-            ws,
-            &params->cParams,
-            params->useRowMatchFinder,
-            crp,
-            needsIndexReset,
-            ZSTD_resetTarget_CCtx), "");
-
-        /* ldm hash table */
-        if (params->ldmParams.enableLdm == ZSTD_ps_enable) {
-            /* TODO: avoid memset? */
-            size_t const ldmHSize = ((size_t)1) << params->ldmParams.hashLog;
-            zc->ldmState.hashTable = (ldmEntry_t*)ZSTD_cwksp_reserve_aligned(ws, ldmHSize * sizeof(ldmEntry_t));
-            ZSTD_memset(zc->ldmState.hashTable, 0, ldmHSize * sizeof(ldmEntry_t));
-            zc->ldmSequences = (rawSeq*)ZSTD_cwksp_reserve_aligned(ws, maxNbLdmSeq * sizeof(rawSeq));
-            zc->maxNbLdmSequences = maxNbLdmSeq;
-
-            ZSTD_window_init(&zc->ldmState.window);
-            zc->ldmState.loadedDictEnd = 0;
-        }
-
-        /* reserve space for block-level external sequences */
-        if (params->useSequenceProducer) {
-            size_t const maxNbExternalSeq = ZSTD_sequenceBound(blockSize);
-            zc->externalMatchCtx.seqBufferCapacity = maxNbExternalSeq;
-            zc->externalMatchCtx.seqBuffer =
-                (ZSTD_Sequence*)ZSTD_cwksp_reserve_aligned(ws, maxNbExternalSeq * sizeof(ZSTD_Sequence));
-        }
 
         DEBUGLOG(3, "wksp: finished allocating, %zd bytes remain available", ZSTD_cwksp_available_space(ws));
-        assert(ZSTD_cwksp_estimated_space_within_bounds(ws, neededSpace, resizeWorkspace));
+        assert(ZSTD_cwksp_estimated_space_within_bounds(ws, neededSpace));
 
         zc->initialized = 1;
 
@@ -2338,10 +2390,11 @@ static size_t ZSTD_resetCCtx_byCopyingCDict(ZSTD_CCtx* cctx,
         }
         /* copy tag table */
         if (ZSTD_rowMatchFinderUsed(cdict_cParams->strategy, cdict->useRowMatchFinder)) {
-            size_t const tagTableSize = hSize*sizeof(U16);
+            size_t const tagTableSize = hSize;
             ZSTD_memcpy(cctx->blockState.matchState.tagTable,
-                cdict->matchState.tagTable,
-                tagTableSize);
+                        cdict->matchState.tagTable,
+                        tagTableSize);
+            cctx->blockState.matchState.hashSalt = cdict->matchState.hashSalt;
         }
     }
 
@@ -3858,9 +3911,10 @@ ZSTD_seqStore_resolveOffCodes(repcodes_t* const dRepcodes, repcodes_t* const cRe
                         const seqStore_t* const seqStore, U32 const nbSeq)
 {
     U32 idx = 0;
+    U32 const longLitLenIdx = seqStore->longLengthType == ZSTD_llt_literalLength ? seqStore->longLengthPos : nbSeq;
     for (; idx < nbSeq; ++idx) {
         seqDef* const seq = seqStore->sequencesStart + idx;
-        U32 const ll0 = (seq->litLength == 0);
+        U32 const ll0 = (seq->litLength == 0) && (idx != longLitLenIdx);
         U32 const offBase = seq->offBase;
         assert(offBase > 0);
         if (OFFBASE_IS_REPCODE(offBase)) {
@@ -4576,31 +4630,51 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,
     }
 }
 
-size_t ZSTD_compressContinue (ZSTD_CCtx* cctx,
-                              void* dst, size_t dstCapacity,
-                        const void* src, size_t srcSize)
+size_t ZSTD_compressContinue_public(ZSTD_CCtx* cctx,
+                                        void* dst, size_t dstCapacity,
+                                  const void* src, size_t srcSize)
 {
     DEBUGLOG(5, "ZSTD_compressContinue (srcSize=%u)", (unsigned)srcSize);
     return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 0 /* last chunk */);
 }
 
+/* NOTE: Must just wrap ZSTD_compressContinue_public() */
+size_t ZSTD_compressContinue(ZSTD_CCtx* cctx,
+                             void* dst, size_t dstCapacity,
+                       const void* src, size_t srcSize)
+{
+    return ZSTD_compressContinue_public(cctx, dst, dstCapacity, src, srcSize);
+}
 
-size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx)
+static size_t ZSTD_getBlockSize_deprecated(const ZSTD_CCtx* cctx)
 {
     ZSTD_compressionParameters const cParams = cctx->appliedParams.cParams;
     assert(!ZSTD_checkCParams(cParams));
     return MIN(cctx->appliedParams.maxBlockSize, (size_t)1 << cParams.windowLog);
 }
 
-size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+/* NOTE: Must just wrap ZSTD_getBlockSize_deprecated() */
+size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx)
+{
+    return ZSTD_getBlockSize_deprecated(cctx);
+}
+
+/* NOTE: Must just wrap ZSTD_compressBlock_deprecated() */
+size_t ZSTD_compressBlock_deprecated(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
 {
     DEBUGLOG(5, "ZSTD_compressBlock: srcSize = %u", (unsigned)srcSize);
-    { size_t const blockSizeMax = ZSTD_getBlockSize(cctx);
+    { size_t const blockSizeMax = ZSTD_getBlockSize_deprecated(cctx);
       RETURN_ERROR_IF(srcSize > blockSizeMax, srcSize_wrong, "input is larger than a block"); }
 
     return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0 /* frame mode */, 0 /* last chunk */);
 }
 
+/* NOTE: Must just wrap ZSTD_compressBlock_deprecated() */
+size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_deprecated(cctx, dst, dstCapacity, src, srcSize);
+}
+
 /*! ZSTD_loadDictionaryContent() :
  *  @return : 0, or an error code
  */
@@ -4644,31 +4718,42 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,
             ip = iend - maxDictSize;
             src = ip;
             srcSize = maxDictSize;
-    }   }
+        }
+    }
 
     if (srcSize > ZSTD_CHUNKSIZE_MAX) {
         /* We must have cleared our windows when our source is this large. */
         assert(ZSTD_window_isEmpty(ms->window));
         if (loadLdmDict) assert(ZSTD_window_isEmpty(ls->window));
     }
+    ZSTD_window_update(&ms->window, src, srcSize, /* forceNonContiguous */ 0);
 
     DEBUGLOG(4, "ZSTD_loadDictionaryContent(): useRowMatchFinder=%d", (int)params->useRowMatchFinder);
-    ZSTD_window_update(&ms->window, src, srcSize, /* forceNonContiguous */ 0);
-    ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base);
-    ms->forceNonContiguous = params->deterministicRefPrefix;
 
-    if (loadLdmDict) {
+    if (loadLdmDict) { /* Load the entire dict into LDM matchfinders. */
         ZSTD_window_update(&ls->window, src, srcSize, /* forceNonContiguous */ 0);
         ls->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ls->window.base);
+        ZSTD_ldm_fillHashTable(ls, ip, iend, &params->ldmParams);
+    }
+
+    /* If the dict is larger than we can reasonably index in our tables, only load the suffix. */
+    if (params->cParams.strategy < ZSTD_btultra) {
+        U32 maxDictSize = 8U << MIN(MAX(params->cParams.hashLog, params->cParams.chainLog), 28);
+        if (srcSize > maxDictSize) {
+            ip = iend - maxDictSize;
+            src = ip;
+            srcSize = maxDictSize;
+        }
     }
 
+    ms->nextToUpdate = (U32)(ip - ms->window.base);
+    ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base);
+    ms->forceNonContiguous = params->deterministicRefPrefix;
+
     if (srcSize <= HASH_READ_SIZE) return 0;
 
     ZSTD_overflowCorrectIfNeeded(ms, ws, params, ip, iend);
 
-    if (loadLdmDict)
-        ZSTD_ldm_fillHashTable(ls, ip, iend, &params->ldmParams);
-
     switch(params->cParams.strategy)
     {
     case ZSTD_fast:
@@ -4688,7 +4773,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,
         } else {
             assert(params->useRowMatchFinder != ZSTD_ps_auto);
             if (params->useRowMatchFinder == ZSTD_ps_enable) {
-                size_t const tagTableSize = ((size_t)1 << params->cParams.hashLog) * sizeof(U16);
+                size_t const tagTableSize = ((size_t)1 << params->cParams.hashLog);
                 ZSTD_memset(ms->tagTable, 0, tagTableSize);
                 ZSTD_row_update(ms, iend-HASH_READ_SIZE);
                 DEBUGLOG(4, "Using row-based hash table for lazy dict");
@@ -4991,8 +5076,8 @@ size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx,
                                             &cctxParams, pledgedSrcSize);
 }
 
-size_t
-ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)
+static size_t
+ZSTD_compressBegin_usingDict_deprecated(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)
 {
     ZSTD_CCtx_params cctxParams;
     {   ZSTD_parameters const params = ZSTD_getParams_internal(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize, ZSTD_cpm_noAttachDict);
@@ -5003,9 +5088,15 @@ ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize,
                                        &cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, ZSTDb_not_buffered);
 }
 
+size_t
+ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)
+{
+    return ZSTD_compressBegin_usingDict_deprecated(cctx, dict, dictSize, compressionLevel);
+}
+
 size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel)
 {
-    return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel);
+    return ZSTD_compressBegin_usingDict_deprecated(cctx, NULL, 0, compressionLevel);
 }
 
 
@@ -5075,9 +5166,9 @@ void ZSTD_CCtx_trace(ZSTD_CCtx* cctx, size_t extraCSize)
 #endif
 }
 
-size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,
-                         void* dst, size_t dstCapacity,
-                   const void* src, size_t srcSize)
+size_t ZSTD_compressEnd_public(ZSTD_CCtx* cctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize)
 {
     size_t endResult;
     size_t const cSize = ZSTD_compressContinue_internal(cctx,
@@ -5101,6 +5192,14 @@ size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,
     return cSize + endResult;
 }
 
+/* NOTE: Must just wrap ZSTD_compressEnd_public() */
+size_t ZSTD_compressEnd(ZSTD_CCtx* cctx,
+                        void* dst, size_t dstCapacity,
+                  const void* src, size_t srcSize)
+{
+    return ZSTD_compressEnd_public(cctx, dst, dstCapacity, src, srcSize);
+}
+
 size_t ZSTD_compress_advanced (ZSTD_CCtx* cctx,
                                void* dst, size_t dstCapacity,
                          const void* src, size_t srcSize,
@@ -5129,7 +5228,7 @@ size_t ZSTD_compress_advanced_internal(
     FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
                          dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,
                          params, srcSize, ZSTDb_not_buffered) , "");
-    return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
+    return ZSTD_compressEnd_public(cctx, dst, dstCapacity, src, srcSize);
 }
 
 size_t ZSTD_compress_usingDict(ZSTD_CCtx* cctx,
@@ -5451,6 +5550,7 @@ const ZSTD_CDict* ZSTD_initStaticCDict(
     params.cParams = cParams;
     params.useRowMatchFinder = useRowMatchFinder;
     cdict->useRowMatchFinder = useRowMatchFinder;
+    cdict->compressionLevel = ZSTD_NO_CLEVEL;
 
     if (ZSTD_isError( ZSTD_initCDict_internal(cdict,
                                               dict, dictSize,
@@ -5530,12 +5630,17 @@ size_t ZSTD_compressBegin_usingCDict_advanced(
 
 /* ZSTD_compressBegin_usingCDict() :
  * cdict must be != NULL */
-size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
+size_t ZSTD_compressBegin_usingCDict_deprecated(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
 {
     ZSTD_frameParameters const fParams = { 0 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };
     return ZSTD_compressBegin_usingCDict_internal(cctx, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN);
 }
 
+size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
+{
+    return ZSTD_compressBegin_usingCDict_deprecated(cctx, cdict);
+}
+
 /*! ZSTD_compress_usingCDict_internal():
  * Implementation of various ZSTD_compress_usingCDict* functions.
  */
@@ -5545,7 +5650,7 @@ static size_t ZSTD_compress_usingCDict_internal(ZSTD_CCtx* cctx,
                                 const ZSTD_CDict* cdict, ZSTD_frameParameters fParams)
 {
     FORWARD_IF_ERROR(ZSTD_compressBegin_usingCDict_internal(cctx, cdict, fParams, srcSize), ""); /* will check if cdict != NULL */
-    return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
+    return ZSTD_compressEnd_public(cctx, dst, dstCapacity, src, srcSize);
 }
 
 /*! ZSTD_compress_usingCDict_advanced():
@@ -5803,7 +5908,7 @@ static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
                 || zcs->appliedParams.outBufferMode == ZSTD_bm_stable)  /* OR we are allowed to return dstSizeTooSmall */
               && (zcs->inBuffPos == 0) ) {
                 /* shortcut to compression pass directly into output buffer */
-                size_t const cSize = ZSTD_compressEnd(zcs,
+                size_t const cSize = ZSTD_compressEnd_public(zcs,
                                                 op, oend-op, ip, iend-ip);
                 DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize);
                 FORWARD_IF_ERROR(cSize, "ZSTD_compressEnd failed");
@@ -5861,9 +5966,9 @@ static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
                 if (inputBuffered) {
                     unsigned const lastBlock = (flushMode == ZSTD_e_end) && (ip==iend);
                     cSize = lastBlock ?
-                            ZSTD_compressEnd(zcs, cDst, oSize,
+                            ZSTD_compressEnd_public(zcs, cDst, oSize,
                                         zcs->inBuff + zcs->inToCompress, iSize) :
-                            ZSTD_compressContinue(zcs, cDst, oSize,
+                            ZSTD_compressContinue_public(zcs, cDst, oSize,
                                         zcs->inBuff + zcs->inToCompress, iSize);
                     FORWARD_IF_ERROR(cSize, "%s", lastBlock ? "ZSTD_compressEnd failed" : "ZSTD_compressContinue failed");
                     zcs->frameEnded = lastBlock;
@@ -5879,8 +5984,8 @@ static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
                 } else { /* !inputBuffered, hence ZSTD_bm_stable */
                     unsigned const lastBlock = (flushMode == ZSTD_e_end) && (ip + iSize == iend);
                     cSize = lastBlock ?
-                            ZSTD_compressEnd(zcs, cDst, oSize, ip, iSize) :
-                            ZSTD_compressContinue(zcs, cDst, oSize, ip, iSize);
+                            ZSTD_compressEnd_public(zcs, cDst, oSize, ip, iSize) :
+                            ZSTD_compressContinue_public(zcs, cDst, oSize, ip, iSize);
                     /* Consume the input prior to error checking to mirror buffered mode. */
                     if (ip) ip += iSize;
                     FORWARD_IF_ERROR(cSize, "%s", lastBlock ? "ZSTD_compressEnd failed" : "ZSTD_compressContinue failed");
diff --git a/contrib/libs/zstd/lib/compress/zstd_compress_internal.h b/contrib/libs/zstd/lib/compress/zstd_compress_internal.h
index cbb85e527e..10f68d010e 100644
--- a/contrib/libs/zstd/lib/compress/zstd_compress_internal.h
+++ b/contrib/libs/zstd/lib/compress/zstd_compress_internal.h
@@ -226,8 +226,10 @@ struct ZSTD_matchState_t {
     U32 hashLog3;           /* dispatch table for matches of len==3 : larger == faster, more memory */
 
     U32 rowHashLog;                          /* For row-based matchfinder: Hashlog based on nb of rows in the hashTable.*/
-    U16* tagTable;                           /* For row-based matchFinder: A row-based table containing the hashes and head index. */
+    BYTE* tagTable;                          /* For row-based matchFinder: A row-based table containing the hashes and head index. */
     U32 hashCache[ZSTD_ROW_HASH_CACHE_SIZE]; /* For row-based matchFinder: a cache of hashes to improve speed */
+    U64 hashSalt;                            /* For row-based matchFinder: salts the hash for re-use of tag table */
+    U32 hashSaltEntropy;                     /* For row-based matchFinder: collects entropy for salt generation */
 
     U32* hashTable;
     U32* hashTable3;
@@ -247,6 +249,13 @@ struct ZSTD_matchState_t {
      * This behavior is controlled from the cctx ms.
      * This parameter has no effect in the cdict ms. */
     int prefetchCDictTables;
+
+    /* When == 0, lazy match finders insert every position.
+     * When != 0, lazy match finders only insert positions they search.
+     * This allows them to skip much faster over incompressible data,
+     * at a small cost to compression ratio.
+     */
+    int lazySkipping;
 };
 
 typedef struct {
@@ -787,28 +796,35 @@ ZSTD_count_2segments(const BYTE* ip, const BYTE* match,
  *  Hashes
  ***************************************/
 static const U32 prime3bytes = 506832829U;
-static U32    ZSTD_hash3(U32 u, U32 h) { assert(h <= 32); 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 U32    ZSTD_hash3(U32 u, U32 h, U32 s) { assert(h <= 32); return (((u << (32-24)) * prime3bytes) ^ s)  >> (32-h) ; }
+MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h, 0); } /* only in zstd_opt.h */
+MEM_STATIC size_t ZSTD_hash3PtrS(const void* ptr, U32 h, U32 s) { return ZSTD_hash3(MEM_readLE32(ptr), h, s); }
 
 static const U32 prime4bytes = 2654435761U;
-static U32    ZSTD_hash4(U32 u, U32 h) { assert(h <= 32); return (u * prime4bytes) >> (32-h) ; }
-static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_readLE32(ptr), h); }
+static U32    ZSTD_hash4(U32 u, U32 h, U32 s) { assert(h <= 32); return ((u * prime4bytes) ^ s) >> (32-h) ; }
+static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_readLE32(ptr), h, 0); }
+static size_t ZSTD_hash4PtrS(const void* ptr, U32 h, U32 s) { return ZSTD_hash4(MEM_readLE32(ptr), h, s); }
 
 static const U64 prime5bytes = 889523592379ULL;
-static size_t ZSTD_hash5(U64 u, U32 h) { assert(h <= 64); 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 size_t ZSTD_hash5(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u  << (64-40)) * prime5bytes) ^ s) >> (64-h)) ; }
+static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h, 0); }
+static size_t ZSTD_hash5PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash5(MEM_readLE64(p), h, s); }
 
 static const U64 prime6bytes = 227718039650203ULL;
-static size_t ZSTD_hash6(U64 u, U32 h) { assert(h <= 64); 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 size_t ZSTD_hash6(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u  << (64-48)) * prime6bytes) ^ s) >> (64-h)) ; }
+static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h, 0); }
+static size_t ZSTD_hash6PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash6(MEM_readLE64(p), h, s); }
 
 static const U64 prime7bytes = 58295818150454627ULL;
-static size_t ZSTD_hash7(U64 u, U32 h) { assert(h <= 64); 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 size_t ZSTD_hash7(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u  << (64-56)) * prime7bytes) ^ s) >> (64-h)) ; }
+static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h, 0); }
+static size_t ZSTD_hash7PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash7(MEM_readLE64(p), h, s); }
 
 static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
-static size_t ZSTD_hash8(U64 u, U32 h) { assert(h <= 64); 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_hash8(U64 u, U32 h, U64 s) { assert(h <= 64); return (size_t)((((u) * prime8bytes)  ^ s) >> (64-h)) ; }
+static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h, 0); }
+static size_t ZSTD_hash8PtrS(const void* p, U32 h, U64 s) { return ZSTD_hash8(MEM_readLE64(p), h, s); }
+
 
 MEM_STATIC FORCE_INLINE_ATTR
 size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)
@@ -828,6 +844,24 @@ size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)
     }
 }
 
+MEM_STATIC FORCE_INLINE_ATTR
+size_t ZSTD_hashPtrSalted(const void* p, U32 hBits, U32 mls, const U64 hashSalt) {
+    /* Although some of these hashes do support hBits up to 64, some do not.
+     * To be on the safe side, always avoid hBits > 32. */
+    assert(hBits <= 32);
+
+    switch(mls)
+    {
+        default:
+        case 4: return ZSTD_hash4PtrS(p, hBits, (U32)hashSalt);
+        case 5: return ZSTD_hash5PtrS(p, hBits, hashSalt);
+        case 6: return ZSTD_hash6PtrS(p, hBits, hashSalt);
+        case 7: return ZSTD_hash7PtrS(p, hBits, hashSalt);
+        case 8: return ZSTD_hash8PtrS(p, hBits, hashSalt);
+    }
+}
+
+
 /** ZSTD_ipow() :
  * Return base^exponent.
  */
@@ -1475,4 +1509,24 @@ ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition*
                                    const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
                                    const void* src, size_t blockSize, ZSTD_paramSwitch_e externalRepSearch);
 
+
+/* ===============================================================
+ * Deprecated definitions that are still used internally to avoid
+ * deprecation warnings. These functions are exactly equivalent to
+ * their public variants, but avoid the deprecation warnings.
+ * =============================================================== */
+
+size_t ZSTD_compressBegin_usingCDict_deprecated(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
+
+size_t ZSTD_compressContinue_public(ZSTD_CCtx* cctx,
+                                    void* dst, size_t dstCapacity,
+                              const void* src, size_t srcSize);
+
+size_t ZSTD_compressEnd_public(ZSTD_CCtx* cctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize);
+
+size_t ZSTD_compressBlock_deprecated(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+
 #endif /* ZSTD_COMPRESS_H */
diff --git a/contrib/libs/zstd/lib/compress/zstd_cwksp.h b/contrib/libs/zstd/lib/compress/zstd_cwksp.h
index 97676693b5..cc7fb1c715 100644
--- a/contrib/libs/zstd/lib/compress/zstd_cwksp.h
+++ b/contrib/libs/zstd/lib/compress/zstd_cwksp.h
@@ -14,7 +14,9 @@
 /*-*************************************
 *  Dependencies
 ***************************************/
+#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customFree */
 #include "../common/zstd_internal.h"
+#include "../common/portability_macros.h"
 
 #if defined (__cplusplus)
 extern "C" {
@@ -44,8 +46,9 @@ extern "C" {
 ***************************************/
 typedef enum {
     ZSTD_cwksp_alloc_objects,
-    ZSTD_cwksp_alloc_buffers,
-    ZSTD_cwksp_alloc_aligned
+    ZSTD_cwksp_alloc_aligned_init_once,
+    ZSTD_cwksp_alloc_aligned,
+    ZSTD_cwksp_alloc_buffers
 } ZSTD_cwksp_alloc_phase_e;
 
 /**
@@ -98,8 +101,8 @@ typedef enum {
  *
  * Workspace Layout:
  *
- * [                        ... workspace ...                         ]
- * [objects][tables ... ->] free space [<- ... aligned][<- ... buffers]
+ * [                        ... workspace ...                           ]
+ * [objects][tables ->] free space [<- buffers][<- aligned][<- init once]
  *
  * The various objects that live in the workspace are divided into the
  * following categories, and are allocated separately:
@@ -123,9 +126,18 @@ typedef enum {
  *   uint32_t arrays, all of whose values are between 0 and (nextSrc - base).
  *   Their sizes depend on the cparams. These tables are 64-byte aligned.
  *
- * - Aligned: these buffers are used for various purposes that require 4 byte
- *   alignment, but don't require any initialization before they're used. These
- *   buffers are each aligned to 64 bytes.
+ * - Init once: these buffers require to be initialized at least once before
+ *   use. They should be used when we want to skip memory initialization
+ *   while not triggering memory checkers (like Valgrind) when reading from
+ *   from this memory without writing to it first.
+ *   These buffers should be used carefully as they might contain data
+ *   from previous compressions.
+ *   Buffers are aligned to 64 bytes.
+ *
+ * - Aligned: these buffers don't require any initialization before they're
+ *   used. The user of the buffer should make sure they write into a buffer
+ *   location before reading from it.
+ *   Buffers are aligned to 64 bytes.
  *
  * - Buffers: these buffers are used for various purposes that don't require
  *   any alignment or initialization before they're used. This means they can
@@ -137,8 +149,9 @@ typedef enum {
  * correctly packed into the workspace buffer. That order is:
  *
  * 1. Objects
- * 2. Buffers
- * 3. Aligned/Tables
+ * 2. Init once / Tables
+ * 3. Aligned / Tables
+ * 4. Buffers / Tables
  *
  * Attempts to reserve objects of different types out of order will fail.
  */
@@ -150,6 +163,7 @@ typedef struct {
     void* tableEnd;
     void* tableValidEnd;
     void* allocStart;
+    void* initOnceStart;
 
     BYTE allocFailed;
     int workspaceOversizedDuration;
@@ -162,6 +176,7 @@ typedef struct {
 ***************************************/
 
 MEM_STATIC size_t ZSTD_cwksp_available_space(ZSTD_cwksp* ws);
+MEM_STATIC void*  ZSTD_cwksp_initialAllocStart(ZSTD_cwksp* ws);
 
 MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) {
     (void)ws;
@@ -171,6 +186,20 @@ MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) {
     assert(ws->tableEnd <= ws->allocStart);
     assert(ws->tableValidEnd <= ws->allocStart);
     assert(ws->allocStart <= ws->workspaceEnd);
+    assert(ws->initOnceStart <= ZSTD_cwksp_initialAllocStart(ws));
+    assert(ws->workspace <= ws->initOnceStart);
+#if ZSTD_MEMORY_SANITIZER
+    {
+        intptr_t const offset = __msan_test_shadow(ws->initOnceStart,
+            (U8*)ZSTD_cwksp_initialAllocStart(ws) - (U8*)ws->initOnceStart);
+#if defined(ZSTD_MSAN_PRINT)
+        if(offset!=-1) {
+            __msan_print_shadow((U8*)ws->initOnceStart + offset - 8, 32);
+        }
+#endif
+        assert(offset==-1);
+    };
+#endif
 }
 
 /**
@@ -217,14 +246,10 @@ MEM_STATIC size_t ZSTD_cwksp_aligned_alloc_size(size_t size) {
  * for internal purposes (currently only alignment).
  */
 MEM_STATIC size_t ZSTD_cwksp_slack_space_required(void) {
-    /* For alignment, the wksp will always allocate an additional n_1=[1, 64] bytes
-     * to align the beginning of tables section, as well as another n_2=[0, 63] bytes
-     * to align the beginning of the aligned section.
-     *
-     * n_1 + n_2 == 64 bytes if the cwksp is freshly allocated, due to tables and
-     * aligneds being sized in multiples of 64 bytes.
+    /* For alignment, the wksp will always allocate an additional 2*ZSTD_CWKSP_ALIGNMENT_BYTES
+     * bytes to align the beginning of tables section and end of buffers;
      */
-    size_t const slackSpace = ZSTD_CWKSP_ALIGNMENT_BYTES;
+    size_t const slackSpace = ZSTD_CWKSP_ALIGNMENT_BYTES * 2;
     return slackSpace;
 }
 
@@ -237,11 +262,19 @@ MEM_STATIC size_t ZSTD_cwksp_bytes_to_align_ptr(void* ptr, const size_t alignByt
     size_t const alignBytesMask = alignBytes - 1;
     size_t const bytes = (alignBytes - ((size_t)ptr & (alignBytesMask))) & alignBytesMask;
     assert((alignBytes & alignBytesMask) == 0);
-    assert(bytes != ZSTD_CWKSP_ALIGNMENT_BYTES);
+    assert(bytes < alignBytes);
     return bytes;
 }
 
 /**
+ * Returns the initial value for allocStart which is used to determine the position from
+ * which we can allocate from the end of the workspace.
+ */
+MEM_STATIC void*  ZSTD_cwksp_initialAllocStart(ZSTD_cwksp* ws) {
+    return (void*)((size_t)ws->workspaceEnd & ~(ZSTD_CWKSP_ALIGNMENT_BYTES-1));
+}
+
+/**
  * Internal function. Do not use directly.
  * Reserves the given number of bytes within the aligned/buffer segment of the wksp,
  * which counts from the end of the wksp (as opposed to the object/table segment).
@@ -281,27 +314,16 @@ ZSTD_cwksp_internal_advance_phase(ZSTD_cwksp* ws, ZSTD_cwksp_alloc_phase_e phase
 {
     assert(phase >= ws->phase);
     if (phase > ws->phase) {
-        /* Going from allocating objects to allocating buffers */
-        if (ws->phase < ZSTD_cwksp_alloc_buffers &&
-                phase >= ZSTD_cwksp_alloc_buffers) {
+        /* Going from allocating objects to allocating initOnce / tables */
+        if (ws->phase < ZSTD_cwksp_alloc_aligned_init_once &&
+            phase >= ZSTD_cwksp_alloc_aligned_init_once) {
             ws->tableValidEnd = ws->objectEnd;
-        }
+            ws->initOnceStart = ZSTD_cwksp_initialAllocStart(ws);
 
-        /* Going from allocating buffers to allocating aligneds/tables */
-        if (ws->phase < ZSTD_cwksp_alloc_aligned &&
-                phase >= ZSTD_cwksp_alloc_aligned) {
-            {   /* Align the start of the "aligned" to 64 bytes. Use [1, 64] bytes. */
-                size_t const bytesToAlign =
-                    ZSTD_CWKSP_ALIGNMENT_BYTES - ZSTD_cwksp_bytes_to_align_ptr(ws->allocStart, ZSTD_CWKSP_ALIGNMENT_BYTES);
-                DEBUGLOG(5, "reserving aligned alignment addtl space: %zu", bytesToAlign);
-                ZSTD_STATIC_ASSERT((ZSTD_CWKSP_ALIGNMENT_BYTES & (ZSTD_CWKSP_ALIGNMENT_BYTES - 1)) == 0); /* power of 2 */
-                RETURN_ERROR_IF(!ZSTD_cwksp_reserve_internal_buffer_space(ws, bytesToAlign),
-                                memory_allocation, "aligned phase - alignment initial allocation failed!");
-            }
             {   /* Align the start of the tables to 64 bytes. Use [0, 63] bytes */
-                void* const alloc = ws->objectEnd;
+                void *const alloc = ws->objectEnd;
                 size_t const bytesToAlign = ZSTD_cwksp_bytes_to_align_ptr(alloc, ZSTD_CWKSP_ALIGNMENT_BYTES);
-                void* const objectEnd = (BYTE*)alloc + bytesToAlign;
+                void *const objectEnd = (BYTE *) alloc + bytesToAlign;
                 DEBUGLOG(5, "reserving table alignment addtl space: %zu", bytesToAlign);
                 RETURN_ERROR_IF(objectEnd > ws->workspaceEnd, memory_allocation,
                                 "table phase - alignment initial allocation failed!");
@@ -309,7 +331,9 @@ ZSTD_cwksp_internal_advance_phase(ZSTD_cwksp* ws, ZSTD_cwksp_alloc_phase_e phase
                 ws->tableEnd = objectEnd;  /* table area starts being empty */
                 if (ws->tableValidEnd < ws->tableEnd) {
                     ws->tableValidEnd = ws->tableEnd;
-        }   }   }
+                }
+            }
+        }
         ws->phase = phase;
         ZSTD_cwksp_assert_internal_consistency(ws);
     }
@@ -321,7 +345,7 @@ ZSTD_cwksp_internal_advance_phase(ZSTD_cwksp* ws, ZSTD_cwksp_alloc_phase_e phase
  */
 MEM_STATIC int ZSTD_cwksp_owns_buffer(const ZSTD_cwksp* ws, const void* ptr)
 {
-    return (ptr != NULL) && (ws->workspace <= ptr) && (ptr <= ws->workspaceEnd);
+    return (ptr != NULL) && (ws->workspace <= ptr) && (ptr < ws->workspaceEnd);
 }
 
 /**
@@ -368,6 +392,36 @@ MEM_STATIC BYTE* ZSTD_cwksp_reserve_buffer(ZSTD_cwksp* ws, size_t bytes)
 
 /**
  * Reserves and returns memory sized on and aligned on ZSTD_CWKSP_ALIGNMENT_BYTES (64 bytes).
+ * This memory has been initialized at least once in the past.
+ * This doesn't mean it has been initialized this time, and it might contain data from previous
+ * operations.
+ * The main usage is for algorithms that might need read access into uninitialized memory.
+ * The algorithm must maintain safety under these conditions and must make sure it doesn't
+ * leak any of the past data (directly or in side channels).
+ */
+MEM_STATIC void* ZSTD_cwksp_reserve_aligned_init_once(ZSTD_cwksp* ws, size_t bytes)
+{
+    size_t const alignedBytes = ZSTD_cwksp_align(bytes, ZSTD_CWKSP_ALIGNMENT_BYTES);
+    void* ptr = ZSTD_cwksp_reserve_internal(ws, alignedBytes, ZSTD_cwksp_alloc_aligned_init_once);
+    assert(((size_t)ptr & (ZSTD_CWKSP_ALIGNMENT_BYTES-1))== 0);
+    if(ptr && ptr < ws->initOnceStart) {
+        /* We assume the memory following the current allocation is either:
+         * 1. Not usable as initOnce memory (end of workspace)
+         * 2. Another initOnce buffer that has been allocated before (and so was previously memset)
+         * 3. An ASAN redzone, in which case we don't want to write on it
+         * For these reasons it should be fine to not explicitly zero every byte up to ws->initOnceStart.
+         * Note that we assume here that MSAN and ASAN cannot run in the same time. */
+        ZSTD_memset(ptr, 0, MIN((size_t)((U8*)ws->initOnceStart - (U8*)ptr), alignedBytes));
+        ws->initOnceStart = ptr;
+    }
+#if ZSTD_MEMORY_SANITIZER
+    assert(__msan_test_shadow(ptr, bytes) == -1);
+#endif
+    return ptr;
+}
+
+/**
+ * Reserves and returns memory sized on and aligned on ZSTD_CWKSP_ALIGNMENT_BYTES (64 bytes).
  */
 MEM_STATIC void* ZSTD_cwksp_reserve_aligned(ZSTD_cwksp* ws, size_t bytes)
 {
@@ -384,13 +438,17 @@ MEM_STATIC void* ZSTD_cwksp_reserve_aligned(ZSTD_cwksp* ws, size_t bytes)
  */
 MEM_STATIC void* ZSTD_cwksp_reserve_table(ZSTD_cwksp* ws, size_t bytes)
 {
-    const ZSTD_cwksp_alloc_phase_e phase = ZSTD_cwksp_alloc_aligned;
+    const ZSTD_cwksp_alloc_phase_e phase = ZSTD_cwksp_alloc_aligned_init_once;
     void* alloc;
     void* end;
     void* top;
 
-    if (ZSTD_isError(ZSTD_cwksp_internal_advance_phase(ws, phase))) {
-        return NULL;
+    /* We can only start allocating tables after we are done reserving space for objects at the
+     * start of the workspace */
+    if(ws->phase < phase) {
+        if (ZSTD_isError(ZSTD_cwksp_internal_advance_phase(ws, phase))) {
+            return NULL;
+        }
     }
     alloc = ws->tableEnd;
     end = (BYTE *)alloc + bytes;
@@ -469,11 +527,19 @@ MEM_STATIC void ZSTD_cwksp_mark_tables_dirty(ZSTD_cwksp* ws)
 #if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE)
     /* To validate that the table re-use logic is sound, and that we don't
      * access table space that we haven't cleaned, we re-"poison" the table
-     * space every time we mark it dirty. */
+     * space every time we mark it dirty.
+     * Since tableValidEnd space and initOnce space may overlap we don't poison
+     * the initOnce portion as it break its promise. This means that this poisoning
+     * check isn't always applied fully. */
     {
         size_t size = (BYTE*)ws->tableValidEnd - (BYTE*)ws->objectEnd;
         assert(__msan_test_shadow(ws->objectEnd, size) == -1);
-        __msan_poison(ws->objectEnd, size);
+        if((BYTE*)ws->tableValidEnd < (BYTE*)ws->initOnceStart) {
+            __msan_poison(ws->objectEnd, size);
+        } else {
+            assert(ws->initOnceStart >= ws->objectEnd);
+            __msan_poison(ws->objectEnd, (BYTE*)ws->initOnceStart - (BYTE*)ws->objectEnd);
+        }
     }
 #endif
 
@@ -538,11 +604,14 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
 #if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE)
     /* To validate that the context re-use logic is sound, and that we don't
      * access stuff that this compression hasn't initialized, we re-"poison"
-     * the workspace (or at least the non-static, non-table parts of it)
-     * every time we start a new compression. */
+     * the workspace except for the areas in which we expect memory re-use
+     * without initialization (objects, valid tables area and init once
+     * memory). */
     {
-        size_t size = (BYTE*)ws->workspaceEnd - (BYTE*)ws->tableValidEnd;
-        __msan_poison(ws->tableValidEnd, size);
+        if((BYTE*)ws->tableValidEnd < (BYTE*)ws->initOnceStart) {
+            size_t size = (BYTE*)ws->initOnceStart - (BYTE*)ws->tableValidEnd;
+            __msan_poison(ws->tableValidEnd, size);
+        }
     }
 #endif
 
@@ -558,10 +627,10 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
 #endif
 
     ws->tableEnd = ws->objectEnd;
-    ws->allocStart = ws->workspaceEnd;
+    ws->allocStart = ZSTD_cwksp_initialAllocStart(ws);
     ws->allocFailed = 0;
-    if (ws->phase > ZSTD_cwksp_alloc_buffers) {
-        ws->phase = ZSTD_cwksp_alloc_buffers;
+    if (ws->phase > ZSTD_cwksp_alloc_aligned_init_once) {
+        ws->phase = ZSTD_cwksp_alloc_aligned_init_once;
     }
     ZSTD_cwksp_assert_internal_consistency(ws);
 }
@@ -578,6 +647,7 @@ MEM_STATIC void ZSTD_cwksp_init(ZSTD_cwksp* ws, void* start, size_t size, ZSTD_c
     ws->workspaceEnd = (BYTE*)start + size;
     ws->objectEnd = ws->workspace;
     ws->tableValidEnd = ws->objectEnd;
+    ws->initOnceStart = ZSTD_cwksp_initialAllocStart(ws);
     ws->phase = ZSTD_cwksp_alloc_objects;
     ws->isStatic = isStatic;
     ZSTD_cwksp_clear(ws);
@@ -630,17 +700,11 @@ MEM_STATIC int ZSTD_cwksp_reserve_failed(const ZSTD_cwksp* ws) {
  * Returns if the estimated space needed for a wksp is within an acceptable limit of the
  * actual amount of space used.
  */
-MEM_STATIC int ZSTD_cwksp_estimated_space_within_bounds(const ZSTD_cwksp* const ws,
-                                                        size_t const estimatedSpace, int resizedWorkspace) {
-    if (resizedWorkspace) {
-        /* Resized/newly allocated wksp should have exact bounds */
-        return ZSTD_cwksp_used(ws) == estimatedSpace;
-    } else {
-        /* Due to alignment, when reusing a workspace, we can actually consume 63 fewer or more bytes
-         * than estimatedSpace. See the comments in zstd_cwksp.h for details.
-         */
-        return (ZSTD_cwksp_used(ws) >= estimatedSpace - 63) && (ZSTD_cwksp_used(ws) <= estimatedSpace + 63);
-    }
+MEM_STATIC int ZSTD_cwksp_estimated_space_within_bounds(const ZSTD_cwksp *const ws, size_t const estimatedSpace) {
+    /* We have an alignment space between objects and tables between tables and buffers, so we can have up to twice
+     * the alignment bytes difference between estimation and actual usage */
+    return (estimatedSpace - ZSTD_cwksp_slack_space_required()) <= ZSTD_cwksp_used(ws) &&
+           ZSTD_cwksp_used(ws) <= estimatedSpace;
 }
 
 
diff --git a/contrib/libs/zstd/lib/compress/zstd_lazy.c b/contrib/libs/zstd/lib/compress/zstd_lazy.c
index a247342729..5ba88e8678 100644
--- a/contrib/libs/zstd/lib/compress/zstd_lazy.c
+++ b/contrib/libs/zstd/lib/compress/zstd_lazy.c
@@ -12,6 +12,8 @@
 #include "zstd_lazy.h"
 #include "../common/bits.h" /* ZSTD_countTrailingZeros64 */
 
+#define kLazySkippingStep 8
+
 
 /*-*************************************
 *  Binary Tree search
@@ -618,7 +620,7 @@ size_t ZSTD_dedicatedDictSearch_lazy_search(size_t* offsetPtr, size_t ml, U32 nb
 FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
                         ZSTD_matchState_t* ms,
                         const ZSTD_compressionParameters* const cParams,
-                        const BYTE* ip, U32 const mls)
+                        const BYTE* ip, U32 const mls, U32 const lazySkipping)
 {
     U32* const hashTable  = ms->hashTable;
     const U32 hashLog = cParams->hashLog;
@@ -633,6 +635,9 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
         NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
         hashTable[h] = idx;
         idx++;
+        /* Stop inserting every position when in the lazy skipping mode. */
+        if (lazySkipping)
+            break;
     }
 
     ms->nextToUpdate = target;
@@ -641,7 +646,7 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
 
 U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) {
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
-    return ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, ms->cParams.minMatch);
+    return ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, ms->cParams.minMatch, /* lazySkipping*/ 0);
 }
 
 /* inlining is important to hardwire a hot branch (template emulation) */
@@ -685,7 +690,7 @@ size_t ZSTD_HcFindBestMatch(
     }
 
     /* HC4 match finder */
-    matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls);
+    matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls, ms->lazySkipping);
 
     for ( ; (matchIndex>=lowLimit) & (nbAttempts>0) ; nbAttempts--) {
         size_t currentMl=0;
@@ -758,7 +763,6 @@ size_t ZSTD_HcFindBestMatch(
 * (SIMD) Row-based matchfinder
 ***********************************/
 /* Constants for row-based hash */
-#define ZSTD_ROW_HASH_TAG_OFFSET 16     /* byte offset of hashes in the match state's tagTable from the beginning of a row */
 #define ZSTD_ROW_HASH_TAG_MASK ((1u << ZSTD_ROW_HASH_TAG_BITS) - 1)
 #define ZSTD_ROW_HASH_MAX_ENTRIES 64    /* absolute maximum number of entries per row, for all configurations */
 
@@ -774,39 +778,15 @@ MEM_STATIC U32 ZSTD_VecMask_next(ZSTD_VecMask val) {
     return ZSTD_countTrailingZeros64(val);
 }
 
-/* ZSTD_rotateRight_*():
- * Rotates a bitfield to the right by "count" bits.
- * https://en.wikipedia.org/w/index.php?title=Circular_shift&oldid=991635599#Implementing_circular_shifts
- */
-FORCE_INLINE_TEMPLATE
-U64 ZSTD_rotateRight_U64(U64 const value, U32 count) {
-    assert(count < 64);
-    count &= 0x3F; /* for fickle pattern recognition */
-    return (value >> count) | (U64)(value << ((0U - count) & 0x3F));
-}
-
-FORCE_INLINE_TEMPLATE
-U32 ZSTD_rotateRight_U32(U32 const value, U32 count) {
-    assert(count < 32);
-    count &= 0x1F; /* for fickle pattern recognition */
-    return (value >> count) | (U32)(value << ((0U - count) & 0x1F));
-}
-
-FORCE_INLINE_TEMPLATE
-U16 ZSTD_rotateRight_U16(U16 const value, U32 count) {
-    assert(count < 16);
-    count &= 0x0F; /* for fickle pattern recognition */
-    return (value >> count) | (U16)(value << ((0U - count) & 0x0F));
-}
-
 /* ZSTD_row_nextIndex():
  * Returns the next index to insert at within a tagTable row, and updates the "head"
- * value to reflect the update. Essentially cycles backwards from [0, {entries per row})
+ * value to reflect the update. Essentially cycles backwards from [1, {entries per row})
  */
 FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextIndex(BYTE* const tagRow, U32 const rowMask) {
-  U32 const next = (*tagRow - 1) & rowMask;
-  *tagRow = (BYTE)next;
-  return next;
+    U32 next = (*tagRow-1) & rowMask;
+    next += (next == 0) ? rowMask : 0; /* skip first position */
+    *tagRow = (BYTE)next;
+    return next;
 }
 
 /* ZSTD_isAligned():
@@ -820,7 +800,7 @@ MEM_STATIC int ZSTD_isAligned(void const* ptr, size_t align) {
 /* ZSTD_row_prefetch():
  * Performs prefetching for the hashTable and tagTable at a given row.
  */
-FORCE_INLINE_TEMPLATE void ZSTD_row_prefetch(U32 const* hashTable, U16 const* tagTable, U32 const relRow, U32 const rowLog) {
+FORCE_INLINE_TEMPLATE void ZSTD_row_prefetch(U32 const* hashTable, BYTE const* tagTable, U32 const relRow, U32 const rowLog) {
     PREFETCH_L1(hashTable + relRow);
     if (rowLog >= 5) {
         PREFETCH_L1(hashTable + relRow + 16);
@@ -844,13 +824,13 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const B
                                    U32 idx, const BYTE* const iLimit)
 {
     U32 const* const hashTable = ms->hashTable;
-    U16 const* const tagTable = ms->tagTable;
+    BYTE const* const tagTable = ms->tagTable;
     U32 const hashLog = ms->rowHashLog;
     U32 const maxElemsToPrefetch = (base + idx) > iLimit ? 0 : (U32)(iLimit - (base + idx) + 1);
     U32 const lim = idx + MIN(ZSTD_ROW_HASH_CACHE_SIZE, maxElemsToPrefetch);
 
     for (; idx < lim; ++idx) {
-        U32 const hash = (U32)ZSTD_hashPtr(base + idx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls);
+        U32 const hash = (U32)ZSTD_hashPtrSalted(base + idx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls, ms->hashSalt);
         U32 const row = (hash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog;
         ZSTD_row_prefetch(hashTable, tagTable, row, rowLog);
         ms->hashCache[idx & ZSTD_ROW_HASH_CACHE_MASK] = hash;
@@ -866,11 +846,12 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const B
  * base + idx + ZSTD_ROW_HASH_CACHE_SIZE. Also prefetches the appropriate rows from hashTable and tagTable.
  */
 FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable,
-                                                  U16 const* tagTable, BYTE const* base,
+                                                  BYTE const* tagTable, BYTE const* base,
                                                   U32 idx, U32 const hashLog,
-                                                  U32 const rowLog, U32 const mls)
+                                                  U32 const rowLog, U32 const mls,
+                                                  U64 const hashSalt)
 {
-    U32 const newHash = (U32)ZSTD_hashPtr(base+idx+ZSTD_ROW_HASH_CACHE_SIZE, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls);
+    U32 const newHash = (U32)ZSTD_hashPtrSalted(base+idx+ZSTD_ROW_HASH_CACHE_SIZE, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls, hashSalt);
     U32 const row = (newHash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog;
     ZSTD_row_prefetch(hashTable, tagTable, row, rowLog);
     {   U32 const hash = cache[idx & ZSTD_ROW_HASH_CACHE_MASK];
@@ -888,22 +869,21 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
                                                         U32 const rowMask, U32 const useCache)
 {
     U32* const hashTable = ms->hashTable;
-    U16* const tagTable = ms->tagTable;
+    BYTE* const tagTable = ms->tagTable;
     U32 const hashLog = ms->rowHashLog;
     const BYTE* const base = ms->window.base;
 
     DEBUGLOG(6, "ZSTD_row_update_internalImpl(): updateStartIdx=%u, updateEndIdx=%u", updateStartIdx, updateEndIdx);
     for (; updateStartIdx < updateEndIdx; ++updateStartIdx) {
-        U32 const hash = useCache ? ZSTD_row_nextCachedHash(ms->hashCache, hashTable, tagTable, base, updateStartIdx, hashLog, rowLog, mls)
-                                  : (U32)ZSTD_hashPtr(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls);
+        U32 const hash = useCache ? ZSTD_row_nextCachedHash(ms->hashCache, hashTable, tagTable, base, updateStartIdx, hashLog, rowLog, mls, ms->hashSalt)
+                                  : (U32)ZSTD_hashPtrSalted(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls, ms->hashSalt);
         U32 const relRow = (hash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog;
         U32* const row = hashTable + relRow;
-        BYTE* tagRow = (BYTE*)(tagTable + relRow);  /* Though tagTable is laid out as a table of U16, each tag is only 1 byte.
-                                                       Explicit cast allows us to get exact desired position within each row */
+        BYTE* tagRow = tagTable + relRow;
         U32 const pos = ZSTD_row_nextIndex(tagRow, rowMask);
 
-        assert(hash == ZSTD_hashPtr(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls));
-        ((BYTE*)tagRow)[pos + ZSTD_ROW_HASH_TAG_OFFSET] = hash & ZSTD_ROW_HASH_TAG_MASK;
+        assert(hash == ZSTD_hashPtrSalted(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls, ms->hashSalt));
+        tagRow[pos] = hash & ZSTD_ROW_HASH_TAG_MASK;
         row[pos] = updateStartIdx;
     }
 }
@@ -1059,7 +1039,7 @@ ZSTD_row_getNEONMask(const U32 rowEntries, const BYTE* const src, const BYTE tag
 FORCE_INLINE_TEMPLATE ZSTD_VecMask
 ZSTD_row_getMatchMask(const BYTE* const tagRow, const BYTE tag, const U32 headGrouped, const U32 rowEntries)
 {
-    const BYTE* const src = tagRow + ZSTD_ROW_HASH_TAG_OFFSET;
+    const BYTE* const src = tagRow;
     assert((rowEntries == 16) || (rowEntries == 32) || rowEntries == 64);
     assert(rowEntries <= ZSTD_ROW_HASH_MAX_ENTRIES);
     assert(ZSTD_row_matchMaskGroupWidth(rowEntries) * rowEntries <= sizeof(ZSTD_VecMask) * 8);
@@ -1144,7 +1124,7 @@ size_t ZSTD_RowFindBestMatch(
                         const U32 rowLog)
 {
     U32* const hashTable = ms->hashTable;
-    U16* const tagTable = ms->tagTable;
+    BYTE* const tagTable = ms->tagTable;
     U32* const hashCache = ms->hashCache;
     const U32 hashLog = ms->rowHashLog;
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -1163,8 +1143,10 @@ size_t ZSTD_RowFindBestMatch(
     const U32 rowMask = rowEntries - 1;
     const U32 cappedSearchLog = MIN(cParams->searchLog, rowLog); /* nb of searches is capped at nb entries per row */
     const U32 groupWidth = ZSTD_row_matchMaskGroupWidth(rowEntries);
+    const U64 hashSalt = ms->hashSalt;
     U32 nbAttempts = 1U << cappedSearchLog;
     size_t ml=4-1;
+    U32 hash;
 
     /* DMS/DDS variables that may be referenced laster */
     const ZSTD_matchState_t* const dms = ms->dictMatchState;
@@ -1188,7 +1170,7 @@ size_t ZSTD_RowFindBestMatch(
     if (dictMode == ZSTD_dictMatchState) {
         /* Prefetch DMS rows */
         U32* const dmsHashTable = dms->hashTable;
-        U16* const dmsTagTable = dms->tagTable;
+        BYTE* const dmsTagTable = dms->tagTable;
         U32 const dmsHash = (U32)ZSTD_hashPtr(ip, dms->rowHashLog + ZSTD_ROW_HASH_TAG_BITS, mls);
         U32 const dmsRelRow = (dmsHash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog;
         dmsTag = dmsHash & ZSTD_ROW_HASH_TAG_MASK;
@@ -1198,9 +1180,19 @@ size_t ZSTD_RowFindBestMatch(
     }
 
     /* Update the hashTable and tagTable up to (but not including) ip */
-    ZSTD_row_update_internal(ms, ip, mls, rowLog, rowMask, 1 /* useCache */);
+    if (!ms->lazySkipping) {
+        ZSTD_row_update_internal(ms, ip, mls, rowLog, rowMask, 1 /* useCache */);
+        hash = ZSTD_row_nextCachedHash(hashCache, hashTable, tagTable, base, curr, hashLog, rowLog, mls, hashSalt);
+    } else {
+        /* Stop inserting every position when in the lazy skipping mode.
+         * The hash cache is also not kept up to date in this mode.
+         */
+        hash = (U32)ZSTD_hashPtrSalted(ip, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls, hashSalt);
+        ms->nextToUpdate = curr;
+    }
+    ms->hashSaltEntropy += hash; /* collect salt entropy */
+
     {   /* Get the hash for ip, compute the appropriate row */
-        U32 const hash = ZSTD_row_nextCachedHash(hashCache, hashTable, tagTable, base, curr, hashLog, rowLog, mls);
         U32 const relRow = (hash >> ZSTD_ROW_HASH_TAG_BITS) << rowLog;
         U32 const tag = hash & ZSTD_ROW_HASH_TAG_MASK;
         U32* const row = hashTable + relRow;
@@ -1212,9 +1204,10 @@ size_t ZSTD_RowFindBestMatch(
         ZSTD_VecMask matches = ZSTD_row_getMatchMask(tagRow, (BYTE)tag, headGrouped, rowEntries);
 
         /* Cycle through the matches and prefetch */
-        for (; (matches > 0) && (nbAttempts > 0); --nbAttempts, matches &= (matches - 1)) {
+        for (; (matches > 0) && (nbAttempts > 0); matches &= (matches - 1)) {
             U32 const matchPos = ((headGrouped + ZSTD_VecMask_next(matches)) / groupWidth) & rowMask;
             U32 const matchIndex = row[matchPos];
+            if(matchPos == 0) continue;
             assert(numMatches < rowEntries);
             if (matchIndex < lowLimit)
                 break;
@@ -1224,13 +1217,14 @@ size_t ZSTD_RowFindBestMatch(
                 PREFETCH_L1(dictBase + matchIndex);
             }
             matchBuffer[numMatches++] = matchIndex;
+            --nbAttempts;
         }
 
         /* Speed opt: insert current byte into hashtable too. This allows us to avoid one iteration of the loop
            in ZSTD_row_update_internal() at the next search. */
         {
             U32 const pos = ZSTD_row_nextIndex(tagRow, rowMask);
-            tagRow[pos + ZSTD_ROW_HASH_TAG_OFFSET] = (BYTE)tag;
+            tagRow[pos] = (BYTE)tag;
             row[pos] = ms->nextToUpdate++;
         }
 
@@ -1281,13 +1275,15 @@ size_t ZSTD_RowFindBestMatch(
             size_t currMatch = 0;
             ZSTD_VecMask matches = ZSTD_row_getMatchMask(dmsTagRow, (BYTE)dmsTag, headGrouped, rowEntries);
 
-            for (; (matches > 0) && (nbAttempts > 0); --nbAttempts, matches &= (matches - 1)) {
+            for (; (matches > 0) && (nbAttempts > 0); matches &= (matches - 1)) {
                 U32 const matchPos = ((headGrouped + ZSTD_VecMask_next(matches)) / groupWidth) & rowMask;
                 U32 const matchIndex = dmsRow[matchPos];
+                if(matchPos == 0) continue;
                 if (matchIndex < dmsLowestIndex)
                     break;
                 PREFETCH_L1(dmsBase + matchIndex);
                 matchBuffer[numMatches++] = matchIndex;
+                --nbAttempts;
             }
 
             /* Return the longest match */
@@ -1544,10 +1540,11 @@ ZSTD_compressBlock_lazy_generic(
         assert(offset_2 <= dictAndPrefixLength);
     }
 
+    /* Reset the lazy skipping state */
+    ms->lazySkipping = 0;
+
     if (searchMethod == search_rowHash) {
-        ZSTD_row_fillHashCache(ms, base, rowLog,
-                            MIN(ms->cParams.minMatch, 6 /* mls caps out at 6 */),
-                            ms->nextToUpdate, ilimit);
+        ZSTD_row_fillHashCache(ms, base, rowLog, mls, ms->nextToUpdate, ilimit);
     }
 
     /* Match Loop */
@@ -1591,7 +1588,16 @@ ZSTD_compressBlock_lazy_generic(
         }
 
         if (matchLength < 4) {
-            ip += ((ip-anchor) >> kSearchStrength) + 1;   /* jump faster over incompressible sections */
+            size_t const step = ((size_t)(ip-anchor) >> kSearchStrength) + 1;   /* jump faster over incompressible sections */;
+            ip += step;
+            /* Enter the lazy skipping mode once we are skipping more than 8 bytes at a time.
+             * In this mode we stop inserting every position into our tables, and only insert
+             * positions that we search, which is one in step positions.
+             * The exact cutoff is flexible, I've just chosen a number that is reasonably high,
+             * so we minimize the compression ratio loss in "normal" scenarios. This mode gets
+             * triggered once we've gone 2KB without finding any matches.
+             */
+            ms->lazySkipping = step > kLazySkippingStep;
             continue;
         }
 
@@ -1695,6 +1701,13 @@ _storeSequence:
             ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offBase, matchLength);
             anchor = ip = start + matchLength;
         }
+        if (ms->lazySkipping) {
+            /* We've found a match, disable lazy skipping mode, and refill the hash cache. */
+            if (searchMethod == search_rowHash) {
+                ZSTD_row_fillHashCache(ms, base, rowLog, mls, ms->nextToUpdate, ilimit);
+            }
+            ms->lazySkipping = 0;
+        }
 
         /* check immediate repcode */
         if (isDxS) {
@@ -1912,12 +1925,13 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
 
     DEBUGLOG(5, "ZSTD_compressBlock_lazy_extDict_generic (searchFunc=%u)", (U32)searchMethod);
 
+    /* Reset the lazy skipping state */
+    ms->lazySkipping = 0;
+
     /* init */
     ip += (ip == prefixStart);
     if (searchMethod == search_rowHash) {
-        ZSTD_row_fillHashCache(ms, base, rowLog,
-                               MIN(ms->cParams.minMatch, 6 /* mls caps out at 6 */),
-                               ms->nextToUpdate, ilimit);
+        ZSTD_row_fillHashCache(ms, base, rowLog, mls, ms->nextToUpdate, ilimit);
     }
 
     /* Match Loop */
@@ -1955,7 +1969,16 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
         }
 
         if (matchLength < 4) {
-            ip += ((ip-anchor) >> kSearchStrength) + 1;   /* jump faster over incompressible sections */
+            size_t const step = ((size_t)(ip-anchor) >> kSearchStrength);
+            ip += step + 1;   /* jump faster over incompressible sections */
+            /* Enter the lazy skipping mode once we are skipping more than 8 bytes at a time.
+             * In this mode we stop inserting every position into our tables, and only insert
+             * positions that we search, which is one in step positions.
+             * The exact cutoff is flexible, I've just chosen a number that is reasonably high,
+             * so we minimize the compression ratio loss in "normal" scenarios. This mode gets
+             * triggered once we've gone 2KB without finding any matches.
+             */
+            ms->lazySkipping = step > kLazySkippingStep;
             continue;
         }
 
@@ -2041,6 +2064,13 @@ _storeSequence:
             ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offBase, matchLength);
             anchor = ip = start + matchLength;
         }
+        if (ms->lazySkipping) {
+            /* We've found a match, disable lazy skipping mode, and refill the hash cache. */
+            if (searchMethod == search_rowHash) {
+                ZSTD_row_fillHashCache(ms, base, rowLog, mls, ms->nextToUpdate, ilimit);
+            }
+            ms->lazySkipping = 0;
+        }
 
         /* check immediate repcode */
         while (ip <= ilimit) {
diff --git a/contrib/libs/zstd/lib/compress/zstd_opt.c b/contrib/libs/zstd/lib/compress/zstd_opt.c
index fdd7f9d8b5..f02a760946 100644
--- a/contrib/libs/zstd/lib/compress/zstd_opt.c
+++ b/contrib/libs/zstd/lib/compress/zstd_opt.c
@@ -1086,6 +1086,8 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
     ZSTD_optimal_t lastSequence;
     ZSTD_optLdm_t optLdm;
 
+    ZSTD_memset(&lastSequence, 0, sizeof(ZSTD_optimal_t));
+
     optLdm.seqStore = ms->ldmSeqStore ? *ms->ldmSeqStore : kNullRawSeqStore;
     optLdm.endPosInBlock = optLdm.startPosInBlock = optLdm.offset = 0;
     ZSTD_opt_getNextMatchAndUpdateSeqStore(&optLdm, (U32)(ip-istart), (U32)(iend-ip));
diff --git a/contrib/libs/zstd/lib/compress/zstdmt_compress.c b/contrib/libs/zstd/lib/compress/zstdmt_compress.c
index 7a2c71720a..6786075569 100644
--- a/contrib/libs/zstd/lib/compress/zstdmt_compress.c
+++ b/contrib/libs/zstd/lib/compress/zstdmt_compress.c
@@ -20,6 +20,7 @@
 
 
 /* ======   Dependencies   ====== */
+#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
 #include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memset, INT_MAX, UINT_MAX */
 #include "../common/mem.h"         /* MEM_STATIC */
 #include "../common/pool.h"        /* threadpool */
@@ -719,7 +720,7 @@ static void ZSTDMT_compressionJob(void* jobDescription)
     ZSTDMT_serialState_update(job->serial, cctx, rawSeqStore, job->src, job->jobID);
 
     if (!job->firstJob) {  /* flush and overwrite frame header when it's not first job */
-        size_t const hSize = ZSTD_compressContinue(cctx, dstBuff.start, dstBuff.capacity, job->src.start, 0);
+        size_t const hSize = ZSTD_compressContinue_public(cctx, dstBuff.start, dstBuff.capacity, job->src.start, 0);
         if (ZSTD_isError(hSize)) JOB_ERROR(hSize);
         DEBUGLOG(5, "ZSTDMT_compressionJob: flush and overwrite %u bytes of frame header (not first job)", (U32)hSize);
         ZSTD_invalidateRepCodes(cctx);
@@ -737,7 +738,7 @@ static void ZSTDMT_compressionJob(void* jobDescription)
         DEBUGLOG(5, "ZSTDMT_compressionJob: compress %u bytes in %i blocks", (U32)job->src.size, nbChunks);
         assert(job->cSize == 0);
         for (chunkNb = 1; chunkNb < nbChunks; chunkNb++) {
-            size_t const cSize = ZSTD_compressContinue(cctx, op, oend-op, ip, chunkSize);
+            size_t const cSize = ZSTD_compressContinue_public(cctx, op, oend-op, ip, chunkSize);
             if (ZSTD_isError(cSize)) JOB_ERROR(cSize);
             ip += chunkSize;
             op += cSize; assert(op < oend);
@@ -757,8 +758,8 @@ static void ZSTDMT_compressionJob(void* jobDescription)
             size_t const lastBlockSize1 = job->src.size & (chunkSize-1);
             size_t const lastBlockSize = ((lastBlockSize1==0) & (job->src.size>=chunkSize)) ? chunkSize : lastBlockSize1;
             size_t const cSize = (job->lastJob) ?
-                 ZSTD_compressEnd     (cctx, op, oend-op, ip, lastBlockSize) :
-                 ZSTD_compressContinue(cctx, op, oend-op, ip, lastBlockSize);
+                 ZSTD_compressEnd_public(cctx, op, oend-op, ip, lastBlockSize) :
+                 ZSTD_compressContinue_public(cctx, op, oend-op, ip, lastBlockSize);
             if (ZSTD_isError(cSize)) JOB_ERROR(cSize);
             lastCBlockSize = cSize;
     }   }
diff --git a/contrib/libs/zstd/lib/decompress/huf_decompress.c b/contrib/libs/zstd/lib/decompress/huf_decompress.c
index c2d1f633a4..5b217ac586 100644
--- a/contrib/libs/zstd/lib/decompress/huf_decompress.c
+++ b/contrib/libs/zstd/lib/decompress/huf_decompress.c
@@ -696,7 +696,7 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
 
     /* Copy the arguments to local variables */
     ZSTD_memcpy(&bits, &args->bits, sizeof(bits));
-    ZSTD_memcpy(&ip, &args->ip, sizeof(ip));
+    ZSTD_memcpy((void*)(&ip), &args->ip, sizeof(ip));
     ZSTD_memcpy(&op, &args->op, sizeof(op));
 
     assert(MEM_isLittleEndian());
@@ -779,7 +779,7 @@ _out:
 
     /* Save the final values of each of the state variables back to args. */
     ZSTD_memcpy(&args->bits, &bits, sizeof(bits));
-    ZSTD_memcpy(&args->ip, &ip, sizeof(ip));
+    ZSTD_memcpy((void*)(&args->ip), &ip, sizeof(ip));
     ZSTD_memcpy(&args->op, &op, sizeof(op));
 }
 
@@ -1476,7 +1476,7 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
 
     /* Copy the arguments to local registers. */
     ZSTD_memcpy(&bits, &args->bits, sizeof(bits));
-    ZSTD_memcpy(&ip, &args->ip, sizeof(ip));
+    ZSTD_memcpy((void*)(&ip), &args->ip, sizeof(ip));
     ZSTD_memcpy(&op, &args->op, sizeof(op));
 
     oend[0] = op[1];
@@ -1599,7 +1599,7 @@ _out:
 
     /* Save the final values of each of the state variables back to args. */
     ZSTD_memcpy(&args->bits, &bits, sizeof(bits));
-    ZSTD_memcpy(&args->ip, &ip, sizeof(ip));
+    ZSTD_memcpy((void*)(&args->ip), &ip, sizeof(ip));
     ZSTD_memcpy(&args->op, &op, sizeof(op));
 }
 
diff --git a/contrib/libs/zstd/lib/decompress/zstd_ddict.c b/contrib/libs/zstd/lib/decompress/zstd_ddict.c
index ad5c34a7fc..309ec0d036 100644
--- a/contrib/libs/zstd/lib/decompress/zstd_ddict.c
+++ b/contrib/libs/zstd/lib/decompress/zstd_ddict.c
@@ -14,6 +14,7 @@
 /*-*******************************************************
 *  Dependencies
 *********************************************************/
+#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customFree */
 #include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
 #include "../common/cpu.h"         /* bmi2 */
 #include "../common/mem.h"         /* low level memory routines */
diff --git a/contrib/libs/zstd/lib/decompress/zstd_decompress.c b/contrib/libs/zstd/lib/decompress/zstd_decompress.c
index 093a32716d..02e254386d 100644
--- a/contrib/libs/zstd/lib/decompress/zstd_decompress.c
+++ b/contrib/libs/zstd/lib/decompress/zstd_decompress.c
@@ -55,6 +55,7 @@
 /*-*******************************************************
 *  Dependencies
 *********************************************************/
+#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
 #include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
 #include "../common/mem.h"         /* low level memory routines */
 #define FSE_STATIC_LINKING_ONLY
@@ -588,49 +589,52 @@ static size_t readSkippableFrameSize(void const* src, size_t srcSize)
     sizeU32 = MEM_readLE32((BYTE const*)src + ZSTD_FRAMEIDSIZE);
     RETURN_ERROR_IF((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32,
                     frameParameter_unsupported, "");
-    {
-        size_t const skippableSize = skippableHeaderSize + sizeU32;
+    {   size_t const skippableSize = skippableHeaderSize + sizeU32;
         RETURN_ERROR_IF(skippableSize > srcSize, srcSize_wrong, "");
         return skippableSize;
     }
 }
 
 /*! ZSTD_readSkippableFrame() :
- * Retrieves a zstd skippable frame containing data given by src, and writes it to dst buffer.
+ * Retrieves content of a skippable frame, and writes it to dst buffer.
  *
  * The parameter magicVariant will receive the magicVariant that was supplied when the frame was written,
  * i.e. magicNumber - ZSTD_MAGIC_SKIPPABLE_START.  This can be NULL if the caller is not interested
  * in the magicVariant.
  *
- * Returns an error if destination buffer is not large enough, or if the frame is not skippable.
+ * Returns an error if destination buffer is not large enough, or if this is not a valid skippable frame.
  *
  * @return : number of bytes written or a ZSTD error.
  */
-ZSTDLIB_API size_t ZSTD_readSkippableFrame(void* dst, size_t dstCapacity, unsigned* magicVariant,
-                                            const void* src, size_t srcSize)
+size_t ZSTD_readSkippableFrame(void* dst, size_t dstCapacity,
+                               unsigned* magicVariant,  /* optional, can be NULL */
+                         const void* src, size_t srcSize)
 {
-    U32 const magicNumber = MEM_readLE32(src);
-    size_t skippableFrameSize = readSkippableFrameSize(src, srcSize);
-    size_t skippableContentSize = skippableFrameSize - ZSTD_SKIPPABLEHEADERSIZE;
-
-    /* check input validity */
-    RETURN_ERROR_IF(!ZSTD_isSkippableFrame(src, srcSize), frameParameter_unsupported, "");
-    RETURN_ERROR_IF(skippableFrameSize < ZSTD_SKIPPABLEHEADERSIZE || skippableFrameSize > srcSize, srcSize_wrong, "");
-    RETURN_ERROR_IF(skippableContentSize > dstCapacity, dstSize_tooSmall, "");
+    RETURN_ERROR_IF(srcSize < ZSTD_SKIPPABLEHEADERSIZE, srcSize_wrong, "");
 
-    /* deliver payload */
-    if (skippableContentSize > 0  && dst != NULL)
-        ZSTD_memcpy(dst, (const BYTE *)src + ZSTD_SKIPPABLEHEADERSIZE, skippableContentSize);
-    if (magicVariant != NULL)
-        *magicVariant = magicNumber - ZSTD_MAGIC_SKIPPABLE_START;
-    return skippableContentSize;
+    {   U32 const magicNumber = MEM_readLE32(src);
+        size_t skippableFrameSize = readSkippableFrameSize(src, srcSize);
+        size_t skippableContentSize = skippableFrameSize - ZSTD_SKIPPABLEHEADERSIZE;
+
+        /* check input validity */
+        RETURN_ERROR_IF(!ZSTD_isSkippableFrame(src, srcSize), frameParameter_unsupported, "");
+        RETURN_ERROR_IF(skippableFrameSize < ZSTD_SKIPPABLEHEADERSIZE || skippableFrameSize > srcSize, srcSize_wrong, "");
+        RETURN_ERROR_IF(skippableContentSize > dstCapacity, dstSize_tooSmall, "");
+
+        /* deliver payload */
+        if (skippableContentSize > 0  && dst != NULL)
+            ZSTD_memcpy(dst, (const BYTE *)src + ZSTD_SKIPPABLEHEADERSIZE, skippableContentSize);
+        if (magicVariant != NULL)
+            *magicVariant = magicNumber - ZSTD_MAGIC_SKIPPABLE_START;
+        return skippableContentSize;
+    }
 }
 
 /** ZSTD_findDecompressedSize() :
- *  compatible with legacy mode
  *  `srcSize` must be the exact length of some number of ZSTD compressed and/or
  *      skippable frames
- *  @return : decompressed size of the frames contained */
+ *  note: compatible with legacy mode
+ * @return : decompressed size of the frames contained */
 unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
 {
     unsigned long long totalDstSize = 0;
@@ -640,9 +644,7 @@ unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
 
         if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
             size_t const skippableSize = readSkippableFrameSize(src, srcSize);
-            if (ZSTD_isError(skippableSize)) {
-                return ZSTD_CONTENTSIZE_ERROR;
-            }
+            if (ZSTD_isError(skippableSize)) return ZSTD_CONTENTSIZE_ERROR;
             assert(skippableSize <= srcSize);
 
             src = (const BYTE *)src + skippableSize;
@@ -650,17 +652,17 @@ unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
             continue;
         }
 
-        {   unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
-            if (ret >= ZSTD_CONTENTSIZE_ERROR) return ret;
+        {   unsigned long long const fcs = ZSTD_getFrameContentSize(src, srcSize);
+            if (fcs >= ZSTD_CONTENTSIZE_ERROR) return fcs;
 
-            /* check for overflow */
-            if (totalDstSize + ret < totalDstSize) return ZSTD_CONTENTSIZE_ERROR;
-            totalDstSize += ret;
+            if (totalDstSize + fcs < totalDstSize)
+                return ZSTD_CONTENTSIZE_ERROR; /* check for overflow */
+            totalDstSize += fcs;
         }
+        /* skip to next frame */
         {   size_t const frameSrcSize = ZSTD_findFrameCompressedSize(src, srcSize);
-            if (ZSTD_isError(frameSrcSize)) {
-                return ZSTD_CONTENTSIZE_ERROR;
-            }
+            if (ZSTD_isError(frameSrcSize)) return ZSTD_CONTENTSIZE_ERROR;
+            assert(frameSrcSize <= srcSize);
 
             src = (const BYTE *)src + frameSrcSize;
             srcSize -= frameSrcSize;
@@ -1090,17 +1092,18 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
         }
 #endif
 
-        {   U32 const magicNumber = MEM_readLE32(src);
-            DEBUGLOG(4, "reading magic number %08X (expecting %08X)",
-                        (unsigned)magicNumber, ZSTD_MAGICNUMBER);
+        if (srcSize >= 4) {
+            U32 const magicNumber = MEM_readLE32(src);
+            DEBUGLOG(5, "reading magic number %08X", (unsigned)magicNumber);
             if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
+                /* skippable frame detected : skip it */
                 size_t const skippableSize = readSkippableFrameSize(src, srcSize);
-                FORWARD_IF_ERROR(skippableSize, "readSkippableFrameSize failed");
+                FORWARD_IF_ERROR(skippableSize, "invalid skippable frame");
                 assert(skippableSize <= srcSize);
 
                 src = (const BYTE *)src + skippableSize;
                 srcSize -= skippableSize;
-                continue;
+                continue; /* check next frame */
         }   }
 
         if (ddict) {
diff --git a/contrib/libs/zstd/lib/decompress/zstd_decompress_block.c b/contrib/libs/zstd/lib/decompress/zstd_decompress_block.c
index 0a06a021e1..09896a931e 100644
--- a/contrib/libs/zstd/lib/decompress/zstd_decompress_block.c
+++ b/contrib/libs/zstd/lib/decompress/zstd_decompress_block.c
@@ -1232,11 +1232,7 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
 
         /* sequence */
         {   size_t offset;
-    #if defined(__clang__)
-            if (LIKELY(ofBits > 1)) {
-    #else
             if (ofBits > 1) {
-    #endif
                 ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
                 ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
                 ZSTD_STATIC_ASSERT(STREAM_ACCUMULATOR_MIN_32 > LONG_OFFSETS_MAX_EXTRA_BITS_32);
@@ -1273,11 +1269,7 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
             seq.offset = offset;
         }
 
-    #if defined(__clang__)
-        if (UNLIKELY(mlBits > 0))
-    #else
         if (mlBits > 0)
-    #endif
             seq.matchLength += BIT_readBitsFast(&seqState->DStream, mlBits/*>0*/);
 
         if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
@@ -1287,11 +1279,7 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
         /* Ensure there are enough bits to read the rest of data in 64-bit mode. */
         ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
 
-    #if defined(__clang__)
-        if (UNLIKELY(llBits > 0))
-    #else
         if (llBits > 0)
-    #endif
             seq.litLength += BIT_readBitsFast(&seqState->DStream, llBits/*>0*/);
 
         if (MEM_32bits())
@@ -1987,7 +1975,7 @@ ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,
 
 
 /**
- * @returns The total size of the history referencable by zstd, including
+ * @returns The total size of the history referenceable by zstd, including
  * both the prefix and the extDict. At @p op any offset larger than this
  * is invalid.
  */
@@ -2124,7 +2112,9 @@ ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
         ip += seqHSize;
         srcSize -= seqHSize;
 
-        RETURN_ERROR_IF(dst == NULL && nbSeq > 0, dstSize_tooSmall, "NULL not handled");
+        RETURN_ERROR_IF((dst == NULL || dstCapacity == 0) && nbSeq > 0, dstSize_tooSmall, "NULL not handled");
+        RETURN_ERROR_IF(MEM_64bits() && sizeof(size_t) == sizeof(void*) && (size_t)(-1) - (size_t)dst < (size_t)(1 << 20), dstSize_tooSmall,
+                "invalid dst");
 
         /* If we could potentially have long offsets, or we might want to use the prefetch decoder,
          * compute information about the share of long offsets, and the maximum nbAdditionalBits.
@@ -2181,9 +2171,9 @@ void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst, size_t dstSize)
 }
 
 
-size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,
-                            void* dst, size_t dstCapacity,
-                      const void* src, size_t srcSize)
+size_t ZSTD_decompressBlock_deprecated(ZSTD_DCtx* dctx,
+                                       void* dst, size_t dstCapacity,
+                                 const void* src, size_t srcSize)
 {
     size_t dSize;
     ZSTD_checkContinuity(dctx, dst, dstCapacity);
@@ -2191,3 +2181,12 @@ size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,
     dctx->previousDstEnd = (char*)dst + dSize;
     return dSize;
 }
+
+
+/* NOTE: Must just wrap ZSTD_decompressBlock_deprecated() */
+size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity,
+                      const void* src, size_t srcSize)
+{
+    return ZSTD_decompressBlock_deprecated(dctx, dst, dstCapacity, src, srcSize);
+}
diff --git a/contrib/libs/zstd/lib/decompress/zstd_decompress_block.h b/contrib/libs/zstd/lib/decompress/zstd_decompress_block.h
index 67791dbc3a..9d1318882d 100644
--- a/contrib/libs/zstd/lib/decompress/zstd_decompress_block.h
+++ b/contrib/libs/zstd/lib/decompress/zstd_decompress_block.h
@@ -64,5 +64,10 @@ void ZSTD_buildFSETable(ZSTD_seqSymbol* dt,
                    unsigned tableLog, void* wksp, size_t wkspSize,
                    int bmi2);
 
+/* Internal definition of ZSTD_decompressBlock() to avoid deprecation warnings. */
+size_t ZSTD_decompressBlock_deprecated(ZSTD_DCtx* dctx,
+                            void* dst, size_t dstCapacity,
+                      const void* src, size_t srcSize);
+
 
 #endif /* ZSTD_DEC_BLOCK_H */
diff --git a/contrib/libs/zstd/lib/dictBuilder/zdict.c b/contrib/libs/zstd/lib/dictBuilder/zdict.c
index 140a0f909c..604223c772 100644
--- a/contrib/libs/zstd/lib/dictBuilder/zdict.c
+++ b/contrib/libs/zstd/lib/dictBuilder/zdict.c
@@ -566,11 +566,11 @@ static void ZDICT_countEStats(EStats_ress_t esr, const ZSTD_parameters* params,
     size_t cSize;
 
     if (srcSize > blockSizeMax) srcSize = blockSizeMax;   /* protection vs large samples */
-    {   size_t const errorCode = ZSTD_compressBegin_usingCDict(esr.zc, esr.dict);
+    {   size_t const errorCode = ZSTD_compressBegin_usingCDict_deprecated(esr.zc, esr.dict);
         if (ZSTD_isError(errorCode)) { DISPLAYLEVEL(1, "warning : ZSTD_compressBegin_usingCDict failed \n"); return; }
 
     }
-    cSize = ZSTD_compressBlock(esr.zc, esr.workPlace, ZSTD_BLOCKSIZE_MAX, src, srcSize);
+    cSize = ZSTD_compressBlock_deprecated(esr.zc, esr.workPlace, ZSTD_BLOCKSIZE_MAX, src, srcSize);
     if (ZSTD_isError(cSize)) { DISPLAYLEVEL(3, "warning : could not compress sample size %u \n", (unsigned)srcSize); return; }
 
     if (cSize) {  /* if == 0; block is not compressible */
diff --git a/contrib/libs/zstd/lib/zstd.h b/contrib/libs/zstd/lib/zstd.h
index 95aac07370..e5c3f8b68b 100644
--- a/contrib/libs/zstd/lib/zstd.h
+++ b/contrib/libs/zstd/lib/zstd.h
@@ -106,7 +106,7 @@ extern "C" {
 /*------   Version   ------*/
 #define ZSTD_VERSION_MAJOR    1
 #define ZSTD_VERSION_MINOR    5
-#define ZSTD_VERSION_RELEASE  4
+#define ZSTD_VERSION_RELEASE  5
 #define ZSTD_VERSION_NUMBER  (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)
 
 /*! ZSTD_versionNumber() :
@@ -148,7 +148,8 @@ ZSTDLIB_API const char* ZSTD_versionString(void);
 ***************************************/
 /*! ZSTD_compress() :
  *  Compresses `src` content as a single zstd compressed frame into already allocated `dst`.
- *  Hint : compression runs faster if `dstCapacity` >=  `ZSTD_compressBound(srcSize)`.
+ *  NOTE: Providing `dstCapacity >= ZSTD_compressBound(srcSize)` guarantees that zstd will have
+ *        enough space to successfully compress the data.
  *  @return : compressed size written into `dst` (<= `dstCapacity),
  *            or an error code if it fails (which can be tested using ZSTD_isError()). */
 ZSTDLIB_API size_t ZSTD_compress( void* dst, size_t dstCapacity,
@@ -578,7 +579,8 @@ ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
  *  Should cctx hold data from a previously unfinished frame, everything about it is forgotten.
  *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
  *  - The function is always blocking, returns when compression is completed.
- *  Hint : compression runs faster if `dstCapacity` >=  `ZSTD_compressBound(srcSize)`.
+ *  NOTE: Providing `dstCapacity >= ZSTD_compressBound(srcSize)` guarantees that zstd will have
+ *        enough space to successfully compress the data, though it is possible it fails for other reasons.
  * @return : compressed size written into `dst` (<= `dstCapacity),
  *           or an error code if it fails (which can be tested using ZSTD_isError()).
  */
@@ -1018,9 +1020,11 @@ ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
  * Advanced dictionary and prefix API (Requires v1.4.0+)
  *
  * This API allows dictionaries to be used with ZSTD_compress2(),
- * ZSTD_compressStream2(), and ZSTD_decompressDCtx(). Dictionaries are sticky, and
- * only reset with the context is reset with ZSTD_reset_parameters or
- * ZSTD_reset_session_and_parameters. Prefixes are single-use.
+ * ZSTD_compressStream2(), and ZSTD_decompressDCtx().
+ * Dictionaries are sticky, they remain valid when same context is re-used,
+ * they only reset when the context is reset
+ * with ZSTD_reset_parameters or ZSTD_reset_session_and_parameters.
+ * In contrast, Prefixes are single-use.
  ******************************************************************************/
 
 
@@ -1041,7 +1045,11 @@ ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
  *           Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead.
  *           In such a case, dictionary buffer must outlive its users.
  *  Note 4 : Use ZSTD_CCtx_loadDictionary_advanced()
- *           to precisely select how dictionary content must be interpreted. */
+ *           to precisely select how dictionary content must be interpreted.
+ *  Note 5 : This method does not benefit from LDM (long distance mode).
+ *           If you want to employ LDM on some large dictionary content,
+ *           prefer employing ZSTD_CCtx_refPrefix() described below.
+ */
 ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
 
 /*! ZSTD_CCtx_refCDict() : Requires v1.4.0+
@@ -1064,6 +1072,7 @@ ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
  *  Decompression will need same prefix to properly regenerate data.
  *  Compressing with a prefix is similar in outcome as performing a diff and compressing it,
  *  but performs much faster, especially during decompression (compression speed is tunable with compression level).
+ *  This method is compatible with LDM (long distance mode).
  * @result : 0, or an error code (which can be tested with ZSTD_isError()).
  *  Special: Adding any prefix (including NULL) invalidates any previous prefix or dictionary
  *  Note 1 : Prefix buffer is referenced. It **must** outlive compression.
@@ -1387,7 +1396,7 @@ typedef enum {
 } ZSTD_paramSwitch_e;
 
 /***************************************
-*  Frame size functions
+*  Frame header and size functions
 ***************************************/
 
 /*! ZSTD_findDecompressedSize() :
@@ -1434,6 +1443,30 @@ ZSTDLIB_STATIC_API unsigned long long ZSTD_decompressBound(const void* src, size
  *           or an error code (if srcSize is too small) */
 ZSTDLIB_STATIC_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
 
+typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
+typedef struct {
+    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;
+    unsigned _reserved1;
+    unsigned _reserved2;
+} ZSTD_frameHeader;
+
+/*! ZSTD_getFrameHeader() :
+ *  decode Frame Header, or requires larger `srcSize`.
+ * @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() */
+ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize);   /**< doesn't consume input */
+/*! ZSTD_getFrameHeader_advanced() :
+ *  same as ZSTD_getFrameHeader(),
+ *  with added capability to select a format (like ZSTD_f_zstd1_magicless) */
+ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);
+
 /*! ZSTD_decompressionMargin() :
  * Zstd supports in-place decompression, where the input and output buffers overlap.
  * In this case, the output buffer must be at least (Margin + Output_Size) bytes large,
@@ -1803,12 +1836,26 @@ ZSTDLIB_STATIC_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
 ZSTDLIB_STATIC_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
 
 /*! ZSTD_CCtx_setCParams() :
- *  Set all parameters provided within @cparams into the working @cctx.
+ *  Set all parameters provided within @p cparams into the working @p cctx.
  *  Note : if modifying parameters during compression (MT mode only),
  *         note that changes to the .windowLog parameter will be ignored.
- * @return 0 on success, or an error code (can be checked with ZSTD_isError()) */
+ * @return 0 on success, or an error code (can be checked with ZSTD_isError()).
+ *         On failure, no parameters are updated.
+ */
 ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setCParams(ZSTD_CCtx* cctx, ZSTD_compressionParameters cparams);
 
+/*! ZSTD_CCtx_setFParams() :
+ *  Set all parameters provided within @p fparams into the working @p cctx.
+ * @return 0 on success, or an error code (can be checked with ZSTD_isError()).
+ */
+ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setFParams(ZSTD_CCtx* cctx, ZSTD_frameParameters fparams);
+
+/*! ZSTD_CCtx_setParams() :
+ *  Set all parameters provided within @p params into the working @p cctx.
+ * @return 0 on success, or an error code (can be checked with ZSTD_isError()).
+ */
+ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setParams(ZSTD_CCtx* cctx, ZSTD_parameters params);
+
 /*! ZSTD_compress_advanced() :
  *  Note : this function is now DEPRECATED.
  *         It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_setParameter() and other parameter setters.
@@ -2134,7 +2181,7 @@ ZSTDLIB_STATIC_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const vo
  * This parameter can be used to set an upper bound on the blocksize
  * that overrides the default ZSTD_BLOCKSIZE_MAX. It cannot be used to set upper
  * bounds greater than ZSTD_BLOCKSIZE_MAX or bounds lower than 1KB (will make
- * compressBound() innacurate). Only currently meant to be used for testing.
+ * compressBound() inaccurate). Only currently meant to be used for testing.
  *
  */
 #define ZSTD_c_maxBlockSize ZSTD_c_experimentalParam18
@@ -2452,12 +2499,9 @@ size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs,
                            int compressionLevel);
 
 /*! ZSTD_initCStream_advanced() :
- * This function is DEPRECATED, and is approximately equivalent to:
+ * This function is DEPRECATED, and is equivalent to:
  *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
- *     // Pseudocode: Set each zstd parameter and leave the rest as-is.
- *     for ((param, value) : params) {
- *         ZSTD_CCtx_setParameter(zcs, param, value);
- *     }
+ *     ZSTD_CCtx_setParams(zcs, params);
  *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
  *     ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
  *
@@ -2486,12 +2530,9 @@ ZSTDLIB_STATIC_API
 size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);
 
 /*! ZSTD_initCStream_usingCDict_advanced() :
- *   This function is DEPRECATED, and is approximately equivalent to:
+ *   This function is DEPRECATED, and is equivalent to:
  *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
- *     // Pseudocode: Set each zstd frame parameter and leave the rest as-is.
- *     for ((fParam, value) : fParams) {
- *         ZSTD_CCtx_setParameter(zcs, fParam, value);
- *     }
+ *     ZSTD_CCtx_setFParams(zcs, fParams);
  *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
  *     ZSTD_CCtx_refCDict(zcs, cdict);
  *
@@ -2598,12 +2639,180 @@ ZSTD_DEPRECATED("use ZSTD_DCtx_reset, see zstd.h for detailed instructions")
 ZSTDLIB_STATIC_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
 
 
+/* ********************* BLOCK-LEVEL SEQUENCE PRODUCER API *********************
+ *
+ * *** OVERVIEW ***
+ * The Block-Level Sequence Producer API allows users to provide their own custom
+ * sequence producer which libzstd invokes to process each block. The produced list
+ * of sequences (literals and matches) is then post-processed by libzstd to produce
+ * valid compressed blocks.
+ *
+ * This block-level offload API is a more granular complement of the existing
+ * frame-level offload API compressSequences() (introduced in v1.5.1). It offers
+ * an easier migration story for applications already integrated with libzstd: the
+ * user application continues to invoke the same compression functions
+ * ZSTD_compress2() or ZSTD_compressStream2() as usual, and transparently benefits
+ * from the specific advantages of the external sequence producer. For example,
+ * the sequence producer could be tuned to take advantage of known characteristics
+ * of the input, to offer better speed / ratio, or could leverage hardware
+ * acceleration not available within libzstd itself.
+ *
+ * See contrib/externalSequenceProducer for an example program employing the
+ * Block-Level Sequence Producer API.
+ *
+ * *** USAGE ***
+ * The user is responsible for implementing a function of type
+ * ZSTD_sequenceProducer_F. For each block, zstd will pass the following
+ * arguments to the user-provided function:
+ *
+ *   - sequenceProducerState: a pointer to a user-managed state for the sequence
+ *     producer.
+ *
+ *   - outSeqs, outSeqsCapacity: an output buffer for the sequence producer.
+ *     outSeqsCapacity is guaranteed >= ZSTD_sequenceBound(srcSize). The memory
+ *     backing outSeqs is managed by the CCtx.
+ *
+ *   - src, srcSize: an input buffer for the sequence producer to parse.
+ *     srcSize is guaranteed to be <= ZSTD_BLOCKSIZE_MAX.
+ *
+ *   - dict, dictSize: a history buffer, which may be empty, which the sequence
+ *     producer may reference as it parses the src buffer. Currently, zstd will
+ *     always pass dictSize == 0 into external sequence producers, but this will
+ *     change in the future.
+ *
+ *   - compressionLevel: a signed integer representing the zstd compression level
+ *     set by the user for the current operation. The sequence producer may choose
+ *     to use this information to change its compression strategy and speed/ratio
+ *     tradeoff. Note: the compression level does not reflect zstd parameters set
+ *     through the advanced API.
+ *
+ *   - windowSize: a size_t representing the maximum allowed offset for external
+ *     sequences. Note that sequence offsets are sometimes allowed to exceed the
+ *     windowSize if a dictionary is present, see doc/zstd_compression_format.md
+ *     for details.
+ *
+ * The user-provided function shall return a size_t representing the number of
+ * sequences written to outSeqs. This return value will be treated as an error
+ * code if it is greater than outSeqsCapacity. The return value must be non-zero
+ * if srcSize is non-zero. The ZSTD_SEQUENCE_PRODUCER_ERROR macro is provided
+ * for convenience, but any value greater than outSeqsCapacity will be treated as
+ * an error code.
+ *
+ * If the user-provided function does not return an error code, the sequences
+ * written to outSeqs must be a valid parse of the src buffer. Data corruption may
+ * occur if the parse is not valid. A parse is defined to be valid if the
+ * following conditions hold:
+ *   - The sum of matchLengths and literalLengths must equal srcSize.
+ *   - All sequences in the parse, except for the final sequence, must have
+ *     matchLength >= ZSTD_MINMATCH_MIN. The final sequence must have
+ *     matchLength >= ZSTD_MINMATCH_MIN or matchLength == 0.
+ *   - All offsets must respect the windowSize parameter as specified in
+ *     doc/zstd_compression_format.md.
+ *   - If the final sequence has matchLength == 0, it must also have offset == 0.
+ *
+ * zstd will only validate these conditions (and fail compression if they do not
+ * hold) if the ZSTD_c_validateSequences cParam is enabled. Note that sequence
+ * validation has a performance cost.
+ *
+ * If the user-provided function returns an error, zstd will either fall back
+ * to an internal sequence producer or fail the compression operation. The user can
+ * choose between the two behaviors by setting the ZSTD_c_enableSeqProducerFallback
+ * cParam. Fallback compression will follow any other cParam settings, such as
+ * compression level, the same as in a normal compression operation.
+ *
+ * The user shall instruct zstd to use a particular ZSTD_sequenceProducer_F
+ * function by calling
+ *         ZSTD_registerSequenceProducer(cctx,
+ *                                       sequenceProducerState,
+ *                                       sequenceProducer)
+ * This setting will persist until the next parameter reset of the CCtx.
+ *
+ * The sequenceProducerState must be initialized by the user before calling
+ * ZSTD_registerSequenceProducer(). The user is responsible for destroying the
+ * sequenceProducerState.
+ *
+ * *** LIMITATIONS ***
+ * This API is compatible with all zstd compression APIs which respect advanced parameters.
+ * However, there are three limitations:
+ *
+ * First, the ZSTD_c_enableLongDistanceMatching cParam is not currently supported.
+ * COMPRESSION WILL FAIL if it is enabled and the user tries to compress with a block-level
+ * external sequence producer.
+ *   - Note that ZSTD_c_enableLongDistanceMatching is auto-enabled by default in some
+ *     cases (see its documentation for details). Users must explicitly set
+ *     ZSTD_c_enableLongDistanceMatching to ZSTD_ps_disable in such cases if an external
+ *     sequence producer is registered.
+ *   - As of this writing, ZSTD_c_enableLongDistanceMatching is disabled by default
+ *     whenever ZSTD_c_windowLog < 128MB, but that's subject to change. Users should
+ *     check the docs on ZSTD_c_enableLongDistanceMatching whenever the Block-Level Sequence
+ *     Producer API is used in conjunction with advanced settings (like ZSTD_c_windowLog).
+ *
+ * Second, history buffers are not currently supported. Concretely, zstd will always pass
+ * dictSize == 0 to the external sequence producer (for now). This has two implications:
+ *   - Dictionaries are not currently supported. Compression will *not* fail if the user
+ *     references a dictionary, but the dictionary won't have any effect.
+ *   - Stream history is not currently supported. All advanced compression APIs, including
+ *     streaming APIs, work with external sequence producers, but each block is treated as
+ *     an independent chunk without history from previous blocks.
+ *
+ * Third, multi-threading within a single compression is not currently supported. In other words,
+ * COMPRESSION WILL FAIL if ZSTD_c_nbWorkers > 0 and an external sequence producer is registered.
+ * Multi-threading across compressions is fine: simply create one CCtx per thread.
+ *
+ * Long-term, we plan to overcome all three limitations. There is no technical blocker to
+ * overcoming them. It is purely a question of engineering effort.
+ */
+
+#define ZSTD_SEQUENCE_PRODUCER_ERROR ((size_t)(-1))
+
+typedef size_t ZSTD_sequenceProducer_F (
+  void* sequenceProducerState,
+  ZSTD_Sequence* outSeqs, size_t outSeqsCapacity,
+  const void* src, size_t srcSize,
+  const void* dict, size_t dictSize,
+  int compressionLevel,
+  size_t windowSize
+);
+
+/*! ZSTD_registerSequenceProducer() :
+ * Instruct zstd to use a block-level external sequence producer function.
+ *
+ * The sequenceProducerState must be initialized by the caller, and the caller is
+ * responsible for managing its lifetime. This parameter is sticky across
+ * compressions. It will remain set until the user explicitly resets compression
+ * parameters.
+ *
+ * Sequence producer registration is considered to be an "advanced parameter",
+ * part of the "advanced API". This means it will only have an effect on compression
+ * APIs which respect advanced parameters, such as compress2() and compressStream2().
+ * Older compression APIs such as compressCCtx(), which predate the introduction of
+ * "advanced parameters", will ignore any external sequence producer setting.
+ *
+ * The sequence producer can be "cleared" by registering a NULL function pointer. This
+ * removes all limitations described above in the "LIMITATIONS" section of the API docs.
+ *
+ * The user is strongly encouraged to read the full API documentation (above) before
+ * calling this function. */
+ZSTDLIB_STATIC_API void
+ZSTD_registerSequenceProducer(
+  ZSTD_CCtx* cctx,
+  void* sequenceProducerState,
+  ZSTD_sequenceProducer_F* sequenceProducer
+);
+
+
 /*********************************************************************
-*  Buffer-less and synchronous inner streaming functions
+*  Buffer-less and synchronous inner streaming functions (DEPRECATED)
 *
-*  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 several restrictions, documented below.
-*  Prefer normal streaming API for an easier experience.
+*  This API is deprecated, and will be removed in a future version.
+*  It allows streaming (de)compression with user allocated buffers.
+*  However, it is hard to use, and not as well tested as the rest of
+*  our API.
+*
+*  Please use the normal streaming API instead: ZSTD_compressStream2,
+*  and ZSTD_decompressStream.
+*  If there is functionality that you need, but it doesn't provide,
+*  please open an issue on our GitHub.
 ********************************************************************* */
 
 /**
@@ -2636,15 +2845,20 @@ ZSTDLIB_STATIC_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
 */
 
 /*=====   Buffer-less streaming compression functions  =====*/
+ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);
+ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);
+ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); /**< note: fails if cdict==NULL */
 
 ZSTD_DEPRECATED("This function will likely be removed in a future release. It is misleading and has very limited utility.")
 ZSTDLIB_STATIC_API
 size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); /**<  note: if pledgedSrcSize is not known, use ZSTD_CONTENTSIZE_UNKNOWN */
 
+ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_compressContinue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
 
 /* The ZSTD_compressBegin_advanced() and ZSTD_compressBegin_usingCDict_advanced() are now DEPRECATED and will generate a compiler warning */
@@ -2728,29 +2942,7 @@ size_t ZSTD_compressBegin_usingCDict_advanced(ZSTD_CCtx* const cctx, const ZSTD_
 */
 
 /*=====   Buffer-less streaming decompression functions  =====*/
-typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
-typedef struct {
-    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;
-    unsigned _reserved1;
-    unsigned _reserved2;
-} ZSTD_frameHeader;
 
-/*! ZSTD_getFrameHeader() :
- *  decode Frame Header, or requires larger `srcSize`.
- * @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() */
-ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize);   /**< doesn't consume input */
-/*! ZSTD_getFrameHeader_advanced() :
- *  same as ZSTD_getFrameHeader(),
- *  with added capability to select a format (like ZSTD_f_zstd1_magicless) */
-ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);
 ZSTDLIB_STATIC_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_STATIC_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx);
@@ -2769,11 +2961,23 @@ ZSTDLIB_STATIC_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);
 
 
 
-/* ============================ */
-/**       Block level API       */
-/* ============================ */
+/* ========================================= */
+/**       Block level API (DEPRECATED)       */
+/* ========================================= */
 
 /*!
+
+    This API is deprecated in favor of the regular compression API.
+    You can get the frame header down to 2 bytes by setting:
+      - ZSTD_c_format = ZSTD_f_zstd1_magicless
+      - ZSTD_c_contentSizeFlag = 0
+      - ZSTD_c_checksumFlag = 0
+      - ZSTD_c_dictIDFlag = 0
+
+    This API is not as well tested as our normal API, so we recommend not using it.
+    We will be removing it in a future version. If the normal API doesn't provide
+    the functionality you need, please open a GitHub issue.
+
     Block functions produce and decode raw zstd blocks, without frame metadata.
     Frame metadata cost is typically ~12 bytes, which can be non-negligible for very small blocks (< 100 bytes).
     But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes.
@@ -2800,173 +3004,15 @@ ZSTDLIB_STATIC_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);
 */
 
 /*=====   Raw zstd block functions  =====*/
+ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_getBlockSize   (const ZSTD_CCtx* cctx);
+ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_compressBlock  (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.")
 ZSTDLIB_STATIC_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.")
 ZSTDLIB_STATIC_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. */
 
-
-/* ********************* BLOCK-LEVEL SEQUENCE PRODUCER API *********************
- *
- * *** OVERVIEW ***
- * The Block-Level Sequence Producer API allows users to provide their own custom
- * sequence producer which libzstd invokes to process each block. The produced list
- * of sequences (literals and matches) is then post-processed by libzstd to produce
- * valid compressed blocks.
- *
- * This block-level offload API is a more granular complement of the existing
- * frame-level offload API compressSequences() (introduced in v1.5.1). It offers
- * an easier migration story for applications already integrated with libzstd: the
- * user application continues to invoke the same compression functions
- * ZSTD_compress2() or ZSTD_compressStream2() as usual, and transparently benefits
- * from the specific advantages of the external sequence producer. For example,
- * the sequence producer could be tuned to take advantage of known characteristics
- * of the input, to offer better speed / ratio, or could leverage hardware
- * acceleration not available within libzstd itself.
- *
- * See contrib/externalSequenceProducer for an example program employing the
- * Block-Level Sequence Producer API.
- *
- * *** USAGE ***
- * The user is responsible for implementing a function of type
- * ZSTD_sequenceProducer_F. For each block, zstd will pass the following
- * arguments to the user-provided function:
- *
- *   - sequenceProducerState: a pointer to a user-managed state for the sequence
- *     producer.
- *
- *   - outSeqs, outSeqsCapacity: an output buffer for the sequence producer.
- *     outSeqsCapacity is guaranteed >= ZSTD_sequenceBound(srcSize). The memory
- *     backing outSeqs is managed by the CCtx.
- *
- *   - src, srcSize: an input buffer for the sequence producer to parse.
- *     srcSize is guaranteed to be <= ZSTD_BLOCKSIZE_MAX.
- *
- *   - dict, dictSize: a history buffer, which may be empty, which the sequence
- *     producer may reference as it parses the src buffer. Currently, zstd will
- *     always pass dictSize == 0 into external sequence producers, but this will
- *     change in the future.
- *
- *   - compressionLevel: a signed integer representing the zstd compression level
- *     set by the user for the current operation. The sequence producer may choose
- *     to use this information to change its compression strategy and speed/ratio
- *     tradeoff. Note: the compression level does not reflect zstd parameters set
- *     through the advanced API.
- *
- *   - windowSize: a size_t representing the maximum allowed offset for external
- *     sequences. Note that sequence offsets are sometimes allowed to exceed the
- *     windowSize if a dictionary is present, see doc/zstd_compression_format.md
- *     for details.
- *
- * The user-provided function shall return a size_t representing the number of
- * sequences written to outSeqs. This return value will be treated as an error
- * code if it is greater than outSeqsCapacity. The return value must be non-zero
- * if srcSize is non-zero. The ZSTD_SEQUENCE_PRODUCER_ERROR macro is provided
- * for convenience, but any value greater than outSeqsCapacity will be treated as
- * an error code.
- *
- * If the user-provided function does not return an error code, the sequences
- * written to outSeqs must be a valid parse of the src buffer. Data corruption may
- * occur if the parse is not valid. A parse is defined to be valid if the
- * following conditions hold:
- *   - The sum of matchLengths and literalLengths must equal srcSize.
- *   - All sequences in the parse, except for the final sequence, must have
- *     matchLength >= ZSTD_MINMATCH_MIN. The final sequence must have
- *     matchLength >= ZSTD_MINMATCH_MIN or matchLength == 0.
- *   - All offsets must respect the windowSize parameter as specified in
- *     doc/zstd_compression_format.md.
- *   - If the final sequence has matchLength == 0, it must also have offset == 0.
- *
- * zstd will only validate these conditions (and fail compression if they do not
- * hold) if the ZSTD_c_validateSequences cParam is enabled. Note that sequence
- * validation has a performance cost.
- *
- * If the user-provided function returns an error, zstd will either fall back
- * to an internal sequence producer or fail the compression operation. The user can
- * choose between the two behaviors by setting the ZSTD_c_enableSeqProducerFallback
- * cParam. Fallback compression will follow any other cParam settings, such as
- * compression level, the same as in a normal compression operation.
- *
- * The user shall instruct zstd to use a particular ZSTD_sequenceProducer_F
- * function by calling
- *         ZSTD_registerSequenceProducer(cctx,
- *                                       sequenceProducerState,
- *                                       sequenceProducer)
- * This setting will persist until the next parameter reset of the CCtx.
- *
- * The sequenceProducerState must be initialized by the user before calling
- * ZSTD_registerSequenceProducer(). The user is responsible for destroying the
- * sequenceProducerState.
- *
- * *** LIMITATIONS ***
- * This API is compatible with all zstd compression APIs which respect advanced parameters.
- * However, there are three limitations:
- *
- * First, the ZSTD_c_enableLongDistanceMatching cParam is not currently supported.
- * COMPRESSION WILL FAIL if it is enabled and the user tries to compress with a block-level
- * external sequence producer.
- *   - Note that ZSTD_c_enableLongDistanceMatching is auto-enabled by default in some
- *     cases (see its documentation for details). Users must explicitly set
- *     ZSTD_c_enableLongDistanceMatching to ZSTD_ps_disable in such cases if an external
- *     sequence producer is registered.
- *   - As of this writing, ZSTD_c_enableLongDistanceMatching is disabled by default
- *     whenever ZSTD_c_windowLog < 128MB, but that's subject to change. Users should
- *     check the docs on ZSTD_c_enableLongDistanceMatching whenever the Block-Level Sequence
- *     Producer API is used in conjunction with advanced settings (like ZSTD_c_windowLog).
- *
- * Second, history buffers are not currently supported. Concretely, zstd will always pass
- * dictSize == 0 to the external sequence producer (for now). This has two implications:
- *   - Dictionaries are not currently supported. Compression will *not* fail if the user
- *     references a dictionary, but the dictionary won't have any effect.
- *   - Stream history is not currently supported. All advanced compression APIs, including
- *     streaming APIs, work with external sequence producers, but each block is treated as
- *     an independent chunk without history from previous blocks.
- *
- * Third, multi-threading within a single compression is not currently supported. In other words,
- * COMPRESSION WILL FAIL if ZSTD_c_nbWorkers > 0 and an external sequence producer is registered.
- * Multi-threading across compressions is fine: simply create one CCtx per thread.
- *
- * Long-term, we plan to overcome all three limitations. There is no technical blocker to
- * overcoming them. It is purely a question of engineering effort.
- */
-
-#define ZSTD_SEQUENCE_PRODUCER_ERROR ((size_t)(-1))
-
-typedef size_t ZSTD_sequenceProducer_F (
-  void* sequenceProducerState,
-  ZSTD_Sequence* outSeqs, size_t outSeqsCapacity,
-  const void* src, size_t srcSize,
-  const void* dict, size_t dictSize,
-  int compressionLevel,
-  size_t windowSize
-);
-
-/*! ZSTD_registerSequenceProducer() :
- * Instruct zstd to use a block-level external sequence producer function.
- *
- * The sequenceProducerState must be initialized by the caller, and the caller is
- * responsible for managing its lifetime. This parameter is sticky across
- * compressions. It will remain set until the user explicitly resets compression
- * parameters.
- *
- * Sequence producer registration is considered to be an "advanced parameter",
- * part of the "advanced API". This means it will only have an effect on compression
- * APIs which respect advanced parameters, such as compress2() and compressStream2().
- * Older compression APIs such as compressCCtx(), which predate the introduction of
- * "advanced parameters", will ignore any external sequence producer setting.
- *
- * The sequence producer can be "cleared" by registering a NULL function pointer. This
- * removes all limitations described above in the "LIMITATIONS" section of the API docs.
- *
- * The user is strongly encouraged to read the full API documentation (above) before
- * calling this function. */
-ZSTDLIB_STATIC_API void
-ZSTD_registerSequenceProducer(
-  ZSTD_CCtx* cctx,
-  void* sequenceProducerState,
-  ZSTD_sequenceProducer_F* sequenceProducer
-);
-
 #endif   /* ZSTD_H_ZSTD_STATIC_LINKING_ONLY */
 
 #if defined (__cplusplus)
diff --git a/contrib/libs/zstd/programs/benchzstd.c b/contrib/libs/zstd/programs/benchzstd.c
index 63ecd99d51..1c809086db 100644
--- a/contrib/libs/zstd/programs/benchzstd.c
+++ b/contrib/libs/zstd/programs/benchzstd.c
@@ -697,9 +697,9 @@ static BMK_benchOutcome_t BMK_benchCLevel(const void* srcBuffer, size_t benchedS
                                 displayLevel, displayName, adv);
 }
 
-BMK_benchOutcome_t BMK_syntheticTest(int cLevel, double compressibility,
-                          const ZSTD_compressionParameters* compressionParams,
-                          int displayLevel, const BMK_advancedParams_t* adv)
+int BMK_syntheticTest(int cLevel, double compressibility,
+                      const ZSTD_compressionParameters* compressionParams,
+                      int displayLevel, const BMK_advancedParams_t* adv)
 {
     char name[20] = {0};
     size_t const benchedSize = 10000000;
@@ -707,12 +707,16 @@ BMK_benchOutcome_t BMK_syntheticTest(int cLevel, double compressibility,
     BMK_benchOutcome_t res;
 
     if (cLevel > ZSTD_maxCLevel()) {
-        RETURN_ERROR(15, BMK_benchOutcome_t, "Invalid Compression Level");
+        DISPLAYLEVEL(1, "Invalid Compression Level");
+        return 15;
     }
 
     /* Memory allocation */
     srcBuffer = malloc(benchedSize);
-    if (!srcBuffer) RETURN_ERROR(21, BMK_benchOutcome_t, "not enough memory");
+    if (!srcBuffer) {
+        DISPLAYLEVEL(1, "allocation error : not enough memory");
+        return 16;
+    }
 
     /* Fill input buffer */
     RDG_genBuffer(srcBuffer, benchedSize, compressibility, 0.0, 0);
@@ -728,7 +732,7 @@ BMK_benchOutcome_t BMK_syntheticTest(int cLevel, double compressibility,
     /* clean up */
     free(srcBuffer);
 
-    return res;
+    return !BMK_isSuccessful_benchOutcome(res);
 }
 
 
@@ -790,7 +794,7 @@ static int BMK_loadFiles(void* buffer, size_t bufferSize,
     return 0;
 }
 
-BMK_benchOutcome_t BMK_benchFilesAdvanced(
+int BMK_benchFilesAdvanced(
                         const char* const * fileNamesTable, unsigned nbFiles,
                         const char* dictFileName, int cLevel,
                         const ZSTD_compressionParameters* compressionParams,
@@ -805,19 +809,25 @@ BMK_benchOutcome_t BMK_benchFilesAdvanced(
     U64 const totalSizeToLoad = UTIL_getTotalFileSize(fileNamesTable, nbFiles);
 
     if (!nbFiles) {
-        RETURN_ERROR(14, BMK_benchOutcome_t, "No Files to Benchmark");
+        DISPLAYLEVEL(1, "No Files to Benchmark");
+        return 13;
     }
 
     if (cLevel > ZSTD_maxCLevel()) {
-        RETURN_ERROR(15, BMK_benchOutcome_t, "Invalid Compression Level");
+        DISPLAYLEVEL(1, "Invalid Compression Level");
+        return 14;
     }
 
     if (totalSizeToLoad == UTIL_FILESIZE_UNKNOWN) {
-        RETURN_ERROR(9, BMK_benchOutcome_t, "Error loading files");
+        DISPLAYLEVEL(1, "Error loading files");
+        return 15;
     }
 
     fileSizes = (size_t*)calloc(nbFiles, sizeof(size_t));
-    if (!fileSizes) RETURN_ERROR(12, BMK_benchOutcome_t, "not enough memory for fileSizes");
+    if (!fileSizes) {
+        DISPLAYLEVEL(1, "not enough memory for fileSizes");
+        return 16;
+    }
 
     /* Load dictionary */
     if (dictFileName != NULL) {
@@ -825,18 +835,21 @@ BMK_benchOutcome_t BMK_benchFilesAdvanced(
         if (dictFileSize == UTIL_FILESIZE_UNKNOWN) {
             DISPLAYLEVEL(1, "error loading %s : %s \n", dictFileName, strerror(errno));
             free(fileSizes);
-            RETURN_ERROR(9, BMK_benchOutcome_t, "benchmark aborted");
+            DISPLAYLEVEL(1, "benchmark aborted");
+            return 17;
         }
         if (dictFileSize > 64 MB) {
             free(fileSizes);
-            RETURN_ERROR(10, BMK_benchOutcome_t, "dictionary file %s too large", dictFileName);
+            DISPLAYLEVEL(1, "dictionary file %s too large", dictFileName);
+            return 18;
         }
         dictBufferSize = (size_t)dictFileSize;
         dictBuffer = malloc(dictBufferSize);
         if (dictBuffer==NULL) {
             free(fileSizes);
-            RETURN_ERROR(11, BMK_benchOutcome_t, "not enough memory for dictionary (%u bytes)",
+            DISPLAYLEVEL(1, "not enough memory for dictionary (%u bytes)",
                             (unsigned)dictBufferSize);
+            return 19;
         }
 
         {   int const errorCode = BMK_loadFiles(dictBuffer, dictBufferSize,
@@ -858,7 +871,8 @@ BMK_benchOutcome_t BMK_benchFilesAdvanced(
     if (!srcBuffer) {
         free(dictBuffer);
         free(fileSizes);
-        RETURN_ERROR(12, BMK_benchOutcome_t, "not enough memory");
+        DISPLAYLEVEL(1, "not enough memory for srcBuffer");
+        return 20;
     }
 
     /* Load input buffer */
@@ -886,12 +900,11 @@ _cleanUp:
     free(srcBuffer);
     free(dictBuffer);
     free(fileSizes);
-    return res;
+    return !BMK_isSuccessful_benchOutcome(res);
 }
 
 
-BMK_benchOutcome_t BMK_benchFiles(
-                    const char* const * fileNamesTable, unsigned nbFiles,
+int BMK_benchFiles(const char* const * fileNamesTable, unsigned nbFiles,
                     const char* dictFileName,
                     int cLevel, const ZSTD_compressionParameters* compressionParams,
                     int displayLevel)
diff --git a/contrib/libs/zstd/programs/benchzstd.h b/contrib/libs/zstd/programs/benchzstd.h
index aa683dfc25..f14a681925 100644
--- a/contrib/libs/zstd/programs/benchzstd.h
+++ b/contrib/libs/zstd/programs/benchzstd.h
@@ -81,21 +81,13 @@ BMK_benchResult_t BMK_extract_benchResult(BMK_benchOutcome_t outcome);
  *      2 : + result + interaction + warnings;
  *      3 : + information;
  *      4 : + debug
- * @return:
- *      a variant, which expresses either an error, or a valid result.
- *      Use BMK_isSuccessful_benchOutcome() to check if function was successful.
- *      If yes, extract the valid result with BMK_extract_benchResult(),
- *      it will contain :
- *          .cSpeed: compression speed in bytes per second,
- *          .dSpeed: decompression speed in bytes per second,
- *          .cSize : compressed size, in bytes
- *          .cMem  : memory budget required for the compression context
+ * @return: 0 on success, !0 on error
  */
-BMK_benchOutcome_t BMK_benchFiles(
-                   const char* const * fileNamesTable, unsigned nbFiles,
-                   const char* dictFileName,
-                   int cLevel, const ZSTD_compressionParameters* compressionParams,
-                   int displayLevel);
+int BMK_benchFiles(
+            const char* const * fileNamesTable, unsigned nbFiles,
+            const char* dictFileName,
+            int cLevel, const ZSTD_compressionParameters* compressionParams,
+            int displayLevel);
 
 
 typedef enum {
@@ -126,11 +118,11 @@ BMK_advancedParams_t BMK_initAdvancedParams(void);
 /*! BMK_benchFilesAdvanced():
  *  Same as BMK_benchFiles(),
  *  with more controls, provided through advancedParams_t structure */
-BMK_benchOutcome_t BMK_benchFilesAdvanced(
-                   const char* const * fileNamesTable, unsigned nbFiles,
-                   const char* dictFileName,
-                   int cLevel, const ZSTD_compressionParameters* compressionParams,
-                   int displayLevel, const BMK_advancedParams_t* adv);
+int BMK_benchFilesAdvanced(
+               const char* const * fileNamesTable, unsigned nbFiles,
+               const char* dictFileName,
+               int cLevel, const ZSTD_compressionParameters* compressionParams,
+               int displayLevel, const BMK_advancedParams_t* adv);
 
 /*! BMK_syntheticTest() -- called from zstdcli */
 /*  Generates a sample with datagen, using compressibility argument */
@@ -139,20 +131,11 @@ BMK_benchOutcome_t BMK_benchFilesAdvanced(
  *  compressionParams - basic compression Parameters
  *  displayLevel - see benchFiles
  *  adv - see advanced_Params_t
- * @return:
- *      a variant, which expresses either an error, or a valid result.
- *      Use BMK_isSuccessful_benchOutcome() to check if function was successful.
- *      If yes, extract the valid result with BMK_extract_benchResult(),
- *      it will contain :
- *          .cSpeed: compression speed in bytes per second,
- *          .dSpeed: decompression speed in bytes per second,
- *          .cSize : compressed size, in bytes
- *          .cMem  : memory budget required for the compression context
+ * @return: 0 on success, !0 on error
  */
-BMK_benchOutcome_t BMK_syntheticTest(
-                          int cLevel, double compressibility,
-                          const ZSTD_compressionParameters* compressionParams,
-                          int displayLevel, const BMK_advancedParams_t* adv);
+int BMK_syntheticTest(int cLevel, double compressibility,
+                      const ZSTD_compressionParameters* compressionParams,
+                      int displayLevel, const BMK_advancedParams_t* adv);
 
 
 
@@ -190,8 +173,8 @@ BMK_benchOutcome_t BMK_benchMem(const void* srcBuffer, size_t srcSize,
                         int displayLevel, const char* displayName);
 
 
-/* BMK_benchMemAdvanced() : same as BMK_benchMem()
- * with following additional options :
+/* BMK_benchMemAdvanced() : used by Paramgrill
+ * same as BMK_benchMem() with following additional options :
  * dstBuffer - destination buffer to write compressed output in, NULL if none provided.
  * dstCapacity - capacity of destination buffer, give 0 if dstBuffer = NULL
  * adv = see advancedParams_t
diff --git a/contrib/libs/zstd/programs/fileio.c b/contrib/libs/zstd/programs/fileio.c
index 3b885bc65f..546fd35622 100644
--- a/contrib/libs/zstd/programs/fileio.c
+++ b/contrib/libs/zstd/programs/fileio.c
@@ -485,6 +485,11 @@ void FIO_setPassThroughFlag(FIO_prefs_t* const prefs, int value) {
     prefs->passThrough = (value != 0);
 }
 
+void FIO_setMMapDict(FIO_prefs_t* const prefs, ZSTD_paramSwitch_e value)
+{
+    prefs->mmapDict = value;
+}
+
 /* FIO_ctx_t functions */
 
 void FIO_setHasStdoutOutput(FIO_ctx_t* const fCtx, int value) {
@@ -576,6 +581,8 @@ FIO_openDstFile(FIO_ctx_t* fCtx, FIO_prefs_t* const prefs,
                 const char* srcFileName, const char* dstFileName,
                 const int mode)
 {
+    int isDstRegFile;
+
     if (prefs->testMode) return NULL;  /* do not open file in test mode */
 
     assert(dstFileName != NULL);
@@ -595,11 +602,16 @@ FIO_openDstFile(FIO_ctx_t* fCtx, FIO_prefs_t* const prefs,
         return NULL;
     }
 
+    isDstRegFile = UTIL_isRegularFile(dstFileName);  /* invoke once */
     if (prefs->sparseFileSupport == 1) {
         prefs->sparseFileSupport = ZSTD_SPARSE_DEFAULT;
+        if (!isDstRegFile) {
+            prefs->sparseFileSupport = 0;
+            DISPLAYLEVEL(4, "Sparse File Support is disabled when output is not a file \n");
+        }
     }
 
-    if (UTIL_isRegularFile(dstFileName)) {
+    if (isDstRegFile) {
         /* Check if destination file already exists */
 #if !defined(_WIN32)
         /* this test does not work on Windows :
@@ -644,32 +656,55 @@ FIO_openDstFile(FIO_ctx_t* fCtx, FIO_prefs_t* const prefs,
 #endif
         if (f == NULL) {
             DISPLAYLEVEL(1, "zstd: %s: %s\n", dstFileName, strerror(errno));
+        } else {
+            /* An increased buffer size can provide a significant performance
+             * boost on some platforms. Note that providing a NULL buf with a
+             * size that's not 0 is not defined in ANSI C, but is defined in an
+             * extension. There are three possibilities here:
+             * 1. Libc supports the extended version and everything is good.
+             * 2. Libc ignores the size when buf is NULL, in which case
+             *    everything will continue as if we didn't call `setvbuf()`.
+             * 3. We fail the call and execution continues but a warning
+             *    message might be shown.
+             * In all cases due execution continues. For now, I believe that
+             * this is a more cost-effective solution than managing the buffers
+             * allocations ourselves (will require an API change).
+             */
+            if (setvbuf(f, NULL, _IOFBF, 1 MB)) {
+                DISPLAYLEVEL(2, "Warning: setvbuf failed for %s\n", dstFileName);
+            }
         }
-        /* An increased buffer size can provide a significant performance boost on some platforms.
-         * Note that providing a NULL buf with a size that's not 0 is not defined in ANSI C, but is defined
-         * in an extension. There are three possibilities here -
-         * 1. Libc supports the extended version and everything is good.
-         * 2. Libc ignores the size when buf is NULL, in which case everything will continue as if we didn't
-         *    call `setvbuf`.
-         * 3. We fail the call and execution continues but a warning message might be shown.
-         * In all cases due execution continues. For now, I believe that this is a more cost-effective
-         * solution than managing the buffers allocations ourselves (will require an API change). */
-        if(setvbuf(f, NULL, _IOFBF, 1 MB))
-            DISPLAYLEVEL(2, "Warning: setvbuf failed for %s\n", dstFileName);
         return f;
     }
 }
 
-/*! FIO_createDictBuffer() :
- *  creates a buffer, pointed by `*bufferPtr`,
+
+/* FIO_getDictFileStat() :
+ */
+static void FIO_getDictFileStat(const char* fileName, stat_t* dictFileStat) {
+    assert(dictFileStat != NULL);
+    if (fileName == NULL) return;
+
+    if (!UTIL_stat(fileName, dictFileStat)) {
+        EXM_THROW(31, "Stat failed on dictionary file %s: %s", fileName, strerror(errno));
+    }
+
+    if (!UTIL_isRegularFileStat(dictFileStat)) {
+        EXM_THROW(32, "Dictionary %s must be a regular file.", fileName);
+    }
+}
+
+/*  FIO_setDictBufferMalloc() :
+ *  allocates a buffer, pointed by `dict->dictBuffer`,
  *  loads `filename` content into it, up to DICTSIZE_MAX bytes.
  * @return : loaded size
  *  if fileName==NULL, returns 0 and a NULL pointer
  */
-static size_t FIO_createDictBuffer(void** bufferPtr, const char* fileName, FIO_prefs_t* const prefs, stat_t* dictFileStat)
+static size_t FIO_setDictBufferMalloc(FIO_Dict_t* dict, const char* fileName, FIO_prefs_t* const prefs, stat_t* dictFileStat)
 {
     FILE* fileHandle;
     U64 fileSize;
+    void** bufferPtr = &dict->dictBuffer;
 
     assert(bufferPtr != NULL);
     assert(dictFileStat != NULL);
@@ -678,14 +713,6 @@ static size_t FIO_createDictBuffer(void** bufferPtr, const char* fileName, FIO_p
 
     DISPLAYLEVEL(4,"Loading %s as dictionary \n", fileName);
 
-    if (!UTIL_stat(fileName, dictFileStat)) {
-        EXM_THROW(31, "Stat failed on dictionary file %s: %s", fileName, strerror(errno));
-    }
-
-    if (!UTIL_isRegularFileStat(dictFileStat)) {
-        EXM_THROW(32, "Dictionary %s must be a regular file.", fileName);
-    }
-
     fileHandle = fopen(fileName, "rb");
 
     if (fileHandle == NULL) {
@@ -712,6 +739,130 @@ static size_t FIO_createDictBuffer(void** bufferPtr, const char* fileName, FIO_p
     return (size_t)fileSize;
 }
 
+#if (PLATFORM_POSIX_VERSION > 0)
+#include <sys/mman.h>
+static void FIO_munmap(FIO_Dict_t* dict)
+{
+    munmap(dict->dictBuffer, dict->dictBufferSize);
+    dict->dictBuffer = NULL;
+    dict->dictBufferSize = 0;
+}
+static size_t FIO_setDictBufferMMap(FIO_Dict_t* dict, const char* fileName, FIO_prefs_t* const prefs, stat_t* dictFileStat)
+{
+    int fileHandle;
+    U64 fileSize;
+    void** bufferPtr = &dict->dictBuffer;
+
+    assert(bufferPtr != NULL);
+    assert(dictFileStat != NULL);
+    *bufferPtr = NULL;
+    if (fileName == NULL) return 0;
+
+    DISPLAYLEVEL(4,"Loading %s as dictionary \n", fileName);
+
+    fileHandle = open(fileName, O_RDONLY);
+
+    if (fileHandle == -1) {
+        EXM_THROW(33, "Couldn't open dictionary %s: %s", fileName, strerror(errno));
+    }
+
+    fileSize = UTIL_getFileSizeStat(dictFileStat);
+    {
+        size_t const dictSizeMax = prefs->patchFromMode ? prefs->memLimit : DICTSIZE_MAX;
+        if (fileSize >  dictSizeMax) {
+            EXM_THROW(34, "Dictionary file %s is too large (> %u bytes)",
+                            fileName,  (unsigned)dictSizeMax);   /* avoid extreme cases */
+        }
+    }
+
+    *bufferPtr = mmap(NULL, (size_t)fileSize, PROT_READ, MAP_PRIVATE, fileHandle, 0);
+    if (*bufferPtr==NULL) EXM_THROW(34, "%s", strerror(errno));
+
+    close(fileHandle);
+    return (size_t)fileSize;
+}
+#elif defined(_MSC_VER) || defined(_WIN32)
+#include <windows.h>
+static void FIO_munmap(FIO_Dict_t* dict)
+{
+    UnmapViewOfFile(dict->dictBuffer);
+    CloseHandle(dict->dictHandle);
+    dict->dictBuffer = NULL;
+    dict->dictBufferSize = 0;
+}
+static size_t FIO_setDictBufferMMap(FIO_Dict_t* dict, const char* fileName, FIO_prefs_t* const prefs, stat_t* dictFileStat)
+{
+    HANDLE fileHandle, mapping;
+    U64 fileSize;
+    void** bufferPtr = &dict->dictBuffer;
+
+    assert(bufferPtr != NULL);
+    assert(dictFileStat != NULL);
+    *bufferPtr = NULL;
+    if (fileName == NULL) return 0;
+
+    DISPLAYLEVEL(4,"Loading %s as dictionary \n", fileName);
+
+    fileHandle = CreateFileA(fileName, GENERIC_READ, FILE_SHARE_READ, NULL, OPEN_EXISTING, FILE_ATTRIBUTE_READONLY, NULL);
+
+    if (fileHandle == INVALID_HANDLE_VALUE) {
+        EXM_THROW(33, "Couldn't open dictionary %s: %s", fileName, strerror(errno));
+    }
+
+    fileSize = UTIL_getFileSizeStat(dictFileStat);
+    {
+        size_t const dictSizeMax = prefs->patchFromMode ? prefs->memLimit : DICTSIZE_MAX;
+        if (fileSize >  dictSizeMax) {
+            EXM_THROW(34, "Dictionary file %s is too large (> %u bytes)",
+                            fileName,  (unsigned)dictSizeMax);   /* avoid extreme cases */
+        }
+    }
+
+    mapping = CreateFileMapping(fileHandle, NULL, PAGE_READONLY, 0, 0, NULL);
+	if (mapping == NULL) {
+        EXM_THROW(35, "Couldn't map dictionary %s: %s", fileName, strerror(errno));
+    }
+
+	*bufferPtr = MapViewOfFile(mapping, FILE_MAP_READ, 0, 0, (DWORD)fileSize); /* we can only cast to DWORD here because dictSize <= 2GB */
+	if (*bufferPtr==NULL) EXM_THROW(36, "%s", strerror(errno));
+
+    dict->dictHandle = fileHandle;
+    return (size_t)fileSize;
+}
+#else
+static size_t FIO_setDictBufferMMap(FIO_Dict_t* dict, const char* fileName, FIO_prefs_t* const prefs, stat_t* dictFileStat)
+{
+   return FIO_setDictBufferMalloc(dict, fileName, prefs, dictFileStat);
+}
+static void FIO_munmap(FIO_Dict_t* dict) {
+   free(dict->dictBuffer);
+   dict->dictBuffer = NULL;
+   dict->dictBufferSize = 0;
+}
+#endif
+
+static void FIO_freeDict(FIO_Dict_t* dict) {
+    if (dict->dictBufferType == FIO_mallocDict) {
+        free(dict->dictBuffer);
+        dict->dictBuffer = NULL;
+        dict->dictBufferSize = 0;
+    } else if (dict->dictBufferType == FIO_mmapDict)  {
+        FIO_munmap(dict);
+    } else {
+        assert(0); /* Should not reach this case */
+    }
+}
+
+static void FIO_initDict(FIO_Dict_t* dict, const char* fileName, FIO_prefs_t* const prefs, stat_t* dictFileStat, FIO_dictBufferType_t dictBufferType) {
+    dict->dictBufferType = dictBufferType;
+    if (dict->dictBufferType == FIO_mallocDict) {
+        dict->dictBufferSize = FIO_setDictBufferMalloc(dict, fileName, prefs, dictFileStat);
+    } else if (dict->dictBufferType == FIO_mmapDict)  {
+        dict->dictBufferSize = FIO_setDictBufferMMap(dict, fileName, prefs, dictFileStat);
+    } else {
+        assert(0); /* Should not reach this case */
+    }
+}
 
 
 /* FIO_checkFilenameCollisions() :
@@ -914,8 +1065,7 @@ static ZSTD_outBuffer setOutBuffer(void* buf, size_t s, size_t pos)
  *  Compression
  ************************************************************************/
 typedef struct {
-    void* dictBuffer;
-    size_t dictBufferSize;
+    FIO_Dict_t dict;
     const char* dictFileName;
     stat_t dictFileStat;
     ZSTD_CStream* cctx;
@@ -961,6 +1111,9 @@ static void FIO_adjustParamsForPatchFromMode(FIO_prefs_t* const prefs,
 static cRess_t FIO_createCResources(FIO_prefs_t* const prefs,
                                     const char* dictFileName, unsigned long long const maxSrcFileSize,
                                     int cLevel, ZSTD_compressionParameters comprParams) {
+    int useMMap = prefs->mmapDict == ZSTD_ps_enable;
+    int forceNoUseMMap = prefs->mmapDict == ZSTD_ps_disable;
+    FIO_dictBufferType_t dictBufferType;
     cRess_t ress;
     memset(&ress, 0, sizeof(ress));
 
@@ -970,19 +1123,25 @@ static cRess_t FIO_createCResources(FIO_prefs_t* const prefs,
         EXM_THROW(30, "allocation error (%s): can't create ZSTD_CCtx",
                     strerror(errno));
 
+    FIO_getDictFileStat(dictFileName, &ress.dictFileStat);
+
     /* need to update memLimit before calling createDictBuffer
      * because of memLimit check inside it */
     if (prefs->patchFromMode) {
+        U64 const dictSize = UTIL_getFileSizeStat(&ress.dictFileStat);
         unsigned long long const ssSize = (unsigned long long)prefs->streamSrcSize;
-        FIO_adjustParamsForPatchFromMode(prefs, &comprParams, UTIL_getFileSize(dictFileName), ssSize > 0 ? ssSize : maxSrcFileSize, cLevel);
+        useMMap |= dictSize > prefs->memLimit;
+        FIO_adjustParamsForPatchFromMode(prefs, &comprParams, dictSize, ssSize > 0 ? ssSize : maxSrcFileSize, cLevel);
     }
-    ress.dictBufferSize = FIO_createDictBuffer(&ress.dictBuffer, dictFileName, prefs, &ress.dictFileStat);   /* works with dictFileName==NULL */
+
+    dictBufferType = (useMMap && !forceNoUseMMap) ? FIO_mmapDict : FIO_mallocDict;
+    FIO_initDict(&ress.dict, dictFileName, prefs, &ress.dictFileStat, dictBufferType);   /* works with dictFileName==NULL */
 
     ress.writeCtx = AIO_WritePool_create(prefs, ZSTD_CStreamOutSize());
     ress.readCtx = AIO_ReadPool_create(prefs, ZSTD_CStreamInSize());
 
     /* Advanced parameters, including dictionary */
-    if (dictFileName && (ress.dictBuffer==NULL))
+    if (dictFileName && (ress.dict.dictBuffer==NULL))
         EXM_THROW(32, "allocation error : can't create dictBuffer");
     ress.dictFileName = dictFileName;
 
@@ -1032,17 +1191,17 @@ static cRess_t FIO_createCResources(FIO_prefs_t* const prefs,
 #endif
     /* dictionary */
     if (prefs->patchFromMode) {
-        CHECK( ZSTD_CCtx_refPrefix(ress.cctx, ress.dictBuffer, ress.dictBufferSize) );
+        CHECK( ZSTD_CCtx_refPrefix(ress.cctx, ress.dict.dictBuffer, ress.dict.dictBufferSize) );
     } else {
-        CHECK( ZSTD_CCtx_loadDictionary(ress.cctx, ress.dictBuffer, ress.dictBufferSize) );
+        CHECK( ZSTD_CCtx_loadDictionary_byReference(ress.cctx, ress.dict.dictBuffer, ress.dict.dictBufferSize) );
     }
 
     return ress;
 }
 
-static void FIO_freeCResources(const cRess_t* const ress)
+static void FIO_freeCResources(cRess_t* const ress)
 {
-    free(ress->dictBuffer);
+    FIO_freeDict(&(ress->dict));
     AIO_WritePool_free(ress->writeCtx);
     AIO_ReadPool_free(ress->readCtx);
     ZSTD_freeCStream(ress->cctx);   /* never fails */
@@ -1173,8 +1332,8 @@ FIO_compressLzmaFrame(cRess_t* ress,
     }
 
     writeJob =AIO_WritePool_acquireJob(ress->writeCtx);
-    strm.next_out = (Bytef*)writeJob->buffer;
-    strm.avail_out = (uInt)writeJob->bufferSize;
+    strm.next_out = (BYTE*)writeJob->buffer;
+    strm.avail_out = writeJob->bufferSize;
     strm.next_in = 0;
     strm.avail_in = 0;
 
@@ -1201,7 +1360,7 @@ FIO_compressLzmaFrame(cRess_t* ress,
                 writeJob->usedBufferSize = compBytes;
                 AIO_WritePool_enqueueAndReacquireWriteJob(&writeJob);
                 outFileSize += compBytes;
-                strm.next_out = (Bytef*)writeJob->buffer;
+                strm.next_out = (BYTE*)writeJob->buffer;
                 strm.avail_out = writeJob->bufferSize;
         }   }
         if (srcFileSize == UTIL_FILESIZE_UNKNOWN)
@@ -1681,6 +1840,7 @@ static int FIO_compressFilename_dstFile(FIO_ctx_t* const fCtx,
     int result;
     int transferStat = 0;
     FILE *dstFile;
+    int dstFd = -1;
 
     assert(AIO_ReadPool_getFile(ress.readCtx) != NULL);
     if (AIO_WritePool_getFile(ress.writeCtx) == NULL) {
@@ -1696,6 +1856,7 @@ static int FIO_compressFilename_dstFile(FIO_ctx_t* const fCtx,
         DISPLAYLEVEL(6, "FIO_compressFilename_dstFile: opening dst: %s \n", dstFileName);
         dstFile = FIO_openDstFile(fCtx, prefs, srcFileName, dstFileName, dstFileInitialPermissions);
         if (dstFile==NULL) return 1;  /* could not open dstFileName */
+        dstFd = fileno(dstFile);
         AIO_WritePool_setFile(ress.writeCtx, dstFile);
         /* Must only be added after FIO_openDstFile() succeeds.
          * Otherwise we may delete the destination file if it already exists,
@@ -1709,14 +1870,20 @@ static int FIO_compressFilename_dstFile(FIO_ctx_t* const fCtx,
     if (closeDstFile) {
         clearHandler();
 
+        if (transferStat) {
+            UTIL_setFDStat(dstFd, dstFileName, srcFileStat);
+        }
+
         DISPLAYLEVEL(6, "FIO_compressFilename_dstFile: closing dst: %s \n", dstFileName);
         if (AIO_WritePool_closeFile(ress.writeCtx)) { /* error closing file */
             DISPLAYLEVEL(1, "zstd: %s: %s \n", dstFileName, strerror(errno));
             result=1;
         }
+
         if (transferStat) {
-            UTIL_setFileStat(dstFileName, srcFileStat);
+            UTIL_utime(dstFileName, srcFileStat);
         }
+
         if ( (result != 0)  /* operation failure */
           && strcmp(dstFileName, stdoutmark)  /* special case : don't remove() stdout */
           ) {
@@ -1873,7 +2040,7 @@ int FIO_compressFilename(FIO_ctx_t* const fCtx, FIO_prefs_t* const prefs, const
                          const char* srcFileName, const char* dictFileName,
                          int compressionLevel, ZSTD_compressionParameters comprParams)
 {
-    cRess_t const ress = FIO_createCResources(prefs, dictFileName, UTIL_getFileSize(srcFileName), compressionLevel, comprParams);
+    cRess_t ress = FIO_createCResources(prefs, dictFileName, UTIL_getFileSize(srcFileName), compressionLevel, comprParams);
     int const result = FIO_compressFilename_srcFile(fCtx, prefs, ress, dstFileName, srcFileName, compressionLevel);
 
 #define DISPLAY_LEVEL_DEFAULT 2
@@ -2043,6 +2210,7 @@ int FIO_compressMultipleFilenames(FIO_ctx_t* const fCtx,
  *  Decompression
  ***************************************************************************/
 typedef struct {
+    FIO_Dict_t dict;
     ZSTD_DStream* dctx;
     WritePoolCtx_t *writeCtx;
     ReadPoolCtx_t *readCtx;
@@ -2050,11 +2218,19 @@ typedef struct {
 
 static dRess_t FIO_createDResources(FIO_prefs_t* const prefs, const char* dictFileName)
 {
+    int useMMap = prefs->mmapDict == ZSTD_ps_enable;
+    int forceNoUseMMap = prefs->mmapDict == ZSTD_ps_disable;
+    stat_t statbuf;
     dRess_t ress;
     memset(&ress, 0, sizeof(ress));
 
-    if (prefs->patchFromMode)
-        FIO_adjustMemLimitForPatchFromMode(prefs, UTIL_getFileSize(dictFileName), 0 /* just use the dict size */);
+    FIO_getDictFileStat(dictFileName, &statbuf);
+
+    if (prefs->patchFromMode){
+        U64 const dictSize = UTIL_getFileSizeStat(&statbuf);
+        useMMap |= dictSize > prefs->memLimit;
+        FIO_adjustMemLimitForPatchFromMode(prefs, dictSize, 0 /* just use the dict size */);
+    }
 
     /* Allocation */
     ress.dctx = ZSTD_createDStream();
@@ -2064,29 +2240,34 @@ static dRess_t FIO_createDResources(FIO_prefs_t* const prefs, const char* dictFi
     CHECK( ZSTD_DCtx_setParameter(ress.dctx, ZSTD_d_forceIgnoreChecksum, !prefs->checksumFlag));
 
     /* dictionary */
-    {   void* dictBuffer;
-        stat_t statbuf;
-        size_t const dictBufferSize = FIO_createDictBuffer(&dictBuffer, dictFileName, prefs, &statbuf);
-        CHECK( ZSTD_DCtx_reset(ress.dctx, ZSTD_reset_session_only) );
-        CHECK( ZSTD_DCtx_loadDictionary(ress.dctx, dictBuffer, dictBufferSize) );
-        free(dictBuffer);
+    {
+        FIO_dictBufferType_t dictBufferType = (useMMap && !forceNoUseMMap) ? FIO_mmapDict : FIO_mallocDict;
+        FIO_initDict(&ress.dict, dictFileName, prefs, &statbuf, dictBufferType);
+
+        CHECK(ZSTD_DCtx_reset(ress.dctx, ZSTD_reset_session_only) );
+
+        if (prefs->patchFromMode){
+            CHECK(ZSTD_DCtx_refPrefix(ress.dctx, ress.dict.dictBuffer, ress.dict.dictBufferSize));
+        } else {
+            CHECK(ZSTD_DCtx_loadDictionary_byReference(ress.dctx, ress.dict.dictBuffer, ress.dict.dictBufferSize));
+        }
     }
 
     ress.writeCtx = AIO_WritePool_create(prefs, ZSTD_DStreamOutSize());
     ress.readCtx = AIO_ReadPool_create(prefs, ZSTD_DStreamInSize());
-
     return ress;
 }
 
 static void FIO_freeDResources(dRess_t ress)
 {
+    FIO_freeDict(&(ress.dict));
     CHECK( ZSTD_freeDStream(ress.dctx) );
     AIO_WritePool_free(ress.writeCtx);
     AIO_ReadPool_free(ress.readCtx);
 }
 
-/** FIO_passThrough() : just copy input into output, for compatibility with gzip -df mode
-    @return : 0 (no error) */
+/* FIO_passThrough() : just copy input into output, for compatibility with gzip -df mode
+ * @return : 0 (no error) */
 static int FIO_passThrough(dRess_t *ress)
 {
     size_t const blockSize = MIN(MIN(64 KB, ZSTD_DStreamInSize()), ZSTD_DStreamOutSize());
@@ -2114,7 +2295,8 @@ static int FIO_passThrough(dRess_t *ress)
 static void
 FIO_zstdErrorHelp(const FIO_prefs_t* const prefs,
                   const dRess_t* ress,
-                  size_t err, const char* srcFileName)
+                  size_t err,
+                  const char* srcFileName)
 {
     ZSTD_frameHeader header;
 
@@ -2316,8 +2498,8 @@ FIO_decompressLzmaFrame(dRess_t* ress,
     }
 
     writeJob = AIO_WritePool_acquireJob(ress->writeCtx);
-    strm.next_out = (Bytef*)writeJob->buffer;
-    strm.avail_out = (uInt)writeJob->bufferSize;
+    strm.next_out = (BYTE*)writeJob->buffer;
+    strm.avail_out = writeJob->bufferSize;
     strm.next_in = (BYTE const*)ress->readCtx->srcBuffer;
     strm.avail_in = ress->readCtx->srcBufferLoaded;
 
@@ -2345,7 +2527,7 @@ FIO_decompressLzmaFrame(dRess_t* ress,
                 writeJob->usedBufferSize = decompBytes;
                 AIO_WritePool_enqueueAndReacquireWriteJob(&writeJob);
                 outFileSize += decompBytes;
-                strm.next_out = (Bytef*)writeJob->buffer;
+                strm.next_out = (BYTE*)writeJob->buffer;
                 strm.avail_out = writeJob->bufferSize;
         }   }
         if (ret == LZMA_STREAM_END) break;
@@ -2540,6 +2722,7 @@ static int FIO_decompressDstFile(FIO_ctx_t* const fCtx,
     int result;
     int releaseDstFile = 0;
     int transferStat = 0;
+    int dstFd = 0;
 
     if ((AIO_WritePool_getFile(ress.writeCtx) == NULL) && (prefs->testMode == 0)) {
         FILE *dstFile;
@@ -2555,6 +2738,7 @@ static int FIO_decompressDstFile(FIO_ctx_t* const fCtx,
 
         dstFile = FIO_openDstFile(fCtx, prefs, srcFileName, dstFileName, dstFilePermissions);
         if (dstFile==NULL) return 1;
+        dstFd = fileno(dstFile);
         AIO_WritePool_setFile(ress.writeCtx, dstFile);
 
         /* Must only be added after FIO_openDstFile() succeeds.
@@ -2568,13 +2752,18 @@ static int FIO_decompressDstFile(FIO_ctx_t* const fCtx,
 
     if (releaseDstFile) {
         clearHandler();
+
+        if (transferStat) {
+            UTIL_setFDStat(dstFd, dstFileName, srcFileStat);
+        }
+
         if (AIO_WritePool_closeFile(ress.writeCtx)) {
             DISPLAYLEVEL(1, "zstd: %s: %s \n", dstFileName, strerror(errno));
             result = 1;
         }
 
         if (transferStat) {
-            UTIL_setFileStat(dstFileName, srcFileStat);
+            UTIL_utime(dstFileName, srcFileStat);
         }
 
         if ( (result != 0)  /* operation failure */
@@ -2655,6 +2844,8 @@ int FIO_decompressFilename(FIO_ctx_t* const fCtx, FIO_prefs_t* const prefs,
 
     int const decodingError = FIO_decompressSrcFile(fCtx, prefs, ress, dstFileName, srcFileName);
 
+
+
     FIO_freeDResources(ress);
     return decodingError;
 }
diff --git a/contrib/libs/zstd/programs/fileio.h b/contrib/libs/zstd/programs/fileio.h
index 291d4d4145..224d89525d 100644
--- a/contrib/libs/zstd/programs/fileio.h
+++ b/contrib/libs/zstd/programs/fileio.h
@@ -106,6 +106,7 @@ void FIO_setContentSize(FIO_prefs_t* const prefs, int value);
 void FIO_displayCompressionParameters(const FIO_prefs_t* prefs);
 void FIO_setAsyncIOFlag(FIO_prefs_t* const prefs, int value);
 void FIO_setPassThroughFlag(FIO_prefs_t* const prefs, int value);
+void FIO_setMMapDict(FIO_prefs_t* const prefs, ZSTD_paramSwitch_e value);
 
 /* FIO_ctx_t functions */
 void FIO_setNbFilesTotal(FIO_ctx_t* const fCtx, int value);
diff --git a/contrib/libs/zstd/programs/fileio_types.h b/contrib/libs/zstd/programs/fileio_types.h
index c1f42f1ad0..2994a60929 100644
--- a/contrib/libs/zstd/programs/fileio_types.h
+++ b/contrib/libs/zstd/programs/fileio_types.h
@@ -69,6 +69,18 @@ typedef struct FIO_prefs_s {
     int contentSize;
     int allowBlockDevices;
     int passThrough;
+    ZSTD_paramSwitch_e mmapDict;
 } FIO_prefs_t;
 
+typedef enum {FIO_mallocDict, FIO_mmapDict} FIO_dictBufferType_t;
+
+typedef struct {
+    void* dictBuffer;
+    size_t dictBufferSize;
+    FIO_dictBufferType_t dictBufferType;
+#if defined(_MSC_VER) || defined(_WIN32)
+    HANDLE dictHandle;
+#endif
+} FIO_Dict_t;
+
 #endif /* FILEIO_TYPES_HEADER */
diff --git a/contrib/libs/zstd/programs/timefn.c b/contrib/libs/zstd/programs/timefn.c
index f941e57e61..4f045226b8 100644
--- a/contrib/libs/zstd/programs/timefn.c
+++ b/contrib/libs/zstd/programs/timefn.c
@@ -88,7 +88,7 @@ UTIL_time_t UTIL_getTime(void)
 /* C11 requires support of timespec_get().
  * However, FreeBSD 11 claims C11 compliance while lacking timespec_get().
  * Double confirm timespec_get() support by checking the definition of TIME_UTC.
- * However, some versions of Android manage to simultanously define TIME_UTC
+ * However, some versions of Android manage to simultaneously define TIME_UTC
  * and lack timespec_get() support... */
 #elif (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */) \
     && defined(TIME_UTC) && !defined(__ANDROID__)
diff --git a/contrib/libs/zstd/programs/util.c b/contrib/libs/zstd/programs/util.c
index e017772ef6..c9031e91d3 100644
--- a/contrib/libs/zstd/programs/util.c
+++ b/contrib/libs/zstd/programs/util.c
@@ -102,6 +102,17 @@ UTIL_STATIC void* UTIL_realloc(void *ptr, size_t size)
     #define chmod _chmod
 #endif
 
+#ifndef ZSTD_HAVE_FCHMOD
+#if PLATFORM_POSIX_VERSION >= 199309L
+#define ZSTD_HAVE_FCHMOD
+#endif
+#endif
+
+#ifndef ZSTD_HAVE_FCHOWN
+#if PLATFORM_POSIX_VERSION >= 200809L
+#define ZSTD_HAVE_FCHOWN
+#endif
+#endif
 
 /*-****************************************
 *  Console log
@@ -147,21 +158,38 @@ void UTIL_traceFileStat(void)
     g_traceFileStat = 1;
 }
 
-int UTIL_stat(const char* filename, stat_t* statbuf)
+int UTIL_fstat(const int fd, const char* filename, stat_t* statbuf)
 {
     int ret;
-    UTIL_TRACE_CALL("UTIL_stat(%s)", filename);
+    UTIL_TRACE_CALL("UTIL_stat(%d, %s)", fd, filename);
 #if defined(_MSC_VER)
-    ret = !_stat64(filename, statbuf);
+    if (fd >= 0) {
+        ret = !_fstat64(fd, statbuf);
+    } else {
+        ret = !_stat64(filename, statbuf);
+    }
 #elif defined(__MINGW32__) && defined (__MSVCRT__)
-    ret = !_stati64(filename, statbuf);
+    if (fd >= 0) {
+        ret = !_fstati64(fd, statbuf);
+    } else {
+        ret = !_stati64(filename, statbuf);
+    }
 #else
-    ret = !stat(filename, statbuf);
+    if (fd >= 0) {
+        ret = !fstat(fd, statbuf);
+    } else {
+        ret = !stat(filename, statbuf);
+    }
 #endif
     UTIL_TRACE_RET(ret);
     return ret;
 }
 
+int UTIL_stat(const char* filename, stat_t* statbuf)
+{
+    return UTIL_fstat(-1, filename, statbuf);
+}
+
 int UTIL_isRegularFile(const char* infilename)
 {
     stat_t statbuf;
@@ -184,10 +212,15 @@ int UTIL_isRegularFileStat(const stat_t* statbuf)
 /* like chmod, but avoid changing permission of /dev/null */
 int UTIL_chmod(char const* filename, const stat_t* statbuf, mode_t permissions)
 {
+    return UTIL_fchmod(-1, filename, statbuf, permissions);
+}
+
+int UTIL_fchmod(const int fd, char const* filename, const stat_t* statbuf, mode_t permissions)
+{
     stat_t localStatBuf;
     UTIL_TRACE_CALL("UTIL_chmod(%s, %#4o)", filename, (unsigned)permissions);
     if (statbuf == NULL) {
-        if (!UTIL_stat(filename, &localStatBuf)) {
+        if (!UTIL_fstat(fd, filename, &localStatBuf)) {
             UTIL_TRACE_RET(0);
             return 0;
         }
@@ -197,9 +230,20 @@ int UTIL_chmod(char const* filename, const stat_t* statbuf, mode_t permissions)
         UTIL_TRACE_RET(0);
         return 0; /* pretend success, but don't change anything */
     }
-    UTIL_TRACE_CALL("chmod");
+#ifdef ZSTD_HAVE_FCHMOD
+    if (fd >= 0) {
+        int ret;
+        UTIL_TRACE_CALL("fchmod");
+        ret = fchmod(fd, permissions);
+        UTIL_TRACE_RET(ret);
+        UTIL_TRACE_RET(ret);
+        return ret;
+    } else
+#endif
     {
-        int const ret = chmod(filename, permissions);
+        int ret;
+        UTIL_TRACE_CALL("chmod");
+        ret = chmod(filename, permissions);
         UTIL_TRACE_RET(ret);
         UTIL_TRACE_RET(ret);
         return ret;
@@ -237,18 +281,20 @@ int UTIL_utime(const char* filename, const stat_t *statbuf)
 
 int UTIL_setFileStat(const char *filename, const stat_t *statbuf)
 {
+    return UTIL_setFDStat(-1, filename, statbuf);
+}
+
+int UTIL_setFDStat(const int fd, const char *filename, const stat_t *statbuf)
+{
     int res = 0;
     stat_t curStatBuf;
-    UTIL_TRACE_CALL("UTIL_setFileStat(%s)", filename);
+    UTIL_TRACE_CALL("UTIL_setFileStat(%d, %s)", fd, filename);
 
-    if (!UTIL_stat(filename, &curStatBuf) || !UTIL_isRegularFileStat(&curStatBuf)) {
+    if (!UTIL_fstat(fd, filename, &curStatBuf) || !UTIL_isRegularFileStat(&curStatBuf)) {
         UTIL_TRACE_RET(-1);
         return -1;
     }
 
-    /* set access and modification times */
-    res += UTIL_utime(filename, statbuf);
-
     /* Mimic gzip's behavior:
      *
      * "Change the group first, then the permissions, then the owner.
@@ -258,13 +304,27 @@ int UTIL_setFileStat(const char *filename, const stat_t *statbuf)
      * setgid bits." */
 
 #if !defined(_WIN32)
-    res += chown(filename, -1, statbuf->st_gid);  /* Apply group ownership */
+#ifdef ZSTD_HAVE_FCHOWN
+    if (fd >= 0) {
+        res += fchown(fd, -1, statbuf->st_gid);  /* Apply group ownership */
+    } else
+#endif
+    {
+        res += chown(filename, -1, statbuf->st_gid);  /* Apply group ownership */
+    }
 #endif
 
-    res += UTIL_chmod(filename, &curStatBuf, statbuf->st_mode & 0777);  /* Copy file permissions */
+    res += UTIL_fchmod(fd, filename, &curStatBuf, statbuf->st_mode & 0777);  /* Copy file permissions */
 
 #if !defined(_WIN32)
-    res += chown(filename, statbuf->st_uid, -1);  /* Apply user ownership */
+#ifdef ZSTD_HAVE_FCHOWN
+    if (fd >= 0) {
+        res += fchown(fd, statbuf->st_uid, -1);  /* Apply user ownership */
+    } else
+#endif
+    {
+        res += chown(filename, statbuf->st_uid, -1);  /* Apply user ownership */
+    }
 #endif
 
     errno = 0;
diff --git a/contrib/libs/zstd/programs/util.h b/contrib/libs/zstd/programs/util.h
index 4ec54137dd..8234646bf3 100644
--- a/contrib/libs/zstd/programs/util.h
+++ b/contrib/libs/zstd/programs/util.h
@@ -126,15 +126,25 @@ int UTIL_requireUserConfirmation(const char* prompt, const char* abortMsg, const
 /**
  * Calls platform's equivalent of stat() on filename and writes info to statbuf.
  * Returns success (1) or failure (0).
+ *
+ * UTIL_fstat() is like UTIL_stat() but takes an optional fd that refers to the
+ * file in question. It turns out that this can be meaningfully faster. If fd is
+ * -1, behaves just like UTIL_stat() (i.e., falls back to using the filename).
  */
 int UTIL_stat(const char* filename, stat_t* statbuf);
+int UTIL_fstat(const int fd, const char* filename, stat_t* statbuf);
 
 /**
  * Instead of getting a file's stats, this updates them with the info in the
  * provided stat_t. Currently sets owner, group, atime, and mtime. Will only
  * update this info for regular files.
+ *
+ * UTIL_setFDStat() also takes an fd, and will preferentially use that to
+ * indicate which file to modify, If fd is -1, it will fall back to using the
+ * filename.
  */
 int UTIL_setFileStat(const char* filename, const stat_t* statbuf);
+int UTIL_setFDStat(const int fd, const char* filename, const stat_t* statbuf);
 
 /**
  * Set atime to now and mtime to the st_mtim in statbuf.
@@ -159,8 +169,11 @@ U64 UTIL_getFileSizeStat(const stat_t* statbuf);
  * Like chmod(), but only modifies regular files. Provided statbuf may be NULL,
  * in which case this function will stat() the file internally, in order to
  * check whether it should be modified.
+ *
+ * If fd is -1, fd is ignored and the filename is used.
  */
 int UTIL_chmod(char const* filename, const stat_t* statbuf, mode_t permissions);
+int UTIL_fchmod(const int fd, char const* filename, const stat_t* statbuf, mode_t permissions);
 
 /*
  * In the absence of a pre-existing stat result on the file in question, these
diff --git a/contrib/libs/zstd/programs/zstdcli.c b/contrib/libs/zstd/programs/zstdcli.c
index 93f75e21d9..d2465456b5 100644
--- a/contrib/libs/zstd/programs/zstdcli.c
+++ b/contrib/libs/zstd/programs/zstdcli.c
@@ -37,7 +37,7 @@
 
 #include "fileio.h"   /* stdinmark, stdoutmark, ZSTD_EXTENSION */
 #ifndef ZSTD_NOBENCH
-#  include "benchzstd.h"  /* BMK_benchFiles */
+#  include "benchzstd.h"  /* BMK_benchFilesAdvanced */
 #endif
 #ifndef ZSTD_NODICT
 #  include "dibio.h"  /* ZDICT_cover_params_t, DiB_trainFromFiles() */
@@ -165,7 +165,7 @@ static void usage(FILE* f, const char* programName)
 #endif
     DISPLAY_F(f, "  -D DICT                       Use DICT as the dictionary for compression or decompression.\n\n");
     DISPLAY_F(f, "  -f, --force                   Disable input and output checks. Allows overwriting existing files,\n");
-    DISPLAY_F(f, "                                receiving input from the console, printing ouput to STDOUT, and\n");
+    DISPLAY_F(f, "                                receiving input from the console, printing output to STDOUT, and\n");
     DISPLAY_F(f, "                                operating on links, block devices, etc. Unrecognized formats will be\n");
     DISPLAY_F(f, "                                passed-through through as-is.\n\n");
 
@@ -254,6 +254,7 @@ static void usage_advanced(const char* programName)
 
     DISPLAYOUT("\n");
     DISPLAYOUT("  --format=zstd                 Compress files to the `.zst` format. [Default]\n");
+    DISPLAYOUT("  --mmap-dict                   Memory-map dictionary file rather than mallocing and loading all at once");
 #ifdef ZSTD_GZCOMPRESS
     DISPLAYOUT("  --format=gzip                 Compress files to the `.gz` format.\n");
 #endif
@@ -851,6 +852,7 @@ int main(int argCount, const char* argv[])
         ultra=0,
         contentSize=1,
         removeSrcFile=0;
+    ZSTD_paramSwitch_e mmapDict=ZSTD_ps_auto;
     ZSTD_paramSwitch_e useRowMatchFinder = ZSTD_ps_auto;
     FIO_compressionType_t cType = FIO_zstdCompression;
     unsigned nbWorkers = 0;
@@ -984,6 +986,8 @@ int main(int argCount, const char* argv[])
                 if (longCommandWArg(&argument, "--adapt=")) { adapt = 1; if (!parseAdaptParameters(argument, &adaptMin, &adaptMax)) { badusage(programName); CLEAN_RETURN(1); } continue; }
                 if (!strcmp(argument, "--single-thread")) { nbWorkers = 0; singleThread = 1; continue; }
                 if (!strcmp(argument, "--format=zstd")) { suffix = ZSTD_EXTENSION; cType = FIO_zstdCompression; continue; }
+                if (!strcmp(argument, "--mmap-dict")) { mmapDict = ZSTD_ps_enable; continue; }
+                if (!strcmp(argument, "--no-mmap-dict")) { mmapDict = ZSTD_ps_disable; continue; }
 #ifdef ZSTD_GZCOMPRESS
                 if (!strcmp(argument, "--format=gzip")) { suffix = GZ_EXTENSION; cType = FIO_gzipCompression; continue; }
                 if (exeNameMatch(programName, ZSTD_GZ)) {     /* behave like gzip */
@@ -1391,18 +1395,15 @@ int main(int argCount, const char* argv[])
                     int c;
                     DISPLAYLEVEL(3, "Benchmarking %s \n", filenames->fileNames[i]);
                     for(c = cLevel; c <= cLevelLast; c++) {
-                        BMK_benchOutcome_t const bo = BMK_benchFilesAdvanced(&filenames->fileNames[i], 1, dictFileName, c, &compressionParams, g_displayLevel, &benchParams);
-                        if (!BMK_isSuccessful_benchOutcome(bo)) return 1;
+                        operationResult = BMK_benchFilesAdvanced(&filenames->fileNames[i], 1, dictFileName, c, &compressionParams, g_displayLevel, &benchParams);
                 }   }
             } else {
                 for(; cLevel <= cLevelLast; cLevel++) {
-                    BMK_benchOutcome_t const bo = BMK_benchFilesAdvanced(filenames->fileNames, (unsigned)filenames->tableSize, dictFileName, cLevel, &compressionParams, g_displayLevel, &benchParams);
-                    if (!BMK_isSuccessful_benchOutcome(bo)) return 1;
+                    operationResult = BMK_benchFilesAdvanced(filenames->fileNames, (unsigned)filenames->tableSize, dictFileName, cLevel, &compressionParams, g_displayLevel, &benchParams);
             }   }
         } else {
             for(; cLevel <= cLevelLast; cLevel++) {
-                BMK_benchOutcome_t const bo = BMK_syntheticTest(cLevel, compressibility, &compressionParams, g_displayLevel, &benchParams);
-                if (!BMK_isSuccessful_benchOutcome(bo)) return 1;
+                operationResult = BMK_syntheticTest(cLevel, compressibility, &compressionParams, g_displayLevel, &benchParams);
         }   }
 
 #else
@@ -1526,6 +1527,7 @@ int main(int argCount, const char* argv[])
     FIO_setNotificationLevel(g_displayLevel);
     FIO_setAllowBlockDevices(prefs, allowBlockDevices);
     FIO_setPatchFromMode(prefs, patchFromDictFileName != NULL);
+    FIO_setMMapDict(prefs, mmapDict);
     if (memLimit == 0) {
         if (compressionParams.windowLog == 0) {
             memLimit = (U32)1 << g_defaultMaxWindowLog;