Update zstd to 1.5.6.

zstd/common/allocations.h

@@ -14,7 +14,7 @@
 #define ZSTD_DEPS_NEED_MALLOC
 #include "zstd_deps.h"   /* ZSTD_malloc, ZSTD_calloc, ZSTD_free, ZSTD_memset */
 
-#include "mem.h" /* MEM_STATIC */
+#include "compiler.h" /* MEM_STATIC */
 #define ZSTD_STATIC_LINKING_ONLY
 #include "../zstd.h" /* ZSTD_customMem */
 

zstd/common/bitstream.h

@@ -90,19 +90,20 @@ MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC);
 /*-********************************************
 *  bitStream decoding API (read backward)
 **********************************************/
+typedef size_t BitContainerType;
 typedef struct {
-    size_t   bitContainer;
+    BitContainerType bitContainer;
     unsigned bitsConsumed;
     const char* ptr;
     const char* start;
     const char* limitPtr;
 } BIT_DStream_t;
 
-typedef enum { BIT_DStream_unfinished = 0,
+typedef enum { BIT_DStream_unfinished = 0,  /* fully refilled */
-               BIT_DStream_endOfBuffer = 1,
+               BIT_DStream_endOfBuffer = 1, /* still some bits left in bitstream */
-               BIT_DStream_completed = 2,
+               BIT_DStream_completed = 2,   /* bitstream entirely consumed, bit-exact */
-               BIT_DStream_overflow = 3 } BIT_DStream_status;  /* result of BIT_reloadDStream() */
+               BIT_DStream_overflow = 3     /* user requested more bits than present in bitstream */
-               /* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */
+    } BIT_DStream_status;  /* result of BIT_reloadDStream() */
 
 MEM_STATIC size_t   BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize);
 MEM_STATIC size_t   BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits);
@@ -112,7 +113,7 @@ MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* bitD);
 
 /* Start by invoking BIT_initDStream().
 *  A chunk of the bitStream is then stored into a local register.
-*  Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
+*  Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (BitContainerType).
 *  You can then retrieve bitFields stored into the local register, **in reverse order**.
 *  Local register is explicitly reloaded from memory by the BIT_reloadDStream() method.
 *  A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished.
@@ -162,7 +163,7 @@ MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,
     return 0;
 }
 
-MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
+FORCE_INLINE_TEMPLATE size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
 {
 #if defined(STATIC_BMI2) && STATIC_BMI2 == 1 && !defined(ZSTD_NO_INTRINSICS)
     return  _bzhi_u64(bitContainer, nbBits);
@@ -267,22 +268,22 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
         bitD->bitContainer = *(const BYTE*)(bitD->start);
         switch(srcSize)
         {
-        case 7: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);
+        case 7: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);
                 ZSTD_FALLTHROUGH;
 
-        case 6: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
+        case 6: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
                 ZSTD_FALLTHROUGH;
 
-        case 5: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
+        case 5: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
                 ZSTD_FALLTHROUGH;
 
-        case 4: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[3]) << 24;
+        case 4: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[3]) << 24;
                 ZSTD_FALLTHROUGH;
 
-        case 3: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[2]) << 16;
+        case 3: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[2]) << 16;
                 ZSTD_FALLTHROUGH;
 
-        case 2: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[1]) <<  8;
+        case 2: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[1]) <<  8;
                 ZSTD_FALLTHROUGH;
 
         default: break;
@@ -297,12 +298,12 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
     return srcSize;
 }
 
-MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getUpperBits(size_t bitContainer, U32 const start)
+FORCE_INLINE_TEMPLATE size_t BIT_getUpperBits(BitContainerType bitContainer, U32 const start)
 {
     return bitContainer >> start;
 }
 
-MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits)
+FORCE_INLINE_TEMPLATE size_t BIT_getMiddleBits(BitContainerType bitContainer, U32 const start, U32 const nbBits)
 {
     U32 const regMask = sizeof(bitContainer)*8 - 1;
     /* if start > regMask, bitstream is corrupted, and result is undefined */
@@ -325,7 +326,7 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 c
  *  On 32-bits, maxNbBits==24.
  *  On 64-bits, maxNbBits==56.
  * @return : value extracted */
-MEM_STATIC  FORCE_INLINE_ATTR size_t BIT_lookBits(const BIT_DStream_t*  bitD, U32 nbBits)
+FORCE_INLINE_TEMPLATE size_t BIT_lookBits(const BIT_DStream_t*  bitD, U32 nbBits)
 {
     /* arbitrate between double-shift and shift+mask */
 #if 1
@@ -348,7 +349,7 @@ MEM_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits)
     return (bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> (((regMask+1)-nbBits) & regMask);
 }
 
-MEM_STATIC FORCE_INLINE_ATTR void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
+FORCE_INLINE_TEMPLATE void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
 {
     bitD->bitsConsumed += nbBits;
 }
@@ -357,7 +358,7 @@ MEM_STATIC FORCE_INLINE_ATTR void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
  *  Read (consume) next n bits from local register and update.
  *  Pay attention to not read more than nbBits contained into local register.
  * @return : extracted value. */
-MEM_STATIC FORCE_INLINE_ATTR size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
+FORCE_INLINE_TEMPLATE size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
 {
     size_t const value = BIT_lookBits(bitD, nbBits);
     BIT_skipBits(bitD, nbBits);
@@ -374,6 +375,21 @@ MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
     return value;
 }
 
+/*! BIT_reloadDStream_internal() :
+ *  Simple variant of BIT_reloadDStream(), with two conditions:
+ *  1. bitstream is valid : bitsConsumed <= sizeof(bitD->bitContainer)*8
+ *  2. look window is valid after shifted down : bitD->ptr >= bitD->start
+ */
+MEM_STATIC BIT_DStream_status BIT_reloadDStream_internal(BIT_DStream_t* bitD)
+{
+    assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
+    bitD->ptr -= bitD->bitsConsumed >> 3;
+    assert(bitD->ptr >= bitD->start);
+    bitD->bitsConsumed &= 7;
+    bitD->bitContainer = MEM_readLEST(bitD->ptr);
+    return BIT_DStream_unfinished;
+}
+
 /*! BIT_reloadDStreamFast() :
  *  Similar to BIT_reloadDStream(), but with two differences:
  *  1. bitsConsumed <= sizeof(bitD->bitContainer)*8 must hold!
@@ -384,31 +400,35 @@ MEM_STATIC BIT_DStream_status BIT_reloadDStreamFast(BIT_DStream_t* bitD)
 {
     if (UNLIKELY(bitD->ptr < bitD->limitPtr))
         return BIT_DStream_overflow;
-    assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
+    return BIT_reloadDStream_internal(bitD);
-    bitD->ptr -= bitD->bitsConsumed >> 3;
-    bitD->bitsConsumed &= 7;
-    bitD->bitContainer = MEM_readLEST(bitD->ptr);
-    return BIT_DStream_unfinished;
 }
 
 /*! BIT_reloadDStream() :
  *  Refill `bitD` from buffer previously set in BIT_initDStream() .
- *  This function is safe, it guarantees it will not read beyond src buffer.
+ *  This function is safe, it guarantees it will not never 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 FORCE_INLINE_ATTR BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
+FORCE_INLINE_TEMPLATE BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
 {
-    if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))  /* overflow detected, like end of stream */
+    /* note : once in overflow mode, a bitstream remains in this mode until it's reset */
+    if (UNLIKELY(bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))) {
+        static const BitContainerType zeroFilled = 0;
+        bitD->ptr = (const char*)&zeroFilled; /* aliasing is allowed for char */
+        /* overflow detected, erroneous scenario or end of stream: no update */
         return BIT_DStream_overflow;
+    }
+
+    assert(bitD->ptr >= bitD->start);
 
     if (bitD->ptr >= bitD->limitPtr) {
-        return BIT_reloadDStreamFast(bitD);
+        return BIT_reloadDStream_internal(bitD);
     }
     if (bitD->ptr == bitD->start) {
+        /* reached end of bitStream => no update */
         if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer;
         return BIT_DStream_completed;
     }
-    /* start < ptr < limitPtr */
+    /* start < ptr < limitPtr => cautious update */
     {   U32 nbBytes = bitD->bitsConsumed >> 3;
         BIT_DStream_status result = BIT_DStream_unfinished;
         if (bitD->ptr - nbBytes < bitD->start) {

zstd/common/compiler.h

@@ -11,6 +11,8 @@
 #ifndef ZSTD_COMPILER_H
 #define ZSTD_COMPILER_H
 
+#include <stddef.h>
+
 #include "portability_macros.h"
 
 /*-*******************************************************
@@ -51,12 +53,19 @@
 #  define WIN_CDECL
 #endif
 
+/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
+#if defined(__GNUC__)
+#  define UNUSED_ATTR __attribute__((unused))
+#else
+#  define UNUSED_ATTR
+#endif
+
 /**
  * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
  * parameters. They must be inlined for the compiler to eliminate the constant
  * branches.
  */
-#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
+#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR UNUSED_ATTR
 /**
  * HINT_INLINE is used to help the compiler generate better code. It is *not*
  * used for "templates", so it can be tweaked based on the compilers
@@ -71,14 +80,28 @@
 #if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5
 #  define HINT_INLINE static INLINE_KEYWORD
 #else
-#  define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR
+#  define HINT_INLINE FORCE_INLINE_TEMPLATE
 #endif
 
-/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
+/* "soft" inline :
+ * The compiler is free to select if it's a good idea to inline or not.
+ * The main objective is to silence compiler warnings
+ * when a defined function in included but not used.
+ *
+ * Note : this macro is prefixed `MEM_` because it used to be provided by `mem.h` unit.
+ * Updating the prefix is probably preferable, but requires a fairly large codemod,
+ * since this name is used everywhere.
+ */
+#ifndef MEM_STATIC  /* already defined in Linux Kernel mem.h */
 #if defined(__GNUC__)
-#  define UNUSED_ATTR __attribute__((unused))
+#  define MEM_STATIC static __inline UNUSED_ATTR
+#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
+#  define MEM_STATIC static inline
+#elif defined(_MSC_VER)
+#  define MEM_STATIC static __inline
 #else
-#  define UNUSED_ATTR
+#  define MEM_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
+#endif
 #endif
 
 /* force no inlining */
@@ -109,10 +132,10 @@
 /* prefetch
  * can be disabled, by declaring NO_PREFETCH build macro */
 #if defined(NO_PREFETCH)
-#  define PREFETCH_L1(ptr)  (void)(ptr)  /* disabled */
+#  define PREFETCH_L1(ptr)  do { (void)(ptr); } while (0)  /* disabled */
-#  define PREFETCH_L2(ptr)  (void)(ptr)  /* disabled */
+#  define PREFETCH_L2(ptr)  do { (void)(ptr); } while (0)  /* disabled */
 #else
-#  if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86))  /* _mm_prefetch() is not defined outside of x86/x64 */
+#  if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) && !defined(_M_ARM64EC)  /* _mm_prefetch() is not defined outside of x86/x64 */
 #    include <mmintrin.h>   /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */
 #    define PREFETCH_L1(ptr)  _mm_prefetch((const char*)(ptr), _MM_HINT_T0)
 #    define PREFETCH_L2(ptr)  _mm_prefetch((const char*)(ptr), _MM_HINT_T1)
@@ -120,24 +143,25 @@
 #    define PREFETCH_L1(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */)
 #    define PREFETCH_L2(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */)
 #  elif defined(__aarch64__)
-#    define PREFETCH_L1(ptr)  __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr)))
+#    define PREFETCH_L1(ptr)  do { __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr))); } while (0)
-#    define PREFETCH_L2(ptr)  __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr)))
+#    define PREFETCH_L2(ptr)  do { __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr))); } while (0)
 #  else
-#    define PREFETCH_L1(ptr) (void)(ptr)  /* disabled */
+#    define PREFETCH_L1(ptr) do { (void)(ptr); } while (0)  /* disabled */
-#    define PREFETCH_L2(ptr) (void)(ptr)  /* disabled */
+#    define PREFETCH_L2(ptr) do { (void)(ptr); } while (0)  /* disabled */
 #  endif
 #endif  /* NO_PREFETCH */
 
 #define CACHELINE_SIZE 64
 
-#define PREFETCH_AREA(p, s)  {            \
+#define PREFETCH_AREA(p, s)                              \
+    do {                                                 \
         const char* const _ptr = (const char*)(p);       \
         size_t const _size = (size_t)(s);                \
         size_t _pos;                                     \
         for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \
             PREFETCH_L2(_ptr + _pos);                    \
         }                                                \
-}
+    } while (0)
 
 /* vectorization
  * older GCC (pre gcc-4.3 picked as the cutoff) uses a different syntax,
@@ -166,9 +190,9 @@
 #endif
 
 #if __has_builtin(__builtin_unreachable) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 5)))
-#  define ZSTD_UNREACHABLE { assert(0), __builtin_unreachable(); }
+#  define ZSTD_UNREACHABLE do { assert(0), __builtin_unreachable(); } while (0)
 #else
-#  define ZSTD_UNREACHABLE { assert(0); }
+#  define ZSTD_UNREACHABLE do { assert(0); } while (0)
 #endif
 
 /* disable warnings */
@@ -281,6 +305,74 @@
 *  Sanitizer
 *****************************************************************/
 
+/**
+ * Zstd relies on pointer overflow in its decompressor.
+ * We add this attribute to functions that rely on pointer overflow.
+ */
+#ifndef ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+#  if __has_attribute(no_sanitize)
+#    if !defined(__clang__) && defined(__GNUC__) && __GNUC__ < 8
+       /* gcc < 8 only has signed-integer-overlow which triggers on pointer overflow */
+#      define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR __attribute__((no_sanitize("signed-integer-overflow")))
+#    else
+       /* older versions of clang [3.7, 5.0) will warn that pointer-overflow is ignored. */
+#      define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR __attribute__((no_sanitize("pointer-overflow")))
+#    endif
+#  else
+#    define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+#  endif
+#endif
+
+/**
+ * Helper function to perform a wrapped pointer difference without trigging
+ * UBSAN.
+ *
+ * @returns lhs - rhs with wrapping
+ */
+MEM_STATIC
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+ptrdiff_t ZSTD_wrappedPtrDiff(unsigned char const* lhs, unsigned char const* rhs)
+{
+    return lhs - rhs;
+}
+
+/**
+ * Helper function to perform a wrapped pointer add without triggering UBSAN.
+ *
+ * @return ptr + add with wrapping
+ */
+MEM_STATIC
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+unsigned char const* ZSTD_wrappedPtrAdd(unsigned char const* ptr, ptrdiff_t add)
+{
+    return ptr + add;
+}
+
+/**
+ * Helper function to perform a wrapped pointer subtraction without triggering
+ * UBSAN.
+ *
+ * @return ptr - sub with wrapping
+ */
+MEM_STATIC
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+unsigned char const* ZSTD_wrappedPtrSub(unsigned char const* ptr, ptrdiff_t sub)
+{
+    return ptr - sub;
+}
+
+/**
+ * Helper function to add to a pointer that works around C's undefined behavior
+ * of adding 0 to NULL.
+ *
+ * @returns `ptr + add` except it defines `NULL + 0 == NULL`.
+ */
+MEM_STATIC
+unsigned char* ZSTD_maybeNullPtrAdd(unsigned char* ptr, ptrdiff_t add)
+{
+    return add > 0 ? ptr + add : ptr;
+}
+
 /* Issue #3240 reports an ASAN failure on an llvm-mingw build. Out of an
  * abundance of caution, disable our custom poisoning on mingw. */
 #ifdef __MINGW32__

zstd/common/cpu.h

@@ -35,6 +35,7 @@ MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
     U32 f7b = 0;
     U32 f7c = 0;
 #if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86))
+#if !defined(__clang__)
     int reg[4];
     __cpuid((int*)reg, 0);
     {
@@ -50,6 +51,41 @@ MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
             f7c = (U32)reg[2];
         }
     }
+#else
+    /* Clang compiler has a bug (fixed in https://reviews.llvm.org/D101338) in
+     * which the `__cpuid` intrinsic does not save and restore `rbx` as it needs
+     * to due to being a reserved register. So in that case, do the `cpuid`
+     * ourselves. Clang supports inline assembly anyway.
+     */
+    U32 n;
+    __asm__(
+        "pushq %%rbx\n\t"
+        "cpuid\n\t"
+        "popq %%rbx\n\t"
+        : "=a"(n)
+        : "a"(0)
+        : "rcx", "rdx");
+    if (n >= 1) {
+      U32 f1a;
+      __asm__(
+          "pushq %%rbx\n\t"
+          "cpuid\n\t"
+          "popq %%rbx\n\t"
+          : "=a"(f1a), "=c"(f1c), "=d"(f1d)
+          : "a"(1)
+          :);
+    }
+    if (n >= 7) {
+      __asm__(
+          "pushq %%rbx\n\t"
+          "cpuid\n\t"
+          "movq %%rbx, %%rax\n\t"
+          "popq %%rbx"
+          : "=a"(f7b), "=c"(f7c)
+          : "a"(7), "c"(0)
+          : "rdx");
+    }
+#endif
 #elif defined(__i386__) && defined(__PIC__) && !defined(__clang__) && defined(__GNUC__)
     /* The following block like the normal cpuid branch below, but gcc
      * reserves ebx for use of its pic register so we must specially

zstd/common/debug.c

@@ -21,4 +21,10 @@
 
 #include "debug.h"
 
+#if !defined(ZSTD_LINUX_KERNEL) || (DEBUGLEVEL>=2)
+/* We only use this when DEBUGLEVEL>=2, but we get -Werror=pedantic errors if a
+ * translation unit is empty. So remove this from Linux kernel builds, but
+ * otherwise just leave it in.
+ */
 int g_debuglevel = DEBUGLEVEL;
+#endif

zstd/common/debug.h

@@ -85,18 +85,27 @@ extern int g_debuglevel; /* the variable is only declared,
                             It's useful when enabling very verbose levels
                             on selective conditions (such as position in src) */
 
-#  define RAWLOG(l, ...) {                                       \
+#  define RAWLOG(l, ...)                   \
+    do {                                   \
         if (l<=g_debuglevel) {             \
             ZSTD_DEBUG_PRINT(__VA_ARGS__); \
-            }   }
+        }                                  \
-#  define DEBUGLOG(l, ...) {                                     \
+    } while (0)
+
+#define STRINGIFY(x) #x
+#define TOSTRING(x) STRINGIFY(x)
+#define LINE_AS_STRING TOSTRING(__LINE__)
+
+#  define DEBUGLOG(l, ...)                               \
+    do {                                                 \
         if (l<=g_debuglevel) {                           \
-                    ZSTD_DEBUG_PRINT(__FILE__ ": " __VA_ARGS__); \
+            ZSTD_DEBUG_PRINT(__FILE__ ":" LINE_AS_STRING ": " __VA_ARGS__); \
             ZSTD_DEBUG_PRINT(" \n");                     \
-            }   }
+        }                                                \
+    } while (0)
 #else
-#  define RAWLOG(l, ...)      {}    /* disabled */
+#  define RAWLOG(l, ...)   do { } while (0)    /* disabled */
-#  define DEBUGLOG(l, ...)    {}    /* disabled */
+#  define DEBUGLOG(l, ...) do { } while (0)    /* disabled */
 #endif
 
 

zstd/common/error_private.h

@@ -60,8 +60,13 @@ ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); }
 ERR_STATIC ERR_enum ERR_getErrorCode(size_t code) { if (!ERR_isError(code)) return (ERR_enum)0; return (ERR_enum) (0-code); }
 
 /* check and forward error code */
-#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return e
+#define CHECK_V_F(e, f)     \
-#define CHECK_F(f)   { CHECK_V_F(_var_err__, f); }
+    size_t const e = f;     \
+    do {                    \
+        if (ERR_isError(e)) \
+            return e;       \
+    } while (0)
+#define CHECK_F(f)   do { CHECK_V_F(_var_err__, f); } while (0)
 
 
 /*-****************************************
@@ -96,9 +101,11 @@ void _force_has_format_string(const char *format, ...) {
  * we don't want to force runtime evaluation of its arguments.
  */
 #define _FORCE_HAS_FORMAT_STRING(...)              \
+    do {                                           \
         if (0) {                                   \
             _force_has_format_string(__VA_ARGS__); \
-  }
+        }                                          \
+    } while (0)
 
 #define ERR_QUOTE(str) #str
 
@@ -110,6 +117,7 @@ void _force_has_format_string(const char *format, ...) {
  * this can't just wrap RETURN_ERROR().
  */
 #define RETURN_ERROR_IF(cond, err, ...)                                        \
+    do {                                                                       \
         if (cond) {                                                            \
             RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s",          \
                   __FILE__, __LINE__, ERR_QUOTE(cond), ERR_QUOTE(ERROR(err))); \
@@ -117,7 +125,8 @@ void _force_has_format_string(const char *format, ...) {
             RAWLOG(3, ": " __VA_ARGS__);                                       \
             RAWLOG(3, "\n");                                                   \
             return ERROR(err);                                                 \
-  }
+        }                                                                      \
+    } while (0)
 
 /**
  * Unconditionally return the specified error.
@@ -132,7 +141,7 @@ void _force_has_format_string(const char *format, ...) {
         RAWLOG(3, ": " __VA_ARGS__);                                         \
         RAWLOG(3, "\n");                                                     \
         return ERROR(err);                                                   \
-  } while(0);
+    } while(0)
 
 /**
  * If the provided expression evaluates to an error code, returns that error code.
@@ -150,7 +159,7 @@ void _force_has_format_string(const char *format, ...) {
             RAWLOG(3, "\n");                                                       \
             return err_code;                                                       \
         }                                                                          \
-  } while(0);
+    } while(0)
 
 #if defined (__cplusplus)
 }

zstd/common/fse.h

@@ -229,6 +229,7 @@ If there is an error, the function will return an error code, which can be teste
 
 #endif  /* FSE_H */
 
+
 #if defined(FSE_STATIC_LINKING_ONLY) && !defined(FSE_H_FSE_STATIC_LINKING_ONLY)
 #define FSE_H_FSE_STATIC_LINKING_ONLY
 
@@ -464,13 +465,13 @@ MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, un
     FSE_symbolCompressionTransform const symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
     const U16* const stateTable = (const U16*)(statePtr->stateTable);
     U32 const nbBitsOut  = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);
-    BIT_addBits(bitC, statePtr->value, nbBitsOut);
+    BIT_addBits(bitC,  (size_t)statePtr->value, nbBitsOut);
     statePtr->value = stateTable[ (statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
 }
 
 MEM_STATIC void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* statePtr)
 {
-    BIT_addBits(bitC, statePtr->value, statePtr->stateLog);
+    BIT_addBits(bitC, (size_t)statePtr->value, statePtr->stateLog);
     BIT_flushBits(bitC);
 }
 

zstd/common/fse_decompress.c

@@ -22,8 +22,7 @@
 #define FSE_STATIC_LINKING_ONLY
 #include "fse.h"
 #include "error_private.h"
-#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"  /* ZSTD_memcpy */
-#include "zstd_deps.h"
 #include "bits.h"       /* ZSTD_highbit32 */
 
 
@@ -84,7 +83,7 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
                     symbolNext[s] = 1;
                 } else {
                     if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;
-                    symbolNext[s] = normalizedCounter[s];
+                    symbolNext[s] = (U16)normalizedCounter[s];
         }   }   }
         ZSTD_memcpy(dt, &DTableH, sizeof(DTableH));
     }
@@ -99,8 +98,7 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
          * all symbols have counts <= 8. We ensure we have 8 bytes at the end of
          * our buffer to handle the over-write.
          */
-        {
+        {   U64 const add = 0x0101010101010101ull;
-            U64 const add = 0x0101010101010101ull;
             size_t pos = 0;
             U64 sv = 0;
             U32 s;
@@ -111,9 +109,8 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
                 for (i = 8; i < n; i += 8) {
                     MEM_write64(spread + pos + i, sv);
                 }
-                pos += n;
+                pos += (size_t)n;
-            }
+        }   }
-        }
         /* Now we spread those positions across the table.
          * The benefit of doing it in two stages is that we avoid the
          * variable size inner loop, which caused lots of branch misses.
@@ -232,12 +229,12 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_usingDTable_generic(
             break;
     }   }
 
-    return op-ostart;
+    assert(op >= ostart);
+    return (size_t)(op-ostart);
 }
 
 typedef struct {
     short ncount[FSE_MAX_SYMBOL_VALUE + 1];
-    FSE_DTable dtable[1]; /* Dynamically sized */
 } FSE_DecompressWksp;
 
 
@@ -252,13 +249,18 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_wksp_body(
     unsigned tableLog;
     unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE;
     FSE_DecompressWksp* const wksp = (FSE_DecompressWksp*)workSpace;
+    size_t const dtablePos = sizeof(FSE_DecompressWksp) / sizeof(FSE_DTable);
+    FSE_DTable* const dtable = (FSE_DTable*)workSpace + dtablePos;
 
-    DEBUG_STATIC_ASSERT((FSE_MAX_SYMBOL_VALUE + 1) % 2 == 0);
+    FSE_STATIC_ASSERT((FSE_MAX_SYMBOL_VALUE + 1) % 2 == 0);
     if (wkspSize < sizeof(*wksp)) return ERROR(GENERIC);
 
+    /* correct offset to dtable depends on this property */
+    FSE_STATIC_ASSERT(sizeof(FSE_DecompressWksp) % sizeof(FSE_DTable) == 0);
+
     /* normal FSE decoding mode */
-    {
+    {   size_t const NCountLength =
-        size_t const NCountLength = FSE_readNCount_bmi2(wksp->ncount, &maxSymbolValue, &tableLog, istart, cSrcSize, bmi2);
+            FSE_readNCount_bmi2(wksp->ncount, &maxSymbolValue, &tableLog, istart, cSrcSize, bmi2);
         if (FSE_isError(NCountLength)) return NCountLength;
         if (tableLog > maxLog) return ERROR(tableLog_tooLarge);
         assert(NCountLength <= cSrcSize);
@@ -271,16 +273,16 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_wksp_body(
     workSpace = (BYTE*)workSpace + sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
     wkspSize -= sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
 
-    CHECK_F( FSE_buildDTable_internal(wksp->dtable, wksp->ncount, maxSymbolValue, tableLog, workSpace, wkspSize) );
+    CHECK_F( FSE_buildDTable_internal(dtable, wksp->ncount, maxSymbolValue, tableLog, workSpace, wkspSize) );
 
     {
-        const void* ptr = wksp->dtable;
+        const void* ptr = dtable;
         const FSE_DTableHeader* DTableH = (const FSE_DTableHeader*)ptr;
         const U32 fastMode = DTableH->fastMode;
 
         /* select fast mode (static) */
-        if (fastMode) return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 1);
+        if (fastMode) return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, dtable, 1);
-        return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 0);
+        return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, dtable, 0);
     }
 }
 

zstd/common/huf.h

@@ -197,9 +197,22 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void
 
 /** HUF_getNbBitsFromCTable() :
  *  Read nbBits from CTable symbolTable, for symbol `symbolValue` presumed <= HUF_SYMBOLVALUE_MAX
- *  Note 1 : is not inlined, as HUF_CElt definition is private */
+ *  Note 1 : If symbolValue > HUF_readCTableHeader(symbolTable).maxSymbolValue, returns 0
+ *  Note 2 : is not inlined, as HUF_CElt definition is private
+ */
 U32 HUF_getNbBitsFromCTable(const HUF_CElt* symbolTable, U32 symbolValue);
 
+typedef struct {
+    BYTE tableLog;
+    BYTE maxSymbolValue;
+    BYTE unused[sizeof(size_t) - 2];
+} HUF_CTableHeader;
+
+/** HUF_readCTableHeader() :
+ *  @returns The header from the CTable specifying the tableLog and the maxSymbolValue.
+ */
+HUF_CTableHeader HUF_readCTableHeader(HUF_CElt const* ctable);
+
 /*
  * HUF_decompress() does the following:
  * 1. select the decompression algorithm (X1, X2) based on pre-computed heuristics

zstd/common/mem.h

@@ -31,15 +31,6 @@ extern "C" {
 #   include <stdlib.h>  /* _byteswap_ulong */
 #   include <intrin.h>  /* _byteswap_* */
 #endif
-#if defined(__GNUC__)
-#  define MEM_STATIC static __inline __attribute__((unused))
-#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
-#  define MEM_STATIC static inline
-#elif defined(_MSC_VER)
-#  define MEM_STATIC static __inline
-#else
-#  define MEM_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
-#endif
 
 /*-**************************************************************
 *  Basic Types

zstd/common/pool.c

@@ -223,7 +223,7 @@ static int POOL_resize_internal(POOL_ctx* ctx, size_t numThreads)
     {   ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
         if (!threadPool) return 1;
         /* replace existing thread pool */
-        ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));
+        ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(ZSTD_pthread_t));
         ZSTD_customFree(ctx->threads, ctx->customMem);
         ctx->threads = threadPool;
         /* Initialize additional threads */

zstd/common/pool.h

@@ -47,7 +47,7 @@ void POOL_joinJobs(POOL_ctx* ctx);
 /*! POOL_resize() :
  *  Expands or shrinks pool's number of threads.
  *  This is more efficient than releasing + creating a new context,
- *  since it tries to preserve and re-use existing threads.
+ *  since it tries to preserve and reuse existing threads.
  * `numThreads` must be at least 1.
  * @return : 0 when resize was successful,
  *           !0 (typically 1) if there is an error.

zstd/common/portability_macros.h

@@ -68,6 +68,8 @@
 /* Mark the internal assembly functions as hidden  */
 #ifdef __ELF__
 # define ZSTD_HIDE_ASM_FUNCTION(func) .hidden func
+#elif defined(__APPLE__)
+# define ZSTD_HIDE_ASM_FUNCTION(func) .private_extern func
 #else
 # define ZSTD_HIDE_ASM_FUNCTION(func)
 #endif

zstd/common/threading.c

@@ -73,10 +73,12 @@ int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
     ZSTD_thread_params_t thread_param;
     (void)unused;
 
+    if (thread==NULL) return -1;
+    *thread = NULL;
+
     thread_param.start_routine = start_routine;
     thread_param.arg = arg;
     thread_param.initialized = 0;
-    *thread = NULL;
 
     /* Setup thread initialization synchronization */
     if(ZSTD_pthread_cond_init(&thread_param.initialized_cond, NULL)) {
@@ -91,7 +93,7 @@ int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
 
     /* Spawn thread */
     *thread = (HANDLE)_beginthreadex(NULL, 0, worker, &thread_param, 0, NULL);
-    if (!thread) {
+    if (*thread==NULL) {
         ZSTD_pthread_mutex_destroy(&thread_param.initialized_mutex);
         ZSTD_pthread_cond_destroy(&thread_param.initialized_cond);
         return errno;
@@ -137,6 +139,7 @@ int ZSTD_pthread_join(ZSTD_pthread_t thread)
 
 int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t const* attr)
 {
+    assert(mutex != NULL);
     *mutex = (pthread_mutex_t*)ZSTD_malloc(sizeof(pthread_mutex_t));
     if (!*mutex)
         return 1;
@@ -145,6 +148,7 @@ int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t con
 
 int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex)
 {
+    assert(mutex != NULL);
     if (!*mutex)
         return 0;
     {
@@ -156,6 +160,7 @@ int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex)
 
 int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const* attr)
 {
+    assert(cond != NULL);
     *cond = (pthread_cond_t*)ZSTD_malloc(sizeof(pthread_cond_t));
     if (!*cond)
         return 1;
@@ -164,6 +169,7 @@ int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const*
 
 int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond)
 {
+    assert(cond != NULL);
     if (!*cond)
         return 0;
     {

zstd/common/xxhash.c

@@ -1,10 +1,6 @@
 /*
- *  xxHash - Fast Hash algorithm
+ * xxHash - Extremely Fast Hash algorithm
- *  Copyright (c) Meta Platforms, Inc. and affiliates.
+ * Copyright (c) Yann Collet - Meta Platforms, Inc
- *
- *  You can contact the author at :
- *  - xxHash homepage: https://cyan4973.github.io/xxHash/
- *  - xxHash source repository : https://github.com/Cyan4973/xxHash
  *
  * 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
@@ -12,8 +8,6 @@
  * You may select, at your option, one of the above-listed licenses.
  */
 
-
-
 /*
  * xxhash.c instantiates functions defined in xxhash.h
  */

zstd/common/zstd_internal.h

@@ -178,7 +178,7 @@ static void ZSTD_copy8(void* dst, const void* src) {
     ZSTD_memcpy(dst, src, 8);
 #endif
 }
-#define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
+#define COPY8(d,s) do { ZSTD_copy8(d,s); d+=8; s+=8; } while (0)
 
 /* Need to use memmove here since the literal buffer can now be located within
    the dst buffer. In circumstances where the op "catches up" to where the
@@ -198,7 +198,7 @@ static void ZSTD_copy16(void* dst, const void* src) {
     ZSTD_memcpy(dst, copy16_buf, 16);
 #endif
 }
-#define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; }
+#define COPY16(d,s) do { ZSTD_copy16(d,s); d+=16; s+=16; } while (0)
 
 #define WILDCOPY_OVERLENGTH 32
 #define WILDCOPY_VECLEN 16
@@ -227,7 +227,7 @@ void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e
     if (ovtype == ZSTD_overlap_src_before_dst && diff < WILDCOPY_VECLEN) {
         /* Handle short offset copies. */
         do {
-            COPY8(op, ip)
+            COPY8(op, ip);
         } while (op < oend);
     } else {
         assert(diff >= WILDCOPY_VECLEN || diff <= -WILDCOPY_VECLEN);
@@ -366,13 +366,13 @@ typedef struct {
 
 /*! ZSTD_getcBlockSize() :
  *  Provides the size of compressed block from block header `src` */
-/* Used by: decompress, fullbench (does not get its definition from here) */
+/*  Used by: decompress, fullbench */
 size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
                           blockProperties_t* bpPtr);
 
 /*! ZSTD_decodeSeqHeaders() :
  *  decode sequence header from src */
-/* Used by: decompress, fullbench (does not get its definition from here) */
+/*  Used by: zstd_decompress_block, fullbench */
 size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
                        const void* src, size_t srcSize);
 

zstd/compress/fse_compress.c

@@ -25,7 +25,7 @@
 #include "../common/error_private.h"
 #define ZSTD_DEPS_NEED_MALLOC
 #define ZSTD_DEPS_NEED_MATH64
-#include "../common/zstd_deps.h"  /* ZSTD_malloc, ZSTD_free, ZSTD_memcpy, ZSTD_memset */
+#include "../common/zstd_deps.h"  /* ZSTD_memset */
 #include "../common/bits.h" /* ZSTD_highbit32 */
 
 
@@ -255,7 +255,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
     /* Init */
     remaining = tableSize+1;   /* +1 for extra accuracy */
     threshold = tableSize;
-    nbBits = tableLog+1;
+    nbBits = (int)tableLog+1;
 
     while ((symbol < alphabetSize) && (remaining>1)) {  /* stops at 1 */
         if (previousIs0) {
@@ -274,7 +274,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
             }
             while (symbol >= start+3) {
                 start+=3;
-                bitStream += 3 << bitCount;
+                bitStream += 3U << bitCount;
                 bitCount += 2;
             }
             bitStream += (symbol-start) << bitCount;
@@ -294,7 +294,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
             count++;   /* +1 for extra accuracy */
             if (count>=threshold)
                 count += max;   /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */
-            bitStream += count << bitCount;
+            bitStream += (U32)count << bitCount;
             bitCount  += nbBits;
             bitCount  -= (count<max);
             previousIs0  = (count==1);
@@ -322,7 +322,8 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
     out[1] = (BYTE)(bitStream>>8);
     out+= (bitCount+7) /8;
 
-    return (out-ostart);
+    assert(out >= ostart);
+    return (size_t)(out-ostart);
 }
 
 

zstd/compress/huf_compress.c

@@ -220,6 +220,25 @@ static void HUF_setValue(HUF_CElt* elt, size_t value)
     }
 }
 
+HUF_CTableHeader HUF_readCTableHeader(HUF_CElt const* ctable)
+{
+    HUF_CTableHeader header;
+    ZSTD_memcpy(&header, ctable, sizeof(header));
+    return header;
+}
+
+static void HUF_writeCTableHeader(HUF_CElt* ctable, U32 tableLog, U32 maxSymbolValue)
+{
+    HUF_CTableHeader header;
+    HUF_STATIC_ASSERT(sizeof(ctable[0]) == sizeof(header));
+    ZSTD_memset(&header, 0, sizeof(header));
+    assert(tableLog < 256);
+    header.tableLog = (BYTE)tableLog;
+    assert(maxSymbolValue < 256);
+    header.maxSymbolValue = (BYTE)maxSymbolValue;
+    ZSTD_memcpy(ctable, &header, sizeof(header));
+}
+
 typedef struct {
     HUF_CompressWeightsWksp wksp;
     BYTE bitsToWeight[HUF_TABLELOG_MAX + 1];   /* precomputed conversion table */
@@ -237,6 +256,9 @@ size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize,
 
     HUF_STATIC_ASSERT(HUF_CTABLE_WORKSPACE_SIZE >= sizeof(HUF_WriteCTableWksp));
 
+    assert(HUF_readCTableHeader(CTable).maxSymbolValue == maxSymbolValue);
+    assert(HUF_readCTableHeader(CTable).tableLog == huffLog);
+
     /* check conditions */
     if (workspaceSize < sizeof(HUF_WriteCTableWksp)) return ERROR(GENERIC);
     if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge);
@@ -283,7 +305,9 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void
     if (tableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);
     if (nbSymbols > *maxSymbolValuePtr+1) return ERROR(maxSymbolValue_tooSmall);
 
-    CTable[0] = tableLog;
+    *maxSymbolValuePtr = nbSymbols - 1;
+
+    HUF_writeCTableHeader(CTable, tableLog, *maxSymbolValuePtr);
 
     /* Prepare base value per rank */
     {   U32 n, nextRankStart = 0;
@@ -315,7 +339,6 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void
         { U32 n; for (n=0; n<nbSymbols; n++) HUF_setValue(ct + n, valPerRank[HUF_getNbBits(ct[n])]++); }
     }
 
-    *maxSymbolValuePtr = nbSymbols - 1;
     return readSize;
 }
 
@@ -323,6 +346,8 @@ U32 HUF_getNbBitsFromCTable(HUF_CElt const* CTable, U32 symbolValue)
 {
     const HUF_CElt* const ct = CTable + 1;
     assert(symbolValue <= HUF_SYMBOLVALUE_MAX);
+    if (symbolValue > HUF_readCTableHeader(CTable).maxSymbolValue)
+        return 0;
     return (U32)HUF_getNbBits(ct[symbolValue]);
 }
 
@@ -723,7 +748,8 @@ static void HUF_buildCTableFromTree(HUF_CElt* CTable, nodeElt const* huffNode, i
         HUF_setNbBits(ct + huffNode[n].byte, huffNode[n].nbBits);   /* push nbBits per symbol, symbol order */
     for (n=0; n<alphabetSize; n++)
         HUF_setValue(ct + n, valPerRank[HUF_getNbBits(ct[n])]++);   /* assign value within rank, symbol order */
-    CTable[0] = maxNbBits;
+
+    HUF_writeCTableHeader(CTable, maxNbBits, maxSymbolValue);
 }
 
 size_t
@@ -776,9 +802,16 @@ size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count,
 }
 
 int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue) {
+    HUF_CTableHeader header = HUF_readCTableHeader(CTable);
     HUF_CElt const* ct = CTable + 1;
     int bad = 0;
     int s;
+
+    assert(header.tableLog <= HUF_TABLELOG_ABSOLUTEMAX);
+
+    if (header.maxSymbolValue < maxSymbolValue)
+        return 0;
+
     for (s = 0; s <= (int)maxSymbolValue; ++s) {
         bad |= (count[s] != 0) & (HUF_getNbBits(ct[s]) == 0);
     }
@@ -1024,17 +1057,17 @@ HUF_compress1X_usingCTable_internal_body(void* dst, size_t dstSize,
                                    const void* src, size_t srcSize,
                                    const HUF_CElt* CTable)
 {
-    U32 const tableLog = (U32)CTable[0];
+    U32 const tableLog = HUF_readCTableHeader(CTable).tableLog;
     HUF_CElt const* ct = CTable + 1;
     const BYTE* ip = (const BYTE*) src;
     BYTE* const ostart = (BYTE*)dst;
     BYTE* const oend = ostart + dstSize;
-    BYTE* op = ostart;
     HUF_CStream_t bitC;
 
     /* init */
     if (dstSize < 8) return 0;   /* not enough space to compress */
-    { size_t const initErr = HUF_initCStream(&bitC, op, (size_t)(oend-op));
+    { BYTE* op = ostart;
+      size_t const initErr = HUF_initCStream(&bitC, op, (size_t)(oend-op));
       if (HUF_isError(initErr)) return 0; }
 
     if (dstSize < HUF_tightCompressBound(srcSize, (size_t)tableLog) || tableLog > 11)
@@ -1255,7 +1288,7 @@ unsigned HUF_optimalTableLog(
 
     {   BYTE* dst = (BYTE*)workSpace + sizeof(HUF_WriteCTableWksp);
         size_t dstSize = wkspSize - sizeof(HUF_WriteCTableWksp);
-        size_t maxBits, hSize, newSize;
+        size_t hSize, newSize;
         const unsigned symbolCardinality = HUF_cardinality(count, maxSymbolValue);
         const unsigned minTableLog = HUF_minTableLog(symbolCardinality);
         size_t optSize = ((size_t) ~0) - 1;
@@ -1266,12 +1299,14 @@ unsigned HUF_optimalTableLog(
         /* Search until size increases */
         for (optLogGuess = minTableLog; optLogGuess <= maxTableLog; optLogGuess++) {
             DEBUGLOG(7, "checking for huffLog=%u", optLogGuess);
-            maxBits = HUF_buildCTable_wksp(table, count, maxSymbolValue, optLogGuess, workSpace, wkspSize);
+
+            {   size_t maxBits = HUF_buildCTable_wksp(table, count, maxSymbolValue, optLogGuess, workSpace, wkspSize);
                 if (ERR_isError(maxBits)) continue;
 
                 if (maxBits < optLogGuess && optLogGuess > minTableLog) break;
 
                 hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits, workSpace, wkspSize);
+            }
 
             if (ERR_isError(hSize)) continue;
 
@@ -1372,12 +1407,6 @@ HUF_compress_internal (void* dst, size_t dstSize,
         huffLog = (U32)maxBits;
         DEBUGLOG(6, "bit distribution completed (%zu symbols)", showCTableBits(table->CTable + 1, maxSymbolValue+1));
     }
-    /* Zero unused symbols in CTable, so we can check it for validity */
-    {
-        size_t const ctableSize = HUF_CTABLE_SIZE_ST(maxSymbolValue);
-        size_t const unusedSize = sizeof(table->CTable) - ctableSize * sizeof(HUF_CElt);
-        ZSTD_memset(table->CTable + ctableSize, 0, unusedSize);
-    }
 
     /* Write table description header */
     {   CHECK_V_F(hSize, HUF_writeCTable_wksp(op, dstSize, table->CTable, maxSymbolValue, huffLog,
@@ -1420,7 +1449,7 @@ size_t HUF_compress1X_repeat (void* dst, size_t dstSize,
 /* HUF_compress4X_repeat():
  * compress input using 4 streams.
  * consider skipping quickly
- * re-use an existing huffman compression table */
+ * reuse an existing huffman compression table */
 size_t HUF_compress4X_repeat (void* dst, size_t dstSize,
                       const void* src, size_t srcSize,
                       unsigned maxSymbolValue, unsigned huffLog,

zstd/compress/zstd_compress.c

@@ -178,6 +178,7 @@ static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx)
 
 size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)
 {
+    DEBUGLOG(3, "ZSTD_freeCCtx (address: %p)", (void*)cctx);
     if (cctx==NULL) return 0;   /* support free on NULL */
     RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
                     "not compatible with static CCtx");
@@ -649,10 +650,11 @@ static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value)
     return 0;
 }
 
-#define BOUNDCHECK(cParam, val) { \
+#define BOUNDCHECK(cParam, val)                                       \
+    do {                                                              \
         RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val),        \
                         parameter_outOfBound, "Param out of bounds"); \
-}
+    } while (0)
 
 
 static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param)
@@ -868,7 +870,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
 #else
         FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value), "");
         CCtxParams->nbWorkers = value;
-        return CCtxParams->nbWorkers;
+        return (size_t)(CCtxParams->nbWorkers);
 #endif
 
     case ZSTD_c_jobSize :
@@ -892,7 +894,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
 #else
         FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value), "");
         CCtxParams->overlapLog = value;
-        return CCtxParams->overlapLog;
+        return (size_t)CCtxParams->overlapLog;
 #endif
 
     case ZSTD_c_rsyncable :
@@ -902,7 +904,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
 #else
         FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value), "");
         CCtxParams->rsyncable = value;
-        return CCtxParams->rsyncable;
+        return (size_t)CCtxParams->rsyncable;
 #endif
 
     case ZSTD_c_enableDedicatedDictSearch :
@@ -939,8 +941,10 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
         return CCtxParams->ldmParams.hashRateLog;
 
     case ZSTD_c_targetCBlockSize :
-        if (value!=0)   /* 0 ==> default */
+        if (value!=0) {  /* 0 ==> default */
+            value = MAX(value, ZSTD_TARGETCBLOCKSIZE_MIN);
             BOUNDCHECK(ZSTD_c_targetCBlockSize, value);
+        }
         CCtxParams->targetCBlockSize = (U32)value;
         return CCtxParams->targetCBlockSize;
 
@@ -968,7 +972,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
     case ZSTD_c_validateSequences:
         BOUNDCHECK(ZSTD_c_validateSequences, value);
         CCtxParams->validateSequences = value;
-        return CCtxParams->validateSequences;
+        return (size_t)CCtxParams->validateSequences;
 
     case ZSTD_c_useBlockSplitter:
         BOUNDCHECK(ZSTD_c_useBlockSplitter, value);
@@ -983,7 +987,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
     case ZSTD_c_deterministicRefPrefix:
         BOUNDCHECK(ZSTD_c_deterministicRefPrefix, value);
         CCtxParams->deterministicRefPrefix = !!value;
-        return CCtxParams->deterministicRefPrefix;
+        return (size_t)CCtxParams->deterministicRefPrefix;
 
     case ZSTD_c_prefetchCDictTables:
         BOUNDCHECK(ZSTD_c_prefetchCDictTables, value);
@@ -993,7 +997,7 @@ size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
     case ZSTD_c_enableSeqProducerFallback:
         BOUNDCHECK(ZSTD_c_enableSeqProducerFallback, value);
         CCtxParams->enableMatchFinderFallback = value;
-        return CCtxParams->enableMatchFinderFallback;
+        return (size_t)CCtxParams->enableMatchFinderFallback;
 
     case ZSTD_c_maxBlockSize:
         if (value!=0)    /* 0 ==> default */
@@ -1363,7 +1367,6 @@ size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset)
         RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong,
                         "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);
     }
     return 0;
@@ -1391,11 +1394,12 @@ size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
 static ZSTD_compressionParameters
 ZSTD_clampCParams(ZSTD_compressionParameters cParams)
 {
-#   define CLAMP_TYPE(cParam, val, type) {                                \
+#   define CLAMP_TYPE(cParam, val, type)                                      \
+        do {                                                                  \
             ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);         \
             if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound;      \
             else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \
-    }
+        } while (0)
 #   define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, unsigned)
     CLAMP(ZSTD_c_windowLog, cParams.windowLog);
     CLAMP(ZSTD_c_chainLog,  cParams.chainLog);
@@ -1467,6 +1471,48 @@ ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar,
     const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1);
     assert(ZSTD_checkCParams(cPar)==0);
 
+    /* Cascade the selected strategy down to the next-highest one built into
+     * this binary. */
+#ifdef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_btultra2) {
+        cPar.strategy = ZSTD_btultra;
+    }
+    if (cPar.strategy == ZSTD_btultra) {
+        cPar.strategy = ZSTD_btopt;
+    }
+#endif
+#ifdef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_btopt) {
+        cPar.strategy = ZSTD_btlazy2;
+    }
+#endif
+#ifdef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_btlazy2) {
+        cPar.strategy = ZSTD_lazy2;
+    }
+#endif
+#ifdef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_lazy2) {
+        cPar.strategy = ZSTD_lazy;
+    }
+#endif
+#ifdef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_lazy) {
+        cPar.strategy = ZSTD_greedy;
+    }
+#endif
+#ifdef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_greedy) {
+        cPar.strategy = ZSTD_dfast;
+    }
+#endif
+#ifdef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
+    if (cPar.strategy == ZSTD_dfast) {
+        cPar.strategy = ZSTD_fast;
+        cPar.targetLength = 0;
+    }
+#endif
+
     switch (mode) {
     case ZSTD_cpm_unknown:
     case ZSTD_cpm_noAttachDict:
@@ -1617,8 +1663,8 @@ ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams,
       + ZSTD_cwksp_aligned_alloc_size((MaxLL+1) * sizeof(U32))
       + ZSTD_cwksp_aligned_alloc_size((MaxOff+1) * sizeof(U32))
       + ZSTD_cwksp_aligned_alloc_size((1<<Litbits) * sizeof(U32))
-      + ZSTD_cwksp_aligned_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t))
+      + ZSTD_cwksp_aligned_alloc_size(ZSTD_OPT_SIZE * sizeof(ZSTD_match_t))
-      + ZSTD_cwksp_aligned_alloc_size((ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t));
+      + ZSTD_cwksp_aligned_alloc_size(ZSTD_OPT_SIZE * sizeof(ZSTD_optimal_t));
     size_t const lazyAdditionalSpace = ZSTD_rowMatchFinderUsed(cParams->strategy, useRowMatchFinder)
                                             ? ZSTD_cwksp_aligned_alloc_size(hSize)
                                             : 0;
@@ -1707,7 +1753,7 @@ size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params)
      * be needed. However, we still allocate two 0-sized buffers, which can
      * take space under ASAN. */
     return ZSTD_estimateCCtxSize_usingCCtxParams_internal(
-        &cParams, &params->ldmParams, 1, useRowMatchFinder, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN, params->useSequenceProducer, params->maxBlockSize);
+        &cParams, &params->ldmParams, 1, useRowMatchFinder, 0, 0, ZSTD_CONTENTSIZE_UNKNOWN, ZSTD_hasExtSeqProd(params), params->maxBlockSize);
 }
 
 size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams)
@@ -1768,7 +1814,7 @@ size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)
 
         return ZSTD_estimateCCtxSize_usingCCtxParams_internal(
             &cParams, &params->ldmParams, 1, useRowMatchFinder, inBuffSize, outBuffSize,
-            ZSTD_CONTENTSIZE_UNKNOWN, params->useSequenceProducer, params->maxBlockSize);
+            ZSTD_CONTENTSIZE_UNKNOWN, ZSTD_hasExtSeqProd(params), params->maxBlockSize);
     }
 }
 
@@ -2001,8 +2047,8 @@ ZSTD_reset_matchState(ZSTD_matchState_t* ms,
         ms->opt.litLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxLL+1) * sizeof(unsigned));
         ms->opt.matchLengthFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxML+1) * sizeof(unsigned));
         ms->opt.offCodeFreq = (unsigned*)ZSTD_cwksp_reserve_aligned(ws, (MaxOff+1) * sizeof(unsigned));
-        ms->opt.matchTable = (ZSTD_match_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_match_t));
+        ms->opt.matchTable = (ZSTD_match_t*)ZSTD_cwksp_reserve_aligned(ws, ZSTD_OPT_SIZE * sizeof(ZSTD_match_t));
-        ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, (ZSTD_OPT_NUM+1) * sizeof(ZSTD_optimal_t));
+        ms->opt.priceTable = (ZSTD_optimal_t*)ZSTD_cwksp_reserve_aligned(ws, ZSTD_OPT_SIZE * sizeof(ZSTD_optimal_t));
     }
 
     ms->cParams = *cParams;
@@ -2074,7 +2120,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
 
     {   size_t const windowSize = MAX(1, (size_t)MIN(((U64)1 << params->cParams.windowLog), pledgedSrcSize));
         size_t const blockSize = MIN(params->maxBlockSize, windowSize);
-        size_t const maxNbSeq = ZSTD_maxNbSeq(blockSize, params->cParams.minMatch, params->useSequenceProducer);
+        size_t const maxNbSeq = ZSTD_maxNbSeq(blockSize, params->cParams.minMatch, ZSTD_hasExtSeqProd(params));
         size_t const buffOutSize = (zbuff == ZSTDb_buffered && params->outBufferMode == ZSTD_bm_buffered)
                 ? ZSTD_compressBound(blockSize) + 1
                 : 0;
@@ -2091,8 +2137,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
         size_t const neededSpace =
             ZSTD_estimateCCtxSize_usingCCtxParams_internal(
                 &params->cParams, &params->ldmParams, zc->staticSize != 0, params->useRowMatchFinder,
-                buffInSize, buffOutSize, pledgedSrcSize, params->useSequenceProducer, params->maxBlockSize);
+                buffInSize, buffOutSize, pledgedSrcSize, ZSTD_hasExtSeqProd(params), params->maxBlockSize);
-        int resizeWorkspace;
 
         FORWARD_IF_ERROR(neededSpace, "cctx size estimate failed!");
 
@@ -2101,7 +2146,7 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
         {   /* Check if workspace is large enough, alloc a new one if needed */
             int const workspaceTooSmall = ZSTD_cwksp_sizeof(ws) < neededSpace;
             int const workspaceWasteful = ZSTD_cwksp_check_wasteful(ws, neededSpace);
-            resizeWorkspace = workspaceTooSmall || workspaceWasteful;
+            int resizeWorkspace = workspaceTooSmall || workspaceWasteful;
             DEBUGLOG(4, "Need %zu B workspace", neededSpace);
             DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize);
 
@@ -2176,10 +2221,10 @@ static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
         }
 
         /* reserve space for block-level external sequences */
-        if (params->useSequenceProducer) {
+        if (ZSTD_hasExtSeqProd(params)) {
             size_t const maxNbExternalSeq = ZSTD_sequenceBound(blockSize);
-            zc->externalMatchCtx.seqBufferCapacity = maxNbExternalSeq;
+            zc->extSeqBufCapacity = maxNbExternalSeq;
-            zc->externalMatchCtx.seqBuffer =
+            zc->extSeqBuf =
                 (ZSTD_Sequence*)ZSTD_cwksp_reserve_aligned(ws, maxNbExternalSeq * sizeof(ZSTD_Sequence));
         }
 
@@ -2564,7 +2609,7 @@ ZSTD_reduceTable_internal (U32* const table, U32 const size, U32 const reducerVa
     assert(size < (1U<<31));   /* can be casted to int */
 
 #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
+    /* To validate that the table reuse 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.
      *
@@ -2992,40 +3037,43 @@ ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_paramS
     static const ZSTD_blockCompressor blockCompressor[4][ZSTD_STRATEGY_MAX+1] = {
         { ZSTD_compressBlock_fast  /* default for 0 */,
           ZSTD_compressBlock_fast,
-          ZSTD_compressBlock_doubleFast,
+          ZSTD_COMPRESSBLOCK_DOUBLEFAST,
-          ZSTD_compressBlock_greedy,
+          ZSTD_COMPRESSBLOCK_GREEDY,
-          ZSTD_compressBlock_lazy,
+          ZSTD_COMPRESSBLOCK_LAZY,
-          ZSTD_compressBlock_lazy2,
+          ZSTD_COMPRESSBLOCK_LAZY2,
-          ZSTD_compressBlock_btlazy2,
+          ZSTD_COMPRESSBLOCK_BTLAZY2,
-          ZSTD_compressBlock_btopt,
+          ZSTD_COMPRESSBLOCK_BTOPT,
-          ZSTD_compressBlock_btultra,
+          ZSTD_COMPRESSBLOCK_BTULTRA,
-          ZSTD_compressBlock_btultra2 },
+          ZSTD_COMPRESSBLOCK_BTULTRA2
+        },
         { ZSTD_compressBlock_fast_extDict  /* default for 0 */,
           ZSTD_compressBlock_fast_extDict,
-          ZSTD_compressBlock_doubleFast_extDict,
+          ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT,
-          ZSTD_compressBlock_greedy_extDict,
+          ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT,
-          ZSTD_compressBlock_lazy_extDict,
+          ZSTD_COMPRESSBLOCK_LAZY_EXTDICT,
-          ZSTD_compressBlock_lazy2_extDict,
+          ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT,
-          ZSTD_compressBlock_btlazy2_extDict,
+          ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT,
-          ZSTD_compressBlock_btopt_extDict,
+          ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT,
-          ZSTD_compressBlock_btultra_extDict,
+          ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT,
-          ZSTD_compressBlock_btultra_extDict },
+          ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT
+        },
         { ZSTD_compressBlock_fast_dictMatchState  /* default for 0 */,
           ZSTD_compressBlock_fast_dictMatchState,
-          ZSTD_compressBlock_doubleFast_dictMatchState,
+          ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE,
-          ZSTD_compressBlock_greedy_dictMatchState,
+          ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE,
-          ZSTD_compressBlock_lazy_dictMatchState,
+          ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE,
-          ZSTD_compressBlock_lazy2_dictMatchState,
+          ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE,
-          ZSTD_compressBlock_btlazy2_dictMatchState,
+          ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE,
-          ZSTD_compressBlock_btopt_dictMatchState,
+          ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE,
-          ZSTD_compressBlock_btultra_dictMatchState,
+          ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE,
-          ZSTD_compressBlock_btultra_dictMatchState },
+          ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE
+        },
         { NULL  /* default for 0 */,
           NULL,
           NULL,
-          ZSTD_compressBlock_greedy_dedicatedDictSearch,
+          ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH,
-          ZSTD_compressBlock_lazy_dedicatedDictSearch,
+          ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH,
-          ZSTD_compressBlock_lazy2_dedicatedDictSearch,
+          ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH,
           NULL,
           NULL,
           NULL,
@@ -3038,18 +3086,26 @@ ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_paramS
     DEBUGLOG(4, "Selected block compressor: dictMode=%d strat=%d rowMatchfinder=%d", (int)dictMode, (int)strat, (int)useRowMatchFinder);
     if (ZSTD_rowMatchFinderUsed(strat, useRowMatchFinder)) {
         static const ZSTD_blockCompressor rowBasedBlockCompressors[4][3] = {
-            { ZSTD_compressBlock_greedy_row,
+            {
-            ZSTD_compressBlock_lazy_row,
+                ZSTD_COMPRESSBLOCK_GREEDY_ROW,
-            ZSTD_compressBlock_lazy2_row },
+                ZSTD_COMPRESSBLOCK_LAZY_ROW,
-            { ZSTD_compressBlock_greedy_extDict_row,
+                ZSTD_COMPRESSBLOCK_LAZY2_ROW
-            ZSTD_compressBlock_lazy_extDict_row,
+            },
-            ZSTD_compressBlock_lazy2_extDict_row },
+            {
-            { ZSTD_compressBlock_greedy_dictMatchState_row,
+                ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW,
-            ZSTD_compressBlock_lazy_dictMatchState_row,
+                ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW,
-            ZSTD_compressBlock_lazy2_dictMatchState_row },
+                ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW
-            { ZSTD_compressBlock_greedy_dedicatedDictSearch_row,
+            },
-            ZSTD_compressBlock_lazy_dedicatedDictSearch_row,
+            {
-            ZSTD_compressBlock_lazy2_dedicatedDictSearch_row }
+                ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW,
+                ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW,
+                ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW
+            },
+            {
+                ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW,
+                ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW,
+                ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW
+            }
         };
         DEBUGLOG(4, "Selecting a row-based matchfinder");
         assert(useRowMatchFinder != ZSTD_ps_auto);
@@ -3192,7 +3248,7 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
             /* External matchfinder + LDM is technically possible, just not implemented yet.
              * We need to revisit soon and implement it. */
             RETURN_ERROR_IF(
-                zc->appliedParams.useSequenceProducer,
+                ZSTD_hasExtSeqProd(&zc->appliedParams),
                 parameter_combination_unsupported,
                 "Long-distance matching with external sequence producer enabled is not currently supported."
             );
@@ -3211,7 +3267,7 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
             /* External matchfinder + LDM is technically possible, just not implemented yet.
              * We need to revisit soon and implement it. */
             RETURN_ERROR_IF(
-                zc->appliedParams.useSequenceProducer,
+                ZSTD_hasExtSeqProd(&zc->appliedParams),
                 parameter_combination_unsupported,
                 "Long-distance matching with external sequence producer enabled is not currently supported."
             );
@@ -3230,18 +3286,18 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
                                        zc->appliedParams.useRowMatchFinder,
                                        src, srcSize);
             assert(ldmSeqStore.pos == ldmSeqStore.size);
-        } else if (zc->appliedParams.useSequenceProducer) {
+        } else if (ZSTD_hasExtSeqProd(&zc->appliedParams)) {
             assert(
-                zc->externalMatchCtx.seqBufferCapacity >= ZSTD_sequenceBound(srcSize)
+                zc->extSeqBufCapacity >= ZSTD_sequenceBound(srcSize)
             );
-            assert(zc->externalMatchCtx.mFinder != NULL);
+            assert(zc->appliedParams.extSeqProdFunc != NULL);
 
             {   U32 const windowSize = (U32)1 << zc->appliedParams.cParams.windowLog;
 
-                size_t const nbExternalSeqs = (zc->externalMatchCtx.mFinder)(
+                size_t const nbExternalSeqs = (zc->appliedParams.extSeqProdFunc)(
-                    zc->externalMatchCtx.mState,
+                    zc->appliedParams.extSeqProdState,
-                    zc->externalMatchCtx.seqBuffer,
+                    zc->extSeqBuf,
-                    zc->externalMatchCtx.seqBufferCapacity,
+                    zc->extSeqBufCapacity,
                     src, srcSize,
                     NULL, 0,  /* dict and dictSize, currently not supported */
                     zc->appliedParams.compressionLevel,
@@ -3249,21 +3305,21 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
                 );
 
                 size_t const nbPostProcessedSeqs = ZSTD_postProcessSequenceProducerResult(
-                    zc->externalMatchCtx.seqBuffer,
+                    zc->extSeqBuf,
                     nbExternalSeqs,
-                    zc->externalMatchCtx.seqBufferCapacity,
+                    zc->extSeqBufCapacity,
                     srcSize
                 );
 
                 /* Return early if there is no error, since we don't need to worry about last literals */
                 if (!ZSTD_isError(nbPostProcessedSeqs)) {
                     ZSTD_sequencePosition seqPos = {0,0,0};
-                    size_t const seqLenSum = ZSTD_fastSequenceLengthSum(zc->externalMatchCtx.seqBuffer, nbPostProcessedSeqs);
+                    size_t const seqLenSum = ZSTD_fastSequenceLengthSum(zc->extSeqBuf, nbPostProcessedSeqs);
                     RETURN_ERROR_IF(seqLenSum > srcSize, externalSequences_invalid, "External sequences imply too large a block!");
                     FORWARD_IF_ERROR(
                         ZSTD_copySequencesToSeqStoreExplicitBlockDelim(
                             zc, &seqPos,
-                            zc->externalMatchCtx.seqBuffer, nbPostProcessedSeqs,
+                            zc->extSeqBuf, nbPostProcessedSeqs,
                             src, srcSize,
                             zc->appliedParams.searchForExternalRepcodes
                         ),
@@ -3280,7 +3336,9 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
                 }
 
                 /* Fallback to software matchfinder */
-                {   ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy,
+                {   ZSTD_blockCompressor const blockCompressor =
+                        ZSTD_selectBlockCompressor(
+                            zc->appliedParams.cParams.strategy,
                             zc->appliedParams.useRowMatchFinder,
                             dictMode);
                     ms->ldmSeqStore = NULL;
@@ -3292,7 +3350,8 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
                     lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize);
             }   }
         } else {   /* not long range mode and no external matchfinder */
-            ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy,
+            ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(
+                    zc->appliedParams.cParams.strategy,
                     zc->appliedParams.useRowMatchFinder,
                     dictMode);
             ms->ldmSeqStore = NULL;
@@ -3304,29 +3363,38 @@ static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
     return ZSTDbss_compress;
 }
 
-static void ZSTD_copyBlockSequences(ZSTD_CCtx* zc)
+static size_t ZSTD_copyBlockSequences(SeqCollector* seqCollector, const seqStore_t* seqStore, const U32 prevRepcodes[ZSTD_REP_NUM])
 {
-    const seqStore_t* seqStore = ZSTD_getSeqStore(zc);
+    const seqDef* inSeqs = seqStore->sequencesStart;
-    const seqDef* seqStoreSeqs = seqStore->sequencesStart;
+    const size_t nbInSequences = seqStore->sequences - inSeqs;
-    size_t seqStoreSeqSize = seqStore->sequences - seqStoreSeqs;
+    const size_t nbInLiterals = (size_t)(seqStore->lit - seqStore->litStart);
-    size_t seqStoreLiteralsSize = (size_t)(seqStore->lit - seqStore->litStart);
-    size_t literalsRead = 0;
-    size_t lastLLSize;
 
-    ZSTD_Sequence* outSeqs = &zc->seqCollector.seqStart[zc->seqCollector.seqIndex];
+    ZSTD_Sequence* outSeqs = seqCollector->seqIndex == 0 ? seqCollector->seqStart : seqCollector->seqStart + seqCollector->seqIndex;
+    const size_t nbOutSequences = nbInSequences + 1;
+    size_t nbOutLiterals = 0;
+    repcodes_t repcodes;
     size_t i;
-    repcodes_t updatedRepcodes;
 
-    assert(zc->seqCollector.seqIndex + 1 < zc->seqCollector.maxSequences);
+    /* Bounds check that we have enough space for every input sequence
-    /* Ensure we have enough space for last literals "sequence" */
+     * and the block delimiter
-    assert(zc->seqCollector.maxSequences >= seqStoreSeqSize + 1);
+     */
-    ZSTD_memcpy(updatedRepcodes.rep, zc->blockState.prevCBlock->rep, sizeof(repcodes_t));
+    assert(seqCollector->seqIndex <= seqCollector->maxSequences);
-    for (i = 0; i < seqStoreSeqSize; ++i) {
+    RETURN_ERROR_IF(
-        U32 rawOffset = seqStoreSeqs[i].offBase - ZSTD_REP_NUM;
+        nbOutSequences > (size_t)(seqCollector->maxSequences - seqCollector->seqIndex),
-        outSeqs[i].litLength = seqStoreSeqs[i].litLength;
+        dstSize_tooSmall,
-        outSeqs[i].matchLength = seqStoreSeqs[i].mlBase + MINMATCH;
+        "Not enough space to copy sequences");
+
+    ZSTD_memcpy(&repcodes, prevRepcodes, sizeof(repcodes));
+    for (i = 0; i < nbInSequences; ++i) {
+        U32 rawOffset;
+        outSeqs[i].litLength = inSeqs[i].litLength;
+        outSeqs[i].matchLength = inSeqs[i].mlBase + MINMATCH;
         outSeqs[i].rep = 0;
 
+        /* Handle the possible single length >= 64K
+         * There can only be one because we add MINMATCH to every match length,
+         * and blocks are at most 128K.
+         */
         if (i == seqStore->longLengthPos) {
             if (seqStore->longLengthType == ZSTD_llt_literalLength) {
                 outSeqs[i].litLength += 0x10000;
@@ -3335,41 +3403,55 @@ static void ZSTD_copyBlockSequences(ZSTD_CCtx* zc)
             }
         }
 
-        if (seqStoreSeqs[i].offBase <= ZSTD_REP_NUM) {
+        /* Determine the raw offset given the offBase, which may be a repcode. */
-            /* Derive the correct offset corresponding to a repcode */
+        if (OFFBASE_IS_REPCODE(inSeqs[i].offBase)) {
-            outSeqs[i].rep = seqStoreSeqs[i].offBase;
+            const U32 repcode = OFFBASE_TO_REPCODE(inSeqs[i].offBase);
+            assert(repcode > 0);
+            outSeqs[i].rep = repcode;
             if (outSeqs[i].litLength != 0) {
-                rawOffset = updatedRepcodes.rep[outSeqs[i].rep - 1];
+                rawOffset = repcodes.rep[repcode - 1];
             } else {
-                if (outSeqs[i].rep == 3) {
+                if (repcode == 3) {
-                    rawOffset = updatedRepcodes.rep[0] - 1;
+                    assert(repcodes.rep[0] > 1);
+                    rawOffset = repcodes.rep[0] - 1;
                 } else {
-                    rawOffset = updatedRepcodes.rep[outSeqs[i].rep];
+                    rawOffset = repcodes.rep[repcode];
                 }
             }
+        } else {
+            rawOffset = OFFBASE_TO_OFFSET(inSeqs[i].offBase);
         }
         outSeqs[i].offset = rawOffset;
-        /* seqStoreSeqs[i].offset == offCode+1, and ZSTD_updateRep() expects offCode
+
-           so we provide seqStoreSeqs[i].offset - 1 */
+        /* Update repcode history for the sequence */
-        ZSTD_updateRep(updatedRepcodes.rep,
+        ZSTD_updateRep(repcodes.rep,
-                       seqStoreSeqs[i].offBase,
+                       inSeqs[i].offBase,
-                       seqStoreSeqs[i].litLength == 0);
+                       inSeqs[i].litLength == 0);
-        literalsRead += outSeqs[i].litLength;
+
+        nbOutLiterals += outSeqs[i].litLength;
     }
     /* Insert last literals (if any exist) in the block as a sequence with ml == off == 0.
      * If there are no last literals, then we'll emit (of: 0, ml: 0, ll: 0), which is a marker
      * for the block boundary, according to the API.
      */
-    assert(seqStoreLiteralsSize >= literalsRead);
+    assert(nbInLiterals >= nbOutLiterals);
-    lastLLSize = seqStoreLiteralsSize - literalsRead;
+    {
-    outSeqs[i].litLength = (U32)lastLLSize;
+        const size_t lastLLSize = nbInLiterals - nbOutLiterals;
-    outSeqs[i].matchLength = outSeqs[i].offset = outSeqs[i].rep = 0;
+        outSeqs[nbInSequences].litLength = (U32)lastLLSize;
-    seqStoreSeqSize++;
+        outSeqs[nbInSequences].matchLength = 0;
-    zc->seqCollector.seqIndex += seqStoreSeqSize;
+        outSeqs[nbInSequences].offset = 0;
+        assert(nbOutSequences == nbInSequences + 1);
+    }
+    seqCollector->seqIndex += nbOutSequences;
+    assert(seqCollector->seqIndex <= seqCollector->maxSequences);
+
+    return 0;
 }
 
 size_t ZSTD_sequenceBound(size_t srcSize) {
-    return (srcSize / ZSTD_MINMATCH_MIN) + 1;
+    const size_t maxNbSeq = (srcSize / ZSTD_MINMATCH_MIN) + 1;
+    const size_t maxNbDelims = (srcSize / ZSTD_BLOCKSIZE_MAX_MIN) + 1;
+    return maxNbSeq + maxNbDelims;
 }
 
 size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs,
@@ -3378,6 +3460,16 @@ size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs,
     const size_t dstCapacity = ZSTD_compressBound(srcSize);
     void* dst = ZSTD_customMalloc(dstCapacity, ZSTD_defaultCMem);
     SeqCollector seqCollector;
+    {
+        int targetCBlockSize;
+        FORWARD_IF_ERROR(ZSTD_CCtx_getParameter(zc, ZSTD_c_targetCBlockSize, &targetCBlockSize), "");
+        RETURN_ERROR_IF(targetCBlockSize != 0, parameter_unsupported, "targetCBlockSize != 0");
+    }
+    {
+        int nbWorkers;
+        FORWARD_IF_ERROR(ZSTD_CCtx_getParameter(zc, ZSTD_c_nbWorkers, &nbWorkers), "");
+        RETURN_ERROR_IF(nbWorkers != 0, parameter_unsupported, "nbWorkers != 0");
+    }
 
     RETURN_ERROR_IF(dst == NULL, memory_allocation, "NULL pointer!");
 
@@ -3387,8 +3479,12 @@ size_t ZSTD_generateSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs,
     seqCollector.maxSequences = outSeqsSize;
     zc->seqCollector = seqCollector;
 
-    ZSTD_compress2(zc, dst, dstCapacity, src, srcSize);
+    {
+        const size_t ret = ZSTD_compress2(zc, dst, dstCapacity, src, srcSize);
         ZSTD_customFree(dst, ZSTD_defaultCMem);
+        FORWARD_IF_ERROR(ret, "ZSTD_compress2 failed");
+    }
+    assert(zc->seqCollector.seqIndex <= ZSTD_sequenceBound(srcSize));
     return zc->seqCollector.seqIndex;
 }
 
@@ -3981,8 +4077,9 @@ ZSTD_compressSeqStore_singleBlock(ZSTD_CCtx* zc,
         cSeqsSize = 1;
     }
 
+    /* Sequence collection not supported when block splitting */
     if (zc->seqCollector.collectSequences) {
-        ZSTD_copyBlockSequences(zc);
+        FORWARD_IF_ERROR(ZSTD_copyBlockSequences(&zc->seqCollector, seqStore, dRepOriginal.rep), "copyBlockSequences failed");
         ZSTD_blockState_confirmRepcodesAndEntropyTables(&zc->blockState);
         return 0;
     }
@@ -4204,6 +4301,7 @@ ZSTD_compressBlock_splitBlock(ZSTD_CCtx* zc,
         if (bss == ZSTDbss_noCompress) {
             if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid)
                 zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check;
+            RETURN_ERROR_IF(zc->seqCollector.collectSequences, sequenceProducer_failed, "Uncompressible block");
             cSize = ZSTD_noCompressBlock(dst, dstCapacity, src, srcSize, lastBlock);
             FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed");
             DEBUGLOG(4, "ZSTD_compressBlock_splitBlock: Nocompress block");
@@ -4236,11 +4334,15 @@ ZSTD_compressBlock_internal(ZSTD_CCtx* zc,
 
     {   const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize);
         FORWARD_IF_ERROR(bss, "ZSTD_buildSeqStore failed");
-        if (bss == ZSTDbss_noCompress) { cSize = 0; goto out; }
+        if (bss == ZSTDbss_noCompress) {
+            RETURN_ERROR_IF(zc->seqCollector.collectSequences, sequenceProducer_failed, "Uncompressible block");
+            cSize = 0;
+            goto out;
+        }
     }
 
     if (zc->seqCollector.collectSequences) {
-        ZSTD_copyBlockSequences(zc);
+        FORWARD_IF_ERROR(ZSTD_copyBlockSequences(&zc->seqCollector, ZSTD_getSeqStore(zc), zc->blockState.prevCBlock->rep), "copyBlockSequences failed");
         ZSTD_blockState_confirmRepcodesAndEntropyTables(&zc->blockState);
         return 0;
     }
@@ -4553,19 +4655,15 @@ size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity)
     }
 }
 
-size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq)
+void ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq)
 {
-    RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong,
+    assert(cctx->stage == ZSTDcs_init);
-                    "wrong cctx stage");
+    assert(nbSeq == 0 || cctx->appliedParams.ldmParams.enableLdm != ZSTD_ps_enable);
-    RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm == ZSTD_ps_enable,
-                    parameter_unsupported,
-                    "incompatible with ldm");
     cctx->externSeqStore.seq = seq;
     cctx->externSeqStore.size = nbSeq;
     cctx->externSeqStore.capacity = nbSeq;
     cctx->externSeqStore.pos = 0;
     cctx->externSeqStore.posInSequence = 0;
-    return 0;
 }
 
 
@@ -4760,12 +4858,19 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,
         ZSTD_fillHashTable(ms, iend, dtlm, tfp);
         break;
     case ZSTD_dfast:
+#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
         ZSTD_fillDoubleHashTable(ms, iend, dtlm, tfp);
+#else
+        assert(0); /* shouldn't be called: cparams should've been adjusted. */
+#endif
         break;
 
     case ZSTD_greedy:
     case ZSTD_lazy:
     case ZSTD_lazy2:
+#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR)
         assert(srcSize >= HASH_READ_SIZE);
         if (ms->dedicatedDictSearch) {
             assert(ms->chainTable != NULL);
@@ -4782,14 +4887,23 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,
                 DEBUGLOG(4, "Using chain-based hash table for lazy dict");
             }
         }
+#else
+        assert(0); /* shouldn't be called: cparams should've been adjusted. */
+#endif
         break;
 
     case ZSTD_btlazy2:   /* we want the dictionary table fully sorted */
     case ZSTD_btopt:
     case ZSTD_btultra:
     case ZSTD_btultra2:
+#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR)
         assert(srcSize >= HASH_READ_SIZE);
         ZSTD_updateTree(ms, iend-HASH_READ_SIZE, iend);
+#else
+        assert(0); /* shouldn't be called: cparams should've been adjusted. */
+#endif
         break;
 
     default:
@@ -4836,11 +4950,10 @@ size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace,
 
         /* We only set the loaded table as valid if it contains all non-zero
          * weights. Otherwise, we set it to check */
-        if (!hasZeroWeights)
+        if (!hasZeroWeights && maxSymbolValue == 255)
             bs->entropy.huf.repeatMode = HUF_repeat_valid;
 
         RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted, "");
-        RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted, "");
         dictPtr += hufHeaderSize;
     }
 
@@ -5107,14 +5220,13 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity)
 {
     BYTE* const ostart = (BYTE*)dst;
     BYTE* op = ostart;
-    size_t fhSize = 0;
 
     DEBUGLOG(4, "ZSTD_writeEpilogue");
     RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing");
 
     /* special case : empty frame */
     if (cctx->stage == ZSTDcs_init) {
-        fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, 0, 0);
+        size_t fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, &cctx->appliedParams, 0, 0);
         FORWARD_IF_ERROR(fhSize, "ZSTD_writeFrameHeader failed");
         dstCapacity -= fhSize;
         op += fhSize;
@@ -5124,8 +5236,9 @@ static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity)
     if (cctx->stage != ZSTDcs_ending) {
         /* write one last empty block, make it the "last" block */
         U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0;
-        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall, "no room for epilogue");
+        ZSTD_STATIC_ASSERT(ZSTD_BLOCKHEADERSIZE == 3);
-        MEM_writeLE32(op, cBlockHeader24);
+        RETURN_ERROR_IF(dstCapacity<3, dstSize_tooSmall, "no room for epilogue");
+        MEM_writeLE24(op, cBlockHeader24);
         op += ZSTD_blockHeaderSize;
         dstCapacity -= ZSTD_blockHeaderSize;
     }
@@ -5455,7 +5568,7 @@ ZSTD_CDict* ZSTD_createCDict_advanced2(
                         cctxParams.useRowMatchFinder, cctxParams.enableDedicatedDictSearch,
                         customMem);
 
-    if (ZSTD_isError( ZSTD_initCDict_internal(cdict,
+    if (!cdict || ZSTD_isError( ZSTD_initCDict_internal(cdict,
                                     dict, dictSize,
                                     dictLoadMethod, dictContentType,
                                     cctxParams) )) {
@@ -5879,7 +5992,7 @@ static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
     if (zcs->appliedParams.inBufferMode == ZSTD_bm_stable) {
         assert(input->pos >= zcs->stableIn_notConsumed);
         input->pos -= zcs->stableIn_notConsumed;
-        ip -= zcs->stableIn_notConsumed;
+        if (ip) ip -= zcs->stableIn_notConsumed;
         zcs->stableIn_notConsumed = 0;
     }
     if (zcs->appliedParams.inBufferMode == ZSTD_bm_buffered) {
@@ -6138,7 +6251,7 @@ static size_t ZSTD_CCtx_init_compressStream2(ZSTD_CCtx* cctx,
 #ifdef ZSTD_MULTITHREAD
     /* If external matchfinder is enabled, make sure to fail before checking job size (for consistency) */
     RETURN_ERROR_IF(
-        params.useSequenceProducer == 1 && params.nbWorkers >= 1,
+        ZSTD_hasExtSeqProd(&params) && params.nbWorkers >= 1,
         parameter_combination_unsupported,
         "External sequence producer isn't supported with nbWorkers >= 1"
     );
@@ -6430,7 +6543,7 @@ ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx,
         if (cctx->appliedParams.validateSequences) {
             seqPos->posInSrc += litLength + matchLength;
             FORWARD_IF_ERROR(ZSTD_validateSequence(offBase, matchLength, cctx->appliedParams.cParams.minMatch, seqPos->posInSrc,
-                                                cctx->appliedParams.cParams.windowLog, dictSize, cctx->appliedParams.useSequenceProducer),
+                                                cctx->appliedParams.cParams.windowLog, dictSize, ZSTD_hasExtSeqProd(&cctx->appliedParams)),
                                                 "Sequence validation failed");
         }
         RETURN_ERROR_IF(idx - seqPos->idx >= cctx->seqStore.maxNbSeq, externalSequences_invalid,
@@ -6568,7 +6681,7 @@ ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition*
         if (cctx->appliedParams.validateSequences) {
             seqPos->posInSrc += litLength + matchLength;
             FORWARD_IF_ERROR(ZSTD_validateSequence(offBase, matchLength, cctx->appliedParams.cParams.minMatch, seqPos->posInSrc,
-                                                   cctx->appliedParams.cParams.windowLog, dictSize, cctx->appliedParams.useSequenceProducer),
+                                                   cctx->appliedParams.cParams.windowLog, dictSize, ZSTD_hasExtSeqProd(&cctx->appliedParams)),
                                                    "Sequence validation failed");
         }
         DEBUGLOG(6, "Storing sequence: (of: %u, ml: %u, ll: %u)", offBase, matchLength, litLength);
@@ -7014,19 +7127,27 @@ ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeH
 }
 
 void ZSTD_registerSequenceProducer(
-    ZSTD_CCtx* zc, void* mState,
+    ZSTD_CCtx* zc,
-    ZSTD_sequenceProducer_F* mFinder
+    void* extSeqProdState,
+    ZSTD_sequenceProducer_F extSeqProdFunc
 ) {
-    if (mFinder != NULL) {
+    assert(zc != NULL);
-        ZSTD_externalMatchCtx emctx;
+    ZSTD_CCtxParams_registerSequenceProducer(
-        emctx.mState = mState;
+        &zc->requestedParams, extSeqProdState, extSeqProdFunc
-        emctx.mFinder = mFinder;
+    );
-        emctx.seqBuffer = NULL;
+}
-        emctx.seqBufferCapacity = 0;
+
-        zc->externalMatchCtx = emctx;
+void ZSTD_CCtxParams_registerSequenceProducer(
-        zc->requestedParams.useSequenceProducer = 1;
+  ZSTD_CCtx_params* params,
+  void* extSeqProdState,
+  ZSTD_sequenceProducer_F extSeqProdFunc
+) {
+    assert(params != NULL);
+    if (extSeqProdFunc != NULL) {
+        params->extSeqProdFunc = extSeqProdFunc;
+        params->extSeqProdState = extSeqProdState;
     } else {
-        ZSTD_memset(&zc->externalMatchCtx, 0, sizeof(zc->externalMatchCtx));
+        params->extSeqProdFunc = NULL;
-        zc->requestedParams.useSequenceProducer = 0;
+        params->extSeqProdState = NULL;
     }
 }

zstd/compress/zstd_compress_internal.h

@@ -39,7 +39,7 @@ extern "C" {
                                        It's not a big deal though : candidate will just be sorted again.
                                        Additionally, candidate position 1 will be lost.
                                        But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss.
-                                       The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy.
+                                       The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table reuse with a different strategy.
                                        This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */
 
 
@@ -159,23 +159,24 @@ typedef struct {
 UNUSED_ATTR static const rawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0, 0};
 
 typedef struct {
-    int price;
+    int price;  /* price from beginning of segment to this position */
-    U32 off;
+    U32 off;    /* offset of previous match */
-    U32 mlen;
+    U32 mlen;   /* length of previous match */
-    U32 litlen;
+    U32 litlen; /* nb of literals since previous match */
-    U32 rep[ZSTD_REP_NUM];
+    U32 rep[ZSTD_REP_NUM];  /* offset history after previous match */
 } ZSTD_optimal_t;
 
 typedef enum { zop_dynamic=0, zop_predef } ZSTD_OptPrice_e;
 
+#define ZSTD_OPT_SIZE (ZSTD_OPT_NUM+3)
 typedef struct {
     /* All tables are allocated inside cctx->workspace by ZSTD_resetCCtx_internal() */
     unsigned* litFreq;           /* table of literals statistics, of size 256 */
     unsigned* litLengthFreq;     /* table of litLength statistics, of size (MaxLL+1) */
     unsigned* matchLengthFreq;   /* table of matchLength statistics, of size (MaxML+1) */
     unsigned* offCodeFreq;       /* table of offCode statistics, of size (MaxOff+1) */
-    ZSTD_match_t* matchTable;    /* list of found matches, of size ZSTD_OPT_NUM+1 */
+    ZSTD_match_t* matchTable;    /* list of found matches, of size ZSTD_OPT_SIZE */
-    ZSTD_optimal_t* priceTable;  /* All positions tracked by optimal parser, of size ZSTD_OPT_NUM+1 */
+    ZSTD_optimal_t* priceTable;  /* All positions tracked by optimal parser, of size ZSTD_OPT_SIZE */
 
     U32  litSum;                 /* nb of literals */
     U32  litLengthSum;           /* nb of litLength codes */
@@ -228,7 +229,7 @@ struct ZSTD_matchState_t {
     U32 rowHashLog;                          /* For row-based matchfinder: Hashlog based on nb of rows in the hashTable.*/
     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 */
+    U64 hashSalt;                            /* For row-based matchFinder: salts the hash for reuse of tag table */
     U32 hashSaltEntropy;                     /* For row-based matchFinder: collects entropy for salt generation */
 
     U32* hashTable;
@@ -360,10 +361,11 @@ struct ZSTD_CCtx_params_s {
      * if the external matchfinder returns an error code. */
     int enableMatchFinderFallback;
 
-    /* Indicates whether an external matchfinder has been referenced.
+    /* Parameters for the external sequence producer API.
-     * Users can't set this externally.
+     * Users set these parameters through ZSTD_registerSequenceProducer().
-     * It is set internally in ZSTD_registerSequenceProducer(). */
+     * It is not possible to set these parameters individually through the public API. */
-    int useSequenceProducer;
+    void* extSeqProdState;
+    ZSTD_sequenceProducer_F extSeqProdFunc;
 
     /* Adjust the max block size*/
     size_t maxBlockSize;
@@ -401,14 +403,6 @@ typedef struct {
     ZSTD_entropyCTablesMetadata_t entropyMetadata;
 } ZSTD_blockSplitCtx;
 
-/* Context for block-level external matchfinder API */
-typedef struct {
-  void* mState;
-  ZSTD_sequenceProducer_F* mFinder;
-  ZSTD_Sequence* seqBuffer;
-  size_t seqBufferCapacity;
-} ZSTD_externalMatchCtx;
-
 struct ZSTD_CCtx_s {
     ZSTD_compressionStage_e stage;
     int cParamsChanged;                  /* == 1 if cParams(except wlog) or compression level are changed in requestedParams. Triggers transmission of new params to ZSTDMT (if available) then reset to 0. */
@@ -479,8 +473,9 @@ struct ZSTD_CCtx_s {
     /* Workspace for block splitter */
     ZSTD_blockSplitCtx blockSplitCtx;
 
-    /* Workspace for external matchfinder */
+    /* Buffer for output from external sequence producer */
-    ZSTD_externalMatchCtx externalMatchCtx;
+    ZSTD_Sequence* extSeqBuf;
+    size_t extSeqBufCapacity;
 };
 
 typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e;
@@ -1053,7 +1048,9 @@ MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window,
  * The least significant cycleLog bits of the indices must remain the same,
  * which may be 0. Every index up to maxDist in the past must be valid.
  */
-MEM_STATIC U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog,
+MEM_STATIC
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog,
                                            U32 maxDist, void const* src)
 {
     /* preemptive overflow correction:
@@ -1246,7 +1243,9 @@ MEM_STATIC void ZSTD_window_init(ZSTD_window_t* window) {
  * forget about the extDict. Handles overlap of the prefix and extDict.
  * Returns non-zero if the segment is contiguous.
  */
-MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window,
+MEM_STATIC
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_window_update(ZSTD_window_t* window,
                                   void const* src, size_t srcSize,
                                   int forceNonContiguous)
 {
@@ -1467,11 +1466,10 @@ size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity);
  * This cannot be used when long range matching is enabled.
  * Zstd will use these sequences, and pass the literals to a secondary block
  * compressor.
- * @return : An error code on failure.
  * NOTE: seqs are not verified! Invalid sequences can cause out-of-bounds memory
  * access and data corruption.
  */
-size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq);
+void ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq);
 
 /** ZSTD_cycleLog() :
  *  condition for correct operation : hashLog > 1 */
@@ -1509,6 +1507,10 @@ 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);
 
+/* Returns 1 if an external sequence producer is registered, otherwise returns 0. */
+MEM_STATIC int ZSTD_hasExtSeqProd(const ZSTD_CCtx_params* params) {
+    return params->extSeqProdFunc != NULL;
+}
 
 /* ===============================================================
  * Deprecated definitions that are still used internally to avoid

zstd/compress/zstd_compress_superblock.c

@@ -76,8 +76,8 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
     }
 
     {   int const flags = bmi2 ? HUF_flags_bmi2 : 0;
-        const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, oend-op, literals, litSize, hufTable, flags)
+        const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags)
-                                          : HUF_compress4X_usingCTable(op, oend-op, literals, litSize, hufTable, flags);
+                                          : HUF_compress4X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags);
         op += cSize;
         cLitSize += cSize;
         if (cSize == 0 || ERR_isError(cSize)) {
@@ -102,7 +102,7 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
     switch(lhSize)
     {
     case 3: /* 2 - 2 - 10 - 10 */
-        {   U32 const lhc = hType + ((!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14);
+        {   U32 const lhc = hType + ((U32)(!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14);
             MEM_writeLE24(ostart, lhc);
             break;
         }
@@ -122,30 +122,30 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
     }
     *entropyWritten = 1;
     DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)litSize, (U32)(op-ostart));
-    return op-ostart;
+    return (size_t)(op-ostart);
 }
 
 static size_t
 ZSTD_seqDecompressedSize(seqStore_t const* seqStore,
-                   const seqDef* sequences, size_t nbSeq,
+                   const seqDef* sequences, size_t nbSeqs,
-                         size_t litSize, int lastSequence)
+                         size_t litSize, int lastSubBlock)
 {
-    const seqDef* const sstart = sequences;
-    const seqDef* const send = sequences + nbSeq;
-    const seqDef* sp = sstart;
     size_t matchLengthSum = 0;
     size_t litLengthSum = 0;
-    (void)(litLengthSum); /* suppress unused variable warning on some environments */
+    size_t n;
-    while (send-sp > 0) {
+    for (n=0; n<nbSeqs; n++) {
-        ZSTD_sequenceLength const seqLen = ZSTD_getSequenceLength(seqStore, sp);
+        const ZSTD_sequenceLength seqLen = ZSTD_getSequenceLength(seqStore, sequences+n);
         litLengthSum += seqLen.litLength;
         matchLengthSum += seqLen.matchLength;
-        sp++;
     }
-    assert(litLengthSum <= litSize);
+    DEBUGLOG(5, "ZSTD_seqDecompressedSize: %u sequences from %p: %u literals + %u matchlength",
-    if (!lastSequence) {
+                (unsigned)nbSeqs, (const void*)sequences,
+                (unsigned)litLengthSum, (unsigned)matchLengthSum);
+    if (!lastSubBlock)
         assert(litLengthSum == litSize);
-    }
+    else
+        assert(litLengthSum <= litSize);
+    (void)litLengthSum;
     return matchLengthSum + litSize;
 }
 
@@ -180,14 +180,14 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
     /* Sequences Header */
     RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,
                     dstSize_tooSmall, "");
-    if (nbSeq < 0x7F)
+    if (nbSeq < 128)
         *op++ = (BYTE)nbSeq;
     else if (nbSeq < LONGNBSEQ)
         op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;
     else
         op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;
     if (nbSeq==0) {
-        return op - ostart;
+        return (size_t)(op - ostart);
     }
 
     /* seqHead : flags for FSE encoding type */
@@ -209,7 +209,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
     }
 
     {   size_t const bitstreamSize = ZSTD_encodeSequences(
-                                        op, oend - op,
+                                        op, (size_t)(oend - op),
                                         fseTables->matchlengthCTable, mlCode,
                                         fseTables->offcodeCTable, ofCode,
                                         fseTables->litlengthCTable, llCode,
@@ -253,7 +253,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
 #endif
 
     *entropyWritten = 1;
-    return op - ostart;
+    return (size_t)(op - ostart);
 }
 
 /** ZSTD_compressSubBlock() :
@@ -279,7 +279,8 @@ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy,
                 litSize, nbSeq, writeLitEntropy, writeSeqEntropy, lastBlock);
     {   size_t cLitSize = ZSTD_compressSubBlock_literal((const HUF_CElt*)entropy->huf.CTable,
                                                         &entropyMetadata->hufMetadata, literals, litSize,
-                                                        op, oend-op, bmi2, writeLitEntropy, litEntropyWritten);
+                                                        op, (size_t)(oend-op),
+                                                        bmi2, writeLitEntropy, litEntropyWritten);
         FORWARD_IF_ERROR(cLitSize, "ZSTD_compressSubBlock_literal failed");
         if (cLitSize == 0) return 0;
         op += cLitSize;
@@ -289,18 +290,18 @@ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy,
                                                   sequences, nbSeq,
                                                   llCode, mlCode, ofCode,
                                                   cctxParams,
-                                                  op, oend-op,
+                                                  op, (size_t)(oend-op),
                                                   bmi2, writeSeqEntropy, seqEntropyWritten);
         FORWARD_IF_ERROR(cSeqSize, "ZSTD_compressSubBlock_sequences failed");
         if (cSeqSize == 0) return 0;
         op += cSeqSize;
     }
     /* Write block header */
-    {   size_t cSize = (op-ostart)-ZSTD_blockHeaderSize;
+    {   size_t cSize = (size_t)(op-ostart) - ZSTD_blockHeaderSize;
         U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
         MEM_writeLE24(ostart, cBlockHeader24);
     }
-    return op-ostart;
+    return (size_t)(op-ostart);
 }
 
 static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t litSize,
@@ -389,7 +390,11 @@ static size_t ZSTD_estimateSubBlockSize_sequences(const BYTE* ofCodeTable,
     return cSeqSizeEstimate + sequencesSectionHeaderSize;
 }
 
-static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
+typedef struct {
+    size_t estLitSize;
+    size_t estBlockSize;
+} EstimatedBlockSize;
+static EstimatedBlockSize ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
                                         const BYTE* ofCodeTable,
                                         const BYTE* llCodeTable,
                                         const BYTE* mlCodeTable,
@@ -397,15 +402,17 @@ static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
                                         const ZSTD_entropyCTables_t* entropy,
                                         const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
                                         void* workspace, size_t wkspSize,
-                                        int writeLitEntropy, int writeSeqEntropy) {
+                                        int writeLitEntropy, int writeSeqEntropy)
-    size_t cSizeEstimate = 0;
+{
-    cSizeEstimate += ZSTD_estimateSubBlockSize_literal(literals, litSize,
+    EstimatedBlockSize ebs;
+    ebs.estLitSize = ZSTD_estimateSubBlockSize_literal(literals, litSize,
                                                         &entropy->huf, &entropyMetadata->hufMetadata,
                                                         workspace, wkspSize, writeLitEntropy);
-    cSizeEstimate += ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable,
+    ebs.estBlockSize = ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable,
                                                          nbSeq, &entropy->fse, &entropyMetadata->fseMetadata,
                                                          workspace, wkspSize, writeSeqEntropy);
-    return cSizeEstimate + ZSTD_blockHeaderSize;
+    ebs.estBlockSize += ebs.estLitSize + ZSTD_blockHeaderSize;
+    return ebs;
 }
 
 static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMetadata)
@@ -419,13 +426,56 @@ static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMe
     return 0;
 }
 
+static size_t countLiterals(seqStore_t const* seqStore, const seqDef* sp, size_t seqCount)
+{
+    size_t n, total = 0;
+    assert(sp != NULL);
+    for (n=0; n<seqCount; n++) {
+        total += ZSTD_getSequenceLength(seqStore, sp+n).litLength;
+    }
+    DEBUGLOG(6, "countLiterals for %zu sequences from %p => %zu bytes", seqCount, (const void*)sp, total);
+    return total;
+}
+
+#define BYTESCALE 256
+
+static size_t sizeBlockSequences(const seqDef* sp, size_t nbSeqs,
+                size_t targetBudget, size_t avgLitCost, size_t avgSeqCost,
+                int firstSubBlock)
+{
+    size_t n, budget = 0, inSize=0;
+    /* entropy headers */
+    size_t const headerSize = (size_t)firstSubBlock * 120 * BYTESCALE; /* generous estimate */
+    assert(firstSubBlock==0 || firstSubBlock==1);
+    budget += headerSize;
+
+    /* first sequence => at least one sequence*/
+    budget += sp[0].litLength * avgLitCost + avgSeqCost;
+    if (budget > targetBudget) return 1;
+    inSize = sp[0].litLength + (sp[0].mlBase+MINMATCH);
+
+    /* loop over sequences */
+    for (n=1; n<nbSeqs; n++) {
+        size_t currentCost = sp[n].litLength * avgLitCost + avgSeqCost;
+        budget += currentCost;
+        inSize += sp[n].litLength + (sp[n].mlBase+MINMATCH);
+        /* stop when sub-block budget is reached */
+        if ( (budget > targetBudget)
+            /* though continue to expand until the sub-block is deemed compressible */
+          && (budget < inSize * BYTESCALE) )
+            break;
+    }
+
+    return n;
+}
+
 /** ZSTD_compressSubBlock_multi() :
  *  Breaks super-block into multiple sub-blocks and compresses them.
- *  Entropy will be written to the first block.
+ *  Entropy will be written into the first block.
- *  The following blocks will use repeat mode to compress.
+ *  The following blocks use repeat_mode to compress.
- *  All sub-blocks are compressed blocks (no raw or rle blocks).
+ *  Sub-blocks are all compressed, except the last one when beneficial.
- *  @return : compressed size of the super block (which is multiple ZSTD blocks)
+ *  @return : compressed size of the super block (which features multiple ZSTD blocks)
- *            Or 0 if it failed to compress. */
+ *            or 0 if it failed to compress. */
 static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
                             const ZSTD_compressedBlockState_t* prevCBlock,
                             ZSTD_compressedBlockState_t* nextCBlock,
@@ -438,10 +488,12 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
 {
     const seqDef* const sstart = seqStorePtr->sequencesStart;
     const seqDef* const send = seqStorePtr->sequences;
-    const seqDef* sp = sstart;
+    const seqDef* sp = sstart; /* tracks progresses within seqStorePtr->sequences */
+    size_t const nbSeqs = (size_t)(send - sstart);
     const BYTE* const lstart = seqStorePtr->litStart;
     const BYTE* const lend = seqStorePtr->lit;
     const BYTE* lp = lstart;
+    size_t const nbLiterals = (size_t)(lend - lstart);
     BYTE const* ip = (BYTE const*)src;
     BYTE const* const iend = ip + srcSize;
     BYTE* const ostart = (BYTE*)dst;
@@ -450,66 +502,73 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
     const BYTE* llCodePtr = seqStorePtr->llCode;
     const BYTE* mlCodePtr = seqStorePtr->mlCode;
     const BYTE* ofCodePtr = seqStorePtr->ofCode;
-    size_t targetCBlockSize = cctxParams->targetCBlockSize;
+    size_t const minTarget = ZSTD_TARGETCBLOCKSIZE_MIN; /* enforce minimum size, to reduce undesirable side effects */
-    size_t litSize, seqCount;
+    size_t const targetCBlockSize = MAX(minTarget, cctxParams->targetCBlockSize);
-    int writeLitEntropy = entropyMetadata->hufMetadata.hType == set_compressed;
+    int writeLitEntropy = (entropyMetadata->hufMetadata.hType == set_compressed);
     int writeSeqEntropy = 1;
-    int lastSequence = 0;
 
-    DEBUGLOG(5, "ZSTD_compressSubBlock_multi (litSize=%u, nbSeq=%u)",
+    DEBUGLOG(5, "ZSTD_compressSubBlock_multi (srcSize=%u, litSize=%u, nbSeq=%u)",
-                (unsigned)(lend-lp), (unsigned)(send-sstart));
+               (unsigned)srcSize, (unsigned)(lend-lstart), (unsigned)(send-sstart));
 
-    litSize = 0;
+        /* let's start by a general estimation for the full block */
-    seqCount = 0;
+    if (nbSeqs > 0) {
-    do {
+        EstimatedBlockSize const ebs =
-        size_t cBlockSizeEstimate = 0;
+                ZSTD_estimateSubBlockSize(lp, nbLiterals,
-        if (sstart == send) {
+                                        ofCodePtr, llCodePtr, mlCodePtr, nbSeqs,
-            lastSequence = 1;
-        } else {
-            const seqDef* const sequence = sp + seqCount;
-            lastSequence = sequence == send - 1;
-            litSize += ZSTD_getSequenceLength(seqStorePtr, sequence).litLength;
-            seqCount++;
-        }
-        if (lastSequence) {
-            assert(lp <= lend);
-            assert(litSize <= (size_t)(lend - lp));
-            litSize = (size_t)(lend - lp);
-        }
-        /* I think there is an optimization opportunity here.
-         * Calling ZSTD_estimateSubBlockSize for every sequence can be wasteful
-         * since it recalculates estimate from scratch.
-         * For example, it would recount literal distribution and symbol codes every time.
-         */
-        cBlockSizeEstimate = ZSTD_estimateSubBlockSize(lp, litSize, ofCodePtr, llCodePtr, mlCodePtr, seqCount,
                                         &nextCBlock->entropy, entropyMetadata,
-                                                       workspace, wkspSize, writeLitEntropy, writeSeqEntropy);
+                                        workspace, wkspSize,
-        if (cBlockSizeEstimate > targetCBlockSize || lastSequence) {
+                                        writeLitEntropy, writeSeqEntropy);
-            int litEntropyWritten = 0;
+        /* quick estimation */
+        size_t const avgLitCost = nbLiterals ? (ebs.estLitSize * BYTESCALE) / nbLiterals : BYTESCALE;
+        size_t const avgSeqCost = ((ebs.estBlockSize - ebs.estLitSize) * BYTESCALE) / nbSeqs;
+        const size_t nbSubBlocks = MAX((ebs.estBlockSize + (targetCBlockSize/2)) / targetCBlockSize, 1);
+        size_t n, avgBlockBudget, blockBudgetSupp=0;
+        avgBlockBudget = (ebs.estBlockSize * BYTESCALE) / nbSubBlocks;
+        DEBUGLOG(5, "estimated fullblock size=%u bytes ; avgLitCost=%.2f ; avgSeqCost=%.2f ; targetCBlockSize=%u, nbSubBlocks=%u ; avgBlockBudget=%.0f bytes",
+                    (unsigned)ebs.estBlockSize, (double)avgLitCost/BYTESCALE, (double)avgSeqCost/BYTESCALE,
+                    (unsigned)targetCBlockSize, (unsigned)nbSubBlocks, (double)avgBlockBudget/BYTESCALE);
+        /* simplification: if estimates states that the full superblock doesn't compress, just bail out immediately
+         * this will result in the production of a single uncompressed block covering @srcSize.*/
+        if (ebs.estBlockSize > srcSize) return 0;
+
+        /* compress and write sub-blocks */
+        assert(nbSubBlocks>0);
+        for (n=0; n < nbSubBlocks-1; n++) {
+            /* determine nb of sequences for current sub-block + nbLiterals from next sequence */
+            size_t const seqCount = sizeBlockSequences(sp, (size_t)(send-sp),
+                                        avgBlockBudget + blockBudgetSupp, avgLitCost, avgSeqCost, n==0);
+            /* if reached last sequence : break to last sub-block (simplification) */
+            assert(seqCount <= (size_t)(send-sp));
+            if (sp + seqCount == send) break;
+            assert(seqCount > 0);
+            /* compress sub-block */
+            {   int litEntropyWritten = 0;
                 int seqEntropyWritten = 0;
-            const size_t decompressedSize = ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, lastSequence);
+                size_t litSize = countLiterals(seqStorePtr, sp, seqCount);
-            const size_t cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
+                const size_t decompressedSize =
+                        ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 0);
+                size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
                                                 sp, seqCount,
                                                 lp, litSize,
                                                 llCodePtr, mlCodePtr, ofCodePtr,
                                                 cctxParams,
-                                                       op, oend-op,
+                                                op, (size_t)(oend-op),
                                                 bmi2, writeLitEntropy, writeSeqEntropy,
                                                 &litEntropyWritten, &seqEntropyWritten,
-                                                       lastBlock && lastSequence);
+                                                0);
                 FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed");
+
+                /* check compressibility, update state components */
                 if (cSize > 0 && cSize < decompressedSize) {
-                DEBUGLOG(5, "Committed the sub-block");
+                    DEBUGLOG(5, "Committed sub-block compressing %u bytes => %u bytes",
+                                (unsigned)decompressedSize, (unsigned)cSize);
                     assert(ip + decompressedSize <= iend);
                     ip += decompressedSize;
-                sp += seqCount;
                     lp += litSize;
                     op += cSize;
                     llCodePtr += seqCount;
                     mlCodePtr += seqCount;
                     ofCodePtr += seqCount;
-                litSize = 0;
-                seqCount = 0;
                     /* Entropy only needs to be written once */
                     if (litEntropyWritten) {
                         writeLitEntropy = 0;
@@ -517,29 +576,78 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
                     if (seqEntropyWritten) {
                         writeSeqEntropy = 0;
                     }
+                    sp += seqCount;
+                    blockBudgetSupp = 0;
+            }   }
+            /* otherwise : do not compress yet, coalesce current sub-block with following one */
+        }
+    } /* if (nbSeqs > 0) */
+
+    /* write last block */
+    DEBUGLOG(5, "Generate last sub-block: %u sequences remaining", (unsigned)(send - sp));
+    {   int litEntropyWritten = 0;
+        int seqEntropyWritten = 0;
+        size_t litSize = (size_t)(lend - lp);
+        size_t seqCount = (size_t)(send - sp);
+        const size_t decompressedSize =
+                ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 1);
+        size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
+                                            sp, seqCount,
+                                            lp, litSize,
+                                            llCodePtr, mlCodePtr, ofCodePtr,
+                                            cctxParams,
+                                            op, (size_t)(oend-op),
+                                            bmi2, writeLitEntropy, writeSeqEntropy,
+                                            &litEntropyWritten, &seqEntropyWritten,
+                                            lastBlock);
+        FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed");
+
+        /* update pointers, the nb of literals borrowed from next sequence must be preserved */
+        if (cSize > 0 && cSize < decompressedSize) {
+            DEBUGLOG(5, "Last sub-block compressed %u bytes => %u bytes",
+                        (unsigned)decompressedSize, (unsigned)cSize);
+            assert(ip + decompressedSize <= iend);
+            ip += decompressedSize;
+            lp += litSize;
+            op += cSize;
+            llCodePtr += seqCount;
+            mlCodePtr += seqCount;
+            ofCodePtr += seqCount;
+            /* Entropy only needs to be written once */
+            if (litEntropyWritten) {
+                writeLitEntropy = 0;
+            }
+            if (seqEntropyWritten) {
+                writeSeqEntropy = 0;
+            }
+            sp += seqCount;
         }
     }
-    } while (!lastSequence);
+
+
     if (writeLitEntropy) {
-        DEBUGLOG(5, "ZSTD_compressSubBlock_multi has literal entropy tables unwritten");
+        DEBUGLOG(5, "Literal entropy tables were never written");
         ZSTD_memcpy(&nextCBlock->entropy.huf, &prevCBlock->entropy.huf, sizeof(prevCBlock->entropy.huf));
     }
     if (writeSeqEntropy && ZSTD_needSequenceEntropyTables(&entropyMetadata->fseMetadata)) {
         /* If we haven't written our entropy tables, then we've violated our contract and
          * must emit an uncompressed block.
          */
-        DEBUGLOG(5, "ZSTD_compressSubBlock_multi has sequence entropy tables unwritten");
+        DEBUGLOG(5, "Sequence entropy tables were never written => cancel, emit an uncompressed block");
         return 0;
     }
+
     if (ip < iend) {
-        size_t const cSize = ZSTD_noCompressBlock(op, oend - op, ip, iend - ip, lastBlock);
+        /* some data left : last part of the block sent uncompressed */
-        DEBUGLOG(5, "ZSTD_compressSubBlock_multi last sub-block uncompressed, %zu bytes", (size_t)(iend - ip));
+        size_t const rSize = (size_t)((iend - ip));
+        size_t const cSize = ZSTD_noCompressBlock(op, (size_t)(oend - op), ip, rSize, lastBlock);
+        DEBUGLOG(5, "Generate last uncompressed sub-block of %u bytes", (unsigned)(rSize));
         FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed");
         assert(cSize != 0);
         op += cSize;
         /* We have to regenerate the repcodes because we've skipped some sequences */
         if (sp < send) {
-            seqDef const* seq;
+            const seqDef* seq;
             repcodes_t rep;
             ZSTD_memcpy(&rep, prevCBlock->rep, sizeof(rep));
             for (seq = sstart; seq < sp; ++seq) {
@@ -548,14 +656,17 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
             ZSTD_memcpy(nextCBlock->rep, &rep, sizeof(rep));
         }
     }
-    DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed");
+
-    return op-ostart;
+    DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed all subBlocks: total compressed size = %u",
+                (unsigned)(op-ostart));
+    return (size_t)(op-ostart);
 }
 
 size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc,
                                void* dst, size_t dstCapacity,
-                               void const* src, size_t srcSize,
+                               const void* src, size_t srcSize,
-                               unsigned lastBlock) {
+                               unsigned lastBlock)
+{
     ZSTD_entropyCTablesMetadata_t entropyMetadata;
 
     FORWARD_IF_ERROR(ZSTD_buildBlockEntropyStats(&zc->seqStore,

zstd/compress/zstd_cwksp.h

@@ -192,6 +192,7 @@ MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) {
     {
         intptr_t const offset = __msan_test_shadow(ws->initOnceStart,
             (U8*)ZSTD_cwksp_initialAllocStart(ws) - (U8*)ws->initOnceStart);
+        (void)offset;
 #if defined(ZSTD_MSAN_PRINT)
         if(offset!=-1) {
             __msan_print_shadow((U8*)ws->initOnceStart + offset - 8, 32);
@@ -433,7 +434,7 @@ MEM_STATIC void* ZSTD_cwksp_reserve_aligned(ZSTD_cwksp* ws, size_t bytes)
 
 /**
  * Aligned on 64 bytes. These buffers have the special property that
- * their values remain constrained, allowing us to re-use them without
+ * their values remain constrained, allowing us to reuse them without
  * memset()-ing them.
  */
 MEM_STATIC void* ZSTD_cwksp_reserve_table(ZSTD_cwksp* ws, size_t bytes)
@@ -525,7 +526,7 @@ MEM_STATIC void ZSTD_cwksp_mark_tables_dirty(ZSTD_cwksp* ws)
     DEBUGLOG(4, "cwksp: ZSTD_cwksp_mark_tables_dirty");
 
 #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
+    /* To validate that the table reuse 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.
      * Since tableValidEnd space and initOnce space may overlap we don't poison
@@ -602,9 +603,9 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
     DEBUGLOG(4, "cwksp: clearing!");
 
 #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
+    /* To validate that the context reuse logic is sound, and that we don't
      * access stuff that this compression hasn't initialized, we re-"poison"
-     * the workspace except for the areas in which we expect memory re-use
+     * the workspace except for the areas in which we expect memory reuse
      * without initialization (objects, valid tables area and init once
      * memory). */
     {
@@ -635,6 +636,15 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
     ZSTD_cwksp_assert_internal_consistency(ws);
 }
 
+MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) {
+    return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace);
+}
+
+MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) {
+    return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace)
+         + (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart);
+}
+
 /**
  * The provided workspace takes ownership of the buffer [start, start+size).
  * Any existing values in the workspace are ignored (the previously managed
@@ -666,6 +676,11 @@ MEM_STATIC size_t ZSTD_cwksp_create(ZSTD_cwksp* ws, size_t size, ZSTD_customMem
 MEM_STATIC void ZSTD_cwksp_free(ZSTD_cwksp* ws, ZSTD_customMem customMem) {
     void *ptr = ws->workspace;
     DEBUGLOG(4, "cwksp: freeing workspace");
+#if ZSTD_MEMORY_SANITIZER && !defined(ZSTD_MSAN_DONT_POISON_WORKSPACE)
+    if (ptr != NULL && customMem.customFree != NULL) {
+        __msan_unpoison(ptr, ZSTD_cwksp_sizeof(ws));
+    }
+#endif
     ZSTD_memset(ws, 0, sizeof(ZSTD_cwksp));
     ZSTD_customFree(ptr, customMem);
 }
@@ -679,15 +694,6 @@ MEM_STATIC void ZSTD_cwksp_move(ZSTD_cwksp* dst, ZSTD_cwksp* src) {
     ZSTD_memset(src, 0, sizeof(ZSTD_cwksp));
 }
 
-MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) {
-    return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace);
-}
-
-MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) {
-    return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace)
-         + (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart);
-}
-
 MEM_STATIC int ZSTD_cwksp_reserve_failed(const ZSTD_cwksp* ws) {
     return ws->allocFailed;
 }

zstd/compress/zstd_double_fast.c

@@ -11,7 +11,11 @@
 #include "zstd_compress_internal.h"
 #include "zstd_double_fast.h"
 
-static void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms,
+#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
+
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms,
                               void const* end, ZSTD_dictTableLoadMethod_e dtlm)
 {
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -47,7 +51,9 @@ static void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms,
     }   }
 }
 
-static void ZSTD_fillDoubleHashTableForCCtx(ZSTD_matchState_t* ms,
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_fillDoubleHashTableForCCtx(ZSTD_matchState_t* ms,
                               void const* end, ZSTD_dictTableLoadMethod_e dtlm)
 {
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -95,6 +101,7 @@ void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
 
 
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_compressBlock_doubleFast_noDict_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize, U32 const mls /* template */)
@@ -305,6 +312,7 @@ _match_stored:
 
 
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize,
@@ -348,8 +356,8 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
     if (ms->prefetchCDictTables) {
         size_t const hashTableBytes = (((size_t)1) << dictCParams->hashLog) * sizeof(U32);
         size_t const chainTableBytes = (((size_t)1) << dictCParams->chainLog) * sizeof(U32);
-        PREFETCH_AREA(dictHashLong, hashTableBytes)
+        PREFETCH_AREA(dictHashLong, hashTableBytes);
-        PREFETCH_AREA(dictHashSmall, chainTableBytes)
+        PREFETCH_AREA(dictHashSmall, chainTableBytes);
     }
 
     /* init */
@@ -589,7 +597,9 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState(
 }
 
 
-static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_compressBlock_doubleFast_extDict_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize,
         U32 const mls /* template */)
@@ -756,3 +766,5 @@ size_t ZSTD_compressBlock_doubleFast_extDict(
         return ZSTD_compressBlock_doubleFast_extDict_7(ms, seqStore, rep, src, srcSize);
     }
 }
+
+#endif /* ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR */

zstd/compress/zstd_double_fast.h

@@ -18,9 +18,12 @@ extern "C" {
 #include "../common/mem.h"      /* U32 */
 #include "zstd_compress_internal.h"     /* ZSTD_CCtx, size_t */
 
+#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
+
 void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
                               void const* end, ZSTD_dictTableLoadMethod_e dtlm,
                               ZSTD_tableFillPurpose_e tfp);
+
 size_t ZSTD_compressBlock_doubleFast(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
@@ -31,6 +34,14 @@ size_t ZSTD_compressBlock_doubleFast_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
 
+#define ZSTD_COMPRESSBLOCK_DOUBLEFAST ZSTD_compressBlock_doubleFast
+#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE ZSTD_compressBlock_doubleFast_dictMatchState
+#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT ZSTD_compressBlock_doubleFast_extDict
+#else
+#define ZSTD_COMPRESSBLOCK_DOUBLEFAST NULL
+#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT NULL
+#endif /* ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR */
 
 #if defined (__cplusplus)
 }

zstd/compress/zstd_fast.c

@@ -11,7 +11,9 @@
 #include "zstd_compress_internal.h"  /* ZSTD_hashPtr, ZSTD_count, ZSTD_storeSeq */
 #include "zstd_fast.h"
 
-static void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms,
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms,
                         const void* const end,
                         ZSTD_dictTableLoadMethod_e dtlm)
 {
@@ -46,7 +48,9 @@ static void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms,
                 }   }   }   }
 }
 
-static void ZSTD_fillHashTableForCCtx(ZSTD_matchState_t* ms,
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_fillHashTableForCCtx(ZSTD_matchState_t* ms,
                         const void* const end,
                         ZSTD_dictTableLoadMethod_e dtlm)
 {
@@ -139,8 +143,9 @@ void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
  *
  * This is also the work we do at the beginning to enter the loop initially.
  */
-FORCE_INLINE_TEMPLATE size_t
+FORCE_INLINE_TEMPLATE
-ZSTD_compressBlock_fast_noDict_generic(
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_compressBlock_fast_noDict_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize,
         U32 const mls, U32 const hasStep)
@@ -456,6 +461,7 @@ size_t ZSTD_compressBlock_fast(
 }
 
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_compressBlock_fast_dictMatchState_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize, U32 const mls, U32 const hasStep)
@@ -502,7 +508,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
 
     if (ms->prefetchCDictTables) {
         size_t const hashTableBytes = (((size_t)1) << dictCParams->hashLog) * sizeof(U32);
-        PREFETCH_AREA(dictHashTable, hashTableBytes)
+        PREFETCH_AREA(dictHashTable, hashTableBytes);
     }
 
     /* init */
@@ -681,7 +687,9 @@ size_t ZSTD_compressBlock_fast_dictMatchState(
 }
 
 
-static size_t ZSTD_compressBlock_fast_extDict_generic(
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_compressBlock_fast_extDict_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize, U32 const mls, U32 const hasStep)
 {

zstd/compress/zstd_lazy.c

@@ -12,6 +12,11 @@
 #include "zstd_lazy.h"
 #include "../common/bits.h" /* ZSTD_countTrailingZeros64 */
 
+#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR)
+
 #define kLazySkippingStep 8
 
 
@@ -19,8 +24,9 @@
 *  Binary Tree search
 ***************************************/
 
-static void
+static
-ZSTD_updateDUBT(ZSTD_matchState_t* ms,
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_updateDUBT(ZSTD_matchState_t* ms,
                 const BYTE* ip, const BYTE* iend,
                 U32 mls)
 {
@@ -63,8 +69,9 @@ ZSTD_updateDUBT(ZSTD_matchState_t* ms,
  *  sort one already inserted but unsorted position
  *  assumption : curr >= btlow == (curr - btmask)
  *  doesn't fail */
-static void
+static
-ZSTD_insertDUBT1(const ZSTD_matchState_t* ms,
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_insertDUBT1(const ZSTD_matchState_t* ms,
                  U32 curr, const BYTE* inputEnd,
                  U32 nbCompares, U32 btLow,
                  const ZSTD_dictMode_e dictMode)
@@ -152,8 +159,9 @@ ZSTD_insertDUBT1(const ZSTD_matchState_t* ms,
 }
 
 
-static size_t
+static
-ZSTD_DUBT_findBetterDictMatch (
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_DUBT_findBetterDictMatch (
         const ZSTD_matchState_t* ms,
         const BYTE* const ip, const BYTE* const iend,
         size_t* offsetPtr,
@@ -230,8 +238,9 @@ ZSTD_DUBT_findBetterDictMatch (
 }
 
 
-static size_t
+static
-ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
                         const BYTE* const ip, const BYTE* const iend,
                         size_t* offBasePtr,
                         U32 const mls,
@@ -381,8 +390,9 @@ ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
 
 
 /** ZSTD_BtFindBestMatch() : Tree updater, providing best match */
-FORCE_INLINE_TEMPLATE size_t
+FORCE_INLINE_TEMPLATE
-ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms,
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms,
                 const BYTE* const ip, const BYTE* const iLimit,
                       size_t* offBasePtr,
                 const U32 mls /* template */,
@@ -617,7 +627,9 @@ size_t ZSTD_dedicatedDictSearch_lazy_search(size_t* offsetPtr, size_t ml, U32 nb
 
 /* Update chains up to ip (excluded)
    Assumption : always within prefix (i.e. not within extDict) */
-FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_insertAndFindFirstIndex_internal(
                         ZSTD_matchState_t* ms,
                         const ZSTD_compressionParameters* const cParams,
                         const BYTE* ip, U32 const mls, U32 const lazySkipping)
@@ -651,6 +663,7 @@ U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) {
 
 /* inlining is important to hardwire a hot branch (template emulation) */
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_HcFindBestMatch(
                         ZSTD_matchState_t* ms,
                         const BYTE* const ip, const BYTE* const iLimit,
@@ -819,7 +832,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_prefetch(U32 const* hashTable, BYTE const* t
  * Fill up the hash cache starting at idx, prefetching up to ZSTD_ROW_HASH_CACHE_SIZE entries,
  * but not beyond iLimit.
  */
-FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const BYTE* base,
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const BYTE* base,
                                    U32 const rowLog, U32 const mls,
                                    U32 idx, const BYTE* const iLimit)
 {
@@ -845,7 +860,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const B
  * Returns the hash of base + idx, and replaces the hash in the hash cache with the byte at
  * 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,
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable,
                                                   BYTE const* tagTable, BYTE const* base,
                                                   U32 idx, U32 const hashLog,
                                                   U32 const rowLog, U32 const mls,
@@ -863,7 +880,9 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTab
 /* ZSTD_row_update_internalImpl():
  * Updates the hash table with positions starting from updateStartIdx until updateEndIdx.
  */
-FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
                                   U32 updateStartIdx, U32 const updateEndIdx,
                                   U32 const mls, U32 const rowLog,
                                   U32 const rowMask, U32 const useCache)
@@ -892,7 +911,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
  * Inserts the byte at ip into the appropriate position in the hash table, and updates ms->nextToUpdate.
  * Skips sections of long matches as is necessary.
  */
-FORCE_INLINE_TEMPLATE void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const BYTE* ip,
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const BYTE* ip,
                               U32 const mls, U32 const rowLog,
                               U32 const rowMask, U32 const useCache)
 {
@@ -1102,20 +1123,21 @@ ZSTD_row_getMatchMask(const BYTE* const tagRow, const BYTE tag, const U32 headGr
 
 /* The high-level approach of the SIMD row based match finder is as follows:
  * - Figure out where to insert the new entry:
- *      - Generate a hash from a byte along with an additional 1-byte "short hash". The additional byte is our "tag"
+ *      - Generate a hash for current input posistion and split it into a one byte of tag and `rowHashLog` bits of index.
- *      - The hashTable is effectively split into groups or "rows" of 16 or 32 entries of U32, and the hash determines
+ *           - The hash is salted by a value that changes on every contex reset, so when the same table is used
+ *             we will avoid collisions that would otherwise slow us down by intorducing phantom matches.
+ *      - The hashTable is effectively split into groups or "rows" of 15 or 31 entries of U32, and the index determines
  *        which row to insert into.
- *      - Determine the correct position within the row to insert the entry into. Each row of 16 or 32 can
+ *      - Determine the correct position within the row to insert the entry into. Each row of 15 or 31 can
- *        be considered as a circular buffer with a "head" index that resides in the tagTable.
+ *        be considered as a circular buffer with a "head" index that resides in the tagTable (overall 16 or 32 bytes
- *      - Also insert the "tag" into the equivalent row and position in the tagTable.
+ *        per row).
- *          - Note: The tagTable has 17 or 33 1-byte entries per row, due to 16 or 32 tags, and 1 "head" entry.
+ * - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte tag calculated for the position and
- *                  The 17 or 33 entry rows are spaced out to occur every 32 or 64 bytes, respectively,
- *                  for alignment/performance reasons, leaving some bytes unused.
- * - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte "short hash" and
  *   generate a bitfield that we can cycle through to check the collisions in the hash table.
  * - Pick the longest match.
+ * - Insert the tag into the equivalent row and position in the tagTable.
  */
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_RowFindBestMatch(
                         ZSTD_matchState_t* ms,
                         const BYTE* const ip, const BYTE* const iLimit,
@@ -1489,8 +1511,9 @@ FORCE_INLINE_TEMPLATE size_t ZSTD_searchMax(
 *  Common parser - lazy strategy
 *********************************/
 
-FORCE_INLINE_TEMPLATE size_t
+FORCE_INLINE_TEMPLATE
-ZSTD_compressBlock_lazy_generic(
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_compressBlock_lazy_generic(
                         ZSTD_matchState_t* ms, seqStore_t* seqStore,
                         U32 rep[ZSTD_REP_NUM],
                         const void* src, size_t srcSize,
@@ -1754,29 +1777,10 @@ _storeSequence:
     /* Return the last literals size */
     return (size_t)(iend - anchor);
 }
+#endif /* build exclusions */
 
 
-size_t ZSTD_compressBlock_btlazy2(
+#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
-}
-
-size_t ZSTD_compressBlock_lazy2(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
-}
-
-size_t ZSTD_compressBlock_lazy(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
-}
-
 size_t ZSTD_compressBlock_greedy(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -1784,27 +1788,6 @@ size_t ZSTD_compressBlock_greedy(
     return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict);
 }
 
-size_t ZSTD_compressBlock_btlazy2_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
-}
-
-size_t ZSTD_compressBlock_lazy2_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
-}
-
-size_t ZSTD_compressBlock_lazy_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
-}
-
 size_t ZSTD_compressBlock_greedy_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -1812,21 +1795,6 @@ size_t ZSTD_compressBlock_greedy_dictMatchState(
     return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState);
 }
 
-
-size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch);
-}
-
-size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch);
-}
-
 size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -1834,21 +1802,6 @@ size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
     return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dedicatedDictSearch);
 }
 
-/* Row-based matchfinder */
-size_t ZSTD_compressBlock_lazy2_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict);
-}
-
-size_t ZSTD_compressBlock_lazy_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict);
-}
-
 size_t ZSTD_compressBlock_greedy_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -1856,11 +1809,48 @@ size_t ZSTD_compressBlock_greedy_row(
     return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_noDict);
 }
 
-size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
+size_t ZSTD_compressBlock_greedy_dictMatchState_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dedicatedDictSearch);
+}
+#endif
+
+#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_lazy(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_lazy_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch);
+}
+
+size_t ZSTD_compressBlock_lazy_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict);
 }
 
 size_t ZSTD_compressBlock_lazy_dictMatchState_row(
@@ -1870,13 +1860,49 @@ size_t ZSTD_compressBlock_lazy_dictMatchState_row(
     return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dictMatchState);
 }
 
-size_t ZSTD_compressBlock_greedy_dictMatchState_row(
+size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dictMatchState);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch);
+}
+#endif
+
+#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_lazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
 }
 
+size_t ZSTD_compressBlock_lazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
+}
+
+size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch);
+}
+
+size_t ZSTD_compressBlock_lazy2_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict);
+}
+
+size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState);
+}
 
 size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
@@ -1884,22 +1910,30 @@ size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
 {
     return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dedicatedDictSearch);
 }
+#endif
 
-size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
+#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_btlazy2(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
 }
 
-size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row(
+size_t ZSTD_compressBlock_btlazy2_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dedicatedDictSearch);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
 }
+#endif
 
+#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR)
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_compressBlock_lazy_extDict_generic(
                         ZSTD_matchState_t* ms, seqStore_t* seqStore,
                         U32 rep[ZSTD_REP_NUM],
@@ -2101,8 +2135,9 @@ _storeSequence:
     /* Return the last literals size */
     return (size_t)(iend - anchor);
 }
+#endif /* build exclusions */
 
-
+#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
 size_t ZSTD_compressBlock_greedy_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -2110,6 +2145,15 @@ size_t ZSTD_compressBlock_greedy_extDict(
     return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0);
 }
 
+size_t ZSTD_compressBlock_greedy_extDict_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0);
+}
+#endif
+
+#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
 size_t ZSTD_compressBlock_lazy_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -2118,29 +2162,6 @@ size_t ZSTD_compressBlock_lazy_extDict(
     return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1);
 }
 
-size_t ZSTD_compressBlock_lazy2_extDict(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-
-{
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2);
-}
-
-size_t ZSTD_compressBlock_btlazy2_extDict(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-
-{
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2);
-}
-
-size_t ZSTD_compressBlock_greedy_extDict_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0);
-}
-
 size_t ZSTD_compressBlock_lazy_extDict_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
@@ -2148,6 +2169,16 @@ size_t ZSTD_compressBlock_lazy_extDict_row(
 {
     return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1);
 }
+#endif
+
+#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_lazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2);
+}
 
 size_t ZSTD_compressBlock_lazy2_extDict_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
@@ -2155,3 +2186,14 @@ size_t ZSTD_compressBlock_lazy2_extDict_row(
 {
     return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2);
 }
+#endif
+
+#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_btlazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+
+{
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2);
+}
+#endif

zstd/compress/zstd_lazy.h

@@ -27,98 +27,173 @@ extern "C" {
 
 #define ZSTD_ROW_HASH_TAG_BITS 8        /* nb bits to use for the tag */
 
+#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR)
 U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip);
 void ZSTD_row_update(ZSTD_matchState_t* const ms, const BYTE* ip);
 
 void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip);
 
 void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue);  /*! used in ZSTD_reduceIndex(). preemptively increase value of ZSTD_DUBT_UNSORTED_MARK */
+#endif
 
-size_t ZSTD_compressBlock_btlazy2(
+#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy2(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy2_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-
-size_t ZSTD_compressBlock_btlazy2_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy2_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy_dictMatchState_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy_dictMatchState_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-
-size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-
 size_t ZSTD_compressBlock_greedy_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy_extDict(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_lazy2_extDict(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_greedy_extDict_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
+
+#define ZSTD_COMPRESSBLOCK_GREEDY ZSTD_compressBlock_greedy
+#define ZSTD_COMPRESSBLOCK_GREEDY_ROW ZSTD_compressBlock_greedy_row
+#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE ZSTD_compressBlock_greedy_dictMatchState
+#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW ZSTD_compressBlock_greedy_dictMatchState_row
+#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH ZSTD_compressBlock_greedy_dedicatedDictSearch
+#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_greedy_dedicatedDictSearch_row
+#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT ZSTD_compressBlock_greedy_extDict
+#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW ZSTD_compressBlock_greedy_extDict_row
+#else
+#define ZSTD_COMPRESSBLOCK_GREEDY NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_ROW NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT NULL
+#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW NULL
+#endif
+
+#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_lazy(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_dictMatchState_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_lazy_extDict_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
+
+#define ZSTD_COMPRESSBLOCK_LAZY ZSTD_compressBlock_lazy
+#define ZSTD_COMPRESSBLOCK_LAZY_ROW ZSTD_compressBlock_lazy_row
+#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE ZSTD_compressBlock_lazy_dictMatchState
+#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW ZSTD_compressBlock_lazy_dictMatchState_row
+#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH ZSTD_compressBlock_lazy_dedicatedDictSearch
+#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_lazy_dedicatedDictSearch_row
+#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT ZSTD_compressBlock_lazy_extDict
+#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW ZSTD_compressBlock_lazy_extDict_row
+#else
+#define ZSTD_COMPRESSBLOCK_LAZY NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_ROW NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT NULL
+#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW NULL
+#endif
+
+#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_lazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_lazy2_extDict(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_lazy2_extDict_row(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
+
+#define ZSTD_COMPRESSBLOCK_LAZY2 ZSTD_compressBlock_lazy2
+#define ZSTD_COMPRESSBLOCK_LAZY2_ROW ZSTD_compressBlock_lazy2_row
+#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE ZSTD_compressBlock_lazy2_dictMatchState
+#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW ZSTD_compressBlock_lazy2_dictMatchState_row
+#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH ZSTD_compressBlock_lazy2_dedicatedDictSearch
+#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_lazy2_dedicatedDictSearch_row
+#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT ZSTD_compressBlock_lazy2_extDict
+#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW ZSTD_compressBlock_lazy2_extDict_row
+#else
+#define ZSTD_COMPRESSBLOCK_LAZY2 NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_ROW NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT NULL
+#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW NULL
+#endif
+
+#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_btlazy2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btlazy2_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_btlazy2_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
 
+#define ZSTD_COMPRESSBLOCK_BTLAZY2 ZSTD_compressBlock_btlazy2
+#define ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE ZSTD_compressBlock_btlazy2_dictMatchState
+#define ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT ZSTD_compressBlock_btlazy2_extDict
+#else
+#define ZSTD_COMPRESSBLOCK_BTLAZY2 NULL
+#define ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT NULL
+#endif
+
 
 #if defined (__cplusplus)
 }

zstd/compress/zstd_ldm.c

@@ -246,7 +246,11 @@ static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms,
         break;
 
     case ZSTD_dfast:
+#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
         ZSTD_fillDoubleHashTable(ms, iend, ZSTD_dtlm_fast, ZSTD_tfp_forCCtx);
+#else
+        assert(0); /* shouldn't be called: cparams should've been adjusted. */
+#endif
         break;
 
     case ZSTD_greedy:
@@ -318,7 +322,9 @@ static void ZSTD_ldm_limitTableUpdate(ZSTD_matchState_t* ms, const BYTE* anchor)
     }
 }
 
-static size_t ZSTD_ldm_generateSequences_internal(
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_ldm_generateSequences_internal(
         ldmState_t* ldmState, rawSeqStore_t* rawSeqStore,
         ldmParams_t const* params, void const* src, size_t srcSize)
 {
@@ -689,7 +695,6 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
         /* maybeSplitSequence updates rawSeqStore->pos */
         rawSeq const sequence = maybeSplitSequence(rawSeqStore,
                                                    (U32)(iend - ip), minMatch);
-        int i;
         /* End signal */
         if (sequence.offset == 0)
             break;
@@ -702,6 +707,7 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
         /* Run the block compressor */
         DEBUGLOG(5, "pos %u : calling block compressor on segment of size %u", (unsigned)(ip-istart), sequence.litLength);
         {
+            int i;
             size_t const newLitLength =
                 blockCompressor(ms, seqStore, rep, ip, sequence.litLength);
             ip += sequence.litLength;

zstd/compress/zstd_opt.c

@@ -12,6 +12,9 @@
 #include "hist.h"
 #include "zstd_opt.h"
 
+#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR)
 
 #define ZSTD_LITFREQ_ADD    2   /* scaling factor for litFreq, so that frequencies adapt faster to new stats */
 #define ZSTD_MAX_PRICE     (1<<30)
@@ -264,6 +267,7 @@ static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength,
                                 const optState_t* const optPtr,
                                 int optLevel)
 {
+    DEBUGLOG(8, "ZSTD_rawLiteralsCost (%u literals)", litLength);
     if (litLength == 0) return 0;
 
     if (!ZSTD_compressedLiterals(optPtr))
@@ -402,7 +406,9 @@ MEM_STATIC U32 ZSTD_readMINMATCH(const void* memPtr, U32 length)
 
 /* Update hashTable3 up to ip (excluded)
    Assumption : always within prefix (i.e. not within extDict) */
-static U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms,
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms,
                                        U32* nextToUpdate3,
                                        const BYTE* const ip)
 {
@@ -431,7 +437,9 @@ static U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms,
  * @param ip assumed <= iend-8 .
  * @param target The target of ZSTD_updateTree_internal() - we are filling to this position
  * @return : nb of positions added */
-static U32 ZSTD_insertBt1(
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_insertBt1(
                 const ZSTD_matchState_t* ms,
                 const BYTE* const ip, const BYTE* const iend,
                 U32 const target,
@@ -550,6 +558,7 @@ static U32 ZSTD_insertBt1(
 }
 
 FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 void ZSTD_updateTree_internal(
                 ZSTD_matchState_t* ms,
                 const BYTE* const ip, const BYTE* const iend,
@@ -558,7 +567,7 @@ void ZSTD_updateTree_internal(
     const BYTE* const base = ms->window.base;
     U32 const target = (U32)(ip - base);
     U32 idx = ms->nextToUpdate;
-    DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u  (dictMode:%u)",
+    DEBUGLOG(7, "ZSTD_updateTree_internal, from %u to %u  (dictMode:%u)",
                 idx, target, dictMode);
 
     while(idx < target) {
@@ -575,7 +584,9 @@ void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) {
     ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict);
 }
 
-FORCE_INLINE_TEMPLATE U32
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32
 ZSTD_insertBtAndGetAllMatches (
                 ZSTD_match_t* matches,  /* store result (found matches) in this table (presumed large enough) */
                 ZSTD_matchState_t* ms,
@@ -816,7 +827,9 @@ typedef U32 (*ZSTD_getAllMatchesFn)(
     U32 const ll0,
     U32 const lengthToBeat);
 
-FORCE_INLINE_TEMPLATE U32 ZSTD_btGetAllMatches_internal(
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+U32 ZSTD_btGetAllMatches_internal(
         ZSTD_match_t* matches,
         ZSTD_matchState_t* ms,
         U32* nextToUpdate3,
@@ -1035,11 +1048,6 @@ ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm,
 *  Optimal parser
 *********************************/
 
-static U32 ZSTD_totalLen(ZSTD_optimal_t sol)
-{
-    return sol.litlen + sol.mlen;
-}
-
 #if 0 /* debug */
 
 static void
@@ -1057,7 +1065,13 @@ listStats(const U32* table, int lastEltID)
 
 #endif
 
-FORCE_INLINE_TEMPLATE size_t
+#define LIT_PRICE(_p) (int)ZSTD_rawLiteralsCost(_p, 1, optStatePtr, optLevel)
+#define LL_PRICE(_l) (int)ZSTD_litLengthPrice(_l, optStatePtr, optLevel)
+#define LL_INCPRICE(_l) (LL_PRICE(_l) - LL_PRICE(_l-1))
+
+FORCE_INLINE_TEMPLATE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t
 ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
                                seqStore_t* seqStore,
                                U32 rep[ZSTD_REP_NUM],
@@ -1083,10 +1097,10 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
 
     ZSTD_optimal_t* const opt = optStatePtr->priceTable;
     ZSTD_match_t* const matches = optStatePtr->matchTable;
-    ZSTD_optimal_t lastSequence;
+    ZSTD_optimal_t lastStretch;
     ZSTD_optLdm_t optLdm;
 
-    ZSTD_memset(&lastSequence, 0, sizeof(ZSTD_optimal_t));
+    ZSTD_memset(&lastStretch, 0, sizeof(ZSTD_optimal_t));
 
     optLdm.seqStore = ms->ldmSeqStore ? *ms->ldmSeqStore : kNullRawSeqStore;
     optLdm.endPosInBlock = optLdm.startPosInBlock = optLdm.offset = 0;
@@ -1109,18 +1123,30 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
             U32 nbMatches = getAllMatches(matches, ms, &nextToUpdate3, ip, iend, rep, ll0, minMatch);
             ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
                                               (U32)(ip-istart), (U32)(iend-ip));
-            if (!nbMatches) { ip++; continue; }
+            if (!nbMatches) {
+                DEBUGLOG(8, "no match found at cPos %u", (unsigned)(ip-istart));
+                ip++;
+                continue;
+            }
+
+            /* Match found: let's store this solution, and eventually find more candidates.
+             * During this forward pass, @opt is used to store stretches,
+             * defined as "a match followed by N literals".
+             * Note how this is different from a Sequence, which is "N literals followed by a match".
+             * Storing stretches allows us to store different match predecessors
+             * for each literal position part of a literals run. */
 
             /* initialize opt[0] */
-            { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
+            opt[0].mlen = 0;  /* there are only literals so far */
-            opt[0].mlen = 0;  /* means is_a_literal */
             opt[0].litlen = litlen;
-            /* We don't need to include the actual price of the literals because
+            /* No need to include the actual price of the literals before the first match
-             * it is static for the duration of the forward pass, and is included
+             * because it is static for the duration of the forward pass, and is included
-             * in every price. We include the literal length to avoid negative
+             * in every subsequent price. But, we include the literal length because
-             * prices when we subtract the previous literal length.
+             * the cost variation of litlen depends on the value of litlen.
              */
-            opt[0].price = (int)ZSTD_litLengthPrice(litlen, optStatePtr, optLevel);
+            opt[0].price = LL_PRICE(litlen);
+            ZSTD_STATIC_ASSERT(sizeof(opt[0].rep[0]) == sizeof(rep[0]));
+            ZSTD_memcpy(&opt[0].rep, rep, sizeof(opt[0].rep));
 
             /* large match -> immediate encoding */
             {   U32 const maxML = matches[nbMatches-1].len;
@@ -1129,82 +1155,106 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
                             nbMatches, maxML, maxOffBase, (U32)(ip-prefixStart));
 
                 if (maxML > sufficient_len) {
-                    lastSequence.litlen = litlen;
+                    lastStretch.litlen = 0;
-                    lastSequence.mlen = maxML;
+                    lastStretch.mlen = maxML;
-                    lastSequence.off = maxOffBase;
+                    lastStretch.off = maxOffBase;
-                    DEBUGLOG(6, "large match (%u>%u), immediate encoding",
+                    DEBUGLOG(6, "large match (%u>%u) => immediate encoding",
                                 maxML, sufficient_len);
                     cur = 0;
-                    last_pos = ZSTD_totalLen(lastSequence);
+                    last_pos = maxML;
                     goto _shortestPath;
             }   }
 
             /* set prices for first matches starting position == 0 */
             assert(opt[0].price >= 0);
-            {   U32 const literalsPrice = (U32)opt[0].price + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
+            {   U32 pos;
-                U32 pos;
                 U32 matchNb;
                 for (pos = 1; pos < minMatch; pos++) {
-                    opt[pos].price = ZSTD_MAX_PRICE;   /* mlen, litlen and price will be fixed during forward scanning */
+                    opt[pos].price = ZSTD_MAX_PRICE;
+                    opt[pos].mlen = 0;
+                    opt[pos].litlen = litlen + pos;
                 }
                 for (matchNb = 0; matchNb < nbMatches; matchNb++) {
                     U32 const offBase = matches[matchNb].off;
                     U32 const end = matches[matchNb].len;
                     for ( ; pos <= end ; pos++ ) {
-                        U32 const matchPrice = ZSTD_getMatchPrice(offBase, pos, optStatePtr, optLevel);
+                        int const matchPrice = (int)ZSTD_getMatchPrice(offBase, pos, optStatePtr, optLevel);
-                        U32 const sequencePrice = literalsPrice + matchPrice;
+                        int const sequencePrice = opt[0].price + matchPrice;
                         DEBUGLOG(7, "rPos:%u => set initial price : %.2f",
-                                    pos, ZSTD_fCost((int)sequencePrice));
+                                    pos, ZSTD_fCost(sequencePrice));
                         opt[pos].mlen = pos;
                         opt[pos].off = offBase;
-                        opt[pos].litlen = litlen;
+                        opt[pos].litlen = 0; /* end of match */
-                        opt[pos].price = (int)sequencePrice;
+                        opt[pos].price = sequencePrice + LL_PRICE(0);
-                }   }
+                    }
+                }
                 last_pos = pos-1;
+                opt[pos].price = ZSTD_MAX_PRICE;
             }
         }
 
         /* check further positions */
         for (cur = 1; cur <= last_pos; cur++) {
             const BYTE* const inr = ip + cur;
-            assert(cur < ZSTD_OPT_NUM);
+            assert(cur <= ZSTD_OPT_NUM);
-            DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur)
+            DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur);
 
             /* Fix current position with one literal if cheaper */
-            {   U32 const litlen = (opt[cur-1].mlen == 0) ? opt[cur-1].litlen + 1 : 1;
+            {   U32 const litlen = opt[cur-1].litlen + 1;
                 int const price = opt[cur-1].price
-                                + (int)ZSTD_rawLiteralsCost(ip+cur-1, 1, optStatePtr, optLevel)
+                                + LIT_PRICE(ip+cur-1)
-                                + (int)ZSTD_litLengthPrice(litlen, optStatePtr, optLevel)
+                                + LL_INCPRICE(litlen);
-                                - (int)ZSTD_litLengthPrice(litlen-1, optStatePtr, optLevel);
                 assert(price < 1000000000); /* overflow check */
                 if (price <= opt[cur].price) {
+                    ZSTD_optimal_t const prevMatch = opt[cur];
                     DEBUGLOG(7, "cPos:%zi==rPos:%u : better price (%.2f<=%.2f) using literal (ll==%u) (hist:%u,%u,%u)",
                                 inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), litlen,
                                 opt[cur-1].rep[0], opt[cur-1].rep[1], opt[cur-1].rep[2]);
-                    opt[cur].mlen = 0;
+                    opt[cur] = opt[cur-1];
-                    opt[cur].off = 0;
                     opt[cur].litlen = litlen;
                     opt[cur].price = price;
+                    if ( (optLevel >= 1) /* additional check only for higher modes */
+                      && (prevMatch.litlen == 0) /* replace a match */
+                      && (LL_INCPRICE(1) < 0) /* ll1 is cheaper than ll0 */
+                      && LIKELY(ip + cur < iend)
+                    ) {
+                        /* check next position, in case it would be cheaper */
+                        int with1literal = prevMatch.price + LIT_PRICE(ip+cur) + LL_INCPRICE(1);
+                        int withMoreLiterals = price + LIT_PRICE(ip+cur) + LL_INCPRICE(litlen+1);
+                        DEBUGLOG(7, "then at next rPos %u : match+1lit %.2f vs %ulits %.2f",
+                                cur+1, ZSTD_fCost(with1literal), litlen+1, ZSTD_fCost(withMoreLiterals));
+                        if ( (with1literal < withMoreLiterals)
+                          && (with1literal < opt[cur+1].price) ) {
+                            /* update offset history - before it disappears */
+                            U32 const prev = cur - prevMatch.mlen;
+                            repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, prevMatch.off, opt[prev].litlen==0);
+                            assert(cur >= prevMatch.mlen);
+                            DEBUGLOG(7, "==> match+1lit is cheaper (%.2f < %.2f) (hist:%u,%u,%u) !",
+                                        ZSTD_fCost(with1literal), ZSTD_fCost(withMoreLiterals),
+                                        newReps.rep[0], newReps.rep[1], newReps.rep[2] );
+                            opt[cur+1] = prevMatch;  /* mlen & offbase */
+                            ZSTD_memcpy(opt[cur+1].rep, &newReps, sizeof(repcodes_t));
+                            opt[cur+1].litlen = 1;
+                            opt[cur+1].price = with1literal;
+                            if (last_pos < cur+1) last_pos = cur+1;
+                        }
+                    }
                 } else {
-                    DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f) (hist:%u,%u,%u)",
+                    DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f)",
-                                inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price),
+                                inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price));
-                                opt[cur].rep[0], opt[cur].rep[1], opt[cur].rep[2]);
                 }
             }
 
-            /* Set the repcodes of the current position. We must do it here
+            /* Offset history is not updated during match comparison.
-             * because we rely on the repcodes of the 2nd to last sequence being
+             * Do it here, now that the match is selected and confirmed.
-             * correct to set the next chunks repcodes during the backward
-             * traversal.
              */
             ZSTD_STATIC_ASSERT(sizeof(opt[cur].rep) == sizeof(repcodes_t));
             assert(cur >= opt[cur].mlen);
-            if (opt[cur].mlen != 0) {
+            if (opt[cur].litlen == 0) {
+                /* just finished a match => alter offset history */
                 U32 const prev = cur - opt[cur].mlen;
-                repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, opt[cur].off, opt[cur].litlen==0);
+                repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, opt[cur].off, opt[prev].litlen==0);
                 ZSTD_memcpy(opt[cur].rep, &newReps, sizeof(repcodes_t));
-            } else {
-                ZSTD_memcpy(opt[cur].rep, opt[cur - 1].rep, sizeof(repcodes_t));
             }
 
             /* last match must start at a minimum distance of 8 from oend */
@@ -1214,15 +1264,14 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
 
             if ( (optLevel==0) /*static_test*/
               && (opt[cur+1].price <= opt[cur].price + (BITCOST_MULTIPLIER/2)) ) {
-                DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1);
+                DEBUGLOG(7, "skip current position : next rPos(%u) price is cheaper", cur+1);
                 continue;  /* skip unpromising positions; about ~+6% speed, -0.01 ratio */
             }
 
             assert(opt[cur].price >= 0);
-            {   U32 const ll0 = (opt[cur].mlen != 0);
+            {   U32 const ll0 = (opt[cur].litlen == 0);
-                U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0;
+                int const previousPrice = opt[cur].price;
-                U32 const previousPrice = (U32)opt[cur].price;
+                int const basePrice = previousPrice + LL_PRICE(0);
-                U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
                 U32 nbMatches = getAllMatches(matches, ms, &nextToUpdate3, inr, iend, opt[cur].rep, ll0, minMatch);
                 U32 matchNb;
 
@@ -1234,18 +1283,17 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
                     continue;
                 }
 
-                {   U32 const maxML = matches[nbMatches-1].len;
+                {   U32 const longestML = matches[nbMatches-1].len;
-                    DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u",
+                    DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of longest ML=%u",
-                                inr-istart, cur, nbMatches, maxML);
+                                inr-istart, cur, nbMatches, longestML);
 
-                    if ( (maxML > sufficient_len)
+                    if ( (longestML > sufficient_len)
-                      || (cur + maxML >= ZSTD_OPT_NUM) ) {
+                      || (cur + longestML >= ZSTD_OPT_NUM)
-                        lastSequence.mlen = maxML;
+                      || (ip + cur + longestML >= iend) ) {
-                        lastSequence.off = matches[nbMatches-1].off;
+                        lastStretch.mlen = longestML;
-                        lastSequence.litlen = litlen;
+                        lastStretch.off = matches[nbMatches-1].off;
-                        cur -= (opt[cur].mlen==0) ? opt[cur].litlen : 0;  /* last sequence is actually only literals, fix cur to last match - note : may underflow, in which case, it's first sequence, and it's okay */
+                        lastStretch.litlen = 0;
-                        last_pos = cur + ZSTD_totalLen(lastSequence);
+                        last_pos = cur + longestML;
-                        if (cur > ZSTD_OPT_NUM) cur = 0;   /* underflow => first match */
                         goto _shortestPath;
                 }   }
 
@@ -1257,19 +1305,24 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
                     U32 mlen;
 
                     DEBUGLOG(7, "testing match %u => offBase=%4u, mlen=%2u, llen=%2u",
-                                matchNb, matches[matchNb].off, lastML, litlen);
+                                matchNb, matches[matchNb].off, lastML, opt[cur].litlen);
 
                     for (mlen = lastML; mlen >= startML; mlen--) {  /* scan downward */
                         U32 const pos = cur + mlen;
-                        int const price = (int)basePrice + (int)ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel);
+                        int const price = basePrice + (int)ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel);
 
                         if ((pos > last_pos) || (price < opt[pos].price)) {
                             DEBUGLOG(7, "rPos:%u (ml=%2u) => new better price (%.2f<%.2f)",
                                         pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price));
-                            while (last_pos < pos) { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; }   /* fill empty positions */
+                            while (last_pos < pos) {
+                                /* fill empty positions, for future comparisons */
+                                last_pos++;
+                                opt[last_pos].price = ZSTD_MAX_PRICE;
+                                opt[last_pos].litlen = !0;  /* just needs to be != 0, to mean "not an end of match" */
+                            }
                             opt[pos].mlen = mlen;
                             opt[pos].off = offset;
-                            opt[pos].litlen = litlen;
+                            opt[pos].litlen = 0;
                             opt[pos].price = price;
                         } else {
                             DEBUGLOG(7, "rPos:%u (ml=%2u) => new price is worse (%.2f>=%.2f)",
@@ -1277,47 +1330,81 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
                             if (optLevel==0) break;  /* early update abort; gets ~+10% speed for about -0.01 ratio loss */
                         }
             }   }   }
+            opt[last_pos+1].price = ZSTD_MAX_PRICE;
         }  /* for (cur = 1; cur <= last_pos; cur++) */
 
-        lastSequence = opt[last_pos];
+        lastStretch = opt[last_pos];
-        cur = last_pos > ZSTD_totalLen(lastSequence) ? last_pos - ZSTD_totalLen(lastSequence) : 0;  /* single sequence, and it starts before `ip` */
+        assert(cur >= lastStretch.mlen);
-        assert(cur < ZSTD_OPT_NUM);  /* control overflow*/
+        cur = last_pos - lastStretch.mlen;
 
 _shortestPath:   /* cur, last_pos, best_mlen, best_off have to be set */
         assert(opt[0].mlen == 0);
+        assert(last_pos >= lastStretch.mlen);
+        assert(cur == last_pos - lastStretch.mlen);
 
-        /* Set the next chunk's repcodes based on the repcodes of the beginning
+        if (lastStretch.mlen==0) {
-         * of the last match, and the last sequence. This avoids us having to
+            /* no solution : all matches have been converted into literals */
-         * update them while traversing the sequences.
+            assert(lastStretch.litlen == (ip - anchor) + last_pos);
-         */
+            ip += last_pos;
-        if (lastSequence.mlen != 0) {
+            continue;
-            repcodes_t const reps = ZSTD_newRep(opt[cur].rep, lastSequence.off, lastSequence.litlen==0);
+        }
-            ZSTD_memcpy(rep, &reps, sizeof(reps));
+        assert(lastStretch.off > 0);
+
+        /* Update offset history */
+        if (lastStretch.litlen == 0) {
+            /* finishing on a match : update offset history */
+            repcodes_t const reps = ZSTD_newRep(opt[cur].rep, lastStretch.off, opt[cur].litlen==0);
+            ZSTD_memcpy(rep, &reps, sizeof(repcodes_t));
         } else {
-            ZSTD_memcpy(rep, opt[cur].rep, sizeof(repcodes_t));
+            ZSTD_memcpy(rep, lastStretch.rep, sizeof(repcodes_t));
+            assert(cur >= lastStretch.litlen);
+            cur -= lastStretch.litlen;
         }
 
-        {   U32 const storeEnd = cur + 1;
+        /* Let's write the shortest path solution.
+         * It is stored in @opt in reverse order,
+         * starting from @storeEnd (==cur+2),
+         * effectively partially @opt overwriting.
+         * Content is changed too:
+         * - So far, @opt stored stretches, aka a match followed by literals
+         * - Now, it will store sequences, aka literals followed by a match
+         */
+        {   U32 const storeEnd = cur + 2;
             U32 storeStart = storeEnd;
-            U32 seqPos = cur;
+            U32 stretchPos = cur;
 
             DEBUGLOG(6, "start reverse traversal (last_pos:%u, cur:%u)",
                         last_pos, cur); (void)last_pos;
-            assert(storeEnd < ZSTD_OPT_NUM);
+            assert(storeEnd < ZSTD_OPT_SIZE);
-            DEBUGLOG(6, "last sequence copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
+            DEBUGLOG(6, "last stretch copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
-                        storeEnd, lastSequence.litlen, lastSequence.mlen, lastSequence.off);
+                        storeEnd, lastStretch.litlen, lastStretch.mlen, lastStretch.off);
-            opt[storeEnd] = lastSequence;
+            if (lastStretch.litlen > 0) {
-            while (seqPos > 0) {
+                /* last "sequence" is unfinished: just a bunch of literals */
-                U32 const backDist = ZSTD_totalLen(opt[seqPos]);
+                opt[storeEnd].litlen = lastStretch.litlen;
+                opt[storeEnd].mlen = 0;
+                storeStart = storeEnd-1;
+                opt[storeStart] = lastStretch;
+            } {
+                opt[storeEnd] = lastStretch;  /* note: litlen will be fixed */
+                storeStart = storeEnd;
+            }
+            while (1) {
+                ZSTD_optimal_t nextStretch = opt[stretchPos];
+                opt[storeStart].litlen = nextStretch.litlen;
+                DEBUGLOG(6, "selected sequence (llen=%u,mlen=%u,ofc=%u)",
+                            opt[storeStart].litlen, opt[storeStart].mlen, opt[storeStart].off);
+                if (nextStretch.mlen == 0) {
+                    /* reaching beginning of segment */
+                    break;
+                }
                 storeStart--;
-                DEBUGLOG(6, "sequence from rPos=%u copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
+                opt[storeStart] = nextStretch; /* note: litlen will be fixed */
-                            seqPos, storeStart, opt[seqPos].litlen, opt[seqPos].mlen, opt[seqPos].off);
+                assert(nextStretch.litlen + nextStretch.mlen <= stretchPos);
-                opt[storeStart] = opt[seqPos];
+                stretchPos -= nextStretch.litlen + nextStretch.mlen;
-                seqPos = (seqPos > backDist) ? seqPos - backDist : 0;
             }
 
             /* save sequences */
-            DEBUGLOG(6, "sending selected sequences into seqStore")
+            DEBUGLOG(6, "sending selected sequences into seqStore");
             {   U32 storePos;
                 for (storePos=storeStart; storePos <= storeEnd; storePos++) {
                     U32 const llen = opt[storePos].litlen;
@@ -1339,6 +1426,9 @@ _shortestPath:   /* cur, last_pos, best_mlen, best_off have to be set */
                     anchor += advance;
                     ip = anchor;
             }   }
+            DEBUGLOG(7, "new offset history : %u, %u, %u", rep[0], rep[1], rep[2]);
+
+            /* update all costs */
             ZSTD_setBasePrices(optStatePtr, optLevel);
         }
     }   /* while (ip < ilimit) */
@@ -1346,21 +1436,27 @@ _shortestPath:   /* cur, last_pos, best_mlen, best_off have to be set */
     /* Return the last literals size */
     return (size_t)(iend - anchor);
 }
+#endif /* build exclusions */
 
+#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
 static size_t ZSTD_compressBlock_opt0(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         const void* src, size_t srcSize, const ZSTD_dictMode_e dictMode)
 {
     return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /* optLevel */, dictMode);
 }
+#endif
 
+#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
 static size_t ZSTD_compressBlock_opt2(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         const void* src, size_t srcSize, const ZSTD_dictMode_e dictMode)
 {
     return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /* optLevel */, dictMode);
 }
+#endif
 
+#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
 size_t ZSTD_compressBlock_btopt(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         const void* src, size_t srcSize)
@@ -1368,17 +1464,20 @@ size_t ZSTD_compressBlock_btopt(
     DEBUGLOG(5, "ZSTD_compressBlock_btopt");
     return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_noDict);
 }
+#endif
 
 
 
 
+#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
 /* ZSTD_initStats_ultra():
  * make a first compression pass, just to seed stats with more accurate starting values.
  * only works on first block, with no dictionary and no ldm.
  * this function cannot error out, its narrow contract must be respected.
  */
-static void
+static
-ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+void ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
                           seqStore_t* seqStore,
                           U32 rep[ZSTD_REP_NUM],
                     const void* src, size_t srcSize)
@@ -1425,7 +1524,7 @@ size_t ZSTD_compressBlock_btultra2(
      * Consequently, this can only work if no data has been previously loaded in tables,
      * aka, no dictionary, no prefix, no ldm preprocessing.
      * The compression ratio gain is generally small (~0.5% on first block),
-    ** the cost is 2x cpu time on first block. */
+     * the cost is 2x cpu time on first block. */
     assert(srcSize <= ZSTD_BLOCKSIZE_MAX);
     if ( (ms->opt.litLengthSum==0)   /* first block */
       && (seqStore->sequences == seqStore->sequencesStart)  /* no ldm */
@@ -1438,7 +1537,9 @@ size_t ZSTD_compressBlock_btultra2(
 
     return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_noDict);
 }
+#endif
 
+#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
 size_t ZSTD_compressBlock_btopt_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         const void* src, size_t srcSize)
@@ -1446,19 +1547,21 @@ size_t ZSTD_compressBlock_btopt_dictMatchState(
     return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState);
 }
 
-size_t ZSTD_compressBlock_btultra_dictMatchState(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        const void* src, size_t srcSize)
-{
-    return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState);
-}
-
 size_t ZSTD_compressBlock_btopt_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         const void* src, size_t srcSize)
 {
     return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_extDict);
 }
+#endif
+
+#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_btultra_dictMatchState(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        const void* src, size_t srcSize)
+{
+    return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState);
+}
 
 size_t ZSTD_compressBlock_btultra_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
@@ -1466,6 +1569,7 @@ size_t ZSTD_compressBlock_btultra_extDict(
 {
     return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_extDict);
 }
+#endif
 
 /* note : no btultra2 variant for extDict nor dictMatchState,
  * because btultra2 is not meant to work with dictionaries

zstd/compress/zstd_opt.h

@@ -17,28 +17,38 @@ extern "C" {
 
 #include "zstd_compress_internal.h"
 
+#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \
+ || !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR)
 /* used in ZSTD_loadDictionaryContent() */
 void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend);
+#endif
 
+#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
 size_t ZSTD_compressBlock_btopt(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_btultra(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_btultra2(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize);
-
-
 size_t ZSTD_compressBlock_btopt_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
-size_t ZSTD_compressBlock_btultra_dictMatchState(
+size_t ZSTD_compressBlock_btopt_extDict(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
 
-size_t ZSTD_compressBlock_btopt_extDict(
+#define ZSTD_COMPRESSBLOCK_BTOPT ZSTD_compressBlock_btopt
+#define ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE ZSTD_compressBlock_btopt_dictMatchState
+#define ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT ZSTD_compressBlock_btopt_extDict
+#else
+#define ZSTD_COMPRESSBLOCK_BTOPT NULL
+#define ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT NULL
+#endif
+
+#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
+size_t ZSTD_compressBlock_btultra(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+size_t ZSTD_compressBlock_btultra_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize);
 size_t ZSTD_compressBlock_btultra_extDict(
@@ -48,6 +58,20 @@ size_t ZSTD_compressBlock_btultra_extDict(
         /* note : no btultra2 variant for extDict nor dictMatchState,
          * because btultra2 is not meant to work with dictionaries
          * and is only specific for the first block (no prefix) */
+size_t ZSTD_compressBlock_btultra2(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize);
+
+#define ZSTD_COMPRESSBLOCK_BTULTRA ZSTD_compressBlock_btultra
+#define ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE ZSTD_compressBlock_btultra_dictMatchState
+#define ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT ZSTD_compressBlock_btultra_extDict
+#define ZSTD_COMPRESSBLOCK_BTULTRA2 ZSTD_compressBlock_btultra2
+#else
+#define ZSTD_COMPRESSBLOCK_BTULTRA NULL
+#define ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE NULL
+#define ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT NULL
+#define ZSTD_COMPRESSBLOCK_BTULTRA2 NULL
+#endif
 
 #if defined (__cplusplus)
 }

zstd/compress/zstdmt_compress.c

@@ -15,10 +15,6 @@
 #endif
 
 
-/* ======   Constants   ====== */
-#define ZSTDMT_OVERLAPLOG_DEFAULT 0
-
-
 /* ======   Dependencies   ====== */
 #include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
 #include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memset, INT_MAX, UINT_MAX */
@@ -44,12 +40,13 @@
 #  include <unistd.h>
 #  include <sys/times.h>
 
-#  define DEBUG_PRINTHEX(l,p,n) {            \
+#  define DEBUG_PRINTHEX(l,p,n)                                       \
+    do {                                                              \
         unsigned debug_u;                                             \
         for (debug_u=0; debug_u<(n); debug_u++)                       \
             RAWLOG(l, "%02X ", ((const unsigned char*)(p))[debug_u]); \
         RAWLOG(l, " \n");                                             \
-}
+    } while (0)
 
 static unsigned long long GetCurrentClockTimeMicroseconds(void)
 {
@@ -61,25 +58,28 @@ static unsigned long long GetCurrentClockTimeMicroseconds(void)
 }  }
 
 #define MUTEX_WAIT_TIME_DLEVEL 6
-#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) {          \
+#define ZSTD_PTHREAD_MUTEX_LOCK(mutex)                                                  \
+    do {                                                                                \
         if (DEBUGLEVEL >= MUTEX_WAIT_TIME_DLEVEL) {                                     \
             unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds();    \
             ZSTD_pthread_mutex_lock(mutex);                                             \
             {   unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \
                 unsigned long long const elapsedTime = (afterTime-beforeTime);          \
-            if (elapsedTime > 1000) {  /* or whatever threshold you like; I'm using 1 millisecond here */ \
+                if (elapsedTime > 1000) {                                               \
-                DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, "Thread took %llu microseconds to acquire mutex %s \n", \
+                    /* or whatever threshold you like; I'm using 1 millisecond here */  \
+                    DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL,                                    \
+                        "Thread took %llu microseconds to acquire mutex %s \n",         \
                         elapsedTime, #mutex);                                           \
             }   }                                                                       \
         } else {                                                                        \
             ZSTD_pthread_mutex_lock(mutex);                                             \
         }                                                                               \
-}
+    } while (0)
 
 #else
 
 #  define ZSTD_PTHREAD_MUTEX_LOCK(m) ZSTD_pthread_mutex_lock(m)
-#  define DEBUG_PRINTHEX(l,p,n) {}
+#  define DEBUG_PRINTHEX(l,p,n) do { } while (0)
 
 #endif
 
@@ -100,18 +100,39 @@ typedef struct ZSTDMT_bufferPool_s {
     unsigned totalBuffers;
     unsigned nbBuffers;
     ZSTD_customMem cMem;
-    buffer_t bTable[1];   /* variable size */
+    buffer_t* buffers;
 } ZSTDMT_bufferPool;
 
+static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool)
+{
+    DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool);
+    if (!bufPool) return;   /* compatibility with free on NULL */
+    if (bufPool->buffers) {
+        unsigned u;
+        for (u=0; u<bufPool->totalBuffers; u++) {
+            DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->buffers[u].start);
+            ZSTD_customFree(bufPool->buffers[u].start, bufPool->cMem);
+        }
+        ZSTD_customFree(bufPool->buffers, bufPool->cMem);
+    }
+    ZSTD_pthread_mutex_destroy(&bufPool->poolMutex);
+    ZSTD_customFree(bufPool, bufPool->cMem);
+}
+
 static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned maxNbBuffers, ZSTD_customMem cMem)
 {
-    ZSTDMT_bufferPool* const bufPool = (ZSTDMT_bufferPool*)ZSTD_customCalloc(
+    ZSTDMT_bufferPool* const bufPool =
-        sizeof(ZSTDMT_bufferPool) + (maxNbBuffers-1) * sizeof(buffer_t), cMem);
+        (ZSTDMT_bufferPool*)ZSTD_customCalloc(sizeof(ZSTDMT_bufferPool), cMem);
     if (bufPool==NULL) return NULL;
     if (ZSTD_pthread_mutex_init(&bufPool->poolMutex, NULL)) {
         ZSTD_customFree(bufPool, cMem);
         return NULL;
     }
+    bufPool->buffers = (buffer_t*)ZSTD_customCalloc(maxNbBuffers * sizeof(buffer_t), cMem);
+    if (bufPool->buffers==NULL) {
+        ZSTDMT_freeBufferPool(bufPool);
+        return NULL;
+    }
     bufPool->bufferSize = 64 KB;
     bufPool->totalBuffers = maxNbBuffers;
     bufPool->nbBuffers = 0;
@@ -119,32 +140,19 @@ static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned maxNbBuffers, ZSTD_cu
     return bufPool;
 }
 
-static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool)
-{
-    unsigned u;
-    DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool);
-    if (!bufPool) return;   /* compatibility with free on NULL */
-    for (u=0; u<bufPool->totalBuffers; u++) {
-        DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->bTable[u].start);
-        ZSTD_customFree(bufPool->bTable[u].start, bufPool->cMem);
-    }
-    ZSTD_pthread_mutex_destroy(&bufPool->poolMutex);
-    ZSTD_customFree(bufPool, bufPool->cMem);
-}
-
 /* only works at initialization, not during compression */
 static size_t ZSTDMT_sizeof_bufferPool(ZSTDMT_bufferPool* bufPool)
 {
-    size_t const poolSize = sizeof(*bufPool)
+    size_t const poolSize = sizeof(*bufPool);
-                          + (bufPool->totalBuffers - 1) * sizeof(buffer_t);
+    size_t const arraySize = bufPool->totalBuffers * sizeof(buffer_t);
     unsigned u;
     size_t totalBufferSize = 0;
     ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
     for (u=0; u<bufPool->totalBuffers; u++)
-        totalBufferSize += bufPool->bTable[u].capacity;
+        totalBufferSize += bufPool->buffers[u].capacity;
     ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
 
-    return poolSize + totalBufferSize;
+    return poolSize + arraySize + totalBufferSize;
 }
 
 /* ZSTDMT_setBufferSize() :
@@ -187,9 +195,9 @@ static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)
     DEBUGLOG(5, "ZSTDMT_getBuffer: bSize = %u", (U32)bufPool->bufferSize);
     ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
     if (bufPool->nbBuffers) {   /* try to use an existing buffer */
-        buffer_t const buf = bufPool->bTable[--(bufPool->nbBuffers)];
+        buffer_t const buf = bufPool->buffers[--(bufPool->nbBuffers)];
         size_t const availBufferSize = buf.capacity;
-        bufPool->bTable[bufPool->nbBuffers] = g_nullBuffer;
+        bufPool->buffers[bufPool->nbBuffers] = g_nullBuffer;
         if ((availBufferSize >= bSize) & ((availBufferSize>>3) <= bSize)) {
             /* large enough, but not too much */
             DEBUGLOG(5, "ZSTDMT_getBuffer: provide buffer %u of size %u",
@@ -250,14 +258,14 @@ static void ZSTDMT_releaseBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buf)
     if (buf.start == NULL) return;   /* compatible with release on NULL */
     ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
     if (bufPool->nbBuffers < bufPool->totalBuffers) {
-        bufPool->bTable[bufPool->nbBuffers++] = buf;  /* stored for later use */
+        bufPool->buffers[bufPool->nbBuffers++] = buf;  /* stored for later use */
         DEBUGLOG(5, "ZSTDMT_releaseBuffer: stored buffer of size %u in slot %u",
                     (U32)buf.capacity, (U32)(bufPool->nbBuffers-1));
         ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
         return;
     }
     ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
-    /* Reached bufferPool capacity (should not happen) */
+    /* Reached bufferPool capacity (note: should not happen) */
     DEBUGLOG(5, "ZSTDMT_releaseBuffer: pool capacity reached => freeing ");
     ZSTD_customFree(buf.start, bufPool->cMem);
 }
@@ -350,16 +358,20 @@ typedef struct {
     int totalCCtx;
     int availCCtx;
     ZSTD_customMem cMem;
-    ZSTD_CCtx* cctx[1];   /* variable size */
+    ZSTD_CCtx** cctxs;
 } ZSTDMT_CCtxPool;
 
-/* note : all CCtx borrowed from the pool should be released back to the pool _before_ freeing the pool */
+/* note : all CCtx borrowed from the pool must be reverted back to the pool _before_ freeing the pool */
 static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool)
 {
+    if (!pool) return;
+    ZSTD_pthread_mutex_destroy(&pool->poolMutex);
+    if (pool->cctxs) {
         int cid;
         for (cid=0; cid<pool->totalCCtx; cid++)
-        ZSTD_freeCCtx(pool->cctx[cid]);  /* note : compatible with free on NULL */
+            ZSTD_freeCCtx(pool->cctxs[cid]);  /* free compatible with NULL */
-    ZSTD_pthread_mutex_destroy(&pool->poolMutex);
+        ZSTD_customFree(pool->cctxs, pool->cMem);
+    }
     ZSTD_customFree(pool, pool->cMem);
 }
 
@@ -368,19 +380,24 @@ static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool)
 static ZSTDMT_CCtxPool* ZSTDMT_createCCtxPool(int nbWorkers,
                                               ZSTD_customMem cMem)
 {
-    ZSTDMT_CCtxPool* const cctxPool = (ZSTDMT_CCtxPool*) ZSTD_customCalloc(
+    ZSTDMT_CCtxPool* const cctxPool =
-        sizeof(ZSTDMT_CCtxPool) + (nbWorkers-1)*sizeof(ZSTD_CCtx*), cMem);
+        (ZSTDMT_CCtxPool*) ZSTD_customCalloc(sizeof(ZSTDMT_CCtxPool), cMem);
     assert(nbWorkers > 0);
     if (!cctxPool) return NULL;
     if (ZSTD_pthread_mutex_init(&cctxPool->poolMutex, NULL)) {
         ZSTD_customFree(cctxPool, cMem);
         return NULL;
     }
-    cctxPool->cMem = cMem;
     cctxPool->totalCCtx = nbWorkers;
+    cctxPool->cctxs = (ZSTD_CCtx**)ZSTD_customCalloc(nbWorkers * sizeof(ZSTD_CCtx*), cMem);
+    if (!cctxPool->cctxs) {
+        ZSTDMT_freeCCtxPool(cctxPool);
+        return NULL;
+    }
+    cctxPool->cMem = cMem;
+    cctxPool->cctxs[0] = ZSTD_createCCtx_advanced(cMem);
+    if (!cctxPool->cctxs[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; }
     cctxPool->availCCtx = 1;   /* at least one cctx for single-thread mode */
-    cctxPool->cctx[0] = ZSTD_createCCtx_advanced(cMem);
-    if (!cctxPool->cctx[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; }
     DEBUGLOG(3, "cctxPool created, with %u workers", nbWorkers);
     return cctxPool;
 }
@@ -402,16 +419,16 @@ static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool)
 {
     ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);
     {   unsigned const nbWorkers = cctxPool->totalCCtx;
-        size_t const poolSize = sizeof(*cctxPool)
+        size_t const poolSize = sizeof(*cctxPool);
-                                + (nbWorkers-1) * sizeof(ZSTD_CCtx*);
+        size_t const arraySize = cctxPool->totalCCtx * sizeof(ZSTD_CCtx*);
-        unsigned u;
         size_t totalCCtxSize = 0;
+        unsigned u;
         for (u=0; u<nbWorkers; u++) {
-            totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctx[u]);
+            totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctxs[u]);
         }
         ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
         assert(nbWorkers > 0);
-        return poolSize + totalCCtxSize;
+        return poolSize + arraySize + totalCCtxSize;
     }
 }
 
@@ -421,7 +438,7 @@ static ZSTD_CCtx* ZSTDMT_getCCtx(ZSTDMT_CCtxPool* cctxPool)
     ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);
     if (cctxPool->availCCtx) {
         cctxPool->availCCtx--;
-        {   ZSTD_CCtx* const cctx = cctxPool->cctx[cctxPool->availCCtx];
+        {   ZSTD_CCtx* const cctx = cctxPool->cctxs[cctxPool->availCCtx];
             ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
             return cctx;
     }   }
@@ -435,7 +452,7 @@ static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx)
     if (cctx==NULL) return;   /* compatibility with release on NULL */
     ZSTD_pthread_mutex_lock(&pool->poolMutex);
     if (pool->availCCtx < pool->totalCCtx)
-        pool->cctx[pool->availCCtx++] = cctx;
+        pool->cctxs[pool->availCCtx++] = cctx;
     else {
         /* pool overflow : should not happen, since totalCCtx==nbWorkers */
         DEBUGLOG(4, "CCtx pool overflow : free cctx");
@@ -601,11 +618,8 @@ static void ZSTDMT_serialState_update(serialState_t* serialState,
     ZSTD_pthread_mutex_unlock(&serialState->mutex);
 
     if (seqStore.size > 0) {
-        size_t const err = ZSTD_referenceExternalSequences(
+        ZSTD_referenceExternalSequences(jobCCtx, seqStore.seq, seqStore.size);
-            jobCCtx, seqStore.seq, seqStore.size);
         assert(serialState->params.ldmParams.enableLdm == ZSTD_ps_enable);
-        assert(!ZSTD_isError(err));
-        (void)err;
     }
 }
 
@@ -657,12 +671,13 @@ typedef struct {
     unsigned frameChecksumNeeded;        /* used only by mtctx */
 } ZSTDMT_jobDescription;
 
-#define JOB_ERROR(e) {                          \
+#define JOB_ERROR(e)                                \
+    do {                                            \
         ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex);   \
         job->cSize = e;                             \
         ZSTD_pthread_mutex_unlock(&job->job_mutex); \
         goto _endJob;                               \
-}
+    } while (0)
 
 /* ZSTDMT_compressionJob() is a POOL_function type */
 static void ZSTDMT_compressionJob(void* jobDescription)
@@ -1091,7 +1106,7 @@ ZSTD_frameProgression ZSTDMT_getFrameProgression(ZSTDMT_CCtx* mtctx)
     {   unsigned jobNb;
         unsigned lastJobNb = mtctx->nextJobID + mtctx->jobReady; assert(mtctx->jobReady <= 1);
         DEBUGLOG(6, "ZSTDMT_getFrameProgression: jobs: from %u to <%u (jobReady:%u)",
-                    mtctx->doneJobID, lastJobNb, mtctx->jobReady)
+                    mtctx->doneJobID, lastJobNb, mtctx->jobReady);
         for (jobNb = mtctx->doneJobID ; jobNb < lastJobNb ; jobNb++) {
             unsigned const wJobID = jobNb & mtctx->jobIDMask;
             ZSTDMT_jobDescription* jobPtr = &mtctx->jobs[wJobID];

zstd/decompress/huf_decompress.c

@@ -34,6 +34,12 @@
 *  Macros
 ****************************************************************/
 
+#ifdef HUF_DISABLE_FAST_DECODE
+# define HUF_ENABLE_FAST_DECODE 0
+#else
+# define HUF_ENABLE_FAST_DECODE 1
+#endif
+
 /* These two optional macros force the use one way or another of the two
  * Huffman decompression implementations. You can't force in both directions
  * at the same time.
@@ -158,17 +164,18 @@ static size_t HUF_initFastDStream(BYTE const* ip) {
  * op [in/out] - The output pointers, must be updated to reflect what is written.
  * bits [in/out] - The bitstream containers, must be updated to reflect the current state.
  * dt [in] - The decoding table.
- * ilimit [in] - The input limit, stop when any input pointer is below ilimit.
+ * ilowest [in] - The beginning of the valid range of the input. Decoders may read
+ *                down to this pointer. It may be below iend[0].
  * oend [in] - The end of the output stream. op[3] must not cross oend.
  * iend [in] - The end of each input stream. ip[i] may cross iend[i],
- *             as long as it is above ilimit, but that indicates corruption.
+ *             as long as it is above ilowest, but that indicates corruption.
  */
 typedef struct {
     BYTE const* ip[4];
     BYTE* op[4];
     U64 bits[4];
     void const* dt;
-    BYTE const* ilimit;
+    BYTE const* ilowest;
     BYTE* oend;
     BYTE const* iend[4];
 } HUF_DecompressFastArgs;
@@ -186,9 +193,9 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
     void const* dt = DTable + 1;
     U32 const dtLog = HUF_getDTableDesc(DTable).tableLog;
 
-    const BYTE* const ilimit = (const BYTE*)src + 6 + 8;
+    const BYTE* const istart = (const BYTE*)src;
 
-    BYTE* const oend = (BYTE*)dst + dstSize;
+    BYTE* const oend = ZSTD_maybeNullPtrAdd((BYTE*)dst, dstSize);
 
     /* The fast decoding loop assumes 64-bit little-endian.
      * This condition is false on x32.
@@ -196,6 +203,11 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
     if (!MEM_isLittleEndian() || MEM_32bits())
         return 0;
 
+    /* Avoid nullptr addition */
+    if (dstSize == 0)
+        return 0;
+    assert(dst != NULL);
+
     /* strict minimum : jump table + 1 byte per stream */
     if (srcSize < 10)
         return ERROR(corruption_detected);
@@ -209,7 +221,6 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
 
     /* Read the jump table. */
     {
-        const BYTE* const istart = (const BYTE*)src;
         size_t const length1 = MEM_readLE16(istart);
         size_t const length2 = MEM_readLE16(istart+2);
         size_t const length3 = MEM_readLE16(istart+4);
@@ -221,10 +232,8 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
 
         /* HUF_initFastDStream() requires this, and this small of an input
          * won't benefit from the ASM loop anyways.
-         * length1 must be >= 16 so that ip[0] >= ilimit before the loop
-         * starts.
          */
-        if (length1 < 16 || length2 < 8 || length3 < 8 || length4 < 8)
+        if (length1 < 8 || length2 < 8 || length3 < 8 || length4 < 8)
             return 0;
         if (length4 > srcSize) return ERROR(corruption_detected);   /* overflow */
     }
@@ -256,11 +265,12 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
     args->bits[2] = HUF_initFastDStream(args->ip[2]);
     args->bits[3] = HUF_initFastDStream(args->ip[3]);
 
-    /* If ip[] >= ilimit, it is guaranteed to be safe to
+    /* The decoders must be sure to never read beyond ilowest.
-        * reload bits[]. It may be beyond its section, but is
+     * This is lower than iend[0], but allowing decoders to read
-        * guaranteed to be valid (>= istart).
+     * down to ilowest can allow an extra iteration or two in the
+     * fast loop.
      */
-    args->ilimit = ilimit;
+    args->ilowest = istart;
 
     args->oend = oend;
     args->dt = dt;
@@ -285,13 +295,31 @@ static size_t HUF_initRemainingDStream(BIT_DStream_t* bit, HUF_DecompressFastArg
     assert(sizeof(size_t) == 8);
     bit->bitContainer = MEM_readLEST(args->ip[stream]);
     bit->bitsConsumed = ZSTD_countTrailingZeros64(args->bits[stream]);
-    bit->start = (const char*)args->iend[0];
+    bit->start = (const char*)args->ilowest;
     bit->limitPtr = bit->start + sizeof(size_t);
     bit->ptr = (const char*)args->ip[stream];
 
     return 0;
 }
 
+/* Calls X(N) for each stream 0, 1, 2, 3. */
+#define HUF_4X_FOR_EACH_STREAM(X) \
+    do {                          \
+        X(0);                     \
+        X(1);                     \
+        X(2);                     \
+        X(3);                     \
+    } while (0)
+
+/* Calls X(N, var) for each stream 0, 1, 2, 3. */
+#define HUF_4X_FOR_EACH_STREAM_WITH_VAR(X, var) \
+    do {                                        \
+        X(0, (var));                            \
+        X(1, (var));                            \
+        X(2, (var));                            \
+        X(3, (var));                            \
+    } while (0)
+
 
 #ifndef HUF_FORCE_DECOMPRESS_X2
 
@@ -500,15 +528,19 @@ HUF_decodeSymbolX1(BIT_DStream_t* Dstream, const HUF_DEltX1* dt, const U32 dtLog
 }
 
 #define HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr) \
-    *ptr++ = HUF_decodeSymbolX1(DStreamPtr, dt, dtLog)
+    do { *ptr++ = HUF_decodeSymbolX1(DStreamPtr, dt, dtLog); } while (0)
 
 #define HUF_DECODE_SYMBOLX1_1(ptr, DStreamPtr)      \
+    do {                                            \
         if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
-        HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr)
+            HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr); \
+    } while (0)
 
 #define HUF_DECODE_SYMBOLX1_2(ptr, DStreamPtr)      \
+    do {                                            \
         if (MEM_64bits())                           \
-        HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr)
+            HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr); \
+    } while (0)
 
 HINT_INLINE size_t
 HUF_decodeStreamX1(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX1* const dt, const U32 dtLog)
@@ -546,7 +578,7 @@ HUF_decompress1X1_usingDTable_internal_body(
     const HUF_DTable* DTable)
 {
     BYTE* op = (BYTE*)dst;
-    BYTE* const oend = op + dstSize;
+    BYTE* const oend = ZSTD_maybeNullPtrAdd(op, dstSize);
     const void* dtPtr = DTable + 1;
     const HUF_DEltX1* const dt = (const HUF_DEltX1*)dtPtr;
     BIT_DStream_t bitD;
@@ -574,6 +606,7 @@ HUF_decompress4X1_usingDTable_internal_body(
 {
     /* Check */
     if (cSrcSize < 10) return ERROR(corruption_detected);  /* strict minimum : jump table + 1 byte per stream */
+    if (dstSize < 6) return ERROR(corruption_detected);         /* stream 4-split doesn't work */
 
     {   const BYTE* const istart = (const BYTE*) cSrc;
         BYTE* const ostart = (BYTE*) dst;
@@ -609,7 +642,7 @@ HUF_decompress4X1_usingDTable_internal_body(
 
         if (length4 > cSrcSize) return ERROR(corruption_detected);   /* overflow */
         if (opStart4 > oend) return ERROR(corruption_detected);      /* overflow */
-        if (dstSize < 6) return ERROR(corruption_detected);         /* stream 4-split doesn't work */
+        assert(dstSize >= 6); /* validated above */
         CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );
         CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );
         CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );
@@ -692,7 +725,7 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
     BYTE* op[4];
     U16 const* const dtable = (U16 const*)args->dt;
     BYTE* const oend = args->oend;
-    BYTE const* const ilimit = args->ilimit;
+    BYTE const* const ilowest = args->ilowest;
 
     /* Copy the arguments to local variables */
     ZSTD_memcpy(&bits, &args->bits, sizeof(bits));
@@ -705,13 +738,12 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
     for (;;) {
         BYTE* olimit;
         int stream;
-        int symbol;
 
         /* Assert loop preconditions */
 #ifndef NDEBUG
         for (stream = 0; stream < 4; ++stream) {
             assert(op[stream] <= (stream == 3 ? oend : op[stream + 1]));
-            assert(ip[stream] >= ilimit);
+            assert(ip[stream] >= ilowest);
         }
 #endif
         /* Compute olimit */
@@ -721,7 +753,7 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
             /* Each iteration consumes up to 11 bits * 5 = 55 bits < 7 bytes
              * per stream.
              */
-            size_t const iiters = (size_t)(ip[0] - ilimit) / 7;
+            size_t const iiters = (size_t)(ip[0] - ilowest) / 7;
             /* We can safely run iters iterations before running bounds checks */
             size_t const iters = MIN(oiters, iiters);
             size_t const symbols = iters * 5;
@@ -732,8 +764,8 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
              */
             olimit = op[3] + symbols;
 
-            /* Exit fast decoding loop once we get close to the end. */
+            /* Exit fast decoding loop once we reach the end. */
-            if (op[3] + 20 > olimit)
+            if (op[3] == olimit)
                 break;
 
             /* Exit the decoding loop if any input pointer has crossed the
@@ -752,27 +784,42 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
         }
 #endif
 
+#define HUF_4X1_DECODE_SYMBOL(_stream, _symbol)                 \
+    do {                                                        \
+        int const index = (int)(bits[(_stream)] >> 53);         \
+        int const entry = (int)dtable[index];                   \
+        bits[(_stream)] <<= (entry & 0x3F);                     \
+        op[(_stream)][(_symbol)] = (BYTE)((entry >> 8) & 0xFF); \
+    } while (0)
+
+#define HUF_4X1_RELOAD_STREAM(_stream)                              \
+    do {                                                            \
+        int const ctz = ZSTD_countTrailingZeros64(bits[(_stream)]); \
+        int const nbBits = ctz & 7;                                 \
+        int const nbBytes = ctz >> 3;                               \
+        op[(_stream)] += 5;                                         \
+        ip[(_stream)] -= nbBytes;                                   \
+        bits[(_stream)] = MEM_read64(ip[(_stream)]) | 1;            \
+        bits[(_stream)] <<= nbBits;                                 \
+    } while (0)
+
+        /* Manually unroll the loop because compilers don't consistently
+         * unroll the inner loops, which destroys performance.
+         */
         do {
             /* Decode 5 symbols in each of the 4 streams */
-            for (symbol = 0; symbol < 5; ++symbol) {
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 0);
-                for (stream = 0; stream < 4; ++stream) {
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 1);
-                    int const index = (int)(bits[stream] >> 53);
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 2);
-                    int const entry = (int)dtable[index];
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 3);
-                    bits[stream] <<= (entry & 63);
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 4);
-                    op[stream][symbol] = (BYTE)((entry >> 8) & 0xFF);
+
-                }
+            /* Reload each of the 4 the bitstreams */
-            }
+            HUF_4X_FOR_EACH_STREAM(HUF_4X1_RELOAD_STREAM);
-            /* Reload the bitstreams */
-            for (stream = 0; stream < 4; ++stream) {
-                int const ctz = ZSTD_countTrailingZeros64(bits[stream]);
-                int const nbBits = ctz & 7;
-                int const nbBytes = ctz >> 3;
-                op[stream] += 5;
-                ip[stream] -= nbBytes;
-                bits[stream] = MEM_read64(ip[stream]) | 1;
-                bits[stream] <<= nbBits;
-            }
         } while (op[3] < olimit);
+
+#undef HUF_4X1_DECODE_SYMBOL
+#undef HUF_4X1_RELOAD_STREAM
     }
 
 _out:
@@ -797,8 +844,8 @@ HUF_decompress4X1_usingDTable_internal_fast(
     HUF_DecompressFastLoopFn loopFn)
 {
     void const* dt = DTable + 1;
-    const BYTE* const iend = (const BYTE*)cSrc + 6;
+    BYTE const* const ilowest = (BYTE const*)cSrc;
-    BYTE* const oend = (BYTE*)dst + dstSize;
+    BYTE* const oend = ZSTD_maybeNullPtrAdd((BYTE*)dst, dstSize);
     HUF_DecompressFastArgs args;
     {   size_t const ret = HUF_DecompressFastArgs_init(&args, dst, dstSize, cSrc, cSrcSize, DTable);
         FORWARD_IF_ERROR(ret, "Failed to init fast loop args");
@@ -806,18 +853,22 @@ HUF_decompress4X1_usingDTable_internal_fast(
             return 0;
     }
 
-    assert(args.ip[0] >= args.ilimit);
+    assert(args.ip[0] >= args.ilowest);
     loopFn(&args);
 
-    /* Our loop guarantees that ip[] >= ilimit and that we haven't
+    /* Our loop guarantees that ip[] >= ilowest and that we haven't
     * overwritten any op[].
     */
-    assert(args.ip[0] >= iend);
+    assert(args.ip[0] >= ilowest);
-    assert(args.ip[1] >= iend);
+    assert(args.ip[0] >= ilowest);
-    assert(args.ip[2] >= iend);
+    assert(args.ip[1] >= ilowest);
-    assert(args.ip[3] >= iend);
+    assert(args.ip[2] >= ilowest);
+    assert(args.ip[3] >= ilowest);
     assert(args.op[3] <= oend);
-    (void)iend;
+
+    assert(ilowest == args.ilowest);
+    assert(ilowest + 6 == args.iend[0]);
+    (void)ilowest;
 
     /* finish bit streams one by one. */
     {   size_t const segmentSize = (dstSize+3) / 4;
@@ -868,7 +919,7 @@ static size_t HUF_decompress4X1_usingDTable_internal(void* dst, size_t dstSize,
     }
 #endif
 
-    if (!(flags & HUF_flags_disableFast)) {
+    if (HUF_ENABLE_FAST_DECODE && !(flags & HUF_flags_disableFast)) {
         size_t const ret = HUF_decompress4X1_usingDTable_internal_fast(dst, dstSize, cSrc, cSrcSize, DTable, loopFn);
         if (ret != 0)
             return ret;
@@ -1239,15 +1290,19 @@ HUF_decodeLastSymbolX2(void* op, BIT_DStream_t* DStream, const HUF_DEltX2* dt, c
 }
 
 #define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \
-    ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
+    do { ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog); } while (0)
 
 #define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr)                     \
+    do {                                                           \
         if (MEM_64bits() || (HUF_TABLELOG_MAX<=12))                \
-        ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
+            ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog); \
+    } while (0)
 
 #define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr)                     \
+    do {                                                           \
         if (MEM_64bits())                                          \
-        ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
+            ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog); \
+    } while (0)
 
 HINT_INLINE size_t
 HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd,
@@ -1307,7 +1362,7 @@ HUF_decompress1X2_usingDTable_internal_body(
 
     /* decode */
     {   BYTE* const ostart = (BYTE*) dst;
-        BYTE* const oend = ostart + dstSize;
+        BYTE* const oend = ZSTD_maybeNullPtrAdd(ostart, dstSize);
         const void* const dtPtr = DTable+1;   /* force compiler to not use strict-aliasing */
         const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
         DTableDesc const dtd = HUF_getDTableDesc(DTable);
@@ -1332,6 +1387,7 @@ HUF_decompress4X2_usingDTable_internal_body(
     const HUF_DTable* DTable)
 {
     if (cSrcSize < 10) return ERROR(corruption_detected);   /* strict minimum : jump table + 1 byte per stream */
+    if (dstSize < 6) return ERROR(corruption_detected);         /* stream 4-split doesn't work */
 
     {   const BYTE* const istart = (const BYTE*) cSrc;
         BYTE* const ostart = (BYTE*) dst;
@@ -1367,7 +1423,7 @@ HUF_decompress4X2_usingDTable_internal_body(
 
         if (length4 > cSrcSize) return ERROR(corruption_detected);  /* overflow */
         if (opStart4 > oend) return ERROR(corruption_detected);     /* overflow */
-        if (dstSize < 6) return ERROR(corruption_detected);         /* stream 4-split doesn't work */
+        assert(dstSize >= 6 /* validated above */);
         CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );
         CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );
         CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );
@@ -1472,7 +1528,7 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
     BYTE* op[4];
     BYTE* oend[4];
     HUF_DEltX2 const* const dtable = (HUF_DEltX2 const*)args->dt;
-    BYTE const* const ilimit = args->ilimit;
+    BYTE const* const ilowest = args->ilowest;
 
     /* Copy the arguments to local registers. */
     ZSTD_memcpy(&bits, &args->bits, sizeof(bits));
@@ -1490,13 +1546,12 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
     for (;;) {
         BYTE* olimit;
         int stream;
-        int symbol;
 
         /* Assert loop preconditions */
 #ifndef NDEBUG
         for (stream = 0; stream < 4; ++stream) {
             assert(op[stream] <= oend[stream]);
-            assert(ip[stream] >= ilimit);
+            assert(ip[stream] >= ilowest);
         }
 #endif
         /* Compute olimit */
@@ -1509,7 +1564,7 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
              * We also know that each input pointer is >= ip[0]. So we can run
              * iters loops before running out of input.
              */
-            size_t iters = (size_t)(ip[0] - ilimit) / 7;
+            size_t iters = (size_t)(ip[0] - ilowest) / 7;
             /* Each iteration can produce up to 10 bytes of output per stream.
              * Each output stream my advance at different rates. So take the
              * minimum number of safe iterations among all the output streams.
@@ -1527,8 +1582,8 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
              */
             olimit = op[3] + (iters * 5);
 
-            /* Exit the fast decoding loop if we are too close to the end. */
+            /* Exit the fast decoding loop once we reach the end. */
-            if (op[3] + 10 > olimit)
+            if (op[3] == olimit)
                 break;
 
             /* Exit the decoding loop if any input pointer has crossed the
@@ -1547,54 +1602,58 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
         }
 #endif
 
+#define HUF_4X2_DECODE_SYMBOL(_stream, _decode3)                      \
+    do {                                                              \
+        if ((_decode3) || (_stream) != 3) {                           \
+            int const index = (int)(bits[(_stream)] >> 53);           \
+            HUF_DEltX2 const entry = dtable[index];                   \
+            MEM_write16(op[(_stream)], entry.sequence); \
+            bits[(_stream)] <<= (entry.nbBits) & 0x3F;                \
+            op[(_stream)] += (entry.length);                          \
+        }                                                             \
+    } while (0)
+
+#define HUF_4X2_RELOAD_STREAM(_stream)                                  \
+    do {                                                                \
+        HUF_4X2_DECODE_SYMBOL(3, 1);                                    \
+        {                                                               \
+            int const ctz = ZSTD_countTrailingZeros64(bits[(_stream)]); \
+            int const nbBits = ctz & 7;                                 \
+            int const nbBytes = ctz >> 3;                               \
+            ip[(_stream)] -= nbBytes;                                   \
+            bits[(_stream)] = MEM_read64(ip[(_stream)]) | 1;            \
+            bits[(_stream)] <<= nbBits;                                 \
+        }                                                               \
+    } while (0)
+
+        /* Manually unroll the loop because compilers don't consistently
+         * unroll the inner loops, which destroys performance.
+         */
         do {
-            /* Do 5 table lookups for each of the first 3 streams */
+            /* Decode 5 symbols from each of the first 3 streams.
-            for (symbol = 0; symbol < 5; ++symbol) {
+             * The final stream will be decoded during the reload phase
-                for (stream = 0; stream < 3; ++stream) {
+             * to reduce register pressure.
-                    int const index = (int)(bits[stream] >> 53);
-                    HUF_DEltX2 const entry = dtable[index];
-                    MEM_write16(op[stream], entry.sequence);
-                    bits[stream] <<= (entry.nbBits);
-                    op[stream] += (entry.length);
-                }
-            }
-            /* Do 1 table lookup from the final stream */
-            {
-                int const index = (int)(bits[3] >> 53);
-                HUF_DEltX2 const entry = dtable[index];
-                MEM_write16(op[3], entry.sequence);
-                bits[3] <<= (entry.nbBits);
-                op[3] += (entry.length);
-            }
-            /* Do 4 table lookups from the final stream & reload bitstreams */
-            for (stream = 0; stream < 4; ++stream) {
-                /* Do a table lookup from the final stream.
-                 * This is interleaved with the reloading to reduce register
-                 * pressure. This shouldn't be necessary, but compilers can
-                 * struggle with codegen with high register pressure.
              */
-                {
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
-                    int const index = (int)(bits[3] >> 53);
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
-                    HUF_DEltX2 const entry = dtable[index];
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
-                    MEM_write16(op[3], entry.sequence);
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
-                    bits[3] <<= (entry.nbBits);
+            HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
-                    op[3] += (entry.length);
+
-                }
+            /* Decode one symbol from the final stream */
-                /* Reload the bistreams. The final bitstream must be reloaded
+            HUF_4X2_DECODE_SYMBOL(3, 1);
-                 * after the 5th symbol was decoded.
+
+            /* Decode 4 symbols from the final stream & reload bitstreams.
+             * The final stream is reloaded last, meaning that all 5 symbols
+             * are decoded from the final stream before it is reloaded.
              */
-                {
+            HUF_4X_FOR_EACH_STREAM(HUF_4X2_RELOAD_STREAM);
-                    int const ctz = ZSTD_countTrailingZeros64(bits[stream]);
-                    int const nbBits = ctz & 7;
-                    int const nbBytes = ctz >> 3;
-                    ip[stream] -= nbBytes;
-                    bits[stream] = MEM_read64(ip[stream]) | 1;
-                    bits[stream] <<= nbBits;
-                }
-            }
         } while (op[3] < olimit);
     }
 
+#undef HUF_4X2_DECODE_SYMBOL
+#undef HUF_4X2_RELOAD_STREAM
+
 _out:
 
     /* Save the final values of each of the state variables back to args. */
@@ -1611,8 +1670,8 @@ HUF_decompress4X2_usingDTable_internal_fast(
     const HUF_DTable* DTable,
     HUF_DecompressFastLoopFn loopFn) {
     void const* dt = DTable + 1;
-    const BYTE* const iend = (const BYTE*)cSrc + 6;
+    const BYTE* const ilowest = (const BYTE*)cSrc;
-    BYTE* const oend = (BYTE*)dst + dstSize;
+    BYTE* const oend = ZSTD_maybeNullPtrAdd((BYTE*)dst, dstSize);
     HUF_DecompressFastArgs args;
     {
         size_t const ret = HUF_DecompressFastArgs_init(&args, dst, dstSize, cSrc, cSrcSize, DTable);
@@ -1621,16 +1680,19 @@ HUF_decompress4X2_usingDTable_internal_fast(
             return 0;
     }
 
-    assert(args.ip[0] >= args.ilimit);
+    assert(args.ip[0] >= args.ilowest);
     loopFn(&args);
 
     /* note : op4 already verified within main loop */
-    assert(args.ip[0] >= iend);
+    assert(args.ip[0] >= ilowest);
-    assert(args.ip[1] >= iend);
+    assert(args.ip[1] >= ilowest);
-    assert(args.ip[2] >= iend);
+    assert(args.ip[2] >= ilowest);
-    assert(args.ip[3] >= iend);
+    assert(args.ip[3] >= ilowest);
     assert(args.op[3] <= oend);
-    (void)iend;
+
+    assert(ilowest == args.ilowest);
+    assert(ilowest + 6 == args.iend[0]);
+    (void)ilowest;
 
     /* finish bitStreams one by one */
     {
@@ -1679,7 +1741,7 @@ static size_t HUF_decompress4X2_usingDTable_internal(void* dst, size_t dstSize,
     }
 #endif
 
-    if (!(flags & HUF_flags_disableFast)) {
+    if (HUF_ENABLE_FAST_DECODE && !(flags & HUF_flags_disableFast)) {
         size_t const ret = HUF_decompress4X2_usingDTable_internal_fast(dst, dstSize, cSrc, cSrcSize, DTable, loopFn);
         if (ret != 0)
             return ret;

zstd/decompress/zstd_decompress.c

@@ -55,18 +55,19 @@
 /*-*******************************************************
 *  Dependencies
 *********************************************************/
-#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
 #include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
+#include "../common/allocations.h"  /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
+#include "../common/error_private.h"
+#include "../common/zstd_internal.h"  /* blockProperties_t */
 #include "../common/mem.h"         /* low level memory routines */
+#include "../common/bits.h"  /* ZSTD_highbit32 */
 #define FSE_STATIC_LINKING_ONLY
 #include "../common/fse.h"
 #include "../common/huf.h"
 #include "../common/xxhash.h" /* XXH64_reset, XXH64_update, XXH64_digest, XXH64 */
-#include "../common/zstd_internal.h"  /* blockProperties_t */
 #include "zstd_decompress_internal.h"   /* ZSTD_DCtx */
 #include "zstd_ddict.h"  /* ZSTD_DDictDictContent */
 #include "zstd_decompress_block.h"   /* ZSTD_decompressBlock_internal */
-#include "../common/bits.h"  /* ZSTD_highbit32 */
 
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
 #  include "../legacy/zstd_legacy.h"
@@ -245,6 +246,7 @@ static void ZSTD_DCtx_resetParameters(ZSTD_DCtx* dctx)
     dctx->forceIgnoreChecksum = ZSTD_d_validateChecksum;
     dctx->refMultipleDDicts = ZSTD_rmd_refSingleDDict;
     dctx->disableHufAsm = 0;
+    dctx->maxBlockSizeParam = 0;
 }
 
 static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)
@@ -265,6 +267,7 @@ static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)
 #endif
     dctx->noForwardProgress = 0;
     dctx->oversizedDuration = 0;
+    dctx->isFrameDecompression = 1;
 #if DYNAMIC_BMI2
     dctx->bmi2 = ZSTD_cpuSupportsBmi2();
 #endif
@@ -726,17 +729,17 @@ static ZSTD_frameSizeInfo ZSTD_errorFrameSizeInfo(size_t ret)
     return frameSizeInfo;
 }
 
-static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize)
+static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize, ZSTD_format_e format)
 {
     ZSTD_frameSizeInfo frameSizeInfo;
     ZSTD_memset(&frameSizeInfo, 0, sizeof(ZSTD_frameSizeInfo));
 
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
-    if (ZSTD_isLegacy(src, srcSize))
+    if (format == ZSTD_f_zstd1 && ZSTD_isLegacy(src, srcSize))
         return ZSTD_findFrameSizeInfoLegacy(src, srcSize);
 #endif
 
-    if ((srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
+    if (format == ZSTD_f_zstd1 && (srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
         && (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
         frameSizeInfo.compressedSize = readSkippableFrameSize(src, srcSize);
         assert(ZSTD_isError(frameSizeInfo.compressedSize) ||
@@ -750,7 +753,7 @@ static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize
         ZSTD_frameHeader zfh;
 
         /* Extract Frame Header */
-        {   size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);
+        {   size_t const ret = ZSTD_getFrameHeader_advanced(&zfh, src, srcSize, format);
             if (ZSTD_isError(ret))
                 return ZSTD_errorFrameSizeInfo(ret);
             if (ret > 0)
@@ -793,15 +796,17 @@ static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize
     }
 }
 
+static size_t ZSTD_findFrameCompressedSize_advanced(const void *src, size_t srcSize, ZSTD_format_e format) {
+    ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize, format);
+    return frameSizeInfo.compressedSize;
+}
+
 /** ZSTD_findFrameCompressedSize() :
- *  compatible with legacy mode
+ * See docs in zstd.h
- *  `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame
+ * Note: compatible with legacy mode */
- *  `srcSize` must be at least as large as the frame contained
- *  @return : the compressed size of the frame starting at `src` */
 size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
 {
-    ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+    return ZSTD_findFrameCompressedSize_advanced(src, srcSize, ZSTD_f_zstd1);
-    return frameSizeInfo.compressedSize;
 }
 
 /** ZSTD_decompressBound() :
@@ -815,7 +820,7 @@ unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize)
     unsigned long long bound = 0;
     /* Iterate over each frame */
     while (srcSize > 0) {
-        ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+        ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize, ZSTD_f_zstd1);
         size_t const compressedSize = frameSizeInfo.compressedSize;
         unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;
         if (ZSTD_isError(compressedSize) || decompressedBound == ZSTD_CONTENTSIZE_ERROR)
@@ -835,7 +840,7 @@ size_t ZSTD_decompressionMargin(void const* src, size_t srcSize)
 
     /* Iterate over each frame */
     while (srcSize > 0) {
-        ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+        ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize, ZSTD_f_zstd1);
         size_t const compressedSize = frameSizeInfo.compressedSize;
         unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;
         ZSTD_frameHeader zfh;
@@ -971,6 +976,10 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
         ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize;
     }
 
+    /* Shrink the blockSizeMax if enabled */
+    if (dctx->maxBlockSizeParam != 0)
+        dctx->fParams.blockSizeMax = MIN(dctx->fParams.blockSizeMax, (unsigned)dctx->maxBlockSizeParam);
+
     /* Loop on each block */
     while (1) {
         BYTE* oBlockEnd = oend;
@@ -1003,7 +1012,8 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
         switch(blockProperties.blockType)
         {
         case bt_compressed:
-            decodedSize = ZSTD_decompressBlock_internal(dctx, op, (size_t)(oBlockEnd-op), ip, cBlockSize, /* frame */ 1, not_streaming);
+            assert(dctx->isFrameDecompression == 1);
+            decodedSize = ZSTD_decompressBlock_internal(dctx, op, (size_t)(oBlockEnd-op), ip, cBlockSize, not_streaming);
             break;
         case bt_raw :
             /* Use oend instead of oBlockEnd because this function is safe to overlap. It uses memmove. */
@@ -1016,12 +1026,14 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
         default:
             RETURN_ERROR(corruption_detected, "invalid block type");
         }
-
+        FORWARD_IF_ERROR(decodedSize, "Block decompression failure");
-        if (ZSTD_isError(decodedSize)) return decodedSize;
+        DEBUGLOG(5, "Decompressed block of dSize = %u", (unsigned)decodedSize);
-        if (dctx->validateChecksum)
+        if (dctx->validateChecksum) {
             XXH64_update(&dctx->xxhState, op, decodedSize);
-        if (decodedSize != 0)
+        }
+        if (decodedSize) /* support dst = NULL,0 */ {
             op += decodedSize;
+        }
         assert(ip != NULL);
         ip += cBlockSize;
         remainingSrcSize -= cBlockSize;
@@ -1051,7 +1063,9 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
     return (size_t)(op-ostart);
 }
 
-static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
+static
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
+size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
                                         void* dst, size_t dstCapacity,
                                   const void* src, size_t srcSize,
                                   const void* dict, size_t dictSize,
@@ -1071,7 +1085,7 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
     while (srcSize >= ZSTD_startingInputLength(dctx->format)) {
 
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
-        if (ZSTD_isLegacy(src, srcSize)) {
+        if (dctx->format == ZSTD_f_zstd1 && ZSTD_isLegacy(src, srcSize)) {
             size_t decodedSize;
             size_t const frameSize = ZSTD_findFrameCompressedSizeLegacy(src, srcSize);
             if (ZSTD_isError(frameSize)) return frameSize;
@@ -1081,6 +1095,15 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
             decodedSize = ZSTD_decompressLegacy(dst, dstCapacity, src, frameSize, dict, dictSize);
             if (ZSTD_isError(decodedSize)) return decodedSize;
 
+            {
+                unsigned long long const expectedSize = ZSTD_getFrameContentSize(src, srcSize);
+                RETURN_ERROR_IF(expectedSize == ZSTD_CONTENTSIZE_ERROR, corruption_detected, "Corrupted frame header!");
+                if (expectedSize != ZSTD_CONTENTSIZE_UNKNOWN) {
+                    RETURN_ERROR_IF(expectedSize != decodedSize, corruption_detected,
+                        "Frame header size does not match decoded size!");
+                }
+            }
+
             assert(decodedSize <= dstCapacity);
             dst = (BYTE*)dst + decodedSize;
             dstCapacity -= decodedSize;
@@ -1092,7 +1115,7 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
         }
 #endif
 
-        if (srcSize >= 4) {
+        if (dctx->format == ZSTD_f_zstd1 && 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) {
@@ -1319,7 +1342,8 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
             {
             case bt_compressed:
                 DEBUGLOG(5, "ZSTD_decompressContinue: case bt_compressed");
-                rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 1, is_streaming);
+                assert(dctx->isFrameDecompression == 1);
+                rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, is_streaming);
                 dctx->expected = 0;  /* Streaming not supported */
                 break;
             case bt_raw :
@@ -1388,6 +1412,7 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
     case ZSTDds_decodeSkippableHeader:
         assert(src != NULL);
         assert(srcSize <= ZSTD_SKIPPABLEHEADERSIZE);
+        assert(dctx->format != ZSTD_f_zstd1_magicless);
         ZSTD_memcpy(dctx->headerBuffer + (ZSTD_SKIPPABLEHEADERSIZE - srcSize), src, srcSize);   /* complete skippable header */
         dctx->expected = MEM_readLE32(dctx->headerBuffer + ZSTD_FRAMEIDSIZE);   /* note : dctx->expected can grow seriously large, beyond local buffer size */
         dctx->stage = ZSTDds_skipFrame;
@@ -1548,6 +1573,7 @@ size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)
     dctx->litEntropy = dctx->fseEntropy = 0;
     dctx->dictID = 0;
     dctx->bType = bt_reserved;
+    dctx->isFrameDecompression = 1;
     ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
     ZSTD_memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue));  /* initial repcodes */
     dctx->LLTptr = dctx->entropy.LLTable;
@@ -1819,6 +1845,10 @@ ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam)
             bounds.lowerBound = 0;
             bounds.upperBound = 1;
             return bounds;
+        case ZSTD_d_maxBlockSize:
+            bounds.lowerBound = ZSTD_BLOCKSIZE_MAX_MIN;
+            bounds.upperBound = ZSTD_BLOCKSIZE_MAX;
+            return bounds;
 
         default:;
     }
@@ -1863,6 +1893,9 @@ size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int* value
         case ZSTD_d_disableHuffmanAssembly:
             *value = (int)dctx->disableHufAsm;
             return 0;
+        case ZSTD_d_maxBlockSize:
+            *value = dctx->maxBlockSizeParam;
+            return 0;
         default:;
     }
     RETURN_ERROR(parameter_unsupported, "");
@@ -1900,6 +1933,10 @@ size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value
             CHECK_DBOUNDS(ZSTD_d_disableHuffmanAssembly, value);
             dctx->disableHufAsm = value != 0;
             return 0;
+        case ZSTD_d_maxBlockSize:
+            if (value != 0) CHECK_DBOUNDS(ZSTD_d_maxBlockSize, value);
+            dctx->maxBlockSizeParam = value;
+            return 0;
         default:;
     }
     RETURN_ERROR(parameter_unsupported, "");
@@ -1911,6 +1948,7 @@ size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset)
       || (reset == ZSTD_reset_session_and_parameters) ) {
         dctx->streamStage = zdss_init;
         dctx->noForwardProgress = 0;
+        dctx->isFrameDecompression = 1;
     }
     if ( (reset == ZSTD_reset_parameters)
       || (reset == ZSTD_reset_session_and_parameters) ) {
@@ -1927,11 +1965,17 @@ size_t ZSTD_sizeof_DStream(const ZSTD_DStream* dctx)
     return ZSTD_sizeof_DCtx(dctx);
 }
 
-size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
+static size_t ZSTD_decodingBufferSize_internal(unsigned long long windowSize, unsigned long long frameContentSize, size_t blockSizeMax)
 {
-    size_t const blockSize = (size_t) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
+    size_t const blockSize = MIN((size_t)MIN(windowSize, ZSTD_BLOCKSIZE_MAX), blockSizeMax);
-    /* space is needed to store the litbuffer after the output of a given block without stomping the extDict of a previous run, as well as to cover both windows against wildcopy*/
+    /* We need blockSize + WILDCOPY_OVERLENGTH worth of buffer so that if a block
-    unsigned long long const neededRBSize = windowSize + blockSize + ZSTD_BLOCKSIZE_MAX + (WILDCOPY_OVERLENGTH * 2);
+     * ends at windowSize + WILDCOPY_OVERLENGTH + 1 bytes, we can start writing
+     * the block at the beginning of the output buffer, and maintain a full window.
+     *
+     * We need another blockSize worth of buffer so that we can store split
+     * literals at the end of the block without overwriting the extDict window.
+     */
+    unsigned long long const neededRBSize = windowSize + (blockSize * 2) + (WILDCOPY_OVERLENGTH * 2);
     unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);
     size_t const minRBSize = (size_t) neededSize;
     RETURN_ERROR_IF((unsigned long long)minRBSize != neededSize,
@@ -1939,6 +1983,11 @@ size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long
     return minRBSize;
 }
 
+size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
+{
+    return ZSTD_decodingBufferSize_internal(windowSize, frameContentSize, ZSTD_BLOCKSIZE_MAX);
+}
+
 size_t ZSTD_estimateDStreamSize(size_t windowSize)
 {
     size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
@@ -2134,12 +2183,12 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
             if (zds->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
                 && zds->fParams.frameType != ZSTD_skippableFrame
                 && (U64)(size_t)(oend-op) >= zds->fParams.frameContentSize) {
-                size_t const cSize = ZSTD_findFrameCompressedSize(istart, (size_t)(iend-istart));
+                size_t const cSize = ZSTD_findFrameCompressedSize_advanced(istart, (size_t)(iend-istart), zds->format);
                 if (cSize <= (size_t)(iend-istart)) {
                     /* shortcut : using single-pass mode */
                     size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, (size_t)(oend-op), istart, cSize, ZSTD_getDDict(zds));
                     if (ZSTD_isError(decompressedSize)) return decompressedSize;
-                    DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()")
+                    DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()");
                     assert(istart != NULL);
                     ip = istart + cSize;
                     op = op ? op + decompressedSize : op; /* can occur if frameContentSize = 0 (empty frame) */
@@ -2161,7 +2210,8 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
             DEBUGLOG(4, "Consume header");
             FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(zds, ZSTD_getDDict(zds)), "");
 
-            if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {  /* skippable frame */
+            if (zds->format == ZSTD_f_zstd1
+                && (MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {  /* skippable frame */
                 zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE);
                 zds->stage = ZSTDds_skipFrame;
             } else {
@@ -2177,11 +2227,13 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
             zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
             RETURN_ERROR_IF(zds->fParams.windowSize > zds->maxWindowSize,
                             frameParameter_windowTooLarge, "");
+            if (zds->maxBlockSizeParam != 0)
+                zds->fParams.blockSizeMax = MIN(zds->fParams.blockSizeMax, (unsigned)zds->maxBlockSizeParam);
 
             /* Adapt buffer sizes to frame header instructions */
             {   size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);
                 size_t const neededOutBuffSize = zds->outBufferMode == ZSTD_bm_buffered
-                        ? ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize)
+                        ? ZSTD_decodingBufferSize_internal(zds->fParams.windowSize, zds->fParams.frameContentSize, zds->fParams.blockSizeMax)
                         : 0;
 
                 ZSTD_DCtx_updateOversizedDuration(zds, neededInBuffSize, neededOutBuffSize);

zstd/decompress/zstd_decompress_block.c

@@ -51,6 +51,13 @@ static void ZSTD_copy4(void* dst, const void* src) { ZSTD_memcpy(dst, src, 4); }
  *   Block decoding
  ***************************************************************/
 
+static size_t ZSTD_blockSizeMax(ZSTD_DCtx const* dctx)
+{
+    size_t const blockSizeMax = dctx->isFrameDecompression ? dctx->fParams.blockSizeMax : ZSTD_BLOCKSIZE_MAX;
+    assert(blockSizeMax <= ZSTD_BLOCKSIZE_MAX);
+    return blockSizeMax;
+}
+
 /*! ZSTD_getcBlockSize() :
  *  Provides the size of compressed block from block header `src` */
 size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
@@ -73,41 +80,49 @@ size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
 static void ZSTD_allocateLiteralsBuffer(ZSTD_DCtx* dctx, void* const dst, const size_t dstCapacity, const size_t litSize,
     const streaming_operation streaming, const size_t expectedWriteSize, const unsigned splitImmediately)
 {
-    if (streaming == not_streaming && dstCapacity > ZSTD_BLOCKSIZE_MAX + WILDCOPY_OVERLENGTH + litSize + WILDCOPY_OVERLENGTH)
+    size_t const blockSizeMax = ZSTD_blockSizeMax(dctx);
-    {
+    assert(litSize <= blockSizeMax);
-        /* room for litbuffer to fit without read faulting */
+    assert(dctx->isFrameDecompression || streaming == not_streaming);
-        dctx->litBuffer = (BYTE*)dst + ZSTD_BLOCKSIZE_MAX + WILDCOPY_OVERLENGTH;
+    assert(expectedWriteSize <= blockSizeMax);
+    if (streaming == not_streaming && dstCapacity > blockSizeMax + WILDCOPY_OVERLENGTH + litSize + WILDCOPY_OVERLENGTH) {
+        /* If we aren't streaming, we can just put the literals after the output
+         * of the current block. We don't need to worry about overwriting the
+         * extDict of our window, because it doesn't exist.
+         * So if we have space after the end of the block, just put it there.
+         */
+        dctx->litBuffer = (BYTE*)dst + blockSizeMax + WILDCOPY_OVERLENGTH;
         dctx->litBufferEnd = dctx->litBuffer + litSize;
         dctx->litBufferLocation = ZSTD_in_dst;
-    }
+    } else if (litSize <= ZSTD_LITBUFFEREXTRASIZE) {
-    else if (litSize > ZSTD_LITBUFFEREXTRASIZE)
+        /* Literals fit entirely within the extra buffer, put them there to avoid
-    {
+         * having to split the literals.
-        /* won't fit in litExtraBuffer, so it will be split between end of dst and extra buffer */
+         */
+        dctx->litBuffer = dctx->litExtraBuffer;
+        dctx->litBufferEnd = dctx->litBuffer + litSize;
+        dctx->litBufferLocation = ZSTD_not_in_dst;
+    } else {
+        assert(blockSizeMax > ZSTD_LITBUFFEREXTRASIZE);
+        /* Literals must be split between the output block and the extra lit
+         * buffer. We fill the extra lit buffer with the tail of the literals,
+         * and put the rest of the literals at the end of the block, with
+         * WILDCOPY_OVERLENGTH of buffer room to allow for overreads.
+         * This MUST not write more than our maxBlockSize beyond dst, because in
+         * streaming mode, that could overwrite part of our extDict window.
+         */
         if (splitImmediately) {
             /* won't fit in litExtraBuffer, so it will be split between end of dst and extra buffer */
             dctx->litBuffer = (BYTE*)dst + expectedWriteSize - litSize + ZSTD_LITBUFFEREXTRASIZE - WILDCOPY_OVERLENGTH;
             dctx->litBufferEnd = dctx->litBuffer + litSize - ZSTD_LITBUFFEREXTRASIZE;
-        }
+        } else {
-        else {
             /* initially this will be stored entirely in dst during huffman decoding, it will partially be shifted to litExtraBuffer after */
             dctx->litBuffer = (BYTE*)dst + expectedWriteSize - litSize;
             dctx->litBufferEnd = (BYTE*)dst + expectedWriteSize;
         }
         dctx->litBufferLocation = ZSTD_split;
-    }
+        assert(dctx->litBufferEnd <= (BYTE*)dst + expectedWriteSize);
-    else
-    {
-        /* fits entirely within litExtraBuffer, so no split is necessary */
-        dctx->litBuffer = dctx->litExtraBuffer;
-        dctx->litBufferEnd = dctx->litBuffer + litSize;
-        dctx->litBufferLocation = ZSTD_not_in_dst;
     }
 }
 
-/* Hidden declaration for fullbench */
-size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
-                          const void* src, size_t srcSize,
-                          void* dst, size_t dstCapacity, const streaming_operation streaming);
 /*! ZSTD_decodeLiteralsBlock() :
  * Where it is possible to do so without being stomped by the output during decompression, the literals block will be stored
  * in the dstBuffer.  If there is room to do so, it will be stored in full in the excess dst space after where the current
@@ -116,7 +131,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
  *
  * @return : nb of bytes read from src (< srcSize )
  *  note : symbol not declared but exposed for fullbench */
-size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
+static size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                           const void* src, size_t srcSize,   /* note : srcSize < BLOCKSIZE */
                           void* dst, size_t dstCapacity, const streaming_operation streaming)
 {
@@ -125,6 +140,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
 
     {   const BYTE* const istart = (const BYTE*) src;
         symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
+        size_t const blockSizeMax = ZSTD_blockSizeMax(dctx);
 
         switch(litEncType)
         {
@@ -140,7 +156,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                 U32 const lhlCode = (istart[0] >> 2) & 3;
                 U32 const lhc = MEM_readLE32(istart);
                 size_t hufSuccess;
-                size_t expectedWriteSize = MIN(ZSTD_BLOCKSIZE_MAX, dstCapacity);
+                size_t expectedWriteSize = MIN(blockSizeMax, dstCapacity);
                 int const flags = 0
                     | (ZSTD_DCtx_get_bmi2(dctx) ? HUF_flags_bmi2 : 0)
                     | (dctx->disableHufAsm ? HUF_flags_disableAsm : 0);
@@ -167,7 +183,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                     break;
                 }
                 RETURN_ERROR_IF(litSize > 0 && dst == NULL, dstSize_tooSmall, "NULL not handled");
-                RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected, "");
+                RETURN_ERROR_IF(litSize > blockSizeMax, corruption_detected, "");
                 if (!singleStream)
                     RETURN_ERROR_IF(litSize < MIN_LITERALS_FOR_4_STREAMS, literals_headerWrong,
                         "Not enough literals (%zu) for the 4-streams mode (min %u)",
@@ -214,10 +230,12 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                 }
                 if (dctx->litBufferLocation == ZSTD_split)
                 {
+                    assert(litSize > ZSTD_LITBUFFEREXTRASIZE);
                     ZSTD_memcpy(dctx->litExtraBuffer, dctx->litBufferEnd - ZSTD_LITBUFFEREXTRASIZE, ZSTD_LITBUFFEREXTRASIZE);
                     ZSTD_memmove(dctx->litBuffer + ZSTD_LITBUFFEREXTRASIZE - WILDCOPY_OVERLENGTH, dctx->litBuffer, litSize - ZSTD_LITBUFFEREXTRASIZE);
                     dctx->litBuffer += ZSTD_LITBUFFEREXTRASIZE - WILDCOPY_OVERLENGTH;
                     dctx->litBufferEnd -= WILDCOPY_OVERLENGTH;
+                    assert(dctx->litBufferEnd <= (BYTE*)dst + blockSizeMax);
                 }
 
                 RETURN_ERROR_IF(HUF_isError(hufSuccess), corruption_detected, "");
@@ -232,7 +250,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
         case set_basic:
             {   size_t litSize, lhSize;
                 U32 const lhlCode = ((istart[0]) >> 2) & 3;
-                size_t expectedWriteSize = MIN(ZSTD_BLOCKSIZE_MAX, dstCapacity);
+                size_t expectedWriteSize = MIN(blockSizeMax, dstCapacity);
                 switch(lhlCode)
                 {
                 case 0: case 2: default:   /* note : default is impossible, since lhlCode into [0..3] */
@@ -251,6 +269,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                 }
 
                 RETURN_ERROR_IF(litSize > 0 && dst == NULL, dstSize_tooSmall, "NULL not handled");
+                RETURN_ERROR_IF(litSize > blockSizeMax, corruption_detected, "");
                 RETURN_ERROR_IF(expectedWriteSize < litSize, dstSize_tooSmall, "");
                 ZSTD_allocateLiteralsBuffer(dctx, dst, dstCapacity, litSize, streaming, expectedWriteSize, 1);
                 if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) {  /* risk reading beyond src buffer with wildcopy */
@@ -279,7 +298,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
         case set_rle:
             {   U32 const lhlCode = ((istart[0]) >> 2) & 3;
                 size_t litSize, lhSize;
-                size_t expectedWriteSize = MIN(ZSTD_BLOCKSIZE_MAX, dstCapacity);
+                size_t expectedWriteSize = MIN(blockSizeMax, dstCapacity);
                 switch(lhlCode)
                 {
                 case 0: case 2: default:   /* note : default is impossible, since lhlCode into [0..3] */
@@ -298,7 +317,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                     break;
                 }
                 RETURN_ERROR_IF(litSize > 0 && dst == NULL, dstSize_tooSmall, "NULL not handled");
-                RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected, "");
+                RETURN_ERROR_IF(litSize > blockSizeMax, corruption_detected, "");
                 RETURN_ERROR_IF(expectedWriteSize < litSize, dstSize_tooSmall, "");
                 ZSTD_allocateLiteralsBuffer(dctx, dst, dstCapacity, litSize, streaming, expectedWriteSize, 1);
                 if (dctx->litBufferLocation == ZSTD_split)
@@ -320,6 +339,18 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
     }
 }
 
+/* Hidden declaration for fullbench */
+size_t ZSTD_decodeLiteralsBlock_wrapper(ZSTD_DCtx* dctx,
+                          const void* src, size_t srcSize,
+                          void* dst, size_t dstCapacity);
+size_t ZSTD_decodeLiteralsBlock_wrapper(ZSTD_DCtx* dctx,
+                          const void* src, size_t srcSize,
+                          void* dst, size_t dstCapacity)
+{
+    dctx->isFrameDecompression = 0;
+    return ZSTD_decodeLiteralsBlock(dctx, src, srcSize, dst, dstCapacity, not_streaming);
+}
+
 /* Default FSE distribution tables.
  * These are pre-calculated FSE decoding tables using default distributions as defined in specification :
  * https://github.com/facebook/zstd/blob/release/doc/zstd_compression_format.md#default-distributions
@@ -675,11 +706,6 @@ size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
 
     /* SeqHead */
     nbSeq = *ip++;
-    if (!nbSeq) {
-        *nbSeqPtr=0;
-        RETURN_ERROR_IF(srcSize != 1, srcSize_wrong, "");
-        return 1;
-    }
     if (nbSeq > 0x7F) {
         if (nbSeq == 0xFF) {
             RETURN_ERROR_IF(ip+2 > iend, srcSize_wrong, "");
@@ -692,8 +718,16 @@ size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
     }
     *nbSeqPtr = nbSeq;
 
+    if (nbSeq == 0) {
+        /* No sequence : section ends immediately */
+        RETURN_ERROR_IF(ip != iend, corruption_detected,
+            "extraneous data present in the Sequences section");
+        return (size_t)(ip - istart);
+    }
+
     /* FSE table descriptors */
     RETURN_ERROR_IF(ip+1 > iend, srcSize_wrong, ""); /* minimum possible size: 1 byte for symbol encoding types */
+    RETURN_ERROR_IF(*ip & 3, corruption_detected, ""); /* The last field, Reserved, must be all-zeroes. */
     {   symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
         symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
         symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
@@ -840,7 +874,7 @@ static void ZSTD_safecopy(BYTE* op, const BYTE* const oend_w, BYTE const* ip, pt
 /* ZSTD_safecopyDstBeforeSrc():
  * This version allows overlap with dst before src, or handles the non-overlap case with dst after src
  * Kept separate from more common ZSTD_safecopy case to avoid performance impact to the safecopy common case */
-static void ZSTD_safecopyDstBeforeSrc(BYTE* op, BYTE const* ip, ptrdiff_t length) {
+static void ZSTD_safecopyDstBeforeSrc(BYTE* op, const BYTE* ip, ptrdiff_t length) {
     ptrdiff_t const diff = op - ip;
     BYTE* const oend = op + length;
 
@@ -869,6 +903,7 @@ static void ZSTD_safecopyDstBeforeSrc(BYTE* op, BYTE const* ip, ptrdiff_t length
  * to be optimized for many small sequences, since those fall into ZSTD_execSequence().
  */
 FORCE_NOINLINE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_execSequenceEnd(BYTE* op,
     BYTE* const oend, seq_t sequence,
     const BYTE** litPtr, const BYTE* const litLimit,
@@ -916,6 +951,7 @@ size_t ZSTD_execSequenceEnd(BYTE* op,
  * This version is intended to be used during instances where the litBuffer is still split.  It is kept separate to avoid performance impact for the good case.
  */
 FORCE_NOINLINE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_execSequenceEndSplitLitBuffer(BYTE* op,
     BYTE* const oend, const BYTE* const oend_w, seq_t sequence,
     const BYTE** litPtr, const BYTE* const litLimit,
@@ -961,6 +997,7 @@ size_t ZSTD_execSequenceEndSplitLitBuffer(BYTE* op,
 }
 
 HINT_INLINE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_execSequence(BYTE* op,
     BYTE* const oend, seq_t sequence,
     const BYTE** litPtr, const BYTE* const litLimit,
@@ -1059,6 +1096,7 @@ size_t ZSTD_execSequence(BYTE* op,
 }
 
 HINT_INLINE
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 size_t ZSTD_execSequenceSplitLitBuffer(BYTE* op,
     BYTE* const oend, const BYTE* const oend_w, seq_t sequence,
     const BYTE** litPtr, const BYTE* const litLimit,
@@ -1181,14 +1219,20 @@ ZSTD_updateFseStateWithDInfo(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, U16
 
 typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
 
+/**
+ * ZSTD_decodeSequence():
+ * @p longOffsets : tells the decoder to reload more bit while decoding large offsets
+ *                  only used in 32-bit mode
+ * @return : Sequence (litL + matchL + offset)
+ */
 FORCE_INLINE_TEMPLATE seq_t
-ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
+ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets, const int isLastSeq)
 {
     seq_t seq;
     /*
-     * ZSTD_seqSymbol is a structure with a total of 64 bits wide. So it can be
+     * ZSTD_seqSymbol is a 64 bits wide structure.
-     * loaded in one operation and extracted its fields by simply shifting or
+     * It can be loaded in one operation
-     * bit-extracting on aarch64.
+     * and its fields extracted by simply shifting or bit-extracting on aarch64.
      * GCC doesn't recognize this and generates more unnecessary ldr/ldrb/ldrh
      * operations that cause performance drop. This can be avoided by using this
      * ZSTD_memcpy hack.
@@ -1261,7 +1305,7 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
                 } else {
                     offset = ofBase + ll0 + BIT_readBitsFast(&seqState->DStream, 1);
                     {   size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
-                        temp += !temp;   /* 0 is not valid; input is corrupted; force offset to 1 */
+                        temp -= !temp; /* 0 is not valid: input corrupted => force offset to -1 => corruption detected at execSequence */
                         if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
                         seqState->prevOffset[1] = seqState->prevOffset[0];
                         seqState->prevOffset[0] = offset = temp;
@@ -1288,17 +1332,22 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
         DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",
                     (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
 
+        if (!isLastSeq) {
+            /* don't update FSE state for last Sequence */
             ZSTD_updateFseStateWithDInfo(&seqState->stateLL, &seqState->DStream, llNext, llnbBits);    /* <=  9 bits */
             ZSTD_updateFseStateWithDInfo(&seqState->stateML, &seqState->DStream, mlNext, mlnbBits);    /* <=  9 bits */
             if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);    /* <= 18 bits */
             ZSTD_updateFseStateWithDInfo(&seqState->stateOffb, &seqState->DStream, ofNext, ofnbBits);  /* <=  8 bits */
+            BIT_reloadDStream(&seqState->DStream);
+        }
     }
 
     return seq;
 }
 
-#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
-MEM_STATIC int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefixStart, BYTE const* oLitEnd)
+#if DEBUGLEVEL >= 1
+static int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefixStart, BYTE const* oLitEnd)
 {
     size_t const windowSize = dctx->fParams.windowSize;
     /* No dictionary used. */
@@ -1312,14 +1361,16 @@ MEM_STATIC int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefix
     /* Dictionary is active. */
     return 1;
 }
+#endif
 
-MEM_STATIC void ZSTD_assertValidSequence(
+static void ZSTD_assertValidSequence(
         ZSTD_DCtx const* dctx,
         BYTE const* op, BYTE const* oend,
         seq_t const seq,
         BYTE const* prefixStart, BYTE const* virtualStart)
 {
 #if DEBUGLEVEL >= 1
+    if (dctx->isFrameDecompression) {
         size_t const windowSize = dctx->fParams.windowSize;
         size_t const sequenceSize = seq.litLength + seq.matchLength;
         BYTE const* const oLitEnd = op + seq.litLength;
@@ -1327,7 +1378,7 @@ MEM_STATIC void ZSTD_assertValidSequence(
                 (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
         assert(op <= oend);
         assert((size_t)(oend - op) >= sequenceSize);
-    assert(sequenceSize <= ZSTD_BLOCKSIZE_MAX);
+        assert(sequenceSize <= ZSTD_blockSizeMax(dctx));
         if (ZSTD_dictionaryIsActive(dctx, prefixStart, oLitEnd)) {
             size_t const dictSize = (size_t)((char const*)dctx->dictContentEndForFuzzing - (char const*)dctx->dictContentBeginForFuzzing);
             /* Offset must be within the dictionary. */
@@ -1337,6 +1388,7 @@ MEM_STATIC void ZSTD_assertValidSequence(
             /* Offset must be within our window. */
             assert(seq.offset <= windowSize);
         }
+    }
 #else
     (void)dctx, (void)op, (void)oend, (void)seq, (void)prefixStart, (void)virtualStart;
 #endif
@@ -1351,23 +1403,21 @@ DONT_VECTORIZE
 ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
                                void* dst, size_t maxDstSize,
                          const void* seqStart, size_t seqSize, int nbSeq,
-                         const ZSTD_longOffset_e isLongOffset,
+                         const ZSTD_longOffset_e isLongOffset)
-                         const int frame)
 {
     const BYTE* ip = (const BYTE*)seqStart;
     const BYTE* const iend = ip + seqSize;
     BYTE* const ostart = (BYTE*)dst;
-    BYTE* const oend = ostart + maxDstSize;
+    BYTE* const oend = ZSTD_maybeNullPtrAdd(ostart, maxDstSize);
     BYTE* op = ostart;
     const BYTE* litPtr = dctx->litPtr;
     const BYTE* litBufferEnd = dctx->litBufferEnd;
     const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
     const BYTE* const vBase = (const BYTE*) (dctx->virtualStart);
     const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
-    DEBUGLOG(5, "ZSTD_decompressSequences_bodySplitLitBuffer");
+    DEBUGLOG(5, "ZSTD_decompressSequences_bodySplitLitBuffer (%i seqs)", nbSeq);
-    (void)frame;
 
-    /* Regen sequences */
+    /* Literals are split between internal buffer & output buffer */
     if (nbSeq) {
         seqState_t seqState;
         dctx->fseEntropy = 1;
@@ -1386,8 +1436,7 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
                 BIT_DStream_completed < BIT_DStream_overflow);
 
         /* decompress without overrunning litPtr begins */
-        {
+        {   seq_t sequence = {0,0,0};  /* some static analyzer believe that @sequence is not initialized (it necessarily is, since for(;;) loop as at least one iteration) */
-            seq_t sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
             /* Align the decompression loop to 32 + 16 bytes.
                 *
                 * zstd compiled with gcc-9 on an Intel i9-9900k shows 10% decompression
@@ -1449,27 +1498,26 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
 #endif
 
             /* Handle the initial state where litBuffer is currently split between dst and litExtraBuffer */
-            for (; litPtr + sequence.litLength <= dctx->litBufferEnd; ) {
+            for ( ; nbSeq; nbSeq--) {
-                size_t const oneSeqSize = ZSTD_execSequenceSplitLitBuffer(op, oend, litPtr + sequence.litLength - WILDCOPY_OVERLENGTH, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
+                sequence = ZSTD_decodeSequence(&seqState, isLongOffset, nbSeq==1);
+                if (litPtr + sequence.litLength > dctx->litBufferEnd) break;
+                {   size_t const oneSeqSize = ZSTD_execSequenceSplitLitBuffer(op, oend, litPtr + sequence.litLength - WILDCOPY_OVERLENGTH, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                     assert(!ZSTD_isError(oneSeqSize));
-                if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
+                    ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
 #endif
                     if (UNLIKELY(ZSTD_isError(oneSeqSize)))
                         return oneSeqSize;
                     DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
                     op += oneSeqSize;
-                if (UNLIKELY(!--nbSeq))
+            }   }
-                    break;
+            DEBUGLOG(6, "reached: (litPtr + sequence.litLength > dctx->litBufferEnd)");
-                BIT_reloadDStream(&(seqState.DStream));
-                sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
-            }
 
             /* If there are more sequences, they will need to read literals from litExtraBuffer; copy over the remainder from dst and update litPtr and litEnd */
             if (nbSeq > 0) {
                 const size_t leftoverLit = dctx->litBufferEnd - litPtr;
-                if (leftoverLit)
+                DEBUGLOG(6, "There are %i sequences left, and %zu/%zu literals left in buffer", nbSeq, leftoverLit, sequence.litLength);
-                {
+                if (leftoverLit) {
                     RETURN_ERROR_IF(leftoverLit > (size_t)(oend - op), dstSize_tooSmall, "remaining lit must fit within dstBuffer");
                     ZSTD_safecopyDstBeforeSrc(op, litPtr, leftoverLit);
                     sequence.litLength -= leftoverLit;
@@ -1478,24 +1526,22 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
                 litPtr = dctx->litExtraBuffer;
                 litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
                 dctx->litBufferLocation = ZSTD_not_in_dst;
-                {
+                {   size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
-                    size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                     assert(!ZSTD_isError(oneSeqSize));
-                    if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
+                    ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
 #endif
                     if (UNLIKELY(ZSTD_isError(oneSeqSize)))
                         return oneSeqSize;
                     DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
                     op += oneSeqSize;
-                    if (--nbSeq)
-                        BIT_reloadDStream(&(seqState.DStream));
                 }
+                nbSeq--;
             }
         }
 
-        if (nbSeq > 0) /* there is remaining lit from extra buffer */
+        if (nbSeq > 0) {
-        {
+            /* there is remaining lit from extra buffer */
 
 #if defined(__GNUC__) && defined(__x86_64__)
             __asm__(".p2align 6");
@@ -1514,35 +1560,34 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
 #  endif
 #endif
 
-            for (; ; ) {
+            for ( ; nbSeq ; nbSeq--) {
-                seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
+                seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, nbSeq==1);
                 size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                 assert(!ZSTD_isError(oneSeqSize));
-                if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
+                ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
 #endif
                 if (UNLIKELY(ZSTD_isError(oneSeqSize)))
                     return oneSeqSize;
                 DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
                 op += oneSeqSize;
-                if (UNLIKELY(!--nbSeq))
-                    break;
-                BIT_reloadDStream(&(seqState.DStream));
             }
         }
 
         /* check if reached exact end */
         DEBUGLOG(5, "ZSTD_decompressSequences_bodySplitLitBuffer: after decode loop, remaining nbSeq : %i", nbSeq);
         RETURN_ERROR_IF(nbSeq, corruption_detected, "");
-        RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected, "");
+        DEBUGLOG(5, "bitStream : start=%p, ptr=%p, bitsConsumed=%u", seqState.DStream.start, seqState.DStream.ptr, seqState.DStream.bitsConsumed);
+        RETURN_ERROR_IF(!BIT_endOfDStream(&seqState.DStream), corruption_detected, "");
         /* save reps for next block */
         { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
     }
 
     /* last literal segment */
-    if (dctx->litBufferLocation == ZSTD_split)  /* split hasn't been reached yet, first get dst then copy litExtraBuffer */
+    if (dctx->litBufferLocation == ZSTD_split) {
-    {
+        /* split hasn't been reached yet, first get dst then copy litExtraBuffer */
-        size_t const lastLLSize = litBufferEnd - litPtr;
+        size_t const lastLLSize = (size_t)(litBufferEnd - litPtr);
+        DEBUGLOG(6, "copy last literals from segment : %u", (U32)lastLLSize);
         RETURN_ERROR_IF(lastLLSize > (size_t)(oend - op), dstSize_tooSmall, "");
         if (op != NULL) {
             ZSTD_memmove(op, litPtr, lastLLSize);
@@ -1552,15 +1597,17 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
         litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
         dctx->litBufferLocation = ZSTD_not_in_dst;
     }
-    {   size_t const lastLLSize = litBufferEnd - litPtr;
+    /* copy last literals from internal buffer */
+    {   size_t const lastLLSize = (size_t)(litBufferEnd - litPtr);
+        DEBUGLOG(6, "copy last literals from internal buffer : %u", (U32)lastLLSize);
         RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall, "");
         if (op != NULL) {
             ZSTD_memcpy(op, litPtr, lastLLSize);
             op += lastLLSize;
-        }
+    }   }
-    }
 
-    return op-ostart;
+    DEBUGLOG(6, "decoded block of size %u bytes", (U32)(op - ostart));
+    return (size_t)(op - ostart);
 }
 
 FORCE_INLINE_TEMPLATE size_t
@@ -1568,13 +1615,12 @@ DONT_VECTORIZE
 ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
     void* dst, size_t maxDstSize,
     const void* seqStart, size_t seqSize, int nbSeq,
-    const ZSTD_longOffset_e isLongOffset,
+    const ZSTD_longOffset_e isLongOffset)
-    const int frame)
 {
     const BYTE* ip = (const BYTE*)seqStart;
     const BYTE* const iend = ip + seqSize;
     BYTE* const ostart = (BYTE*)dst;
-    BYTE* const oend = dctx->litBufferLocation == ZSTD_not_in_dst ? ostart + maxDstSize : dctx->litBuffer;
+    BYTE* const oend = dctx->litBufferLocation == ZSTD_not_in_dst ? ZSTD_maybeNullPtrAdd(ostart, maxDstSize) : dctx->litBuffer;
     BYTE* op = ostart;
     const BYTE* litPtr = dctx->litPtr;
     const BYTE* const litEnd = litPtr + dctx->litSize;
@@ -1582,7 +1628,6 @@ ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
     const BYTE* const vBase = (const BYTE*)(dctx->virtualStart);
     const BYTE* const dictEnd = (const BYTE*)(dctx->dictEnd);
     DEBUGLOG(5, "ZSTD_decompressSequences_body: nbSeq = %d", nbSeq);
-    (void)frame;
 
     /* Regen sequences */
     if (nbSeq) {
@@ -1597,11 +1642,6 @@ ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
         ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
         assert(dst != NULL);
 
-        ZSTD_STATIC_ASSERT(
-            BIT_DStream_unfinished < BIT_DStream_completed &&
-            BIT_DStream_endOfBuffer < BIT_DStream_completed &&
-            BIT_DStream_completed < BIT_DStream_overflow);
-
 #if defined(__GNUC__) && defined(__x86_64__)
             __asm__(".p2align 6");
             __asm__("nop");
@@ -1616,73 +1656,70 @@ ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
 #  endif
 #endif
 
-        for ( ; ; ) {
+        for ( ; nbSeq ; nbSeq--) {
-            seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
+            seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, nbSeq==1);
             size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, prefixStart, vBase, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
             assert(!ZSTD_isError(oneSeqSize));
-            if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
+            ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
 #endif
             if (UNLIKELY(ZSTD_isError(oneSeqSize)))
                 return oneSeqSize;
             DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
             op += oneSeqSize;
-            if (UNLIKELY(!--nbSeq))
-                break;
-            BIT_reloadDStream(&(seqState.DStream));
         }
 
         /* check if reached exact end */
-        DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq);
+        assert(nbSeq == 0);
-        RETURN_ERROR_IF(nbSeq, corruption_detected, "");
+        RETURN_ERROR_IF(!BIT_endOfDStream(&seqState.DStream), corruption_detected, "");
-        RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected, "");
         /* save reps for next block */
         { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
     }
 
     /* last literal segment */
-    {   size_t const lastLLSize = litEnd - litPtr;
+    {   size_t const lastLLSize = (size_t)(litEnd - litPtr);
+        DEBUGLOG(6, "copy last literals : %u", (U32)lastLLSize);
         RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall, "");
         if (op != NULL) {
             ZSTD_memcpy(op, litPtr, lastLLSize);
             op += lastLLSize;
-        }
+    }   }
-    }
 
-    return op-ostart;
+    DEBUGLOG(6, "decoded block of size %u bytes", (U32)(op - ostart));
+    return (size_t)(op - ostart);
 }
 
 static size_t
 ZSTD_decompressSequences_default(ZSTD_DCtx* dctx,
                                  void* dst, size_t maxDstSize,
                            const void* seqStart, size_t seqSize, int nbSeq,
-                           const ZSTD_longOffset_e isLongOffset,
+                           const ZSTD_longOffset_e isLongOffset)
-                           const int frame)
 {
-    return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 
 static size_t
 ZSTD_decompressSequencesSplitLitBuffer_default(ZSTD_DCtx* dctx,
                                                void* dst, size_t maxDstSize,
                                          const void* seqStart, size_t seqSize, int nbSeq,
-                                         const ZSTD_longOffset_e isLongOffset,
+                                         const ZSTD_longOffset_e isLongOffset)
-                                         const int frame)
 {
-    return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
 
 #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
 
-FORCE_INLINE_TEMPLATE size_t
+FORCE_INLINE_TEMPLATE
-ZSTD_prefetchMatch(size_t prefetchPos, seq_t const sequence,
+
+size_t ZSTD_prefetchMatch(size_t prefetchPos, seq_t const sequence,
                    const BYTE* const prefixStart, const BYTE* const dictEnd)
 {
     prefetchPos += sequence.litLength;
     {   const BYTE* const matchBase = (sequence.offset > prefetchPos) ? dictEnd : prefixStart;
-        const BYTE* const match = matchBase + prefetchPos - sequence.offset; /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
+        /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
          * No consequence though : memory address is only used for prefetching, not for dereferencing */
+        const BYTE* const match = ZSTD_wrappedPtrSub(ZSTD_wrappedPtrAdd(matchBase, prefetchPos), sequence.offset);
         PREFETCH_L1(match); PREFETCH_L1(match+CACHELINE_SIZE);   /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
     }
     return prefetchPos + sequence.matchLength;
@@ -1697,20 +1734,18 @@ ZSTD_decompressSequencesLong_body(
                                ZSTD_DCtx* dctx,
                                void* dst, size_t maxDstSize,
                          const void* seqStart, size_t seqSize, int nbSeq,
-                         const ZSTD_longOffset_e isLongOffset,
+                         const ZSTD_longOffset_e isLongOffset)
-                         const int frame)
 {
     const BYTE* ip = (const BYTE*)seqStart;
     const BYTE* const iend = ip + seqSize;
     BYTE* const ostart = (BYTE*)dst;
-    BYTE* const oend = dctx->litBufferLocation == ZSTD_in_dst ? dctx->litBuffer : ostart + maxDstSize;
+    BYTE* const oend = dctx->litBufferLocation == ZSTD_in_dst ? dctx->litBuffer : ZSTD_maybeNullPtrAdd(ostart, maxDstSize);
     BYTE* op = ostart;
     const BYTE* litPtr = dctx->litPtr;
     const BYTE* litBufferEnd = dctx->litBufferEnd;
     const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
     const BYTE* const dictStart = (const BYTE*) (dctx->virtualStart);
     const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
-    (void)frame;
 
     /* Regen sequences */
     if (nbSeq) {
@@ -1735,20 +1770,17 @@ ZSTD_decompressSequencesLong_body(
         ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
 
         /* prepare in advance */
-        for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {
+        for (seqNb=0; seqNb<seqAdvance; seqNb++) {
-            seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
+            seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, seqNb == nbSeq-1);
             prefetchPos = ZSTD_prefetchMatch(prefetchPos, sequence, prefixStart, dictEnd);
             sequences[seqNb] = sequence;
         }
-        RETURN_ERROR_IF(seqNb<seqAdvance, corruption_detected, "");
 
         /* decompress without stomping litBuffer */
-        for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb < nbSeq); seqNb++) {
+        for (; seqNb < nbSeq; seqNb++) {
-            seq_t sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
+            seq_t sequence = ZSTD_decodeSequence(&seqState, isLongOffset, seqNb == nbSeq-1);
-            size_t oneSeqSize;
 
-            if (dctx->litBufferLocation == ZSTD_split && litPtr + sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK].litLength > dctx->litBufferEnd)
+            if (dctx->litBufferLocation == ZSTD_split && litPtr + sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK].litLength > dctx->litBufferEnd) {
-            {
                 /* lit buffer is reaching split point, empty out the first buffer and transition to litExtraBuffer */
                 const size_t leftoverLit = dctx->litBufferEnd - litPtr;
                 if (leftoverLit)
@@ -1761,26 +1793,26 @@ ZSTD_decompressSequencesLong_body(
                 litPtr = dctx->litExtraBuffer;
                 litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
                 dctx->litBufferLocation = ZSTD_not_in_dst;
-                oneSeqSize = ZSTD_execSequence(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
+                {   size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                     assert(!ZSTD_isError(oneSeqSize));
-                if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
+                    ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
 #endif
                     if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
 
                     prefetchPos = ZSTD_prefetchMatch(prefetchPos, sequence, prefixStart, dictEnd);
                     sequences[seqNb & STORED_SEQS_MASK] = sequence;
                     op += oneSeqSize;
-            }
+            }   }
             else
             {
                 /* lit buffer is either wholly contained in first or second split, or not split at all*/
-                oneSeqSize = dctx->litBufferLocation == ZSTD_split ?
+                size_t const oneSeqSize = dctx->litBufferLocation == ZSTD_split ?
                     ZSTD_execSequenceSplitLitBuffer(op, oend, litPtr + sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK].litLength - WILDCOPY_OVERLENGTH, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd) :
                     ZSTD_execSequence(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                 assert(!ZSTD_isError(oneSeqSize));
-                if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
+                ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
 #endif
                 if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
 
@@ -1789,17 +1821,15 @@ ZSTD_decompressSequencesLong_body(
                 op += oneSeqSize;
             }
         }
-        RETURN_ERROR_IF(seqNb<nbSeq, corruption_detected, "");
+        RETURN_ERROR_IF(!BIT_endOfDStream(&seqState.DStream), corruption_detected, "");
 
         /* finish queue */
         seqNb -= seqAdvance;
         for ( ; seqNb<nbSeq ; seqNb++) {
             seq_t *sequence = &(sequences[seqNb&STORED_SEQS_MASK]);
-            if (dctx->litBufferLocation == ZSTD_split && litPtr + sequence->litLength > dctx->litBufferEnd)
+            if (dctx->litBufferLocation == ZSTD_split && litPtr + sequence->litLength > dctx->litBufferEnd) {
-            {
                 const size_t leftoverLit = dctx->litBufferEnd - litPtr;
-                if (leftoverLit)
+                if (leftoverLit) {
-                {
                     RETURN_ERROR_IF(leftoverLit > (size_t)(oend - op), dstSize_tooSmall, "remaining lit must fit within dstBuffer");
                     ZSTD_safecopyDstBeforeSrc(op, litPtr, leftoverLit);
                     sequence->litLength -= leftoverLit;
@@ -1808,11 +1838,10 @@ ZSTD_decompressSequencesLong_body(
                 litPtr = dctx->litExtraBuffer;
                 litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
                 dctx->litBufferLocation = ZSTD_not_in_dst;
-                {
+                {   size_t const oneSeqSize = ZSTD_execSequence(op, oend, *sequence, &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
-                    size_t const oneSeqSize = ZSTD_execSequence(op, oend, *sequence, &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                     assert(!ZSTD_isError(oneSeqSize));
-                    if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
+                    ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
 #endif
                     if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
                     op += oneSeqSize;
@@ -1825,7 +1854,7 @@ ZSTD_decompressSequencesLong_body(
                     ZSTD_execSequence(op, oend, *sequence, &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
 #if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
                 assert(!ZSTD_isError(oneSeqSize));
-                if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
+                ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
 #endif
                 if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
                 op += oneSeqSize;
@@ -1837,8 +1866,7 @@ ZSTD_decompressSequencesLong_body(
     }
 
     /* last literal segment */
-    if (dctx->litBufferLocation == ZSTD_split)  /* first deplete literal buffer in dst, then copy litExtraBuffer */
+    if (dctx->litBufferLocation == ZSTD_split) { /* first deplete literal buffer in dst, then copy litExtraBuffer */
-    {
         size_t const lastLLSize = litBufferEnd - litPtr;
         RETURN_ERROR_IF(lastLLSize > (size_t)(oend - op), dstSize_tooSmall, "");
         if (op != NULL) {
@@ -1856,17 +1884,16 @@ ZSTD_decompressSequencesLong_body(
         }
     }
 
-    return op-ostart;
+    return (size_t)(op - ostart);
 }
 
 static size_t
 ZSTD_decompressSequencesLong_default(ZSTD_DCtx* dctx,
                                  void* dst, size_t maxDstSize,
                            const void* seqStart, size_t seqSize, int nbSeq,
-                           const ZSTD_longOffset_e isLongOffset,
+                           const ZSTD_longOffset_e isLongOffset)
-                           const int frame)
 {
-    return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
 
@@ -1880,20 +1907,18 @@ DONT_VECTORIZE
 ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx,
                                  void* dst, size_t maxDstSize,
                            const void* seqStart, size_t seqSize, int nbSeq,
-                           const ZSTD_longOffset_e isLongOffset,
+                           const ZSTD_longOffset_e isLongOffset)
-                           const int frame)
 {
-    return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 static BMI2_TARGET_ATTRIBUTE size_t
 DONT_VECTORIZE
 ZSTD_decompressSequencesSplitLitBuffer_bmi2(ZSTD_DCtx* dctx,
                                  void* dst, size_t maxDstSize,
                            const void* seqStart, size_t seqSize, int nbSeq,
-                           const ZSTD_longOffset_e isLongOffset,
+                           const ZSTD_longOffset_e isLongOffset)
-                           const int frame)
 {
-    return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
 
@@ -1902,10 +1927,9 @@ static BMI2_TARGET_ATTRIBUTE size_t
 ZSTD_decompressSequencesLong_bmi2(ZSTD_DCtx* dctx,
                                  void* dst, size_t maxDstSize,
                            const void* seqStart, size_t seqSize, int nbSeq,
-                           const ZSTD_longOffset_e isLongOffset,
+                           const ZSTD_longOffset_e isLongOffset)
-                           const int frame)
 {
-    return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
 
@@ -1915,37 +1939,34 @@ typedef size_t (*ZSTD_decompressSequences_t)(
                             ZSTD_DCtx* dctx,
                             void* dst, size_t maxDstSize,
                             const void* seqStart, size_t seqSize, int nbSeq,
-                            const ZSTD_longOffset_e isLongOffset,
+                            const ZSTD_longOffset_e isLongOffset);
-                            const int frame);
 
 #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
 static size_t
 ZSTD_decompressSequences(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
                    const void* seqStart, size_t seqSize, int nbSeq,
-                   const ZSTD_longOffset_e isLongOffset,
+                   const ZSTD_longOffset_e isLongOffset)
-                   const int frame)
 {
     DEBUGLOG(5, "ZSTD_decompressSequences");
 #if DYNAMIC_BMI2
     if (ZSTD_DCtx_get_bmi2(dctx)) {
-        return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+        return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
     }
 #endif
-    return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 static size_t
 ZSTD_decompressSequencesSplitLitBuffer(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
                                  const void* seqStart, size_t seqSize, int nbSeq,
-                                 const ZSTD_longOffset_e isLongOffset,
+                                 const ZSTD_longOffset_e isLongOffset)
-                                 const int frame)
 {
     DEBUGLOG(5, "ZSTD_decompressSequencesSplitLitBuffer");
 #if DYNAMIC_BMI2
     if (ZSTD_DCtx_get_bmi2(dctx)) {
-        return ZSTD_decompressSequencesSplitLitBuffer_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+        return ZSTD_decompressSequencesSplitLitBuffer_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
     }
 #endif
-    return ZSTD_decompressSequencesSplitLitBuffer_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+    return ZSTD_decompressSequencesSplitLitBuffer_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
 
@@ -1960,16 +1981,15 @@ static size_t
 ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,
                              void* dst, size_t maxDstSize,
                              const void* seqStart, size_t seqSize, int nbSeq,
-                             const ZSTD_longOffset_e isLongOffset,
+                             const ZSTD_longOffset_e isLongOffset)
-                             const int frame)
 {
     DEBUGLOG(5, "ZSTD_decompressSequencesLong");
 #if DYNAMIC_BMI2
     if (ZSTD_DCtx_get_bmi2(dctx)) {
-        return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+        return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
     }
 #endif
-  return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
+  return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
 }
 #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
 
@@ -2051,20 +2071,20 @@ static size_t ZSTD_maxShortOffset(void)
 size_t
 ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
                               void* dst, size_t dstCapacity,
-                        const void* src, size_t srcSize, const int frame, const streaming_operation streaming)
+                        const void* src, size_t srcSize, const streaming_operation streaming)
 {   /* blockType == blockCompressed */
     const BYTE* ip = (const BYTE*)src;
-    DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);
+    DEBUGLOG(5, "ZSTD_decompressBlock_internal (cSize : %u)", (unsigned)srcSize);
 
     /* Note : the wording of the specification
-     * allows compressed block to be sized exactly ZSTD_BLOCKSIZE_MAX.
+     * allows compressed block to be sized exactly ZSTD_blockSizeMax(dctx).
      * This generally does not happen, as it makes little sense,
      * since an uncompressed block would feature same size and have no decompression cost.
      * Also, note that decoder from reference libzstd before < v1.5.4
      * would consider this edge case as an error.
-     * As a consequence, avoid generating compressed blocks of size ZSTD_BLOCKSIZE_MAX
+     * As a consequence, avoid generating compressed blocks of size ZSTD_blockSizeMax(dctx)
      * for broader compatibility with the deployed ecosystem of zstd decoders */
-    RETURN_ERROR_IF(srcSize > ZSTD_BLOCKSIZE_MAX, srcSize_wrong, "");
+    RETURN_ERROR_IF(srcSize > ZSTD_blockSizeMax(dctx), srcSize_wrong, "");
 
     /* Decode literals section */
     {   size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize, dst, dstCapacity, streaming);
@@ -2079,8 +2099,8 @@ ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
         /* Compute the maximum block size, which must also work when !frame and fParams are unset.
          * Additionally, take the min with dstCapacity to ensure that the totalHistorySize fits in a size_t.
          */
-        size_t const blockSizeMax = MIN(dstCapacity, (frame ? dctx->fParams.blockSizeMax : ZSTD_BLOCKSIZE_MAX));
+        size_t const blockSizeMax = MIN(dstCapacity, ZSTD_blockSizeMax(dctx));
-        size_t const totalHistorySize = ZSTD_totalHistorySize((BYTE*)dst + blockSizeMax, (BYTE const*)dctx->virtualStart);
+        size_t const totalHistorySize = ZSTD_totalHistorySize(ZSTD_maybeNullPtrAdd((BYTE*)dst, blockSizeMax), (BYTE const*)dctx->virtualStart);
         /* isLongOffset must be true if there are long offsets.
          * Offsets are long if they are larger than ZSTD_maxShortOffset().
          * We don't expect that to be the case in 64-bit mode.
@@ -2145,21 +2165,22 @@ ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
         {
 #endif
 #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
-            return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
+            return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);
 #endif
         }
 
 #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
         /* else */
         if (dctx->litBufferLocation == ZSTD_split)
-            return ZSTD_decompressSequencesSplitLitBuffer(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
+            return ZSTD_decompressSequencesSplitLitBuffer(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);
         else
-            return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
+            return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);
 #endif
     }
 }
 
 
+ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
 void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst, size_t dstSize)
 {
     if (dst != dctx->previousDstEnd && dstSize > 0) {   /* not contiguous */
@@ -2176,8 +2197,10 @@ size_t ZSTD_decompressBlock_deprecated(ZSTD_DCtx* dctx,
                                  const void* src, size_t srcSize)
 {
     size_t dSize;
+    dctx->isFrameDecompression = 0;
     ZSTD_checkContinuity(dctx, dst, dstCapacity);
-    dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 0, not_streaming);
+    dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, not_streaming);
+    FORWARD_IF_ERROR(dSize, "");
     dctx->previousDstEnd = (char*)dst + dSize;
     return dSize;
 }

zstd/decompress/zstd_decompress_block.h

@@ -47,7 +47,7 @@ typedef enum {
  */
 size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
                                void* dst, size_t dstCapacity,
-                         const void* src, size_t srcSize, const int frame, const streaming_operation streaming);
+                         const void* src, size_t srcSize, const streaming_operation streaming);
 
 /* ZSTD_buildFSETable() :
  * generate FSE decoding table for one symbol (ll, ml or off)

zstd/decompress/zstd_decompress_internal.h

@@ -153,6 +153,7 @@ struct ZSTD_DCtx_s
     size_t litSize;
     size_t rleSize;
     size_t staticSize;
+    int isFrameDecompression;
 #if DYNAMIC_BMI2 != 0
     int bmi2;                     /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */
 #endif
@@ -166,6 +167,7 @@ struct ZSTD_DCtx_s
     ZSTD_DDictHashSet* ddictSet;                    /* Hash set for multiple ddicts */
     ZSTD_refMultipleDDicts_e refMultipleDDicts;     /* User specified: if == 1, will allow references to multiple DDicts. Default == 0 (disabled) */
     int disableHufAsm;
+    int maxBlockSizeParam;
 
     /* streaming */
     ZSTD_dStreamStage streamStage;

zstd/dictBuilder/cover.c

@@ -31,8 +31,8 @@
 #endif
 
 #include "../common/mem.h" /* read */
-#include "../common/pool.h"
+#include "../common/pool.h" /* POOL_ctx */
-#include "../common/threading.h"
+#include "../common/threading.h" /* ZSTD_pthread_mutex_t */
 #include "../common/zstd_internal.h" /* includes zstd.h */
 #include "../common/bits.h" /* ZSTD_highbit32 */
 #include "../zdict.h"
@@ -303,7 +303,8 @@ static int WIN_CDECL COVER_strict_cmp8(const void *lp, const void *rp) {
  */
 static const size_t *COVER_lower_bound(const size_t* first, const size_t* last,
                                        size_t value) {
-  size_t count = last - first;
+  size_t count = (size_t)(last - first);
+  assert(last >= first);
   while (count != 0) {
     size_t step = count / 2;
     const size_t *ptr = first;
@@ -549,7 +550,8 @@ static void COVER_ctx_destroy(COVER_ctx_t *ctx) {
  */
 static size_t COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
                           const size_t *samplesSizes, unsigned nbSamples,
-                          unsigned d, double splitPoint) {
+                          unsigned d, double splitPoint)
+{
   const BYTE *const samples = (const BYTE *)samplesBuffer;
   const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);
   /* Split samples into testing and training sets */
@@ -733,7 +735,7 @@ static size_t COVER_buildDictionary(const COVER_ctx_t *ctx, U32 *freqs,
   return tail;
 }
 
-ZDICTLIB_API size_t ZDICT_trainFromBuffer_cover(
+ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_cover(
     void *dictBuffer, size_t dictBufferCapacity,
     const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
     ZDICT_cover_params_t parameters)
@@ -907,8 +909,10 @@ void COVER_best_start(COVER_best_t *best) {
  * Decrements liveJobs and signals any waiting threads if liveJobs == 0.
  * If this dictionary is the best so far save it and its parameters.
  */
-void COVER_best_finish(COVER_best_t *best, ZDICT_cover_params_t parameters,
+void COVER_best_finish(COVER_best_t* best,
-                              COVER_dictSelection_t selection) {
+                      ZDICT_cover_params_t parameters,
+                      COVER_dictSelection_t selection)
+{
   void* dict = selection.dictContent;
   size_t compressedSize = selection.totalCompressedSize;
   size_t dictSize = selection.dictSize;
@@ -1119,7 +1123,7 @@ _cleanup:
   free(freqs);
 }
 
-ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(
+ZDICTLIB_STATIC_API size_t ZDICT_optimizeTrainFromBuffer_cover(
     void* dictBuffer, size_t dictBufferCapacity, const void* samplesBuffer,
     const size_t* samplesSizes, unsigned nbSamples,
     ZDICT_cover_params_t* parameters)

zstd/dictBuilder/cover.h

@@ -12,14 +12,8 @@
 #  define ZDICT_STATIC_LINKING_ONLY
 #endif
 
-#include <stdio.h>  /* fprintf */
+#include "../common/threading.h" /* ZSTD_pthread_mutex_t */
-#include <stdlib.h> /* malloc, free, qsort */
+#include "../common/mem.h"   /* U32, BYTE */
-#include <string.h> /* memset */
-#include <time.h>   /* clock */
-#include "../common/mem.h" /* read */
-#include "../common/pool.h"
-#include "../common/threading.h"
-#include "../common/zstd_internal.h" /* includes zstd.h */
 #include "../zdict.h"
 
 /**

zstd/dictBuilder/fastcover.c

@@ -545,7 +545,7 @@ FASTCOVER_convertToFastCoverParams(ZDICT_cover_params_t coverParams,
 }
 
 
-ZDICTLIB_API size_t
+ZDICTLIB_STATIC_API size_t
 ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity,
                                 const void* samplesBuffer,
                                 const size_t* samplesSizes, unsigned nbSamples,
@@ -614,7 +614,7 @@ ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity,
 }
 
 
-ZDICTLIB_API size_t
+ZDICTLIB_STATIC_API size_t
 ZDICT_optimizeTrainFromBuffer_fastCover(
                     void* dictBuffer, size_t dictBufferCapacity,
                     const void* samplesBuffer,

zstd/dictBuilder/zdict.c

@@ -74,9 +74,9 @@ static const U32 g_selectivity_default = 9;
 *  Console display
 ***************************************/
 #undef  DISPLAY
-#define DISPLAY(...)         { fprintf(stderr, __VA_ARGS__); fflush( stderr ); }
+#define DISPLAY(...)         do { fprintf(stderr, __VA_ARGS__); fflush( stderr ); } while (0)
 #undef  DISPLAYLEVEL
-#define DISPLAYLEVEL(l, ...) if (notificationLevel>=l) { DISPLAY(__VA_ARGS__); }    /* 0 : no display;   1: errors;   2: default;  3: details;  4: debug */
+#define DISPLAYLEVEL(l, ...) do { if (notificationLevel>=l) { DISPLAY(__VA_ARGS__); } } while (0)    /* 0 : no display;   1: errors;   2: default;  3: details;  4: debug */
 
 static clock_t ZDICT_clockSpan(clock_t nPrevious) { return clock() - nPrevious; }
 
@@ -477,10 +477,16 @@ static size_t ZDICT_trainBuffer_legacy(dictItem* dictList, U32 dictListSize,
     clock_t const refreshRate = CLOCKS_PER_SEC * 3 / 10;
 
 #   undef  DISPLAYUPDATE
-#   define DISPLAYUPDATE(l, ...) if (notificationLevel>=l) { \
+#   define DISPLAYUPDATE(l, ...)                                   \
-            if (ZDICT_clockSpan(displayClock) > refreshRate)  \
+        do {                                                       \
-            { displayClock = clock(); DISPLAY(__VA_ARGS__); \
+            if (notificationLevel>=l) {                            \
-            if (notificationLevel>=4) fflush(stderr); } }
+                if (ZDICT_clockSpan(displayClock) > refreshRate) { \
+                    displayClock = clock();                        \
+                    DISPLAY(__VA_ARGS__);                          \
+                }                                                  \
+                if (notificationLevel>=4) fflush(stderr);          \
+            }                                                      \
+        } while (0)
 
     /* init */
     DISPLAYLEVEL(2, "\r%70s\r", "");   /* clean display line */

zstd/zstd.h

@@ -106,7 +106,7 @@ extern "C" {
 /*------   Version   ------*/
 #define ZSTD_VERSION_MAJOR    1
 #define ZSTD_VERSION_MINOR    5
-#define ZSTD_VERSION_RELEASE  5
+#define ZSTD_VERSION_RELEASE  6
 #define ZSTD_VERSION_NUMBER  (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)
 
 /*! ZSTD_versionNumber() :
@@ -228,7 +228,7 @@ ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize)
  * for example to size a static array on stack.
  * Will produce constant value 0 if srcSize too large.
  */
-#define ZSTD_MAX_INPUT_SIZE ((sizeof(size_t)==8) ? 0xFF00FF00FF00FF00LLU : 0xFF00FF00U)
+#define ZSTD_MAX_INPUT_SIZE ((sizeof(size_t)==8) ? 0xFF00FF00FF00FF00ULL : 0xFF00FF00U)
 #define ZSTD_COMPRESSBOUND(srcSize)   (((size_t)(srcSize) >= ZSTD_MAX_INPUT_SIZE) ? 0 : (srcSize) + ((srcSize)>>8) + (((srcSize) < (128<<10)) ? (((128<<10) - (srcSize)) >> 11) /* margin, from 64 to 0 */ : 0))  /* this formula ensures that bound(A) + bound(B) <= bound(A+B) as long as A and B >= 128 KB */
 ZSTDLIB_API size_t ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case single-pass scenario */
 /* ZSTD_isError() :
@@ -249,7 +249,7 @@ ZSTDLIB_API int         ZSTD_defaultCLevel(void);           /*!< default compres
 /*= Compression context
  *  When compressing many times,
  *  it is recommended to allocate a context just once,
- *  and re-use it for each successive compression operation.
+ *  and reuse it for each successive compression operation.
  *  This will make workload friendlier for system's memory.
  *  Note : re-using context is just a speed / resource optimization.
  *         It doesn't change the compression ratio, which remains identical.
@@ -262,9 +262,9 @@ ZSTDLIB_API size_t     ZSTD_freeCCtx(ZSTD_CCtx* cctx);  /* accept NULL pointer *
 
 /*! ZSTD_compressCCtx() :
  *  Same as ZSTD_compress(), using an explicit ZSTD_CCtx.
- *  Important : in order to behave similarly to `ZSTD_compress()`,
+ *  Important : in order to mirror `ZSTD_compress()` behavior,
- *  this function compresses at requested compression level,
+ *  this function compresses at the requested compression level,
- *  __ignoring any other parameter__ .
+ *  __ignoring any other advanced parameter__ .
  *  If any advanced parameter was set using the advanced API,
  *  they will all be reset. Only `compressionLevel` remains.
  */
@@ -276,7 +276,7 @@ ZSTDLIB_API size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,
 /*= Decompression context
  *  When decompressing many times,
  *  it is recommended to allocate a context only once,
- *  and re-use it for each successive compression operation.
+ *  and reuse it for each successive compression operation.
  *  This will make workload friendlier for system's memory.
  *  Use one context per thread for parallel execution. */
 typedef struct ZSTD_DCtx_s ZSTD_DCtx;
@@ -286,7 +286,7 @@ ZSTDLIB_API size_t     ZSTD_freeDCtx(ZSTD_DCtx* dctx);  /* accept NULL pointer *
 /*! ZSTD_decompressDCtx() :
  *  Same as ZSTD_decompress(),
  *  requires an allocated ZSTD_DCtx.
- *  Compatible with sticky parameters.
+ *  Compatible with sticky parameters (see below).
  */
 ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,
                                        void* dst, size_t dstCapacity,
@@ -302,12 +302,12 @@ ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,
  *   using ZSTD_CCtx_set*() functions.
  *   Pushed parameters are sticky : they are valid for next compressed frame, and any subsequent frame.
  *   "sticky" parameters are applicable to `ZSTD_compress2()` and `ZSTD_compressStream*()` !
- *   __They do not apply to "simple" one-shot variants such as ZSTD_compressCCtx()__ .
+ *   __They do not apply to one-shot variants such as ZSTD_compressCCtx()__ .
  *
  *   It's possible to reset all parameters to "default" using ZSTD_CCtx_reset().
  *
  *   This API supersedes all other "advanced" API entry points in the experimental section.
- *   In the future, we expect to remove from experimental API entry points which are redundant with this API.
+ *   In the future, we expect to remove API entry points from experimental which are redundant with this API.
  */
 
 
@@ -390,6 +390,19 @@ typedef enum {
                               * The higher the value of selected strategy, the more complex it is,
                               * resulting in stronger and slower compression.
                               * Special: value 0 means "use default strategy". */
+
+    ZSTD_c_targetCBlockSize=130, /* v1.5.6+
+                                  * Attempts to fit compressed block size into approximatively targetCBlockSize.
+                                  * Bound by ZSTD_TARGETCBLOCKSIZE_MIN and ZSTD_TARGETCBLOCKSIZE_MAX.
+                                  * Note that it's not a guarantee, just a convergence target (default:0).
+                                  * No target when targetCBlockSize == 0.
+                                  * This is helpful in low bandwidth streaming environments to improve end-to-end latency,
+                                  * when a client can make use of partial documents (a prominent example being Chrome).
+                                  * Note: this parameter is stable since v1.5.6.
+                                  * It was present as an experimental parameter in earlier versions,
+                                  * but it's not recommended using it with earlier library versions
+                                  * due to massive performance regressions.
+                                  */
     /* LDM mode parameters */
     ZSTD_c_enableLongDistanceMatching=160, /* Enable long distance matching.
                                      * This parameter is designed to improve compression ratio
@@ -469,7 +482,6 @@ typedef enum {
      * ZSTD_c_forceMaxWindow
      * ZSTD_c_forceAttachDict
      * ZSTD_c_literalCompressionMode
-     * ZSTD_c_targetCBlockSize
      * ZSTD_c_srcSizeHint
      * ZSTD_c_enableDedicatedDictSearch
      * ZSTD_c_stableInBuffer
@@ -490,7 +502,7 @@ typedef enum {
      ZSTD_c_experimentalParam3=1000,
      ZSTD_c_experimentalParam4=1001,
      ZSTD_c_experimentalParam5=1002,
-     ZSTD_c_experimentalParam6=1003,
+     /* was ZSTD_c_experimentalParam6=1003; is now ZSTD_c_targetCBlockSize */
      ZSTD_c_experimentalParam7=1004,
      ZSTD_c_experimentalParam8=1005,
      ZSTD_c_experimentalParam9=1006,
@@ -575,6 +587,7 @@ ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
 
 /*! ZSTD_compress2() :
  *  Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.
+ *  (note that this entry point doesn't even expose a compression level parameter).
  *  ZSTD_compress2() always starts a new frame.
  *  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*()
@@ -618,6 +631,7 @@ typedef enum {
      * ZSTD_d_forceIgnoreChecksum
      * ZSTD_d_refMultipleDDicts
      * ZSTD_d_disableHuffmanAssembly
+     * ZSTD_d_maxBlockSize
      * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
      * note : never ever use experimentalParam? names directly
      */
@@ -625,7 +639,8 @@ typedef enum {
      ZSTD_d_experimentalParam2=1001,
      ZSTD_d_experimentalParam3=1002,
      ZSTD_d_experimentalParam4=1003,
-     ZSTD_d_experimentalParam5=1004
+     ZSTD_d_experimentalParam5=1004,
+     ZSTD_d_experimentalParam6=1005
 
 } ZSTD_dParameter;
 
@@ -680,14 +695,14 @@ typedef struct ZSTD_outBuffer_s {
 *  A ZSTD_CStream object is required to track streaming operation.
 *  Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.
 *  ZSTD_CStream objects can be reused multiple times on consecutive compression operations.
-*  It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
+*  It is recommended to reuse ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
 *
 *  For parallel execution, use one separate ZSTD_CStream per thread.
 *
 *  note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.
 *
 *  Parameters are sticky : when starting a new compression on the same context,
-*  it will re-use the same sticky parameters as previous compression session.
+*  it will reuse the same sticky parameters as previous compression session.
 *  When in doubt, it's recommended to fully initialize the context before usage.
 *  Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),
 *  ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to
@@ -776,6 +791,11 @@ typedef enum {
  *            only ZSTD_e_end or ZSTD_e_flush operations are allowed.
  *            Before starting a new compression job, or changing compression parameters,
  *            it is required to fully flush internal buffers.
+ *  - note: if an operation ends with an error, it may leave @cctx in an undefined state.
+ *          Therefore, it's UB to invoke ZSTD_compressStream2() of ZSTD_compressStream() on such a state.
+ *          In order to be re-employed after an error, a state must be reset,
+ *          which can be done explicitly (ZSTD_CCtx_reset()),
+ *          or is sometimes implied by methods starting a new compression job (ZSTD_initCStream(), ZSTD_compressCCtx())
  */
 ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
                                          ZSTD_outBuffer* output,
@@ -835,7 +855,7 @@ ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
 *
 *  A ZSTD_DStream object is required to track streaming operations.
 *  Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
-*  ZSTD_DStream objects can be re-used multiple times.
+*  ZSTD_DStream objects can be reused multiple times.
 *
 *  Use ZSTD_initDStream() to start a new decompression operation.
 * @return : recommended first input size
@@ -889,6 +909,12 @@ ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);
  * @return : 0 when a frame is completely decoded and fully flushed,
  *           or an error code, which can be tested using ZSTD_isError(),
  *           or any other value > 0, which means there is some decoding or flushing to do to complete current frame.
+ *
+ * Note: when an operation returns with an error code, the @zds state may be left in undefined state.
+ *       It's UB to invoke `ZSTD_decompressStream()` on such a state.
+ *       In order to re-use such a state, it must be first reset,
+ *       which can be done explicitly (`ZSTD_DCtx_reset()`),
+ *       or is implied for operations starting some new decompression job (`ZSTD_initDStream`, `ZSTD_decompressDCtx()`, `ZSTD_decompress_usingDict()`)
  */
 ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
 
@@ -1021,7 +1047,7 @@ ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
  *
  * This API allows dictionaries to be used with ZSTD_compress2(),
  * ZSTD_compressStream2(), and ZSTD_decompressDCtx().
- * Dictionaries are sticky, they remain valid when same context is re-used,
+ * Dictionaries are sticky, they remain valid when same context is reused,
  * they only reset when the context is reset
  * with ZSTD_reset_parameters or ZSTD_reset_session_and_parameters.
  * In contrast, Prefixes are single-use.
@@ -1239,7 +1265,7 @@ ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
 #define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)
 
 /* Advanced parameter bounds */
-#define ZSTD_TARGETCBLOCKSIZE_MIN   64
+#define ZSTD_TARGETCBLOCKSIZE_MIN   1340 /* suitable to fit into an ethernet / wifi / 4G transport frame */
 #define ZSTD_TARGETCBLOCKSIZE_MAX   ZSTD_BLOCKSIZE_MAX
 #define ZSTD_SRCSIZEHINT_MIN        0
 #define ZSTD_SRCSIZEHINT_MAX        INT_MAX
@@ -1527,21 +1553,34 @@ typedef enum {
 ZSTDLIB_STATIC_API size_t ZSTD_sequenceBound(size_t srcSize);
 
 /*! ZSTD_generateSequences() :
+ * WARNING: This function is meant for debugging and informational purposes ONLY!
+ * Its implementation is flawed, and it will be deleted in a future version.
+ * It is not guaranteed to succeed, as there are several cases where it will give
+ * up and fail. You should NOT use this function in production code.
+ *
+ * This function is deprecated, and will be removed in a future version.
+ *
  * Generate sequences using ZSTD_compress2(), given a source buffer.
  *
+ * @param zc The compression context to be used for ZSTD_compress2(). Set any
+ *           compression parameters you need on this context.
+ * @param outSeqs The output sequences buffer of size @p outSeqsSize
+ * @param outSeqsSize The size of the output sequences buffer.
+ *                    ZSTD_sequenceBound(srcSize) is an upper bound on the number
+ *                    of sequences that can be generated.
+ * @param src The source buffer to generate sequences from of size @p srcSize.
+ * @param srcSize The size of the source buffer.
+ *
  * Each block will end with a dummy sequence
  * with offset == 0, matchLength == 0, and litLength == length of last literals.
  * litLength may be == 0, and if so, then the sequence of (of: 0 ml: 0 ll: 0)
  * simply acts as a block delimiter.
  *
- * @zc can be used to insert custom compression params.
+ * @returns The number of sequences generated, necessarily less than
- * This function invokes ZSTD_compress2().
+ *          ZSTD_sequenceBound(srcSize), or an error code that can be checked
- *
+ *          with ZSTD_isError().
- * The output of this function can be fed into ZSTD_compressSequences() with CCtx
- * setting of ZSTD_c_blockDelimiters as ZSTD_sf_explicitBlockDelimiters
- * @return : number of sequences generated
  */
-
+ZSTD_DEPRECATED("For debugging only, will be replaced by ZSTD_extractSequences()")
 ZSTDLIB_STATIC_API size_t
 ZSTD_generateSequences(ZSTD_CCtx* zc,
                        ZSTD_Sequence* outSeqs, size_t outSeqsSize,
@@ -1640,56 +1679,59 @@ ZSTDLIB_API unsigned ZSTD_isSkippableFrame(const void* buffer, size_t size);
 /*! ZSTD_estimate*() :
  *  These functions make it possible to estimate memory usage
  *  of a future {D,C}Ctx, before its creation.
+ *  This is useful in combination with ZSTD_initStatic(),
+ *  which makes it possible to employ a static buffer for ZSTD_CCtx* state.
  *
  *  ZSTD_estimateCCtxSize() will provide a memory budget large enough
- *  for any compression level up to selected one.
+ *  to compress data of any size using one-shot compression ZSTD_compressCCtx() or ZSTD_compress2()
- *  Note : Unlike ZSTD_estimateCStreamSize*(), this estimate
+ *  associated with any compression level up to max specified one.
- *         does not include space for a window buffer.
- *         Therefore, the estimation is only guaranteed for single-shot compressions, not streaming.
  *  The estimate will assume the input may be arbitrarily large,
  *  which is the worst case.
  *
+ *  Note that the size estimation is specific for one-shot compression,
+ *  it is not valid for streaming (see ZSTD_estimateCStreamSize*())
+ *  nor other potential ways of using a ZSTD_CCtx* state.
+ *
  *  When srcSize can be bound by a known and rather "small" value,
- *  this fact can be used to provide a tighter estimation
+ *  this knowledge can be used to provide a tighter budget estimation
- *  because the CCtx compression context will need less memory.
+ *  because the ZSTD_CCtx* state will need less memory for small inputs.
- *  This tighter estimation can be provided by more advanced functions
+ *  This tighter estimation can be provided by employing more advanced functions
  *  ZSTD_estimateCCtxSize_usingCParams(), which can be used in tandem with ZSTD_getCParams(),
  *  and ZSTD_estimateCCtxSize_usingCCtxParams(), which can be used in tandem with ZSTD_CCtxParams_setParameter().
  *  Both can be used to estimate memory using custom compression parameters and arbitrary srcSize limits.
  *
  *  Note : only single-threaded compression is supported.
  *  ZSTD_estimateCCtxSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1.
- *
- *  Note 2 : ZSTD_estimateCCtxSize* functions are not compatible with the Block-Level Sequence Producer API at this time.
- *  Size estimates assume that no external sequence producer is registered.
  */
-ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize(int compressionLevel);
+ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize(int maxCompressionLevel);
 ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
 ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
 ZSTDLIB_STATIC_API size_t ZSTD_estimateDCtxSize(void);
 
 /*! ZSTD_estimateCStreamSize() :
- *  ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one.
+ *  ZSTD_estimateCStreamSize() will provide a memory budget large enough for streaming compression
- *  It will also consider src size to be arbitrarily "large", which is worst case.
+ *  using any compression level up to the max specified one.
+ *  It will also consider src size to be arbitrarily "large", which is a worst case scenario.
  *  If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.
  *  ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
  *  ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
  *  Note : CStream size estimation is only correct for single-threaded compression.
- *  ZSTD_DStream memory budget depends on window Size.
+ *  ZSTD_estimateCStreamSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1.
+ *  Note 2 : ZSTD_estimateCStreamSize* functions are not compatible with the Block-Level Sequence Producer API at this time.
+ *  Size estimates assume that no external sequence producer is registered.
+ *
+ *  ZSTD_DStream memory budget depends on frame's window Size.
  *  This information can be passed manually, using ZSTD_estimateDStreamSize,
  *  or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame();
+ *  Any frame requesting a window size larger than max specified one will be rejected.
  *  Note : if streaming is init with function ZSTD_init?Stream_usingDict(),
  *         an internal ?Dict will be created, which additional size is not estimated here.
  *         In this case, get total size by adding ZSTD_estimate?DictSize
- *  Note 2 : only single-threaded compression is supported.
- *  ZSTD_estimateCStreamSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1.
- *  Note 3 : ZSTD_estimateCStreamSize* functions are not compatible with the Block-Level Sequence Producer API at this time.
- *  Size estimates assume that no external sequence producer is registered.
  */
-ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize(int compressionLevel);
+ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize(int maxCompressionLevel);
 ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);
 ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);
-ZSTDLIB_STATIC_API size_t ZSTD_estimateDStreamSize(size_t windowSize);
+ZSTDLIB_STATIC_API size_t ZSTD_estimateDStreamSize(size_t maxWindowSize);
 ZSTDLIB_STATIC_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);
 
 /*! ZSTD_estimate?DictSize() :
@@ -1946,11 +1988,6 @@ ZSTDLIB_STATIC_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const vo
  */
 #define ZSTD_c_literalCompressionMode ZSTD_c_experimentalParam5
 
-/* Tries to fit compressed block size to be around targetCBlockSize.
- * No target when targetCBlockSize == 0.
- * There is no guarantee on compressed block size (default:0) */
-#define ZSTD_c_targetCBlockSize ZSTD_c_experimentalParam6
-
 /* User's best guess of source size.
  * Hint is not valid when srcSizeHint == 0.
  * There is no guarantee that hint is close to actual source size,
@@ -2430,6 +2467,22 @@ ZSTDLIB_STATIC_API size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParamete
  */
 #define ZSTD_d_disableHuffmanAssembly ZSTD_d_experimentalParam5
 
+/* ZSTD_d_maxBlockSize
+ * Allowed values are between 1KB and ZSTD_BLOCKSIZE_MAX (128KB).
+ * The default is ZSTD_BLOCKSIZE_MAX, and setting to 0 will set to the default.
+ *
+ * Forces the decompressor to reject blocks whose content size is
+ * larger than the configured maxBlockSize. When maxBlockSize is
+ * larger than the windowSize, the windowSize is used instead.
+ * This saves memory on the decoder when you know all blocks are small.
+ *
+ * This option is typically used in conjunction with ZSTD_c_maxBlockSize.
+ *
+ * WARNING: This causes the decoder to reject otherwise valid frames
+ * that have block sizes larger than the configured maxBlockSize.
+ */
+#define ZSTD_d_maxBlockSize ZSTD_d_experimentalParam6
+
 
 /*! ZSTD_DCtx_setFormat() :
  *  This function is REDUNDANT. Prefer ZSTD_DCtx_setParameter().
@@ -2557,7 +2610,7 @@ size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
  *       explicitly specified.
  *
  *  start a new frame, using same parameters from previous frame.
- *  This is typically useful to skip dictionary loading stage, since it will re-use it in-place.
+ *  This is typically useful to skip dictionary loading stage, since it will reuse it in-place.
  *  Note that zcs must be init at least once before using ZSTD_resetCStream().
  *  If pledgedSrcSize is not known at reset time, use macro ZSTD_CONTENTSIZE_UNKNOWN.
  *  If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end.
@@ -2633,7 +2686,7 @@ ZSTDLIB_STATIC_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const Z
  *
  *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
  *
- * re-use decompression parameters from previous init; saves dictionary loading
+ * reuse decompression parameters from previous init; saves dictionary loading
  */
 ZSTD_DEPRECATED("use ZSTD_DCtx_reset, see zstd.h for detailed instructions")
 ZSTDLIB_STATIC_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
@@ -2765,7 +2818,7 @@ ZSTDLIB_STATIC_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
 
 #define ZSTD_SEQUENCE_PRODUCER_ERROR ((size_t)(-1))
 
-typedef size_t ZSTD_sequenceProducer_F (
+typedef size_t (*ZSTD_sequenceProducer_F) (
   void* sequenceProducerState,
   ZSTD_Sequence* outSeqs, size_t outSeqsCapacity,
   const void* src, size_t srcSize,
@@ -2797,7 +2850,23 @@ ZSTDLIB_STATIC_API void
 ZSTD_registerSequenceProducer(
   ZSTD_CCtx* cctx,
   void* sequenceProducerState,
-  ZSTD_sequenceProducer_F* sequenceProducer
+  ZSTD_sequenceProducer_F sequenceProducer
+);
+
+/*! ZSTD_CCtxParams_registerSequenceProducer() :
+ * Same as ZSTD_registerSequenceProducer(), but operates on ZSTD_CCtx_params.
+ * This is used for accurate size estimation with ZSTD_estimateCCtxSize_usingCCtxParams(),
+ * which is needed when creating a ZSTD_CCtx with ZSTD_initStaticCCtx().
+ *
+ * If you are using the external sequence producer API in a scenario where ZSTD_initStaticCCtx()
+ * is required, then this function is for you. Otherwise, you probably don't need it.
+ *
+ * See tests/zstreamtest.c for example usage. */
+ZSTDLIB_STATIC_API void
+ZSTD_CCtxParams_registerSequenceProducer(
+  ZSTD_CCtx_params* params,
+  void* sequenceProducerState,
+  ZSTD_sequenceProducer_F sequenceProducer
 );
 
 
@@ -2820,7 +2889,7 @@ ZSTD_registerSequenceProducer(
 
   A ZSTD_CCtx object is required to track streaming operations.
   Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource.
-  ZSTD_CCtx object can be re-used multiple times within successive compression operations.
+  ZSTD_CCtx object can be reused multiple times within successive compression operations.
 
   Start by initializing a context.
   Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression.
@@ -2841,7 +2910,7 @@ ZSTD_registerSequenceProducer(
   It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame.
   Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders.
 
-  `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress again.
+  `ZSTD_CCtx` object can be reused (ZSTD_compressBegin()) to compress again.
 */
 
 /*=====   Buffer-less streaming compression functions  =====*/
@@ -2873,7 +2942,7 @@ size_t ZSTD_compressBegin_usingCDict_advanced(ZSTD_CCtx* const cctx, const ZSTD_
 
   A ZSTD_DCtx object is required to track streaming operations.
   Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.
-  A ZSTD_DCtx object can be re-used multiple times.
+  A ZSTD_DCtx object can be reused multiple times.
 
   First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader().
   Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough.