inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 1 | /* |
W. Felix Handte | 5d693cc | 2022-12-20 12:49:47 -0500 | [diff] [blame] | 2 | * Copyright (c) Meta Platforms, Inc. and affiliates. |
Yann Collet | 4ded9e5 | 2016-08-30 10:04:33 -0700 | [diff] [blame] | 3 | * All rights reserved. |
| 4 | * |
Yann Collet | 32fb407 | 2017-08-18 16:52:05 -0700 | [diff] [blame] | 5 | * This source code is licensed under both the BSD-style license (found in the |
| 6 | * LICENSE file in the root directory of this source tree) and the GPLv2 (found |
| 7 | * in the COPYING file in the root directory of this source tree). |
Yann Collet | 3128e03 | 2017-09-08 00:09:23 -0700 | [diff] [blame] | 8 | * You may select, at your option, one of the above-listed licenses. |
Yann Collet | 4ded9e5 | 2016-08-30 10:04:33 -0700 | [diff] [blame] | 9 | */ |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 10 | #if defined (__cplusplus) |
| 11 | extern "C" { |
| 12 | #endif |
| 13 | |
Nick Terrell | 05c00f2 | 2016-11-29 11:46:37 -0800 | [diff] [blame] | 14 | #ifndef ZSTD_H_235446 |
| 15 | #define ZSTD_H_235446 |
| 16 | |
Yann Collet | 39ceef2 | 2023-01-30 19:06:39 -0800 | [diff] [blame] | 17 | /* ====== Dependencies ======*/ |
Nick Magerko | fee8fbc | 2019-08-19 12:58:54 -0700 | [diff] [blame] | 18 | #include <limits.h> /* INT_MAX */ |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 19 | #include <stddef.h> /* size_t */ |
| 20 | |
| 21 | |
Yann Collet | 426a9d4 | 2016-12-07 16:39:34 -0800 | [diff] [blame] | 22 | /* ===== ZSTDLIB_API : control library symbols visibility ===== */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 23 | #ifndef ZSTDLIB_VISIBLE |
Nick Terrell | 358a237 | 2022-12-15 15:46:34 -0800 | [diff] [blame] | 24 | /* Backwards compatibility with old macro name */ |
| 25 | # ifdef ZSTDLIB_VISIBILITY |
| 26 | # define ZSTDLIB_VISIBLE ZSTDLIB_VISIBILITY |
| 27 | # elif defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__) |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 28 | # define ZSTDLIB_VISIBLE __attribute__ ((visibility ("default"))) |
Nick Terrell | a128040 | 2017-05-19 18:01:59 -0700 | [diff] [blame] | 29 | # else |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 30 | # define ZSTDLIB_VISIBLE |
Nick Terrell | 358a237 | 2022-12-15 15:46:34 -0800 | [diff] [blame] | 31 | # endif |
| 32 | #endif |
| 33 | |
| 34 | #ifndef ZSTDLIB_HIDDEN |
| 35 | # if defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__) |
| 36 | # define ZSTDLIB_HIDDEN __attribute__ ((visibility ("hidden"))) |
| 37 | # else |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 38 | # define ZSTDLIB_HIDDEN |
Nick Terrell | a128040 | 2017-05-19 18:01:59 -0700 | [diff] [blame] | 39 | # endif |
Nick Terrell | 8de46ab | 2016-12-16 13:27:30 -0800 | [diff] [blame] | 40 | #endif |
Nick Terrell | 358a237 | 2022-12-15 15:46:34 -0800 | [diff] [blame] | 41 | |
Nick Terrell | 8de46ab | 2016-12-16 13:27:30 -0800 | [diff] [blame] | 42 | #if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1) |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 43 | # define ZSTDLIB_API __declspec(dllexport) ZSTDLIB_VISIBLE |
Nick Terrell | 8de46ab | 2016-12-16 13:27:30 -0800 | [diff] [blame] | 44 | #elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1) |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 45 | # define ZSTDLIB_API __declspec(dllimport) ZSTDLIB_VISIBLE /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/ |
Nick Terrell | 8de46ab | 2016-12-16 13:27:30 -0800 | [diff] [blame] | 46 | #else |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 47 | # define ZSTDLIB_API ZSTDLIB_VISIBLE |
Christophe Chevalier | c6e8453 | 2015-12-07 17:44:09 +0100 | [diff] [blame] | 48 | #endif |
| 49 | |
Nick Terrell | 0f4fd28 | 2022-08-01 11:52:14 -0700 | [diff] [blame] | 50 | /* Deprecation warnings : |
| 51 | * Should these warnings be a problem, it is generally possible to disable them, |
| 52 | * typically with -Wno-deprecated-declarations for gcc or _CRT_SECURE_NO_WARNINGS in Visual. |
| 53 | * Otherwise, it's also possible to define ZSTD_DISABLE_DEPRECATE_WARNINGS. |
| 54 | */ |
| 55 | #ifdef ZSTD_DISABLE_DEPRECATE_WARNINGS |
| 56 | # define ZSTD_DEPRECATED(message) /* disable deprecation warnings */ |
| 57 | #else |
| 58 | # if defined (__cplusplus) && (__cplusplus >= 201402) /* C++14 or greater */ |
| 59 | # define ZSTD_DEPRECATED(message) [[deprecated(message)]] |
| 60 | # elif (defined(GNUC) && (GNUC > 4 || (GNUC == 4 && GNUC_MINOR >= 5))) || defined(__clang__) |
| 61 | # define ZSTD_DEPRECATED(message) __attribute__((deprecated(message))) |
| 62 | # elif defined(__GNUC__) && (__GNUC__ >= 3) |
| 63 | # define ZSTD_DEPRECATED(message) __attribute__((deprecated)) |
| 64 | # elif defined(_MSC_VER) |
| 65 | # define ZSTD_DEPRECATED(message) __declspec(deprecated(message)) |
| 66 | # else |
| 67 | # pragma message("WARNING: You need to implement ZSTD_DEPRECATED for this compiler") |
| 68 | # define ZSTD_DEPRECATED(message) |
| 69 | # endif |
| 70 | #endif /* ZSTD_DISABLE_DEPRECATE_WARNINGS */ |
| 71 | |
Christophe Chevalier | c6e8453 | 2015-12-07 17:44:09 +0100 | [diff] [blame] | 72 | |
W. Felix Handte | e589ac6 | 2018-08-13 14:57:19 -0700 | [diff] [blame] | 73 | /******************************************************************************* |
inikep | 2d26133 | 2016-10-06 16:28:21 +0200 | [diff] [blame] | 74 | Introduction |
| 75 | |
W. Felix Handte | e589ac6 | 2018-08-13 14:57:19 -0700 | [diff] [blame] | 76 | zstd, short for Zstandard, is a fast lossless compression algorithm, targeting |
| 77 | real-time compression scenarios at zlib-level and better compression ratios. |
| 78 | The zstd compression library provides in-memory compression and decompression |
| 79 | functions. |
| 80 | |
| 81 | The library supports regular compression levels from 1 up to ZSTD_maxCLevel(), |
| 82 | which is currently 22. Levels >= 20, labeled `--ultra`, should be used with |
| 83 | caution, as they require more memory. The library also offers negative |
Yann Collet | 7269fe6 | 2018-09-14 16:06:35 -0700 | [diff] [blame] | 84 | compression levels, which extend the range of speed vs. ratio preferences. |
| 85 | The lower the level, the faster the speed (at the cost of compression). |
W. Felix Handte | e589ac6 | 2018-08-13 14:57:19 -0700 | [diff] [blame] | 86 | |
inikep | 2d26133 | 2016-10-06 16:28:21 +0200 | [diff] [blame] | 87 | Compression can be done in: |
| 88 | - a single step (described as Simple API) |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 89 | - a single step, reusing a context (described as Explicit context) |
Yann Collet | 37d1300 | 2016-10-24 17:22:12 -0700 | [diff] [blame] | 90 | - unbounded multiple steps (described as Streaming compression) |
inikep | 2d26133 | 2016-10-06 16:28:21 +0200 | [diff] [blame] | 91 | |
W. Felix Handte | e589ac6 | 2018-08-13 14:57:19 -0700 | [diff] [blame] | 92 | The compression ratio achievable on small data can be highly improved using |
| 93 | a dictionary. Dictionary compression can be performed in: |
| 94 | - a single step (described as Simple dictionary API) |
| 95 | - a single step, reusing a dictionary (described as Bulk-processing |
| 96 | dictionary API) |
| 97 | |
| 98 | Advanced experimental functions can be accessed using |
| 99 | `#define ZSTD_STATIC_LINKING_ONLY` before including zstd.h. |
| 100 | |
| 101 | Advanced experimental APIs should never be used with a dynamically-linked |
| 102 | library. They are not "stable"; their definitions or signatures may change in |
| 103 | the future. Only static linking is allowed. |
| 104 | *******************************************************************************/ |
inikep | 2d26133 | 2016-10-06 16:28:21 +0200 | [diff] [blame] | 105 | |
| 106 | /*------ Version ------*/ |
Yann Collet | 901e85f | 2016-08-31 07:51:25 -0700 | [diff] [blame] | 107 | #define ZSTD_VERSION_MAJOR 1 |
sen | 9c23ea9 | 2021-05-11 16:32:09 -0400 | [diff] [blame] | 108 | #define ZSTD_VERSION_MINOR 5 |
Yann Collet | 686e7e4 | 2024-03-14 15:38:14 -0700 | [diff] [blame] | 109 | #define ZSTD_VERSION_RELEASE 6 |
Yann Collet | bfff899 | 2017-05-16 16:12:23 -0700 | [diff] [blame] | 110 | #define ZSTD_VERSION_NUMBER (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE) |
animalize | 6365e0e | 2020-08-27 13:51:22 +0800 | [diff] [blame] | 111 | |
| 112 | /*! ZSTD_versionNumber() : |
| 113 | * Return runtime library version, the value is (MAJOR*100*100 + MINOR*100 + RELEASE). */ |
| 114 | ZSTDLIB_API unsigned ZSTD_versionNumber(void); |
Yann Collet | bfff899 | 2017-05-16 16:12:23 -0700 | [diff] [blame] | 115 | |
Yann Collet | e02808f | 2016-04-20 22:46:16 +0200 | [diff] [blame] | 116 | #define ZSTD_LIB_VERSION ZSTD_VERSION_MAJOR.ZSTD_VERSION_MINOR.ZSTD_VERSION_RELEASE |
| 117 | #define ZSTD_QUOTE(str) #str |
| 118 | #define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str) |
| 119 | #define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION) |
animalize | 6365e0e | 2020-08-27 13:51:22 +0800 | [diff] [blame] | 120 | |
| 121 | /*! ZSTD_versionString() : |
| 122 | * Return runtime library version, like "1.4.5". Requires v1.3.0+. */ |
| 123 | ZSTDLIB_API const char* ZSTD_versionString(void); |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 124 | |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 125 | /* ************************************* |
| 126 | * Default constant |
| 127 | ***************************************/ |
Duc Ngo | e8ef725 | 2018-06-11 10:01:35 -0700 | [diff] [blame] | 128 | #ifndef ZSTD_CLEVEL_DEFAULT |
| 129 | # define ZSTD_CLEVEL_DEFAULT 3 |
| 130 | #endif |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 131 | |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 132 | /* ************************************* |
| 133 | * Constants |
| 134 | ***************************************/ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 135 | |
| 136 | /* All magic numbers are supposed read/written to/from files/memory using little-endian convention */ |
| 137 | #define ZSTD_MAGICNUMBER 0xFD2FB528 /* valid since v0.8.0 */ |
| 138 | #define ZSTD_MAGIC_DICTIONARY 0xEC30A437 /* valid since v0.7.0 */ |
| 139 | #define ZSTD_MAGIC_SKIPPABLE_START 0x184D2A50 /* all 16 values, from 0x184D2A50 to 0x184D2A5F, signal the beginning of a skippable frame */ |
| 140 | #define ZSTD_MAGIC_SKIPPABLE_MASK 0xFFFFFFF0 |
| 141 | |
| 142 | #define ZSTD_BLOCKSIZELOG_MAX 17 |
| 143 | #define ZSTD_BLOCKSIZE_MAX (1<<ZSTD_BLOCKSIZELOG_MAX) |
| 144 | |
| 145 | |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 146 | /*************************************** |
Yann Collet | cf05b9d | 2016-07-18 16:52:10 +0200 | [diff] [blame] | 147 | * Simple API |
Yann Collet | 7010c27 | 2015-10-21 09:07:25 +0100 | [diff] [blame] | 148 | ***************************************/ |
Yann Collet | 953ce72 | 2016-02-04 15:28:14 +0100 | [diff] [blame] | 149 | /*! ZSTD_compress() : |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 150 | * Compresses `src` content as a single zstd compressed frame into already allocated `dst`. |
Nick Terrell | c40c737 | 2023-03-07 15:42:51 -0800 | [diff] [blame] | 151 | * NOTE: Providing `dstCapacity >= ZSTD_compressBound(srcSize)` guarantees that zstd will have |
| 152 | * enough space to successfully compress the data. |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 153 | * @return : compressed size written into `dst` (<= `dstCapacity), |
| 154 | * or an error code if it fails (which can be tested using ZSTD_isError()). */ |
Yann Collet | cf05b9d | 2016-07-18 16:52:10 +0200 | [diff] [blame] | 155 | ZSTDLIB_API size_t ZSTD_compress( void* dst, size_t dstCapacity, |
| 156 | const void* src, size_t srcSize, |
| 157 | int compressionLevel); |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 158 | |
Yann Collet | 953ce72 | 2016-02-04 15:28:14 +0100 | [diff] [blame] | 159 | /*! ZSTD_decompress() : |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 160 | * `compressedSize` : must be the _exact_ size of some number of compressed and/or skippable frames. |
Yann Collet | fe234bf | 2017-06-19 15:23:19 -0700 | [diff] [blame] | 161 | * `dstCapacity` is an upper bound of originalSize to regenerate. |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 162 | * If user cannot imply a maximum upper bound, it's better to use streaming mode to decompress data. |
| 163 | * @return : the number of bytes decompressed into `dst` (<= `dstCapacity`), |
| 164 | * or an errorCode if it fails (which can be tested using ZSTD_isError()). */ |
Yann Collet | 7d968c7 | 2016-02-03 02:11:32 +0100 | [diff] [blame] | 165 | ZSTDLIB_API size_t ZSTD_decompress( void* dst, size_t dstCapacity, |
Christophe Chevalier | c6e8453 | 2015-12-07 17:44:09 +0100 | [diff] [blame] | 166 | const void* src, size_t compressedSize); |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 167 | |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 168 | /*! ZSTD_getFrameContentSize() : requires v1.3.0+ |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 169 | * `src` should point to the start of a ZSTD encoded frame. |
| 170 | * `srcSize` must be at least as large as the frame header. |
| 171 | * hint : any size >= `ZSTD_frameHeaderSize_max` is large enough. |
Yann Collet | 9b979d0 | 2018-05-31 11:12:18 -0700 | [diff] [blame] | 172 | * @return : - decompressed size of `src` frame content, if known |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 173 | * - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined |
| 174 | * - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small) |
| 175 | * note 1 : a 0 return value means the frame is valid but "empty". |
| 176 | * note 2 : decompressed size is an optional field, it may not be present, typically in streaming mode. |
| 177 | * When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size. |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 178 | * In which case, it's necessary to use streaming mode to decompress data. |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 179 | * Optionally, application can rely on some implicit limit, |
| 180 | * as ZSTD_decompress() only needs an upper bound of decompressed size. |
| 181 | * (For example, data could be necessarily cut into blocks <= 16 KB). |
Yann Collet | 9b979d0 | 2018-05-31 11:12:18 -0700 | [diff] [blame] | 182 | * note 3 : decompressed size is always present when compression is completed using single-pass functions, |
| 183 | * such as ZSTD_compress(), ZSTD_compressCCtx() ZSTD_compress_usingDict() or ZSTD_compress_usingCDict(). |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 184 | * note 4 : decompressed size can be very large (64-bits value), |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 185 | * potentially larger than what local system can handle as a single memory segment. |
| 186 | * In which case, it's necessary to use streaming mode to decompress data. |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 187 | * note 5 : If source is untrusted, decompressed size could be wrong or intentionally modified. |
| 188 | * Always ensure return value fits within application's authorized limits. |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 189 | * Each application can set its own limits. |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 190 | * note 6 : This function replaces ZSTD_getDecompressedSize() */ |
| 191 | #define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1) |
| 192 | #define ZSTD_CONTENTSIZE_ERROR (0ULL - 2) |
| 193 | ZSTDLIB_API unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize); |
| 194 | |
| 195 | /*! ZSTD_getDecompressedSize() : |
| 196 | * NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize(). |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 197 | * Both functions work the same way, but ZSTD_getDecompressedSize() blends |
| 198 | * "empty", "unknown" and "error" results to the same return value (0), |
| 199 | * while ZSTD_getFrameContentSize() gives them separate return values. |
Yann Collet | 9b979d0 | 2018-05-31 11:12:18 -0700 | [diff] [blame] | 200 | * @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise. */ |
Nick Terrell | 0f4fd28 | 2022-08-01 11:52:14 -0700 | [diff] [blame] | 201 | ZSTD_DEPRECATED("Replaced by ZSTD_getFrameContentSize") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 202 | ZSTDLIB_API |
Nick Terrell | 0f4fd28 | 2022-08-01 11:52:14 -0700 | [diff] [blame] | 203 | unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize); |
Yann Collet | ac175d4 | 2016-09-13 00:51:47 +0200 | [diff] [blame] | 204 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 205 | /*! ZSTD_findFrameCompressedSize() : Requires v1.4.0+ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 206 | * `src` should point to the start of a ZSTD frame or skippable frame. |
| 207 | * `srcSize` must be >= first frame size |
| 208 | * @return : the compressed size of the first frame starting at `src`, |
| 209 | * suitable to pass as `srcSize` to `ZSTD_decompress` or similar, |
| 210 | * or an error code if input is invalid */ |
| 211 | ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize); |
| 212 | |
Yann Collet | 4110534 | 2016-07-27 15:09:11 +0200 | [diff] [blame] | 213 | |
Yann Collet | cf05b9d | 2016-07-18 16:52:10 +0200 | [diff] [blame] | 214 | /*====== Helper functions ======*/ |
Yann Collet | 45ed0df | 2022-12-15 15:23:15 -0800 | [diff] [blame] | 215 | /* ZSTD_compressBound() : |
| 216 | * maximum compressed size in worst case single-pass scenario. |
| 217 | * When invoking `ZSTD_compress()` or any other one-pass compression function, |
Yann Collet | 97f63ce | 2022-12-16 12:28:26 -0800 | [diff] [blame] | 218 | * it's recommended to provide @dstCapacity >= ZSTD_compressBound(srcSize) |
| 219 | * as it eliminates one potential failure scenario, |
| 220 | * aka not enough room in dst buffer to write the compressed frame. |
| 221 | * Note : ZSTD_compressBound() itself can fail, if @srcSize > ZSTD_MAX_INPUT_SIZE . |
Yann Collet | 45ed0df | 2022-12-15 15:23:15 -0800 | [diff] [blame] | 222 | * In which case, ZSTD_compressBound() will return an error code |
| 223 | * which can be tested using ZSTD_isError(). |
| 224 | * |
| 225 | * ZSTD_COMPRESSBOUND() : |
| 226 | * same as ZSTD_compressBound(), but as a macro. |
| 227 | * It can be used to produce constants, which can be useful for static allocation, |
| 228 | * for example to size a static array on stack. |
| 229 | * Will produce constant value 0 if srcSize too large. |
| 230 | */ |
Yann Collet | 94a2f27 | 2023-05-31 13:29:53 -0700 | [diff] [blame] | 231 | #define ZSTD_MAX_INPUT_SIZE ((sizeof(size_t)==8) ? 0xFF00FF00FF00FF00ULL : 0xFF00FF00U) |
Yann Collet | 2f4238e | 2022-12-16 15:58:25 -0800 | [diff] [blame] | 232 | #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 */ |
Yann Collet | 45ed0df | 2022-12-15 15:23:15 -0800 | [diff] [blame] | 233 | ZSTDLIB_API size_t ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case single-pass scenario */ |
| 234 | /* ZSTD_isError() : |
| 235 | * Most ZSTD_* functions returning a size_t value can be tested for error, |
| 236 | * using ZSTD_isError(). |
| 237 | * @return 1 if error, 0 otherwise |
| 238 | */ |
shakeelrao | 8930c3c | 2019-02-28 22:55:18 -0800 | [diff] [blame] | 239 | ZSTDLIB_API unsigned ZSTD_isError(size_t code); /*!< tells if a `size_t` function result is an error code */ |
| 240 | ZSTDLIB_API const char* ZSTD_getErrorName(size_t code); /*!< provides readable string from an error code */ |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 241 | ZSTDLIB_API int ZSTD_minCLevel(void); /*!< minimum negative compression level allowed, requires v1.4.0+ */ |
shakeelrao | 8930c3c | 2019-02-28 22:55:18 -0800 | [diff] [blame] | 242 | ZSTDLIB_API int ZSTD_maxCLevel(void); /*!< maximum compression level available */ |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 243 | ZSTDLIB_API int ZSTD_defaultCLevel(void); /*!< default compression level, specified by ZSTD_CLEVEL_DEFAULT, requires v1.5.0+ */ |
Yann Collet | 7010c27 | 2015-10-21 09:07:25 +0100 | [diff] [blame] | 244 | |
| 245 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 246 | /*************************************** |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 247 | * Explicit context |
Yann Collet | 7010c27 | 2015-10-21 09:07:25 +0100 | [diff] [blame] | 248 | ***************************************/ |
Yann Collet | 37d1300 | 2016-10-24 17:22:12 -0700 | [diff] [blame] | 249 | /*= Compression context |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 250 | * When compressing many times, |
Yann Collet | b8ec4b0 | 2019-07-09 11:43:59 -0700 | [diff] [blame] | 251 | * it is recommended to allocate a context just once, |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 252 | * and reuse it for each successive compression operation. |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 253 | * This will make workload friendlier for system's memory. |
Yann Collet | b8ec4b0 | 2019-07-09 11:43:59 -0700 | [diff] [blame] | 254 | * Note : re-using context is just a speed / resource optimization. |
| 255 | * It doesn't change the compression ratio, which remains identical. |
| 256 | * Note 2 : In multi-threaded environments, |
| 257 | * use one different context per thread for parallel execution. |
| 258 | */ |
Yann Collet | 87c18b2 | 2016-08-26 01:43:47 +0200 | [diff] [blame] | 259 | typedef struct ZSTD_CCtx_s ZSTD_CCtx; |
Christophe Chevalier | c6e8453 | 2015-12-07 17:44:09 +0100 | [diff] [blame] | 260 | ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void); |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 261 | ZSTDLIB_API size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx); /* accept NULL pointer */ |
Yann Collet | 7010c27 | 2015-10-21 09:07:25 +0100 | [diff] [blame] | 262 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 263 | /*! ZSTD_compressCCtx() : |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 264 | * Same as ZSTD_compress(), using an explicit ZSTD_CCtx. |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 265 | * Important : in order to mirror `ZSTD_compress()` behavior, |
| 266 | * this function compresses at the requested compression level, |
| 267 | * __ignoring any other advanced parameter__ . |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 268 | * If any advanced parameter was set using the advanced API, |
| 269 | * they will all be reset. Only `compressionLevel` remains. |
| 270 | */ |
Yann Collet | 3e042d5 | 2018-12-04 17:30:58 -0800 | [diff] [blame] | 271 | ZSTDLIB_API size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx, |
Yann Collet | a5ffe3d | 2017-05-12 16:29:19 -0700 | [diff] [blame] | 272 | void* dst, size_t dstCapacity, |
| 273 | const void* src, size_t srcSize, |
| 274 | int compressionLevel); |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 275 | |
Sean Purcell | dec2b96 | 2017-03-14 11:24:09 -0700 | [diff] [blame] | 276 | /*= Decompression context |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 277 | * When decompressing many times, |
Yann Collet | a5ffe3d | 2017-05-12 16:29:19 -0700 | [diff] [blame] | 278 | * it is recommended to allocate a context only once, |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 279 | * and reuse it for each successive compression operation. |
Yann Collet | 71ddeb6 | 2017-04-20 22:54:54 -0700 | [diff] [blame] | 280 | * This will make workload friendlier for system's memory. |
Yann Collet | a5ffe3d | 2017-05-12 16:29:19 -0700 | [diff] [blame] | 281 | * Use one context per thread for parallel execution. */ |
Yann Collet | 87c18b2 | 2016-08-26 01:43:47 +0200 | [diff] [blame] | 282 | typedef struct ZSTD_DCtx_s ZSTD_DCtx; |
Yann Collet | ecd651b | 2016-01-07 15:35:18 +0100 | [diff] [blame] | 283 | ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx(void); |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 284 | ZSTDLIB_API size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx); /* accept NULL pointer */ |
Yann Collet | ecd651b | 2016-01-07 15:35:18 +0100 | [diff] [blame] | 285 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 286 | /*! ZSTD_decompressDCtx() : |
Yann Collet | 3e042d5 | 2018-12-04 17:30:58 -0800 | [diff] [blame] | 287 | * Same as ZSTD_decompress(), |
| 288 | * requires an allocated ZSTD_DCtx. |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 289 | * Compatible with sticky parameters (see below). |
Yann Collet | 3e042d5 | 2018-12-04 17:30:58 -0800 | [diff] [blame] | 290 | */ |
| 291 | ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx, |
Yann Collet | a5ffe3d | 2017-05-12 16:29:19 -0700 | [diff] [blame] | 292 | void* dst, size_t dstCapacity, |
| 293 | const void* src, size_t srcSize); |
Yann Collet | 7d968c7 | 2016-02-03 02:11:32 +0100 | [diff] [blame] | 294 | |
| 295 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 296 | /********************************************* |
| 297 | * Advanced compression API (Requires v1.4.0+) |
| 298 | **********************************************/ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 299 | |
| 300 | /* API design : |
Yann Collet | 34e146f | 2018-12-04 10:28:36 -0800 | [diff] [blame] | 301 | * Parameters are pushed one by one into an existing context, |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 302 | * using ZSTD_CCtx_set*() functions. |
Yann Collet | 96d8874 | 2018-12-06 10:14:34 -0800 | [diff] [blame] | 303 | * Pushed parameters are sticky : they are valid for next compressed frame, and any subsequent frame. |
Yann Collet | 34e146f | 2018-12-04 10:28:36 -0800 | [diff] [blame] | 304 | * "sticky" parameters are applicable to `ZSTD_compress2()` and `ZSTD_compressStream*()` ! |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 305 | * __They do not apply to one-shot variants such as ZSTD_compressCCtx()__ . |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 306 | * |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 307 | * It's possible to reset all parameters to "default" using ZSTD_CCtx_reset(). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 308 | * |
Dimitris Apostolou | ebbd675 | 2021-11-13 10:04:04 +0200 | [diff] [blame] | 309 | * This API supersedes all other "advanced" API entry points in the experimental section. |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 310 | * In the future, we expect to remove API entry points from experimental which are redundant with this API. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 311 | */ |
Yann Collet | 2c8fde5 | 2018-11-13 17:36:35 -0800 | [diff] [blame] | 312 | |
| 313 | |
Yann Collet | 2c8fde5 | 2018-11-13 17:36:35 -0800 | [diff] [blame] | 314 | /* Compression strategies, listed from fastest to strongest */ |
| 315 | typedef enum { ZSTD_fast=1, |
| 316 | ZSTD_dfast=2, |
| 317 | ZSTD_greedy=3, |
| 318 | ZSTD_lazy=4, |
| 319 | ZSTD_lazy2=5, |
| 320 | ZSTD_btlazy2=6, |
| 321 | ZSTD_btopt=7, |
Yann Collet | e9448cd | 2018-12-06 13:38:09 -0800 | [diff] [blame] | 322 | ZSTD_btultra=8, |
| 323 | ZSTD_btultra2=9 |
Yann Collet | be9e561 | 2018-12-06 15:00:52 -0800 | [diff] [blame] | 324 | /* note : new strategies _might_ be added in the future. |
| 325 | Only the order (from fast to strong) is guaranteed */ |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 326 | } ZSTD_strategy; |
| 327 | |
Sen Huang | 3c595a4 | 2021-04-16 10:10:26 -0700 | [diff] [blame] | 328 | typedef enum { |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 329 | |
Nick Terrell | 787b769 | 2019-03-13 15:23:24 -0700 | [diff] [blame] | 330 | /* compression parameters |
| 331 | * Note: When compressing with a ZSTD_CDict these parameters are superseded |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 332 | * by the parameters used to construct the ZSTD_CDict. |
| 333 | * See ZSTD_CCtx_refCDict() for more info (superseded-by-cdict). */ |
| 334 | ZSTD_c_compressionLevel=100, /* Set compression parameters according to pre-defined cLevel table. |
| 335 | * Note that exact compression parameters are dynamically determined, |
| 336 | * depending on both compression level and srcSize (when known). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 337 | * Default level is ZSTD_CLEVEL_DEFAULT==3. |
| 338 | * Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 339 | * Note 1 : it's possible to pass a negative compression level. |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 340 | * Note 2 : setting a level does not automatically set all other compression parameters |
| 341 | * to default. Setting this will however eventually dynamically impact the compression |
| 342 | * parameters which have not been manually set. The manually set |
Bimba Shrestha | cba46e9 | 2020-03-03 13:12:02 -0800 | [diff] [blame] | 343 | * ones will 'stick'. */ |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 344 | /* Advanced compression parameters : |
| 345 | * It's possible to pin down compression parameters to some specific values. |
| 346 | * In which case, these values are no longer dynamically selected by the compressor */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 347 | ZSTD_c_windowLog=101, /* Maximum allowed back-reference distance, expressed as power of 2. |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 348 | * This will set a memory budget for streaming decompression, |
| 349 | * with larger values requiring more memory |
| 350 | * and typically compressing more. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 351 | * Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX. |
| 352 | * Special: value 0 means "use default windowLog". |
Yann Collet | 96d8874 | 2018-12-06 10:14:34 -0800 | [diff] [blame] | 353 | * Note: Using a windowLog greater than ZSTD_WINDOWLOG_LIMIT_DEFAULT |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 354 | * requires explicitly allowing such size at streaming decompression stage. */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 355 | ZSTD_c_hashLog=102, /* Size of the initial probe table, as a power of 2. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 356 | * Resulting memory usage is (1 << (hashLog+2)). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 357 | * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX. |
| 358 | * Larger tables improve compression ratio of strategies <= dFast, |
| 359 | * and improve speed of strategies > dFast. |
| 360 | * Special: value 0 means "use default hashLog". */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 361 | ZSTD_c_chainLog=103, /* Size of the multi-probe search table, as a power of 2. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 362 | * Resulting memory usage is (1 << (chainLog+2)). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 363 | * Must be clamped between ZSTD_CHAINLOG_MIN and ZSTD_CHAINLOG_MAX. |
| 364 | * Larger tables result in better and slower compression. |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 365 | * This parameter is useless for "fast" strategy. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 366 | * It's still useful when using "dfast" strategy, |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 367 | * in which case it defines a secondary probe table. |
| 368 | * Special: value 0 means "use default chainLog". */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 369 | ZSTD_c_searchLog=104, /* Number of search attempts, as a power of 2. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 370 | * More attempts result in better and slower compression. |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 371 | * This parameter is useless for "fast" and "dFast" strategies. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 372 | * Special: value 0 means "use default searchLog". */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 373 | ZSTD_c_minMatch=105, /* Minimum size of searched matches. |
Yann Collet | e874dac | 2018-11-20 14:56:07 -0800 | [diff] [blame] | 374 | * Note that Zstandard can still find matches of smaller size, |
| 375 | * it just tweaks its search algorithm to look for this size and larger. |
| 376 | * Larger values increase compression and decompression speed, but decrease ratio. |
| 377 | * Must be clamped between ZSTD_MINMATCH_MIN and ZSTD_MINMATCH_MAX. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 378 | * Note that currently, for all strategies < btopt, effective minimum is 4. |
| 379 | * , for all strategies > fast, effective maximum is 6. |
| 380 | * Special: value 0 means "use default minMatchLength". */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 381 | ZSTD_c_targetLength=106, /* Impact of this field depends on strategy. |
Yann Collet | 34aa401 | 2018-12-06 17:22:19 -0800 | [diff] [blame] | 382 | * For strategies btopt, btultra & btultra2: |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 383 | * Length of Match considered "good enough" to stop search. |
| 384 | * Larger values make compression stronger, and slower. |
| 385 | * For strategy fast: |
| 386 | * Distance between match sampling. |
| 387 | * Larger values make compression faster, and weaker. |
| 388 | * Special: value 0 means "use default targetLength". */ |
Yann Collet | be9e561 | 2018-12-06 15:00:52 -0800 | [diff] [blame] | 389 | ZSTD_c_strategy=107, /* See ZSTD_strategy enum definition. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 390 | * The higher the value of selected strategy, the more complex it is, |
| 391 | * resulting in stronger and slower compression. |
| 392 | * Special: value 0 means "use default strategy". */ |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 393 | |
| 394 | ZSTD_c_targetCBlockSize=130, /* v1.5.6+ |
| 395 | * Attempts to fit compressed block size into approximatively targetCBlockSize. |
| 396 | * Bound by ZSTD_TARGETCBLOCKSIZE_MIN and ZSTD_TARGETCBLOCKSIZE_MAX. |
| 397 | * Note that it's not a guarantee, just a convergence target (default:0). |
| 398 | * No target when targetCBlockSize == 0. |
| 399 | * This is helpful in low bandwidth streaming environments to improve end-to-end latency, |
| 400 | * when a client can make use of partial documents (a prominent example being Chrome). |
| 401 | * Note: this parameter is stable since v1.5.6. |
| 402 | * It was present as an experimental parameter in earlier versions, |
Yann Collet | c5da438 | 2024-03-18 12:33:22 -0700 | [diff] [blame] | 403 | * but it's not recommended using it with earlier library versions |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 404 | * due to massive performance regressions. |
| 405 | */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 406 | /* LDM mode parameters */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 407 | ZSTD_c_enableLongDistanceMatching=160, /* Enable long distance matching. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 408 | * This parameter is designed to improve compression ratio |
| 409 | * for large inputs, by finding large matches at long distance. |
| 410 | * It increases memory usage and window size. |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 411 | * Note: enabling this parameter increases default ZSTD_c_windowLog to 128 MB |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 412 | * except when expressly set to a different value. |
senhuang42 | aad436d | 2020-10-19 11:02:29 -0400 | [diff] [blame] | 413 | * Note: will be enabled by default if ZSTD_c_windowLog >= 128 MB and |
| 414 | * compression strategy >= ZSTD_btopt (== compression level 16+) */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 415 | ZSTD_c_ldmHashLog=161, /* Size of the table for long distance matching, as a power of 2. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 416 | * Larger values increase memory usage and compression ratio, |
| 417 | * but decrease compression speed. |
| 418 | * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX |
| 419 | * default: windowlog - 7. |
| 420 | * Special: value 0 means "automatically determine hashlog". */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 421 | ZSTD_c_ldmMinMatch=162, /* Minimum match size for long distance matcher. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 422 | * Larger/too small values usually decrease compression ratio. |
| 423 | * Must be clamped between ZSTD_LDM_MINMATCH_MIN and ZSTD_LDM_MINMATCH_MAX. |
| 424 | * Special: value 0 means "use default value" (default: 64). */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 425 | ZSTD_c_ldmBucketSizeLog=163, /* Log size of each bucket in the LDM hash table for collision resolution. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 426 | * Larger values improve collision resolution but decrease compression speed. |
Yann Collet | 5c6d4b1 | 2018-11-20 16:06:00 -0800 | [diff] [blame] | 427 | * The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 428 | * Special: value 0 means "use default value" (default: 3). */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 429 | ZSTD_c_ldmHashRateLog=164, /* Frequency of inserting/looking up entries into the LDM hash table. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 430 | * Must be clamped between 0 and (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN). |
| 431 | * Default is MAX(0, (windowLog - ldmHashLog)), optimizing hash table usage. |
| 432 | * Larger values improve compression speed. |
| 433 | * Deviating far from default value will likely result in a compression ratio decrease. |
Yann Collet | 41c7d0b | 2018-11-21 14:36:57 -0800 | [diff] [blame] | 434 | * Special: value 0 means "automatically determine hashRateLog". */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 435 | |
| 436 | /* frame parameters */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 437 | ZSTD_c_contentSizeFlag=200, /* Content size will be written into frame header _whenever known_ (default:1) |
Yann Collet | 52b94f9 | 2018-12-11 12:08:21 -0800 | [diff] [blame] | 438 | * Content size must be known at the beginning of compression. |
| 439 | * This is automatically the case when using ZSTD_compress2(), |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 440 | * For streaming scenarios, content size must be provided with ZSTD_CCtx_setPledgedSrcSize() */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 441 | ZSTD_c_checksumFlag=201, /* A 32-bits checksum of content is written at end of frame (default:0) */ |
| 442 | ZSTD_c_dictIDFlag=202, /* When applicable, dictionary's ID is written into frame header (default:1) */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 443 | |
| 444 | /* multi-threading parameters */ |
Yann Collet | 16b353b | 2020-07-11 02:16:52 -0700 | [diff] [blame] | 445 | /* These parameters are only active if multi-threading is enabled (compiled with build macro ZSTD_MULTITHREAD). |
| 446 | * Otherwise, trying to set any other value than default (0) will be a no-op and return an error. |
| 447 | * In a situation where it's unknown if the linked library supports multi-threading or not, |
| 448 | * setting ZSTD_c_nbWorkers to any value >= 1 and consulting the return value provides a quick way to check this property. |
| 449 | */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 450 | ZSTD_c_nbWorkers=400, /* Select how many threads will be spawned to compress in parallel. |
Yann Collet | 16b353b | 2020-07-11 02:16:52 -0700 | [diff] [blame] | 451 | * When nbWorkers >= 1, triggers asynchronous mode when invoking ZSTD_compressStream*() : |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 452 | * ZSTD_compressStream*() consumes input and flush output if possible, but immediately gives back control to caller, |
Yann Collet | 16b353b | 2020-07-11 02:16:52 -0700 | [diff] [blame] | 453 | * while compression is performed in parallel, within worker thread(s). |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 454 | * (note : a strong exception to this rule is when first invocation of ZSTD_compressStream2() sets ZSTD_e_end : |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 455 | * in which case, ZSTD_compressStream2() delegates to ZSTD_compress2(), which is always a blocking call). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 456 | * More workers improve speed, but also increase memory usage. |
Yann Collet | 16b353b | 2020-07-11 02:16:52 -0700 | [diff] [blame] | 457 | * Default value is `0`, aka "single-threaded mode" : no worker is spawned, |
| 458 | * compression is performed inside Caller's thread, and all invocations are blocking */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 459 | ZSTD_c_jobSize=401, /* Size of a compression job. This value is enforced only when nbWorkers >= 1. |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 460 | * Each compression job is completed in parallel, so this value can indirectly impact the nb of active threads. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 461 | * 0 means default, which is dynamically determined based on compression parameters. |
Yann Collet | 455fd1a | 2021-05-05 09:03:11 -0700 | [diff] [blame] | 462 | * Job size must be a minimum of overlap size, or ZSTDMT_JOBSIZE_MIN (= 512 KB), whichever is largest. |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 463 | * The minimum size is automatically and transparently enforced. */ |
Yann Collet | 9b784de | 2018-12-11 16:55:33 -0800 | [diff] [blame] | 464 | ZSTD_c_overlapLog=402, /* Control the overlap size, as a fraction of window size. |
| 465 | * The overlap size is an amount of data reloaded from previous job at the beginning of a new job. |
| 466 | * It helps preserve compression ratio, while each job is compressed in parallel. |
Yann Collet | 5c6d4b1 | 2018-11-20 16:06:00 -0800 | [diff] [blame] | 467 | * This value is enforced only when nbWorkers >= 1. |
| 468 | * Larger values increase compression ratio, but decrease speed. |
Yann Collet | 1993f5d | 2018-12-12 21:09:14 -0800 | [diff] [blame] | 469 | * Possible values range from 0 to 9 : |
| 470 | * - 0 means "default" : value will be determined by the library, depending on strategy |
Yann Collet | 9b784de | 2018-12-11 16:55:33 -0800 | [diff] [blame] | 471 | * - 1 means "no overlap" |
| 472 | * - 9 means "full overlap", using a full window size. |
| 473 | * Each intermediate rank increases/decreases load size by a factor 2 : |
Yann Collet | 1993f5d | 2018-12-12 21:09:14 -0800 | [diff] [blame] | 474 | * 9: full window; 8: w/2; 7: w/4; 6: w/8; 5:w/16; 4: w/32; 3:w/64; 2:w/128; 1:no overlap; 0:default |
Yann Collet | 9b784de | 2018-12-11 16:55:33 -0800 | [diff] [blame] | 475 | * default value varies between 6 and 9, depending on strategy */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 476 | |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 477 | /* note : additional experimental parameters are also available |
| 478 | * within the experimental section of the API. |
| 479 | * At the time of this writing, they include : |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 480 | * ZSTD_c_rsyncable |
| 481 | * ZSTD_c_format |
| 482 | * ZSTD_c_forceMaxWindow |
| 483 | * ZSTD_c_forceAttachDict |
sen | c730b8c | 2021-05-12 16:09:48 -0400 | [diff] [blame] | 484 | * ZSTD_c_literalCompressionMode |
Nick Magerko | dffbac5 | 2019-08-19 08:52:08 -0700 | [diff] [blame] | 485 | * ZSTD_c_srcSizeHint |
W. Felix Handte | c204110 | 2020-08-11 18:03:38 -0400 | [diff] [blame] | 486 | * ZSTD_c_enableDedicatedDictSearch |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 487 | * ZSTD_c_stableInBuffer |
| 488 | * ZSTD_c_stableOutBuffer |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 489 | * ZSTD_c_blockDelimiters |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 490 | * ZSTD_c_validateSequences |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 491 | * ZSTD_c_useBlockSplitter |
Nick Terrell | 4694423 | 2020-11-02 17:52:29 -0800 | [diff] [blame] | 492 | * ZSTD_c_useRowMatchFinder |
Elliot Gorokhovsky | 2a12811 | 2022-06-21 11:59:27 -0400 | [diff] [blame] | 493 | * ZSTD_c_prefetchCDictTables |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 494 | * ZSTD_c_enableSeqProducerFallback |
Danielle Rozenblit | 908e812 | 2023-01-04 13:01:54 -0800 | [diff] [blame] | 495 | * ZSTD_c_maxBlockSize |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 496 | * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them. |
Yann Collet | 5a1e01e | 2018-12-10 17:36:20 -0800 | [diff] [blame] | 497 | * note : never ever use experimentalParam? names directly; |
| 498 | * also, the enums values themselves are unstable and can still change. |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 499 | */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 500 | ZSTD_c_experimentalParam1=500, |
| 501 | ZSTD_c_experimentalParam2=10, |
| 502 | ZSTD_c_experimentalParam3=1000, |
Nick Terrell | f951311 | 2019-02-13 14:59:22 -0800 | [diff] [blame] | 503 | ZSTD_c_experimentalParam4=1001, |
sen | c730b8c | 2021-05-12 16:09:48 -0400 | [diff] [blame] | 504 | ZSTD_c_experimentalParam5=1002, |
W. Felix Handte | 3613448 | 2024-03-13 16:56:54 -0400 | [diff] [blame] | 505 | /* was ZSTD_c_experimentalParam6=1003; is now ZSTD_c_targetCBlockSize */ |
W. Felix Handte | c204110 | 2020-08-11 18:03:38 -0400 | [diff] [blame] | 506 | ZSTD_c_experimentalParam7=1004, |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 507 | ZSTD_c_experimentalParam8=1005, |
| 508 | ZSTD_c_experimentalParam9=1006, |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 509 | ZSTD_c_experimentalParam10=1007, |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 510 | ZSTD_c_experimentalParam11=1008, |
senhuang42 | c56d6e4 | 2020-12-07 17:11:23 -0500 | [diff] [blame] | 511 | ZSTD_c_experimentalParam12=1009, |
Nick Terrell | 4694423 | 2020-11-02 17:52:29 -0800 | [diff] [blame] | 512 | ZSTD_c_experimentalParam13=1010, |
Nick Terrell | 172b4b6 | 2021-05-05 12:18:47 -0700 | [diff] [blame] | 513 | ZSTD_c_experimentalParam14=1011, |
Elliot Gorokhovsky | 2a12811 | 2022-06-21 11:59:27 -0400 | [diff] [blame] | 514 | ZSTD_c_experimentalParam15=1012, |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 515 | ZSTD_c_experimentalParam16=1013, |
Danielle Rozenblit | 908e812 | 2023-01-04 13:01:54 -0800 | [diff] [blame] | 516 | ZSTD_c_experimentalParam17=1014, |
Elliot Gorokhovsky | 7f8189c | 2023-01-31 13:51:32 -0800 | [diff] [blame] | 517 | ZSTD_c_experimentalParam18=1015, |
| 518 | ZSTD_c_experimentalParam19=1016 |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 519 | } ZSTD_cParameter; |
| 520 | |
Yann Collet | e874dac | 2018-11-20 14:56:07 -0800 | [diff] [blame] | 521 | typedef struct { |
| 522 | size_t error; |
| 523 | int lowerBound; |
| 524 | int upperBound; |
| 525 | } ZSTD_bounds; |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 526 | |
Yann Collet | e874dac | 2018-11-20 14:56:07 -0800 | [diff] [blame] | 527 | /*! ZSTD_cParam_getBounds() : |
| 528 | * All parameters must belong to an interval with lower and upper bounds, |
| 529 | * otherwise they will either trigger an error or be automatically clamped. |
| 530 | * @return : a structure, ZSTD_bounds, which contains |
| 531 | * - an error status field, which must be tested using ZSTD_isError() |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 532 | * - lower and upper bounds, both inclusive |
Yann Collet | e874dac | 2018-11-20 14:56:07 -0800 | [diff] [blame] | 533 | */ |
| 534 | ZSTDLIB_API ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter cParam); |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 535 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 536 | /*! ZSTD_CCtx_setParameter() : |
| 537 | * Set one compression parameter, selected by enum ZSTD_cParameter. |
Yann Collet | 3b838ab | 2018-11-20 11:53:01 -0800 | [diff] [blame] | 538 | * All parameters have valid bounds. Bounds can be queried using ZSTD_cParam_getBounds(). |
| 539 | * Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 540 | * Setting a parameter is generally only possible during frame initialization (before starting compression). |
| 541 | * Exception : when using multi-threading mode (nbWorkers >= 1), |
| 542 | * the following parameters can be updated _during_ compression (within same frame): |
| 543 | * => compressionLevel, hashLog, chainLog, searchLog, minMatch, targetLength and strategy. |
| 544 | * new parameters will be active for next job only (after a flush()). |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 545 | * @return : an error code (which can be tested using ZSTD_isError()). |
| 546 | */ |
Yann Collet | 3b838ab | 2018-11-20 11:53:01 -0800 | [diff] [blame] | 547 | ZSTDLIB_API size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 548 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 549 | /*! ZSTD_CCtx_setPledgedSrcSize() : |
| 550 | * Total input data size to be compressed as a single frame. |
Yann Collet | 52b94f9 | 2018-12-11 12:08:21 -0800 | [diff] [blame] | 551 | * Value will be written in frame header, unless if explicitly forbidden using ZSTD_c_contentSizeFlag. |
| 552 | * This value will also be controlled at end of frame, and trigger an error if not respected. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 553 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 554 | * Note 1 : pledgedSrcSize==0 actually means zero, aka an empty frame. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 555 | * In order to mean "unknown content size", pass constant ZSTD_CONTENTSIZE_UNKNOWN. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 556 | * ZSTD_CONTENTSIZE_UNKNOWN is default value for any new frame. |
| 557 | * Note 2 : pledgedSrcSize is only valid once, for the next frame. |
Yann Collet | 52b94f9 | 2018-12-11 12:08:21 -0800 | [diff] [blame] | 558 | * It's discarded at the end of the frame, and replaced by ZSTD_CONTENTSIZE_UNKNOWN. |
| 559 | * Note 3 : Whenever all input data is provided and consumed in a single round, |
| 560 | * for example with ZSTD_compress2(), |
| 561 | * or invoking immediately ZSTD_compressStream2(,,,ZSTD_e_end), |
Josh Soref | a880ca2 | 2019-04-12 14:18:11 -0400 | [diff] [blame] | 562 | * this value is automatically overridden by srcSize instead. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 563 | */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 564 | ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize); |
| 565 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 566 | typedef enum { |
Yann Collet | 5c68639 | 2018-11-15 16:12:39 -0800 | [diff] [blame] | 567 | ZSTD_reset_session_only = 1, |
| 568 | ZSTD_reset_parameters = 2, |
| 569 | ZSTD_reset_session_and_parameters = 3 |
Yann Collet | 114bd43 | 2018-11-20 12:00:20 -0800 | [diff] [blame] | 570 | } ZSTD_ResetDirective; |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 571 | |
| 572 | /*! ZSTD_CCtx_reset() : |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 573 | * There are 2 different things that can be reset, independently or jointly : |
| 574 | * - The session : will stop compressing current frame, and make CCtx ready to start a new one. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 575 | * Useful after an error, or to interrupt any ongoing compression. |
| 576 | * Any internal data not yet flushed is cancelled. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 577 | * Compression parameters and dictionary remain unchanged. |
| 578 | * They will be used to compress next frame. |
| 579 | * Resetting session never fails. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 580 | * - The parameters : changes all parameters back to "default". |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 581 | * This also removes any reference to any dictionary or external sequence producer. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 582 | * Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing) |
| 583 | * otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError()) |
| 584 | * - Both : similar to resetting the session, followed by resetting parameters. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 585 | */ |
Yann Collet | 114bd43 | 2018-11-20 12:00:20 -0800 | [diff] [blame] | 586 | ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 587 | |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 588 | /*! ZSTD_compress2() : |
| 589 | * Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API. |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 590 | * (note that this entry point doesn't even expose a compression level parameter). |
Yann Collet | c226a7b | 2018-12-10 17:33:49 -0800 | [diff] [blame] | 591 | * ZSTD_compress2() always starts a new frame. |
| 592 | * Should cctx hold data from a previously unfinished frame, everything about it is forgotten. |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 593 | * - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*() |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 594 | * - The function is always blocking, returns when compression is completed. |
Nick Terrell | c40c737 | 2023-03-07 15:42:51 -0800 | [diff] [blame] | 595 | * NOTE: Providing `dstCapacity >= ZSTD_compressBound(srcSize)` guarantees that zstd will have |
| 596 | * enough space to successfully compress the data, though it is possible it fails for other reasons. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 597 | * @return : compressed size written into `dst` (<= `dstCapacity), |
| 598 | * or an error code if it fails (which can be tested using ZSTD_isError()). |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 599 | */ |
| 600 | ZSTDLIB_API size_t ZSTD_compress2( ZSTD_CCtx* cctx, |
| 601 | void* dst, size_t dstCapacity, |
| 602 | const void* src, size_t srcSize); |
| 603 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 604 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 605 | /*********************************************** |
| 606 | * Advanced decompression API (Requires v1.4.0+) |
| 607 | ************************************************/ |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 608 | |
Yann Collet | 34e146f | 2018-12-04 10:28:36 -0800 | [diff] [blame] | 609 | /* The advanced API pushes parameters one by one into an existing DCtx context. |
Yann Collet | 96d8874 | 2018-12-06 10:14:34 -0800 | [diff] [blame] | 610 | * Parameters are sticky, and remain valid for all following frames |
Yann Collet | 3e042d5 | 2018-12-04 17:30:58 -0800 | [diff] [blame] | 611 | * using the same DCtx context. |
Yann Collet | 34e146f | 2018-12-04 10:28:36 -0800 | [diff] [blame] | 612 | * It's possible to reset parameters to default values using ZSTD_DCtx_reset(). |
Yann Collet | 3e042d5 | 2018-12-04 17:30:58 -0800 | [diff] [blame] | 613 | * Note : This API is compatible with existing ZSTD_decompressDCtx() and ZSTD_decompressStream(). |
| 614 | * Therefore, no new decompression function is necessary. |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 615 | */ |
| 616 | |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 617 | typedef enum { |
| 618 | |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 619 | ZSTD_d_windowLogMax=100, /* Select a size limit (in power of 2) beyond which |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 620 | * the streaming API will refuse to allocate memory buffer |
| 621 | * in order to protect the host from unreasonable memory requirements. |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 622 | * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode. |
Nick Terrell | 00679da | 2019-04-02 19:20:52 -0700 | [diff] [blame] | 623 | * By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT). |
| 624 | * Special: value 0 means "use default maximum windowLog". */ |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 625 | |
| 626 | /* note : additional experimental parameters are also available |
| 627 | * within the experimental section of the API. |
| 628 | * At the time of this writing, they include : |
Nick Terrell | b104f8e | 2020-04-27 12:12:28 -0700 | [diff] [blame] | 629 | * ZSTD_d_format |
Nick Terrell | a4ff217 | 2020-04-27 17:42:03 -0700 | [diff] [blame] | 630 | * ZSTD_d_stableOutBuffer |
senhuang42 | 2f39124 | 2020-08-21 16:23:39 -0400 | [diff] [blame] | 631 | * ZSTD_d_forceIgnoreChecksum |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 632 | * ZSTD_d_refMultipleDDicts |
Nick Terrell | 8957fef | 2023-01-13 16:34:52 -0800 | [diff] [blame] | 633 | * ZSTD_d_disableHuffmanAssembly |
Nick Terrell | 61efb2a | 2023-04-14 17:06:24 -0700 | [diff] [blame] | 634 | * ZSTD_d_maxBlockSize |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 635 | * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them. |
Yann Collet | 34e146f | 2018-12-04 10:28:36 -0800 | [diff] [blame] | 636 | * note : never ever use experimentalParam? names directly |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 637 | */ |
Nick Terrell | a4ff217 | 2020-04-27 17:42:03 -0700 | [diff] [blame] | 638 | ZSTD_d_experimentalParam1=1000, |
senhuang42 | b5cddda | 2020-08-21 15:59:03 -0400 | [diff] [blame] | 639 | ZSTD_d_experimentalParam2=1001, |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 640 | ZSTD_d_experimentalParam3=1002, |
Nick Terrell | 8957fef | 2023-01-13 16:34:52 -0800 | [diff] [blame] | 641 | ZSTD_d_experimentalParam4=1003, |
Nick Terrell | 61efb2a | 2023-04-14 17:06:24 -0700 | [diff] [blame] | 642 | ZSTD_d_experimentalParam5=1004, |
| 643 | ZSTD_d_experimentalParam6=1005 |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 644 | |
| 645 | } ZSTD_dParameter; |
| 646 | |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 647 | /*! ZSTD_dParam_getBounds() : |
| 648 | * All parameters must belong to an interval with lower and upper bounds, |
| 649 | * otherwise they will either trigger an error or be automatically clamped. |
| 650 | * @return : a structure, ZSTD_bounds, which contains |
| 651 | * - an error status field, which must be tested using ZSTD_isError() |
| 652 | * - both lower and upper bounds, inclusive |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 653 | */ |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 654 | ZSTDLIB_API ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam); |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 655 | |
| 656 | /*! ZSTD_DCtx_setParameter() : |
| 657 | * Set one compression parameter, selected by enum ZSTD_dParameter. |
| 658 | * All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds(). |
| 659 | * Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter). |
| 660 | * Setting a parameter is only possible during frame initialization (before starting decompression). |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 661 | * @return : 0, or an error code (which can be tested using ZSTD_isError()). |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 662 | */ |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 663 | ZSTDLIB_API size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value); |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 664 | |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 665 | /*! ZSTD_DCtx_reset() : |
| 666 | * Return a DCtx to clean state. |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 667 | * Session and parameters can be reset jointly or separately. |
Yann Collet | 5c68639 | 2018-11-15 16:12:39 -0800 | [diff] [blame] | 668 | * Parameters can only be reset when no active frame is being decompressed. |
| 669 | * @return : 0, or an error code, which can be tested with ZSTD_isError() |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 670 | */ |
Yann Collet | 114bd43 | 2018-11-20 12:00:20 -0800 | [diff] [blame] | 671 | ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset); |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 672 | |
| 673 | |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 674 | /**************************** |
| 675 | * Streaming |
| 676 | ****************************/ |
| 677 | |
| 678 | typedef struct ZSTD_inBuffer_s { |
| 679 | const void* src; /**< start of input buffer */ |
| 680 | size_t size; /**< size of input buffer */ |
| 681 | size_t pos; /**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */ |
| 682 | } ZSTD_inBuffer; |
| 683 | |
| 684 | typedef struct ZSTD_outBuffer_s { |
| 685 | void* dst; /**< start of output buffer */ |
| 686 | size_t size; /**< size of output buffer */ |
| 687 | size_t pos; /**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */ |
| 688 | } ZSTD_outBuffer; |
| 689 | |
| 690 | |
| 691 | |
| 692 | /*-*********************************************************************** |
| 693 | * Streaming compression - HowTo |
| 694 | * |
| 695 | * A ZSTD_CStream object is required to track streaming operation. |
| 696 | * Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources. |
| 697 | * ZSTD_CStream objects can be reused multiple times on consecutive compression operations. |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 698 | * It is recommended to reuse ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 699 | * |
| 700 | * For parallel execution, use one separate ZSTD_CStream per thread. |
| 701 | * |
| 702 | * note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing. |
| 703 | * |
| 704 | * Parameters are sticky : when starting a new compression on the same context, |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 705 | * it will reuse the same sticky parameters as previous compression session. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 706 | * When in doubt, it's recommended to fully initialize the context before usage. |
Nick Terrell | 7231ea7 | 2019-04-03 19:21:05 -0700 | [diff] [blame] | 707 | * Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(), |
| 708 | * ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to |
| 709 | * set more specific parameters, the pledged source size, or load a dictionary. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 710 | * |
Nick Terrell | 7231ea7 | 2019-04-03 19:21:05 -0700 | [diff] [blame] | 711 | * Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to |
| 712 | * consume input stream. The function will automatically update both `pos` |
| 713 | * fields within `input` and `output`. |
| 714 | * Note that the function may not consume the entire input, for example, because |
| 715 | * the output buffer is already full, in which case `input.pos < input.size`. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 716 | * The caller must check if input has been entirely consumed. |
| 717 | * If not, the caller must make some room to receive more compressed data, |
| 718 | * and then present again remaining input data. |
Nick Terrell | 48a6427 | 2019-04-09 16:24:17 -0700 | [diff] [blame] | 719 | * note: ZSTD_e_continue is guaranteed to make some forward progress when called, |
| 720 | * but doesn't guarantee maximal forward progress. This is especially relevant |
| 721 | * when compressing with multiple threads. The call won't block if it can |
| 722 | * consume some input, but if it can't it will wait for some, but not all, |
| 723 | * output to be flushed. |
Nick Terrell | ce388fe | 2019-04-05 17:44:07 -0700 | [diff] [blame] | 724 | * @return : provides a minimum amount of data remaining to be flushed from internal buffers |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 725 | * or an error code, which can be tested using ZSTD_isError(). |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 726 | * |
| 727 | * At any moment, it's possible to flush whatever data might remain stuck within internal buffer, |
Nick Terrell | 7231ea7 | 2019-04-03 19:21:05 -0700 | [diff] [blame] | 728 | * using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated. |
| 729 | * Note that, if `output->size` is too small, a single invocation with ZSTD_e_flush might not be enough (return code > 0). |
| 730 | * In which case, make some room to receive more compressed data, and call again ZSTD_compressStream2() with ZSTD_e_flush. |
| 731 | * You must continue calling ZSTD_compressStream2() with ZSTD_e_flush until it returns 0, at which point you can change the |
| 732 | * operation. |
Nick Terrell | 48a6427 | 2019-04-09 16:24:17 -0700 | [diff] [blame] | 733 | * note: ZSTD_e_flush will flush as much output as possible, meaning when compressing with multiple threads, it will |
| 734 | * block until the flush is complete or the output buffer is full. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 735 | * @return : 0 if internal buffers are entirely flushed, |
| 736 | * >0 if some data still present within internal buffer (the value is minimal estimation of remaining size), |
| 737 | * or an error code, which can be tested using ZSTD_isError(). |
| 738 | * |
Nick Terrell | 7231ea7 | 2019-04-03 19:21:05 -0700 | [diff] [blame] | 739 | * Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 740 | * It will perform a flush and write frame epilogue. |
| 741 | * The epilogue is required for decoders to consider a frame completed. |
Nick Terrell | 7231ea7 | 2019-04-03 19:21:05 -0700 | [diff] [blame] | 742 | * flush operation is the same, and follows same rules as calling ZSTD_compressStream2() with ZSTD_e_flush. |
| 743 | * You must continue calling ZSTD_compressStream2() with ZSTD_e_end until it returns 0, at which point you are free to |
| 744 | * start a new frame. |
Nick Terrell | 48a6427 | 2019-04-09 16:24:17 -0700 | [diff] [blame] | 745 | * note: ZSTD_e_end will flush as much output as possible, meaning when compressing with multiple threads, it will |
| 746 | * block until the flush is complete or the output buffer is full. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 747 | * @return : 0 if frame fully completed and fully flushed, |
| 748 | * >0 if some data still present within internal buffer (the value is minimal estimation of remaining size), |
| 749 | * or an error code, which can be tested using ZSTD_isError(). |
| 750 | * |
| 751 | * *******************************************************************/ |
| 752 | |
| 753 | typedef ZSTD_CCtx ZSTD_CStream; /**< CCtx and CStream are now effectively same object (>= v1.3.0) */ |
| 754 | /* Continue to distinguish them for compatibility with older versions <= v1.2.0 */ |
| 755 | /*===== ZSTD_CStream management functions =====*/ |
| 756 | ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void); |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 757 | ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs); /* accept NULL pointer */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 758 | |
| 759 | /*===== Streaming compression functions =====*/ |
| 760 | typedef enum { |
| 761 | ZSTD_e_continue=0, /* collect more data, encoder decides when to output compressed result, for optimal compression ratio */ |
| 762 | ZSTD_e_flush=1, /* flush any data provided so far, |
| 763 | * it creates (at least) one new block, that can be decoded immediately on reception; |
Nick Terrell | 48a6427 | 2019-04-09 16:24:17 -0700 | [diff] [blame] | 764 | * frame will continue: any future data can still reference previously compressed data, improving compression. |
| 765 | * note : multithreaded compression will block to flush as much output as possible. */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 766 | ZSTD_e_end=2 /* flush any remaining data _and_ close current frame. |
| 767 | * note that frame is only closed after compressed data is fully flushed (return value == 0). |
| 768 | * After that point, any additional data starts a new frame. |
Nick Terrell | 48a6427 | 2019-04-09 16:24:17 -0700 | [diff] [blame] | 769 | * note : each frame is independent (does not reference any content from previous frame). |
| 770 | : note : multithreaded compression will block to flush as much output as possible. */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 771 | } ZSTD_EndDirective; |
| 772 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 773 | /*! ZSTD_compressStream2() : Requires v1.4.0+ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 774 | * Behaves about the same as ZSTD_compressStream, with additional control on end directive. |
| 775 | * - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*() |
| 776 | * - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode) |
Josh Soref | a880ca2 | 2019-04-12 14:18:11 -0400 | [diff] [blame] | 777 | * - output->pos must be <= dstCapacity, input->pos must be <= srcSize |
| 778 | * - output->pos and input->pos will be updated. They are guaranteed to remain below their respective limit. |
Yann Collet | c91a085 | 2020-09-14 10:56:08 -0700 | [diff] [blame] | 779 | * - endOp must be a valid directive |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 780 | * - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller. |
Yann Collet | c91a085 | 2020-09-14 10:56:08 -0700 | [diff] [blame] | 781 | * - When nbWorkers>=1, function is non-blocking : it copies a portion of input, distributes jobs to internal worker threads, flush to output whatever is available, |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 782 | * and then immediately returns, just indicating that there is some data remaining to be flushed. |
| 783 | * The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte. |
| 784 | * - Exception : if the first call requests a ZSTD_e_end directive and provides enough dstCapacity, the function delegates to ZSTD_compress2() which is always blocking. |
| 785 | * - @return provides a minimum amount of data remaining to be flushed from internal buffers |
| 786 | * or an error code, which can be tested using ZSTD_isError(). |
| 787 | * if @return != 0, flush is not fully completed, there is still some data left within internal buffers. |
| 788 | * This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers. |
| 789 | * For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed. |
| 790 | * - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0), |
| 791 | * only ZSTD_e_end or ZSTD_e_flush operations are allowed. |
| 792 | * Before starting a new compression job, or changing compression parameters, |
| 793 | * it is required to fully flush internal buffers. |
Yann Collet | 902c7ec | 2024-03-18 12:30:35 -0700 | [diff] [blame] | 794 | * - note: if an operation ends with an error, it may leave @cctx in an undefined state. |
| 795 | * Therefore, it's UB to invoke ZSTD_compressStream2() of ZSTD_compressStream() on such a state. |
| 796 | * In order to be re-employed after an error, a state must be reset, |
| 797 | * which can be done explicitly (ZSTD_CCtx_reset()), |
| 798 | * or is sometimes implied by methods starting a new compression job (ZSTD_initCStream(), ZSTD_compressCCtx()) |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 799 | */ |
| 800 | ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx, |
| 801 | ZSTD_outBuffer* output, |
| 802 | ZSTD_inBuffer* input, |
| 803 | ZSTD_EndDirective endOp); |
| 804 | |
Yann Collet | b5c98fb | 2019-06-04 10:26:16 -0700 | [diff] [blame] | 805 | |
| 806 | /* These buffer sizes are softly recommended. |
| 807 | * They are not required : ZSTD_compressStream*() happily accepts any buffer size, for both input and output. |
| 808 | * Respecting the recommended size just makes it a bit easier for ZSTD_compressStream*(), |
| 809 | * reducing the amount of memory shuffling and buffering, resulting in minor performance savings. |
| 810 | * |
| 811 | * However, note that these recommendations are from the perspective of a C caller program. |
| 812 | * If the streaming interface is invoked from some other language, |
| 813 | * especially managed ones such as Java or Go, through a foreign function interface such as jni or cgo, |
| 814 | * a major performance rule is to reduce crossing such interface to an absolute minimum. |
| 815 | * It's not rare that performance ends being spent more into the interface, rather than compression itself. |
| 816 | * In which cases, prefer using large buffers, as large as practical, |
| 817 | * for both input and output, to reduce the nb of roundtrips. |
| 818 | */ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 819 | ZSTDLIB_API size_t ZSTD_CStreamInSize(void); /**< recommended size for input buffer */ |
Yann Collet | b5c98fb | 2019-06-04 10:26:16 -0700 | [diff] [blame] | 820 | ZSTDLIB_API size_t ZSTD_CStreamOutSize(void); /**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block. */ |
| 821 | |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 822 | |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 823 | /* ***************************************************************************** |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 824 | * This following is a legacy streaming API, available since v1.0+ . |
Yann Collet | b5c98fb | 2019-06-04 10:26:16 -0700 | [diff] [blame] | 825 | * It can be replaced by ZSTD_CCtx_reset() and ZSTD_compressStream2(). |
| 826 | * It is redundant, but remains fully supported. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 827 | ******************************************************************************/ |
| 828 | |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 829 | /*! |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 830 | * Equivalent to: |
| 831 | * |
| 832 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
| 833 | * ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any) |
| 834 | * ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel); |
Elliot Gorokhovsky | 3bfd3be | 2023-01-23 13:28:36 -0500 | [diff] [blame] | 835 | * |
| 836 | * Note that ZSTD_initCStream() clears any previously set dictionary. Use the new API |
| 837 | * to compress with a dictionary. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 838 | */ |
| 839 | ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel); |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 840 | /*! |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 841 | * Alternative for ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue). |
| 842 | * NOTE: The return value is different. ZSTD_compressStream() returns a hint for |
| 843 | * the next read size (if non-zero and not an error). ZSTD_compressStream2() |
Yann Collet | b5c98fb | 2019-06-04 10:26:16 -0700 | [diff] [blame] | 844 | * returns the minimum nb of bytes left to flush (if non-zero and not an error). |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 845 | */ |
| 846 | ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input); |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 847 | /*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_flush). */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 848 | ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output); |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 849 | /*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_end). */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 850 | ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output); |
| 851 | |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 852 | |
| 853 | /*-*************************************************************************** |
| 854 | * Streaming decompression - HowTo |
| 855 | * |
| 856 | * A ZSTD_DStream object is required to track streaming operations. |
| 857 | * Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources. |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 858 | * ZSTD_DStream objects can be reused multiple times. |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 859 | * |
| 860 | * Use ZSTD_initDStream() to start a new decompression operation. |
| 861 | * @return : recommended first input size |
| 862 | * Alternatively, use advanced API to set specific properties. |
| 863 | * |
| 864 | * Use ZSTD_decompressStream() repetitively to consume your input. |
| 865 | * The function will update both `pos` fields. |
| 866 | * If `input.pos < input.size`, some input has not been consumed. |
| 867 | * It's up to the caller to present again remaining data. |
| 868 | * The function tries to flush all data decoded immediately, respecting output buffer size. |
| 869 | * If `output.pos < output.size`, decoder has flushed everything it could. |
| 870 | * But if `output.pos == output.size`, there might be some data left within internal buffers., |
| 871 | * In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer. |
| 872 | * Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX. |
| 873 | * @return : 0 when a frame is completely decoded and fully flushed, |
| 874 | * or an error code, which can be tested using ZSTD_isError(), |
| 875 | * or any other value > 0, which means there is still some decoding or flushing to do to complete current frame : |
| 876 | * the return value is a suggested next input size (just a hint for better latency) |
| 877 | * that will never request more than the remaining frame size. |
| 878 | * *******************************************************************************/ |
| 879 | |
| 880 | typedef ZSTD_DCtx ZSTD_DStream; /**< DCtx and DStream are now effectively same object (>= v1.3.0) */ |
| 881 | /* For compatibility with versions <= v1.2.0, prefer differentiating them. */ |
| 882 | /*===== ZSTD_DStream management functions =====*/ |
| 883 | ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void); |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 884 | ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds); /* accept NULL pointer */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 885 | |
| 886 | /*===== Streaming decompression functions =====*/ |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 887 | |
Konstantin Gribov | 1c847e2 | 2022-08-01 23:50:54 +0300 | [diff] [blame] | 888 | /*! ZSTD_initDStream() : |
| 889 | * Initialize/reset DStream state for new decompression operation. |
| 890 | * Call before new decompression operation using same DStream. |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 891 | * |
Konstantin Gribov | 1c847e2 | 2022-08-01 23:50:54 +0300 | [diff] [blame] | 892 | * Note : This function is redundant with the advanced API and equivalent to: |
W. Felix Handte | 722149c | 2019-11-19 01:57:15 -0500 | [diff] [blame] | 893 | * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only); |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 894 | * ZSTD_DCtx_refDDict(zds, NULL); |
| 895 | */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 896 | ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds); |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 897 | |
Konstantin Gribov | 1c847e2 | 2022-08-01 23:50:54 +0300 | [diff] [blame] | 898 | /*! ZSTD_decompressStream() : |
| 899 | * Streaming decompression function. |
| 900 | * Call repetitively to consume full input updating it as necessary. |
Dimitri Papadopoulos | 0015308 | 2022-09-08 23:12:23 +0200 | [diff] [blame] | 901 | * Function will update both input and output `pos` fields exposing current state via these fields: |
Konstantin Gribov | 1c847e2 | 2022-08-01 23:50:54 +0300 | [diff] [blame] | 902 | * - `input.pos < input.size`, some input remaining and caller should provide remaining input |
| 903 | * on the next call. |
| 904 | * - `output.pos < output.size`, decoder finished and flushed all remaining buffers. |
| 905 | * - `output.pos == output.size`, potentially uncflushed data present in the internal buffers, |
| 906 | * call ZSTD_decompressStream() again to flush remaining data to output. |
| 907 | * Note : with no additional input, amount of data flushed <= ZSTD_BLOCKSIZE_MAX. |
| 908 | * |
Dimitri Papadopoulos | 0015308 | 2022-09-08 23:12:23 +0200 | [diff] [blame] | 909 | * @return : 0 when a frame is completely decoded and fully flushed, |
Konstantin Gribov | 1c847e2 | 2022-08-01 23:50:54 +0300 | [diff] [blame] | 910 | * or an error code, which can be tested using ZSTD_isError(), |
| 911 | * or any other value > 0, which means there is some decoding or flushing to do to complete current frame. |
Yann Collet | 5d82c2b | 2024-03-18 12:17:41 -0700 | [diff] [blame] | 912 | * |
Yann Collet | 902c7ec | 2024-03-18 12:30:35 -0700 | [diff] [blame] | 913 | * Note: when an operation returns with an error code, the @zds state may be left in undefined state. |
Yann Collet | 5d82c2b | 2024-03-18 12:17:41 -0700 | [diff] [blame] | 914 | * It's UB to invoke `ZSTD_decompressStream()` on such a state. |
Yann Collet | 902c7ec | 2024-03-18 12:30:35 -0700 | [diff] [blame] | 915 | * In order to re-use such a state, it must be first reset, |
Yann Collet | 5d82c2b | 2024-03-18 12:17:41 -0700 | [diff] [blame] | 916 | * which can be done explicitly (`ZSTD_DCtx_reset()`), |
Yann Collet | 902c7ec | 2024-03-18 12:30:35 -0700 | [diff] [blame] | 917 | * or is implied for operations starting some new decompression job (`ZSTD_initDStream`, `ZSTD_decompressDCtx()`, `ZSTD_decompress_usingDict()`) |
Konstantin Gribov | 1c847e2 | 2022-08-01 23:50:54 +0300 | [diff] [blame] | 918 | */ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 919 | ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input); |
| 920 | |
| 921 | ZSTDLIB_API size_t ZSTD_DStreamInSize(void); /*!< recommended size for input buffer */ |
| 922 | ZSTDLIB_API size_t ZSTD_DStreamOutSize(void); /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */ |
| 923 | |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 924 | |
| 925 | /************************** |
| 926 | * Simple dictionary API |
| 927 | ***************************/ |
| 928 | /*! ZSTD_compress_usingDict() : |
| 929 | * Compression at an explicit compression level using a Dictionary. |
| 930 | * A dictionary can be any arbitrary data segment (also called a prefix), |
Nick Terrell | 09149be | 2021-04-30 15:02:12 -0700 | [diff] [blame] | 931 | * or a buffer with specified information (see zdict.h). |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 932 | * Note : This function loads the dictionary, resulting in significant startup delay. |
| 933 | * It's intended for a dictionary used only once. |
| 934 | * Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used. */ |
| 935 | ZSTDLIB_API size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx, |
| 936 | void* dst, size_t dstCapacity, |
| 937 | const void* src, size_t srcSize, |
| 938 | const void* dict,size_t dictSize, |
| 939 | int compressionLevel); |
| 940 | |
| 941 | /*! ZSTD_decompress_usingDict() : |
| 942 | * Decompression using a known Dictionary. |
| 943 | * Dictionary must be identical to the one used during compression. |
| 944 | * Note : This function loads the dictionary, resulting in significant startup delay. |
| 945 | * It's intended for a dictionary used only once. |
| 946 | * Note : When `dict == NULL || dictSize < 8` no dictionary is used. */ |
| 947 | ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx, |
| 948 | void* dst, size_t dstCapacity, |
| 949 | const void* src, size_t srcSize, |
| 950 | const void* dict,size_t dictSize); |
| 951 | |
| 952 | |
| 953 | /*********************************** |
| 954 | * Bulk processing dictionary API |
| 955 | **********************************/ |
| 956 | typedef struct ZSTD_CDict_s ZSTD_CDict; |
| 957 | |
| 958 | /*! ZSTD_createCDict() : |
Yann Collet | 74065da | 2019-10-28 11:15:41 -0700 | [diff] [blame] | 959 | * When compressing multiple messages or blocks using the same dictionary, |
| 960 | * it's recommended to digest the dictionary only once, since it's a costly operation. |
| 961 | * ZSTD_createCDict() will create a state from digesting a dictionary. |
| 962 | * The resulting state can be used for future compression operations with very limited startup cost. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 963 | * ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only. |
Yann Collet | 74065da | 2019-10-28 11:15:41 -0700 | [diff] [blame] | 964 | * @dictBuffer can be released after ZSTD_CDict creation, because its content is copied within CDict. |
| 965 | * Note 1 : Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate @dictBuffer content. |
| 966 | * Note 2 : A ZSTD_CDict can be created from an empty @dictBuffer, |
| 967 | * in which case the only thing that it transports is the @compressionLevel. |
| 968 | * This can be useful in a pipeline featuring ZSTD_compress_usingCDict() exclusively, |
| 969 | * expecting a ZSTD_CDict parameter with any data, including those without a known dictionary. */ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 970 | ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize, |
| 971 | int compressionLevel); |
| 972 | |
| 973 | /*! ZSTD_freeCDict() : |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 974 | * Function frees memory allocated by ZSTD_createCDict(). |
| 975 | * If a NULL pointer is passed, no operation is performed. */ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 976 | ZSTDLIB_API size_t ZSTD_freeCDict(ZSTD_CDict* CDict); |
| 977 | |
| 978 | /*! ZSTD_compress_usingCDict() : |
| 979 | * Compression using a digested Dictionary. |
| 980 | * Recommended when same dictionary is used multiple times. |
| 981 | * Note : compression level is _decided at dictionary creation time_, |
| 982 | * and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */ |
| 983 | ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx, |
| 984 | void* dst, size_t dstCapacity, |
| 985 | const void* src, size_t srcSize, |
| 986 | const ZSTD_CDict* cdict); |
| 987 | |
| 988 | |
| 989 | typedef struct ZSTD_DDict_s ZSTD_DDict; |
| 990 | |
| 991 | /*! ZSTD_createDDict() : |
| 992 | * Create a digested dictionary, ready to start decompression operation without startup delay. |
| 993 | * dictBuffer can be released after DDict creation, as its content is copied inside DDict. */ |
| 994 | ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize); |
| 995 | |
| 996 | /*! ZSTD_freeDDict() : |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 997 | * Function frees memory allocated with ZSTD_createDDict() |
| 998 | * If a NULL pointer is passed, no operation is performed. */ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 999 | ZSTDLIB_API size_t ZSTD_freeDDict(ZSTD_DDict* ddict); |
| 1000 | |
| 1001 | /*! ZSTD_decompress_usingDDict() : |
| 1002 | * Decompression using a digested Dictionary. |
| 1003 | * Recommended when same dictionary is used multiple times. */ |
| 1004 | ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx, |
| 1005 | void* dst, size_t dstCapacity, |
| 1006 | const void* src, size_t srcSize, |
| 1007 | const ZSTD_DDict* ddict); |
| 1008 | |
| 1009 | |
Nick Terrell | 1dfe37f | 2019-04-05 17:44:56 -0700 | [diff] [blame] | 1010 | /******************************** |
| 1011 | * Dictionary helper functions |
| 1012 | *******************************/ |
| 1013 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1014 | /*! ZSTD_getDictID_fromDict() : Requires v1.4.0+ |
Nick Terrell | 1dfe37f | 2019-04-05 17:44:56 -0700 | [diff] [blame] | 1015 | * Provides the dictID stored within dictionary. |
| 1016 | * if @return == 0, the dictionary is not conformant with Zstandard specification. |
| 1017 | * It can still be loaded, but as a content-only dictionary. */ |
| 1018 | ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize); |
| 1019 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1020 | /*! ZSTD_getDictID_fromCDict() : Requires v1.5.0+ |
Sen Huang | 9c1ca3c | 2021-04-16 09:52:36 -0700 | [diff] [blame] | 1021 | * Provides the dictID of the dictionary loaded into `cdict`. |
| 1022 | * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty. |
| 1023 | * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */ |
| 1024 | ZSTDLIB_API unsigned ZSTD_getDictID_fromCDict(const ZSTD_CDict* cdict); |
| 1025 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1026 | /*! ZSTD_getDictID_fromDDict() : Requires v1.4.0+ |
Nick Terrell | 1dfe37f | 2019-04-05 17:44:56 -0700 | [diff] [blame] | 1027 | * Provides the dictID of the dictionary loaded into `ddict`. |
| 1028 | * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty. |
| 1029 | * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */ |
| 1030 | ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict); |
| 1031 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1032 | /*! ZSTD_getDictID_fromFrame() : Requires v1.4.0+ |
Nick Terrell | 1dfe37f | 2019-04-05 17:44:56 -0700 | [diff] [blame] | 1033 | * Provides the dictID required to decompressed the frame stored within `src`. |
| 1034 | * If @return == 0, the dictID could not be decoded. |
| 1035 | * This could for one of the following reasons : |
| 1036 | * - The frame does not require a dictionary to be decoded (most common case). |
Dominique Pelle | b772f53 | 2022-03-12 08:52:40 +0100 | [diff] [blame] | 1037 | * - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden piece of information. |
Nick Terrell | 1dfe37f | 2019-04-05 17:44:56 -0700 | [diff] [blame] | 1038 | * Note : this use case also happens when using a non-conformant dictionary. |
| 1039 | * - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`). |
| 1040 | * - This is not a Zstandard frame. |
| 1041 | * When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code. */ |
| 1042 | ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize); |
| 1043 | |
| 1044 | |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1045 | /******************************************************************************* |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1046 | * Advanced dictionary and prefix API (Requires v1.4.0+) |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1047 | * |
| 1048 | * This API allows dictionaries to be used with ZSTD_compress2(), |
Yann Collet | f4563d8 | 2023-03-14 17:17:21 -0700 | [diff] [blame] | 1049 | * ZSTD_compressStream2(), and ZSTD_decompressDCtx(). |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 1050 | * Dictionaries are sticky, they remain valid when same context is reused, |
Yann Collet | f4563d8 | 2023-03-14 17:17:21 -0700 | [diff] [blame] | 1051 | * they only reset when the context is reset |
| 1052 | * with ZSTD_reset_parameters or ZSTD_reset_session_and_parameters. |
| 1053 | * In contrast, Prefixes are single-use. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1054 | ******************************************************************************/ |
| 1055 | |
| 1056 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1057 | /*! ZSTD_CCtx_loadDictionary() : Requires v1.4.0+ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1058 | * Create an internal CDict from `dict` buffer. |
| 1059 | * Decompression will have to use same dictionary. |
| 1060 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
| 1061 | * Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary, |
| 1062 | * meaning "return to no-dictionary mode". |
Nick Terrell | e4018c4 | 2022-12-19 18:08:35 -0800 | [diff] [blame] | 1063 | * Note 1 : Dictionary is sticky, it will be used for all future compressed frames, |
| 1064 | * until parameters are reset, a new dictionary is loaded, or the dictionary |
| 1065 | * is explicitly invalidated by loading a NULL dictionary. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1066 | * Note 2 : Loading a dictionary involves building tables. |
| 1067 | * It's also a CPU consuming operation, with non-negligible impact on latency. |
| 1068 | * Tables are dependent on compression parameters, and for this reason, |
| 1069 | * compression parameters can no longer be changed after loading a dictionary. |
| 1070 | * Note 3 :`dict` content will be copied internally. |
| 1071 | * Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead. |
| 1072 | * In such a case, dictionary buffer must outlive its users. |
| 1073 | * Note 4 : Use ZSTD_CCtx_loadDictionary_advanced() |
Yann Collet | f4563d8 | 2023-03-14 17:17:21 -0700 | [diff] [blame] | 1074 | * to precisely select how dictionary content must be interpreted. |
| 1075 | * Note 5 : This method does not benefit from LDM (long distance mode). |
| 1076 | * If you want to employ LDM on some large dictionary content, |
| 1077 | * prefer employing ZSTD_CCtx_refPrefix() described below. |
| 1078 | */ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1079 | ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize); |
| 1080 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1081 | /*! ZSTD_CCtx_refCDict() : Requires v1.4.0+ |
Nick Terrell | e4018c4 | 2022-12-19 18:08:35 -0800 | [diff] [blame] | 1082 | * Reference a prepared dictionary, to be used for all future compressed frames. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1083 | * Note that compression parameters are enforced from within CDict, |
| 1084 | * and supersede any compression parameter previously set within CCtx. |
Thomas Waldmann | f9802d8 | 2021-01-07 18:47:23 +0100 | [diff] [blame] | 1085 | * The parameters ignored are labelled as "superseded-by-cdict" in the ZSTD_cParameter enum docs. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1086 | * The ignored parameters will be used again if the CCtx is returned to no-dictionary mode. |
| 1087 | * The dictionary will remain valid for future compressed frames using same CCtx. |
| 1088 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
| 1089 | * Special : Referencing a NULL CDict means "return to no-dictionary mode". |
| 1090 | * Note 1 : Currently, only one dictionary can be managed. |
| 1091 | * Referencing a new dictionary effectively "discards" any previous one. |
| 1092 | * Note 2 : CDict is just referenced, its lifetime must outlive its usage within CCtx. */ |
| 1093 | ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); |
| 1094 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1095 | /*! ZSTD_CCtx_refPrefix() : Requires v1.4.0+ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1096 | * Reference a prefix (single-usage dictionary) for next compressed frame. |
| 1097 | * A prefix is **only used once**. Tables are discarded at end of frame (ZSTD_e_end). |
| 1098 | * Decompression will need same prefix to properly regenerate data. |
| 1099 | * Compressing with a prefix is similar in outcome as performing a diff and compressing it, |
| 1100 | * but performs much faster, especially during decompression (compression speed is tunable with compression level). |
Yann Collet | f4563d8 | 2023-03-14 17:17:21 -0700 | [diff] [blame] | 1101 | * This method is compatible with LDM (long distance mode). |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1102 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
| 1103 | * Special: Adding any prefix (including NULL) invalidates any previous prefix or dictionary |
| 1104 | * Note 1 : Prefix buffer is referenced. It **must** outlive compression. |
| 1105 | * Its content must remain unmodified during compression. |
| 1106 | * Note 2 : If the intention is to diff some large src data blob with some prior version of itself, |
| 1107 | * ensure that the window size is large enough to contain the entire source. |
| 1108 | * See ZSTD_c_windowLog. |
| 1109 | * Note 3 : Referencing a prefix involves building tables, which are dependent on compression parameters. |
| 1110 | * It's a CPU consuming operation, with non-negligible impact on latency. |
| 1111 | * If there is a need to use the same prefix multiple times, consider loadDictionary instead. |
Yann Collet | 6323966 | 2019-10-16 16:14:04 -0700 | [diff] [blame] | 1112 | * Note 4 : By default, the prefix is interpreted as raw content (ZSTD_dct_rawContent). |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1113 | * Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation. */ |
| 1114 | ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, |
| 1115 | const void* prefix, size_t prefixSize); |
| 1116 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1117 | /*! ZSTD_DCtx_loadDictionary() : Requires v1.4.0+ |
Nick Terrell | e4018c4 | 2022-12-19 18:08:35 -0800 | [diff] [blame] | 1118 | * Create an internal DDict from dict buffer, to be used to decompress all future frames. |
| 1119 | * The dictionary remains valid for all future frames, until explicitly invalidated, or |
| 1120 | * a new dictionary is loaded. |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1121 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
| 1122 | * Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary, |
| 1123 | * meaning "return to no-dictionary mode". |
| 1124 | * Note 1 : Loading a dictionary involves building tables, |
| 1125 | * which has a non-negligible impact on CPU usage and latency. |
| 1126 | * It's recommended to "load once, use many times", to amortize the cost |
| 1127 | * Note 2 :`dict` content will be copied internally, so `dict` can be released after loading. |
| 1128 | * Use ZSTD_DCtx_loadDictionary_byReference() to reference dictionary content instead. |
| 1129 | * Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to take control of |
| 1130 | * how dictionary content is loaded and interpreted. |
| 1131 | */ |
| 1132 | ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize); |
| 1133 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1134 | /*! ZSTD_DCtx_refDDict() : Requires v1.4.0+ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1135 | * Reference a prepared dictionary, to be used to decompress next frames. |
| 1136 | * The dictionary remains active for decompression of future frames using same DCtx. |
Nick Terrell | 54a4998 | 2021-02-03 19:53:00 -0800 | [diff] [blame] | 1137 | * |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 1138 | * If called with ZSTD_d_refMultipleDDicts enabled, repeated calls of this function |
| 1139 | * will store the DDict references in a table, and the DDict used for decompression |
| 1140 | * will be determined at decompression time, as per the dict ID in the frame. |
| 1141 | * The memory for the table is allocated on the first call to refDDict, and can be |
| 1142 | * freed with ZSTD_freeDCtx(). |
Nick Terrell | 54a4998 | 2021-02-03 19:53:00 -0800 | [diff] [blame] | 1143 | * |
Nick Terrell | e4018c4 | 2022-12-19 18:08:35 -0800 | [diff] [blame] | 1144 | * If called with ZSTD_d_refMultipleDDicts disabled (the default), only one dictionary |
| 1145 | * will be managed, and referencing a dictionary effectively "discards" any previous one. |
| 1146 | * |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1147 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1148 | * Special: referencing a NULL DDict means "return to no-dictionary mode". |
| 1149 | * Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx. |
| 1150 | */ |
| 1151 | ZSTDLIB_API size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict); |
| 1152 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1153 | /*! ZSTD_DCtx_refPrefix() : Requires v1.4.0+ |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1154 | * Reference a prefix (single-usage dictionary) to decompress next frame. |
| 1155 | * This is the reverse operation of ZSTD_CCtx_refPrefix(), |
| 1156 | * and must use the same prefix as the one used during compression. |
| 1157 | * Prefix is **only used once**. Reference is discarded at end of frame. |
| 1158 | * End of frame is reached when ZSTD_decompressStream() returns 0. |
| 1159 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
| 1160 | * Note 1 : Adding any prefix (including NULL) invalidates any previously set prefix or dictionary |
| 1161 | * Note 2 : Prefix buffer is referenced. It **must** outlive decompression. |
| 1162 | * Prefix buffer must remain unmodified up to the end of frame, |
| 1163 | * reached when ZSTD_decompressStream() returns 0. |
Yann Collet | 6323966 | 2019-10-16 16:14:04 -0700 | [diff] [blame] | 1164 | * Note 3 : By default, the prefix is treated as raw content (ZSTD_dct_rawContent). |
Nick Terrell | cf7d601 | 2019-04-03 19:01:51 -0700 | [diff] [blame] | 1165 | * Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section) |
| 1166 | * Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost. |
| 1167 | * A full dictionary is more costly, as it requires building tables. |
| 1168 | */ |
| 1169 | ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, |
| 1170 | const void* prefix, size_t prefixSize); |
| 1171 | |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 1172 | /* === Memory management === */ |
| 1173 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 1174 | /*! ZSTD_sizeof_*() : Requires v1.4.0+ |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 1175 | * These functions give the _current_ memory usage of selected object. |
| 1176 | * Note that object memory usage can evolve (increase or decrease) over time. */ |
| 1177 | ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx); |
| 1178 | ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx); |
| 1179 | ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs); |
| 1180 | ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds); |
| 1181 | ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict); |
| 1182 | ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict); |
| 1183 | |
| 1184 | #endif /* ZSTD_H_235446 */ |
| 1185 | |
| 1186 | |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 1187 | /* ************************************************************************************** |
Nick Terrell | d7d8951 | 2019-03-28 19:04:32 -0600 | [diff] [blame] | 1188 | * ADVANCED AND EXPERIMENTAL FUNCTIONS |
| 1189 | **************************************************************************************** |
| 1190 | * The definitions in the following section are considered experimental. |
| 1191 | * They are provided for advanced scenarios. |
| 1192 | * They should never be used with a dynamic library, as prototypes may change in the future. |
| 1193 | * Use them only in association with static linking. |
| 1194 | * ***************************************************************************************/ |
| 1195 | |
| 1196 | #if defined(ZSTD_STATIC_LINKING_ONLY) && !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY) |
| 1197 | #define ZSTD_H_ZSTD_STATIC_LINKING_ONLY |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 1198 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1199 | /* This can be overridden externally to hide static symbols. */ |
| 1200 | #ifndef ZSTDLIB_STATIC_API |
| 1201 | # if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1) |
| 1202 | # define ZSTDLIB_STATIC_API __declspec(dllexport) ZSTDLIB_VISIBLE |
| 1203 | # elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1) |
| 1204 | # define ZSTDLIB_STATIC_API __declspec(dllimport) ZSTDLIB_VISIBLE |
| 1205 | # else |
| 1206 | # define ZSTDLIB_STATIC_API ZSTDLIB_VISIBLE |
| 1207 | # endif |
| 1208 | #endif |
| 1209 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1210 | /**************************************************************************************** |
| 1211 | * experimental API (static linking only) |
| 1212 | **************************************************************************************** |
| 1213 | * The following symbols and constants |
Yann Collet | fea9206 | 2018-11-21 14:47:23 -0800 | [diff] [blame] | 1214 | * are not planned to join "stable API" status in the near future. |
| 1215 | * They can still change in future versions. |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 1216 | * Some of them are planned to remain in the static_only section indefinitely. |
Yann Collet | fea9206 | 2018-11-21 14:47:23 -0800 | [diff] [blame] | 1217 | * Some of them might be removed in the future (especially when redundant with existing stable functions) |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1218 | * ***************************************************************************************/ |
| 1219 | |
Nick Terrell | b1ec94e | 2019-10-21 19:42:14 -0700 | [diff] [blame] | 1220 | #define ZSTD_FRAMEHEADERSIZE_PREFIX(format) ((format) == ZSTD_f_zstd1 ? 5 : 1) /* minimum input size required to query frame header size */ |
| 1221 | #define ZSTD_FRAMEHEADERSIZE_MIN(format) ((format) == ZSTD_f_zstd1 ? 6 : 2) |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 1222 | #define ZSTD_FRAMEHEADERSIZE_MAX 18 /* can be useful for static allocation */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1223 | #define ZSTD_SKIPPABLEHEADERSIZE 8 |
| 1224 | |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 1225 | /* compression parameter bounds */ |
Yann Collet | 2e7fd6a | 2018-11-20 15:13:27 -0800 | [diff] [blame] | 1226 | #define ZSTD_WINDOWLOG_MAX_32 30 |
| 1227 | #define ZSTD_WINDOWLOG_MAX_64 31 |
| 1228 | #define ZSTD_WINDOWLOG_MAX ((int)(sizeof(size_t) == 4 ? ZSTD_WINDOWLOG_MAX_32 : ZSTD_WINDOWLOG_MAX_64)) |
| 1229 | #define ZSTD_WINDOWLOG_MIN 10 |
| 1230 | #define ZSTD_HASHLOG_MAX ((ZSTD_WINDOWLOG_MAX < 30) ? ZSTD_WINDOWLOG_MAX : 30) |
| 1231 | #define ZSTD_HASHLOG_MIN 6 |
| 1232 | #define ZSTD_CHAINLOG_MAX_32 29 |
| 1233 | #define ZSTD_CHAINLOG_MAX_64 30 |
| 1234 | #define ZSTD_CHAINLOG_MAX ((int)(sizeof(size_t) == 4 ? ZSTD_CHAINLOG_MAX_32 : ZSTD_CHAINLOG_MAX_64)) |
| 1235 | #define ZSTD_CHAINLOG_MIN ZSTD_HASHLOG_MIN |
| 1236 | #define ZSTD_SEARCHLOG_MAX (ZSTD_WINDOWLOG_MAX-1) |
| 1237 | #define ZSTD_SEARCHLOG_MIN 1 |
| 1238 | #define ZSTD_MINMATCH_MAX 7 /* only for ZSTD_fast, other strategies are limited to 6 */ |
| 1239 | #define ZSTD_MINMATCH_MIN 3 /* only for ZSTD_btopt+, faster strategies are limited to 4 */ |
| 1240 | #define ZSTD_TARGETLENGTH_MAX ZSTD_BLOCKSIZE_MAX |
| 1241 | #define ZSTD_TARGETLENGTH_MIN 0 /* note : comparing this constant to an unsigned results in a tautological test */ |
Yann Collet | 39e2898 | 2018-12-06 16:16:16 -0800 | [diff] [blame] | 1242 | #define ZSTD_STRATEGY_MIN ZSTD_fast |
| 1243 | #define ZSTD_STRATEGY_MAX ZSTD_btultra2 |
Danielle Rozenblit | 14b8def | 2023-01-13 07:00:50 -0800 | [diff] [blame] | 1244 | #define ZSTD_BLOCKSIZE_MAX_MIN (1 << 10) /* The minimum valid max blocksize. Maximum blocksizes smaller than this make compressBound() inaccurate. */ |
Yann Collet | 39e2898 | 2018-12-06 16:16:16 -0800 | [diff] [blame] | 1245 | |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 1246 | |
Yann Collet | 5c6d4b1 | 2018-11-20 16:06:00 -0800 | [diff] [blame] | 1247 | #define ZSTD_OVERLAPLOG_MIN 0 |
| 1248 | #define ZSTD_OVERLAPLOG_MAX 9 |
| 1249 | |
Yann Collet | 0c404a4 | 2018-12-06 10:57:19 -0800 | [diff] [blame] | 1250 | #define ZSTD_WINDOWLOG_LIMIT_DEFAULT 27 /* by default, the streaming decoder will refuse any frame |
| 1251 | * requiring larger than (1<<ZSTD_WINDOWLOG_LIMIT_DEFAULT) window size, |
| 1252 | * to preserve host's memory from unreasonable requirements. |
Josh Soref | a880ca2 | 2019-04-12 14:18:11 -0400 | [diff] [blame] | 1253 | * This limit can be overridden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,). |
Yann Collet | 0c404a4 | 2018-12-06 10:57:19 -0800 | [diff] [blame] | 1254 | * The limit does not apply for one-pass decoders (such as ZSTD_decompress()), since no additional memory is allocated */ |
| 1255 | |
Yann Collet | 5c6d4b1 | 2018-11-20 16:06:00 -0800 | [diff] [blame] | 1256 | |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 1257 | /* LDM parameter bounds */ |
Yann Collet | 5c6d4b1 | 2018-11-20 16:06:00 -0800 | [diff] [blame] | 1258 | #define ZSTD_LDM_HASHLOG_MIN ZSTD_HASHLOG_MIN |
| 1259 | #define ZSTD_LDM_HASHLOG_MAX ZSTD_HASHLOG_MAX |
| 1260 | #define ZSTD_LDM_MINMATCH_MIN 4 |
| 1261 | #define ZSTD_LDM_MINMATCH_MAX 4096 |
| 1262 | #define ZSTD_LDM_BUCKETSIZELOG_MIN 1 |
| 1263 | #define ZSTD_LDM_BUCKETSIZELOG_MAX 8 |
Yann Collet | 41c7d0b | 2018-11-21 14:36:57 -0800 | [diff] [blame] | 1264 | #define ZSTD_LDM_HASHRATELOG_MIN 0 |
| 1265 | #define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN) |
Yann Collet | 5c6d4b1 | 2018-11-20 16:06:00 -0800 | [diff] [blame] | 1266 | |
Ephraim Park | 9007701 | 2019-06-24 13:40:52 -0700 | [diff] [blame] | 1267 | /* Advanced parameter bounds */ |
Yann Collet | f5728da | 2024-03-18 12:04:02 -0700 | [diff] [blame] | 1268 | #define ZSTD_TARGETCBLOCKSIZE_MIN 1340 /* suitable to fit into an ethernet / wifi / 4G transport frame */ |
Ephraim Park | 9007701 | 2019-06-24 13:40:52 -0700 | [diff] [blame] | 1269 | #define ZSTD_TARGETCBLOCKSIZE_MAX ZSTD_BLOCKSIZE_MAX |
Nick Magerko | de6a6c7 | 2019-08-20 13:07:51 -0700 | [diff] [blame] | 1270 | #define ZSTD_SRCSIZEHINT_MIN 0 |
Nick Magerko | fee8fbc | 2019-08-19 12:58:54 -0700 | [diff] [blame] | 1271 | #define ZSTD_SRCSIZEHINT_MAX INT_MAX |
Ephraim Park | 9007701 | 2019-06-24 13:40:52 -0700 | [diff] [blame] | 1272 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1273 | |
| 1274 | /* --- Advanced types --- */ |
| 1275 | |
| 1276 | typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params; |
| 1277 | |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1278 | typedef struct { |
senhuang42 | 3ed5d05 | 2020-10-27 17:45:23 -0400 | [diff] [blame] | 1279 | unsigned int offset; /* The offset of the match. (NOT the same as the offset code) |
| 1280 | * If offset == 0 and matchLength == 0, this sequence represents the last |
| 1281 | * literals in the block of litLength size. |
senhuang42 | 9171f92 | 2020-10-27 10:43:37 -0400 | [diff] [blame] | 1282 | */ |
senhuang42 | 761f40d | 2020-10-27 09:41:32 -0400 | [diff] [blame] | 1283 | |
senhuang42 | dc44856 | 2020-10-27 12:28:46 -0400 | [diff] [blame] | 1284 | unsigned int litLength; /* Literal length of the sequence. */ |
| 1285 | unsigned int matchLength; /* Match length of the sequence. */ |
| 1286 | |
| 1287 | /* Note: Users of this API may provide a sequence with matchLength == litLength == offset == 0. |
senhuang42 | 3ed5d05 | 2020-10-27 17:45:23 -0400 | [diff] [blame] | 1288 | * In this case, we will treat the sequence as a marker for a block boundary. |
senhuang42 | dc44856 | 2020-10-27 12:28:46 -0400 | [diff] [blame] | 1289 | */ |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1290 | |
senhuang42 | 3ed5d05 | 2020-10-27 17:45:23 -0400 | [diff] [blame] | 1291 | unsigned int rep; /* Represents which repeat offset is represented by the field 'offset'. |
| 1292 | * Ranges from [0, 3]. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1293 | * |
senhuang42 | 3ed5d05 | 2020-10-27 17:45:23 -0400 | [diff] [blame] | 1294 | * Repeat offsets are essentially previous offsets from previous sequences sorted in |
| 1295 | * recency order. For more detail, see doc/zstd_compression_format.md |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1296 | * |
senhuang42 | 3ed5d05 | 2020-10-27 17:45:23 -0400 | [diff] [blame] | 1297 | * If rep == 0, then 'offset' does not contain a repeat offset. |
senhuang42 | dc44856 | 2020-10-27 12:28:46 -0400 | [diff] [blame] | 1298 | * If rep > 0: |
senhuang42 | 9171f92 | 2020-10-27 10:43:37 -0400 | [diff] [blame] | 1299 | * If litLength != 0: |
senhuang42 | dc44856 | 2020-10-27 12:28:46 -0400 | [diff] [blame] | 1300 | * rep == 1 --> offset == repeat_offset_1 |
| 1301 | * rep == 2 --> offset == repeat_offset_2 |
| 1302 | * rep == 3 --> offset == repeat_offset_3 |
senhuang42 | 9171f92 | 2020-10-27 10:43:37 -0400 | [diff] [blame] | 1303 | * If litLength == 0: |
senhuang42 | dc44856 | 2020-10-27 12:28:46 -0400 | [diff] [blame] | 1304 | * rep == 1 --> offset == repeat_offset_2 |
| 1305 | * rep == 2 --> offset == repeat_offset_3 |
| 1306 | * rep == 3 --> offset == repeat_offset_1 - 1 |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1307 | * |
senhuang42 | 51abd58 | 2020-11-06 10:53:22 -0500 | [diff] [blame] | 1308 | * Note: This field is optional. ZSTD_generateSequences() will calculate the value of |
senhuang42 | 3ed5d05 | 2020-10-27 17:45:23 -0400 | [diff] [blame] | 1309 | * 'rep', but repeat offsets do not necessarily need to be calculated from an external |
senhuang42 | 3c44547 | 2020-11-18 09:52:24 -0500 | [diff] [blame] | 1310 | * sequence provider's perspective. For example, ZSTD_compressSequences() does not |
| 1311 | * use this 'rep' field at all (as of now). |
senhuang42 | 9171f92 | 2020-10-27 10:43:37 -0400 | [diff] [blame] | 1312 | */ |
Bimba Shrestha | 5f8b0f6 | 2019-08-30 09:18:44 -0700 | [diff] [blame] | 1313 | } ZSTD_Sequence; |
| 1314 | |
| 1315 | typedef struct { |
Yann Collet | 2c8fde5 | 2018-11-13 17:36:35 -0800 | [diff] [blame] | 1316 | unsigned windowLog; /**< largest match distance : larger == more compression, more memory needed during decompression */ |
| 1317 | unsigned chainLog; /**< fully searched segment : larger == more compression, slower, more memory (useless for fast) */ |
| 1318 | unsigned hashLog; /**< dispatch table : larger == faster, more memory */ |
| 1319 | unsigned searchLog; /**< nb of searches : larger == more compression, slower */ |
Yann Collet | 2e7fd6a | 2018-11-20 15:13:27 -0800 | [diff] [blame] | 1320 | unsigned minMatch; /**< match length searched : larger == faster decompression, sometimes less compression */ |
Yann Collet | 2c8fde5 | 2018-11-13 17:36:35 -0800 | [diff] [blame] | 1321 | unsigned targetLength; /**< acceptable match size for optimal parser (only) : larger == more compression, slower */ |
| 1322 | ZSTD_strategy strategy; /**< see ZSTD_strategy definition above */ |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1323 | } ZSTD_compressionParameters; |
| 1324 | |
| 1325 | typedef struct { |
Yann Collet | d4d4e10 | 2018-11-21 15:37:26 -0800 | [diff] [blame] | 1326 | int contentSizeFlag; /**< 1: content size will be in frame header (when known) */ |
| 1327 | int checksumFlag; /**< 1: generate a 32-bits checksum using XXH64 algorithm at end of frame, for error detection */ |
| 1328 | int noDictIDFlag; /**< 1: no dictID will be saved into frame header (dictID is only useful for dictionary compression) */ |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1329 | } ZSTD_frameParameters; |
| 1330 | |
| 1331 | typedef struct { |
| 1332 | ZSTD_compressionParameters cParams; |
| 1333 | ZSTD_frameParameters fParams; |
| 1334 | } ZSTD_parameters; |
| 1335 | |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1336 | typedef enum { |
Yann Collet | 2c8fde5 | 2018-11-13 17:36:35 -0800 | [diff] [blame] | 1337 | ZSTD_dct_auto = 0, /* dictionary is "full" when starting with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */ |
| 1338 | ZSTD_dct_rawContent = 1, /* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */ |
| 1339 | ZSTD_dct_fullDict = 2 /* refuses to load a dictionary if it does not respect Zstandard's specification, starting with ZSTD_MAGIC_DICTIONARY */ |
Yann Collet | 6873fec | 2018-03-20 15:13:14 -0700 | [diff] [blame] | 1340 | } ZSTD_dictContentType_e; |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1341 | |
| 1342 | typedef enum { |
Yann Collet | 2c8fde5 | 2018-11-13 17:36:35 -0800 | [diff] [blame] | 1343 | ZSTD_dlm_byCopy = 0, /**< Copy dictionary content internally */ |
Yann Collet | cb18fff | 2019-09-24 17:50:58 -0700 | [diff] [blame] | 1344 | ZSTD_dlm_byRef = 1 /**< Reference dictionary content -- the dictionary buffer must outlive its users. */ |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1345 | } ZSTD_dictLoadMethod_e; |
| 1346 | |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1347 | typedef enum { |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1348 | ZSTD_f_zstd1 = 0, /* zstd frame format, specified in zstd_compression_format.md (default) */ |
Yann Collet | cb18fff | 2019-09-24 17:50:58 -0700 | [diff] [blame] | 1349 | ZSTD_f_zstd1_magicless = 1 /* Variant of zstd frame format, without initial 4-bytes magic number. |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1350 | * Useful to save 4 bytes per generated frame. |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 1351 | * Decoder cannot recognise automatically this format, requiring this instruction. */ |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1352 | } ZSTD_format_e; |
| 1353 | |
| 1354 | typedef enum { |
senhuang42 | 44c54a3 | 2020-08-24 16:14:19 -0400 | [diff] [blame] | 1355 | /* Note: this enum controls ZSTD_d_forceIgnoreChecksum */ |
senhuang42 | 47685ac | 2020-08-21 18:18:53 -0400 | [diff] [blame] | 1356 | ZSTD_d_validateChecksum = 0, |
| 1357 | ZSTD_d_ignoreChecksum = 1 |
| 1358 | } ZSTD_forceIgnoreChecksum_e; |
| 1359 | |
| 1360 | typedef enum { |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 1361 | /* Note: this enum controls ZSTD_d_refMultipleDDicts */ |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 1362 | ZSTD_rmd_refSingleDDict = 0, |
| 1363 | ZSTD_rmd_refMultipleDDicts = 1 |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 1364 | } ZSTD_refMultipleDDicts_e; |
| 1365 | |
| 1366 | typedef enum { |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1367 | /* Note: this enum and the behavior it controls are effectively internal |
| 1368 | * implementation details of the compressor. They are expected to continue |
| 1369 | * to evolve and should be considered only in the context of extremely |
| 1370 | * advanced performance tuning. |
| 1371 | * |
Sen Huang | e4de8b0 | 2019-10-20 14:03:09 -0400 | [diff] [blame] | 1372 | * Zstd currently supports the use of a CDict in three ways: |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1373 | * |
| 1374 | * - The contents of the CDict can be copied into the working context. This |
| 1375 | * means that the compression can search both the dictionary and input |
| 1376 | * while operating on a single set of internal tables. This makes |
| 1377 | * the compression faster per-byte of input. However, the initial copy of |
| 1378 | * the CDict's tables incurs a fixed cost at the beginning of the |
| 1379 | * compression. For small compressions (< 8 KB), that copy can dominate |
| 1380 | * the cost of the compression. |
| 1381 | * |
| 1382 | * - The CDict's tables can be used in-place. In this model, compression is |
| 1383 | * slower per input byte, because the compressor has to search two sets of |
| 1384 | * tables. However, this model incurs no start-up cost (as long as the |
| 1385 | * working context's tables can be reused). For small inputs, this can be |
| 1386 | * faster than copying the CDict's tables. |
Yann Collet | 74065da | 2019-10-28 11:15:41 -0700 | [diff] [blame] | 1387 | * |
Sen Huang | e4de8b0 | 2019-10-20 14:03:09 -0400 | [diff] [blame] | 1388 | * - The CDict's tables are not used at all, and instead we use the working |
Sen Huang | 676f899 | 2019-10-21 15:29:55 -0400 | [diff] [blame] | 1389 | * context alone to reload the dictionary and use params based on the source |
| 1390 | * size. See ZSTD_compress_insertDictionary() and ZSTD_compress_usingDict(). |
| 1391 | * This method is effective when the dictionary sizes are very small relative |
| 1392 | * to the input size, and the input size is fairly large to begin with. |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1393 | * |
| 1394 | * Zstd has a simple internal heuristic that selects which strategy to use |
| 1395 | * at the beginning of a compression. However, if experimentation shows that |
| 1396 | * Zstd is making poor choices, it is possible to override that choice with |
| 1397 | * this enum. |
| 1398 | */ |
| 1399 | ZSTD_dictDefaultAttach = 0, /* Use the default heuristic. */ |
| 1400 | ZSTD_dictForceAttach = 1, /* Never copy the dictionary. */ |
| 1401 | ZSTD_dictForceCopy = 2, /* Always copy the dictionary. */ |
Sen Huang | 59c81aa | 2019-10-21 19:12:15 -0400 | [diff] [blame] | 1402 | ZSTD_dictForceLoad = 3 /* Always reload the dictionary */ |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 1403 | } ZSTD_dictAttachPref_e; |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1404 | |
Nick Terrell | f951311 | 2019-02-13 14:59:22 -0800 | [diff] [blame] | 1405 | typedef enum { |
sen | c730b8c | 2021-05-12 16:09:48 -0400 | [diff] [blame] | 1406 | ZSTD_lcm_auto = 0, /**< Automatically determine the compression mode based on the compression level. |
| 1407 | * Negative compression levels will be uncompressed, and positive compression |
| 1408 | * levels will be compressed. */ |
| 1409 | ZSTD_lcm_huffman = 1, /**< Always attempt Huffman compression. Uncompressed literals will still be |
| 1410 | * emitted if Huffman compression is not profitable. */ |
| 1411 | ZSTD_lcm_uncompressed = 2 /**< Always emit uncompressed literals. */ |
| 1412 | } ZSTD_literalCompressionMode_e; |
| 1413 | |
| 1414 | typedef enum { |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 1415 | /* Note: This enum controls features which are conditionally beneficial. Zstd typically will make a final |
| 1416 | * decision on whether or not to enable the feature (ZSTD_ps_auto), but setting the switch to ZSTD_ps_enable |
| 1417 | * or ZSTD_ps_disable allow for a force enable/disable the feature. |
| 1418 | */ |
| 1419 | ZSTD_ps_auto = 0, /* Let the library automatically determine whether the feature shall be enabled */ |
| 1420 | ZSTD_ps_enable = 1, /* Force-enable the feature */ |
| 1421 | ZSTD_ps_disable = 2 /* Do not use the feature */ |
| 1422 | } ZSTD_paramSwitch_e; |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1423 | |
Sean Purcell | 2db7249 | 2017-02-09 10:50:43 -0800 | [diff] [blame] | 1424 | /*************************************** |
Nick Terrell | fbd97f3 | 2023-01-24 12:21:49 -0800 | [diff] [blame] | 1425 | * Frame header and size functions |
Sean Purcell | d7bfcac | 2017-02-10 11:38:57 -0800 | [diff] [blame] | 1426 | ***************************************/ |
| 1427 | |
Sean Purcell | 2db7249 | 2017-02-09 10:50:43 -0800 | [diff] [blame] | 1428 | /*! ZSTD_findDecompressedSize() : |
shakeelrao | 9ad3f31 | 2019-03-02 17:56:10 -0800 | [diff] [blame] | 1429 | * `src` should point to the start of a series of ZSTD encoded and/or skippable frames |
Yann Collet | 97f2bf6 | 2017-06-29 11:31:40 -0700 | [diff] [blame] | 1430 | * `srcSize` must be the _exact_ size of this series |
shakeelrao | 9ad3f31 | 2019-03-02 17:56:10 -0800 | [diff] [blame] | 1431 | * (i.e. there should be a frame boundary at `src + srcSize`) |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1432 | * @return : - decompressed size of all data in all successive frames |
| 1433 | * - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN |
| 1434 | * - if an error occurred: ZSTD_CONTENTSIZE_ERROR |
| 1435 | * |
| 1436 | * note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode. |
| 1437 | * When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size. |
| 1438 | * In which case, it's necessary to use streaming mode to decompress data. |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1439 | * note 2 : decompressed size is always present when compression is done with ZSTD_compress() |
| 1440 | * note 3 : decompressed size can be very large (64-bits value), |
| 1441 | * potentially larger than what local system can handle as a single memory segment. |
| 1442 | * In which case, it's necessary to use streaming mode to decompress data. |
| 1443 | * note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified. |
| 1444 | * Always ensure result fits within application's authorized limits. |
| 1445 | * Each application can set its own limits. |
| 1446 | * note 5 : ZSTD_findDecompressedSize handles multiple frames, and so it must traverse the input to |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 1447 | * read each contained frame header. This is fast as most of the data is skipped, |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1448 | * however it does mean that all frame data must be present and valid. */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1449 | ZSTDLIB_STATIC_API unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize); |
Sean Purcell | 2db7249 | 2017-02-09 10:50:43 -0800 | [diff] [blame] | 1450 | |
Yann Collet | b3af187 | 2019-06-04 10:35:40 -0700 | [diff] [blame] | 1451 | /*! ZSTD_decompressBound() : |
shakeelrao | 9ad3f31 | 2019-03-02 17:56:10 -0800 | [diff] [blame] | 1452 | * `src` should point to the start of a series of ZSTD encoded and/or skippable frames |
shakeelrao | 1e08c49 | 2019-03-01 18:29:35 -0800 | [diff] [blame] | 1453 | * `srcSize` must be the _exact_ size of this series |
shakeelrao | 9ad3f31 | 2019-03-02 17:56:10 -0800 | [diff] [blame] | 1454 | * (i.e. there should be a frame boundary at `src + srcSize`) |
| 1455 | * @return : - upper-bound for the decompressed size of all data in all successive frames |
Thomas Waldmann | f9802d8 | 2021-01-07 18:47:23 +0100 | [diff] [blame] | 1456 | * - if an error occurred: ZSTD_CONTENTSIZE_ERROR |
shakeelrao | 8930c3c | 2019-02-28 22:55:18 -0800 | [diff] [blame] | 1457 | * |
shakeelrao | 0a3fa6f | 2019-03-18 20:33:15 -0700 | [diff] [blame] | 1458 | * note 1 : an error can occur if `src` contains an invalid or incorrectly formatted frame. |
shakeelrao | 9ad3f31 | 2019-03-02 17:56:10 -0800 | [diff] [blame] | 1459 | * note 2 : the upper-bound is exact when the decompressed size field is available in every ZSTD encoded frame of `src`. |
| 1460 | * in this case, `ZSTD_findDecompressedSize` and `ZSTD_decompressBound` return the same value. |
| 1461 | * note 3 : when the decompressed size field isn't available, the upper-bound for that frame is calculated by: |
shakeelrao | 1e08c49 | 2019-03-01 18:29:35 -0800 | [diff] [blame] | 1462 | * upper-bound = # blocks * min(128 KB, Window_Size) |
shakeelrao | 8930c3c | 2019-02-28 22:55:18 -0800 | [diff] [blame] | 1463 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1464 | ZSTDLIB_STATIC_API unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize); |
shakeelrao | 8930c3c | 2019-02-28 22:55:18 -0800 | [diff] [blame] | 1465 | |
Yann Collet | a54c86c | 2018-09-20 16:17:49 -0700 | [diff] [blame] | 1466 | /*! ZSTD_frameHeaderSize() : |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 1467 | * srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX. |
Yann Collet | 8be984e | 2018-03-30 20:09:27 -0700 | [diff] [blame] | 1468 | * @return : size of the Frame Header, |
| 1469 | * or an error code (if srcSize is too small) */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1470 | ZSTDLIB_STATIC_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize); |
Yann Collet | acbef3d | 2017-06-29 05:18:09 -0700 | [diff] [blame] | 1471 | |
Nick Terrell | fbd97f3 | 2023-01-24 12:21:49 -0800 | [diff] [blame] | 1472 | typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e; |
| 1473 | typedef struct { |
| 1474 | unsigned long long frameContentSize; /* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */ |
| 1475 | unsigned long long windowSize; /* can be very large, up to <= frameContentSize */ |
| 1476 | unsigned blockSizeMax; |
| 1477 | ZSTD_frameType_e frameType; /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */ |
| 1478 | unsigned headerSize; |
| 1479 | unsigned dictID; |
| 1480 | unsigned checksumFlag; |
| 1481 | unsigned _reserved1; |
| 1482 | unsigned _reserved2; |
| 1483 | } ZSTD_frameHeader; |
| 1484 | |
| 1485 | /*! ZSTD_getFrameHeader() : |
| 1486 | * decode Frame Header, or requires larger `srcSize`. |
| 1487 | * @return : 0, `zfhPtr` is correctly filled, |
| 1488 | * >0, `srcSize` is too small, value is wanted `srcSize` amount, |
| 1489 | * or an error code, which can be tested using ZSTD_isError() */ |
| 1490 | ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize); /**< doesn't consume input */ |
| 1491 | /*! ZSTD_getFrameHeader_advanced() : |
| 1492 | * same as ZSTD_getFrameHeader(), |
| 1493 | * with added capability to select a format (like ZSTD_f_zstd1_magicless) */ |
| 1494 | ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format); |
| 1495 | |
Nick Terrell | 5b26619 | 2023-01-11 18:14:40 -0800 | [diff] [blame] | 1496 | /*! ZSTD_decompressionMargin() : |
| 1497 | * Zstd supports in-place decompression, where the input and output buffers overlap. |
| 1498 | * In this case, the output buffer must be at least (Margin + Output_Size) bytes large, |
| 1499 | * and the input buffer must be at the end of the output buffer. |
| 1500 | * |
| 1501 | * _______________________ Output Buffer ________________________ |
| 1502 | * | | |
| 1503 | * | ____ Input Buffer ____| |
| 1504 | * | | | |
| 1505 | * v v v |
| 1506 | * |---------------------------------------|-----------|----------| |
| 1507 | * ^ ^ ^ |
| 1508 | * |___________________ Output_Size ___________________|_ Margin _| |
| 1509 | * |
| 1510 | * NOTE: See also ZSTD_DECOMPRESSION_MARGIN(). |
| 1511 | * NOTE: This applies only to single-pass decompression through ZSTD_decompress() or |
| 1512 | * ZSTD_decompressDCtx(). |
| 1513 | * NOTE: This function supports multi-frame input. |
| 1514 | * |
| 1515 | * @param src The compressed frame(s) |
| 1516 | * @param srcSize The size of the compressed frame(s) |
| 1517 | * @returns The decompression margin or an error that can be checked with ZSTD_isError(). |
| 1518 | */ |
| 1519 | ZSTDLIB_STATIC_API size_t ZSTD_decompressionMargin(const void* src, size_t srcSize); |
| 1520 | |
| 1521 | /*! ZSTD_DECOMPRESS_MARGIN() : |
| 1522 | * Similar to ZSTD_decompressionMargin(), but instead of computing the margin from |
| 1523 | * the compressed frame, compute it from the original size and the blockSizeLog. |
| 1524 | * See ZSTD_decompressionMargin() for details. |
| 1525 | * |
| 1526 | * WARNING: This macro does not support multi-frame input, the input must be a single |
| 1527 | * zstd frame. If you need that support use the function, or implement it yourself. |
| 1528 | * |
| 1529 | * @param originalSize The original uncompressed size of the data. |
| 1530 | * @param blockSize The block size == MIN(windowSize, ZSTD_BLOCKSIZE_MAX). |
| 1531 | * Unless you explicitly set the windowLog smaller than |
| 1532 | * ZSTD_BLOCKSIZELOG_MAX you can just use ZSTD_BLOCKSIZE_MAX. |
| 1533 | */ |
| 1534 | #define ZSTD_DECOMPRESSION_MARGIN(originalSize, blockSize) ((size_t)( \ |
| 1535 | ZSTD_FRAMEHEADERSIZE_MAX /* Frame header */ + \ |
| 1536 | 4 /* checksum */ + \ |
| 1537 | ((originalSize) == 0 ? 0 : 3 * (((originalSize) + (blockSize) - 1) / blockSize)) /* 3 bytes per block */ + \ |
| 1538 | (blockSize) /* One block of margin */ \ |
| 1539 | )) |
| 1540 | |
senhuang42 | 7397d01 | 2020-11-02 10:15:53 -0500 | [diff] [blame] | 1541 | typedef enum { |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 1542 | ZSTD_sf_noBlockDelimiters = 0, /* Representation of ZSTD_Sequence has no block delimiters, sequences only */ |
| 1543 | ZSTD_sf_explicitBlockDelimiters = 1 /* Representation of ZSTD_Sequence contains explicit block delimiters */ |
| 1544 | } ZSTD_sequenceFormat_e; |
senhuang42 | 7397d01 | 2020-11-02 10:15:53 -0500 | [diff] [blame] | 1545 | |
Danielle Rozenblit | aa82998 | 2022-09-09 12:34:25 -0700 | [diff] [blame] | 1546 | /*! ZSTD_sequenceBound() : |
| 1547 | * `srcSize` : size of the input buffer |
| 1548 | * @return : upper-bound for the number of sequences that can be generated |
| 1549 | * from a buffer of srcSize bytes |
| 1550 | * |
| 1551 | * note : returns number of sequences - to get bytes, multiply by sizeof(ZSTD_Sequence). |
| 1552 | */ |
| 1553 | ZSTDLIB_STATIC_API size_t ZSTD_sequenceBound(size_t srcSize); |
| 1554 | |
senhuang42 | 51abd58 | 2020-11-06 10:53:22 -0500 | [diff] [blame] | 1555 | /*! ZSTD_generateSequences() : |
Nick Terrell | 731f4b7 | 2024-03-19 12:37:55 -0700 | [diff] [blame] | 1556 | * WARNING: This function is meant for debugging and informational purposes ONLY! |
| 1557 | * Its implementation is flawed, and it will be deleted in a future version. |
| 1558 | * It is not guaranteed to succeed, as there are several cases where it will give |
| 1559 | * up and fail. You should NOT use this function in production code. |
| 1560 | * |
| 1561 | * This function is deprecated, and will be removed in a future version. |
| 1562 | * |
Yann Collet | 87dcd33 | 2022-01-22 02:30:08 -0800 | [diff] [blame] | 1563 | * Generate sequences using ZSTD_compress2(), given a source buffer. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1564 | * |
Nick Terrell | 731f4b7 | 2024-03-19 12:37:55 -0700 | [diff] [blame] | 1565 | * @param zc The compression context to be used for ZSTD_compress2(). Set any |
| 1566 | * compression parameters you need on this context. |
| 1567 | * @param outSeqs The output sequences buffer of size @p outSeqsSize |
| 1568 | * @param outSeqsSize The size of the output sequences buffer. |
| 1569 | * ZSTD_sequenceBound(srcSize) is an upper bound on the number |
| 1570 | * of sequences that can be generated. |
| 1571 | * @param src The source buffer to generate sequences from of size @p srcSize. |
| 1572 | * @param srcSize The size of the source buffer. |
| 1573 | * |
senhuang42 | 261ea69 | 2020-11-06 10:52:34 -0500 | [diff] [blame] | 1574 | * Each block will end with a dummy sequence |
| 1575 | * with offset == 0, matchLength == 0, and litLength == length of last literals. |
senhuang42 | 7d1dea0 | 2020-11-06 10:56:56 -0500 | [diff] [blame] | 1576 | * litLength may be == 0, and if so, then the sequence of (of: 0 ml: 0 ll: 0) |
| 1577 | * simply acts as a block delimiter. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1578 | * |
Nick Terrell | 731f4b7 | 2024-03-19 12:37:55 -0700 | [diff] [blame] | 1579 | * @returns The number of sequences generated, necessarily less than |
| 1580 | * ZSTD_sequenceBound(srcSize), or an error code that can be checked |
| 1581 | * with ZSTD_isError(). |
Bimba Shrestha | bff6072 | 2019-09-16 08:26:21 -0700 | [diff] [blame] | 1582 | */ |
Nick Terrell | 731f4b7 | 2024-03-19 12:37:55 -0700 | [diff] [blame] | 1583 | ZSTD_DEPRECATED("For debugging only, will be replaced by ZSTD_extractSequences()") |
Yann Collet | 87dcd33 | 2022-01-22 02:30:08 -0800 | [diff] [blame] | 1584 | ZSTDLIB_STATIC_API size_t |
Nick Terrell | 731f4b7 | 2024-03-19 12:37:55 -0700 | [diff] [blame] | 1585 | ZSTD_generateSequences(ZSTD_CCtx* zc, |
| 1586 | ZSTD_Sequence* outSeqs, size_t outSeqsSize, |
| 1587 | const void* src, size_t srcSize); |
Bimba Shrestha | 5f8b0f6 | 2019-08-30 09:18:44 -0700 | [diff] [blame] | 1588 | |
senhuang42 | 7d1dea0 | 2020-11-06 10:56:56 -0500 | [diff] [blame] | 1589 | /*! ZSTD_mergeBlockDelimiters() : |
| 1590 | * Given an array of ZSTD_Sequence, remove all sequences that represent block delimiters/last literals |
Tom Rix | 2b957af | 2022-01-24 12:43:39 -0800 | [diff] [blame] | 1591 | * by merging them into the literals of the next sequence. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1592 | * |
senhuang42 | 261ea69 | 2020-11-06 10:52:34 -0500 | [diff] [blame] | 1593 | * As such, the final generated result has no explicit representation of block boundaries, |
| 1594 | * and the final last literals segment is not represented in the sequences. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1595 | * |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 1596 | * The output of this function can be fed into ZSTD_compressSequences() with CCtx |
| 1597 | * setting of ZSTD_c_blockDelimiters as ZSTD_sf_noBlockDelimiters |
senhuang42 | 7d1dea0 | 2020-11-06 10:56:56 -0500 | [diff] [blame] | 1598 | * @return : number of sequences left after merging |
senhuang42 | 261ea69 | 2020-11-06 10:52:34 -0500 | [diff] [blame] | 1599 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1600 | ZSTDLIB_STATIC_API size_t ZSTD_mergeBlockDelimiters(ZSTD_Sequence* sequences, size_t seqsSize); |
senhuang42 | c675fb4 | 2020-11-03 13:37:50 -0500 | [diff] [blame] | 1601 | |
senhuang42 | b8e16a2 | 2020-11-06 11:35:10 -0500 | [diff] [blame] | 1602 | /*! ZSTD_compressSequences() : |
Dominique Pelle | b772f53 | 2022-03-12 08:52:40 +0100 | [diff] [blame] | 1603 | * Compress an array of ZSTD_Sequence, associated with @src buffer, into dst. |
Yann Collet | 7616e39 | 2022-01-20 21:23:48 -0800 | [diff] [blame] | 1604 | * @src contains the entire input (not just the literals). |
| 1605 | * If @srcSize > sum(sequence.length), the remaining bytes are considered all literals |
senhuang42 | 3c44547 | 2020-11-18 09:52:24 -0500 | [diff] [blame] | 1606 | * If a dictionary is included, then the cctx should reference the dict. (see: ZSTD_CCtx_refCDict(), ZSTD_CCtx_loadDictionary(), etc.) |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1607 | * The entire source is compressed into a single frame. |
| 1608 | * |
senhuang42 | 3c44547 | 2020-11-18 09:52:24 -0500 | [diff] [blame] | 1609 | * The compression behavior changes based on cctx params. In particular: |
| 1610 | * If ZSTD_c_blockDelimiters == ZSTD_sf_noBlockDelimiters, the array of ZSTD_Sequence is expected to contain |
| 1611 | * no block delimiters (defined in ZSTD_Sequence). Block boundaries are roughly determined based on |
| 1612 | * the block size derived from the cctx, and sequences may be split. This is the default setting. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1613 | * |
senhuang42 | 3c44547 | 2020-11-18 09:52:24 -0500 | [diff] [blame] | 1614 | * If ZSTD_c_blockDelimiters == ZSTD_sf_explicitBlockDelimiters, the array of ZSTD_Sequence is expected to contain |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 1615 | * block delimiters (defined in ZSTD_Sequence). Behavior is undefined if no block delimiters are provided. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1616 | * |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 1617 | * If ZSTD_c_validateSequences == 0, this function will blindly accept the sequences provided. Invalid sequences cause undefined |
| 1618 | * behavior. If ZSTD_c_validateSequences == 1, then if sequence is invalid (see doc/zstd_compression_format.md for |
| 1619 | * specifics regarding offset/matchlength requirements) then the function will bail out and return an error. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1620 | * |
senhuang42 | 3efe9c9 | 2020-12-01 10:53:30 -0500 | [diff] [blame] | 1621 | * In addition to the two adjustable experimental params, there are other important cctx params. |
| 1622 | * - ZSTD_c_minMatch MUST be set as less than or equal to the smallest match generated by the match finder. It has a minimum value of ZSTD_MINMATCH_MIN. |
| 1623 | * - ZSTD_c_compressionLevel accordingly adjusts the strength of the entropy coder, as it would in typical compression. |
| 1624 | * - ZSTD_c_windowLog affects offset validation: this function will return an error at higher debug levels if a provided offset |
senhuang42 | 3c44547 | 2020-11-18 09:52:24 -0500 | [diff] [blame] | 1625 | * is larger than what the spec allows for a given window log and dictionary (if present). See: doc/zstd_compression_format.md |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1626 | * |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 1627 | * Note: Repcodes are, as of now, always re-calculated within this function, so ZSTD_Sequence::rep is unused. |
| 1628 | * Note 2: Once we integrate ability to ingest repcodes, the explicit block delims mode must respect those repcodes exactly, |
| 1629 | * and cannot emit an RLE block that disagrees with the repcode history |
Yann Collet | 87dcd33 | 2022-01-22 02:30:08 -0800 | [diff] [blame] | 1630 | * @return : final compressed size, or a ZSTD error code. |
senhuang42 | 5fd69f8 | 2020-11-03 13:52:21 -0500 | [diff] [blame] | 1631 | */ |
Yann Collet | 87dcd33 | 2022-01-22 02:30:08 -0800 | [diff] [blame] | 1632 | ZSTDLIB_STATIC_API size_t |
| 1633 | ZSTD_compressSequences( ZSTD_CCtx* cctx, void* dst, size_t dstSize, |
| 1634 | const ZSTD_Sequence* inSeqs, size_t inSeqsSize, |
| 1635 | const void* src, size_t srcSize); |
senhuang42 | a9327b1 | 2020-11-13 09:55:05 -0500 | [diff] [blame] | 1636 | |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1637 | |
senhuang42 | 5c41490 | 2020-12-21 11:33:41 -0500 | [diff] [blame] | 1638 | /*! ZSTD_writeSkippableFrame() : |
senhuang42 | 7e11bd0 | 2020-12-21 11:13:22 -0500 | [diff] [blame] | 1639 | * Generates a zstd skippable frame containing data given by src, and writes it to dst buffer. |
Yann Collet | 9866148 | 2021-01-06 15:37:46 -0800 | [diff] [blame] | 1640 | * |
Tom Rix | 2b957af | 2022-01-24 12:43:39 -0800 | [diff] [blame] | 1641 | * Skippable frames begin with a 4-byte magic number. There are 16 possible choices of magic number, |
senhuang42 | 5c41490 | 2020-12-21 11:33:41 -0500 | [diff] [blame] | 1642 | * ranging from ZSTD_MAGIC_SKIPPABLE_START to ZSTD_MAGIC_SKIPPABLE_START+15. |
senhuang42 | 7e11bd0 | 2020-12-21 11:13:22 -0500 | [diff] [blame] | 1643 | * As such, the parameter magicVariant controls the exact skippable frame magic number variant used, so |
senhuang42 | 5c41490 | 2020-12-21 11:33:41 -0500 | [diff] [blame] | 1644 | * the magic number used will be ZSTD_MAGIC_SKIPPABLE_START + magicVariant. |
Yann Collet | 9866148 | 2021-01-06 15:37:46 -0800 | [diff] [blame] | 1645 | * |
senhuang42 | 7e11bd0 | 2020-12-21 11:13:22 -0500 | [diff] [blame] | 1646 | * Returns an error if destination buffer is not large enough, if the source size is not representable |
senhuang42 | 5c41490 | 2020-12-21 11:33:41 -0500 | [diff] [blame] | 1647 | * with a 4-byte unsigned int, or if the parameter magicVariant is greater than 15 (and therefore invalid). |
Yann Collet | 9866148 | 2021-01-06 15:37:46 -0800 | [diff] [blame] | 1648 | * |
senhuang42 | 7e11bd0 | 2020-12-21 11:13:22 -0500 | [diff] [blame] | 1649 | * @return : number of bytes written or a ZSTD error. |
| 1650 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1651 | ZSTDLIB_STATIC_API size_t ZSTD_writeSkippableFrame(void* dst, size_t dstCapacity, |
senhuang42 | 5c41490 | 2020-12-21 11:33:41 -0500 | [diff] [blame] | 1652 | const void* src, size_t srcSize, unsigned magicVariant); |
senhuang42 | 7e11bd0 | 2020-12-21 11:13:22 -0500 | [diff] [blame] | 1653 | |
Binh Vo | 9d9f768 | 2021-06-11 12:11:58 -0400 | [diff] [blame] | 1654 | /*! ZSTD_readSkippableFrame() : |
| 1655 | * Retrieves a zstd skippable frame containing data given by src, and writes it to dst buffer. |
| 1656 | * |
| 1657 | * The parameter magicVariant will receive the magicVariant that was supplied when the frame was written, |
| 1658 | * i.e. magicNumber - ZSTD_MAGIC_SKIPPABLE_START. This can be NULL if the caller is not interested |
| 1659 | * in the magicVariant. |
| 1660 | * |
| 1661 | * Returns an error if destination buffer is not large enough, or if the frame is not skippable. |
| 1662 | * |
| 1663 | * @return : number of bytes written or a ZSTD error. |
| 1664 | */ |
| 1665 | ZSTDLIB_API size_t ZSTD_readSkippableFrame(void* dst, size_t dstCapacity, unsigned* magicVariant, |
| 1666 | const void* src, size_t srcSize); |
| 1667 | |
| 1668 | /*! ZSTD_isSkippableFrame() : |
| 1669 | * Tells if the content of `buffer` starts with a valid Frame Identifier for a skippable frame. |
| 1670 | */ |
| 1671 | ZSTDLIB_API unsigned ZSTD_isSkippableFrame(const void* buffer, size_t size); |
| 1672 | |
| 1673 | |
senhuang42 | 7e11bd0 | 2020-12-21 11:13:22 -0500 | [diff] [blame] | 1674 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 1675 | /*************************************** |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1676 | * Memory management |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1677 | ***************************************/ |
| 1678 | |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1679 | /*! ZSTD_estimate*() : |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 1680 | * These functions make it possible to estimate memory usage |
| 1681 | * of a future {D,C}Ctx, before its creation. |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1682 | * This is useful in combination with ZSTD_initStatic(), |
| 1683 | * which makes it possible to employ a static buffer for ZSTD_CCtx* state. |
W. Felix Handte | f7d9b36 | 2019-09-20 14:11:29 -0400 | [diff] [blame] | 1684 | * |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 1685 | * ZSTD_estimateCCtxSize() will provide a memory budget large enough |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1686 | * to compress data of any size using one-shot compression ZSTD_compressCCtx() or ZSTD_compress2() |
| 1687 | * associated with any compression level up to max specified one. |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 1688 | * The estimate will assume the input may be arbitrarily large, |
| 1689 | * which is the worst case. |
W. Felix Handte | f7d9b36 | 2019-09-20 14:11:29 -0400 | [diff] [blame] | 1690 | * |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1691 | * Note that the size estimation is specific for one-shot compression, |
| 1692 | * it is not valid for streaming (see ZSTD_estimateCStreamSize*()) |
| 1693 | * nor other potential ways of using a ZSTD_CCtx* state. |
| 1694 | * |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 1695 | * When srcSize can be bound by a known and rather "small" value, |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1696 | * this knowledge can be used to provide a tighter budget estimation |
| 1697 | * because the ZSTD_CCtx* state will need less memory for small inputs. |
| 1698 | * This tighter estimation can be provided by employing more advanced functions |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 1699 | * ZSTD_estimateCCtxSize_usingCParams(), which can be used in tandem with ZSTD_getCParams(), |
| 1700 | * and ZSTD_estimateCCtxSize_usingCCtxParams(), which can be used in tandem with ZSTD_CCtxParams_setParameter(). |
| 1701 | * Both can be used to estimate memory using custom compression parameters and arbitrary srcSize limits. |
| 1702 | * |
Elliot Gorokhovsky | 3bfd3be | 2023-01-23 13:28:36 -0500 | [diff] [blame] | 1703 | * Note : only single-threaded compression is supported. |
Yann Collet | 76e726e | 2020-05-11 19:21:50 -0700 | [diff] [blame] | 1704 | * ZSTD_estimateCCtxSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1. |
| 1705 | */ |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1706 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize(int maxCompressionLevel); |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1707 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams); |
| 1708 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params); |
| 1709 | ZSTDLIB_STATIC_API size_t ZSTD_estimateDCtxSize(void); |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1710 | |
Stella Lau | b5b9275 | 2017-08-29 10:49:29 -0700 | [diff] [blame] | 1711 | /*! ZSTD_estimateCStreamSize() : |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1712 | * ZSTD_estimateCStreamSize() will provide a memory budget large enough for streaming compression |
| 1713 | * using any compression level up to the max specified one. |
| 1714 | * It will also consider src size to be arbitrarily "large", which is a worst case scenario. |
Yann Collet | 96f0cde | 2017-09-24 16:47:02 -0700 | [diff] [blame] | 1715 | * If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation. |
| 1716 | * ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel. |
Nick Terrell | 7ad7ba3 | 2019-02-19 17:41:56 -0800 | [diff] [blame] | 1717 | * 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. |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1718 | * Note : CStream size estimation is only correct for single-threaded compression. |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1719 | * ZSTD_estimateCStreamSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1. |
| 1720 | * Note 2 : ZSTD_estimateCStreamSize* functions are not compatible with the Block-Level Sequence Producer API at this time. |
| 1721 | * Size estimates assume that no external sequence producer is registered. |
| 1722 | * |
| 1723 | * ZSTD_DStream memory budget depends on frame's window Size. |
Yann Collet | dde10b2 | 2017-06-26 17:44:26 -0700 | [diff] [blame] | 1724 | * This information can be passed manually, using ZSTD_estimateDStreamSize, |
| 1725 | * or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame(); |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1726 | * Any frame requesting a window size larger than max specified one will be rejected. |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1727 | * Note : if streaming is init with function ZSTD_init?Stream_usingDict(), |
Yann Collet | dde10b2 | 2017-06-26 17:44:26 -0700 | [diff] [blame] | 1728 | * an internal ?Dict will be created, which additional size is not estimated here. |
Yann Collet | 39ceef2 | 2023-01-30 19:06:39 -0800 | [diff] [blame] | 1729 | * In this case, get total size by adding ZSTD_estimate?DictSize |
Elliot Gorokhovsky | 3bfd3be | 2023-01-23 13:28:36 -0500 | [diff] [blame] | 1730 | */ |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1731 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize(int maxCompressionLevel); |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1732 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams); |
| 1733 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params); |
Yann Collet | 3fc14e4 | 2023-09-13 11:35:19 -0700 | [diff] [blame] | 1734 | ZSTDLIB_STATIC_API size_t ZSTD_estimateDStreamSize(size_t maxWindowSize); |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1735 | ZSTDLIB_STATIC_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize); |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1736 | |
| 1737 | /*! ZSTD_estimate?DictSize() : |
Yann Collet | 09ae03a | 2017-06-26 16:47:32 -0700 | [diff] [blame] | 1738 | * ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict(). |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1739 | * ZSTD_estimateCDictSize_advanced() makes it possible to control compression parameters precisely, like ZSTD_createCDict_advanced(). |
| 1740 | * Note : dictionaries created by reference (`ZSTD_dlm_byRef`) are logically smaller. |
Stella Lau | c88fb92 | 2017-08-29 11:55:02 -0700 | [diff] [blame] | 1741 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1742 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel); |
| 1743 | ZSTDLIB_STATIC_API size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod); |
| 1744 | ZSTDLIB_STATIC_API size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod); |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1745 | |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1746 | /*! ZSTD_initStatic*() : |
| 1747 | * Initialize an object using a pre-allocated fixed-size buffer. |
| 1748 | * workspace: The memory area to emplace the object into. |
| 1749 | * Provided pointer *must be 8-bytes aligned*. |
| 1750 | * Buffer must outlive object. |
| 1751 | * workspaceSize: Use ZSTD_estimate*Size() to determine |
| 1752 | * how large workspace must be to support target scenario. |
| 1753 | * @return : pointer to object (same address as workspace, just different type), |
| 1754 | * or NULL if error (size too small, incorrect alignment, etc.) |
| 1755 | * Note : zstd will never resize nor malloc() when using a static buffer. |
| 1756 | * If the object requires more memory than available, |
| 1757 | * zstd will just error out (typically ZSTD_error_memory_allocation). |
| 1758 | * Note 2 : there is no corresponding "free" function. |
| 1759 | * Since workspace is allocated externally, it must be freed externally too. |
| 1760 | * Note 3 : cParams : use ZSTD_getCParams() to convert a compression level |
| 1761 | * into its associated cParams. |
| 1762 | * Limitation 1 : currently not compatible with internal dictionary creation, triggered by |
| 1763 | * ZSTD_CCtx_loadDictionary(), ZSTD_initCStream_usingDict() or ZSTD_initDStream_usingDict(). |
| 1764 | * Limitation 2 : static cctx currently not compatible with multi-threading. |
| 1765 | * Limitation 3 : static dctx is incompatible with legacy support. |
| 1766 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1767 | ZSTDLIB_STATIC_API ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize); |
| 1768 | ZSTDLIB_STATIC_API ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize); /**< same as ZSTD_initStaticCCtx() */ |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1769 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1770 | ZSTDLIB_STATIC_API ZSTD_DCtx* ZSTD_initStaticDCtx(void* workspace, size_t workspaceSize); |
| 1771 | ZSTDLIB_STATIC_API ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize); /**< same as ZSTD_initStaticDCtx() */ |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1772 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1773 | ZSTDLIB_STATIC_API const ZSTD_CDict* ZSTD_initStaticCDict( |
Yann Collet | f3b8f90 | 2018-01-17 14:08:48 -0800 | [diff] [blame] | 1774 | void* workspace, size_t workspaceSize, |
| 1775 | const void* dict, size_t dictSize, |
| 1776 | ZSTD_dictLoadMethod_e dictLoadMethod, |
Yann Collet | 6873fec | 2018-03-20 15:13:14 -0700 | [diff] [blame] | 1777 | ZSTD_dictContentType_e dictContentType, |
Yann Collet | f3b8f90 | 2018-01-17 14:08:48 -0800 | [diff] [blame] | 1778 | ZSTD_compressionParameters cParams); |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1779 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1780 | ZSTDLIB_STATIC_API const ZSTD_DDict* ZSTD_initStaticDDict( |
Yann Collet | f3b8f90 | 2018-01-17 14:08:48 -0800 | [diff] [blame] | 1781 | void* workspace, size_t workspaceSize, |
| 1782 | const void* dict, size_t dictSize, |
Yann Collet | 31b54b6 | 2018-03-20 14:52:02 -0700 | [diff] [blame] | 1783 | ZSTD_dictLoadMethod_e dictLoadMethod, |
Yann Collet | 6873fec | 2018-03-20 15:13:14 -0700 | [diff] [blame] | 1784 | ZSTD_dictContentType_e dictContentType); |
Yann Collet | 3ea1563 | 2018-01-10 08:47:19 -0800 | [diff] [blame] | 1785 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1786 | |
Yann Collet | 292eeb6 | 2018-01-10 09:07:47 -0800 | [diff] [blame] | 1787 | /*! Custom memory allocation : |
| 1788 | * These prototypes make it possible to pass your own allocation/free functions. |
| 1789 | * ZSTD_customMem is provided at creation time, using ZSTD_create*_advanced() variants listed below. |
| 1790 | * All allocation/free operations will be completed using these custom variants instead of regular <stdlib.h> ones. |
| 1791 | */ |
| 1792 | typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size); |
| 1793 | typedef void (*ZSTD_freeFunction) (void* opaque, void* address); |
| 1794 | typedef struct { ZSTD_allocFunction customAlloc; ZSTD_freeFunction customFree; void* opaque; } ZSTD_customMem; |
Nick Terrell | dec7fb0 | 2020-09-23 12:58:38 -0700 | [diff] [blame] | 1795 | static |
| 1796 | #ifdef __GNUC__ |
| 1797 | __attribute__((__unused__)) |
| 1798 | #endif |
| 1799 | ZSTD_customMem const ZSTD_defaultCMem = { NULL, NULL, NULL }; /**< this constant defers to stdlib's functions */ |
Yann Collet | 292eeb6 | 2018-01-10 09:07:47 -0800 | [diff] [blame] | 1800 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1801 | ZSTDLIB_STATIC_API ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem); |
| 1802 | ZSTDLIB_STATIC_API ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem); |
| 1803 | ZSTDLIB_STATIC_API ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem); |
| 1804 | ZSTDLIB_STATIC_API ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem); |
Yann Collet | 292eeb6 | 2018-01-10 09:07:47 -0800 | [diff] [blame] | 1805 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1806 | ZSTDLIB_STATIC_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize, |
Yann Collet | 292eeb6 | 2018-01-10 09:07:47 -0800 | [diff] [blame] | 1807 | ZSTD_dictLoadMethod_e dictLoadMethod, |
Yann Collet | 6873fec | 2018-03-20 15:13:14 -0700 | [diff] [blame] | 1808 | ZSTD_dictContentType_e dictContentType, |
Yann Collet | 292eeb6 | 2018-01-10 09:07:47 -0800 | [diff] [blame] | 1809 | ZSTD_compressionParameters cParams, |
| 1810 | ZSTD_customMem customMem); |
| 1811 | |
Martin Liska | 1c2b02e | 2021-10-15 17:55:08 +0200 | [diff] [blame] | 1812 | /*! Thread pool : |
| 1813 | * These prototypes make it possible to share a thread pool among multiple compression contexts. |
| 1814 | * This can limit resources for applications with multiple threads where each one uses |
| 1815 | * a threaded compression mode (via ZSTD_c_nbWorkers parameter). |
| 1816 | * ZSTD_createThreadPool creates a new thread pool with a given number of threads. |
| 1817 | * Note that the lifetime of such pool must exist while being used. |
| 1818 | * ZSTD_CCtx_refThreadPool assigns a thread pool to a context (use NULL argument value |
| 1819 | * to use an internal thread pool). |
| 1820 | * ZSTD_freeThreadPool frees a thread pool, accepts NULL pointer. |
Martin Liska | b684900 | 2020-09-25 14:12:14 +0200 | [diff] [blame] | 1821 | */ |
| 1822 | typedef struct POOL_ctx_s ZSTD_threadPool; |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1823 | ZSTDLIB_STATIC_API ZSTD_threadPool* ZSTD_createThreadPool(size_t numThreads); |
| 1824 | ZSTDLIB_STATIC_API void ZSTD_freeThreadPool (ZSTD_threadPool* pool); /* accept NULL pointer */ |
| 1825 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_refThreadPool(ZSTD_CCtx* cctx, ZSTD_threadPool* pool); |
Martin Liska | b684900 | 2020-09-25 14:12:14 +0200 | [diff] [blame] | 1826 | |
Yann Collet | 9866148 | 2021-01-06 15:37:46 -0800 | [diff] [blame] | 1827 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 1828 | /* |
W. Felix Handte | c09454e | 2020-08-28 12:31:01 -0400 | [diff] [blame] | 1829 | * This API is temporary and is expected to change or disappear in the future! |
W. Felix Handte | ae4ebf6 | 2020-08-07 16:41:16 -0400 | [diff] [blame] | 1830 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1831 | ZSTDLIB_STATIC_API ZSTD_CDict* ZSTD_createCDict_advanced2( |
W. Felix Handte | c09454e | 2020-08-28 12:31:01 -0400 | [diff] [blame] | 1832 | const void* dict, size_t dictSize, |
| 1833 | ZSTD_dictLoadMethod_e dictLoadMethod, |
| 1834 | ZSTD_dictContentType_e dictContentType, |
W. Felix Handte | bc6521a | 2020-09-15 14:06:10 -0400 | [diff] [blame] | 1835 | const ZSTD_CCtx_params* cctxParams, |
W. Felix Handte | c09454e | 2020-08-28 12:31:01 -0400 | [diff] [blame] | 1836 | ZSTD_customMem customMem); |
Bimba Shrestha | 75b6360 | 2020-06-11 18:12:48 -0700 | [diff] [blame] | 1837 | |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1838 | ZSTDLIB_STATIC_API ZSTD_DDict* ZSTD_createDDict_advanced( |
Yann Collet | 9866148 | 2021-01-06 15:37:46 -0800 | [diff] [blame] | 1839 | const void* dict, size_t dictSize, |
| 1840 | ZSTD_dictLoadMethod_e dictLoadMethod, |
| 1841 | ZSTD_dictContentType_e dictContentType, |
| 1842 | ZSTD_customMem customMem); |
| 1843 | |
Yann Collet | 292eeb6 | 2018-01-10 09:07:47 -0800 | [diff] [blame] | 1844 | |
Yann Collet | 5a36c06 | 2017-05-09 15:11:30 -0700 | [diff] [blame] | 1845 | /*************************************** |
Yann Collet | 81e13ef | 2016-06-07 00:51:51 +0200 | [diff] [blame] | 1846 | * Advanced compression functions |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1847 | ***************************************/ |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1848 | |
Yann Collet | 1f57c2e | 2016-12-21 16:20:11 +0100 | [diff] [blame] | 1849 | /*! ZSTD_createCDict_byReference() : |
| 1850 | * Create a digested dictionary for compression |
Yann Collet | c205331 | 2018-12-05 16:23:00 -0800 | [diff] [blame] | 1851 | * Dictionary content is just referenced, not duplicated. |
| 1852 | * As a consequence, `dictBuffer` **must** outlive CDict, |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 1853 | * and its content must remain unmodified throughout the lifetime of CDict. |
| 1854 | * note: equivalent to ZSTD_createCDict_advanced(), with dictLoadMethod==ZSTD_dlm_byRef */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1855 | ZSTDLIB_STATIC_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel); |
Yann Collet | 1f57c2e | 2016-12-21 16:20:11 +0100 | [diff] [blame] | 1856 | |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1857 | /*! ZSTD_getCParams() : |
Yann Collet | f9e4f89 | 2019-01-02 12:14:36 -0800 | [diff] [blame] | 1858 | * @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize. |
| 1859 | * `estimatedSrcSize` value is optional, select 0 if not known */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1860 | ZSTDLIB_STATIC_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize); |
Yann Collet | 2b36b23 | 2016-12-13 17:59:55 +0100 | [diff] [blame] | 1861 | |
| 1862 | /*! ZSTD_getParams() : |
Yann Collet | f9e4f89 | 2019-01-02 12:14:36 -0800 | [diff] [blame] | 1863 | * same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`. |
| 1864 | * All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0 */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1865 | ZSTDLIB_STATIC_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize); |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1866 | |
Yann Collet | 3d2cd7f | 2016-06-27 15:12:26 +0200 | [diff] [blame] | 1867 | /*! ZSTD_checkCParams() : |
Yann Collet | f9e4f89 | 2019-01-02 12:14:36 -0800 | [diff] [blame] | 1868 | * Ensure param values remain within authorized range. |
| 1869 | * @return 0 on success, or an error code (can be checked with ZSTD_isError()) */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1870 | ZSTDLIB_STATIC_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params); |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1871 | |
Yann Collet | 3d2cd7f | 2016-06-27 15:12:26 +0200 | [diff] [blame] | 1872 | /*! ZSTD_adjustCParams() : |
Yann Collet | 381e66c | 2017-06-16 17:29:35 -0700 | [diff] [blame] | 1873 | * optimize params for a given `srcSize` and `dictSize`. |
Yann Collet | f9e4f89 | 2019-01-02 12:14:36 -0800 | [diff] [blame] | 1874 | * `srcSize` can be unknown, in which case use ZSTD_CONTENTSIZE_UNKNOWN. |
| 1875 | * `dictSize` must be `0` when there is no dictionary. |
| 1876 | * cPar can be invalid : all parameters will be clamped within valid range in the @return struct. |
| 1877 | * This function never fails (wide contract) */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1878 | ZSTDLIB_STATIC_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize); |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1879 | |
Yann Collet | 89342d1 | 2022-12-27 23:40:34 -0800 | [diff] [blame] | 1880 | /*! ZSTD_CCtx_setCParams() : |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 1881 | * Set all parameters provided within @p cparams into the working @p cctx. |
Yann Collet | 00c85b2 | 2022-12-28 15:08:18 -0800 | [diff] [blame] | 1882 | * Note : if modifying parameters during compression (MT mode only), |
| 1883 | * note that changes to the .windowLog parameter will be ignored. |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 1884 | * @return 0 on success, or an error code (can be checked with ZSTD_isError()). |
| 1885 | * On failure, no parameters are updated. |
| 1886 | */ |
Yann Collet | 89342d1 | 2022-12-27 23:40:34 -0800 | [diff] [blame] | 1887 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setCParams(ZSTD_CCtx* cctx, ZSTD_compressionParameters cparams); |
| 1888 | |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 1889 | /*! ZSTD_CCtx_setFParams() : |
| 1890 | * Set all parameters provided within @p fparams into the working @p cctx. |
| 1891 | * @return 0 on success, or an error code (can be checked with ZSTD_isError()). |
| 1892 | */ |
| 1893 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setFParams(ZSTD_CCtx* cctx, ZSTD_frameParameters fparams); |
| 1894 | |
| 1895 | /*! ZSTD_CCtx_setParams() : |
| 1896 | * Set all parameters provided within @p params into the working @p cctx. |
| 1897 | * @return 0 on success, or an error code (can be checked with ZSTD_isError()). |
| 1898 | */ |
| 1899 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setParams(ZSTD_CCtx* cctx, ZSTD_parameters params); |
| 1900 | |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1901 | /*! ZSTD_compress_advanced() : |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 1902 | * Note : this function is now DEPRECATED. |
| 1903 | * It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_setParameter() and other parameter setters. |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 1904 | * This prototype will generate compilation warnings. */ |
| 1905 | ZSTD_DEPRECATED("use ZSTD_compress2") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 1906 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 1907 | size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx, |
Yann Collet | 89342d1 | 2022-12-27 23:40:34 -0800 | [diff] [blame] | 1908 | void* dst, size_t dstCapacity, |
| 1909 | const void* src, size_t srcSize, |
| 1910 | const void* dict,size_t dictSize, |
| 1911 | ZSTD_parameters params); |
Yann Collet | f4bd857 | 2017-04-27 11:31:55 -0700 | [diff] [blame] | 1912 | |
| 1913 | /*! ZSTD_compress_usingCDict_advanced() : |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 1914 | * Note : this function is now DEPRECATED. |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 1915 | * It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_loadDictionary() and other parameter setters. |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 1916 | * This prototype will generate compilation warnings. */ |
| 1917 | ZSTD_DEPRECATED("use ZSTD_compress2 with ZSTD_CCtx_loadDictionary") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 1918 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 1919 | size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx, |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 1920 | void* dst, size_t dstCapacity, |
| 1921 | const void* src, size_t srcSize, |
| 1922 | const ZSTD_CDict* cdict, |
| 1923 | ZSTD_frameParameters fParams); |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 1924 | |
Yann Collet | 45c03c5 | 2016-06-14 13:46:11 +0200 | [diff] [blame] | 1925 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1926 | /*! ZSTD_CCtx_loadDictionary_byReference() : |
| 1927 | * Same as ZSTD_CCtx_loadDictionary(), but dictionary content is referenced, instead of being copied into CCtx. |
| 1928 | * It saves some memory, but also requires that `dict` outlives its usage within `cctx` */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1929 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1930 | |
| 1931 | /*! ZSTD_CCtx_loadDictionary_advanced() : |
| 1932 | * Same as ZSTD_CCtx_loadDictionary(), but gives finer control over |
| 1933 | * how to load the dictionary (by copy ? by reference ?) |
| 1934 | * and how to interpret it (automatic ? force raw mode ? full mode only ?) */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1935 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1936 | |
| 1937 | /*! ZSTD_CCtx_refPrefix_advanced() : |
| 1938 | * Same as ZSTD_CCtx_refPrefix(), but gives finer control over |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 1939 | * how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 1940 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 1941 | |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 1942 | /* === experimental parameters === */ |
| 1943 | /* these parameters can be used with ZSTD_setParameter() |
| 1944 | * they are not guaranteed to remain supported in the future */ |
| 1945 | |
| 1946 | /* Enables rsyncable mode, |
| 1947 | * which makes compressed files more rsync friendly |
| 1948 | * by adding periodic synchronization points to the compressed data. |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 1949 | * The target average block size is ZSTD_c_jobSize / 2. |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 1950 | * It's possible to modify the job size to increase or decrease |
| 1951 | * the granularity of the synchronization point. |
| 1952 | * Once the jobSize is smaller than the window size, |
| 1953 | * it will result in compression ratio degradation. |
| 1954 | * NOTE 1: rsyncable mode only works when multithreading is enabled. |
| 1955 | * NOTE 2: rsyncable performs poorly in combination with long range mode, |
| 1956 | * since it will decrease the effectiveness of synchronization points, |
| 1957 | * though mileage may vary. |
| 1958 | * NOTE 3: Rsyncable mode limits maximum compression speed to ~400 MB/s. |
| 1959 | * If the selected compression level is already running significantly slower, |
| 1960 | * the overall speed won't be significantly impacted. |
| 1961 | */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 1962 | #define ZSTD_c_rsyncable ZSTD_c_experimentalParam1 |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 1963 | |
| 1964 | /* Select a compression format. |
| 1965 | * The value must be of type ZSTD_format_e. |
| 1966 | * See ZSTD_format_e enum definition for details */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 1967 | #define ZSTD_c_format ZSTD_c_experimentalParam2 |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 1968 | |
| 1969 | /* Force back-reference distances to remain < windowSize, |
| 1970 | * even when referencing into Dictionary content (default:0) */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 1971 | #define ZSTD_c_forceMaxWindow ZSTD_c_experimentalParam3 |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 1972 | |
| 1973 | /* Controls whether the contents of a CDict |
| 1974 | * are used in place, or copied into the working context. |
| 1975 | * Accepts values from the ZSTD_dictAttachPref_e enum. |
| 1976 | * See the comments on that enum for an explanation of the feature. */ |
Yann Collet | 3583d19 | 2018-12-05 17:26:02 -0800 | [diff] [blame] | 1977 | #define ZSTD_c_forceAttachDict ZSTD_c_experimentalParam4 |
Yann Collet | d3a0c71 | 2018-11-21 16:18:55 -0800 | [diff] [blame] | 1978 | |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 1979 | /* Controlled with ZSTD_paramSwitch_e enum. |
| 1980 | * Default is ZSTD_ps_auto. |
| 1981 | * Set to ZSTD_ps_disable to never compress literals. |
| 1982 | * Set to ZSTD_ps_enable to always compress literals. (Note: uncompressed literals |
| 1983 | * may still be emitted if huffman is not beneficial to use.) |
| 1984 | * |
| 1985 | * By default, in ZSTD_ps_auto, the library will decide at runtime whether to use |
| 1986 | * literals compression based on the compression parameters - specifically, |
| 1987 | * negative compression levels do not use literal compression. |
sen | c730b8c | 2021-05-12 16:09:48 -0400 | [diff] [blame] | 1988 | */ |
| 1989 | #define ZSTD_c_literalCompressionMode ZSTD_c_experimentalParam5 |
| 1990 | |
Nick Magerko | dffbac5 | 2019-08-19 08:52:08 -0700 | [diff] [blame] | 1991 | /* User's best guess of source size. |
| 1992 | * Hint is not valid when srcSizeHint == 0. |
Nick Magerko | 09894dc | 2019-08-19 13:08:41 -0700 | [diff] [blame] | 1993 | * There is no guarantee that hint is close to actual source size, |
| 1994 | * but compression ratio may regress significantly if guess considerably underestimates */ |
Nick Magerko | dffbac5 | 2019-08-19 08:52:08 -0700 | [diff] [blame] | 1995 | #define ZSTD_c_srcSizeHint ZSTD_c_experimentalParam7 |
| 1996 | |
W. Felix Handte | ad9f98a | 2020-08-18 16:58:11 -0400 | [diff] [blame] | 1997 | /* Controls whether the new and experimental "dedicated dictionary search |
W. Felix Handte | 032010f | 2020-09-10 16:36:28 -0400 | [diff] [blame] | 1998 | * structure" can be used. This feature is still rough around the edges, be |
| 1999 | * prepared for surprising behavior! |
W. Felix Handte | ad9f98a | 2020-08-18 16:58:11 -0400 | [diff] [blame] | 2000 | * |
| 2001 | * How to use it: |
| 2002 | * |
| 2003 | * When using a CDict, whether to use this feature or not is controlled at |
| 2004 | * CDict creation, and it must be set in a CCtxParams set passed into that |
W. Felix Handte | 032010f | 2020-09-10 16:36:28 -0400 | [diff] [blame] | 2005 | * construction (via ZSTD_createCDict_advanced2()). A compression will then |
| 2006 | * use the feature or not based on how the CDict was constructed; the value of |
| 2007 | * this param, set in the CCtx, will have no effect. |
W. Felix Handte | ad9f98a | 2020-08-18 16:58:11 -0400 | [diff] [blame] | 2008 | * |
| 2009 | * However, when a dictionary buffer is passed into a CCtx, such as via |
| 2010 | * ZSTD_CCtx_loadDictionary(), this param can be set on the CCtx to control |
| 2011 | * whether the CDict that is created internally can use the feature or not. |
| 2012 | * |
| 2013 | * What it does: |
| 2014 | * |
| 2015 | * Normally, the internal data structures of the CDict are analogous to what |
| 2016 | * would be stored in a CCtx after compressing the contents of a dictionary. |
| 2017 | * To an approximation, a compression using a dictionary can then use those |
| 2018 | * data structures to simply continue what is effectively a streaming |
| 2019 | * compression where the simulated compression of the dictionary left off. |
| 2020 | * Which is to say, the search structures in the CDict are normally the same |
| 2021 | * format as in the CCtx. |
| 2022 | * |
| 2023 | * It is possible to do better, since the CDict is not like a CCtx: the search |
| 2024 | * structures are written once during CDict creation, and then are only read |
| 2025 | * after that, while the search structures in the CCtx are both read and |
| 2026 | * written as the compression goes along. This means we can choose a search |
| 2027 | * structure for the dictionary that is read-optimized. |
| 2028 | * |
W. Felix Handte | 032010f | 2020-09-10 16:36:28 -0400 | [diff] [blame] | 2029 | * This feature enables the use of that different structure. |
| 2030 | * |
| 2031 | * Note that some of the members of the ZSTD_compressionParameters struct have |
| 2032 | * different semantics and constraints in the dedicated search structure. It is |
| 2033 | * highly recommended that you simply set a compression level in the CCtxParams |
| 2034 | * you pass into the CDict creation call, and avoid messing with the cParams |
| 2035 | * directly. |
W. Felix Handte | ad9f98a | 2020-08-18 16:58:11 -0400 | [diff] [blame] | 2036 | * |
| 2037 | * Effects: |
| 2038 | * |
| 2039 | * This will only have any effect when the selected ZSTD_strategy |
| 2040 | * implementation supports this feature. Currently, that's limited to |
| 2041 | * ZSTD_greedy, ZSTD_lazy, and ZSTD_lazy2. |
| 2042 | * |
W. Felix Handte | 032010f | 2020-09-10 16:36:28 -0400 | [diff] [blame] | 2043 | * Note that this means that the CDict tables can no longer be copied into the |
| 2044 | * CCtx, so the dict attachment mode ZSTD_dictForceCopy will no longer be |
Dimitris Apostolou | ebbd675 | 2021-11-13 10:04:04 +0200 | [diff] [blame] | 2045 | * usable. The dictionary can only be attached or reloaded. |
W. Felix Handte | 032010f | 2020-09-10 16:36:28 -0400 | [diff] [blame] | 2046 | * |
| 2047 | * In general, you should expect compression to be faster--sometimes very much |
| 2048 | * so--and CDict creation to be slightly slower. Eventually, we will probably |
| 2049 | * make this mode the default. |
W. Felix Handte | c204110 | 2020-08-11 18:03:38 -0400 | [diff] [blame] | 2050 | */ |
| 2051 | #define ZSTD_c_enableDedicatedDictSearch ZSTD_c_experimentalParam8 |
| 2052 | |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 2053 | /* ZSTD_c_stableInBuffer |
| 2054 | * Experimental parameter. |
| 2055 | * Default is 0 == disabled. Set to 1 to enable. |
| 2056 | * |
Yann Collet | 37b87ad | 2022-01-01 23:15:34 -0800 | [diff] [blame] | 2057 | * Tells the compressor that input data presented with ZSTD_inBuffer |
| 2058 | * will ALWAYS be the same between calls. |
| 2059 | * Technically, the @src pointer must never be changed, |
| 2060 | * and the @pos field can only be updated by zstd. |
| 2061 | * However, it's possible to increase the @size field, |
| 2062 | * allowing scenarios where more data can be appended after compressions starts. |
| 2063 | * These conditions are checked by the compressor, |
| 2064 | * and compression will fail if they are not respected. |
| 2065 | * Also, data in the ZSTD_inBuffer within the range [src, src + pos) |
| 2066 | * MUST not be modified during compression or it will result in data corruption. |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 2067 | * |
| 2068 | * When this flag is enabled zstd won't allocate an input window buffer, |
| 2069 | * because the user guarantees it can reference the ZSTD_inBuffer until |
| 2070 | * the frame is complete. But, it will still allocate an output buffer |
| 2071 | * large enough to fit a block (see ZSTD_c_stableOutBuffer). This will also |
| 2072 | * avoid the memcpy() from the input buffer to the input window buffer. |
| 2073 | * |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 2074 | * NOTE: So long as the ZSTD_inBuffer always points to valid memory, using |
| 2075 | * this flag is ALWAYS memory safe, and will never access out-of-bounds |
Yann Collet | 37b87ad | 2022-01-01 23:15:34 -0800 | [diff] [blame] | 2076 | * memory. However, compression WILL fail if conditions are not respected. |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 2077 | * |
Yann Collet | c0c5ffa | 2021-12-31 18:20:36 -0800 | [diff] [blame] | 2078 | * WARNING: The data in the ZSTD_inBuffer in the range [src, src + pos) MUST |
Yann Collet | 27d336b | 2022-01-02 20:06:46 -0800 | [diff] [blame] | 2079 | * not be modified during compression or it will result in data corruption. |
Yann Collet | 37b87ad | 2022-01-01 23:15:34 -0800 | [diff] [blame] | 2080 | * This is because zstd needs to reference data in the ZSTD_inBuffer to find |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 2081 | * matches. Normally zstd maintains its own window buffer for this purpose, |
Yann Collet | 37b87ad | 2022-01-01 23:15:34 -0800 | [diff] [blame] | 2082 | * but passing this flag tells zstd to rely on user provided buffer instead. |
Nick Terrell | e3e0775 | 2020-10-12 13:15:39 -0700 | [diff] [blame] | 2083 | */ |
| 2084 | #define ZSTD_c_stableInBuffer ZSTD_c_experimentalParam9 |
| 2085 | |
| 2086 | /* ZSTD_c_stableOutBuffer |
| 2087 | * Experimental parameter. |
| 2088 | * Default is 0 == disabled. Set to 1 to enable. |
| 2089 | * |
| 2090 | * Tells he compressor that the ZSTD_outBuffer will not be resized between |
| 2091 | * calls. Specifically: (out.size - out.pos) will never grow. This gives the |
| 2092 | * compressor the freedom to say: If the compressed data doesn't fit in the |
| 2093 | * output buffer then return ZSTD_error_dstSizeTooSmall. This allows us to |
| 2094 | * always decompress directly into the output buffer, instead of decompressing |
| 2095 | * into an internal buffer and copying to the output buffer. |
| 2096 | * |
| 2097 | * When this flag is enabled zstd won't allocate an output buffer, because |
| 2098 | * it can write directly to the ZSTD_outBuffer. It will still allocate the |
| 2099 | * input window buffer (see ZSTD_c_stableInBuffer). |
| 2100 | * |
| 2101 | * Zstd will check that (out.size - out.pos) never grows and return an error |
| 2102 | * if it does. While not strictly necessary, this should prevent surprises. |
| 2103 | */ |
| 2104 | #define ZSTD_c_stableOutBuffer ZSTD_c_experimentalParam10 |
| 2105 | |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 2106 | /* ZSTD_c_blockDelimiters |
| 2107 | * Default is 0 == ZSTD_sf_noBlockDelimiters. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2108 | * |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 2109 | * For use with sequence compression API: ZSTD_compressSequences(). |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2110 | * |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 2111 | * Designates whether or not the given array of ZSTD_Sequence contains block delimiters |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 2112 | * and last literals, which are defined as sequences with offset == 0 and matchLength == 0. |
| 2113 | * See the definition of ZSTD_Sequence for more specifics. |
senhuang42 | 7f563b0 | 2020-11-16 10:44:57 -0500 | [diff] [blame] | 2114 | */ |
| 2115 | #define ZSTD_c_blockDelimiters ZSTD_c_experimentalParam11 |
| 2116 | |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 2117 | /* ZSTD_c_validateSequences |
| 2118 | * Default is 0 == disabled. Set to 1 to enable sequence validation. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2119 | * |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 2120 | * For use with sequence compression API: ZSTD_compressSequences(). |
| 2121 | * Designates whether or not we validate sequences provided to ZSTD_compressSequences() |
| 2122 | * during function execution. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2123 | * |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 2124 | * Without validation, providing a sequence that does not conform to the zstd spec will cause |
| 2125 | * undefined behavior, and may produce a corrupted block. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2126 | * |
Dominique Pelle | b772f53 | 2022-03-12 08:52:40 +0100 | [diff] [blame] | 2127 | * With validation enabled, if sequence is invalid (see doc/zstd_compression_format.md for |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 2128 | * specifics regarding offset/matchlength requirements) then the function will bail out and |
| 2129 | * return an error. |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2130 | * |
senhuang42 | 7742f07 | 2020-11-20 11:23:22 -0500 | [diff] [blame] | 2131 | */ |
| 2132 | #define ZSTD_c_validateSequences ZSTD_c_experimentalParam12 |
| 2133 | |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 2134 | /* ZSTD_c_useBlockSplitter |
| 2135 | * Controlled with ZSTD_paramSwitch_e enum. |
| 2136 | * Default is ZSTD_ps_auto. |
| 2137 | * Set to ZSTD_ps_disable to never use block splitter. |
| 2138 | * Set to ZSTD_ps_enable to always use block splitter. |
Nick Terrell | 09149be | 2021-04-30 15:02:12 -0700 | [diff] [blame] | 2139 | * |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 2140 | * By default, in ZSTD_ps_auto, the library will decide at runtime whether to use |
| 2141 | * block splitting based on the compression parameters. |
senhuang42 | c56d6e4 | 2020-12-07 17:11:23 -0500 | [diff] [blame] | 2142 | */ |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 2143 | #define ZSTD_c_useBlockSplitter ZSTD_c_experimentalParam13 |
senhuang42 | c56d6e4 | 2020-12-07 17:11:23 -0500 | [diff] [blame] | 2144 | |
Nick Terrell | 4694423 | 2020-11-02 17:52:29 -0800 | [diff] [blame] | 2145 | /* ZSTD_c_useRowMatchFinder |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 2146 | * Controlled with ZSTD_paramSwitch_e enum. |
| 2147 | * Default is ZSTD_ps_auto. |
| 2148 | * Set to ZSTD_ps_disable to never use row-based matchfinder. |
| 2149 | * Set to ZSTD_ps_enable to force usage of row-based matchfinder. |
Nick Terrell | 09149be | 2021-04-30 15:02:12 -0700 | [diff] [blame] | 2150 | * |
senhuang42 | b5c35d7 | 2021-09-20 09:04:07 -0400 | [diff] [blame] | 2151 | * By default, in ZSTD_ps_auto, the library will decide at runtime whether to use |
| 2152 | * the row-based matchfinder based on support for SIMD instructions and the window log. |
| 2153 | * Note that this only pertains to compression strategies: greedy, lazy, and lazy2 |
Nick Terrell | 4694423 | 2020-11-02 17:52:29 -0800 | [diff] [blame] | 2154 | */ |
| 2155 | #define ZSTD_c_useRowMatchFinder ZSTD_c_experimentalParam14 |
| 2156 | |
Nick Terrell | 172b4b6 | 2021-05-05 12:18:47 -0700 | [diff] [blame] | 2157 | /* ZSTD_c_deterministicRefPrefix |
| 2158 | * Default is 0 == disabled. Set to 1 to enable. |
| 2159 | * |
| 2160 | * Zstd produces different results for prefix compression when the prefix is |
| 2161 | * directly adjacent to the data about to be compressed vs. when it isn't. |
| 2162 | * This is because zstd detects that the two buffers are contiguous and it can |
| 2163 | * use a more efficient match finding algorithm. However, this produces different |
| 2164 | * results than when the two buffers are non-contiguous. This flag forces zstd |
| 2165 | * to always load the prefix in non-contiguous mode, even if it happens to be |
| 2166 | * adjacent to the data, to guarantee determinism. |
| 2167 | * |
| 2168 | * If you really care about determinism when using a dictionary or prefix, |
| 2169 | * like when doing delta compression, you should select this option. It comes |
| 2170 | * at a speed penalty of about ~2.5% if the dictionary and data happened to be |
| 2171 | * contiguous, and is free if they weren't contiguous. We don't expect that |
| 2172 | * intentionally making the dictionary and data contiguous will be worth the |
| 2173 | * cost to memcpy() the data. |
| 2174 | */ |
| 2175 | #define ZSTD_c_deterministicRefPrefix ZSTD_c_experimentalParam15 |
| 2176 | |
Elliot Gorokhovsky | 93b89fb | 2022-06-21 18:06:48 -0400 | [diff] [blame] | 2177 | /* ZSTD_c_prefetchCDictTables |
| 2178 | * Controlled with ZSTD_paramSwitch_e enum. Default is ZSTD_ps_auto. |
| 2179 | * |
| 2180 | * In some situations, zstd uses CDict tables in-place rather than copying them |
| 2181 | * into the working context. (See docs on ZSTD_dictAttachPref_e above for details). |
| 2182 | * In such situations, compression speed is seriously impacted when CDict tables are |
| 2183 | * "cold" (outside CPU cache). This parameter instructs zstd to prefetch CDict tables |
| 2184 | * when they are used in-place. |
| 2185 | * |
| 2186 | * For sufficiently small inputs, the cost of the prefetch will outweigh the benefit. |
| 2187 | * For sufficiently large inputs, zstd will by default memcpy() CDict tables |
| 2188 | * into the working context, so there is no need to prefetch. This parameter is |
| 2189 | * targeted at a middle range of input sizes, where a prefetch is cheap enough to be |
| 2190 | * useful but memcpy() is too expensive. The exact range of input sizes where this |
| 2191 | * makes sense is best determined by careful experimentation. |
| 2192 | * |
| 2193 | * Note: for this parameter, ZSTD_ps_auto is currently equivalent to ZSTD_ps_disable, |
| 2194 | * but in the future zstd may conditionally enable this feature via an auto-detection |
| 2195 | * heuristic for cold CDicts. |
| 2196 | * Use ZSTD_ps_disable to opt out of prefetching under any circumstances. |
| 2197 | */ |
Elliot Gorokhovsky | 2a12811 | 2022-06-21 11:59:27 -0400 | [diff] [blame] | 2198 | #define ZSTD_c_prefetchCDictTables ZSTD_c_experimentalParam16 |
| 2199 | |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2200 | /* ZSTD_c_enableSeqProducerFallback |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2201 | * Allowed values are 0 (disable) and 1 (enable). The default setting is 0. |
| 2202 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2203 | * Controls whether zstd will fall back to an internal sequence producer if an |
| 2204 | * external sequence producer is registered and returns an error code. This fallback |
| 2205 | * is block-by-block: the internal sequence producer will only be called for blocks |
| 2206 | * where the external sequence producer returns an error code. Fallback parsing will |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2207 | * follow any other cParam settings, such as compression level, the same as in a |
| 2208 | * normal (fully-internal) compression operation. |
| 2209 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2210 | * The user is strongly encouraged to read the full Block-Level Sequence Producer API |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2211 | * documentation (below) before setting this parameter. */ |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2212 | #define ZSTD_c_enableSeqProducerFallback ZSTD_c_experimentalParam17 |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2213 | |
Elliot Gorokhovsky | 7f8189c | 2023-01-31 13:51:32 -0800 | [diff] [blame] | 2214 | /* ZSTD_c_maxBlockSize |
| 2215 | * Allowed values are between 1KB and ZSTD_BLOCKSIZE_MAX (128KB). |
| 2216 | * The default is ZSTD_BLOCKSIZE_MAX, and setting to 0 will set to the default. |
Danielle Rozenblit | 908e812 | 2023-01-04 13:01:54 -0800 | [diff] [blame] | 2217 | * |
Elliot Gorokhovsky | 7f8189c | 2023-01-31 13:51:32 -0800 | [diff] [blame] | 2218 | * This parameter can be used to set an upper bound on the blocksize |
| 2219 | * that overrides the default ZSTD_BLOCKSIZE_MAX. It cannot be used to set upper |
| 2220 | * bounds greater than ZSTD_BLOCKSIZE_MAX or bounds lower than 1KB (will make |
Dimitri Papadopoulos | 547794e | 2023-02-18 10:30:23 +0100 | [diff] [blame] | 2221 | * compressBound() inaccurate). Only currently meant to be used for testing. |
Danielle Rozenblit | 908e812 | 2023-01-04 13:01:54 -0800 | [diff] [blame] | 2222 | * |
| 2223 | */ |
| 2224 | #define ZSTD_c_maxBlockSize ZSTD_c_experimentalParam18 |
| 2225 | |
Elliot Gorokhovsky | 7f8189c | 2023-01-31 13:51:32 -0800 | [diff] [blame] | 2226 | /* ZSTD_c_searchForExternalRepcodes |
| 2227 | * This parameter affects how zstd parses external sequences, such as sequences |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2228 | * provided through the compressSequences() API or from an external block-level |
| 2229 | * sequence producer. |
Elliot Gorokhovsky | 7f8189c | 2023-01-31 13:51:32 -0800 | [diff] [blame] | 2230 | * |
| 2231 | * If set to ZSTD_ps_enable, the library will check for repeated offsets in |
| 2232 | * external sequences, even if those repcodes are not explicitly indicated in |
| 2233 | * the "rep" field. Note that this is the only way to exploit repcode matches |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2234 | * while using compressSequences() or an external sequence producer, since zstd |
Elliot Gorokhovsky | 7f8189c | 2023-01-31 13:51:32 -0800 | [diff] [blame] | 2235 | * currently ignores the "rep" field of external sequences. |
| 2236 | * |
| 2237 | * If set to ZSTD_ps_disable, the library will not exploit repeated offsets in |
| 2238 | * external sequences, regardless of whether the "rep" field has been set. This |
| 2239 | * reduces sequence compression overhead by about 25% while sacrificing some |
| 2240 | * compression ratio. |
| 2241 | * |
| 2242 | * The default value is ZSTD_ps_auto, for which the library will enable/disable |
| 2243 | * based on compression level. |
| 2244 | * |
| 2245 | * Note: for now, this param only has an effect if ZSTD_c_blockDelimiters is |
| 2246 | * set to ZSTD_sf_explicitBlockDelimiters. That may change in the future. |
| 2247 | */ |
| 2248 | #define ZSTD_c_searchForExternalRepcodes ZSTD_c_experimentalParam19 |
| 2249 | |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 2250 | /*! ZSTD_CCtx_getParameter() : |
Yann Collet | 0c404a4 | 2018-12-06 10:57:19 -0800 | [diff] [blame] | 2251 | * Get the requested compression parameter value, selected by enum ZSTD_cParameter, |
| 2252 | * and store it into int* value. |
| 2253 | * @return : 0, or an error code (which can be tested with ZSTD_isError()). |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 2254 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2255 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_getParameter(const ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value); |
Yann Collet | cf9f4b6 | 2018-11-14 14:46:49 -0800 | [diff] [blame] | 2256 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2257 | |
| 2258 | /*! ZSTD_CCtx_params : |
| 2259 | * Quick howto : |
| 2260 | * - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure |
Nick Terrell | 7ad7ba3 | 2019-02-19 17:41:56 -0800 | [diff] [blame] | 2261 | * - ZSTD_CCtxParams_setParameter() : Push parameters one by one into |
| 2262 | * an existing ZSTD_CCtx_params structure. |
| 2263 | * This is similar to |
| 2264 | * ZSTD_CCtx_setParameter(). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2265 | * - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to |
| 2266 | * an existing CCtx. |
| 2267 | * These parameters will be applied to |
Yann Collet | 96d8874 | 2018-12-06 10:14:34 -0800 | [diff] [blame] | 2268 | * all subsequent frames. |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 2269 | * - ZSTD_compressStream2() : Do compression using the CCtx. |
animalize | 0933775 | 2021-03-04 11:47:43 +0800 | [diff] [blame] | 2270 | * - ZSTD_freeCCtxParams() : Free the memory, accept NULL pointer. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2271 | * |
| 2272 | * This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams() |
Yann Collet | 0c404a4 | 2018-12-06 10:57:19 -0800 | [diff] [blame] | 2273 | * for static allocation of CCtx for single-threaded compression. |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2274 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2275 | ZSTDLIB_STATIC_API ZSTD_CCtx_params* ZSTD_createCCtxParams(void); |
| 2276 | ZSTDLIB_STATIC_API size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params); /* accept NULL pointer */ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2277 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2278 | /*! ZSTD_CCtxParams_reset() : |
| 2279 | * Reset params to default values. |
| 2280 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2281 | ZSTDLIB_STATIC_API size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2282 | |
| 2283 | /*! ZSTD_CCtxParams_init() : |
| 2284 | * Initializes the compression parameters of cctxParams according to |
| 2285 | * compression level. All other parameters are reset to their default values. |
| 2286 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2287 | ZSTDLIB_STATIC_API size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2288 | |
| 2289 | /*! ZSTD_CCtxParams_init_advanced() : |
| 2290 | * Initializes the compression and frame parameters of cctxParams according to |
| 2291 | * params. All other parameters are reset to their default values. |
| 2292 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2293 | ZSTDLIB_STATIC_API size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2294 | |
Yann Collet | 8aafbd3 | 2021-05-04 09:05:22 -0700 | [diff] [blame] | 2295 | /*! ZSTD_CCtxParams_setParameter() : Requires v1.4.0+ |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2296 | * Similar to ZSTD_CCtx_setParameter. |
| 2297 | * Set one compression parameter, selected by enum ZSTD_cParameter. |
W. Felix Handte | 99746ee | 2020-08-14 14:44:08 -0400 | [diff] [blame] | 2298 | * Parameters must be applied to a ZSTD_CCtx using |
| 2299 | * ZSTD_CCtx_setParametersUsingCCtxParams(). |
| 2300 | * @result : a code representing success or failure (which can be tested with |
| 2301 | * ZSTD_isError()). |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2302 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2303 | ZSTDLIB_STATIC_API size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2304 | |
Nick Terrell | 7ad7ba3 | 2019-02-19 17:41:56 -0800 | [diff] [blame] | 2305 | /*! ZSTD_CCtxParams_getParameter() : |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2306 | * Similar to ZSTD_CCtx_getParameter. |
| 2307 | * Get the requested value of one compression parameter, selected by enum ZSTD_cParameter. |
| 2308 | * @result : 0, or an error code (which can be tested with ZSTD_isError()). |
| 2309 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2310 | ZSTDLIB_STATIC_API size_t ZSTD_CCtxParams_getParameter(const ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value); |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2311 | |
| 2312 | /*! ZSTD_CCtx_setParametersUsingCCtxParams() : |
| 2313 | * Apply a set of ZSTD_CCtx_params to the compression context. |
| 2314 | * This can be done even after compression is started, |
| 2315 | * if nbWorkers==0, this will have no impact until a new compression is started. |
| 2316 | * if nbWorkers>=1, new parameters will be picked up at next job, |
| 2317 | * with a few restrictions (windowLog, pledgedSrcSize, nbWorkers, jobSize, and overlapLog are not updated). |
| 2318 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2319 | ZSTDLIB_STATIC_API size_t ZSTD_CCtx_setParametersUsingCCtxParams( |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2320 | ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params); |
| 2321 | |
Yann Collet | d8e215c | 2018-11-30 11:16:26 -0800 | [diff] [blame] | 2322 | /*! ZSTD_compressStream2_simpleArgs() : |
| 2323 | * Same as ZSTD_compressStream2(), |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 2324 | * but using only integral types as arguments. |
| 2325 | * This variant might be helpful for binders from dynamic languages |
| 2326 | * which have troubles handling structures containing memory pointers. |
| 2327 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2328 | ZSTDLIB_STATIC_API size_t ZSTD_compressStream2_simpleArgs ( |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 2329 | ZSTD_CCtx* cctx, |
| 2330 | void* dst, size_t dstCapacity, size_t* dstPos, |
| 2331 | const void* src, size_t srcSize, size_t* srcPos, |
| 2332 | ZSTD_EndDirective endOp); |
| 2333 | |
Yann Collet | ff8d371 | 2018-11-14 12:33:57 -0800 | [diff] [blame] | 2334 | |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2335 | /*************************************** |
| 2336 | * Advanced decompression functions |
| 2337 | ***************************************/ |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 2338 | |
Yann Collet | 179b197 | 2016-11-02 17:30:49 -0700 | [diff] [blame] | 2339 | /*! ZSTD_isFrame() : |
| 2340 | * Tells if the content of `buffer` starts with a valid Frame Identifier. |
| 2341 | * Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0. |
| 2342 | * Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled. |
| 2343 | * Note 3 : Skippable Frame Identifiers are considered valid. */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2344 | ZSTDLIB_STATIC_API unsigned ZSTD_isFrame(const void* buffer, size_t size); |
Yann Collet | 179b197 | 2016-11-02 17:30:49 -0700 | [diff] [blame] | 2345 | |
Yann Collet | 4e5eea6 | 2016-12-21 16:44:35 +0100 | [diff] [blame] | 2346 | /*! ZSTD_createDDict_byReference() : |
| 2347 | * Create a digested dictionary, ready to start decompression operation without startup delay. |
Yann Collet | 57827f9 | 2017-05-25 15:44:06 -0700 | [diff] [blame] | 2348 | * Dictionary content is referenced, and therefore stays in dictBuffer. |
| 2349 | * It is important that dictBuffer outlives DDict, |
| 2350 | * it must remain read accessible throughout the lifetime of DDict */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2351 | ZSTDLIB_STATIC_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize); |
Yann Collet | 4e5eea6 | 2016-12-21 16:44:35 +0100 | [diff] [blame] | 2352 | |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2353 | /*! ZSTD_DCtx_loadDictionary_byReference() : |
| 2354 | * Same as ZSTD_DCtx_loadDictionary(), |
| 2355 | * but references `dict` content instead of copying it into `dctx`. |
| 2356 | * This saves memory if `dict` remains around., |
| 2357 | * However, it's imperative that `dict` remains accessible (and unmodified) while being used, so it must outlive decompression. */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2358 | ZSTDLIB_STATIC_API size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize); |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2359 | |
| 2360 | /*! ZSTD_DCtx_loadDictionary_advanced() : |
| 2361 | * Same as ZSTD_DCtx_loadDictionary(), |
| 2362 | * but gives direct control over |
| 2363 | * how to load the dictionary (by copy ? by reference ?) |
| 2364 | * and how to interpret it (automatic ? force raw mode ? full mode only ?). */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2365 | ZSTDLIB_STATIC_API size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType); |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2366 | |
| 2367 | /*! ZSTD_DCtx_refPrefix_advanced() : |
| 2368 | * Same as ZSTD_DCtx_refPrefix(), but gives finer control over |
| 2369 | * how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2370 | ZSTDLIB_STATIC_API size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType); |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2371 | |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 2372 | /*! ZSTD_DCtx_setMaxWindowSize() : |
| 2373 | * Refuses allocating internal buffers for frames requiring a window size larger than provided limit. |
| 2374 | * This protects a decoder context from reserving too much memory for itself (potential attack scenario). |
| 2375 | * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode. |
| 2376 | * By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT) |
| 2377 | * @return : 0, or an error code (which can be tested using ZSTD_isError()). |
| 2378 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2379 | ZSTDLIB_STATIC_API size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize); |
Yann Collet | da1f306 | 2018-11-30 15:59:50 -0800 | [diff] [blame] | 2380 | |
Nick Terrell | 2e7d174 | 2020-09-29 16:25:03 -0700 | [diff] [blame] | 2381 | /*! ZSTD_DCtx_getParameter() : |
| 2382 | * Get the requested decompression parameter value, selected by enum ZSTD_dParameter, |
| 2383 | * and store it into int* value. |
| 2384 | * @return : 0, or an error code (which can be tested with ZSTD_isError()). |
| 2385 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2386 | ZSTDLIB_STATIC_API size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int* value); |
Nick Terrell | 2e7d174 | 2020-09-29 16:25:03 -0700 | [diff] [blame] | 2387 | |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 2388 | /* ZSTD_d_format |
| 2389 | * experimental parameter, |
| 2390 | * allowing selection between ZSTD_format_e input compression formats |
| 2391 | */ |
| 2392 | #define ZSTD_d_format ZSTD_d_experimentalParam1 |
Nick Terrell | a4ff217 | 2020-04-27 17:42:03 -0700 | [diff] [blame] | 2393 | /* ZSTD_d_stableOutBuffer |
| 2394 | * Experimental parameter. |
| 2395 | * Default is 0 == disabled. Set to 1 to enable. |
| 2396 | * |
| 2397 | * Tells the decompressor that the ZSTD_outBuffer will ALWAYS be the same |
| 2398 | * between calls, except for the modifications that zstd makes to pos (the |
| 2399 | * caller must not modify pos). This is checked by the decompressor, and |
| 2400 | * decompression will fail if it ever changes. Therefore the ZSTD_outBuffer |
| 2401 | * MUST be large enough to fit the entire decompressed frame. This will be |
Nick Terrell | 55a57d4 | 2020-04-28 12:07:42 -0700 | [diff] [blame] | 2402 | * checked when the frame content size is known. The data in the ZSTD_outBuffer |
| 2403 | * in the range [dst, dst + pos) MUST not be modified during decompression |
| 2404 | * or you will get data corruption. |
Nick Terrell | a4ff217 | 2020-04-27 17:42:03 -0700 | [diff] [blame] | 2405 | * |
Dominique Pelle | b772f53 | 2022-03-12 08:52:40 +0100 | [diff] [blame] | 2406 | * When this flag is enabled zstd won't allocate an output buffer, because |
Nick Terrell | a4ff217 | 2020-04-27 17:42:03 -0700 | [diff] [blame] | 2407 | * it can write directly to the ZSTD_outBuffer, but it will still allocate |
| 2408 | * an input buffer large enough to fit any compressed block. This will also |
| 2409 | * avoid the memcpy() from the internal output buffer to the ZSTD_outBuffer. |
| 2410 | * If you need to avoid the input buffer allocation use the buffer-less |
| 2411 | * streaming API. |
| 2412 | * |
| 2413 | * NOTE: So long as the ZSTD_outBuffer always points to valid memory, using |
| 2414 | * this flag is ALWAYS memory safe, and will never access out-of-bounds |
| 2415 | * memory. However, decompression WILL fail if you violate the preconditions. |
Nick Terrell | 55a57d4 | 2020-04-28 12:07:42 -0700 | [diff] [blame] | 2416 | * |
| 2417 | * WARNING: The data in the ZSTD_outBuffer in the range [dst, dst + pos) MUST |
| 2418 | * not be modified during decompression or you will get data corruption. This |
| 2419 | * is because zstd needs to reference data in the ZSTD_outBuffer to regenerate |
| 2420 | * matches. Normally zstd maintains its own buffer for this purpose, but passing |
| 2421 | * this flag tells zstd to use the user provided buffer. |
Nick Terrell | a4ff217 | 2020-04-27 17:42:03 -0700 | [diff] [blame] | 2422 | */ |
| 2423 | #define ZSTD_d_stableOutBuffer ZSTD_d_experimentalParam2 |
Yann Collet | aec945f | 2018-12-04 15:35:37 -0800 | [diff] [blame] | 2424 | |
senhuang42 | 2f39124 | 2020-08-21 16:23:39 -0400 | [diff] [blame] | 2425 | /* ZSTD_d_forceIgnoreChecksum |
senhuang42 | b5cddda | 2020-08-21 15:59:03 -0400 | [diff] [blame] | 2426 | * Experimental parameter. |
| 2427 | * Default is 0 == disabled. Set to 1 to enable |
| 2428 | * |
senhuang42 | ffaa0df | 2020-08-22 16:58:41 -0400 | [diff] [blame] | 2429 | * Tells the decompressor to skip checksum validation during decompression, regardless |
| 2430 | * of whether checksumming was specified during compression. This offers some |
senhuang42 | b5cddda | 2020-08-21 15:59:03 -0400 | [diff] [blame] | 2431 | * slight performance benefits, and may be useful for debugging. |
senhuang42 | 44c54a3 | 2020-08-24 16:14:19 -0400 | [diff] [blame] | 2432 | * Param has values of type ZSTD_forceIgnoreChecksum_e |
senhuang42 | b5cddda | 2020-08-21 15:59:03 -0400 | [diff] [blame] | 2433 | */ |
senhuang42 | 2f39124 | 2020-08-21 16:23:39 -0400 | [diff] [blame] | 2434 | #define ZSTD_d_forceIgnoreChecksum ZSTD_d_experimentalParam3 |
senhuang42 | b5cddda | 2020-08-21 15:59:03 -0400 | [diff] [blame] | 2435 | |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 2436 | /* ZSTD_d_refMultipleDDicts |
| 2437 | * Experimental parameter. |
| 2438 | * Default is 0 == disabled. Set to 1 to enable |
| 2439 | * |
| 2440 | * If enabled and dctx is allocated on the heap, then additional memory will be allocated |
| 2441 | * to store references to multiple ZSTD_DDict. That is, multiple calls of ZSTD_refDDict() |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 2442 | * using a given ZSTD_DCtx, rather than overwriting the previous DDict reference, will instead |
| 2443 | * store all references. At decompression time, the appropriate dictID is selected |
| 2444 | * from the set of DDicts based on the dictID in the frame. |
Nick Terrell | 54a4998 | 2021-02-03 19:53:00 -0800 | [diff] [blame] | 2445 | * |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 2446 | * Usage is simply calling ZSTD_refDDict() on multiple dict buffers. |
Nick Terrell | 54a4998 | 2021-02-03 19:53:00 -0800 | [diff] [blame] | 2447 | * |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 2448 | * Param has values of byte ZSTD_refMultipleDDicts_e |
Nick Terrell | 54a4998 | 2021-02-03 19:53:00 -0800 | [diff] [blame] | 2449 | * |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 2450 | * WARNING: Enabling this parameter and calling ZSTD_DCtx_refDDict(), will trigger memory |
| 2451 | * allocation for the hash table. ZSTD_freeDCtx() also frees this memory. |
| 2452 | * Memory is allocated as per ZSTD_DCtx::customMem. |
Nick Terrell | 54a4998 | 2021-02-03 19:53:00 -0800 | [diff] [blame] | 2453 | * |
senhuang42 | 22b7bff | 2020-12-28 16:43:04 -0500 | [diff] [blame] | 2454 | * Although this function allocates memory for the table, the user is still responsible for |
| 2455 | * memory management of the underlying ZSTD_DDict* themselves. |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 2456 | */ |
| 2457 | #define ZSTD_d_refMultipleDDicts ZSTD_d_experimentalParam4 |
| 2458 | |
Nick Terrell | 8957fef | 2023-01-13 16:34:52 -0800 | [diff] [blame] | 2459 | /* ZSTD_d_disableHuffmanAssembly |
| 2460 | * Set to 1 to disable the Huffman assembly implementation. |
| 2461 | * The default value is 0, which allows zstd to use the Huffman assembly |
| 2462 | * implementation if available. |
| 2463 | * |
| 2464 | * This parameter can be used to disable Huffman assembly at runtime. |
| 2465 | * If you want to disable it at compile time you can define the macro |
| 2466 | * ZSTD_DISABLE_ASM. |
| 2467 | */ |
| 2468 | #define ZSTD_d_disableHuffmanAssembly ZSTD_d_experimentalParam5 |
| 2469 | |
Nick Terrell | 61efb2a | 2023-04-14 17:06:24 -0700 | [diff] [blame] | 2470 | /* ZSTD_d_maxBlockSize |
| 2471 | * Allowed values are between 1KB and ZSTD_BLOCKSIZE_MAX (128KB). |
| 2472 | * The default is ZSTD_BLOCKSIZE_MAX, and setting to 0 will set to the default. |
| 2473 | * |
| 2474 | * Forces the decompressor to reject blocks whose content size is |
| 2475 | * larger than the configured maxBlockSize. When maxBlockSize is |
| 2476 | * larger than the windowSize, the windowSize is used instead. |
| 2477 | * This saves memory on the decoder when you know all blocks are small. |
| 2478 | * |
| 2479 | * This option is typically used in conjunction with ZSTD_c_maxBlockSize. |
| 2480 | * |
| 2481 | * WARNING: This causes the decoder to reject otherwise valid frames |
| 2482 | * that have block sizes larger than the configured maxBlockSize. |
| 2483 | */ |
| 2484 | #define ZSTD_d_maxBlockSize ZSTD_d_experimentalParam6 |
| 2485 | |
senhuang42 | fd5b608 | 2020-12-23 16:09:31 -0500 | [diff] [blame] | 2486 | |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2487 | /*! ZSTD_DCtx_setFormat() : |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2488 | * This function is REDUNDANT. Prefer ZSTD_DCtx_setParameter(). |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2489 | * Instruct the decoder context about what kind of data to decode next. |
| 2490 | * This instruction is mandatory to decode data without a fully-formed header, |
| 2491 | * such ZSTD_f_zstd1_magicless for example. |
| 2492 | * @return : 0, or an error code (which can be tested using ZSTD_isError()). */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2493 | ZSTD_DEPRECATED("use ZSTD_DCtx_setParameter() instead") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2494 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2495 | size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format); |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2496 | |
Yann Collet | 34e146f | 2018-12-04 10:28:36 -0800 | [diff] [blame] | 2497 | /*! ZSTD_decompressStream_simpleArgs() : |
| 2498 | * Same as ZSTD_decompressStream(), |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 2499 | * but using only integral types as arguments. |
| 2500 | * This can be helpful for binders from dynamic languages |
| 2501 | * which have troubles handling structures containing memory pointers. |
| 2502 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2503 | ZSTDLIB_STATIC_API size_t ZSTD_decompressStream_simpleArgs ( |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 2504 | ZSTD_DCtx* dctx, |
| 2505 | void* dst, size_t dstCapacity, size_t* dstPos, |
| 2506 | const void* src, size_t srcSize, size_t* srcPos); |
Yann Collet | 21a42bf | 2018-11-14 16:54:54 -0800 | [diff] [blame] | 2507 | |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 2508 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 2509 | /******************************************************************** |
Przemyslaw Skibinski | 984b66c | 2016-10-24 15:59:51 +0200 | [diff] [blame] | 2510 | * Advanced streaming functions |
Yann Collet | 19e5f2a | 2018-11-19 17:38:15 -0800 | [diff] [blame] | 2511 | * Warning : most of these functions are now redundant with the Advanced API. |
| 2512 | * Once Advanced API reaches "stable" status, |
| 2513 | * redundant functions will be deprecated, and then at some point removed. |
Yann Collet | 5a0c8e2 | 2016-08-12 01:20:36 +0200 | [diff] [blame] | 2514 | ********************************************************************/ |
| 2515 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 2516 | /*===== Advanced Streaming compression functions =====*/ |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2517 | |
| 2518 | /*! ZSTD_initCStream_srcSize() : |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2519 | * This function is DEPRECATED, and equivalent to: |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2520 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
| 2521 | * ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any) |
| 2522 | * ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel); |
| 2523 | * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize); |
| 2524 | * |
| 2525 | * pledgedSrcSize must be correct. If it is not known at init time, use |
| 2526 | * ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs, |
| 2527 | * "0" also disables frame content size field. It may be enabled in the future. |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2528 | * This prototype will generate compilation warnings. |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2529 | */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2530 | ZSTD_DEPRECATED("use ZSTD_CCtx_reset, see zstd.h for detailed instructions") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2531 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2532 | size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2533 | int compressionLevel, |
| 2534 | unsigned long long pledgedSrcSize); |
| 2535 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2536 | /*! ZSTD_initCStream_usingDict() : |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2537 | * This function is DEPRECATED, and is equivalent to: |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2538 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
| 2539 | * ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel); |
| 2540 | * ZSTD_CCtx_loadDictionary(zcs, dict, dictSize); |
| 2541 | * |
| 2542 | * Creates of an internal CDict (incompatible with static CCtx), except if |
| 2543 | * dict == NULL or dictSize < 8, in which case no dict is used. |
Yann Collet | 6323966 | 2019-10-16 16:14:04 -0700 | [diff] [blame] | 2544 | * Note: dict is loaded with ZSTD_dct_auto (treated as a full zstd dictionary if |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2545 | * it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy. |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2546 | * This prototype will generate compilation warnings. |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2547 | */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2548 | ZSTD_DEPRECATED("use ZSTD_CCtx_reset, see zstd.h for detailed instructions") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2549 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2550 | size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2551 | const void* dict, size_t dictSize, |
| 2552 | int compressionLevel); |
| 2553 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2554 | /*! ZSTD_initCStream_advanced() : |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 2555 | * This function is DEPRECATED, and is equivalent to: |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2556 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 2557 | * ZSTD_CCtx_setParams(zcs, params); |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2558 | * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize); |
| 2559 | * ZSTD_CCtx_loadDictionary(zcs, dict, dictSize); |
| 2560 | * |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2561 | * dict is loaded with ZSTD_dct_auto and ZSTD_dlm_byCopy. |
| 2562 | * pledgedSrcSize must be correct. |
| 2563 | * If srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN. |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2564 | * This prototype will generate compilation warnings. |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2565 | */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2566 | ZSTD_DEPRECATED("use ZSTD_CCtx_reset, see zstd.h for detailed instructions") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2567 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2568 | size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2569 | const void* dict, size_t dictSize, |
| 2570 | ZSTD_parameters params, |
| 2571 | unsigned long long pledgedSrcSize); |
| 2572 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2573 | /*! ZSTD_initCStream_usingCDict() : |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2574 | * This function is DEPRECATED, and equivalent to: |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2575 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
| 2576 | * ZSTD_CCtx_refCDict(zcs, cdict); |
Yann Collet | 2de4217 | 2021-08-20 09:52:42 -0700 | [diff] [blame] | 2577 | * |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2578 | * note : cdict will just be referenced, and must outlive compression session |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2579 | * This prototype will generate compilation warnings. |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2580 | */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2581 | ZSTD_DEPRECATED("use ZSTD_CCtx_reset and ZSTD_CCtx_refCDict, see zstd.h for detailed instructions") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2582 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2583 | size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict); |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2584 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2585 | /*! ZSTD_initCStream_usingCDict_advanced() : |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 2586 | * This function is DEPRECATED, and is equivalent to: |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2587 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
Nick Terrell | 07a2a33 | 2023-03-07 15:15:40 -0800 | [diff] [blame] | 2588 | * ZSTD_CCtx_setFParams(zcs, fParams); |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2589 | * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize); |
| 2590 | * ZSTD_CCtx_refCDict(zcs, cdict); |
| 2591 | * |
| 2592 | * same as ZSTD_initCStream_usingCDict(), with control over frame parameters. |
| 2593 | * pledgedSrcSize must be correct. If srcSize is not known at init time, use |
| 2594 | * value ZSTD_CONTENTSIZE_UNKNOWN. |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2595 | * This prototype will generate compilation warnings. |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2596 | */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2597 | ZSTD_DEPRECATED("use ZSTD_CCtx_reset and ZSTD_CCtx_refCDict, see zstd.h for detailed instructions") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2598 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2599 | size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2600 | const ZSTD_CDict* cdict, |
| 2601 | ZSTD_frameParameters fParams, |
| 2602 | unsigned long long pledgedSrcSize); |
Yann Collet | 36c2a03 | 2017-04-05 22:06:21 -0700 | [diff] [blame] | 2603 | |
| 2604 | /*! ZSTD_resetCStream() : |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2605 | * This function is DEPRECATED, and is equivalent to: |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2606 | * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); |
| 2607 | * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize); |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2608 | * Note: ZSTD_resetCStream() interprets pledgedSrcSize == 0 as ZSTD_CONTENTSIZE_UNKNOWN, but |
| 2609 | * ZSTD_CCtx_setPledgedSrcSize() does not do the same, so ZSTD_CONTENTSIZE_UNKNOWN must be |
| 2610 | * explicitly specified. |
Nick Terrell | 20f9ff7 | 2019-03-15 18:03:31 -0700 | [diff] [blame] | 2611 | * |
Yann Collet | 96d8874 | 2018-12-06 10:14:34 -0800 | [diff] [blame] | 2612 | * start a new frame, using same parameters from previous frame. |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 2613 | * This is typically useful to skip dictionary loading stage, since it will reuse it in-place. |
Yann Collet | 36c2a03 | 2017-04-05 22:06:21 -0700 | [diff] [blame] | 2614 | * Note that zcs must be init at least once before using ZSTD_resetCStream(). |
Yann Collet | 5eed8e7 | 2017-10-14 00:32:06 -0700 | [diff] [blame] | 2615 | * If pledgedSrcSize is not known at reset time, use macro ZSTD_CONTENTSIZE_UNKNOWN. |
Yann Collet | 36c2a03 | 2017-04-05 22:06:21 -0700 | [diff] [blame] | 2616 | * If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end. |
Yann Collet | 5eed8e7 | 2017-10-14 00:32:06 -0700 | [diff] [blame] | 2617 | * For the time being, pledgedSrcSize==0 is interpreted as "srcSize unknown" for compatibility with older programs, |
Yann Collet | 394eec6 | 2018-01-17 16:39:02 -0800 | [diff] [blame] | 2618 | * but it will change to mean "empty" in future version, so use macro ZSTD_CONTENTSIZE_UNKNOWN instead. |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2619 | * @return : 0, or an error code (which can be tested using ZSTD_isError()) |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2620 | * This prototype will generate compilation warnings. |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2621 | */ |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2622 | ZSTD_DEPRECATED("use ZSTD_CCtx_reset, see zstd.h for detailed instructions") |
Fangrui Song | 5635827 | 2022-09-22 12:30:21 -0700 | [diff] [blame] | 2623 | ZSTDLIB_STATIC_API |
sen | 698f261 | 2021-05-06 17:59:32 -0400 | [diff] [blame] | 2624 | size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize); |
Yann Collet | cb32763 | 2016-08-23 00:30:31 +0200 | [diff] [blame] | 2625 | |
Yann Collet | 5a0c8e2 | 2016-08-12 01:20:36 +0200 | [diff] [blame] | 2626 | |
Yann Collet | 394eec6 | 2018-01-17 16:39:02 -0800 | [diff] [blame] | 2627 | typedef struct { |
Yann Collet | 3e4617e | 2018-08-14 11:49:25 -0700 | [diff] [blame] | 2628 | unsigned long long ingested; /* nb input bytes read and buffered */ |
| 2629 | unsigned long long consumed; /* nb input bytes actually compressed */ |
| 2630 | unsigned long long produced; /* nb of compressed bytes generated and buffered */ |
| 2631 | unsigned long long flushed; /* nb of compressed bytes flushed : not provided; can be tracked from caller side */ |
| 2632 | unsigned currentJobID; /* MT only : latest started job nb */ |
| 2633 | unsigned nbActiveWorkers; /* MT only : nb of workers actively compressing at probe time */ |
Yann Collet | 394eec6 | 2018-01-17 16:39:02 -0800 | [diff] [blame] | 2634 | } ZSTD_frameProgression; |
| 2635 | |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2636 | /* ZSTD_getFrameProgression() : |
Yann Collet | c7190c6 | 2018-01-18 11:15:23 -0800 | [diff] [blame] | 2637 | * tells how much data has been ingested (read from input) |
| 2638 | * consumed (input actually compressed) and produced (output) for current frame. |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2639 | * Note : (ingested - consumed) is amount of input data buffered internally, not yet compressed. |
| 2640 | * Aggregates progression inside active worker threads. |
Yann Collet | 394eec6 | 2018-01-17 16:39:02 -0800 | [diff] [blame] | 2641 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2642 | ZSTDLIB_STATIC_API ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx); |
Yann Collet | 105677c | 2018-08-17 18:11:54 -0700 | [diff] [blame] | 2643 | |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2644 | /*! ZSTD_toFlushNow() : |
Yann Collet | 105677c | 2018-08-17 18:11:54 -0700 | [diff] [blame] | 2645 | * Tell how many bytes are ready to be flushed immediately. |
| 2646 | * Useful for multithreading scenarios (nbWorkers >= 1). |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2647 | * Probe the oldest active job, defined as oldest job not yet entirely flushed, |
| 2648 | * and check its output buffer. |
| 2649 | * @return : amount of data stored in oldest job and ready to be flushed immediately. |
| 2650 | * if @return == 0, it means either : |
| 2651 | * + there is no active job (could be checked with ZSTD_frameProgression()), or |
| 2652 | * + oldest job is still actively compressing data, |
| 2653 | * but everything it has produced has also been flushed so far, |
Yann Collet | d7e10a7 | 2018-11-13 18:09:03 -0800 | [diff] [blame] | 2654 | * therefore flush speed is limited by production speed of oldest job |
| 2655 | * irrespective of the speed of concurrent (and newer) jobs. |
Yann Collet | 89bc309 | 2018-09-19 14:49:13 -0700 | [diff] [blame] | 2656 | */ |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2657 | ZSTDLIB_STATIC_API size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx); |
Yann Collet | 394eec6 | 2018-01-17 16:39:02 -0800 | [diff] [blame] | 2658 | |
| 2659 | |
inikep | 82057aa | 2016-10-06 13:23:52 +0200 | [diff] [blame] | 2660 | /*===== Advanced Streaming decompression functions =====*/ |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2661 | |
| 2662 | /*! |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 2663 | * This function is deprecated, and is equivalent to: |
| 2664 | * |
| 2665 | * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only); |
| 2666 | * ZSTD_DCtx_loadDictionary(zds, dict, dictSize); |
| 2667 | * |
| 2668 | * note: no dictionary will be used if dict == NULL or dictSize < 8 |
| 2669 | */ |
Elliot Gorokhovsky | 5d8cfa6 | 2023-01-13 14:51:47 -0500 | [diff] [blame] | 2670 | ZSTD_DEPRECATED("use ZSTD_DCtx_reset + ZSTD_DCtx_loadDictionary, see zstd.h for detailed instructions") |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2671 | ZSTDLIB_STATIC_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize); |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2672 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2673 | /*! |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 2674 | * This function is deprecated, and is equivalent to: |
| 2675 | * |
| 2676 | * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only); |
| 2677 | * ZSTD_DCtx_refDDict(zds, ddict); |
| 2678 | * |
| 2679 | * note : ddict is referenced, it must outlive decompression session |
| 2680 | */ |
Elliot Gorokhovsky | 5d8cfa6 | 2023-01-13 14:51:47 -0500 | [diff] [blame] | 2681 | ZSTD_DEPRECATED("use ZSTD_DCtx_reset + ZSTD_DCtx_refDDict, see zstd.h for detailed instructions") |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2682 | ZSTDLIB_STATIC_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict); |
Yann Collet | 111b0c5 | 2019-10-22 13:51:18 -0700 | [diff] [blame] | 2683 | |
Yann Collet | 9f8b180 | 2020-12-02 00:15:07 -0800 | [diff] [blame] | 2684 | /*! |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 2685 | * This function is deprecated, and is equivalent to: |
| 2686 | * |
| 2687 | * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only); |
| 2688 | * |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 2689 | * reuse decompression parameters from previous init; saves dictionary loading |
Nick Terrell | 50b9c41 | 2019-04-10 12:34:21 -0700 | [diff] [blame] | 2690 | */ |
Elliot Gorokhovsky | 5d8cfa6 | 2023-01-13 14:51:47 -0500 | [diff] [blame] | 2691 | ZSTD_DEPRECATED("use ZSTD_DCtx_reset, see zstd.h for detailed instructions") |
Stephen Kitt | e81d567 | 2021-02-17 20:24:21 +0100 | [diff] [blame] | 2692 | ZSTDLIB_STATIC_API size_t ZSTD_resetDStream(ZSTD_DStream* zds); |
Yann Collet | 5a0c8e2 | 2016-08-12 01:20:36 +0200 | [diff] [blame] | 2693 | |
| 2694 | |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2695 | /* ********************* BLOCK-LEVEL SEQUENCE PRODUCER API ********************* |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2696 | * |
| 2697 | * *** OVERVIEW *** |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2698 | * The Block-Level Sequence Producer API allows users to provide their own custom |
| 2699 | * sequence producer which libzstd invokes to process each block. The produced list |
| 2700 | * of sequences (literals and matches) is then post-processed by libzstd to produce |
| 2701 | * valid compressed blocks. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2702 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2703 | * This block-level offload API is a more granular complement of the existing |
| 2704 | * frame-level offload API compressSequences() (introduced in v1.5.1). It offers |
| 2705 | * an easier migration story for applications already integrated with libzstd: the |
| 2706 | * user application continues to invoke the same compression functions |
| 2707 | * ZSTD_compress2() or ZSTD_compressStream2() as usual, and transparently benefits |
| 2708 | * from the specific advantages of the external sequence producer. For example, |
| 2709 | * the sequence producer could be tuned to take advantage of known characteristics |
| 2710 | * of the input, to offer better speed / ratio, or could leverage hardware |
| 2711 | * acceleration not available within libzstd itself. |
| 2712 | * |
| 2713 | * See contrib/externalSequenceProducer for an example program employing the |
| 2714 | * Block-Level Sequence Producer API. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2715 | * |
| 2716 | * *** USAGE *** |
| 2717 | * The user is responsible for implementing a function of type |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2718 | * ZSTD_sequenceProducer_F. For each block, zstd will pass the following |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2719 | * arguments to the user-provided function: |
| 2720 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2721 | * - sequenceProducerState: a pointer to a user-managed state for the sequence |
| 2722 | * producer. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2723 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2724 | * - outSeqs, outSeqsCapacity: an output buffer for the sequence producer. |
| 2725 | * outSeqsCapacity is guaranteed >= ZSTD_sequenceBound(srcSize). The memory |
| 2726 | * backing outSeqs is managed by the CCtx. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2727 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2728 | * - src, srcSize: an input buffer for the sequence producer to parse. |
| 2729 | * srcSize is guaranteed to be <= ZSTD_BLOCKSIZE_MAX. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2730 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2731 | * - dict, dictSize: a history buffer, which may be empty, which the sequence |
| 2732 | * producer may reference as it parses the src buffer. Currently, zstd will |
| 2733 | * always pass dictSize == 0 into external sequence producers, but this will |
| 2734 | * change in the future. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2735 | * |
| 2736 | * - compressionLevel: a signed integer representing the zstd compression level |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2737 | * set by the user for the current operation. The sequence producer may choose |
| 2738 | * to use this information to change its compression strategy and speed/ratio |
| 2739 | * tradeoff. Note: the compression level does not reflect zstd parameters set |
| 2740 | * through the advanced API. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2741 | * |
| 2742 | * - windowSize: a size_t representing the maximum allowed offset for external |
| 2743 | * sequences. Note that sequence offsets are sometimes allowed to exceed the |
| 2744 | * windowSize if a dictionary is present, see doc/zstd_compression_format.md |
| 2745 | * for details. |
| 2746 | * |
| 2747 | * The user-provided function shall return a size_t representing the number of |
| 2748 | * sequences written to outSeqs. This return value will be treated as an error |
| 2749 | * code if it is greater than outSeqsCapacity. The return value must be non-zero |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2750 | * if srcSize is non-zero. The ZSTD_SEQUENCE_PRODUCER_ERROR macro is provided |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2751 | * for convenience, but any value greater than outSeqsCapacity will be treated as |
| 2752 | * an error code. |
| 2753 | * |
| 2754 | * If the user-provided function does not return an error code, the sequences |
| 2755 | * written to outSeqs must be a valid parse of the src buffer. Data corruption may |
| 2756 | * occur if the parse is not valid. A parse is defined to be valid if the |
| 2757 | * following conditions hold: |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2758 | * - The sum of matchLengths and literalLengths must equal srcSize. |
| 2759 | * - All sequences in the parse, except for the final sequence, must have |
| 2760 | * matchLength >= ZSTD_MINMATCH_MIN. The final sequence must have |
| 2761 | * matchLength >= ZSTD_MINMATCH_MIN or matchLength == 0. |
| 2762 | * - All offsets must respect the windowSize parameter as specified in |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2763 | * doc/zstd_compression_format.md. |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2764 | * - If the final sequence has matchLength == 0, it must also have offset == 0. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2765 | * |
| 2766 | * zstd will only validate these conditions (and fail compression if they do not |
| 2767 | * hold) if the ZSTD_c_validateSequences cParam is enabled. Note that sequence |
| 2768 | * validation has a performance cost. |
| 2769 | * |
| 2770 | * If the user-provided function returns an error, zstd will either fall back |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2771 | * to an internal sequence producer or fail the compression operation. The user can |
| 2772 | * choose between the two behaviors by setting the ZSTD_c_enableSeqProducerFallback |
| 2773 | * cParam. Fallback compression will follow any other cParam settings, such as |
| 2774 | * compression level, the same as in a normal compression operation. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2775 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2776 | * The user shall instruct zstd to use a particular ZSTD_sequenceProducer_F |
| 2777 | * function by calling |
| 2778 | * ZSTD_registerSequenceProducer(cctx, |
| 2779 | * sequenceProducerState, |
| 2780 | * sequenceProducer) |
| 2781 | * This setting will persist until the next parameter reset of the CCtx. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2782 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2783 | * The sequenceProducerState must be initialized by the user before calling |
| 2784 | * ZSTD_registerSequenceProducer(). The user is responsible for destroying the |
| 2785 | * sequenceProducerState. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2786 | * |
| 2787 | * *** LIMITATIONS *** |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2788 | * This API is compatible with all zstd compression APIs which respect advanced parameters. |
| 2789 | * However, there are three limitations: |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2790 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2791 | * First, the ZSTD_c_enableLongDistanceMatching cParam is not currently supported. |
| 2792 | * COMPRESSION WILL FAIL if it is enabled and the user tries to compress with a block-level |
| 2793 | * external sequence producer. |
| 2794 | * - Note that ZSTD_c_enableLongDistanceMatching is auto-enabled by default in some |
| 2795 | * cases (see its documentation for details). Users must explicitly set |
| 2796 | * ZSTD_c_enableLongDistanceMatching to ZSTD_ps_disable in such cases if an external |
| 2797 | * sequence producer is registered. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2798 | * - As of this writing, ZSTD_c_enableLongDistanceMatching is disabled by default |
| 2799 | * whenever ZSTD_c_windowLog < 128MB, but that's subject to change. Users should |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2800 | * check the docs on ZSTD_c_enableLongDistanceMatching whenever the Block-Level Sequence |
| 2801 | * Producer API is used in conjunction with advanced settings (like ZSTD_c_windowLog). |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2802 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2803 | * Second, history buffers are not currently supported. Concretely, zstd will always pass |
| 2804 | * dictSize == 0 to the external sequence producer (for now). This has two implications: |
| 2805 | * - Dictionaries are not currently supported. Compression will *not* fail if the user |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2806 | * references a dictionary, but the dictionary won't have any effect. |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2807 | * - Stream history is not currently supported. All advanced compression APIs, including |
| 2808 | * streaming APIs, work with external sequence producers, but each block is treated as |
| 2809 | * an independent chunk without history from previous blocks. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2810 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2811 | * Third, multi-threading within a single compression is not currently supported. In other words, |
| 2812 | * COMPRESSION WILL FAIL if ZSTD_c_nbWorkers > 0 and an external sequence producer is registered. |
Elliot Gorokhovsky | bce0382 | 2023-01-19 10:41:24 -0500 | [diff] [blame] | 2813 | * Multi-threading across compressions is fine: simply create one CCtx per thread. |
| 2814 | * |
| 2815 | * Long-term, we plan to overcome all three limitations. There is no technical blocker to |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2816 | * overcoming them. It is purely a question of engineering effort. |
| 2817 | */ |
| 2818 | |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2819 | #define ZSTD_SEQUENCE_PRODUCER_ERROR ((size_t)(-1)) |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2820 | |
Elliot Gorokhovsky | 809c7eb | 2023-11-27 06:52:36 -0800 | [diff] [blame] | 2821 | typedef size_t (*ZSTD_sequenceProducer_F) ( |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2822 | void* sequenceProducerState, |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2823 | ZSTD_Sequence* outSeqs, size_t outSeqsCapacity, |
| 2824 | const void* src, size_t srcSize, |
| 2825 | const void* dict, size_t dictSize, |
| 2826 | int compressionLevel, |
| 2827 | size_t windowSize |
| 2828 | ); |
| 2829 | |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2830 | /*! ZSTD_registerSequenceProducer() : |
| 2831 | * Instruct zstd to use a block-level external sequence producer function. |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2832 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2833 | * The sequenceProducerState must be initialized by the caller, and the caller is |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2834 | * responsible for managing its lifetime. This parameter is sticky across |
| 2835 | * compressions. It will remain set until the user explicitly resets compression |
| 2836 | * parameters. |
| 2837 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2838 | * Sequence producer registration is considered to be an "advanced parameter", |
| 2839 | * part of the "advanced API". This means it will only have an effect on compression |
| 2840 | * APIs which respect advanced parameters, such as compress2() and compressStream2(). |
| 2841 | * Older compression APIs such as compressCCtx(), which predate the introduction of |
| 2842 | * "advanced parameters", will ignore any external sequence producer setting. |
Elliot Gorokhovsky | bce0382 | 2023-01-19 10:41:24 -0500 | [diff] [blame] | 2843 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2844 | * The sequence producer can be "cleared" by registering a NULL function pointer. This |
| 2845 | * removes all limitations described above in the "LIMITATIONS" section of the API docs. |
Elliot Gorokhovsky | bce0382 | 2023-01-19 10:41:24 -0500 | [diff] [blame] | 2846 | * |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2847 | * The user is strongly encouraged to read the full API documentation (above) before |
| 2848 | * calling this function. */ |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2849 | ZSTDLIB_STATIC_API void |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2850 | ZSTD_registerSequenceProducer( |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2851 | ZSTD_CCtx* cctx, |
Elliot Gorokhovsky | ff42ed1 | 2023-02-09 17:01:17 -0500 | [diff] [blame] | 2852 | void* sequenceProducerState, |
Elliot Gorokhovsky | 809c7eb | 2023-11-27 06:52:36 -0800 | [diff] [blame] | 2853 | ZSTD_sequenceProducer_F sequenceProducer |
Elliot Gorokhovsky | 2a40262 | 2022-12-28 16:45:14 -0500 | [diff] [blame] | 2854 | ); |
| 2855 | |
Elliot Gorokhovsky | c6cabf9 | 2023-12-28 14:48:46 -0500 | [diff] [blame] | 2856 | /*! ZSTD_CCtxParams_registerSequenceProducer() : |
| 2857 | * Same as ZSTD_registerSequenceProducer(), but operates on ZSTD_CCtx_params. |
| 2858 | * This is used for accurate size estimation with ZSTD_estimateCCtxSize_usingCCtxParams(), |
| 2859 | * which is needed when creating a ZSTD_CCtx with ZSTD_initStaticCCtx(). |
| 2860 | * |
| 2861 | * If you are using the external sequence producer API in a scenario where ZSTD_initStaticCCtx() |
| 2862 | * is required, then this function is for you. Otherwise, you probably don't need it. |
| 2863 | * |
| 2864 | * See tests/zstreamtest.c for example usage. */ |
| 2865 | ZSTDLIB_STATIC_API void |
| 2866 | ZSTD_CCtxParams_registerSequenceProducer( |
| 2867 | ZSTD_CCtx_params* params, |
| 2868 | void* sequenceProducerState, |
| 2869 | ZSTD_sequenceProducer_F sequenceProducer |
| 2870 | ); |
| 2871 | |
Nick Terrell | fbd97f3 | 2023-01-24 12:21:49 -0800 | [diff] [blame] | 2872 | |
| 2873 | /********************************************************************* |
| 2874 | * Buffer-less and synchronous inner streaming functions (DEPRECATED) |
| 2875 | * |
| 2876 | * This API is deprecated, and will be removed in a future version. |
| 2877 | * It allows streaming (de)compression with user allocated buffers. |
| 2878 | * However, it is hard to use, and not as well tested as the rest of |
| 2879 | * our API. |
| 2880 | * |
| 2881 | * Please use the normal streaming API instead: ZSTD_compressStream2, |
| 2882 | * and ZSTD_decompressStream. |
| 2883 | * If there is functionality that you need, but it doesn't provide, |
| 2884 | * please open an issue on our GitHub. |
| 2885 | ********************************************************************* */ |
| 2886 | |
| 2887 | /** |
| 2888 | Buffer-less streaming compression (synchronous mode) |
| 2889 | |
| 2890 | A ZSTD_CCtx object is required to track streaming operations. |
| 2891 | Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource. |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 2892 | ZSTD_CCtx object can be reused multiple times within successive compression operations. |
Nick Terrell | fbd97f3 | 2023-01-24 12:21:49 -0800 | [diff] [blame] | 2893 | |
| 2894 | Start by initializing a context. |
| 2895 | Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression. |
| 2896 | |
| 2897 | Then, consume your input using ZSTD_compressContinue(). |
| 2898 | There are some important considerations to keep in mind when using this advanced function : |
| 2899 | - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffers only. |
| 2900 | - Interface is synchronous : input is consumed entirely and produces 1+ compressed blocks. |
| 2901 | - Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario. |
| 2902 | Worst case evaluation is provided by ZSTD_compressBound(). |
| 2903 | ZSTD_compressContinue() doesn't guarantee recover after a failed compression. |
| 2904 | - ZSTD_compressContinue() presumes prior input ***is still accessible and unmodified*** (up to maximum distance size, see WindowLog). |
| 2905 | It remembers all previous contiguous blocks, plus one separated memory segment (which can itself consists of multiple contiguous blocks) |
| 2906 | - ZSTD_compressContinue() detects that prior input has been overwritten when `src` buffer overlaps. |
| 2907 | In which case, it will "discard" the relevant memory section from its history. |
| 2908 | |
| 2909 | Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum. |
| 2910 | It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame. |
| 2911 | Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders. |
| 2912 | |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 2913 | `ZSTD_CCtx` object can be reused (ZSTD_compressBegin()) to compress again. |
Nick Terrell | fbd97f3 | 2023-01-24 12:21:49 -0800 | [diff] [blame] | 2914 | */ |
| 2915 | |
| 2916 | /*===== Buffer-less streaming compression functions =====*/ |
| 2917 | ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.") |
| 2918 | ZSTDLIB_STATIC_API size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel); |
| 2919 | ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.") |
| 2920 | ZSTDLIB_STATIC_API size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel); |
| 2921 | ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.") |
| 2922 | ZSTDLIB_STATIC_API size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); /**< note: fails if cdict==NULL */ |
| 2923 | |
| 2924 | ZSTD_DEPRECATED("This function will likely be removed in a future release. It is misleading and has very limited utility.") |
| 2925 | ZSTDLIB_STATIC_API |
| 2926 | size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); /**< note: if pledgedSrcSize is not known, use ZSTD_CONTENTSIZE_UNKNOWN */ |
| 2927 | |
| 2928 | ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.") |
| 2929 | ZSTDLIB_STATIC_API size_t ZSTD_compressContinue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); |
| 2930 | ZSTD_DEPRECATED("The buffer-less API is deprecated in favor of the normal streaming API. See docs.") |
| 2931 | ZSTDLIB_STATIC_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); |
| 2932 | |
| 2933 | /* The ZSTD_compressBegin_advanced() and ZSTD_compressBegin_usingCDict_advanced() are now DEPRECATED and will generate a compiler warning */ |
| 2934 | ZSTD_DEPRECATED("use advanced API to access custom parameters") |
| 2935 | ZSTDLIB_STATIC_API |
| 2936 | size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize); /**< pledgedSrcSize : If srcSize is not known at init time, use ZSTD_CONTENTSIZE_UNKNOWN */ |
| 2937 | ZSTD_DEPRECATED("use advanced API to access custom parameters") |
| 2938 | ZSTDLIB_STATIC_API |
| 2939 | size_t ZSTD_compressBegin_usingCDict_advanced(ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict, ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize); /* compression parameters are already set within cdict. pledgedSrcSize must be correct. If srcSize is not known, use macro ZSTD_CONTENTSIZE_UNKNOWN */ |
| 2940 | /** |
| 2941 | Buffer-less streaming decompression (synchronous mode) |
| 2942 | |
| 2943 | A ZSTD_DCtx object is required to track streaming operations. |
| 2944 | Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it. |
Dimitri Papadopoulos | fe34776 | 2023-09-23 18:56:01 +0200 | [diff] [blame] | 2945 | A ZSTD_DCtx object can be reused multiple times. |
Nick Terrell | fbd97f3 | 2023-01-24 12:21:49 -0800 | [diff] [blame] | 2946 | |
| 2947 | First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader(). |
| 2948 | Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough. |
| 2949 | Data fragment must be large enough to ensure successful decoding. |
| 2950 | `ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough. |
| 2951 | result : 0 : successful decoding, the `ZSTD_frameHeader` structure is correctly filled. |
| 2952 | >0 : `srcSize` is too small, please provide at least result bytes on next attempt. |
| 2953 | errorCode, which can be tested using ZSTD_isError(). |
| 2954 | |
| 2955 | It fills a ZSTD_frameHeader structure with important information to correctly decode the frame, |
| 2956 | such as the dictionary ID, content size, or maximum back-reference distance (`windowSize`). |
| 2957 | Note that these values could be wrong, either because of data corruption, or because a 3rd party deliberately spoofs false information. |
| 2958 | As a consequence, check that values remain within valid application range. |
| 2959 | For example, do not allocate memory blindly, check that `windowSize` is within expectation. |
| 2960 | Each application can set its own limits, depending on local restrictions. |
| 2961 | For extended interoperability, it is recommended to support `windowSize` of at least 8 MB. |
| 2962 | |
| 2963 | ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize` bytes. |
| 2964 | ZSTD_decompressContinue() is very sensitive to contiguity, |
| 2965 | if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place, |
| 2966 | or that previous contiguous segment is large enough to properly handle maximum back-reference distance. |
| 2967 | There are multiple ways to guarantee this condition. |
| 2968 | |
| 2969 | The most memory efficient way is to use a round buffer of sufficient size. |
| 2970 | Sufficient size is determined by invoking ZSTD_decodingBufferSize_min(), |
| 2971 | which can return an error code if required value is too large for current system (in 32-bits mode). |
| 2972 | In a round buffer methodology, ZSTD_decompressContinue() decompresses each block next to previous one, |
| 2973 | up to the moment there is not enough room left in the buffer to guarantee decoding another full block, |
| 2974 | which maximum size is provided in `ZSTD_frameHeader` structure, field `blockSizeMax`. |
| 2975 | At which point, decoding can resume from the beginning of the buffer. |
| 2976 | Note that already decoded data stored in the buffer should be flushed before being overwritten. |
| 2977 | |
| 2978 | There are alternatives possible, for example using two or more buffers of size `windowSize` each, though they consume more memory. |
| 2979 | |
| 2980 | Finally, if you control the compression process, you can also ignore all buffer size rules, |
| 2981 | as long as the encoder and decoder progress in "lock-step", |
| 2982 | aka use exactly the same buffer sizes, break contiguity at the same place, etc. |
| 2983 | |
| 2984 | Once buffers are setup, start decompression, with ZSTD_decompressBegin(). |
| 2985 | If decompression requires a dictionary, use ZSTD_decompressBegin_usingDict() or ZSTD_decompressBegin_usingDDict(). |
| 2986 | |
| 2987 | Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively. |
| 2988 | ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue(). |
| 2989 | ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail. |
| 2990 | |
| 2991 | result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity). |
| 2992 | It can be zero : it just means ZSTD_decompressContinue() has decoded some metadata item. |
| 2993 | It can also be an error code, which can be tested with ZSTD_isError(). |
| 2994 | |
| 2995 | A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero. |
| 2996 | Context can then be reset to start a new decompression. |
| 2997 | |
| 2998 | Note : it's possible to know if next input to present is a header or a block, using ZSTD_nextInputType(). |
| 2999 | This information is not required to properly decode a frame. |
| 3000 | |
| 3001 | == Special case : skippable frames == |
| 3002 | |
| 3003 | Skippable frames allow integration of user-defined data into a flow of concatenated frames. |
| 3004 | Skippable frames will be ignored (skipped) by decompressor. |
| 3005 | The format of skippable frames is as follows : |
| 3006 | a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F |
| 3007 | b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits |
| 3008 | c) Frame Content - any content (User Data) of length equal to Frame Size |
| 3009 | For skippable frames ZSTD_getFrameHeader() returns zfhPtr->frameType==ZSTD_skippableFrame. |
| 3010 | For skippable frames ZSTD_decompressContinue() always returns 0 : it only skips the content. |
| 3011 | */ |
| 3012 | |
| 3013 | /*===== Buffer-less streaming decompression functions =====*/ |
| 3014 | |
| 3015 | ZSTDLIB_STATIC_API size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize); /**< when frame content size is not known, pass in frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN */ |
| 3016 | |
| 3017 | ZSTDLIB_STATIC_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx); |
| 3018 | ZSTDLIB_STATIC_API size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize); |
| 3019 | ZSTDLIB_STATIC_API size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict); |
| 3020 | |
| 3021 | ZSTDLIB_STATIC_API size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx); |
| 3022 | ZSTDLIB_STATIC_API size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); |
| 3023 | |
| 3024 | /* misc */ |
| 3025 | ZSTD_DEPRECATED("This function will likely be removed in the next minor release. It is misleading and has very limited utility.") |
| 3026 | ZSTDLIB_STATIC_API void ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx); |
| 3027 | typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e; |
| 3028 | ZSTDLIB_STATIC_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx); |
| 3029 | |
| 3030 | |
| 3031 | |
| 3032 | |
| 3033 | /* ========================================= */ |
| 3034 | /** Block level API (DEPRECATED) */ |
| 3035 | /* ========================================= */ |
| 3036 | |
| 3037 | /*! |
| 3038 | |
| 3039 | This API is deprecated in favor of the regular compression API. |
| 3040 | You can get the frame header down to 2 bytes by setting: |
| 3041 | - ZSTD_c_format = ZSTD_f_zstd1_magicless |
| 3042 | - ZSTD_c_contentSizeFlag = 0 |
| 3043 | - ZSTD_c_checksumFlag = 0 |
| 3044 | - ZSTD_c_dictIDFlag = 0 |
| 3045 | |
| 3046 | This API is not as well tested as our normal API, so we recommend not using it. |
| 3047 | We will be removing it in a future version. If the normal API doesn't provide |
| 3048 | the functionality you need, please open a GitHub issue. |
| 3049 | |
| 3050 | Block functions produce and decode raw zstd blocks, without frame metadata. |
| 3051 | Frame metadata cost is typically ~12 bytes, which can be non-negligible for very small blocks (< 100 bytes). |
| 3052 | But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes. |
| 3053 | |
| 3054 | A few rules to respect : |
| 3055 | - Compressing and decompressing require a context structure |
| 3056 | + Use ZSTD_createCCtx() and ZSTD_createDCtx() |
| 3057 | - It is necessary to init context before starting |
| 3058 | + compression : any ZSTD_compressBegin*() variant, including with dictionary |
| 3059 | + decompression : any ZSTD_decompressBegin*() variant, including with dictionary |
| 3060 | - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB |
| 3061 | + If input is larger than a block size, it's necessary to split input data into multiple blocks |
| 3062 | + For inputs larger than a single block, consider using regular ZSTD_compress() instead. |
| 3063 | Frame metadata is not that costly, and quickly becomes negligible as source size grows larger than a block. |
| 3064 | - When a block is considered not compressible enough, ZSTD_compressBlock() result will be 0 (zero) ! |
| 3065 | ===> In which case, nothing is produced into `dst` ! |
| 3066 | + User __must__ test for such outcome and deal directly with uncompressed data |
| 3067 | + A block cannot be declared incompressible if ZSTD_compressBlock() return value was != 0. |
| 3068 | Doing so would mess up with statistics history, leading to potential data corruption. |
| 3069 | + ZSTD_decompressBlock() _doesn't accept uncompressed data as input_ !! |
| 3070 | + In case of multiple successive blocks, should some of them be uncompressed, |
| 3071 | decoder must be informed of their existence in order to follow proper history. |
| 3072 | Use ZSTD_insertBlock() for such a case. |
| 3073 | */ |
| 3074 | |
| 3075 | /*===== Raw zstd block functions =====*/ |
| 3076 | ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.") |
| 3077 | ZSTDLIB_STATIC_API size_t ZSTD_getBlockSize (const ZSTD_CCtx* cctx); |
| 3078 | ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.") |
| 3079 | ZSTDLIB_STATIC_API size_t ZSTD_compressBlock (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); |
| 3080 | ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.") |
| 3081 | ZSTDLIB_STATIC_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize); |
| 3082 | ZSTD_DEPRECATED("The block API is deprecated in favor of the normal compression API. See docs.") |
| 3083 | ZSTDLIB_STATIC_API size_t ZSTD_insertBlock (ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize); /**< insert uncompressed block into `dctx` history. Useful for multi-blocks decompression. */ |
| 3084 | |
Nick Terrell | 05c00f2 | 2016-11-29 11:46:37 -0800 | [diff] [blame] | 3085 | #endif /* ZSTD_H_ZSTD_STATIC_LINKING_ONLY */ |
Yann Collet | d3b7f8d | 2016-06-04 19:47:02 +0200 | [diff] [blame] | 3086 | |
Yann Collet | 4856a00 | 2015-01-24 01:58:16 +0100 | [diff] [blame] | 3087 | #if defined (__cplusplus) |
| 3088 | } |
| 3089 | #endif |