zstdlib 0.10.0-x64-mingw32 → 0.12.0-x64-mingw32

Sign up to get free protection for your applications and to get access to all the features.
Files changed (110) hide show
  1. checksums.yaml +4 -4
  2. data/CHANGES.md +17 -0
  3. data/ext/zstdlib_c/extconf.rb +9 -4
  4. data/ext/zstdlib_c/ruby/zlib-3.2/zstdlib.c +5090 -0
  5. data/ext/zstdlib_c/ruby/zlib-3.3/zstdlib.c +5090 -0
  6. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/adler32.c +5 -27
  7. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/compress.c +5 -16
  8. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/crc32.c +94 -161
  9. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/deflate.c +362 -434
  10. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/deflate.h +43 -12
  11. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/gzclose.c +1 -3
  12. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/gzguts.h +13 -18
  13. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/gzlib.c +28 -85
  14. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/gzread.c +23 -73
  15. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/gzwrite.c +19 -65
  16. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/infback.c +17 -30
  17. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inffast.c +1 -4
  18. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inffast.h +1 -1
  19. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inflate.c +36 -102
  20. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inftrees.c +6 -11
  21. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inftrees.h +6 -6
  22. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/trees.c +290 -355
  23. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/uncompr.c +4 -12
  24. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/zconf.h +23 -14
  25. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/zlib.h +202 -199
  26. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/zutil.c +18 -44
  27. data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/zutil.h +13 -33
  28. data/ext/zstdlib_c/zstd-1.5.5/lib/common/allocations.h +55 -0
  29. data/ext/zstdlib_c/zstd-1.5.5/lib/common/bits.h +200 -0
  30. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/bitstream.h +19 -60
  31. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/compiler.h +26 -3
  32. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/cpu.h +1 -1
  33. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/debug.c +1 -1
  34. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/debug.h +1 -1
  35. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/entropy_common.c +12 -40
  36. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/error_private.c +9 -2
  37. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/error_private.h +1 -1
  38. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/fse.h +5 -83
  39. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/fse_decompress.c +7 -99
  40. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/huf.h +65 -156
  41. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/mem.h +39 -46
  42. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/pool.c +26 -10
  43. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/pool.h +7 -1
  44. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/portability_macros.h +22 -3
  45. data/ext/zstdlib_c/zstd-1.5.5/lib/common/threading.c +176 -0
  46. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/threading.h +5 -10
  47. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/xxhash.c +2 -2
  48. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/xxhash.h +8 -8
  49. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/zstd_common.c +1 -36
  50. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/zstd_deps.h +1 -1
  51. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/zstd_internal.h +17 -118
  52. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/common/zstd_trace.h +3 -3
  53. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/clevels.h +1 -1
  54. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/fse_compress.c +7 -124
  55. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/hist.c +1 -1
  56. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/hist.h +1 -1
  57. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/huf_compress.c +234 -169
  58. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress.c +1243 -538
  59. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress_internal.h +225 -151
  60. data/ext/zstdlib_c/zstd-1.5.5/lib/compress/zstd_compress_literals.c +235 -0
  61. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress_literals.h +16 -8
  62. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress_sequences.c +3 -3
  63. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress_sequences.h +1 -1
  64. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress_superblock.c +25 -21
  65. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_compress_superblock.h +1 -1
  66. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_cwksp.h +128 -62
  67. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_double_fast.c +95 -33
  68. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_double_fast.h +3 -2
  69. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_fast.c +433 -148
  70. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_fast.h +3 -2
  71. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_lazy.c +398 -345
  72. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_lazy.h +4 -2
  73. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_ldm.c +5 -5
  74. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_ldm.h +1 -1
  75. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_ldm_geartab.h +1 -1
  76. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_opt.c +106 -80
  77. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstd_opt.h +1 -1
  78. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstdmt_compress.c +17 -9
  79. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/compress/zstdmt_compress.h +1 -1
  80. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/huf_decompress.c +434 -441
  81. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/huf_decompress_amd64.S +30 -39
  82. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/zstd_ddict.c +4 -4
  83. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/zstd_ddict.h +1 -1
  84. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/zstd_decompress.c +205 -80
  85. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/zstd_decompress_block.c +201 -81
  86. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/zstd_decompress_block.h +6 -1
  87. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/decompress/zstd_decompress_internal.h +4 -2
  88. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/zdict.h +53 -31
  89. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/zstd.h +580 -135
  90. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/lib/zstd_errors.h +27 -8
  91. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/gzclose.c +1 -1
  92. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/gzcompatibility.h +8 -8
  93. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/gzguts.h +10 -10
  94. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/gzlib.c +3 -3
  95. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/gzread.c +10 -10
  96. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/gzwrite.c +5 -5
  97. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/zstd_zlibwrapper.c +46 -44
  98. data/ext/zstdlib_c/{zstd-1.5.2 → zstd-1.5.5}/zlibWrapper/zstd_zlibwrapper.h +4 -1
  99. data/lib/2.4/zstdlib_c.so +0 -0
  100. data/lib/2.5/zstdlib_c.so +0 -0
  101. data/lib/2.6/zstdlib_c.so +0 -0
  102. data/lib/2.7/zstdlib_c.so +0 -0
  103. data/lib/3.0/zstdlib_c.so +0 -0
  104. metadata +108 -104
  105. data/ext/zstdlib_c/zstd-1.5.2/lib/common/threading.c +0 -122
  106. data/ext/zstdlib_c/zstd-1.5.2/lib/compress/zstd_compress_literals.c +0 -159
  107. /data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/crc32.h +0 -0
  108. /data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inffixed.h +0 -0
  109. /data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/inflate.h +0 -0
  110. /data/ext/zstdlib_c/{zlib-1.2.12 → zlib-1.3.1}/trees.h +0 -0
@@ -1,7 +1,7 @@
1
1
  /* ******************************************************************
2
2
  * huff0 huffman codec,
3
3
  * part of Finite State Entropy library
4
- * Copyright (c) Yann Collet, Facebook, Inc.
4
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
5
5
  *
6
6
  * You can contact the author at :
7
7
  * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
@@ -21,99 +21,22 @@ extern "C" {
21
21
 
22
22
  /* *** Dependencies *** */
23
23
  #include "zstd_deps.h" /* size_t */
24
-
25
-
26
- /* *** library symbols visibility *** */
27
- /* Note : when linking with -fvisibility=hidden on gcc, or by default on Visual,
28
- * HUF symbols remain "private" (internal symbols for library only).
29
- * Set macro FSE_DLL_EXPORT to 1 if you want HUF symbols visible on DLL interface */
30
- #if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)
31
- # define HUF_PUBLIC_API __attribute__ ((visibility ("default")))
32
- #elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) /* Visual expected */
33
- # define HUF_PUBLIC_API __declspec(dllexport)
34
- #elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)
35
- # define HUF_PUBLIC_API __declspec(dllimport) /* not required, just to generate faster code (saves a function pointer load from IAT and an indirect jump) */
36
- #else
37
- # define HUF_PUBLIC_API
38
- #endif
39
-
40
-
41
- /* ========================== */
42
- /* *** simple functions *** */
43
- /* ========================== */
44
-
45
- /** HUF_compress() :
46
- * Compress content from buffer 'src', of size 'srcSize', into buffer 'dst'.
47
- * 'dst' buffer must be already allocated.
48
- * Compression runs faster if `dstCapacity` >= HUF_compressBound(srcSize).
49
- * `srcSize` must be <= `HUF_BLOCKSIZE_MAX` == 128 KB.
50
- * @return : size of compressed data (<= `dstCapacity`).
51
- * Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!
52
- * if HUF_isError(return), compression failed (more details using HUF_getErrorName())
53
- */
54
- HUF_PUBLIC_API size_t HUF_compress(void* dst, size_t dstCapacity,
55
- const void* src, size_t srcSize);
56
-
57
- /** HUF_decompress() :
58
- * Decompress HUF data from buffer 'cSrc', of size 'cSrcSize',
59
- * into already allocated buffer 'dst', of minimum size 'dstSize'.
60
- * `originalSize` : **must** be the ***exact*** size of original (uncompressed) data.
61
- * Note : in contrast with FSE, HUF_decompress can regenerate
62
- * RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data,
63
- * because it knows size to regenerate (originalSize).
64
- * @return : size of regenerated data (== originalSize),
65
- * or an error code, which can be tested using HUF_isError()
66
- */
67
- HUF_PUBLIC_API size_t HUF_decompress(void* dst, size_t originalSize,
68
- const void* cSrc, size_t cSrcSize);
24
+ #include "mem.h" /* U32 */
25
+ #define FSE_STATIC_LINKING_ONLY
26
+ #include "fse.h"
69
27
 
70
28
 
71
29
  /* *** Tool functions *** */
72
- #define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */
73
- HUF_PUBLIC_API size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */
30
+ #define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */
31
+ size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */
74
32
 
75
33
  /* Error Management */
76
- HUF_PUBLIC_API unsigned HUF_isError(size_t code); /**< tells if a return value is an error code */
77
- HUF_PUBLIC_API const char* HUF_getErrorName(size_t code); /**< provides error code string (useful for debugging) */
34
+ unsigned HUF_isError(size_t code); /**< tells if a return value is an error code */
35
+ const char* HUF_getErrorName(size_t code); /**< provides error code string (useful for debugging) */
78
36
 
79
37
 
80
- /* *** Advanced function *** */
81
-
82
- /** HUF_compress2() :
83
- * Same as HUF_compress(), but offers control over `maxSymbolValue` and `tableLog`.
84
- * `maxSymbolValue` must be <= HUF_SYMBOLVALUE_MAX .
85
- * `tableLog` must be `<= HUF_TABLELOG_MAX` . */
86
- HUF_PUBLIC_API size_t HUF_compress2 (void* dst, size_t dstCapacity,
87
- const void* src, size_t srcSize,
88
- unsigned maxSymbolValue, unsigned tableLog);
89
-
90
- /** HUF_compress4X_wksp() :
91
- * Same as HUF_compress2(), but uses externally allocated `workSpace`.
92
- * `workspace` must be at least as large as HUF_WORKSPACE_SIZE */
93
38
  #define HUF_WORKSPACE_SIZE ((8 << 10) + 512 /* sorting scratch space */)
94
39
  #define HUF_WORKSPACE_SIZE_U64 (HUF_WORKSPACE_SIZE / sizeof(U64))
95
- HUF_PUBLIC_API size_t HUF_compress4X_wksp (void* dst, size_t dstCapacity,
96
- const void* src, size_t srcSize,
97
- unsigned maxSymbolValue, unsigned tableLog,
98
- void* workSpace, size_t wkspSize);
99
-
100
- #endif /* HUF_H_298734234 */
101
-
102
- /* ******************************************************************
103
- * WARNING !!
104
- * The following section contains advanced and experimental definitions
105
- * which shall never be used in the context of a dynamic library,
106
- * because they are not guaranteed to remain stable in the future.
107
- * Only consider them in association with static linking.
108
- * *****************************************************************/
109
- #if defined(HUF_STATIC_LINKING_ONLY) && !defined(HUF_H_HUF_STATIC_LINKING_ONLY)
110
- #define HUF_H_HUF_STATIC_LINKING_ONLY
111
-
112
- /* *** Dependencies *** */
113
- #include "mem.h" /* U32 */
114
- #define FSE_STATIC_LINKING_ONLY
115
- #include "fse.h"
116
-
117
40
 
118
41
  /* *** Constants *** */
119
42
  #define HUF_TABLELOG_MAX 12 /* max runtime value of tableLog (due to static allocation); can be modified up to HUF_TABLELOG_ABSOLUTEMAX */
@@ -154,25 +77,49 @@ typedef U32 HUF_DTable;
154
77
  /* ****************************************
155
78
  * Advanced decompression functions
156
79
  ******************************************/
157
- size_t HUF_decompress4X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */
158
- #ifndef HUF_FORCE_DECOMPRESS_X1
159
- size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */
160
- #endif
161
80
 
162
- size_t HUF_decompress4X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< decodes RLE and uncompressed */
163
- size_t HUF_decompress4X_hufOnly(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< considers RLE and uncompressed as errors */
164
- size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< considers RLE and uncompressed as errors */
165
- size_t HUF_decompress4X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */
166
- size_t HUF_decompress4X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< single-symbol decoder */
167
- #ifndef HUF_FORCE_DECOMPRESS_X1
168
- size_t HUF_decompress4X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */
169
- size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< double-symbols decoder */
170
- #endif
81
+ /**
82
+ * Huffman flags bitset.
83
+ * For all flags, 0 is the default value.
84
+ */
85
+ typedef enum {
86
+ /**
87
+ * If compiled with DYNAMIC_BMI2: Set flag only if the CPU supports BMI2 at runtime.
88
+ * Otherwise: Ignored.
89
+ */
90
+ HUF_flags_bmi2 = (1 << 0),
91
+ /**
92
+ * If set: Test possible table depths to find the one that produces the smallest header + encoded size.
93
+ * If unset: Use heuristic to find the table depth.
94
+ */
95
+ HUF_flags_optimalDepth = (1 << 1),
96
+ /**
97
+ * If set: If the previous table can encode the input, always reuse the previous table.
98
+ * If unset: If the previous table can encode the input, reuse the previous table if it results in a smaller output.
99
+ */
100
+ HUF_flags_preferRepeat = (1 << 2),
101
+ /**
102
+ * If set: Sample the input and check if the sample is uncompressible, if it is then don't attempt to compress.
103
+ * If unset: Always histogram the entire input.
104
+ */
105
+ HUF_flags_suspectUncompressible = (1 << 3),
106
+ /**
107
+ * If set: Don't use assembly implementations
108
+ * If unset: Allow using assembly implementations
109
+ */
110
+ HUF_flags_disableAsm = (1 << 4),
111
+ /**
112
+ * If set: Don't use the fast decoding loop, always use the fallback decoding loop.
113
+ * If unset: Use the fast decoding loop when possible.
114
+ */
115
+ HUF_flags_disableFast = (1 << 5)
116
+ } HUF_flags_e;
171
117
 
172
118
 
173
119
  /* ****************************************
174
120
  * HUF detailed API
175
121
  * ****************************************/
122
+ #define HUF_OPTIMAL_DEPTH_THRESHOLD ZSTD_btultra
176
123
 
177
124
  /*! HUF_compress() does the following:
178
125
  * 1. count symbol occurrence from source[] into table count[] using FSE_count() (exposed within "fse.h")
@@ -185,12 +132,12 @@ size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
185
132
  * For example, it's possible to compress several blocks using the same 'CTable',
186
133
  * or to save and regenerate 'CTable' using external methods.
187
134
  */
188
- unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);
189
- size_t HUF_buildCTable (HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue, unsigned maxNbBits); /* @return : maxNbBits; CTable and count can overlap. In which case, CTable will overwrite count content */
190
- size_t HUF_writeCTable (void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog);
135
+ unsigned HUF_minTableLog(unsigned symbolCardinality);
136
+ unsigned HUF_cardinality(const unsigned* count, unsigned maxSymbolValue);
137
+ unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, void* workSpace,
138
+ size_t wkspSize, HUF_CElt* table, const unsigned* count, int flags); /* table is used as scratch space for building and testing tables, not a return value */
191
139
  size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog, void* workspace, size_t workspaceSize);
192
- size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable);
193
- size_t HUF_compress4X_usingCTable_bmi2(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int bmi2);
140
+ size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int flags);
194
141
  size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
195
142
  int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
196
143
 
@@ -199,6 +146,7 @@ typedef enum {
199
146
  HUF_repeat_check, /**< Can use the previous table but it must be checked. Note : The previous table must have been constructed by HUF_compress{1, 4}X_repeat */
200
147
  HUF_repeat_valid /**< Can use the previous table and it is assumed to be valid */
201
148
  } HUF_repeat;
149
+
202
150
  /** HUF_compress4X_repeat() :
203
151
  * Same as HUF_compress4X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
204
152
  * If it uses hufTable it does not modify hufTable or repeat.
@@ -209,13 +157,13 @@ size_t HUF_compress4X_repeat(void* dst, size_t dstSize,
209
157
  const void* src, size_t srcSize,
210
158
  unsigned maxSymbolValue, unsigned tableLog,
211
159
  void* workSpace, size_t wkspSize, /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
212
- HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2, unsigned suspectUncompressible);
160
+ HUF_CElt* hufTable, HUF_repeat* repeat, int flags);
213
161
 
214
162
  /** HUF_buildCTable_wksp() :
215
163
  * Same as HUF_buildCTable(), but using externally allocated scratch buffer.
216
164
  * `workSpace` must be aligned on 4-bytes boundaries, and its size must be >= HUF_CTABLE_WORKSPACE_SIZE.
217
165
  */
218
- #define HUF_CTABLE_WORKSPACE_SIZE_U32 (2*HUF_SYMBOLVALUE_MAX +1 +1)
166
+ #define HUF_CTABLE_WORKSPACE_SIZE_U32 ((4 * (HUF_SYMBOLVALUE_MAX + 1)) + 192)
219
167
  #define HUF_CTABLE_WORKSPACE_SIZE (HUF_CTABLE_WORKSPACE_SIZE_U32 * sizeof(unsigned))
220
168
  size_t HUF_buildCTable_wksp (HUF_CElt* tree,
221
169
  const unsigned* count, U32 maxSymbolValue, U32 maxNbBits,
@@ -241,7 +189,7 @@ size_t HUF_readStats_wksp(BYTE* huffWeight, size_t hwSize,
241
189
  U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr,
242
190
  const void* src, size_t srcSize,
243
191
  void* workspace, size_t wkspSize,
244
- int bmi2);
192
+ int flags);
245
193
 
246
194
  /** HUF_readCTable() :
247
195
  * Loading a CTable saved with HUF_writeCTable() */
@@ -279,32 +227,12 @@ U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize);
279
227
  #define HUF_DECOMPRESS_WORKSPACE_SIZE ((2 << 10) + (1 << 9))
280
228
  #define HUF_DECOMPRESS_WORKSPACE_SIZE_U32 (HUF_DECOMPRESS_WORKSPACE_SIZE / sizeof(U32))
281
229
 
282
- #ifndef HUF_FORCE_DECOMPRESS_X2
283
- size_t HUF_readDTableX1 (HUF_DTable* DTable, const void* src, size_t srcSize);
284
- size_t HUF_readDTableX1_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);
285
- #endif
286
- #ifndef HUF_FORCE_DECOMPRESS_X1
287
- size_t HUF_readDTableX2 (HUF_DTable* DTable, const void* src, size_t srcSize);
288
- size_t HUF_readDTableX2_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);
289
- #endif
290
-
291
- size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
292
- #ifndef HUF_FORCE_DECOMPRESS_X2
293
- size_t HUF_decompress4X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
294
- #endif
295
- #ifndef HUF_FORCE_DECOMPRESS_X1
296
- size_t HUF_decompress4X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
297
- #endif
298
-
299
230
 
300
231
  /* ====================== */
301
232
  /* single stream variants */
302
233
  /* ====================== */
303
234
 
304
- size_t HUF_compress1X (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);
305
- size_t HUF_compress1X_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize); /**< `workSpace` must be a table of at least HUF_WORKSPACE_SIZE_U64 U64 */
306
- size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable);
307
- size_t HUF_compress1X_usingCTable_bmi2(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int bmi2);
235
+ size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int flags);
308
236
  /** HUF_compress1X_repeat() :
309
237
  * Same as HUF_compress1X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
310
238
  * If it uses hufTable it does not modify hufTable or repeat.
@@ -315,49 +243,30 @@ size_t HUF_compress1X_repeat(void* dst, size_t dstSize,
315
243
  const void* src, size_t srcSize,
316
244
  unsigned maxSymbolValue, unsigned tableLog,
317
245
  void* workSpace, size_t wkspSize, /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
318
- HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2, unsigned suspectUncompressible);
319
-
320
- size_t HUF_decompress1X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* single-symbol decoder */
321
- #ifndef HUF_FORCE_DECOMPRESS_X1
322
- size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* double-symbol decoder */
323
- #endif
324
-
325
- size_t HUF_decompress1X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);
326
- size_t HUF_decompress1X_DCtx_wksp (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);
327
- #ifndef HUF_FORCE_DECOMPRESS_X2
328
- size_t HUF_decompress1X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */
329
- size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< single-symbol decoder */
330
- #endif
331
- #ifndef HUF_FORCE_DECOMPRESS_X1
332
- size_t HUF_decompress1X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */
333
- size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< double-symbols decoder */
334
- #endif
246
+ HUF_CElt* hufTable, HUF_repeat* repeat, int flags);
335
247
 
336
- size_t HUF_decompress1X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable); /**< automatic selection of sing or double symbol decoder, based on DTable */
337
- #ifndef HUF_FORCE_DECOMPRESS_X2
338
- size_t HUF_decompress1X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
339
- #endif
248
+ size_t HUF_decompress1X_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int flags);
340
249
  #ifndef HUF_FORCE_DECOMPRESS_X1
341
- size_t HUF_decompress1X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
250
+ size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int flags); /**< double-symbols decoder */
342
251
  #endif
343
252
 
344
253
  /* BMI2 variants.
345
254
  * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.
346
255
  */
347
- size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
256
+ size_t HUF_decompress1X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int flags);
348
257
  #ifndef HUF_FORCE_DECOMPRESS_X2
349
- size_t HUF_decompress1X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
258
+ size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int flags);
350
259
  #endif
351
- size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
352
- size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
260
+ size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int flags);
261
+ size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int flags);
353
262
  #ifndef HUF_FORCE_DECOMPRESS_X2
354
- size_t HUF_readDTableX1_wksp_bmi2(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int bmi2);
263
+ size_t HUF_readDTableX1_wksp(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int flags);
355
264
  #endif
356
265
  #ifndef HUF_FORCE_DECOMPRESS_X1
357
- size_t HUF_readDTableX2_wksp_bmi2(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int bmi2);
266
+ size_t HUF_readDTableX2_wksp(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int flags);
358
267
  #endif
359
268
 
360
- #endif /* HUF_STATIC_LINKING_ONLY */
269
+ #endif /* HUF_H_298734234 */
361
270
 
362
271
  #if defined (__cplusplus)
363
272
  }
@@ -1,5 +1,5 @@
1
1
  /*
2
- * Copyright (c) Yann Collet, Facebook, Inc.
2
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
3
3
  * All rights reserved.
4
4
  *
5
5
  * This source code is licensed under both the BSD-style license (found in the
@@ -133,21 +133,15 @@ MEM_STATIC size_t MEM_swapST(size_t in);
133
133
  /*-**************************************************************
134
134
  * Memory I/O Implementation
135
135
  *****************************************************************/
136
- /* MEM_FORCE_MEMORY_ACCESS :
137
- * By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
138
- * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
139
- * The below switch allow to select different access method for improved performance.
140
- * Method 0 (default) : use `memcpy()`. Safe and portable.
141
- * Method 1 : `__packed` statement. It depends on compiler extension (i.e., not portable).
142
- * This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
136
+ /* MEM_FORCE_MEMORY_ACCESS : For accessing unaligned memory:
137
+ * Method 0 : always use `memcpy()`. Safe and portable.
138
+ * Method 1 : Use compiler extension to set unaligned access.
143
139
  * Method 2 : direct access. This method is portable but violate C standard.
144
140
  * It can generate buggy code on targets depending on alignment.
145
- * In some circumstances, it's the only known way to get the most performance (i.e. GCC + ARMv6)
146
- * See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
147
- * Prefer these methods in priority order (0 > 1 > 2)
141
+ * Default : method 1 if supported, else method 0
148
142
  */
149
143
  #ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
150
- # if defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
144
+ # ifdef __GNUC__
151
145
  # define MEM_FORCE_MEMORY_ACCESS 1
152
146
  # endif
153
147
  #endif
@@ -190,30 +184,19 @@ MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; }
190
184
 
191
185
  #elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
192
186
 
193
- /* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
194
- /* currently only defined for gcc and icc */
195
- #if defined(_MSC_VER) || (defined(__INTEL_COMPILER) && defined(WIN32))
196
- __pragma( pack(push, 1) )
197
- typedef struct { U16 v; } unalign16;
198
- typedef struct { U32 v; } unalign32;
199
- typedef struct { U64 v; } unalign64;
200
- typedef struct { size_t v; } unalignArch;
201
- __pragma( pack(pop) )
202
- #else
203
- typedef struct { U16 v; } __attribute__((packed)) unalign16;
204
- typedef struct { U32 v; } __attribute__((packed)) unalign32;
205
- typedef struct { U64 v; } __attribute__((packed)) unalign64;
206
- typedef struct { size_t v; } __attribute__((packed)) unalignArch;
207
- #endif
187
+ typedef __attribute__((aligned(1))) U16 unalign16;
188
+ typedef __attribute__((aligned(1))) U32 unalign32;
189
+ typedef __attribute__((aligned(1))) U64 unalign64;
190
+ typedef __attribute__((aligned(1))) size_t unalignArch;
208
191
 
209
- MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign16*)ptr)->v; }
210
- MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign32*)ptr)->v; }
211
- MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign64*)ptr)->v; }
212
- MEM_STATIC size_t MEM_readST(const void* ptr) { return ((const unalignArch*)ptr)->v; }
192
+ MEM_STATIC U16 MEM_read16(const void* ptr) { return *(const unalign16*)ptr; }
193
+ MEM_STATIC U32 MEM_read32(const void* ptr) { return *(const unalign32*)ptr; }
194
+ MEM_STATIC U64 MEM_read64(const void* ptr) { return *(const unalign64*)ptr; }
195
+ MEM_STATIC size_t MEM_readST(const void* ptr) { return *(const unalignArch*)ptr; }
213
196
 
214
- MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign16*)memPtr)->v = value; }
215
- MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign32*)memPtr)->v = value; }
216
- MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign64*)memPtr)->v = value; }
197
+ MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(unalign16*)memPtr = value; }
198
+ MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(unalign32*)memPtr = value; }
199
+ MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(unalign64*)memPtr = value; }
217
200
 
218
201
  #else
219
202
 
@@ -257,6 +240,14 @@ MEM_STATIC void MEM_write64(void* memPtr, U64 value)
257
240
 
258
241
  #endif /* MEM_FORCE_MEMORY_ACCESS */
259
242
 
243
+ MEM_STATIC U32 MEM_swap32_fallback(U32 in)
244
+ {
245
+ return ((in << 24) & 0xff000000 ) |
246
+ ((in << 8) & 0x00ff0000 ) |
247
+ ((in >> 8) & 0x0000ff00 ) |
248
+ ((in >> 24) & 0x000000ff );
249
+ }
250
+
260
251
  MEM_STATIC U32 MEM_swap32(U32 in)
261
252
  {
262
253
  #if defined(_MSC_VER) /* Visual Studio */
@@ -265,22 +256,13 @@ MEM_STATIC U32 MEM_swap32(U32 in)
265
256
  || (defined(__clang__) && __has_builtin(__builtin_bswap32))
266
257
  return __builtin_bswap32(in);
267
258
  #else
268
- return ((in << 24) & 0xff000000 ) |
269
- ((in << 8) & 0x00ff0000 ) |
270
- ((in >> 8) & 0x0000ff00 ) |
271
- ((in >> 24) & 0x000000ff );
259
+ return MEM_swap32_fallback(in);
272
260
  #endif
273
261
  }
274
262
 
275
- MEM_STATIC U64 MEM_swap64(U64 in)
263
+ MEM_STATIC U64 MEM_swap64_fallback(U64 in)
276
264
  {
277
- #if defined(_MSC_VER) /* Visual Studio */
278
- return _byteswap_uint64(in);
279
- #elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
280
- || (defined(__clang__) && __has_builtin(__builtin_bswap64))
281
- return __builtin_bswap64(in);
282
- #else
283
- return ((in << 56) & 0xff00000000000000ULL) |
265
+ return ((in << 56) & 0xff00000000000000ULL) |
284
266
  ((in << 40) & 0x00ff000000000000ULL) |
285
267
  ((in << 24) & 0x0000ff0000000000ULL) |
286
268
  ((in << 8) & 0x000000ff00000000ULL) |
@@ -288,6 +270,17 @@ MEM_STATIC U64 MEM_swap64(U64 in)
288
270
  ((in >> 24) & 0x0000000000ff0000ULL) |
289
271
  ((in >> 40) & 0x000000000000ff00ULL) |
290
272
  ((in >> 56) & 0x00000000000000ffULL);
273
+ }
274
+
275
+ MEM_STATIC U64 MEM_swap64(U64 in)
276
+ {
277
+ #if defined(_MSC_VER) /* Visual Studio */
278
+ return _byteswap_uint64(in);
279
+ #elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
280
+ || (defined(__clang__) && __has_builtin(__builtin_bswap64))
281
+ return __builtin_bswap64(in);
282
+ #else
283
+ return MEM_swap64_fallback(in);
291
284
  #endif
292
285
  }
293
286
 
@@ -1,5 +1,5 @@
1
1
  /*
2
- * Copyright (c) Yann Collet, Facebook, Inc.
2
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
3
3
  * All rights reserved.
4
4
  *
5
5
  * This source code is licensed under both the BSD-style license (found in the
@@ -10,9 +10,9 @@
10
10
 
11
11
 
12
12
  /* ====== Dependencies ======= */
13
+ #include "../common/allocations.h" /* ZSTD_customCalloc, ZSTD_customFree */
13
14
  #include "zstd_deps.h" /* size_t */
14
15
  #include "debug.h" /* assert */
15
- #include "zstd_internal.h" /* ZSTD_customMalloc, ZSTD_customFree */
16
16
  #include "pool.h"
17
17
 
18
18
  /* ====== Compiler specifics ====== */
@@ -96,9 +96,7 @@ static void* POOL_thread(void* opaque) {
96
96
  /* If the intended queue size was 0, signal after finishing job */
97
97
  ZSTD_pthread_mutex_lock(&ctx->queueMutex);
98
98
  ctx->numThreadsBusy--;
99
- if (ctx->queueSize == 1) {
100
- ZSTD_pthread_cond_signal(&ctx->queuePushCond);
101
- }
99
+ ZSTD_pthread_cond_signal(&ctx->queuePushCond);
102
100
  ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
103
101
  }
104
102
  } /* for (;;) */
@@ -128,7 +126,7 @@ POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
128
126
  * empty and full queues.
129
127
  */
130
128
  ctx->queueSize = queueSize + 1;
131
- ctx->queue = (POOL_job*)ZSTD_customMalloc(ctx->queueSize * sizeof(POOL_job), customMem);
129
+ ctx->queue = (POOL_job*)ZSTD_customCalloc(ctx->queueSize * sizeof(POOL_job), customMem);
132
130
  ctx->queueHead = 0;
133
131
  ctx->queueTail = 0;
134
132
  ctx->numThreadsBusy = 0;
@@ -142,7 +140,7 @@ POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
142
140
  }
143
141
  ctx->shutdown = 0;
144
142
  /* Allocate space for the thread handles */
145
- ctx->threads = (ZSTD_pthread_t*)ZSTD_customMalloc(numThreads * sizeof(ZSTD_pthread_t), customMem);
143
+ ctx->threads = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), customMem);
146
144
  ctx->threadCapacity = 0;
147
145
  ctx->customMem = customMem;
148
146
  /* Check for errors */
@@ -175,7 +173,7 @@ static void POOL_join(POOL_ctx* ctx) {
175
173
  /* Join all of the threads */
176
174
  { size_t i;
177
175
  for (i = 0; i < ctx->threadCapacity; ++i) {
178
- ZSTD_pthread_join(ctx->threads[i], NULL); /* note : could fail */
176
+ ZSTD_pthread_join(ctx->threads[i]); /* note : could fail */
179
177
  } }
180
178
  }
181
179
 
@@ -190,6 +188,17 @@ void POOL_free(POOL_ctx *ctx) {
190
188
  ZSTD_customFree(ctx, ctx->customMem);
191
189
  }
192
190
 
191
+ /*! POOL_joinJobs() :
192
+ * Waits for all queued jobs to finish executing.
193
+ */
194
+ void POOL_joinJobs(POOL_ctx* ctx) {
195
+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);
196
+ while(!ctx->queueEmpty || ctx->numThreadsBusy > 0) {
197
+ ZSTD_pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);
198
+ }
199
+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
200
+ }
201
+
193
202
  void ZSTD_freeThreadPool (ZSTD_threadPool* pool) {
194
203
  POOL_free (pool);
195
204
  }
@@ -211,7 +220,7 @@ static int POOL_resize_internal(POOL_ctx* ctx, size_t numThreads)
211
220
  return 0;
212
221
  }
213
222
  /* numThreads > threadCapacity */
214
- { ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_customMalloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
223
+ { ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
215
224
  if (!threadPool) return 1;
216
225
  /* replace existing thread pool */
217
226
  ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));
@@ -262,7 +271,9 @@ static int isQueueFull(POOL_ctx const* ctx) {
262
271
  static void
263
272
  POOL_add_internal(POOL_ctx* ctx, POOL_function function, void *opaque)
264
273
  {
265
- POOL_job const job = {function, opaque};
274
+ POOL_job job;
275
+ job.function = function;
276
+ job.opaque = opaque;
266
277
  assert(ctx != NULL);
267
278
  if (ctx->shutdown) return;
268
279
 
@@ -330,6 +341,11 @@ void POOL_free(POOL_ctx* ctx) {
330
341
  (void)ctx;
331
342
  }
332
343
 
344
+ void POOL_joinJobs(POOL_ctx* ctx){
345
+ assert(!ctx || ctx == &g_poolCtx);
346
+ (void)ctx;
347
+ }
348
+
333
349
  int POOL_resize(POOL_ctx* ctx, size_t numThreads) {
334
350
  (void)ctx; (void)numThreads;
335
351
  return 0;
@@ -1,5 +1,5 @@
1
1
  /*
2
- * Copyright (c) Yann Collet, Facebook, Inc.
2
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
3
3
  * All rights reserved.
4
4
  *
5
5
  * This source code is licensed under both the BSD-style license (found in the
@@ -38,6 +38,12 @@ POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
38
38
  */
39
39
  void POOL_free(POOL_ctx* ctx);
40
40
 
41
+
42
+ /*! POOL_joinJobs() :
43
+ * Waits for all queued jobs to finish executing.
44
+ */
45
+ void POOL_joinJobs(POOL_ctx* ctx);
46
+
41
47
  /*! POOL_resize() :
42
48
  * Expands or shrinks pool's number of threads.
43
49
  * This is more efficient than releasing + creating a new context,
@@ -1,5 +1,5 @@
1
1
  /*
2
- * Copyright (c) Facebook, Inc.
2
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
3
3
  * All rights reserved.
4
4
  *
5
5
  * This source code is licensed under both the BSD-style license (found in the
@@ -12,7 +12,7 @@
12
12
  #define ZSTD_PORTABILITY_MACROS_H
13
13
 
14
14
  /**
15
- * This header file contains macro defintions to support portability.
15
+ * This header file contains macro definitions to support portability.
16
16
  * This header is shared between C and ASM code, so it MUST only
17
17
  * contain macro definitions. It MUST not contain any C code.
18
18
  *
@@ -88,7 +88,7 @@
88
88
  #endif
89
89
 
90
90
  /**
91
- * Only enable assembly for GNUC comptabile compilers,
91
+ * Only enable assembly for GNUC compatible compilers,
92
92
  * because other platforms may not support GAS assembly syntax.
93
93
  *
94
94
  * Only enable assembly for Linux / MacOS, other platforms may
@@ -134,4 +134,23 @@
134
134
  # define ZSTD_ENABLE_ASM_X86_64_BMI2 0
135
135
  #endif
136
136
 
137
+ /*
138
+ * For x86 ELF targets, add .note.gnu.property section for Intel CET in
139
+ * assembly sources when CET is enabled.
140
+ *
141
+ * Additionally, any function that may be called indirectly must begin
142
+ * with ZSTD_CET_ENDBRANCH.
143
+ */
144
+ #if defined(__ELF__) && (defined(__x86_64__) || defined(__i386__)) \
145
+ && defined(__has_include)
146
+ # if __has_include(<cet.h>)
147
+ # include <cet.h>
148
+ # define ZSTD_CET_ENDBRANCH _CET_ENDBR
149
+ # endif
150
+ #endif
151
+
152
+ #ifndef ZSTD_CET_ENDBRANCH
153
+ # define ZSTD_CET_ENDBRANCH
154
+ #endif
155
+
137
156
  #endif /* ZSTD_PORTABILITY_MACROS_H */