zstd 0.1.0.0 → 0.1.1.2
raw patch · 73 files changed
+22492/−9227 lines, 73 filesnew-uploader
Files
- ChangeLog +13/−0
- Codec/Compression/Zstd/Base.hs +0/−6
- Codec/Compression/Zstd/FFI.hs +1/−1
- Codec/Compression/Zstd/Internal.hs +2/−2
- Codec/Compression/Zstd/Streaming.hs +3/−2
- Codec/Compression/Zstd/Types.hs +1/−0
- README.md +2/−0
- changelog.md +0/−3
- examples/low-level/DictionaryCompression.hs +0/−62
- examples/zstd-examples.cabal +0/−23
- tests/Properties.hs +9/−3
- zstd.cabal +22/−11
- zstd/CODE_OF_CONDUCT.md +5/−0
- zstd/CONTRIBUTING.md +1/−1
- zstd/LICENSE-examples +0/−11
- zstd/NEWS +0/−236
- zstd/PATENTS +0/−33
- zstd/README.md +116/−70
- zstd/TESTING.md +44/−0
- zstd/lib/README.md +105/−34
- zstd/lib/common/bitstream.h +136/−95
- zstd/lib/common/compiler.h +140/−0
- zstd/lib/common/cpu.h +215/−0
- zstd/lib/common/debug.c +44/−0
- zstd/lib/common/debug.h +134/−0
- zstd/lib/common/entropy_common.c +29/−20
- zstd/lib/common/error_private.c +22/−11
- zstd/lib/common/error_private.h +8/−8
- zstd/lib/common/fse.h +109/−69
- zstd/lib/common/fse_decompress.c +6/−26
- zstd/lib/common/huf.h +220/−100
- zstd/lib/common/mem.h +52/−44
- zstd/lib/common/pool.c +340/−0
- zstd/lib/common/pool.h +84/−0
- zstd/lib/common/threading.c +75/−0
- zstd/lib/common/threading.h +123/−0
- zstd/lib/common/xxhash.c +34/−25
- zstd/lib/common/xxhash.h +11/−15
- zstd/lib/common/zstd_common.c +38/−28
- zstd/lib/common/zstd_errors.h +54/−35
- zstd/lib/common/zstd_internal.h +171/−122
- zstd/lib/compress/fse_compress.c +116/−245
- zstd/lib/compress/hist.c +203/−0
- zstd/lib/compress/hist.h +95/−0
- zstd/lib/compress/huf_compress.c +261/−72
- zstd/lib/compress/zstd_compress.c +4415/−3304
- zstd/lib/compress/zstd_compress_internal.h +863/−0
- zstd/lib/compress/zstd_double_fast.c +499/−0
- zstd/lib/compress/zstd_double_fast.h +38/−0
- zstd/lib/compress/zstd_fast.c +470/−0
- zstd/lib/compress/zstd_fast.h +37/−0
- zstd/lib/compress/zstd_lazy.c +1106/−0
- zstd/lib/compress/zstd_lazy.h +67/−0
- zstd/lib/compress/zstd_ldm.c +597/−0
- zstd/lib/compress/zstd_ldm.h +105/−0
- zstd/lib/compress/zstd_opt.c +1233/−0
- zstd/lib/compress/zstd_opt.h +41/−904
- zstd/lib/compress/zstdmt_compress.c +2099/−0
- zstd/lib/compress/zstdmt_compress.h +191/−0
- zstd/lib/decompress/huf_decompress.c +616/−269
- zstd/lib/decompress/zstd_ddict.c +240/−0
- zstd/lib/decompress/zstd_ddict.h +44/−0
- zstd/lib/decompress/zstd_decompress.c +1767/−2170
- zstd/lib/decompress/zstd_decompress_block.c +1311/−0
- zstd/lib/decompress/zstd_decompress_block.h +59/−0
- zstd/lib/decompress/zstd_decompress_internal.h +175/−0
- zstd/lib/dictBuilder/cover.c +307/−202
- zstd/lib/dictBuilder/cover.h +112/−0
- zstd/lib/dictBuilder/divsufsort.c +3/−3
- zstd/lib/dictBuilder/fastcover.c +740/−0
- zstd/lib/dictBuilder/zdict.c +233/−182
- zstd/lib/dictBuilder/zdict.h +178/−99
- zstd/lib/zstd.h +1902/−681
+ ChangeLog view
@@ -0,0 +1,13 @@+Version 0.1.1.2 2019-05-20 by luispedro+ * Re-release v0.1.1.1 with updated info++Version 0.1.1.1 2019-05-18 by luispedro (released as hs-zstd)+ * Add missing files to package++Version 0.1.1.0 2019-05-18 by luispedro (release as hs-zstd)+ * Fix bugs+ * Update to newer zstd version++# 0.1.0.0++* Initial release
Codec/Compression/Zstd/Base.hs view
@@ -25,7 +25,6 @@ ( -- * One-shot functions compress- , compressBound , maxCLevel , decompress , getDecompressedSize@@ -110,11 +109,6 @@ -- | The maximum compression level supported by the library. maxCLevel :: Int maxCLevel = fromIntegral FFI.maxCLevel---- | Compute the maximum compressed size of given source buffer.-compressBound :: Ptr src -- ^ Source buffer.- -> IO Int-compressBound src = fromIntegral `fmap` FFI.compressBound src -- | Decompress a buffer. The destination buffer must be already -- allocated.
Codec/Compression/Zstd/FFI.hs view
@@ -140,7 +140,7 @@ -- | Compute the maximum compressed size of given source buffer. foreign import ccall unsafe "ZSTD_compressBound"- compressBound :: Ptr src -- ^ Source buffer.+ compressBound :: CSize -- ^ Size of input. -> IO CSize foreign import ccall unsafe "ZSTD_isError"
Codec/Compression/Zstd/Internal.hs view
@@ -56,10 +56,10 @@ = bail name "unsupported compression level" | otherwise = withForeignPtr sfp $ \sp -> do- let src = sp `plusPtr` off- maxSize <- C.compressBound src+ maxSize <- C.compressBound (fromIntegral len) dfp <- B.mallocByteString (fromIntegral maxSize) withForeignPtr dfp $ \dst -> do+ let src = sp `plusPtr` off csz <- compressor dst maxSize src (fromIntegral len) (fromIntegral level) handleError csz name $ do let size = fromIntegral csz
Codec/Compression/Zstd/Streaming.hs view
@@ -95,14 +95,15 @@ check "endStream" (endStream cptr obuf) $ \leftover -> do touchForeignPtr cfp touchForeignPtr obfp- if | leftover == 0 -> do+ if | leftover <= 0 -> do -- leftover will never be <0, but compiler does not know that opos1 <- fromIntegral `fmap` peekPos obuf Done `fmap` shrink outSize dfp opos1 | leftover > 0 -> do+ opos1 <- fromIntegral `fmap` peekPos obuf dfp1 <- mallocByteString (fromIntegral leftover) poke obuf (buffer (unsafeForeignPtrToPtr dfp1) leftover) touchForeignPtr obfp- bs <- shrink outSize dfp opos+ bs <- shrink outSize dfp opos1 return (Produce bs (finish cfp obfp 0 dfp1)) type ConsumeBlock ctx io = Ptr ctx -> Ptr (Buffer Out)
Codec/Compression/Zstd/Types.hs view
@@ -1,4 +1,5 @@ -- Copyright (c) 2016-present, Facebook, Inc.+-- Copyright (c) 2019-present, Luis Pedro Coelho -- All rights reserved. -- -- This source code is licensed under the BSD-style license found in
README.md view
@@ -1,5 +1,7 @@ # Zstandard bindings for Haskell +[](https://travis-ci.org/luispedro/hs-zstd)+ This library provides Haskell bindings to the [Zstandard compression library](http://facebook.github.io/zstd/).
− changelog.md
@@ -1,3 +0,0 @@-# 0.1.0.0--* Initial release
− examples/low-level/DictionaryCompression.hs
@@ -1,62 +0,0 @@-module Main (main) where--import Codec.Compression.Zstd.Efficient-import Control.Monad (forM_)-import System.Environment (getArgs, getProgName)-import System.Exit (exitFailure)-import System.FilePath (addExtension, splitExtension)-import System.IO (hPutStrLn, stderr)-import qualified Data.ByteString as B--main :: IO ()-main = do- args <- getArgs- case args of- "-c":dictName:files -> do- dict <- (createCDict 3 . mkDict) <$> B.readFile dictName- compress dict files- "-d":dictName:files -> do- dict <- (createDDict . mkDict) <$> B.readFile dictName- decompress dict files- "-t":dictSize:dictName:files ->- case reads dictSize of- [(size,['k'])] -> train (size * 1024) dictName files- [(size,['m'])] -> train (size * 1048576) dictName files- [(size,"")] -> train size dictName files- _ -> do- program <- getProgName- hPutStrLn stderr $ "usage: " ++ program ++ " -t SIZE DICT [FILES]"- exitFailure- _ -> do- program <- getProgName- hPutStrLn stderr $ "usage: " ++ program ++ " -[cdt] ARGS"- exitFailure--compress :: CDict -> [FilePath] -> IO ()-compress dict files = do- withCCtx $ \ctx ->- forM_ files $ \file -> do- content <- compressUsingCDict ctx dict =<< B.readFile file- B.writeFile (addExtension file "zst") content--decompress :: DDict -> [FilePath] -> IO ()-decompress dict files = do- withDCtx $ \ctx ->- forM_ files $ \file -> do- result <- decompressUsingDDict ctx dict =<< B.readFile file- let newFile = case splitExtension file of- (prefix,"zst") -> prefix- _ -> addExtension file "dzst"- case result of- Skip -> hPutStrLn stderr ("NOTE: skipping " ++ file)- Error desc -> hPutStrLn stderr ("ERROR (" ++ file ++ "): " ++ desc)- Decompress bs -> B.writeFile newFile bs--train :: Int -> FilePath -> [FilePath] -> IO ()-train size dictName files = do- md <- trainFromSamples size <$> mapM B.readFile files- case md of- Left err -> do- hPutStrLn stderr ("ERROR: " ++ err)- exitFailure- Right dict -> B.writeFile dictName (fromDict dict)
− examples/zstd-examples.cabal
@@ -1,23 +0,0 @@-name: zstd-examples-version: 0-synopsis: Examples of how to use the zstd library-description: Examples of how to use the zstd library-license: OtherLicense-license-file: ../LICENSE-examples-author: Bryan O'Sullivan-maintainer: bryano@fb.com-copyright: (c) 2016-present, Facebook, Inc. All rights reserved.-build-type: Simple-category: Codec-cabal-version: >=1.10--executable low-level-dictionary-compression- hs-source-dirs: low-level- ghc-options: -Wall -fwarn-tabs- main-is: DictionaryCompression.hs- default-language: Haskell2010- build-depends:- base >= 4.8 && < 5,- bytestring,- filepath,- zstd
tests/Properties.hs view
@@ -17,19 +17,25 @@ import qualified Codec.Compression.Zstd.Lazy as L import qualified Codec.Compression.Zstd.Streaming as S import Codec.Compression.Zstd.Streaming (Result(..))-import Data.Bits (xor) import Data.ByteString (ByteString, pack, unpack)-import Data.Monoid ((<>)) import System.IO.Unsafe (unsafePerformIO) import Test.Framework (Test, testGroup) import Test.Framework.Providers.QuickCheck2 (testProperty) import Test.QuickCheck import qualified Control.Exception as E import qualified Data.ByteString as B+import qualified Data.ByteString.Char8 as B8 import qualified Data.ByteString.Lazy as L import QuickCheckUtils -Right dict = trainFromSamples 400 (replicate 100 "noooooooooo")+dict = unsafePerformIO $ do+ training0 <- B8.lines <$> B.readFile "tests/Properties.hs"+ let training = concat (replicate 20 training0)+ case trainFromSamples 1000 training of+ Left e -> do+ print e+ error e+ Right d -> return d t_rechunk cs bs = L.toStrict (rechunk cs bs) == bs
zstd.cabal view
@@ -1,29 +1,25 @@ name: zstd-version: 0.1.0.0+version: 0.1.1.2 synopsis: Haskell bindings to the Zstandard compression algorithm description: A fast lossless compression algorithm, targeting real-time compression scenarios at zlib-level and better compression ratios.-homepage: https://github.com/facebookexperimental/hs-zstd-bug-reports: https://github.com/facebookexperimental/hs-zstd/issues+homepage: https://github.com/luispedro/hs-zstd+bug-reports: https://github.com/luispedro/hs-zstd/issues license: BSD3 license-file: LICENSE author: Bryan O'Sullivan-maintainer: bryano@fb.com-copyright: (c) 2016-present, Facebook, Inc. All rights reserved.+maintainer: luis@luispedro.org+copyright: (c) 2016-present, Facebook, Inc. All rights reserved.; 2019 Luis Pedro Coelho category: Codec build-type: Simple cabal-version: >=1.10 extra-source-files: *.md+ ChangeLog PATENTS- examples/*.cabal- examples/low-level/*.hs zstd/*.md zstd/LICENSE- zstd/LICENSE-examples- zstd/NEWS- zstd/PATENTS zstd/lib/*.h zstd/lib/README.md zstd/lib/common/*.c@@ -31,6 +27,7 @@ zstd/lib/compress/*.c zstd/lib/compress/*.h zstd/lib/decompress/*.c+ zstd/lib/decompress/*.h zstd/lib/dictBuilder/*.c zstd/lib/dictBuilder/*.h @@ -55,17 +52,31 @@ if flag(standalone) c-sources:+ zstd/lib/common/debug.c zstd/lib/common/entropy_common.c zstd/lib/common/error_private.c zstd/lib/common/fse_decompress.c+ zstd/lib/common/pool.c+ zstd/lib/common/threading.c zstd/lib/common/xxhash.c zstd/lib/common/zstd_common.c zstd/lib/compress/fse_compress.c+ zstd/lib/compress/hist.c zstd/lib/compress/huf_compress.c zstd/lib/compress/zstd_compress.c+ zstd/lib/compress/zstd_double_fast.c+ zstd/lib/compress/zstd_fast.c+ zstd/lib/compress/zstd_lazy.c+ zstd/lib/compress/zstd_ldm.c+ zstd/lib/compress/zstdmt_compress.c+ zstd/lib/compress/zstd_opt.c zstd/lib/decompress/huf_decompress.c+ zstd/lib/decompress/zstd_ddict.c+ zstd/lib/decompress/zstd_decompress_block.c zstd/lib/decompress/zstd_decompress.c+ zstd/lib/dictBuilder/cover.c zstd/lib/dictBuilder/divsufsort.c+ zstd/lib/dictBuilder/fastcover.c zstd/lib/dictBuilder/zdict.c if flag(standalone)@@ -122,4 +133,4 @@ source-repository head type: git- location: https://github.com/facebookexperimental//hs-zstd+ location: https://github.com/luispedro/hs-zstd
+ zstd/CODE_OF_CONDUCT.md view
@@ -0,0 +1,5 @@+# Code of Conduct++Facebook has adopted a Code of Conduct that we expect project participants to adhere to.+Please read the [full text](https://code.fb.com/codeofconduct/)+so that you can understand what actions will and will not be tolerated.
zstd/CONTRIBUTING.md view
@@ -39,4 +39,4 @@ ## License By contributing to Zstandard, you agree that your contributions will be licensed-under the [LICENSE](LICENSE) file in the root directory of this source tree.+under both the [LICENSE](LICENSE) file and the [COPYING](COPYING) file in the root directory of this source tree.
− zstd/LICENSE-examples
@@ -1,11 +0,0 @@-Copyright (c) 2016-present, Facebook, Inc. All rights reserved.--The examples provided by Facebook are for non-commercial testing and evaluation-purposes only. Facebook reserves all rights not expressly granted.--THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL-FACEBOOK BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN-ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION-WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
− zstd/NEWS
@@ -1,236 +0,0 @@-v1.1.3-cli : new : advanced commands for detailed parameters, by Przemyslaw Skibinski-cli : fix zstdless on Mac OS-X, by Andrew Janke-API : fix : all symbols properly exposed in libzstd, by Nick Terrell-API : new : ZSTD_create?Dict_byReference(), requested by Bartosz Taudul-API : new : ZDICT_finalizeDictionary()--v1.1.2-API : streaming : decompression : changed : automatic implicit reset when chain-decoding new frames without init-API : experimental : added : dictID retrieval functions, and ZSTD_initCStream_srcSize()-API : zbuff : changed : prototypes now generate deprecation warnings-lib : improved : faster decompression speed at ultra compression settings and 32-bits mode-lib : changed : only public ZSTD_ symbols are now exposed-lib : changed : reduced usage of stack memory-lib : fixed : several corner case bugs, by Nick Terrell-cli : new : gzstd, experimental version able to decode .gz files, by Przemyslaw Skibinski-cli : new : preserve file attributes-cli : new : added zstdless and zstdgrep tools-cli : fixed : status displays total amount decoded, even for file consisting of multiple frames (like pzstd)-cli : fixed : zstdcat-zlib_wrapper : added support for gz* functions, by Przemyslaw Skibinski-install : better compatibility with FreeBSD, by Dimitry Andric-source tree : changed : zbuff source files moved to lib/deprecated--v1.1.1-New : command -M#, --memory=, --memlimit=, --memlimit-decompress= to limit allowed memory consumption-New : doc/zstd_manual.html, by Przemyslaw Skibinski-Improved : slightly better compression ratio at --ultra levels (>= 20)-Improved : better memory usage when using streaming compression API, thanks to @Rogier-5 report-Added : API : ZSTD_initCStream_usingCDict(), ZSTD_initDStream_usingDDict() (experimental section)-Added : example/multiple_streaming_compression.c-Changed : zstd_errors.h is now installed within /include (and replaces errors_public.h)-Updated man page-Fixed : zstd-small, zstd-compress and zstd-decompress compilation targets--v1.1.0-New : contrib/pzstd, parallel version of zstd, by Nick Terrell-added : NetBSD install target (#338)-Improved : speed for batches of small files-Improved : speed of zlib wrapper, by Przemyslaw Skibinski-Changed : libzstd on Windows supports legacy formats, by Christophe Chevalier-Fixed : CLI -d output to stdout by default when input is stdin (#322)-Fixed : CLI correctly detects console on Mac OS-X-Fixed : CLI supports recursive mode `-r` on Mac OS-X-Fixed : Legacy decoders use unified error codes, reported by benrg (#341), fixed by Przemyslaw Skibinski-Fixed : compatibility with OpenBSD, reported by Juan Francisco Cantero Hurtado (#319)-Fixed : compatibility with Hurd, by Przemyslaw Skibinski (#365)-Fixed : zstd-pgo, reported by octoploid (#329)--v1.0.0-Change Licensing, all project is now BSD, Copyright Facebook-Small decompression speed improvement-API : Streaming API supports legacy format-API : ZDICT_getDictID(), ZSTD_sizeof_{CCtx, DCtx, CStream, DStream}(), ZSTD_setDStreamParamter()-CLI supports legacy formats v0.4+-Fixed : compression fails on certain huge files, reported by Jesse McGrew-Enhanced documentation, by Przemyslaw Skibinski--v0.8.1-New streaming API-Changed : --ultra now enables levels beyond 19-Changed : -i# now selects benchmark time in second-Fixed : ZSTD_compress* can now compress > 4 GB in a single pass, reported by Nick Terrell-Fixed : speed regression on specific patterns (#272)-Fixed : support for Z_SYNC_FLUSH, by Dmitry Krot (#291)-Fixed : ICC compilation, by Przemyslaw Skibinski--v0.8.0-Improved : better speed on clang and gcc -O2, thanks to Eric Biggers-New : Build on FreeBSD and DragonFly, thanks to JrMarino-Changed : modified API : ZSTD_compressEnd()-Fixed : legacy mode with ZSTD_HEAPMODE=0, by Christopher Bergqvist-Fixed : premature end of frame when zero-sized raw block, reported by Eric Biggers-Fixed : large dictionaries (> 384 KB), reported by Ilona Papava-Fixed : checksum correctly checked in single-pass mode-Fixed : combined --test amd --rm, reported by Andreas M. Nilsson-Modified : minor compression level adaptations-Updated : compression format specification to v0.2.0-changed : zstd.h moved to /lib directory--v0.7.5-Transition version, supporting decoding of v0.8.x--v0.7.4-Added : homebrew for Mac, by Daniel Cade-Added : more examples-Fixed : segfault when using small dictionaries, reported by Felix Handte-Modified : default compression level for CLI is now 3-Updated : specification, to v0.1.1--v0.7.3-New : compression format specification-New : `--` separator, stating that all following arguments are file names. Suggested by Chip Turner.-New : `ZSTD_getDecompressedSize()`-New : OpenBSD target, by Juan Francisco Cantero Hurtado-New : `examples` directory-fixed : dictBuilder using HC levels, reported by Bartosz Taudul-fixed : legacy support from ZSTD_decompress_usingDDict(), reported by Felix Handte-fixed : multi-blocks decoding with intermediate uncompressed blocks, reported by Greg Slazinski-modified : removed "mem.h" and "error_public.h" dependencies from "zstd.h" (experimental section)-modified : legacy functions no longer need magic number--v0.7.2-fixed : ZSTD_decompressBlock() using multiple consecutive blocks. Reported by Greg Slazinski.-fixed : potential segfault on very large files (many gigabytes). Reported by Chip Turner.-fixed : CLI displays system error message when destination file cannot be created (#231). Reported by Chip Turner.--v0.7.1-fixed : ZBUFF_compressEnd() called multiple times with too small `dst` buffer, reported by Christophe Chevalier-fixed : dictBuilder fails if first sample is too small, reported by Руслан Ковалёв-fixed : corruption issue, reported by cj-modified : checksum enabled by default in command line mode--v0.7.0-New : Support for directory compression, using `-r`, thanks to Przemyslaw Skibinski-New : Command `--rm`, to remove source file after successful de/compression-New : Visual build scripts, by Christophe Chevalier-New : Support for Sparse File-systems (do not use space for zero-filled sectors)-New : Frame checksum support-New : Support pass-through mode (when using `-df`)-API : more efficient Dictionary API : `ZSTD_compress_usingCDict()`, `ZSTD_decompress_usingDDict()`-API : create dictionary files from custom content, by Giuseppe Ottaviano-API : support for custom malloc/free functions-New : controllable Dictionary ID-New : Support for skippable frames--v0.6.1-New : zlib wrapper API, thanks to Przemyslaw Skibinski-New : Ability to compile compressor / decompressor separately-Changed : new lib directory structure-Fixed : Legacy codec v0.5 compatible with dictionary decompression-Fixed : Decoder corruption error (#173)-Fixed : null-string roundtrip (#176)-New : benchmark mode can select directory as input-Experimental : midipix support, VMS support--v0.6.0-Stronger high compression modes, thanks to Przemyslaw Skibinski-API : ZSTD_getFrameParams() provides size of decompressed content-New : highest compression modes require `--ultra` command to fully unleash their capacity-Fixed : zstd cli return error code > 0 and removes dst file artifact when decompression fails, thanks to Chip Turner--v0.5.1-New : Optimal parsing => Very high compression modes, thanks to Przemyslaw Skibinski-Changed : Dictionary builder integrated into libzstd and zstd cli-Changed (!) : zstd cli now uses "multiple input files" as default mode. See `zstd -h`.-Fix : high compression modes for big-endian platforms-New : zstd cli : `-t` | `--test` command--v0.5.0-New : dictionary builder utility-Changed : streaming & dictionary API-Improved : better compression of small data--v0.4.7-Improved : small compression speed improvement in HC mode-Changed : `zstd_decompress.c` has ZSTD_LEGACY_SUPPORT to 0 by default-fix : bt search bug--v0.4.6-fix : fast compression mode on Windows-New : cmake configuration file, thanks to Artyom Dymchenko-Improved : high compression mode on repetitive data-New : block-level API-New : ZSTD_duplicateCCtx()--v0.4.5-new : -m/--multiple : compress/decompress multiple files--v0.4.4-Fixed : high compression modes for Windows 32 bits-new : external dictionary API extended to buffered mode and accessible through command line-new : windows DLL project, thanks to Christophe Chevalier--v0.4.3 :-new : external dictionary API-new : zstd-frugal--v0.4.2 :-Generic minor improvements for small blocks-Fixed : big-endian compatibility, by Peter Harris (#85)--v0.4.1-Fixed : ZSTD_LEGACY_SUPPORT=0 build mode (reported by Luben)-removed `zstd.c`--v0.4.0-Command line utility compatible with high compression levels-Removed zstdhc => merged into zstd-Added : ZBUFF API (see zstd_buffered.h)-Rolling buffer support--v0.3.6-small blocks params--v0.3.5-minor generic compression improvements--v0.3.4-Faster fast cLevels--v0.3.3-Small compression ratio improvement--v0.3.2-Fixed Visual Studio--v0.3.1 :-Small compression ratio improvement--v0.3-HC mode : compression levels 2-26--v0.2.2-Fix : Visual Studio 2013 & 2015 release compilation, by Christophe Chevalier--v0.2.1-Fix : Read errors, advanced fuzzer tests, by Hanno Böck--v0.2.0-**Breaking format change**-Faster decompression speed-Can still decode v0.1 format--v0.1.3-fix uninitialization warning, reported by Evan Nemerson--v0.1.2-frame concatenation support--v0.1.1-fix compression bug-detects write-flush errors--v0.1.0-first release
− zstd/PATENTS
@@ -1,33 +0,0 @@-Additional Grant of Patent Rights Version 2--"Software" means the Zstandard software distributed by Facebook, Inc.--Facebook, Inc. ("Facebook") hereby grants to each recipient of the Software-("you") a perpetual, worldwide, royalty-free, non-exclusive, irrevocable-(subject to the termination provision below) license under any Necessary-Claims, to make, have made, use, sell, offer to sell, import, and otherwise-transfer the Software. For avoidance of doubt, no license is granted under-Facebook’s rights in any patent claims that are infringed by (i) modifications-to the Software made by you or any third party or (ii) the Software in-combination with any software or other technology.--The license granted hereunder will terminate, automatically and without notice,-if you (or any of your subsidiaries, corporate affiliates or agents) initiate-directly or indirectly, or take a direct financial interest in, any Patent-Assertion: (i) against Facebook or any of its subsidiaries or corporate-affiliates, (ii) against any party if such Patent Assertion arises in whole or-in part from any software, technology, product or service of Facebook or any of-its subsidiaries or corporate affiliates, or (iii) against any party relating-to the Software. Notwithstanding the foregoing, if Facebook or any of its-subsidiaries or corporate affiliates files a lawsuit alleging patent-infringement against you in the first instance, and you respond by filing a-patent infringement counterclaim in that lawsuit against that party that is-unrelated to the Software, the license granted hereunder will not terminate-under section (i) of this paragraph due to such counterclaim.--A "Necessary Claim" is a claim of a patent owned by Facebook that is-necessarily infringed by the Software standing alone.--A "Patent Assertion" is any lawsuit or other action alleging direct, indirect,-or contributory infringement or inducement to infringe any patent, including a-cross-claim or counterclaim.
zstd/README.md view
@@ -1,124 +1,170 @@- __Zstandard__, or `zstd` as short version, is a fast lossless compression algorithm,- targeting real-time compression scenarios at zlib-level and better compression ratios.+<p align="center"><img src="https://raw.githubusercontent.com/facebook/zstd/dev/doc/images/zstd_logo86.png" alt="Zstandard"></p> -It is provided as an open-source BSD-licensed **C** library,-and a command line utility producing and decoding `.zst` compressed files.-For other programming languages,-you can consult a list of known ports on [Zstandard homepage](http://www.zstd.net/#other-languages).+__Zstandard__, or `zstd` as short version, is a fast lossless compression algorithm,+targeting real-time compression scenarios at zlib-level and better compression ratios.+It's backed by a very fast entropy stage, provided by [Huff0 and FSE library](https://github.com/Cyan4973/FiniteStateEntropy). -|Branch |Status |-|------------|---------|-|master | [](https://travis-ci.org/facebook/zstd) |-|dev | [](https://travis-ci.org/facebook/zstd) |+The project is provided as an open-source dual [BSD](LICENSE) and [GPLv2](COPYING) licensed **C** library,+and a command line utility producing and decoding `.zst`, `.gz`, `.xz` and `.lz4` files.+Should your project require another programming language,+a list of known ports and bindings is provided on [Zstandard homepage](http://www.zstd.net/#other-languages). -As a reference, several fast compression algorithms were tested and compared on a Core i7-3930K CPU @ 4.5GHz, using [lzbench], an open-source in-memory benchmark by @inikep compiled with GCC 5.4.0, with the [Silesia compression corpus].+**Development branch status:** +[![Build Status][travisDevBadge]][travisLink]+[![Build status][AppveyorDevBadge]][AppveyorLink]+[![Build status][CircleDevBadge]][CircleLink]+[![Build status][CirrusDevBadge]][CirrusLink]++[travisDevBadge]: https://travis-ci.org/facebook/zstd.svg?branch=dev "Continuous Integration test suite"+[travisLink]: https://travis-ci.org/facebook/zstd+[AppveyorDevBadge]: https://ci.appveyor.com/api/projects/status/xt38wbdxjk5mrbem/branch/dev?svg=true "Windows test suite"+[AppveyorLink]: https://ci.appveyor.com/project/YannCollet/zstd-p0yf0+[CircleDevBadge]: https://circleci.com/gh/facebook/zstd/tree/dev.svg?style=shield "Short test suite"+[CircleLink]: https://circleci.com/gh/facebook/zstd+[CirrusDevBadge]: https://api.cirrus-ci.com/github/facebook/zstd.svg?branch=dev+[CirrusLink]: https://cirrus-ci.com/github/facebook/zstd++## Benchmarks++For reference, several fast compression algorithms were tested and compared+on a server running Arch Linux (`Linux version 5.0.5-arch1-1`),+with a Core i9-9900K CPU @ 5.0GHz,+using [lzbench], an open-source in-memory benchmark by @inikep+compiled with [gcc] 8.2.1,+on the [Silesia compression corpus].+ [lzbench]: https://github.com/inikep/lzbench [Silesia compression corpus]: http://sun.aei.polsl.pl/~sdeor/index.php?page=silesia-+[gcc]: https://gcc.gnu.org/ -| Name | Ratio | C.speed | D.speed |-|-------------------------|-------|--------:|--------:|-| | | MB/s | MB/s |-| **zstd 0.8.2 -1** |**2.877**| **330** | **940** |-| [zlib] 1.2.8 deflate -1 | 2.730 | 95 | 360 |-| brotli 0.4 -0 | 2.708 | 320 | 375 |-| QuickLZ 1.5 | 2.237 | 510 | 605 |-| LZO 2.09 | 2.106 | 610 | 870 |-| [LZ4] r131 | 2.101 | 620 | 3100 |-| Snappy 1.1.3 | 2.091 | 480 | 1600 |-| LZF 3.6 | 2.077 | 375 | 790 |+| Compressor name | Ratio | Compression| Decompress.|+| --------------- | ------| -----------| ---------- |+| **zstd 1.4.0 -1** | 2.884 | 530 MB/s | 1360 MB/s |+| zlib 1.2.11 -1 | 2.743 | 110 MB/s | 440 MB/s |+| brotli 1.0.7 -0 | 2.701 | 430 MB/s | 470 MB/s |+| quicklz 1.5.0 -1 | 2.238 | 600 MB/s | 800 MB/s |+| lzo1x 2.09 -1 | 2.106 | 680 MB/s | 950 MB/s |+| lz4 1.8.3 | 2.101 | 800 MB/s | 4220 MB/s |+| snappy 1.1.4 | 2.073 | 580 MB/s | 2020 MB/s |+| lzf 3.6 -1 | 2.077 | 440 MB/s | 930 MB/s | -[zlib]:http://www.zlib.net/+[zlib]: http://www.zlib.net/ [LZ4]: http://www.lz4.org/ Zstd can also offer stronger compression ratios at the cost of compression speed.-Speed vs Compression trade-off is configurable by small increments. Decompression speed is preserved and remains roughly the same at all settings, a property shared by most LZ compression algorithms, such as [zlib] or lzma.+Speed vs Compression trade-off is configurable by small increments.+Decompression speed is preserved and remains roughly the same at all settings,+a property shared by most LZ compression algorithms, such as [zlib] or lzma. -The following tests were run on a Core i7-3930K CPU @ 4.5GHz, using [lzbench], an open-source in-memory benchmark by @inikep compiled with GCC 5.2.1, on the [Silesia compression corpus].+The following tests were run+on a server running Linux Debian (`Linux version 4.14.0-3-amd64`)+with a Core i7-6700K CPU @ 4.0GHz,+using [lzbench], an open-source in-memory benchmark by @inikep+compiled with [gcc] 7.3.0,+on the [Silesia compression corpus]. Compression Speed vs Ratio | Decompression Speed ---------------------------|--------------------- | + |  -Several algorithms can produce higher compression ratios, but at slower speeds, falling outside of the graph.-For a larger picture including very slow modes, [click on this link](doc/images/DCspeed5.png) .+A few other algorithms can produce higher compression ratios at slower speeds, falling outside of the graph.+For a larger picture including slow modes, [click on this link](doc/images/DCspeed5.png). -### The case for Small Data compression+## The case for Small Data compression -Previous charts provide results applicable to typical file and stream scenarios (several MB). Small data comes with different perspectives. The smaller the amount of data to compress, the more difficult it is to achieve any significant compression.+Previous charts provide results applicable to typical file and stream scenarios (several MB). Small data comes with different perspectives. -This problem is common to many compression algorithms. The reason is, compression algorithms learn from past data how to compress future data. But at the beginning of a new file, there is no "past" to build upon.+The smaller the amount of data to compress, the more difficult it is to compress. This problem is common to all compression algorithms, and reason is, compression algorithms learn from past data how to compress future data. But at the beginning of a new data set, there is no "past" to build upon. -To solve this situation, Zstd offers a __training mode__, which can be used to tune the algorithm for a selected type of data, by providing it with a few samples. The result of the training is stored in a file called "dictionary", which can be loaded before compression and decompression. Using this dictionary, the compression ratio achievable on small data improves dramatically:+To solve this situation, Zstd offers a __training mode__, which can be used to tune the algorithm for a selected type of data.+Training Zstandard is achieved by providing it with a few samples (one file per sample). The result of this training is stored in a file called "dictionary", which must be loaded before compression and decompression.+Using this dictionary, the compression ratio achievable on small data improves dramatically. -+The following example uses the `github-users` [sample set](https://github.com/facebook/zstd/releases/tag/v1.1.3), created from [github public API](https://developer.github.com/v3/users/#get-all-users).+It consists of roughly 10K records weighing about 1KB each. -These compression gains are achieved while simultaneously providing faster compression and decompression speeds.+Compression Ratio | Compression Speed | Decompression Speed+------------------|-------------------|--------------------+ |  |  -Dictionary works if there is some correlation in a family of small data (there is no _universal dictionary_).-Hence, deploying one dictionary per type of data will provide the greatest benefits. Dictionary gains are mostly effective in the first few KB. Then, the compression algorithm will rely more and more on previously decoded content to compress the rest of the file. -#### Dictionary compression How To :+These compression gains are achieved while simultaneously providing _faster_ compression and decompression speeds. -1) Create the dictionary+Training works if there is some correlation in a family of small data samples. The more data-specific a dictionary is, the more efficient it is (there is no _universal dictionary_).+Hence, deploying one dictionary per type of data will provide the greatest benefits.+Dictionary gains are mostly effective in the first few KB. Then, the compression algorithm will gradually use previously decoded content to better compress the rest of the file. -`zstd --train FullPathToTrainingSet/* -o dictionaryName`+### Dictionary compression How To: -2) Compress with dictionary+1. Create the dictionary -`zstd FILE -D dictionaryName`+ `zstd --train FullPathToTrainingSet/* -o dictionaryName` -3) Decompress with dictionary+2. Compress with dictionary -`zstd --decompress FILE.zst -D dictionaryName`+ `zstd -D dictionaryName FILE` -### Build+3. Decompress with dictionary -Once you have the repository cloned, there are multiple ways provided to build Zstandard.+ `zstd -D dictionaryName --decompress FILE.zst` -#### Makefile -If your system is compatible with a standard `make` (or `gmake`) binary generator,-you can simply run it at the root directory.-It will generate `zstd` within root directory.+## Build instructions -Other available options include :-- `make install` : create and install zstd binary, library and man page-- `make test` : create and run `zstd` and test tools on local platform+### Makefile -#### cmake+If your system is compatible with standard `make` (or `gmake`),+invoking `make` in root directory will generate `zstd` cli in root directory. +Other available options include:+- `make install` : create and install zstd cli, library and man pages+- `make check` : create and run `zstd`, tests its behavior on local platform++### cmake+ A `cmake` project generator is provided within `build/cmake`. It can generate Makefiles or other build scripts to create `zstd` binary, and `libzstd` dynamic and static libraries. -#### Visual (Windows)+By default, `CMAKE_BUILD_TYPE` is set to `Release`. -Going into `build` directory, you will find additional possibilities :-- Projects for Visual Studio 2005, 2008 and 2010- + VS2010 project is compatible with VS2012, VS2013 and VS2015-- Automated build scripts for Visual compiler by @KrzysFR , in `build/VS_scripts`,+### Meson++A Meson project is provided within [`build/meson`](build/meson). Follow+build instructions in that directory.++You can also take a look at [`.travis.yml`](.travis.yml) file for an+example about how Meson is used to build this project.++Note that default build type is **release**.++### Visual Studio (Windows)++Going into `build` directory, you will find additional possibilities:+- Projects for Visual Studio 2005, 2008 and 2010.+ + VS2010 project is compatible with VS2012, VS2013, VS2015 and VS2017.+- Automated build scripts for Visual compiler by [@KrzysFR](https://github.com/KrzysFR), in `build/VS_scripts`, which will build `zstd` cli and `libzstd` library without any need to open Visual Studio solution. +### Buck -### Status+You can build the zstd binary via buck by executing: `buck build programs:zstd` from the root of the repo.+The output binary will be in `buck-out/gen/programs/`. -Zstandard is currently deployed within Facebook. It is used daily to compress and decompress very large amounts of data in multiple formats and use cases.+## Status++Zstandard is currently deployed within Facebook. It is used continuously to compress large amounts of data in multiple formats and use cases. Zstandard is considered safe for production environments. -### License+## License -Zstandard is [BSD-licensed](LICENSE). We also provide an [additional patent grant](PATENTS).+Zstandard is dual-licensed under [BSD](LICENSE) and [GPLv2](COPYING). -### Contributing+## Contributing -The "dev" branch is the one where all contributions will be merged before reaching "master".-If you plan to propose a patch, please commit into the "dev" branch or its own feature branch.+The "dev" branch is the one where all contributions are merged before reaching "master".+If you plan to propose a patch, please commit into the "dev" branch, or its own feature branch. Direct commit to "master" are not permitted. For more information, please read [CONTRIBUTING](CONTRIBUTING.md).--### Miscellaneous--Zstd entropy stage is provided by [Huff0 and FSE, from Finite State Entropy library](https://github.com/Cyan4973/FiniteStateEntropy).
+ zstd/TESTING.md view
@@ -0,0 +1,44 @@+Testing+=======++Zstandard CI testing is split up into three sections:+short, medium, and long tests.++Short Tests+-----------+Short tests run on CircleCI for new commits on every branch and pull request.+They consist of the following tests:+- Compilation on all supported targets (x86, x86_64, ARM, AArch64, PowerPC, and PowerPC64)+- Compilation on various versions of gcc, clang, and g+++- `tests/playTests.sh` on x86_64, without the tests on long data (CLI tests)+- Small tests (`tests/legacy.c`, `tests/longmatch.c`, `tests/symbols.c`) on x64_64++Medium Tests+------------+Medium tests run on every commit and pull request to `dev` branch, on TravisCI.+They consist of the following tests:+- The following tests run with UBsan and Asan on x86_64 and x86, as well as with+ Msan on x86_64+ - `tests/playTests.sh --test-long-data`+ - Fuzzer tests: `tests/fuzzer.c`, `tests/zstreamtest.c`, and `tests/decodecorpus.c`+- `tests/zstreamtest.c` under Tsan (streaming mode, including multithreaded mode)+- Valgrind Test (`make -C tests valgrindTest`) (testing CLI and fuzzer under valgrind)+- Fuzzer tests (see above) on ARM, AArch64, PowerPC, and PowerPC64++Long Tests+----------+Long tests run on all commits to `master` branch,+and once a day on the current version of `dev` branch,+on TravisCI.+They consist of the following tests:+- Entire test suite (including fuzzers and some other specialized tests) on:+ - x86_64 and x86 with UBsan and Asan+ - x86_64 with Msan+ - ARM, AArch64, PowerPC, and PowerPC64+- Streaming mode fuzzer with Tsan (for the `zstdmt` testing)+- ZlibWrapper tests, including under valgrind+- Versions test (ensuring `zstd` can decode files from all previous versions)+- `pzstd` with asan and tsan, as well as in 32-bits mode+- Testing `zstd` with legacy mode off+- Testing `zbuff` (old streaming API)+- Entire test suite and make install on macOS
zstd/lib/README.md view
@@ -1,50 +1,120 @@ Zstandard library files ================================ -The __lib__ directory contains several directories.-Depending on target use case, it's enough to include only files from relevant directories.+The __lib__ directory is split into several sub-directories,+in order to make it easier to select or exclude features. +#### Building++`Makefile` script is provided, supporting [Makefile conventions](https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html#Makefile-Conventions),+including commands variables, staged install, directory variables and standard targets.+- `make` : generates both static and dynamic libraries+- `make install` : install libraries and headers in target system directories++`libzstd` default scope is pretty large, including compression, decompression, dictionary builder,+and support for decoding legacy formats >= v0.5.0.+The scope can be reduced on demand (see paragraph _modular build_).+++#### Multithreading support++Multithreading is disabled by default when building with `make`.+Enabling multithreading requires 2 conditions :+- set build macro `ZSTD_MULTITHREAD` (`-DZSTD_MULTITHREAD` for `gcc`)+- for POSIX systems : compile with pthread (`-pthread` compilation flag for `gcc`)++Both conditions are automatically applied when invoking `make lib-mt` target.++When linking a POSIX program with a multithreaded version of `libzstd`,+note that it's necessary to request the `-pthread` flag during link stage.++Multithreading capabilities are exposed+via the [advanced API defined in `lib/zstd.h`](https://github.com/facebook/zstd/blob/v1.3.8/lib/zstd.h#L592).++ #### API -Zstandard's stable API is exposed within [zstd.h](zstd.h),-at the root of `lib` directory.+Zstandard's stable API is exposed within [lib/zstd.h](zstd.h). #### Advanced API -Some additional API may be useful if you're looking into advanced features :-- common/error_public.h : transforms `size_t` function results into an `enum`,- for precise error handling.-- ZSTD_STATIC_LINKING_ONLY : if you define this macro _before_ including `zstd.h`,- it will give access to advanced and experimental API.- These APIs shall ___never be used with dynamic library___ !- They are not "stable", their definition may change in the future.+Optional advanced features are exposed via :++- `lib/common/zstd_errors.h` : translates `size_t` function results+ into a `ZSTD_ErrorCode`, for accurate error handling.++- `ZSTD_STATIC_LINKING_ONLY` : if this macro is defined _before_ including `zstd.h`,+ it unlocks access to the experimental API,+ exposed in the second part of `zstd.h`.+ All definitions in the experimental APIs are unstable,+ they may still change in the future, or even be removed.+ As a consequence, experimental definitions shall ___never be used with dynamic library___ ! Only static linking is allowed. #### Modular build -Directory `common/` is required in all circumstances.-You can select to support compression only, by just adding files from the `compress/` directory,-In a similar way, you can build a decompressor-only library with the `decompress/` directory.+It's possible to compile only a limited set of features within `libzstd`.+The file structure is designed to make this selection manually achievable for any build system : -Other optional functionalities provided are :+- Directory `lib/common` is always required, for all variants. -- `dictBuilder/` : source files to create dictionaries.- The API can be consulted in `dictBuilder/zdict.h`.- This module also depends on `common/` and `compress/` .+- Compression source code lies in `lib/compress` -- `legacy/` : source code to decompress previous versions of zstd, starting from `v0.1`.- This module also depends on `common/` and `decompress/` .- Library compilation must include directive `ZSTD_LEGACY_SUPPORT = 1` .- The main API can be consulted in `legacy/zstd_legacy.h`.- Advanced API from each version can be found in their relevant header file.- For example, advanced API for version `v0.4` is in `legacy/zstd_v04.h` .+- Decompression source code lies in `lib/decompress` +- It's possible to include only `compress` or only `decompress`, they don't depend on each other. -#### Using MinGW+MSYS to create DLL+- `lib/dictBuilder` : makes it possible to generate dictionaries from a set of samples.+ The API is exposed in `lib/dictBuilder/zdict.h`.+ This module depends on both `lib/common` and `lib/compress` . +- `lib/legacy` : makes it possible to decompress legacy zstd formats, starting from `v0.1.0`.+ This module depends on `lib/common` and `lib/decompress`.+ To enable this feature, define `ZSTD_LEGACY_SUPPORT` during compilation.+ Specifying a number limits versions supported to that version onward.+ For example, `ZSTD_LEGACY_SUPPORT=2` means : "support legacy formats >= v0.2.0".+ Conversely, `ZSTD_LEGACY_SUPPORT=0` means "do __not__ support legacy formats".+ By default, this build macro is set as `ZSTD_LEGACY_SUPPORT=5`.+ Decoding supported legacy format is a transparent capability triggered within decompression functions.+ It's also allowed to invoke legacy API directly, exposed in `lib/legacy/zstd_legacy.h`.+ Each version does also provide its own set of advanced API.+ For example, advanced API for version `v0.4` is exposed in `lib/legacy/zstd_v04.h` .++- While invoking `make libzstd`, it's possible to define build macros+ `ZSTD_LIB_COMPRESSION, ZSTD_LIB_DECOMPRESSION`, `ZSTD_LIB_DICTBUILDER`,+ and `ZSTD_LIB_DEPRECATED` as `0` to forgo compilation of the corresponding features.+ This will also disable compilation of all dependencies+ (eg. `ZSTD_LIB_COMPRESSION=0` will also disable dictBuilder).++- There are some additional build macros that can be used to minify the decoder.++ Zstandard often has more than one implementation of a piece of functionality,+ where each implementation optimizes for different scenarios. For example, the+ Huffman decoder has complementary implementations that decode the stream one+ symbol at a time or two symbols at a time. Zstd normally includes both (and+ dispatches between them at runtime), but by defining `HUF_FORCE_DECOMPRESS_X1`+ or `HUF_FORCE_DECOMPRESS_X2`, you can force the use of one or the other, avoiding+ compilation of the other. Similarly, `ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT`+ and `ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG` force the compilation and use of+ only one or the other of two decompression implementations. The smallest+ binary is achieved by using `HUF_FORCE_DECOMPRESS_X1` and+ `ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT`.++ For squeezing the last ounce of size out, you can also define+ `ZSTD_NO_INLINE`, which disables inlining, and `ZSTD_STRIP_ERROR_STRINGS`,+ which removes the error messages that are otherwise returned by+ `ZSTD_getErrorName`.++- While invoking `make libzstd`, the build macro `ZSTD_LEGACY_MULTITHREADED_API=1`+ will expose the deprecated `ZSTDMT` API exposed by `zstdmt_compress.h` in+ the shared library, which is now hidden by default.+++#### Windows : using MinGW+MSYS to create DLL+ DLL can be created using MinGW+MSYS with the `make libzstd` command. This command creates `dll\libzstd.dll` and the import library `dll\libzstd.lib`. The import library is only required with Visual C++.@@ -59,19 +129,20 @@ The compiled executable will require ZSTD DLL which is available at `dll\libzstd.dll`. -#### Obsolete streaming API+#### Deprecated API -Streaming is now provided within `zstd.h`.-Older streaming API is still available within `deprecated/zbuff.h`.-It will be removed in a future version.-Consider migrating code towards newer streaming API in `zstd.h`.+Obsolete API on their way out are stored in directory `lib/deprecated`.+At this stage, it contains older streaming prototypes, in `lib/deprecated/zbuff.h`.+These prototypes will be removed in some future version.+Consider migrating code towards supported streaming API exposed in `zstd.h`. #### Miscellaneous The other files are not source code. There are : - - LICENSE : contains the BSD license text- - Makefile : script to compile or install zstd library (static and dynamic)- - libzstd.pc.in : for pkg-config (`make install`)- - README.md : this file+ - `BUCK` : support for `buck` build system (https://buckbuild.com/)+ - `Makefile` : `make` script to build and install zstd library (static and dynamic)+ - `README.md` : this file+ - `dll/` : resources directory for Windows compilation+ - `libzstd.pc.in` : script for `pkg-config` (used in `make install`)
zstd/lib/common/bitstream.h view
@@ -1,8 +1,7 @@ /* ****************************************************************** bitstream Part of FSE library- header file (to include)- Copyright (C) 2013-2016, Yann Collet.+ Copyright (C) 2013-present, Yann Collet. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -39,7 +38,6 @@ extern "C" { #endif - /* * This API consists of small unitary functions, which must be inlined for best performance. * Since link-time-optimization is not available for all compilers,@@ -50,6 +48,7 @@ * Dependencies ******************************************/ #include "mem.h" /* unaligned access routines */+#include "debug.h" /* assert(), DEBUGLOG(), RAWLOG() */ #include "error_private.h" /* error codes and messages */ @@ -60,18 +59,21 @@ # include <immintrin.h> /* support for bextr (experimental) */ #endif +#define STREAM_ACCUMULATOR_MIN_32 25+#define STREAM_ACCUMULATOR_MIN_64 57+#define STREAM_ACCUMULATOR_MIN ((U32)(MEM_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64)) + /*-****************************************** * bitStream encoding API (write forward) ********************************************/ /* bitStream can mix input from multiple sources.-* A critical property of these streams is that they encode and decode in **reverse** direction.-* So the first bit sequence you add will be the last to be read, like a LIFO stack.-*/-typedef struct-{+ * A critical property of these streams is that they encode and decode in **reverse** direction.+ * So the first bit sequence you add will be the last to be read, like a LIFO stack.+ */+typedef struct { size_t bitContainer;- int bitPos;+ unsigned bitPos; char* startPtr; char* ptr; char* endPtr;@@ -103,12 +105,12 @@ /*-******************************************** * bitStream decoding API (read backward) **********************************************/-typedef struct-{+typedef struct { size_t bitContainer; unsigned bitsConsumed; const char* ptr; const char* start;+ const char* limitPtr; } BIT_DStream_t; typedef enum { BIT_DStream_unfinished = 0,@@ -151,140 +153,179 @@ /*-************************************************************** * Internal functions ****************************************************************/-MEM_STATIC unsigned BIT_highbit32 (register U32 val)+MEM_STATIC unsigned BIT_highbit32 (U32 val) {+ assert(val != 0);+ { # if defined(_MSC_VER) /* Visual */- unsigned long r=0;- _BitScanReverse ( &r, val );- return (unsigned) r;+ unsigned long r=0;+ _BitScanReverse ( &r, val );+ return (unsigned) r; # elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */- return 31 - __builtin_clz (val);+ return 31 - __builtin_clz (val); # else /* Software version */- static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };- U32 v = val;- v |= v >> 1;- v |= v >> 2;- v |= v >> 4;- v |= v >> 8;- v |= v >> 16;- return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27];+ static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29,+ 11, 14, 16, 18, 22, 25, 3, 30,+ 8, 12, 20, 28, 15, 17, 24, 7,+ 19, 27, 23, 6, 26, 5, 4, 31 };+ U32 v = val;+ v |= v >> 1;+ v |= v >> 2;+ v |= v >> 4;+ v |= v >> 8;+ v |= v >> 16;+ return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27]; # endif+ } } /*===== Local Constants =====*/-static const unsigned BIT_mask[] = { 0, 1, 3, 7, 0xF, 0x1F, 0x3F, 0x7F, 0xFF, 0x1FF, 0x3FF, 0x7FF, 0xFFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF, 0x1FFFF, 0x3FFFF, 0x7FFFF, 0xFFFFF, 0x1FFFFF, 0x3FFFFF, 0x7FFFFF, 0xFFFFFF, 0x1FFFFFF, 0x3FFFFFF }; /* up to 26 bits */-+static const unsigned BIT_mask[] = {+ 0, 1, 3, 7, 0xF, 0x1F,+ 0x3F, 0x7F, 0xFF, 0x1FF, 0x3FF, 0x7FF,+ 0xFFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF, 0x1FFFF,+ 0x3FFFF, 0x7FFFF, 0xFFFFF, 0x1FFFFF, 0x3FFFFF, 0x7FFFFF,+ 0xFFFFFF, 0x1FFFFFF, 0x3FFFFFF, 0x7FFFFFF, 0xFFFFFFF, 0x1FFFFFFF,+ 0x3FFFFFFF, 0x7FFFFFFF}; /* up to 31 bits */+#define BIT_MASK_SIZE (sizeof(BIT_mask) / sizeof(BIT_mask[0])) /*-************************************************************** * bitStream encoding ****************************************************************/ /*! BIT_initCStream() :- * `dstCapacity` must be > sizeof(void*)+ * `dstCapacity` must be > sizeof(size_t) * @return : 0 if success,- otherwise an error code (can be tested using ERR_isError() ) */-MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC, void* startPtr, size_t dstCapacity)+ * otherwise an error code (can be tested using ERR_isError()) */+MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,+ void* startPtr, size_t dstCapacity) { bitC->bitContainer = 0; bitC->bitPos = 0; bitC->startPtr = (char*)startPtr; bitC->ptr = bitC->startPtr;- bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->ptr);- if (dstCapacity <= sizeof(bitC->ptr)) return ERROR(dstSize_tooSmall);+ bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->bitContainer);+ if (dstCapacity <= sizeof(bitC->bitContainer)) return ERROR(dstSize_tooSmall); return 0; } /*! BIT_addBits() :- can add up to 26 bits into `bitC`.- Does not check for register overflow ! */-MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC, size_t value, unsigned nbBits)+ * can add up to 31 bits into `bitC`.+ * Note : does not check for register overflow ! */+MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,+ size_t value, unsigned nbBits) {+ MEM_STATIC_ASSERT(BIT_MASK_SIZE == 32);+ assert(nbBits < BIT_MASK_SIZE);+ assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8); bitC->bitContainer |= (value & BIT_mask[nbBits]) << bitC->bitPos; bitC->bitPos += nbBits; } /*! BIT_addBitsFast() :- * works only if `value` is _clean_, meaning all high bits above nbBits are 0 */-MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC, size_t value, unsigned nbBits)+ * works only if `value` is _clean_,+ * meaning all high bits above nbBits are 0 */+MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,+ size_t value, unsigned nbBits) {+ assert((value>>nbBits) == 0);+ assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8); bitC->bitContainer |= value << bitC->bitPos; bitC->bitPos += nbBits; } /*! BIT_flushBitsFast() :+ * assumption : bitContainer has not overflowed * unsafe version; does not check buffer overflow */ MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC) { size_t const nbBytes = bitC->bitPos >> 3;+ assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8); MEM_writeLEST(bitC->ptr, bitC->bitContainer); bitC->ptr += nbBytes;+ assert(bitC->ptr <= bitC->endPtr); bitC->bitPos &= 7;- bitC->bitContainer >>= nbBytes*8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */+ bitC->bitContainer >>= nbBytes*8; } /*! BIT_flushBits() :+ * assumption : bitContainer has not overflowed * safe version; check for buffer overflow, and prevents it.- * note : does not signal buffer overflow. This will be revealed later on using BIT_closeCStream() */+ * note : does not signal buffer overflow.+ * overflow will be revealed later on using BIT_closeCStream() */ MEM_STATIC void BIT_flushBits(BIT_CStream_t* bitC) { size_t const nbBytes = bitC->bitPos >> 3;+ assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8); MEM_writeLEST(bitC->ptr, bitC->bitContainer); bitC->ptr += nbBytes; if (bitC->ptr > bitC->endPtr) bitC->ptr = bitC->endPtr; bitC->bitPos &= 7;- bitC->bitContainer >>= nbBytes*8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */+ bitC->bitContainer >>= nbBytes*8; } /*! BIT_closeCStream() : * @return : size of CStream, in bytes,- or 0 if it could not fit into dstBuffer */+ * or 0 if it could not fit into dstBuffer */ MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC) { BIT_addBitsFast(bitC, 1, 1); /* endMark */ BIT_flushBits(bitC);-- if (bitC->ptr >= bitC->endPtr) return 0; /* doesn't fit within authorized budget : cancel */-+ if (bitC->ptr >= bitC->endPtr) return 0; /* overflow detected */ return (bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0); } /*-********************************************************-* bitStream decoding+* bitStream decoding **********************************************************/ /*! BIT_initDStream() :-* Initialize a BIT_DStream_t.-* `bitD` : a pointer to an already allocated BIT_DStream_t structure.-* `srcSize` must be the *exact* size of the bitStream, in bytes.-* @return : size of stream (== srcSize) or an errorCode if a problem is detected-*/+ * Initialize a BIT_DStream_t.+ * `bitD` : a pointer to an already allocated BIT_DStream_t structure.+ * `srcSize` must be the *exact* size of the bitStream, in bytes.+ * @return : size of stream (== srcSize), or an errorCode if a problem is detected+ */ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize) { if (srcSize < 1) { memset(bitD, 0, sizeof(*bitD)); return ERROR(srcSize_wrong); } + bitD->start = (const char*)srcBuffer;+ bitD->limitPtr = bitD->start + sizeof(bitD->bitContainer);+ if (srcSize >= sizeof(bitD->bitContainer)) { /* normal case */- bitD->start = (const char*)srcBuffer; bitD->ptr = (const char*)srcBuffer + srcSize - sizeof(bitD->bitContainer); bitD->bitContainer = MEM_readLEST(bitD->ptr); { BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1]; bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */ if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ } } else {- bitD->start = (const char*)srcBuffer; bitD->ptr = bitD->start; bitD->bitContainer = *(const BYTE*)(bitD->start); switch(srcSize) {- case 7: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);- case 6: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);- case 5: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);- case 4: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[3]) << 24;- case 3: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[2]) << 16;- case 2: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[1]) << 8;- default:;+ case 7: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);+ /* fall-through */++ case 6: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);+ /* fall-through */++ case 5: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);+ /* fall-through */++ case 4: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[3]) << 24;+ /* fall-through */++ case 3: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[2]) << 16;+ /* fall-through */++ case 2: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[1]) << 8;+ /* fall-through */++ default: break; }- { BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];- bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;- if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ }+ { BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];+ bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;+ if (lastByte == 0) return ERROR(corruption_detected); /* endMark not present */+ } bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize)*8; } @@ -298,20 +339,15 @@ MEM_STATIC size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits) {-#if defined(__BMI__) && defined(__GNUC__) && __GNUC__*1000+__GNUC_MINOR__ >= 4008 /* experimental */-# if defined(__x86_64__)- if (sizeof(bitContainer)==8)- return _bextr_u64(bitContainer, start, nbBits);- else-# endif- return _bextr_u32(bitContainer, start, nbBits);-#else- return (bitContainer >> start) & BIT_mask[nbBits];-#endif+ U32 const regMask = sizeof(bitContainer)*8 - 1;+ /* if start > regMask, bitstream is corrupted, and result is undefined */+ assert(nbBits < BIT_MASK_SIZE);+ return (bitContainer >> (start & regMask)) & BIT_mask[nbBits]; } MEM_STATIC size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits) {+ assert(nbBits < BIT_MASK_SIZE); return bitContainer & BIT_mask[nbBits]; } @@ -320,24 +356,28 @@ * local register is not modified. * On 32-bits, maxNbBits==24. * On 64-bits, maxNbBits==56.- * @return : value extracted- */- MEM_STATIC size_t BIT_lookBits(const BIT_DStream_t* bitD, U32 nbBits)+ * @return : value extracted */+MEM_STATIC size_t BIT_lookBits(const BIT_DStream_t* bitD, U32 nbBits) {-#if defined(__BMI__) && defined(__GNUC__) /* experimental; fails if bitD->bitsConsumed + nbBits > sizeof(bitD->bitContainer)*8 */+ /* arbitrate between double-shift and shift+mask */+#if 1+ /* if bitD->bitsConsumed + nbBits > sizeof(bitD->bitContainer)*8,+ * bitstream is likely corrupted, and result is undefined */ return BIT_getMiddleBits(bitD->bitContainer, (sizeof(bitD->bitContainer)*8) - bitD->bitsConsumed - nbBits, nbBits); #else- U32 const bitMask = sizeof(bitD->bitContainer)*8 - 1;- return ((bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> 1) >> ((bitMask-nbBits) & bitMask);+ /* this code path is slower on my os-x laptop */+ U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;+ return ((bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> 1) >> ((regMask-nbBits) & regMask); #endif } /*! BIT_lookBitsFast() :-* unsafe version; only works only if nbBits >= 1 */+ * unsafe version; only works if nbBits >= 1 */ MEM_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits) {- U32 const bitMask = sizeof(bitD->bitContainer)*8 - 1;- return (bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> (((bitMask+1)-nbBits) & bitMask);+ U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;+ assert(nbBits >= 1);+ return (bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> (((regMask+1)-nbBits) & regMask); } MEM_STATIC void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)@@ -348,9 +388,8 @@ /*! BIT_readBits() : * Read (consume) next n bits from local register and update. * Pay attention to not read more than nbBits contained into local register.- * @return : extracted value.- */-MEM_STATIC size_t BIT_readBits(BIT_DStream_t* bitD, U32 nbBits)+ * @return : extracted value. */+MEM_STATIC size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits) { size_t const value = BIT_lookBits(bitD, nbBits); BIT_skipBits(bitD, nbBits);@@ -358,25 +397,26 @@ } /*! BIT_readBitsFast() :-* unsafe version; only works only if nbBits >= 1 */-MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, U32 nbBits)+ * unsafe version; only works only if nbBits >= 1 */+MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits) { size_t const value = BIT_lookBitsFast(bitD, nbBits);+ assert(nbBits >= 1); BIT_skipBits(bitD, nbBits); return value; } /*! BIT_reloadDStream() :-* Refill `bitD` from buffer previously set in BIT_initDStream() .-* This function is safe, it guarantees it will not read beyond src buffer.-* @return : status of `BIT_DStream_t` internal register.- if status == BIT_DStream_unfinished, internal register is filled with >= (sizeof(bitD->bitContainer)*8 - 7) bits */+ * Refill `bitD` from buffer previously set in BIT_initDStream() .+ * This function is safe, it guarantees it will not read beyond src buffer.+ * @return : status of `BIT_DStream_t` internal register.+ * when status == BIT_DStream_unfinished, internal register is filled with at least 25 or 57 bits */ MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD) {- if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8)) /* should not happen => corruption detected */- return BIT_DStream_overflow;+ if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8)) /* overflow detected, like end of stream */+ return BIT_DStream_overflow; - if (bitD->ptr >= bitD->start + sizeof(bitD->bitContainer)) {+ if (bitD->ptr >= bitD->limitPtr) { bitD->ptr -= bitD->bitsConsumed >> 3; bitD->bitsConsumed &= 7; bitD->bitContainer = MEM_readLEST(bitD->ptr);@@ -386,6 +426,7 @@ if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer; return BIT_DStream_completed; }+ /* start < ptr < limitPtr */ { U32 nbBytes = bitD->bitsConsumed >> 3; BIT_DStream_status result = BIT_DStream_unfinished; if (bitD->ptr - nbBytes < bitD->start) {@@ -394,14 +435,14 @@ } bitD->ptr -= nbBytes; bitD->bitsConsumed -= nbBytes*8;- bitD->bitContainer = MEM_readLEST(bitD->ptr); /* reminder : srcSize > sizeof(bitD) */+ bitD->bitContainer = MEM_readLEST(bitD->ptr); /* reminder : srcSize > sizeof(bitD->bitContainer), otherwise bitD->ptr == bitD->start */ return result; } } /*! BIT_endOfDStream() :-* @return Tells if DStream has exactly reached its end (all bits consumed).-*/+ * @return : 1 if DStream has _exactly_ reached its end (all bits consumed).+ */ MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* DStream) { return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer)*8));
+ zstd/lib/common/compiler.h view
@@ -0,0 +1,140 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#ifndef ZSTD_COMPILER_H+#define ZSTD_COMPILER_H++/*-*******************************************************+* Compiler specifics+*********************************************************/+/* force inlining */++#if !defined(ZSTD_NO_INLINE)+#if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */+# define INLINE_KEYWORD inline+#else+# define INLINE_KEYWORD+#endif++#if defined(__GNUC__)+# define FORCE_INLINE_ATTR __attribute__((always_inline))+#elif defined(_MSC_VER)+# define FORCE_INLINE_ATTR __forceinline+#else+# define FORCE_INLINE_ATTR+#endif++#else++#define INLINE_KEYWORD+#define FORCE_INLINE_ATTR++#endif++/**+ * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant+ * parameters. They must be inlined for the compiler to eliminate the constant+ * branches.+ */+#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR+/**+ * HINT_INLINE is used to help the compiler generate better code. It is *not*+ * used for "templates", so it can be tweaked based on the compilers+ * performance.+ *+ * gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the+ * always_inline attribute.+ *+ * clang up to 5.0.0 (trunk) benefit tremendously from the always_inline+ * attribute.+ */+#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5+# define HINT_INLINE static INLINE_KEYWORD+#else+# define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR+#endif++/* force no inlining */+#ifdef _MSC_VER+# define FORCE_NOINLINE static __declspec(noinline)+#else+# ifdef __GNUC__+# define FORCE_NOINLINE static __attribute__((__noinline__))+# else+# define FORCE_NOINLINE static+# endif+#endif++/* target attribute */+#ifndef __has_attribute+ #define __has_attribute(x) 0 /* Compatibility with non-clang compilers. */+#endif+#if defined(__GNUC__)+# define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))+#else+# define TARGET_ATTRIBUTE(target)+#endif++/* Enable runtime BMI2 dispatch based on the CPU.+ * Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default.+ */+#ifndef DYNAMIC_BMI2+ #if ((defined(__clang__) && __has_attribute(__target__)) \+ || (defined(__GNUC__) \+ && (__GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)))) \+ && (defined(__x86_64__) || defined(_M_X86)) \+ && !defined(__BMI2__)+ # define DYNAMIC_BMI2 1+ #else+ # define DYNAMIC_BMI2 0+ #endif+#endif++/* prefetch+ * can be disabled, by declaring NO_PREFETCH build macro */+#if defined(NO_PREFETCH)+# define PREFETCH_L1(ptr) (void)(ptr) /* disabled */+# define PREFETCH_L2(ptr) (void)(ptr) /* disabled */+#else+# if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) /* _mm_prefetch() is not defined outside of x86/x64 */+# include <mmintrin.h> /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */+# define PREFETCH_L1(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T0)+# define PREFETCH_L2(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T1)+# elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) )+# define PREFETCH_L1(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */)+# define PREFETCH_L2(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */)+# else+# define PREFETCH_L1(ptr) (void)(ptr) /* disabled */+# define PREFETCH_L2(ptr) (void)(ptr) /* disabled */+# endif+#endif /* NO_PREFETCH */++#define CACHELINE_SIZE 64++#define PREFETCH_AREA(p, s) { \+ const char* const _ptr = (const char*)(p); \+ size_t const _size = (size_t)(s); \+ size_t _pos; \+ for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \+ PREFETCH_L2(_ptr + _pos); \+ } \+}++/* disable warnings */+#ifdef _MSC_VER /* Visual Studio */+# include <intrin.h> /* For Visual 2005 */+# pragma warning(disable : 4100) /* disable: C4100: unreferenced formal parameter */+# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */+# pragma warning(disable : 4204) /* disable: C4204: non-constant aggregate initializer */+# pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */+# pragma warning(disable : 4324) /* disable: C4324: padded structure */+#endif++#endif /* ZSTD_COMPILER_H */
+ zstd/lib/common/cpu.h view
@@ -0,0 +1,215 @@+/*+ * Copyright (c) 2018-present, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#ifndef ZSTD_COMMON_CPU_H+#define ZSTD_COMMON_CPU_H++/**+ * Implementation taken from folly/CpuId.h+ * https://github.com/facebook/folly/blob/master/folly/CpuId.h+ */++#include <string.h>++#include "mem.h"++#ifdef _MSC_VER+#include <intrin.h>+#endif++typedef struct {+ U32 f1c;+ U32 f1d;+ U32 f7b;+ U32 f7c;+} ZSTD_cpuid_t;++MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {+ U32 f1c = 0;+ U32 f1d = 0;+ U32 f7b = 0;+ U32 f7c = 0;+#if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86))+ int reg[4];+ __cpuid((int*)reg, 0);+ {+ int const n = reg[0];+ if (n >= 1) {+ __cpuid((int*)reg, 1);+ f1c = (U32)reg[2];+ f1d = (U32)reg[3];+ }+ if (n >= 7) {+ __cpuidex((int*)reg, 7, 0);+ f7b = (U32)reg[1];+ f7c = (U32)reg[2];+ }+ }+#elif defined(__i386__) && defined(__PIC__) && !defined(__clang__) && defined(__GNUC__)+ /* The following block like the normal cpuid branch below, but gcc+ * reserves ebx for use of its pic register so we must specially+ * handle the save and restore to avoid clobbering the register+ */+ U32 n;+ __asm__(+ "pushl %%ebx\n\t"+ "cpuid\n\t"+ "popl %%ebx\n\t"+ : "=a"(n)+ : "a"(0)+ : "ecx", "edx");+ if (n >= 1) {+ U32 f1a;+ __asm__(+ "pushl %%ebx\n\t"+ "cpuid\n\t"+ "popl %%ebx\n\t"+ : "=a"(f1a), "=c"(f1c), "=d"(f1d)+ : "a"(1));+ }+ if (n >= 7) {+ __asm__(+ "pushl %%ebx\n\t"+ "cpuid\n\t"+ "movl %%ebx, %%eax\n\t"+ "popl %%ebx"+ : "=a"(f7b), "=c"(f7c)+ : "a"(7), "c"(0)+ : "edx");+ }+#elif defined(__x86_64__) || defined(_M_X64) || defined(__i386__)+ U32 n;+ __asm__("cpuid" : "=a"(n) : "a"(0) : "ebx", "ecx", "edx");+ if (n >= 1) {+ U32 f1a;+ __asm__("cpuid" : "=a"(f1a), "=c"(f1c), "=d"(f1d) : "a"(1) : "ebx");+ }+ if (n >= 7) {+ U32 f7a;+ __asm__("cpuid"+ : "=a"(f7a), "=b"(f7b), "=c"(f7c)+ : "a"(7), "c"(0)+ : "edx");+ }+#endif+ {+ ZSTD_cpuid_t cpuid;+ cpuid.f1c = f1c;+ cpuid.f1d = f1d;+ cpuid.f7b = f7b;+ cpuid.f7c = f7c;+ return cpuid;+ }+}++#define X(name, r, bit) \+ MEM_STATIC int ZSTD_cpuid_##name(ZSTD_cpuid_t const cpuid) { \+ return ((cpuid.r) & (1U << bit)) != 0; \+ }++/* cpuid(1): Processor Info and Feature Bits. */+#define C(name, bit) X(name, f1c, bit)+ C(sse3, 0)+ C(pclmuldq, 1)+ C(dtes64, 2)+ C(monitor, 3)+ C(dscpl, 4)+ C(vmx, 5)+ C(smx, 6)+ C(eist, 7)+ C(tm2, 8)+ C(ssse3, 9)+ C(cnxtid, 10)+ C(fma, 12)+ C(cx16, 13)+ C(xtpr, 14)+ C(pdcm, 15)+ C(pcid, 17)+ C(dca, 18)+ C(sse41, 19)+ C(sse42, 20)+ C(x2apic, 21)+ C(movbe, 22)+ C(popcnt, 23)+ C(tscdeadline, 24)+ C(aes, 25)+ C(xsave, 26)+ C(osxsave, 27)+ C(avx, 28)+ C(f16c, 29)+ C(rdrand, 30)+#undef C+#define D(name, bit) X(name, f1d, bit)+ D(fpu, 0)+ D(vme, 1)+ D(de, 2)+ D(pse, 3)+ D(tsc, 4)+ D(msr, 5)+ D(pae, 6)+ D(mce, 7)+ D(cx8, 8)+ D(apic, 9)+ D(sep, 11)+ D(mtrr, 12)+ D(pge, 13)+ D(mca, 14)+ D(cmov, 15)+ D(pat, 16)+ D(pse36, 17)+ D(psn, 18)+ D(clfsh, 19)+ D(ds, 21)+ D(acpi, 22)+ D(mmx, 23)+ D(fxsr, 24)+ D(sse, 25)+ D(sse2, 26)+ D(ss, 27)+ D(htt, 28)+ D(tm, 29)+ D(pbe, 31)+#undef D++/* cpuid(7): Extended Features. */+#define B(name, bit) X(name, f7b, bit)+ B(bmi1, 3)+ B(hle, 4)+ B(avx2, 5)+ B(smep, 7)+ B(bmi2, 8)+ B(erms, 9)+ B(invpcid, 10)+ B(rtm, 11)+ B(mpx, 14)+ B(avx512f, 16)+ B(avx512dq, 17)+ B(rdseed, 18)+ B(adx, 19)+ B(smap, 20)+ B(avx512ifma, 21)+ B(pcommit, 22)+ B(clflushopt, 23)+ B(clwb, 24)+ B(avx512pf, 26)+ B(avx512er, 27)+ B(avx512cd, 28)+ B(sha, 29)+ B(avx512bw, 30)+ B(avx512vl, 31)+#undef B+#define C(name, bit) X(name, f7c, bit)+ C(prefetchwt1, 0)+ C(avx512vbmi, 1)+#undef C++#undef X++#endif /* ZSTD_COMMON_CPU_H */
+ zstd/lib/common/debug.c view
@@ -0,0 +1,44 @@+/* ******************************************************************+ debug+ Part of FSE library+ Copyright (C) 2013-present, Yann Collet.++ BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)++ Redistribution and use in source and binary forms, with or without+ modification, are permitted provided that the following conditions are+ met:++ * Redistributions of source code must retain the above copyright+ notice, this list of conditions and the following disclaimer.+ * Redistributions in binary form must reproduce the above+ copyright notice, this list of conditions and the following disclaimer+ in the documentation and/or other materials provided with the+ distribution.++ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS+ "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT+ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR+ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT+ OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,+ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT+ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,+ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY+ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.++ You can contact the author at :+ - Source repository : https://github.com/Cyan4973/FiniteStateEntropy+****************************************************************** */+++/*+ * This module only hosts one global variable+ * which can be used to dynamically influence the verbosity of traces,+ * such as DEBUGLOG and RAWLOG+ */++#include "debug.h"++int g_debuglevel = DEBUGLEVEL;
+ zstd/lib/common/debug.h view
@@ -0,0 +1,134 @@+/* ******************************************************************+ debug+ Part of FSE library+ Copyright (C) 2013-present, Yann Collet.++ BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)++ Redistribution and use in source and binary forms, with or without+ modification, are permitted provided that the following conditions are+ met:++ * Redistributions of source code must retain the above copyright+ notice, this list of conditions and the following disclaimer.+ * Redistributions in binary form must reproduce the above+ copyright notice, this list of conditions and the following disclaimer+ in the documentation and/or other materials provided with the+ distribution.++ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS+ "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT+ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR+ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT+ OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,+ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT+ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,+ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY+ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.++ You can contact the author at :+ - Source repository : https://github.com/Cyan4973/FiniteStateEntropy+****************************************************************** */+++/*+ * The purpose of this header is to enable debug functions.+ * They regroup assert(), DEBUGLOG() and RAWLOG() for run-time,+ * and DEBUG_STATIC_ASSERT() for compile-time.+ *+ * By default, DEBUGLEVEL==0, which means run-time debug is disabled.+ *+ * Level 1 enables assert() only.+ * Starting level 2, traces can be generated and pushed to stderr.+ * The higher the level, the more verbose the traces.+ *+ * It's possible to dynamically adjust level using variable g_debug_level,+ * which is only declared if DEBUGLEVEL>=2,+ * and is a global variable, not multi-thread protected (use with care)+ */++#ifndef DEBUG_H_12987983217+#define DEBUG_H_12987983217++#if defined (__cplusplus)+extern "C" {+#endif+++/* static assert is triggered at compile time, leaving no runtime artefact.+ * static assert only works with compile-time constants.+ * Also, this variant can only be used inside a function. */+#define DEBUG_STATIC_ASSERT(c) (void)sizeof(char[(c) ? 1 : -1])+++/* DEBUGLEVEL is expected to be defined externally,+ * typically through compiler command line.+ * Value must be a number. */+#ifndef DEBUGLEVEL+# define DEBUGLEVEL 0+#endif+++/* DEBUGFILE can be defined externally,+ * typically through compiler command line.+ * note : currently useless.+ * Value must be stderr or stdout */+#ifndef DEBUGFILE+# define DEBUGFILE stderr+#endif+++/* recommended values for DEBUGLEVEL :+ * 0 : release mode, no debug, all run-time checks disabled+ * 1 : enables assert() only, no display+ * 2 : reserved, for currently active debug path+ * 3 : events once per object lifetime (CCtx, CDict, etc.)+ * 4 : events once per frame+ * 5 : events once per block+ * 6 : events once per sequence (verbose)+ * 7+: events at every position (*very* verbose)+ *+ * It's generally inconvenient to output traces > 5.+ * In which case, it's possible to selectively trigger high verbosity levels+ * by modifying g_debug_level.+ */++#if (DEBUGLEVEL>=1)+# include <assert.h>+#else+# ifndef assert /* assert may be already defined, due to prior #include <assert.h> */+# define assert(condition) ((void)0) /* disable assert (default) */+# endif+#endif++#if (DEBUGLEVEL>=2)+# include <stdio.h>+extern int g_debuglevel; /* the variable is only declared,+ it actually lives in debug.c,+ and is shared by the whole process.+ It's not thread-safe.+ It's useful when enabling very verbose levels+ on selective conditions (such as position in src) */++# define RAWLOG(l, ...) { \+ if (l<=g_debuglevel) { \+ fprintf(stderr, __VA_ARGS__); \+ } }+# define DEBUGLOG(l, ...) { \+ if (l<=g_debuglevel) { \+ fprintf(stderr, __FILE__ ": " __VA_ARGS__); \+ fprintf(stderr, " \n"); \+ } }+#else+# define RAWLOG(l, ...) {} /* disabled */+# define DEBUGLOG(l, ...) {} /* disabled */+#endif+++#if defined (__cplusplus)+}+#endif++#endif /* DEBUG_H_12987983217 */
zstd/lib/common/entropy_common.c view
@@ -43,27 +43,21 @@ #include "huf.h" -/*-****************************************-* FSE Error Management-******************************************/-unsigned FSE_isError(size_t code) { return ERR_isError(code); }+/*=== Version ===*/+unsigned FSE_versionNumber(void) { return FSE_VERSION_NUMBER; } -const char* FSE_getErrorName(size_t code) { return ERR_getErrorName(code); } +/*=== Error Management ===*/+unsigned FSE_isError(size_t code) { return ERR_isError(code); }+const char* FSE_getErrorName(size_t code) { return ERR_getErrorName(code); } -/* **************************************************************-* HUF Error Management-****************************************************************/ unsigned HUF_isError(size_t code) { return ERR_isError(code); }- const char* HUF_getErrorName(size_t code) { return ERR_getErrorName(code); } /*-************************************************************** * FSE NCount encoding-decoding ****************************************************************/-static short FSE_abs(short a) { return (short)(a<0 ? -a : a); }- size_t FSE_readNCount (short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr, const void* headerBuffer, size_t hbSize) {@@ -78,7 +72,21 @@ unsigned charnum = 0; int previous0 = 0; - if (hbSize < 4) return ERROR(srcSize_wrong);+ if (hbSize < 4) {+ /* This function only works when hbSize >= 4 */+ char buffer[4];+ memset(buffer, 0, sizeof(buffer));+ memcpy(buffer, headerBuffer, hbSize);+ { size_t const countSize = FSE_readNCount(normalizedCounter, maxSVPtr, tableLogPtr,+ buffer, sizeof(buffer));+ if (FSE_isError(countSize)) return countSize;+ if (countSize > hbSize) return ERROR(corruption_detected);+ return countSize;+ } }+ assert(hbSize >= 4);++ /* init */+ memset(normalizedCounter, 0, (*maxSVPtr+1) * sizeof(normalizedCounter[0])); /* all symbols not present in NCount have a frequency of 0 */ bitStream = MEM_readLE32(ip); nbBits = (bitStream & 0xF) + FSE_MIN_TABLELOG; /* extract tableLog */ if (nbBits > FSE_TABLELOG_ABSOLUTE_MAX) return ERROR(tableLog_tooLarge);@@ -111,27 +119,28 @@ if (n0 > *maxSVPtr) return ERROR(maxSymbolValue_tooSmall); while (charnum < n0) normalizedCounter[charnum++] = 0; if ((ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) {+ assert((bitCount >> 3) <= 3); /* For first condition to work */ ip += bitCount>>3; bitCount &= 7; bitStream = MEM_readLE32(ip) >> bitCount; } else { bitStream >>= 2; } }- { short const max = (short)((2*threshold-1)-remaining);- short count;+ { int const max = (2*threshold-1) - remaining;+ int count; if ((bitStream & (threshold-1)) < (U32)max) {- count = (short)(bitStream & (threshold-1));- bitCount += nbBits-1;+ count = bitStream & (threshold-1);+ bitCount += nbBits-1; } else {- count = (short)(bitStream & (2*threshold-1));+ count = bitStream & (2*threshold-1); if (count >= threshold) count -= max;- bitCount += nbBits;+ bitCount += nbBits; } count--; /* extra accuracy */- remaining -= FSE_abs(count);- normalizedCounter[charnum++] = count;+ remaining -= count < 0 ? -count : count; /* -1 means +1 */+ normalizedCounter[charnum++] = (short)count; previous0 = !count; while (remaining < threshold) { nbBits--;
zstd/lib/common/error_private.c view
@@ -1,10 +1,11 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ /* The purpose of this file is to have a single list of error strings embedded in binary */@@ -13,6 +14,10 @@ const char* ERR_getErrorString(ERR_enum code) {+#ifdef ZSTD_STRIP_ERROR_STRINGS+ (void)code;+ return "Error strings stripped";+#else static const char* const notErrorCode = "Unspecified error code"; switch( code ) {@@ -20,24 +25,30 @@ case PREFIX(GENERIC): return "Error (generic)"; case PREFIX(prefix_unknown): return "Unknown frame descriptor"; case PREFIX(version_unsupported): return "Version not supported";- case PREFIX(parameter_unknown): return "Unknown parameter type"; case PREFIX(frameParameter_unsupported): return "Unsupported frame parameter";- case PREFIX(frameParameter_unsupportedBy32bits): return "Frame parameter unsupported in 32-bits mode"; case PREFIX(frameParameter_windowTooLarge): return "Frame requires too much memory for decoding";- case PREFIX(compressionParameter_unsupported): return "Compression parameter is out of bound";+ case PREFIX(corruption_detected): return "Corrupted block detected";+ case PREFIX(checksum_wrong): return "Restored data doesn't match checksum";+ case PREFIX(parameter_unsupported): return "Unsupported parameter";+ case PREFIX(parameter_outOfBound): return "Parameter is out of bound"; case PREFIX(init_missing): return "Context should be init first"; case PREFIX(memory_allocation): return "Allocation error : not enough memory";+ case PREFIX(workSpace_tooSmall): return "workSpace buffer is not large enough"; case PREFIX(stage_wrong): return "Operation not authorized at current processing stage";- case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";- case PREFIX(srcSize_wrong): return "Src size incorrect";- case PREFIX(corruption_detected): return "Corrupted block detected";- case PREFIX(checksum_wrong): return "Restored data doesn't match checksum"; case PREFIX(tableLog_tooLarge): return "tableLog requires too much memory : unsupported"; case PREFIX(maxSymbolValue_tooLarge): return "Unsupported max Symbol Value : too large"; case PREFIX(maxSymbolValue_tooSmall): return "Specified maxSymbolValue is too small"; case PREFIX(dictionary_corrupted): return "Dictionary is corrupted"; case PREFIX(dictionary_wrong): return "Dictionary mismatch";+ case PREFIX(dictionaryCreation_failed): return "Cannot create Dictionary from provided samples";+ case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";+ case PREFIX(srcSize_wrong): return "Src size is incorrect";+ case PREFIX(dstBuffer_null): return "Operation on NULL destination buffer";+ /* following error codes are not stable and may be removed or changed in a future version */+ case PREFIX(frameIndex_tooLarge): return "Frame index is too large";+ case PREFIX(seekableIO): return "An I/O error occurred when reading/seeking"; case PREFIX(maxCode): default: return notErrorCode; }+#endif }
zstd/lib/common/error_private.h view
@@ -1,10 +1,11 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ /* Note : this module is expected to remain private, do not expose it */@@ -48,10 +49,9 @@ /*-**************************************** * Error codes handling ******************************************/-#ifdef ERROR-# undef ERROR /* reported already defined on VS 2015 (Rich Geldreich) */-#endif-#define ERROR(name) ((size_t)-PREFIX(name))+#undef ERROR /* reported already defined on VS 2015 (Rich Geldreich) */+#define ERROR(name) ZSTD_ERROR(name)+#define ZSTD_ERROR(name) ((size_t)-PREFIX(name)) ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); }
zstd/lib/common/fse.h view
@@ -31,20 +31,48 @@ You can contact the author at : - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ****************************************************************** */-#ifndef FSE_H-#define FSE_H #if defined (__cplusplus) extern "C" { #endif +#ifndef FSE_H+#define FSE_H + /*-***************************************** * Dependencies ******************************************/ #include <stddef.h> /* size_t, ptrdiff_t */ +/*-*****************************************+* FSE_PUBLIC_API : control library symbols visibility+******************************************/+#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)+# define FSE_PUBLIC_API __attribute__ ((visibility ("default")))+#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) /* Visual expected */+# define FSE_PUBLIC_API __declspec(dllexport)+#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)+# define FSE_PUBLIC_API __declspec(dllimport) /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/+#else+# define FSE_PUBLIC_API+#endif++/*------ Version ------*/+#define FSE_VERSION_MAJOR 0+#define FSE_VERSION_MINOR 9+#define FSE_VERSION_RELEASE 0++#define FSE_LIB_VERSION FSE_VERSION_MAJOR.FSE_VERSION_MINOR.FSE_VERSION_RELEASE+#define FSE_QUOTE(str) #str+#define FSE_EXPAND_AND_QUOTE(str) FSE_QUOTE(str)+#define FSE_VERSION_STRING FSE_EXPAND_AND_QUOTE(FSE_LIB_VERSION)++#define FSE_VERSION_NUMBER (FSE_VERSION_MAJOR *100*100 + FSE_VERSION_MINOR *100 + FSE_VERSION_RELEASE)+FSE_PUBLIC_API unsigned FSE_versionNumber(void); /**< library version number; to be used when checking dll version */++ /*-**************************************** * FSE simple functions ******************************************/@@ -56,8 +84,8 @@ if return == 1, srcData is a single byte symbol * srcSize times. Use RLE compression instead. if FSE_isError(return), compression failed (more details using FSE_getErrorName()) */-size_t FSE_compress(void* dst, size_t dstCapacity,- const void* src, size_t srcSize);+FSE_PUBLIC_API size_t FSE_compress(void* dst, size_t dstCapacity,+ const void* src, size_t srcSize); /*! FSE_decompress(): Decompress FSE data from buffer 'cSrc', of size 'cSrcSize',@@ -69,18 +97,18 @@ Why ? : making this distinction requires a header. Header management is intentionally delegated to the user layer, which can better manage special cases. */-size_t FSE_decompress(void* dst, size_t dstCapacity,- const void* cSrc, size_t cSrcSize);+FSE_PUBLIC_API size_t FSE_decompress(void* dst, size_t dstCapacity,+ const void* cSrc, size_t cSrcSize); /*-***************************************** * Tool functions ******************************************/-size_t FSE_compressBound(size_t size); /* maximum compressed size */+FSE_PUBLIC_API size_t FSE_compressBound(size_t size); /* maximum compressed size */ /* Error Management */-unsigned FSE_isError(size_t code); /* tells if a return value is an error code */-const char* FSE_getErrorName(size_t code); /* provides error code string (useful for debugging) */+FSE_PUBLIC_API unsigned FSE_isError(size_t code); /* tells if a return value is an error code */+FSE_PUBLIC_API const char* FSE_getErrorName(size_t code); /* provides error code string (useful for debugging) */ /*-*****************************************@@ -94,7 +122,7 @@ if return == 1, srcData is a single byte symbol * srcSize times. Use RLE compression. if FSE_isError(return), it's an error code. */-size_t FSE_compress2 (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);+FSE_PUBLIC_API size_t FSE_compress2 (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog); /*-*****************************************@@ -102,7 +130,7 @@ ******************************************/ /*! FSE_compress() does the following:-1. count symbol occurrence from source[] into table count[]+1. count symbol occurrence from source[] into table count[] (see hist.h) 2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog) 3. save normalized counters to memory buffer using writeNCount() 4. build encoding table 'CTable' from normalized counters@@ -120,57 +148,50 @@ /* *** COMPRESSION *** */ -/*! FSE_count():- Provides the precise count of each byte within a table 'count'.- 'count' is a table of unsigned int, of minimum size (*maxSymbolValuePtr+1).- *maxSymbolValuePtr will be updated if detected smaller than initial value.- @return : the count of the most frequent symbol (which is not identified).- if return == srcSize, there is only one symbol.- Can also return an error code, which can be tested with FSE_isError(). */-size_t FSE_count(unsigned* count, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize);- /*! FSE_optimalTableLog(): dynamically downsize 'tableLog' when conditions are met. It saves CPU time, by using smaller tables, while preserving or even improving compression ratio. @return : recommended tableLog (necessarily <= 'maxTableLog') */-unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);+FSE_PUBLIC_API unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue); /*! FSE_normalizeCount(): normalize counts so that sum(count[]) == Power_of_2 (2^tableLog) 'normalizedCounter' is a table of short, of minimum size (maxSymbolValue+1). @return : tableLog, or an errorCode, which can be tested using FSE_isError() */-size_t FSE_normalizeCount(short* normalizedCounter, unsigned tableLog, const unsigned* count, size_t srcSize, unsigned maxSymbolValue);+FSE_PUBLIC_API size_t FSE_normalizeCount(short* normalizedCounter, unsigned tableLog,+ const unsigned* count, size_t srcSize, unsigned maxSymbolValue); /*! FSE_NCountWriteBound(): Provides the maximum possible size of an FSE normalized table, given 'maxSymbolValue' and 'tableLog'. Typically useful for allocation purpose. */-size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog);+FSE_PUBLIC_API size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog); /*! FSE_writeNCount(): Compactly save 'normalizedCounter' into 'buffer'. @return : size of the compressed table, or an errorCode, which can be tested using FSE_isError(). */-size_t FSE_writeNCount (void* buffer, size_t bufferSize, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);-+FSE_PUBLIC_API size_t FSE_writeNCount (void* buffer, size_t bufferSize,+ const short* normalizedCounter,+ unsigned maxSymbolValue, unsigned tableLog); /*! Constructor and Destructor of FSE_CTable. Note that FSE_CTable size depends on 'tableLog' and 'maxSymbolValue' */ typedef unsigned FSE_CTable; /* don't allocate that. It's only meant to be more restrictive than void* */-FSE_CTable* FSE_createCTable (unsigned tableLog, unsigned maxSymbolValue);-void FSE_freeCTable (FSE_CTable* ct);+FSE_PUBLIC_API FSE_CTable* FSE_createCTable (unsigned maxSymbolValue, unsigned tableLog);+FSE_PUBLIC_API void FSE_freeCTable (FSE_CTable* ct); /*! FSE_buildCTable(): Builds `ct`, which must be already allocated, using FSE_createCTable(). @return : 0, or an errorCode, which can be tested using FSE_isError() */-size_t FSE_buildCTable(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);+FSE_PUBLIC_API size_t FSE_buildCTable(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog); /*! FSE_compress_usingCTable(): Compress `src` using `ct` into `dst` which must be already allocated. @return : size of compressed data (<= `dstCapacity`), or 0 if compressed data could not fit into `dst`, or an errorCode, which can be tested using FSE_isError() */-size_t FSE_compress_usingCTable (void* dst, size_t dstCapacity, const void* src, size_t srcSize, const FSE_CTable* ct);+FSE_PUBLIC_API size_t FSE_compress_usingCTable (void* dst, size_t dstCapacity, const void* src, size_t srcSize, const FSE_CTable* ct); /*! Tutorial :@@ -223,25 +244,27 @@ @return : size read from 'rBuffer', or an errorCode, which can be tested using FSE_isError(). maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */-size_t FSE_readNCount (short* normalizedCounter, unsigned* maxSymbolValuePtr, unsigned* tableLogPtr, const void* rBuffer, size_t rBuffSize);+FSE_PUBLIC_API size_t FSE_readNCount (short* normalizedCounter,+ unsigned* maxSymbolValuePtr, unsigned* tableLogPtr,+ const void* rBuffer, size_t rBuffSize); /*! Constructor and Destructor of FSE_DTable. Note that its size depends on 'tableLog' */ typedef unsigned FSE_DTable; /* don't allocate that. It's just a way to be more restrictive than void* */-FSE_DTable* FSE_createDTable(unsigned tableLog);-void FSE_freeDTable(FSE_DTable* dt);+FSE_PUBLIC_API FSE_DTable* FSE_createDTable(unsigned tableLog);+FSE_PUBLIC_API void FSE_freeDTable(FSE_DTable* dt); /*! FSE_buildDTable(): Builds 'dt', which must be already allocated, using FSE_createDTable(). return : 0, or an errorCode, which can be tested using FSE_isError() */-size_t FSE_buildDTable (FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);+FSE_PUBLIC_API size_t FSE_buildDTable (FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog); /*! FSE_decompress_usingDTable(): Decompress compressed source `cSrc` of size `cSrcSize` using `dt` into `dst` which must be already allocated. @return : size of regenerated data (necessarily <= `dstCapacity`), or an errorCode, which can be tested using FSE_isError() */-size_t FSE_decompress_usingDTable(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, const FSE_DTable* dt);+FSE_PUBLIC_API size_t FSE_decompress_usingDTable(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, const FSE_DTable* dt); /*! Tutorial :@@ -271,8 +294,10 @@ If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small) */ +#endif /* FSE_H */ -#ifdef FSE_STATIC_LINKING_ONLY+#if defined(FSE_STATIC_LINKING_ONLY) && !defined(FSE_H_FSE_STATIC_LINKING_ONLY)+#define FSE_H_FSE_STATIC_LINKING_ONLY /* *** Dependency *** */ #include "bitstream.h"@@ -290,35 +315,14 @@ #define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) (1 + (1<<(maxTableLog-1)) + ((maxSymbolValue+1)*2)) #define FSE_DTABLE_SIZE_U32(maxTableLog) (1 + (1<<maxTableLog)) --/* *****************************************-* FSE advanced API-*******************************************/-/* FSE_count_wksp() :- * Same as FSE_count(), but using an externally provided scratch buffer.- * `workSpace` size must be table of >= `1024` unsigned- */-size_t FSE_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,- const void* source, size_t sourceSize, unsigned* workSpace);--/** FSE_countFast() :- * same as FSE_count(), but blindly trusts that all byte values within src are <= *maxSymbolValuePtr- */-size_t FSE_countFast(unsigned* count, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize);--/* FSE_countFast_wksp() :- * Same as FSE_countFast(), but using an externally provided scratch buffer.- * `workSpace` must be a table of minimum `1024` unsigned- */-size_t FSE_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize, unsigned* workSpace);--/*! FSE_count_simple- * Same as FSE_countFast(), but does not use any additional memory (not even on stack).- * This function is unsafe, and will segfault if any value within `src` is `> *maxSymbolValuePtr` (presuming it's also the size of `count`).-*/-size_t FSE_count_simple(unsigned* count, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize);+/* or use the size to malloc() space directly. Pay attention to alignment restrictions though */+#define FSE_CTABLE_SIZE(maxTableLog, maxSymbolValue) (FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(FSE_CTable))+#define FSE_DTABLE_SIZE(maxTableLog) (FSE_DTABLE_SIZE_U32(maxTableLog) * sizeof(FSE_DTable)) +/* *****************************************+ * FSE advanced API+ ***************************************** */ unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus); /**< same as FSE_optimalTableLog(), which used `minus==2` */@@ -327,7 +331,7 @@ * Same as FSE_compress2(), but using an externally allocated scratch buffer (`workSpace`). * FSE_WKSP_SIZE_U32() provides the minimum size required for `workSpace` as a table of FSE_CTable. */-#define FSE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) ( FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) + (1<<((maxTableLog>2)?(maxTableLog-2):0)) )+#define FSE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) ( FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) + ((maxTableLog > 12) ? (1 << (maxTableLog - 2)) : 1024) ) size_t FSE_compress_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize); size_t FSE_buildCTable_raw (FSE_CTable* ct, unsigned nbBits);@@ -351,6 +355,11 @@ size_t FSE_decompress_wksp(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, FSE_DTable* workSpace, unsigned maxLog); /**< same as FSE_decompress(), using an externally allocated `workSpace` produced with `FSE_DTABLE_SIZE_U32(maxLog)` */ +typedef enum {+ FSE_repeat_none, /**< Cannot use the previous table */+ FSE_repeat_check, /**< Can use the previous table but it must be checked */+ FSE_repeat_valid /**< Can use the previous table and it is assumed to be valid */+ } FSE_repeat; /* ***************************************** * FSE symbol compression API@@ -503,7 +512,7 @@ const U32 tableLog = MEM_read16(ptr); statePtr->value = (ptrdiff_t)1<<tableLog; statePtr->stateTable = u16ptr+2;- statePtr->symbolTT = ((const U32*)ct + 1 + (tableLog ? (1<<(tableLog-1)) : 1));+ statePtr->symbolTT = ct + 1 + (tableLog ? (1<<(tableLog-1)) : 1); statePtr->stateLog = tableLog; } @@ -522,11 +531,11 @@ } } -MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, U32 symbol)+MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, unsigned symbol) {- const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];+ FSE_symbolCompressionTransform const symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol]; const U16* const stateTable = (const U16*)(statePtr->stateTable);- U32 nbBitsOut = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);+ U32 const nbBitsOut = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16); BIT_addBits(bitC, statePtr->value, nbBitsOut); statePtr->value = stateTable[ (statePtr->value >> nbBitsOut) + symbolTT.deltaFindState]; }@@ -538,6 +547,39 @@ } +/* FSE_getMaxNbBits() :+ * Approximate maximum cost of a symbol, in bits.+ * Fractional get rounded up (i.e : a symbol with a normalized frequency of 3 gives the same result as a frequency of 2)+ * note 1 : assume symbolValue is valid (<= maxSymbolValue)+ * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */+MEM_STATIC U32 FSE_getMaxNbBits(const void* symbolTTPtr, U32 symbolValue)+{+ const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;+ return (symbolTT[symbolValue].deltaNbBits + ((1<<16)-1)) >> 16;+}++/* FSE_bitCost() :+ * Approximate symbol cost, as fractional value, using fixed-point format (accuracyLog fractional bits)+ * note 1 : assume symbolValue is valid (<= maxSymbolValue)+ * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */+MEM_STATIC U32 FSE_bitCost(const void* symbolTTPtr, U32 tableLog, U32 symbolValue, U32 accuracyLog)+{+ const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;+ U32 const minNbBits = symbolTT[symbolValue].deltaNbBits >> 16;+ U32 const threshold = (minNbBits+1) << 16;+ assert(tableLog < 16);+ assert(accuracyLog < 31-tableLog); /* ensure enough room for renormalization double shift */+ { U32 const tableSize = 1 << tableLog;+ U32 const deltaFromThreshold = threshold - (symbolTT[symbolValue].deltaNbBits + tableSize);+ U32 const normalizedDeltaFromThreshold = (deltaFromThreshold << accuracyLog) >> tableLog; /* linear interpolation (very approximate) */+ U32 const bitMultiplier = 1 << accuracyLog;+ assert(symbolTT[symbolValue].deltaNbBits + tableSize <= threshold);+ assert(normalizedDeltaFromThreshold <= bitMultiplier);+ return (minNbBits+1)*bitMultiplier - normalizedDeltaFromThreshold;+ }+}++ /* ====== Decompression ====== */ typedef struct {@@ -664,5 +706,3 @@ #if defined (__cplusplus) } #endif--#endif /* FSE_H */
zstd/lib/common/fse_decompress.c view
@@ -34,42 +34,22 @@ /* **************************************************************-* Compiler specifics-****************************************************************/-#ifdef _MSC_VER /* Visual Studio */-# define FORCE_INLINE static __forceinline-# include <intrin.h> /* For Visual 2005 */-# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */-# pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */-#else-# if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */-# ifdef __GNUC__-# define FORCE_INLINE static inline __attribute__((always_inline))-# else-# define FORCE_INLINE static inline-# endif-# else-# define FORCE_INLINE static-# endif /* __STDC_VERSION__ */-#endif---/* ************************************************************** * Includes ****************************************************************/ #include <stdlib.h> /* malloc, free, qsort */ #include <string.h> /* memcpy, memset */-#include <stdio.h> /* printf (debug) */ #include "bitstream.h"+#include "compiler.h" #define FSE_STATIC_LINKING_ONLY #include "fse.h"+#include "error_private.h" /* ************************************************************** * Error Management ****************************************************************/ #define FSE_isError ERR_isError-#define FSE_STATIC_ASSERT(c) { enum { FSE_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */+#define FSE_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c) /* use only *after* variable declarations */ /* check and forward error code */ #define CHECK_F(f) { size_t const e = f; if (FSE_isError(e)) return e; }@@ -159,8 +139,8 @@ { U32 u; for (u=0; u<tableSize; u++) { FSE_FUNCTION_TYPE const symbol = (FSE_FUNCTION_TYPE)(tableDecode[u].symbol);- U16 nextState = symbolNext[symbol]++;- tableDecode[u].nbBits = (BYTE) (tableLog - BIT_highbit32 ((U32)nextState) );+ U32 const nextState = symbolNext[symbol]++;+ tableDecode[u].nbBits = (BYTE) (tableLog - BIT_highbit32(nextState) ); tableDecode[u].newState = (U16) ( (nextState << tableDecode[u].nbBits) - tableSize); } } @@ -217,7 +197,7 @@ return 0; } -FORCE_INLINE size_t FSE_decompress_usingDTable_generic(+FORCE_INLINE_TEMPLATE size_t FSE_decompress_usingDTable_generic( void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const FSE_DTable* dt, const unsigned fast)
zstd/lib/common/huf.h view
@@ -1,7 +1,7 @@ /* ******************************************************************- Huffman coder, part of New Generation Entropy library- header file- Copyright (C) 2013-2016, Yann Collet.+ huff0 huffman codec,+ part of Finite State Entropy library+ Copyright (C) 2013-present, Yann Collet. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -31,81 +31,114 @@ You can contact the author at : - Source repository : https://github.com/Cyan4973/FiniteStateEntropy ****************************************************************** */-#ifndef HUF_H_298734234-#define HUF_H_298734234 #if defined (__cplusplus) extern "C" { #endif +#ifndef HUF_H_298734234+#define HUF_H_298734234 /* *** Dependencies *** */ #include <stddef.h> /* size_t */ -/* *** simple functions *** */-/**-HUF_compress() :- Compress content from buffer 'src', of size 'srcSize', into buffer 'dst'.- 'dst' buffer must be already allocated.- Compression runs faster if `dstCapacity` >= HUF_compressBound(srcSize).- `srcSize` must be <= `HUF_BLOCKSIZE_MAX` == 128 KB.- @return : size of compressed data (<= `dstCapacity`).- Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!- if return == 1, srcData is a single repeated byte symbol (RLE compression).- if HUF_isError(return), compression failed (more details using HUF_getErrorName())-*/-size_t HUF_compress(void* dst, size_t dstCapacity,- const void* src, size_t srcSize);+/* *** library symbols visibility *** */+/* Note : when linking with -fvisibility=hidden on gcc, or by default on Visual,+ * HUF symbols remain "private" (internal symbols for library only).+ * Set macro FSE_DLL_EXPORT to 1 if you want HUF symbols visible on DLL interface */+#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)+# define HUF_PUBLIC_API __attribute__ ((visibility ("default")))+#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) /* Visual expected */+# define HUF_PUBLIC_API __declspec(dllexport)+#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)+# define HUF_PUBLIC_API __declspec(dllimport) /* not required, just to generate faster code (saves a function pointer load from IAT and an indirect jump) */+#else+# define HUF_PUBLIC_API+#endif -/**-HUF_decompress() :- Decompress HUF data from buffer 'cSrc', of size 'cSrcSize',- into already allocated buffer 'dst', of minimum size 'dstSize'.- `originalSize` : **must** be the ***exact*** size of original (uncompressed) data.- Note : in contrast with FSE, HUF_decompress can regenerate- RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data,- because it knows size to regenerate.- @return : size of regenerated data (== originalSize),- or an error code, which can be tested using HUF_isError()-*/-size_t HUF_decompress(void* dst, size_t originalSize,- const void* cSrc, size_t cSrcSize); +/* ========================== */+/* *** simple functions *** */+/* ========================== */ +/** HUF_compress() :+ * Compress content from buffer 'src', of size 'srcSize', into buffer 'dst'.+ * 'dst' buffer must be already allocated.+ * Compression runs faster if `dstCapacity` >= HUF_compressBound(srcSize).+ * `srcSize` must be <= `HUF_BLOCKSIZE_MAX` == 128 KB.+ * @return : size of compressed data (<= `dstCapacity`).+ * Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!+ * if HUF_isError(return), compression failed (more details using HUF_getErrorName())+ */+HUF_PUBLIC_API size_t HUF_compress(void* dst, size_t dstCapacity,+ const void* src, size_t srcSize);++/** HUF_decompress() :+ * Decompress HUF data from buffer 'cSrc', of size 'cSrcSize',+ * into already allocated buffer 'dst', of minimum size 'dstSize'.+ * `originalSize` : **must** be the ***exact*** size of original (uncompressed) data.+ * Note : in contrast with FSE, HUF_decompress can regenerate+ * RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data,+ * because it knows size to regenerate (originalSize).+ * @return : size of regenerated data (== originalSize),+ * or an error code, which can be tested using HUF_isError()+ */+HUF_PUBLIC_API size_t HUF_decompress(void* dst, size_t originalSize,+ const void* cSrc, size_t cSrcSize);++ /* *** Tool functions *** */-#define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */-size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */+#define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */+HUF_PUBLIC_API size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */ /* Error Management */-unsigned HUF_isError(size_t code); /**< tells if a return value is an error code */-const char* HUF_getErrorName(size_t code); /**< provides error code string (useful for debugging) */+HUF_PUBLIC_API unsigned HUF_isError(size_t code); /**< tells if a return value is an error code */+HUF_PUBLIC_API const char* HUF_getErrorName(size_t code); /**< provides error code string (useful for debugging) */ /* *** Advanced function *** */ /** HUF_compress2() :- * Same as HUF_compress(), but offers direct control over `maxSymbolValue` and `tableLog` .- * `tableLog` must be `<= HUF_TABLELOG_MAX` . */-size_t HUF_compress2 (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);+ * Same as HUF_compress(), but offers control over `maxSymbolValue` and `tableLog`.+ * `maxSymbolValue` must be <= HUF_SYMBOLVALUE_MAX .+ * `tableLog` must be `<= HUF_TABLELOG_MAX` . */+HUF_PUBLIC_API size_t HUF_compress2 (void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned tableLog); /** HUF_compress4X_wksp() :-* Same as HUF_compress2(), but uses externally allocated `workSpace`, which must be a table of >= 1024 unsigned */-size_t HUF_compress4X_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize); /**< `workSpace` must be a table of at least 1024 unsigned */-+ * Same as HUF_compress2(), but uses externally allocated `workSpace`.+ * `workspace` must have minimum alignment of 4, and be at least as large as HUF_WORKSPACE_SIZE */+#define HUF_WORKSPACE_SIZE (6 << 10)+#define HUF_WORKSPACE_SIZE_U32 (HUF_WORKSPACE_SIZE / sizeof(U32))+HUF_PUBLIC_API size_t HUF_compress4X_wksp (void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned tableLog,+ void* workSpace, size_t wkspSize); +#endif /* HUF_H_298734234 */ -#ifdef HUF_STATIC_LINKING_ONLY+/* ******************************************************************+ * WARNING !!+ * The following section contains advanced and experimental definitions+ * which shall never be used in the context of a dynamic library,+ * because they are not guaranteed to remain stable in the future.+ * Only consider them in association with static linking.+ * *****************************************************************/+#if defined(HUF_STATIC_LINKING_ONLY) && !defined(HUF_H_HUF_STATIC_LINKING_ONLY)+#define HUF_H_HUF_STATIC_LINKING_ONLY /* *** Dependencies *** */ #include "mem.h" /* U32 */ /* *** Constants *** */-#define HUF_TABLELOG_ABSOLUTEMAX 15 /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */-#define HUF_TABLELOG_MAX 12 /* max configured tableLog (for static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */-#define HUF_TABLELOG_DEFAULT 11 /* tableLog by default, when not specified */-#define HUF_SYMBOLVALUE_MAX 255+#define HUF_TABLELOG_MAX 12 /* max runtime value of tableLog (due to static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */+#define HUF_TABLELOG_DEFAULT 11 /* default tableLog value when none specified */+#define HUF_SYMBOLVALUE_MAX 255++#define HUF_TABLELOG_ABSOLUTEMAX 15 /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */ #if (HUF_TABLELOG_MAX > HUF_TABLELOG_ABSOLUTEMAX) # error "HUF_TABLELOG_MAX is too large !" #endif@@ -116,123 +149,210 @@ ******************************************/ /* HUF buffer bounds */ #define HUF_CTABLEBOUND 129-#define HUF_BLOCKBOUND(size) (size + (size>>8) + 8) /* only true if incompressible pre-filtered with fast heuristic */+#define HUF_BLOCKBOUND(size) (size + (size>>8) + 8) /* only true when incompressible is pre-filtered with fast heuristic */ #define HUF_COMPRESSBOUND(size) (HUF_CTABLEBOUND + HUF_BLOCKBOUND(size)) /* Macro version, useful for static allocation */ /* static allocation of HUF's Compression Table */+#define HUF_CTABLE_SIZE_U32(maxSymbolValue) ((maxSymbolValue)+1) /* Use tables of U32, for proper alignment */+#define HUF_CTABLE_SIZE(maxSymbolValue) (HUF_CTABLE_SIZE_U32(maxSymbolValue) * sizeof(U32)) #define HUF_CREATE_STATIC_CTABLE(name, maxSymbolValue) \- U32 name##hb[maxSymbolValue+1]; \+ U32 name##hb[HUF_CTABLE_SIZE_U32(maxSymbolValue)]; \ void* name##hv = &(name##hb); \ HUF_CElt* name = (HUF_CElt*)(name##hv) /* no final ; */ /* static allocation of HUF's DTable */ typedef U32 HUF_DTable; #define HUF_DTABLE_SIZE(maxTableLog) (1 + (1<<(maxTableLog)))-#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) \+#define HUF_CREATE_STATIC_DTABLEX1(DTable, maxTableLog) \ HUF_DTable DTable[HUF_DTABLE_SIZE((maxTableLog)-1)] = { ((U32)((maxTableLog)-1) * 0x01000001) }-#define HUF_CREATE_STATIC_DTABLEX4(DTable, maxTableLog) \+#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) \ HUF_DTable DTable[HUF_DTABLE_SIZE(maxTableLog)] = { ((U32)(maxTableLog) * 0x01000001) } /* **************************************** * Advanced decompression functions ******************************************/-size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */-size_t HUF_decompress4X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */+size_t HUF_decompress4X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */+#ifndef HUF_FORCE_DECOMPRESS_X1+size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */+#endif size_t HUF_decompress4X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< decodes RLE and uncompressed */ size_t HUF_decompress4X_hufOnly(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< considers RLE and uncompressed as errors */-size_t HUF_decompress4X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */-size_t HUF_decompress4X4_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */+size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< considers RLE and uncompressed as errors */+size_t HUF_decompress4X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */+size_t HUF_decompress4X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< single-symbol decoder */+#ifndef HUF_FORCE_DECOMPRESS_X1+size_t HUF_decompress4X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< double-symbols decoder */+#endif /* ****************************************-* HUF detailed API-******************************************/-/*!-HUF_compress() does the following:-1. count symbol occurrence from source[] into table count[] using FSE_count()-2. (optional) refine tableLog using HUF_optimalTableLog()-3. build Huffman table from count using HUF_buildCTable()-4. save Huffman table to memory buffer using HUF_writeCTable()-5. encode the data stream using HUF_compress4X_usingCTable()+ * HUF detailed API+ * ****************************************/ -The following API allows targeting specific sub-functions for advanced tasks.-For example, it's possible to compress several blocks using the same 'CTable',-or to save and regenerate 'CTable' using external methods.-*/-/* FSE_count() : find it within "fse.h" */+/*! HUF_compress() does the following:+ * 1. count symbol occurrence from source[] into table count[] using FSE_count() (exposed within "fse.h")+ * 2. (optional) refine tableLog using HUF_optimalTableLog()+ * 3. build Huffman table from count using HUF_buildCTable()+ * 4. save Huffman table to memory buffer using HUF_writeCTable()+ * 5. encode the data stream using HUF_compress4X_usingCTable()+ *+ * The following API allows targeting specific sub-functions for advanced tasks.+ * For example, it's possible to compress several blocks using the same 'CTable',+ * or to save and regenerate 'CTable' using external methods.+ */ unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue); typedef struct HUF_CElt_s HUF_CElt; /* incomplete type */-size_t HUF_buildCTable (HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue, unsigned maxNbBits);+size_t HUF_buildCTable (HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue, unsigned maxNbBits); /* @return : maxNbBits; CTable and count can overlap. In which case, CTable will overwrite count content */ size_t HUF_writeCTable (void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog); size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable); +typedef enum {+ HUF_repeat_none, /**< Cannot use the previous table */+ HUF_repeat_check, /**< Can use the previous table but it must be checked. Note : The previous table must have been constructed by HUF_compress{1, 4}X_repeat */+ HUF_repeat_valid /**< Can use the previous table and it is assumed to be valid */+ } HUF_repeat;+/** HUF_compress4X_repeat() :+ * Same as HUF_compress4X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.+ * If it uses hufTable it does not modify hufTable or repeat.+ * If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.+ * If preferRepeat then the old table will always be used if valid. */+size_t HUF_compress4X_repeat(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned tableLog,+ void* workSpace, size_t wkspSize, /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */+ HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2); /** HUF_buildCTable_wksp() : * Same as HUF_buildCTable(), but using externally allocated scratch buffer.- * `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as a table of 1024 unsigned.+ * `workSpace` must be aligned on 4-bytes boundaries, and its size must be >= HUF_CTABLE_WORKSPACE_SIZE. */-size_t HUF_buildCTable_wksp (HUF_CElt* tree, const U32* count, U32 maxSymbolValue, U32 maxNbBits, void* workSpace, size_t wkspSize);+#define HUF_CTABLE_WORKSPACE_SIZE_U32 (2*HUF_SYMBOLVALUE_MAX +1 +1)+#define HUF_CTABLE_WORKSPACE_SIZE (HUF_CTABLE_WORKSPACE_SIZE_U32 * sizeof(unsigned))+size_t HUF_buildCTable_wksp (HUF_CElt* tree,+ const unsigned* count, U32 maxSymbolValue, U32 maxNbBits,+ void* workSpace, size_t wkspSize); /*! HUF_readStats() :- Read compact Huffman tree, saved by HUF_writeCTable().- `huffWeight` is destination buffer.- @return : size read from `src` , or an error Code .- Note : Needed by HUF_readCTable() and HUF_readDTableXn() . */-size_t HUF_readStats(BYTE* huffWeight, size_t hwSize, U32* rankStats,- U32* nbSymbolsPtr, U32* tableLogPtr,+ * Read compact Huffman tree, saved by HUF_writeCTable().+ * `huffWeight` is destination buffer.+ * @return : size read from `src` , or an error Code .+ * Note : Needed by HUF_readCTable() and HUF_readDTableXn() . */+size_t HUF_readStats(BYTE* huffWeight, size_t hwSize,+ U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr, const void* src, size_t srcSize); /** HUF_readCTable() :-* Loading a CTable saved with HUF_writeCTable() */-size_t HUF_readCTable (HUF_CElt* CTable, unsigned maxSymbolValue, const void* src, size_t srcSize);+ * Loading a CTable saved with HUF_writeCTable() */+size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize); +/** HUF_getNbBits() :+ * Read nbBits from CTable symbolTable, for symbol `symbolValue` presumed <= HUF_SYMBOLVALUE_MAX+ * Note 1 : is not inlined, as HUF_CElt definition is private+ * Note 2 : const void* used, so that it can provide a statically allocated table as argument (which uses type U32) */+U32 HUF_getNbBits(const void* symbolTable, U32 symbolValue); /*-HUF_decompress() does the following:-1. select the decompression algorithm (X2, X4) based on pre-computed heuristics-2. build Huffman table from save, using HUF_readDTableXn()-3. decode 1 or 4 segments in parallel using HUF_decompressSXn_usingDTable-*/+ * HUF_decompress() does the following:+ * 1. select the decompression algorithm (X1, X2) based on pre-computed heuristics+ * 2. build Huffman table from save, using HUF_readDTableX?()+ * 3. decode 1 or 4 segments in parallel using HUF_decompress?X?_usingDTable()+ */ /** HUF_selectDecoder() :-* Tells which decoder is likely to decode faster,-* based on a set of pre-determined metrics.-* @return : 0==HUF_decompress4X2, 1==HUF_decompress4X4 .-* Assumption : 0 < cSrcSize < dstSize <= 128 KB */+ * Tells which decoder is likely to decode faster,+ * based on a set of pre-computed metrics.+ * @return : 0==HUF_decompress4X1, 1==HUF_decompress4X2 .+ * Assumption : 0 < dstSize <= 128 KB */ U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize); +/**+ * The minimum workspace size for the `workSpace` used in+ * HUF_readDTableX1_wksp() and HUF_readDTableX2_wksp().+ *+ * The space used depends on HUF_TABLELOG_MAX, ranging from ~1500 bytes when+ * HUF_TABLE_LOG_MAX=12 to ~1850 bytes when HUF_TABLE_LOG_MAX=15.+ * Buffer overflow errors may potentially occur if code modifications result in+ * a required workspace size greater than that specified in the following+ * macro.+ */+#define HUF_DECOMPRESS_WORKSPACE_SIZE (2 << 10)+#define HUF_DECOMPRESS_WORKSPACE_SIZE_U32 (HUF_DECOMPRESS_WORKSPACE_SIZE / sizeof(U32))++#ifndef HUF_FORCE_DECOMPRESS_X2+size_t HUF_readDTableX1 (HUF_DTable* DTable, const void* src, size_t srcSize);+size_t HUF_readDTableX1_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);+#endif+#ifndef HUF_FORCE_DECOMPRESS_X1 size_t HUF_readDTableX2 (HUF_DTable* DTable, const void* src, size_t srcSize);-size_t HUF_readDTableX4 (HUF_DTable* DTable, const void* src, size_t srcSize);+size_t HUF_readDTableX2_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);+#endif size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);+#ifndef HUF_FORCE_DECOMPRESS_X2+size_t HUF_decompress4X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);+#endif+#ifndef HUF_FORCE_DECOMPRESS_X1 size_t HUF_decompress4X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);-size_t HUF_decompress4X4_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);+#endif +/* ====================== */ /* single stream variants */+/* ====================== */ size_t HUF_compress1X (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);-size_t HUF_compress1X_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize); /**< `workSpace` must be a table of at least 1024 unsigned */+size_t HUF_compress1X_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize); /**< `workSpace` must be a table of at least HUF_WORKSPACE_SIZE_U32 unsigned */ size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable);+/** HUF_compress1X_repeat() :+ * Same as HUF_compress1X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.+ * If it uses hufTable it does not modify hufTable or repeat.+ * If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.+ * If preferRepeat then the old table will always be used if valid. */+size_t HUF_compress1X_repeat(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned tableLog,+ void* workSpace, size_t wkspSize, /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */+ HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2); -size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* single-symbol decoder */-size_t HUF_decompress1X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* double-symbol decoder */+size_t HUF_decompress1X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* single-symbol decoder */+#ifndef HUF_FORCE_DECOMPRESS_X1+size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /* double-symbol decoder */+#endif size_t HUF_decompress1X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);-size_t HUF_decompress1X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */-size_t HUF_decompress1X4_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */+size_t HUF_decompress1X_DCtx_wksp (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);+#ifndef HUF_FORCE_DECOMPRESS_X2+size_t HUF_decompress1X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< single-symbol decoder */+size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< single-symbol decoder */+#endif+#ifndef HUF_FORCE_DECOMPRESS_X1+size_t HUF_decompress1X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< double-symbols decoder */+size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< double-symbols decoder */+#endif size_t HUF_decompress1X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable); /**< automatic selection of sing or double symbol decoder, based on DTable */+#ifndef HUF_FORCE_DECOMPRESS_X2+size_t HUF_decompress1X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);+#endif+#ifndef HUF_FORCE_DECOMPRESS_X1 size_t HUF_decompress1X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);-size_t HUF_decompress1X4_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);+#endif -#endif /* HUF_STATIC_LINKING_ONLY */+/* BMI2 variants.+ * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.+ */+size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);+#ifndef HUF_FORCE_DECOMPRESS_X2+size_t HUF_decompress1X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);+#endif+size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);+size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2); +#endif /* HUF_STATIC_LINKING_ONLY */ #if defined (__cplusplus) } #endif--#endif /* HUF_H_298734234 */
zstd/lib/common/mem.h view
@@ -1,10 +1,11 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ #ifndef MEM_H_MODULE@@ -38,6 +39,10 @@ # define MEM_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */ #endif +#ifndef __has_builtin+# define __has_builtin(x) 0 /* compat. with non-clang compilers */+#endif+ /* code only tested on 32 and 64 bits systems */ #define MEM_STATIC_ASSERT(c) { enum { MEM_static_assert = 1/(int)(!!(c)) }; } MEM_STATIC void MEM_check(void) { MEM_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }@@ -48,23 +53,33 @@ *****************************************************************/ #if !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) # include <stdint.h>- typedef uint8_t BYTE;- typedef uint16_t U16;- typedef int16_t S16;- typedef uint32_t U32;- typedef int32_t S32;- typedef uint64_t U64;- typedef int64_t S64;- typedef intptr_t iPtrDiff;+ typedef uint8_t BYTE;+ typedef uint16_t U16;+ typedef int16_t S16;+ typedef uint32_t U32;+ typedef int32_t S32;+ typedef uint64_t U64;+ typedef int64_t S64; #else+# include <limits.h>+#if CHAR_BIT != 8+# error "this implementation requires char to be exactly 8-bit type"+#endif typedef unsigned char BYTE;+#if USHRT_MAX != 65535+# error "this implementation requires short to be exactly 16-bit type"+#endif typedef unsigned short U16; typedef signed short S16;+#if UINT_MAX != 4294967295+# error "this implementation requires int to be exactly 32-bit type"+#endif typedef unsigned int U32; typedef signed int S32;+/* note : there are no limits defined for long long type in C90.+ * limits exist in C99, however, in such case, <stdint.h> is preferred */ typedef unsigned long long U64; typedef signed long long S64;- typedef ptrdiff_t iPtrDiff; #endif @@ -76,19 +91,18 @@ * Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal. * The below switch allow to select different access method for improved performance. * Method 0 (default) : use `memcpy()`. Safe and portable.- * Method 1 : `__packed` statement. It depends on compiler extension (ie, not portable).+ * Method 1 : `__packed` statement. It depends on compiler extension (i.e., not portable). * This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`. * Method 2 : direct access. This method is portable but violate C standard. * It can generate buggy code on targets depending on alignment.- * In some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)+ * In some circumstances, it's the only known way to get the most performance (i.e. GCC + ARMv6) * See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details. * Prefer these methods in priority order (0 > 1 > 2) */ #ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ # if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) ) # define MEM_FORCE_MEMORY_ACCESS 2-# elif defined(__INTEL_COMPILER) /*|| defined(_MSC_VER)*/ || \- (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) ))+# elif defined(__INTEL_COMPILER) || defined(__GNUC__) # define MEM_FORCE_MEMORY_ACCESS 1 # endif #endif@@ -109,7 +123,7 @@ MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; } MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; } MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }-MEM_STATIC U64 MEM_readST(const void* memPtr) { return *(const size_t*) memPtr; }+MEM_STATIC size_t MEM_readST(const void* memPtr) { return *(const size_t*) memPtr; } MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; } MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }@@ -120,21 +134,27 @@ /* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */ /* currently only defined for gcc and icc */ #if defined(_MSC_VER) || (defined(__INTEL_COMPILER) && defined(WIN32))- __pragma( pack(push, 1) )- typedef union { U16 u16; U32 u32; U64 u64; size_t st; } unalign;+ __pragma( pack(push, 1) )+ typedef struct { U16 v; } unalign16;+ typedef struct { U32 v; } unalign32;+ typedef struct { U64 v; } unalign64;+ typedef struct { size_t v; } unalignArch; __pragma( pack(pop) ) #else- typedef union { U16 u16; U32 u32; U64 u64; size_t st; } __attribute__((packed)) unalign;+ typedef struct { U16 v; } __attribute__((packed)) unalign16;+ typedef struct { U32 v; } __attribute__((packed)) unalign32;+ typedef struct { U64 v; } __attribute__((packed)) unalign64;+ typedef struct { size_t v; } __attribute__((packed)) unalignArch; #endif -MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }-MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }-MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }-MEM_STATIC U64 MEM_readST(const void* ptr) { return ((const unalign*)ptr)->st; }+MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign16*)ptr)->v; }+MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign32*)ptr)->v; }+MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign64*)ptr)->v; }+MEM_STATIC size_t MEM_readST(const void* ptr) { return ((const unalignArch*)ptr)->v; } -MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }-MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign*)memPtr)->u32 = value; }-MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign*)memPtr)->u64 = value; }+MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign16*)memPtr)->v = value; }+MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign32*)memPtr)->v = value; }+MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign64*)memPtr)->v = value; } #else @@ -182,7 +202,8 @@ { #if defined(_MSC_VER) /* Visual Studio */ return _byteswap_ulong(in);-#elif defined (__GNUC__)+#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \+ || (defined(__clang__) && __has_builtin(__builtin_bswap32)) return __builtin_bswap32(in); #else return ((in << 24) & 0xff000000 ) |@@ -196,7 +217,8 @@ { #if defined(_MSC_VER) /* Visual Studio */ return _byteswap_uint64(in);-#elif defined (__GNUC__)+#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \+ || (defined(__clang__) && __has_builtin(__builtin_bswap64)) return __builtin_bswap64(in); #else return ((in << 56) & 0xff00000000000000ULL) |@@ -350,20 +372,6 @@ MEM_writeBE64(memPtr, (U64)val); } --/* function safe only for comparisons */-MEM_STATIC U32 MEM_readMINMATCH(const void* memPtr, U32 length)-{- switch (length)- {- default :- case 4 : return MEM_read32(memPtr);- case 3 : if (MEM_isLittleEndian())- return MEM_read32(memPtr)<<8;- else- return MEM_read32(memPtr)>>8;- }-} #if defined (__cplusplus) }
+ zstd/lib/common/pool.c view
@@ -0,0 +1,340 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+++/* ====== Dependencies ======= */+#include <stddef.h> /* size_t */+#include "debug.h" /* assert */+#include "zstd_internal.h" /* ZSTD_malloc, ZSTD_free */+#include "pool.h"++/* ====== Compiler specifics ====== */+#if defined(_MSC_VER)+# pragma warning(disable : 4204) /* disable: C4204: non-constant aggregate initializer */+#endif+++#ifdef ZSTD_MULTITHREAD++#include "threading.h" /* pthread adaptation */++/* A job is a function and an opaque argument */+typedef struct POOL_job_s {+ POOL_function function;+ void *opaque;+} POOL_job;++struct POOL_ctx_s {+ ZSTD_customMem customMem;+ /* Keep track of the threads */+ ZSTD_pthread_t* threads;+ size_t threadCapacity;+ size_t threadLimit;++ /* The queue is a circular buffer */+ POOL_job *queue;+ size_t queueHead;+ size_t queueTail;+ size_t queueSize;++ /* The number of threads working on jobs */+ size_t numThreadsBusy;+ /* Indicates if the queue is empty */+ int queueEmpty;++ /* The mutex protects the queue */+ ZSTD_pthread_mutex_t queueMutex;+ /* Condition variable for pushers to wait on when the queue is full */+ ZSTD_pthread_cond_t queuePushCond;+ /* Condition variables for poppers to wait on when the queue is empty */+ ZSTD_pthread_cond_t queuePopCond;+ /* Indicates if the queue is shutting down */+ int shutdown;+};++/* POOL_thread() :+ * Work thread for the thread pool.+ * Waits for jobs and executes them.+ * @returns : NULL on failure else non-null.+ */+static void* POOL_thread(void* opaque) {+ POOL_ctx* const ctx = (POOL_ctx*)opaque;+ if (!ctx) { return NULL; }+ for (;;) {+ /* Lock the mutex and wait for a non-empty queue or until shutdown */+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);++ while ( ctx->queueEmpty+ || (ctx->numThreadsBusy >= ctx->threadLimit) ) {+ if (ctx->shutdown) {+ /* even if !queueEmpty, (possible if numThreadsBusy >= threadLimit),+ * a few threads will be shutdown while !queueEmpty,+ * but enough threads will remain active to finish the queue */+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+ return opaque;+ }+ ZSTD_pthread_cond_wait(&ctx->queuePopCond, &ctx->queueMutex);+ }+ /* Pop a job off the queue */+ { POOL_job const job = ctx->queue[ctx->queueHead];+ ctx->queueHead = (ctx->queueHead + 1) % ctx->queueSize;+ ctx->numThreadsBusy++;+ ctx->queueEmpty = ctx->queueHead == ctx->queueTail;+ /* Unlock the mutex, signal a pusher, and run the job */+ ZSTD_pthread_cond_signal(&ctx->queuePushCond);+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);++ job.function(job.opaque);++ /* If the intended queue size was 0, signal after finishing job */+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);+ ctx->numThreadsBusy--;+ if (ctx->queueSize == 1) {+ ZSTD_pthread_cond_signal(&ctx->queuePushCond);+ }+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+ }+ } /* for (;;) */+ assert(0); /* Unreachable */+}++POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {+ return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);+}++POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,+ ZSTD_customMem customMem) {+ POOL_ctx* ctx;+ /* Check parameters */+ if (!numThreads) { return NULL; }+ /* Allocate the context and zero initialize */+ ctx = (POOL_ctx*)ZSTD_calloc(sizeof(POOL_ctx), customMem);+ if (!ctx) { return NULL; }+ /* Initialize the job queue.+ * It needs one extra space since one space is wasted to differentiate+ * empty and full queues.+ */+ ctx->queueSize = queueSize + 1;+ ctx->queue = (POOL_job*)ZSTD_malloc(ctx->queueSize * sizeof(POOL_job), customMem);+ ctx->queueHead = 0;+ ctx->queueTail = 0;+ ctx->numThreadsBusy = 0;+ ctx->queueEmpty = 1;+ (void)ZSTD_pthread_mutex_init(&ctx->queueMutex, NULL);+ (void)ZSTD_pthread_cond_init(&ctx->queuePushCond, NULL);+ (void)ZSTD_pthread_cond_init(&ctx->queuePopCond, NULL);+ ctx->shutdown = 0;+ /* Allocate space for the thread handles */+ ctx->threads = (ZSTD_pthread_t*)ZSTD_malloc(numThreads * sizeof(ZSTD_pthread_t), customMem);+ ctx->threadCapacity = 0;+ ctx->customMem = customMem;+ /* Check for errors */+ if (!ctx->threads || !ctx->queue) { POOL_free(ctx); return NULL; }+ /* Initialize the threads */+ { size_t i;+ for (i = 0; i < numThreads; ++i) {+ if (ZSTD_pthread_create(&ctx->threads[i], NULL, &POOL_thread, ctx)) {+ ctx->threadCapacity = i;+ POOL_free(ctx);+ return NULL;+ } }+ ctx->threadCapacity = numThreads;+ ctx->threadLimit = numThreads;+ }+ return ctx;+}++/*! POOL_join() :+ Shutdown the queue, wake any sleeping threads, and join all of the threads.+*/+static void POOL_join(POOL_ctx* ctx) {+ /* Shut down the queue */+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);+ ctx->shutdown = 1;+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+ /* Wake up sleeping threads */+ ZSTD_pthread_cond_broadcast(&ctx->queuePushCond);+ ZSTD_pthread_cond_broadcast(&ctx->queuePopCond);+ /* Join all of the threads */+ { size_t i;+ for (i = 0; i < ctx->threadCapacity; ++i) {+ ZSTD_pthread_join(ctx->threads[i], NULL); /* note : could fail */+ } }+}++void POOL_free(POOL_ctx *ctx) {+ if (!ctx) { return; }+ POOL_join(ctx);+ ZSTD_pthread_mutex_destroy(&ctx->queueMutex);+ ZSTD_pthread_cond_destroy(&ctx->queuePushCond);+ ZSTD_pthread_cond_destroy(&ctx->queuePopCond);+ ZSTD_free(ctx->queue, ctx->customMem);+ ZSTD_free(ctx->threads, ctx->customMem);+ ZSTD_free(ctx, ctx->customMem);+}++++size_t POOL_sizeof(POOL_ctx *ctx) {+ if (ctx==NULL) return 0; /* supports sizeof NULL */+ return sizeof(*ctx)+ + ctx->queueSize * sizeof(POOL_job)+ + ctx->threadCapacity * sizeof(ZSTD_pthread_t);+}+++/* @return : 0 on success, 1 on error */+static int POOL_resize_internal(POOL_ctx* ctx, size_t numThreads)+{+ if (numThreads <= ctx->threadCapacity) {+ if (!numThreads) return 1;+ ctx->threadLimit = numThreads;+ return 0;+ }+ /* numThreads > threadCapacity */+ { ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_malloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);+ if (!threadPool) return 1;+ /* replace existing thread pool */+ memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));+ ZSTD_free(ctx->threads, ctx->customMem);+ ctx->threads = threadPool;+ /* Initialize additional threads */+ { size_t threadId;+ for (threadId = ctx->threadCapacity; threadId < numThreads; ++threadId) {+ if (ZSTD_pthread_create(&threadPool[threadId], NULL, &POOL_thread, ctx)) {+ ctx->threadCapacity = threadId;+ return 1;+ } }+ } }+ /* successfully expanded */+ ctx->threadCapacity = numThreads;+ ctx->threadLimit = numThreads;+ return 0;+}++/* @return : 0 on success, 1 on error */+int POOL_resize(POOL_ctx* ctx, size_t numThreads)+{+ int result;+ if (ctx==NULL) return 1;+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);+ result = POOL_resize_internal(ctx, numThreads);+ ZSTD_pthread_cond_broadcast(&ctx->queuePopCond);+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+ return result;+}++/**+ * Returns 1 if the queue is full and 0 otherwise.+ *+ * When queueSize is 1 (pool was created with an intended queueSize of 0),+ * then a queue is empty if there is a thread free _and_ no job is waiting.+ */+static int isQueueFull(POOL_ctx const* ctx) {+ if (ctx->queueSize > 1) {+ return ctx->queueHead == ((ctx->queueTail + 1) % ctx->queueSize);+ } else {+ return (ctx->numThreadsBusy == ctx->threadLimit) ||+ !ctx->queueEmpty;+ }+}+++static void POOL_add_internal(POOL_ctx* ctx, POOL_function function, void *opaque)+{+ POOL_job const job = {function, opaque};+ assert(ctx != NULL);+ if (ctx->shutdown) return;++ ctx->queueEmpty = 0;+ ctx->queue[ctx->queueTail] = job;+ ctx->queueTail = (ctx->queueTail + 1) % ctx->queueSize;+ ZSTD_pthread_cond_signal(&ctx->queuePopCond);+}++void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque)+{+ assert(ctx != NULL);+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);+ /* Wait until there is space in the queue for the new job */+ while (isQueueFull(ctx) && (!ctx->shutdown)) {+ ZSTD_pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);+ }+ POOL_add_internal(ctx, function, opaque);+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+}+++int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque)+{+ assert(ctx != NULL);+ ZSTD_pthread_mutex_lock(&ctx->queueMutex);+ if (isQueueFull(ctx)) {+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+ return 0;+ }+ POOL_add_internal(ctx, function, opaque);+ ZSTD_pthread_mutex_unlock(&ctx->queueMutex);+ return 1;+}+++#else /* ZSTD_MULTITHREAD not defined */++/* ========================== */+/* No multi-threading support */+/* ========================== */+++/* We don't need any data, but if it is empty, malloc() might return NULL. */+struct POOL_ctx_s {+ int dummy;+};+static POOL_ctx g_ctx;++POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {+ return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);+}++POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize, ZSTD_customMem customMem) {+ (void)numThreads;+ (void)queueSize;+ (void)customMem;+ return &g_ctx;+}++void POOL_free(POOL_ctx* ctx) {+ assert(!ctx || ctx == &g_ctx);+ (void)ctx;+}++int POOL_resize(POOL_ctx* ctx, size_t numThreads) {+ (void)ctx; (void)numThreads;+ return 0;+}++void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque) {+ (void)ctx;+ function(opaque);+}++int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque) {+ (void)ctx;+ function(opaque);+ return 1;+}++size_t POOL_sizeof(POOL_ctx* ctx) {+ if (ctx==NULL) return 0; /* supports sizeof NULL */+ assert(ctx == &g_ctx);+ return sizeof(*ctx);+}++#endif /* ZSTD_MULTITHREAD */
+ zstd/lib/common/pool.h view
@@ -0,0 +1,84 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#ifndef POOL_H+#define POOL_H++#if defined (__cplusplus)+extern "C" {+#endif+++#include <stddef.h> /* size_t */+#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_customMem */+#include "zstd.h"++typedef struct POOL_ctx_s POOL_ctx;++/*! POOL_create() :+ * Create a thread pool with at most `numThreads` threads.+ * `numThreads` must be at least 1.+ * The maximum number of queued jobs before blocking is `queueSize`.+ * @return : POOL_ctx pointer on success, else NULL.+*/+POOL_ctx* POOL_create(size_t numThreads, size_t queueSize);++POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,+ ZSTD_customMem customMem);++/*! POOL_free() :+ * Free a thread pool returned by POOL_create().+ */+void POOL_free(POOL_ctx* ctx);++/*! POOL_resize() :+ * Expands or shrinks pool's number of threads.+ * This is more efficient than releasing + creating a new context,+ * since it tries to preserve and re-use existing threads.+ * `numThreads` must be at least 1.+ * @return : 0 when resize was successful,+ * !0 (typically 1) if there is an error.+ * note : only numThreads can be resized, queueSize remains unchanged.+ */+int POOL_resize(POOL_ctx* ctx, size_t numThreads);++/*! POOL_sizeof() :+ * @return threadpool memory usage+ * note : compatible with NULL (returns 0 in this case)+ */+size_t POOL_sizeof(POOL_ctx* ctx);++/*! POOL_function :+ * The function type that can be added to a thread pool.+ */+typedef void (*POOL_function)(void*);++/*! POOL_add() :+ * Add the job `function(opaque)` to the thread pool. `ctx` must be valid.+ * Possibly blocks until there is room in the queue.+ * Note : The function may be executed asynchronously,+ * therefore, `opaque` must live until function has been completed.+ */+void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque);+++/*! POOL_tryAdd() :+ * Add the job `function(opaque)` to thread pool _if_ a worker is available.+ * Returns immediately even if not (does not block).+ * @return : 1 if successful, 0 if not.+ */+int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque);+++#if defined (__cplusplus)+}+#endif++#endif
+ zstd/lib/common/threading.c view
@@ -0,0 +1,75 @@+/**+ * Copyright (c) 2016 Tino Reichardt+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ *+ * You can contact the author at:+ * - zstdmt source repository: https://github.com/mcmilk/zstdmt+ */++/**+ * This file will hold wrapper for systems, which do not support pthreads+ */++/* create fake symbol to avoid empty translation unit warning */+int g_ZSTD_threading_useless_symbol;++#if defined(ZSTD_MULTITHREAD) && defined(_WIN32)++/**+ * Windows minimalist Pthread Wrapper, based on :+ * http://www.cse.wustl.edu/~schmidt/win32-cv-1.html+ */+++/* === Dependencies === */+#include <process.h>+#include <errno.h>+#include "threading.h"+++/* === Implementation === */++static unsigned __stdcall worker(void *arg)+{+ ZSTD_pthread_t* const thread = (ZSTD_pthread_t*) arg;+ thread->arg = thread->start_routine(thread->arg);+ return 0;+}++int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,+ void* (*start_routine) (void*), void* arg)+{+ (void)unused;+ thread->arg = arg;+ thread->start_routine = start_routine;+ thread->handle = (HANDLE) _beginthreadex(NULL, 0, worker, thread, 0, NULL);++ if (!thread->handle)+ return errno;+ else+ return 0;+}++int ZSTD_pthread_join(ZSTD_pthread_t thread, void **value_ptr)+{+ DWORD result;++ if (!thread.handle) return 0;++ result = WaitForSingleObject(thread.handle, INFINITE);+ switch (result) {+ case WAIT_OBJECT_0:+ if (value_ptr) *value_ptr = thread.arg;+ return 0;+ case WAIT_ABANDONED:+ return EINVAL;+ default:+ return GetLastError();+ }+}++#endif /* ZSTD_MULTITHREAD */
+ zstd/lib/common/threading.h view
@@ -0,0 +1,123 @@+/**+ * Copyright (c) 2016 Tino Reichardt+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ *+ * You can contact the author at:+ * - zstdmt source repository: https://github.com/mcmilk/zstdmt+ */++#ifndef THREADING_H_938743+#define THREADING_H_938743++#if defined (__cplusplus)+extern "C" {+#endif++#if defined(ZSTD_MULTITHREAD) && defined(_WIN32)++/**+ * Windows minimalist Pthread Wrapper, based on :+ * http://www.cse.wustl.edu/~schmidt/win32-cv-1.html+ */+#ifdef WINVER+# undef WINVER+#endif+#define WINVER 0x0600++#ifdef _WIN32_WINNT+# undef _WIN32_WINNT+#endif+#define _WIN32_WINNT 0x0600++#ifndef WIN32_LEAN_AND_MEAN+# define WIN32_LEAN_AND_MEAN+#endif++#undef ERROR /* reported already defined on VS 2015 (Rich Geldreich) */+#include <windows.h>+#undef ERROR+#define ERROR(name) ZSTD_ERROR(name)+++/* mutex */+#define ZSTD_pthread_mutex_t CRITICAL_SECTION+#define ZSTD_pthread_mutex_init(a, b) ((void)(b), InitializeCriticalSection((a)), 0)+#define ZSTD_pthread_mutex_destroy(a) DeleteCriticalSection((a))+#define ZSTD_pthread_mutex_lock(a) EnterCriticalSection((a))+#define ZSTD_pthread_mutex_unlock(a) LeaveCriticalSection((a))++/* condition variable */+#define ZSTD_pthread_cond_t CONDITION_VARIABLE+#define ZSTD_pthread_cond_init(a, b) ((void)(b), InitializeConditionVariable((a)), 0)+#define ZSTD_pthread_cond_destroy(a) ((void)(a))+#define ZSTD_pthread_cond_wait(a, b) SleepConditionVariableCS((a), (b), INFINITE)+#define ZSTD_pthread_cond_signal(a) WakeConditionVariable((a))+#define ZSTD_pthread_cond_broadcast(a) WakeAllConditionVariable((a))++/* ZSTD_pthread_create() and ZSTD_pthread_join() */+typedef struct {+ HANDLE handle;+ void* (*start_routine)(void*);+ void* arg;+} ZSTD_pthread_t;++int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,+ void* (*start_routine) (void*), void* arg);++int ZSTD_pthread_join(ZSTD_pthread_t thread, void** value_ptr);++/**+ * add here more wrappers as required+ */+++#elif defined(ZSTD_MULTITHREAD) /* posix assumed ; need a better detection method */+/* === POSIX Systems === */+# include <pthread.h>++#define ZSTD_pthread_mutex_t pthread_mutex_t+#define ZSTD_pthread_mutex_init(a, b) pthread_mutex_init((a), (b))+#define ZSTD_pthread_mutex_destroy(a) pthread_mutex_destroy((a))+#define ZSTD_pthread_mutex_lock(a) pthread_mutex_lock((a))+#define ZSTD_pthread_mutex_unlock(a) pthread_mutex_unlock((a))++#define ZSTD_pthread_cond_t pthread_cond_t+#define ZSTD_pthread_cond_init(a, b) pthread_cond_init((a), (b))+#define ZSTD_pthread_cond_destroy(a) pthread_cond_destroy((a))+#define ZSTD_pthread_cond_wait(a, b) pthread_cond_wait((a), (b))+#define ZSTD_pthread_cond_signal(a) pthread_cond_signal((a))+#define ZSTD_pthread_cond_broadcast(a) pthread_cond_broadcast((a))++#define ZSTD_pthread_t pthread_t+#define ZSTD_pthread_create(a, b, c, d) pthread_create((a), (b), (c), (d))+#define ZSTD_pthread_join(a, b) pthread_join((a),(b))++#else /* ZSTD_MULTITHREAD not defined */+/* No multithreading support */++typedef int ZSTD_pthread_mutex_t;+#define ZSTD_pthread_mutex_init(a, b) ((void)(a), (void)(b), 0)+#define ZSTD_pthread_mutex_destroy(a) ((void)(a))+#define ZSTD_pthread_mutex_lock(a) ((void)(a))+#define ZSTD_pthread_mutex_unlock(a) ((void)(a))++typedef int ZSTD_pthread_cond_t;+#define ZSTD_pthread_cond_init(a, b) ((void)(a), (void)(b), 0)+#define ZSTD_pthread_cond_destroy(a) ((void)(a))+#define ZSTD_pthread_cond_wait(a, b) ((void)(a), (void)(b))+#define ZSTD_pthread_cond_signal(a) ((void)(a))+#define ZSTD_pthread_cond_broadcast(a) ((void)(a))++/* do not use ZSTD_pthread_t */++#endif /* ZSTD_MULTITHREAD */++#if defined (__cplusplus)+}+#endif++#endif /* THREADING_H_938743 */
zstd/lib/common/xxhash.c view
@@ -66,10 +66,10 @@ /* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */ /*!XXH_FORCE_NATIVE_FORMAT :- * By default, xxHash library provides endian-independant Hash values, based on little-endian convention.+ * By default, xxHash library provides endian-independent Hash values, based on little-endian convention. * Results are therefore identical for little-endian and big-endian CPU. * This comes at a performance cost for big-endian CPU, since some swapping is required to emulate little-endian format.- * Should endian-independance be of no importance for your application, you may set the #define below to 1,+ * Should endian-independence be of no importance for your application, you may set the #define below to 1, * to improve speed for Big-endian CPU. * This option has no impact on Little_Endian CPU. */@@ -98,35 +98,44 @@ /* Modify the local functions below should you wish to use some other memory routines */ /* for malloc(), free() */ #include <stdlib.h>+#include <stddef.h> /* size_t */ static void* XXH_malloc(size_t s) { return malloc(s); } static void XXH_free (void* p) { free(p); } /* for memcpy() */ #include <string.h> static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcpy(dest,src,size); } -#define XXH_STATIC_LINKING_ONLY+#ifndef XXH_STATIC_LINKING_ONLY+# define XXH_STATIC_LINKING_ONLY+#endif #include "xxhash.h" /* ************************************* * Compiler Specific Options ***************************************/-#ifdef _MSC_VER /* Visual Studio */-# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */-# define FORCE_INLINE static __forceinline+#if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */+# define INLINE_KEYWORD inline #else-# if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */-# ifdef __GNUC__-# define FORCE_INLINE static inline __attribute__((always_inline))-# else-# define FORCE_INLINE static inline-# endif-# else-# define FORCE_INLINE static-# endif /* __STDC_VERSION__ */+# define INLINE_KEYWORD #endif +#if defined(__GNUC__)+# define FORCE_INLINE_ATTR __attribute__((always_inline))+#elif defined(_MSC_VER)+# define FORCE_INLINE_ATTR __forceinline+#else+# define FORCE_INLINE_ATTR+#endif +#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR+++#ifdef _MSC_VER+# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */+#endif++ /* ************************************* * Basic Types ***************************************/@@ -246,7 +255,7 @@ *****************************/ typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment; -FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align)+FORCE_INLINE_TEMPLATE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align) { if (align==XXH_unaligned) return endian==XXH_littleEndian ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr));@@ -254,7 +263,7 @@ return endian==XXH_littleEndian ? *(const U32*)ptr : XXH_swap32(*(const U32*)ptr); } -FORCE_INLINE U32 XXH_readLE32(const void* ptr, XXH_endianess endian)+FORCE_INLINE_TEMPLATE U32 XXH_readLE32(const void* ptr, XXH_endianess endian) { return XXH_readLE32_align(ptr, endian, XXH_unaligned); }@@ -264,7 +273,7 @@ return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); } -FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align)+FORCE_INLINE_TEMPLATE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align) { if (align==XXH_unaligned) return endian==XXH_littleEndian ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr));@@ -272,7 +281,7 @@ return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr); } -FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian)+FORCE_INLINE_TEMPLATE U64 XXH_readLE64(const void* ptr, XXH_endianess endian) { return XXH_readLE64_align(ptr, endian, XXH_unaligned); }@@ -333,7 +342,7 @@ return seed; } -FORCE_INLINE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, XXH_endianess endian, XXH_alignment align)+FORCE_INLINE_TEMPLATE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, XXH_endianess endian, XXH_alignment align) { const BYTE* p = (const BYTE*)input; const BYTE* bEnd = p + len;@@ -433,7 +442,7 @@ return acc; } -FORCE_INLINE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, XXH_endianess endian, XXH_alignment align)+FORCE_INLINE_TEMPLATE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, XXH_endianess endian, XXH_alignment align) { const BYTE* p = (const BYTE*)input; const BYTE* const bEnd = p + len;@@ -582,7 +591,7 @@ } -FORCE_INLINE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian)+FORCE_INLINE_TEMPLATE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian) { const BYTE* p = (const BYTE*)input; const BYTE* const bEnd = p + len;@@ -652,7 +661,7 @@ -FORCE_INLINE U32 XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian)+FORCE_INLINE_TEMPLATE U32 XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian) { const BYTE * p = (const BYTE*)state->mem32; const BYTE* const bEnd = (const BYTE*)(state->mem32) + state->memsize;@@ -702,7 +711,7 @@ /* **** XXH64 **** */ -FORCE_INLINE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian)+FORCE_INLINE_TEMPLATE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian) { const BYTE* p = (const BYTE*)input; const BYTE* const bEnd = p + len;@@ -769,7 +778,7 @@ -FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian)+FORCE_INLINE_TEMPLATE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian) { const BYTE * p = (const BYTE*)state->mem64; const BYTE* const bEnd = (const BYTE*)state->mem64 + state->memsize;
zstd/lib/common/xxhash.h view
@@ -64,16 +64,12 @@ XXH32 6.8 GB/s 6.0 GB/s */ -#ifndef XXHASH_H_5627135585666179-#define XXHASH_H_5627135585666179 1- #if defined (__cplusplus) extern "C" { #endif -#ifndef XXH_NAMESPACE-# define XXH_NAMESPACE ZSTD_ /* Zstandard specific */-#endif+#ifndef XXHASH_H_5627135585666179+#define XXHASH_H_5627135585666179 1 /* ****************************@@ -242,6 +238,11 @@ /* ************************** * Canonical representation ****************************/+/* Default result type for XXH functions are primitive unsigned 32 and 64 bits.+* The canonical representation uses human-readable write convention, aka big-endian (large digits first).+* These functions allow transformation of hash result into and from its canonical format.+* This way, hash values can be written into a file / memory, and remain comparable on different systems and programs.+*/ typedef struct { unsigned char digest[4]; } XXH32_canonical_t; typedef struct { unsigned char digest[8]; } XXH64_canonical_t; @@ -251,14 +252,9 @@ XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src); XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src); -/* Default result type for XXH functions are primitive unsigned 32 and 64 bits.-* The canonical representation uses human-readable write convention, aka big-endian (large digits first).-* These functions allow transformation of hash result into and from its canonical format.-* This way, hash values can be written into a file / memory, and remain comparable on different systems and programs.-*/+#endif /* XXHASH_H_5627135585666179 */ -#ifdef XXH_STATIC_LINKING_ONLY /* ================================================================================================ This section contains definitions which are not guaranteed to remain stable.@@ -266,6 +262,8 @@ They shall only be used with static linking. Never use these definitions in association with dynamic linking ! =================================================================================================== */+#if defined(XXH_STATIC_LINKING_ONLY) && !defined(XXH_STATIC_H_3543687687345)+#define XXH_STATIC_H_3543687687345 /* These definitions are only meant to allow allocation of XXH state statically, on stack, or in a struct for example.@@ -299,11 +297,9 @@ # include "xxhash.c" /* include xxhash functions as `static`, for inlining */ # endif -#endif /* XXH_STATIC_LINKING_ONLY */+#endif /* XXH_STATIC_LINKING_ONLY && XXH_STATIC_H_3543687687345 */ #if defined (__cplusplus) } #endif--#endif /* XXHASH_H_5627135585666179 */
zstd/lib/common/zstd_common.c view
@@ -1,10 +1,11 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ @@ -12,62 +13,71 @@ /*-************************************* * Dependencies ***************************************/-#include <stdlib.h> /* malloc */+#include <stdlib.h> /* malloc, calloc, free */+#include <string.h> /* memset */ #include "error_private.h"-#define ZSTD_STATIC_LINKING_ONLY-#include "zstd.h" /* declaration of ZSTD_isError, ZSTD_getErrorName, ZSTD_getErrorCode, ZSTD_getErrorString, ZSTD_versionNumber */+#include "zstd_internal.h" /*-**************************************** * Version ******************************************/-unsigned ZSTD_versionNumber (void) { return ZSTD_VERSION_NUMBER; }+unsigned ZSTD_versionNumber(void) { return ZSTD_VERSION_NUMBER; } +const char* ZSTD_versionString(void) { return ZSTD_VERSION_STRING; } + /*-**************************************** * ZSTD Error Management ******************************************/+#undef ZSTD_isError /* defined within zstd_internal.h */ /*! ZSTD_isError() :-* tells if a return value is an error code */+ * tells if a return value is an error code+ * symbol is required for external callers */ unsigned ZSTD_isError(size_t code) { return ERR_isError(code); } /*! ZSTD_getErrorName() :-* provides error code string from function result (useful for debugging) */+ * provides error code string from function result (useful for debugging) */ const char* ZSTD_getErrorName(size_t code) { return ERR_getErrorName(code); } /*! ZSTD_getError() :-* convert a `size_t` function result into a proper ZSTD_errorCode enum */+ * convert a `size_t` function result into a proper ZSTD_errorCode enum */ ZSTD_ErrorCode ZSTD_getErrorCode(size_t code) { return ERR_getErrorCode(code); } /*! ZSTD_getErrorString() :-* provides error code string from enum */-const char* ZSTD_getErrorString(ZSTD_ErrorCode code) { return ERR_getErrorName(code); }+ * provides error code string from enum */+const char* ZSTD_getErrorString(ZSTD_ErrorCode code) { return ERR_getErrorString(code); } + /*=************************************************************** * Custom allocator ****************************************************************/-/* default uses stdlib */-void* ZSTD_defaultAllocFunction(void* opaque, size_t size)-{- void* address = malloc(size);- (void)opaque;- return address;-}--void ZSTD_defaultFreeFunction(void* opaque, void* address)+void* ZSTD_malloc(size_t size, ZSTD_customMem customMem) {- (void)opaque;- free(address);+ if (customMem.customAlloc)+ return customMem.customAlloc(customMem.opaque, size);+ return malloc(size); } -void* ZSTD_malloc(size_t size, ZSTD_customMem customMem)+void* ZSTD_calloc(size_t size, ZSTD_customMem customMem) {- return customMem.customAlloc(customMem.opaque, size);+ if (customMem.customAlloc) {+ /* calloc implemented as malloc+memset;+ * not as efficient as calloc, but next best guess for custom malloc */+ void* const ptr = customMem.customAlloc(customMem.opaque, size);+ memset(ptr, 0, size);+ return ptr;+ }+ return calloc(1, size); } void ZSTD_free(void* ptr, ZSTD_customMem customMem) {- if (ptr!=NULL)- customMem.customFree(customMem.opaque, ptr);+ if (ptr!=NULL) {+ if (customMem.customFree)+ customMem.customFree(customMem.opaque, ptr);+ else+ free(ptr);+ } }
zstd/lib/common/zstd_errors.h view
@@ -1,10 +1,11 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ #ifndef ZSTD_ERRORS_H_398273423@@ -19,10 +20,12 @@ /* ===== ZSTDERRORLIB_API : control library symbols visibility ===== */-#if defined(__GNUC__) && (__GNUC__ >= 4)-# define ZSTDERRORLIB_VISIBILITY __attribute__ ((visibility ("default")))-#else-# define ZSTDERRORLIB_VISIBILITY+#ifndef ZSTDERRORLIB_VISIBILITY+# if defined(__GNUC__) && (__GNUC__ >= 4)+# define ZSTDERRORLIB_VISIBILITY __attribute__ ((visibility ("default")))+# else+# define ZSTDERRORLIB_VISIBILITY+# endif #endif #if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1) # define ZSTDERRORLIB_API __declspec(dllexport) ZSTDERRORLIB_VISIBILITY@@ -32,39 +35,55 @@ # define ZSTDERRORLIB_API ZSTDERRORLIB_VISIBILITY #endif -/*-****************************************-* error codes list-******************************************/+/*-*********************************************+ * Error codes list+ *-*********************************************+ * Error codes _values_ are pinned down since v1.3.1 only.+ * Therefore, don't rely on values if you may link to any version < v1.3.1.+ *+ * Only values < 100 are considered stable.+ *+ * note 1 : this API shall be used with static linking only.+ * dynamic linking is not yet officially supported.+ * note 2 : Prefer relying on the enum than on its value whenever possible+ * This is the only supported way to use the error list < v1.3.1+ * note 3 : ZSTD_isError() is always correct, whatever the library version.+ **********************************************/ typedef enum {- ZSTD_error_no_error,- ZSTD_error_GENERIC,- ZSTD_error_prefix_unknown,- ZSTD_error_version_unsupported,- ZSTD_error_parameter_unknown,- ZSTD_error_frameParameter_unsupported,- ZSTD_error_frameParameter_unsupportedBy32bits,- ZSTD_error_frameParameter_windowTooLarge,- ZSTD_error_compressionParameter_unsupported,- ZSTD_error_init_missing,- ZSTD_error_memory_allocation,- ZSTD_error_stage_wrong,- ZSTD_error_dstSize_tooSmall,- ZSTD_error_srcSize_wrong,- ZSTD_error_corruption_detected,- ZSTD_error_checksum_wrong,- ZSTD_error_tableLog_tooLarge,- ZSTD_error_maxSymbolValue_tooLarge,- ZSTD_error_maxSymbolValue_tooSmall,- ZSTD_error_dictionary_corrupted,- ZSTD_error_dictionary_wrong,- ZSTD_error_maxCode+ ZSTD_error_no_error = 0,+ ZSTD_error_GENERIC = 1,+ ZSTD_error_prefix_unknown = 10,+ ZSTD_error_version_unsupported = 12,+ ZSTD_error_frameParameter_unsupported = 14,+ ZSTD_error_frameParameter_windowTooLarge = 16,+ ZSTD_error_corruption_detected = 20,+ ZSTD_error_checksum_wrong = 22,+ ZSTD_error_dictionary_corrupted = 30,+ ZSTD_error_dictionary_wrong = 32,+ ZSTD_error_dictionaryCreation_failed = 34,+ ZSTD_error_parameter_unsupported = 40,+ ZSTD_error_parameter_outOfBound = 42,+ ZSTD_error_tableLog_tooLarge = 44,+ ZSTD_error_maxSymbolValue_tooLarge = 46,+ ZSTD_error_maxSymbolValue_tooSmall = 48,+ ZSTD_error_stage_wrong = 60,+ ZSTD_error_init_missing = 62,+ ZSTD_error_memory_allocation = 64,+ ZSTD_error_workSpace_tooSmall= 66,+ ZSTD_error_dstSize_tooSmall = 70,+ ZSTD_error_srcSize_wrong = 72,+ ZSTD_error_dstBuffer_null = 74,+ /* following error codes are __NOT STABLE__, they can be removed or changed in future versions */+ ZSTD_error_frameIndex_tooLarge = 100,+ ZSTD_error_seekableIO = 102,+ ZSTD_error_maxCode = 120 /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */ } ZSTD_ErrorCode; /*! ZSTD_getErrorCode() : convert a `size_t` function result into a `ZSTD_ErrorCode` enum type,- which can be used to compare directly with enum list published into "error_public.h" */+ which can be used to compare with enum list published above */ ZSTDERRORLIB_API ZSTD_ErrorCode ZSTD_getErrorCode(size_t functionResult);-ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code);+ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code); /**< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */ #if defined (__cplusplus)
zstd/lib/common/zstd_internal.h view
@@ -1,75 +1,111 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ #ifndef ZSTD_CCOMMON_H_MODULE #define ZSTD_CCOMMON_H_MODULE -/*-*******************************************************-* Compiler specifics-*********************************************************/-#ifdef _MSC_VER /* Visual Studio */-# define FORCE_INLINE static __forceinline-# include <intrin.h> /* For Visual 2005 */-# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */-# pragma warning(disable : 4324) /* disable: C4324: padded structure */-# pragma warning(disable : 4100) /* disable: C4100: unreferenced formal parameter */-#else-# if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */-# ifdef __GNUC__-# define FORCE_INLINE static inline __attribute__((always_inline))-# else-# define FORCE_INLINE static inline-# endif-# else-# define FORCE_INLINE static-# endif /* __STDC_VERSION__ */-#endif--#ifdef _MSC_VER-# define FORCE_NOINLINE static __declspec(noinline)-#else-# ifdef __GNUC__-# define FORCE_NOINLINE static __attribute__((__noinline__))-# else-# define FORCE_NOINLINE static-# endif-#endif-+/* this module contains definitions which must be identical+ * across compression, decompression and dictBuilder.+ * It also contains a few functions useful to at least 2 of them+ * and which benefit from being inlined */ /*-************************************* * Dependencies ***************************************/+#include "compiler.h" #include "mem.h"+#include "debug.h" /* assert, DEBUGLOG, RAWLOG, g_debuglevel */ #include "error_private.h" #define ZSTD_STATIC_LINKING_ONLY #include "zstd.h"+#define FSE_STATIC_LINKING_ONLY+#include "fse.h"+#define HUF_STATIC_LINKING_ONLY+#include "huf.h"+#ifndef XXH_STATIC_LINKING_ONLY+# define XXH_STATIC_LINKING_ONLY /* XXH64_state_t */+#endif+#include "xxhash.h" /* XXH_reset, update, digest */ +#if defined (__cplusplus)+extern "C" {+#endif++/* ---- static assert (debug) --- */+#define ZSTD_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)+#define ZSTD_isError ERR_isError /* for inlining */+#define FSE_isError ERR_isError+#define HUF_isError ERR_isError++ /*-************************************* * shared macros ***************************************/+#undef MIN+#undef MAX #define MIN(a,b) ((a)<(b) ? (a) : (b)) #define MAX(a,b) ((a)>(b) ? (a) : (b))-#define CHECK_F(f) { size_t const errcod = f; if (ERR_isError(errcod)) return errcod; } /* check and Forward error code */-#define CHECK_E(f, e) { size_t const errcod = f; if (ERR_isError(errcod)) return ERROR(e); } /* check and send Error code */ +/**+ * Return the specified error if the condition evaluates to true.+ *+ * In debug modes, prints additional information. In order to do that+ * (particularly, printing the conditional that failed), this can't just wrap+ * RETURN_ERROR().+ */+#define RETURN_ERROR_IF(cond, err, ...) \+ if (cond) { \+ RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s", __FILE__, __LINE__, ZSTD_QUOTE(cond), ZSTD_QUOTE(ERROR(err))); \+ RAWLOG(3, ": " __VA_ARGS__); \+ RAWLOG(3, "\n"); \+ return ERROR(err); \+ } +/**+ * Unconditionally return the specified error.+ *+ * In debug modes, prints additional information.+ */+#define RETURN_ERROR(err, ...) \+ do { \+ RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", __FILE__, __LINE__, ZSTD_QUOTE(ERROR(err))); \+ RAWLOG(3, ": " __VA_ARGS__); \+ RAWLOG(3, "\n"); \+ return ERROR(err); \+ } while(0);++/**+ * If the provided expression evaluates to an error code, returns that error code.+ *+ * In debug modes, prints additional information.+ */+#define FORWARD_IF_ERROR(err, ...) \+ do { \+ size_t const err_code = (err); \+ if (ERR_isError(err_code)) { \+ RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s", __FILE__, __LINE__, ZSTD_QUOTE(err), ERR_getErrorName(err_code)); \+ RAWLOG(3, ": " __VA_ARGS__); \+ RAWLOG(3, "\n"); \+ return err_code; \+ } \+ } while(0);++ /*-************************************* * Common constants ***************************************/ #define ZSTD_OPT_NUM (1<<12)-#define ZSTD_DICT_MAGIC 0xEC30A437 /* v0.7+ */ #define ZSTD_REP_NUM 3 /* number of repcodes */-#define ZSTD_REP_CHECK (ZSTD_REP_NUM) /* number of repcodes to check by the optimal parser */ #define ZSTD_REP_MOVE (ZSTD_REP_NUM-1)-#define ZSTD_REP_MOVE_OPT (ZSTD_REP_NUM) static const U32 repStartValue[ZSTD_REP_NUM] = { 1, 4, 8 }; #define KB *(1 <<10)@@ -87,6 +123,8 @@ static const size_t ZSTD_fcs_fieldSize[4] = { 0, 2, 4, 8 }; static const size_t ZSTD_did_fieldSize[4] = { 0, 1, 2, 4 }; +#define ZSTD_FRAMEIDSIZE 4 /* magic number size */+ #define ZSTD_BLOCKHEADERSIZE 3 /* C standard doesn't allow `static const` variable to be init using another `static const` variable */ static const size_t ZSTD_blockHeaderSize = ZSTD_BLOCKHEADERSIZE; typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e;@@ -100,40 +138,53 @@ #define LONGNBSEQ 0x7F00 #define MINMATCH 3-#define EQUAL_READ32 4 #define Litbits 8 #define MaxLit ((1<<Litbits) - 1)-#define MaxML 52-#define MaxLL 35-#define MaxOff 28+#define MaxML 52+#define MaxLL 35+#define DefaultMaxOff 28+#define MaxOff 31 #define MaxSeq MAX(MaxLL, MaxML) /* Assumption : MaxOff < MaxLL,MaxML */ #define MLFSELog 9 #define LLFSELog 9 #define OffFSELog 8+#define MaxFSELog MAX(MAX(MLFSELog, LLFSELog), OffFSELog) -static const U32 LL_bits[MaxLL+1] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,- 1, 1, 1, 1, 2, 2, 3, 3, 4, 6, 7, 8, 9,10,11,12,+static const U32 LL_bits[MaxLL+1] = { 0, 0, 0, 0, 0, 0, 0, 0,+ 0, 0, 0, 0, 0, 0, 0, 0,+ 1, 1, 1, 1, 2, 2, 3, 3,+ 4, 6, 7, 8, 9,10,11,12, 13,14,15,16 };-static const S16 LL_defaultNorm[MaxLL+1] = { 4, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1,- 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 2, 1, 1, 1, 1, 1,+static const S16 LL_defaultNorm[MaxLL+1] = { 4, 3, 2, 2, 2, 2, 2, 2,+ 2, 2, 2, 2, 2, 1, 1, 1,+ 2, 2, 2, 2, 2, 2, 2, 2,+ 2, 3, 2, 1, 1, 1, 1, 1, -1,-1,-1,-1 }; #define LL_DEFAULTNORMLOG 6 /* for static allocation */ static const U32 LL_defaultNormLog = LL_DEFAULTNORMLOG; -static const U32 ML_bits[MaxML+1] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,- 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,- 1, 1, 1, 1, 2, 2, 3, 3, 4, 4, 5, 7, 8, 9,10,11,+static const U32 ML_bits[MaxML+1] = { 0, 0, 0, 0, 0, 0, 0, 0,+ 0, 0, 0, 0, 0, 0, 0, 0,+ 0, 0, 0, 0, 0, 0, 0, 0,+ 0, 0, 0, 0, 0, 0, 0, 0,+ 1, 1, 1, 1, 2, 2, 3, 3,+ 4, 4, 5, 7, 8, 9,10,11, 12,13,14,15,16 };-static const S16 ML_defaultNorm[MaxML+1] = { 1, 4, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1,- 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,- 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,-1,-1,+static const S16 ML_defaultNorm[MaxML+1] = { 1, 4, 3, 2, 2, 2, 2, 2,+ 2, 1, 1, 1, 1, 1, 1, 1,+ 1, 1, 1, 1, 1, 1, 1, 1,+ 1, 1, 1, 1, 1, 1, 1, 1,+ 1, 1, 1, 1, 1, 1, 1, 1,+ 1, 1, 1, 1, 1, 1,-1,-1, -1,-1,-1,-1,-1 }; #define ML_DEFAULTNORMLOG 6 /* for static allocation */ static const U32 ML_defaultNormLog = ML_DEFAULTNORMLOG; -static const S16 OF_defaultNorm[MaxOff+1] = { 1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1,- 1, 1, 1, 1, 1, 1, 1, 1,-1,-1,-1,-1,-1 };+static const S16 OF_defaultNorm[DefaultMaxOff+1] = { 1, 1, 1, 1, 1, 1, 2, 2,+ 2, 1, 1, 1, 1, 1, 1, 1,+ 1, 1, 1, 1, 1, 1, 1, 1,+ -1,-1,-1,-1,-1 }; #define OF_DEFAULTNORMLOG 5 /* for static allocation */ static const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG; @@ -145,7 +196,7 @@ #define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; } /*! ZSTD_wildcopy() :-* custom version of memcpy(), can copy up to 7 bytes too many (8 bytes if length==0) */+ * custom version of memcpy(), can overwrite up to WILDCOPY_OVERLENGTH bytes (if length==0) */ #define WILDCOPY_OVERLENGTH 8 MEM_STATIC void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length) {@@ -169,31 +220,14 @@ /*-*******************************************-* Private interfaces+* Private declarations *********************************************/-typedef struct ZSTD_stats_s ZSTD_stats_t;--typedef struct {- U32 off;- U32 len;-} ZSTD_match_t;--typedef struct {- U32 price;- U32 off;- U32 mlen;- U32 litlen;- U32 rep[ZSTD_REP_NUM];-} ZSTD_optimal_t;-- typedef struct seqDef_s { U32 offset; U16 litLength; U16 matchLength; } seqDef; - typedef struct { seqDef* sequencesStart; seqDef* sequences;@@ -202,69 +236,84 @@ BYTE* llCode; BYTE* mlCode; BYTE* ofCode;+ size_t maxNbSeq;+ size_t maxNbLit; U32 longLengthID; /* 0 == no longLength; 1 == Lit.longLength; 2 == Match.longLength; */ U32 longLengthPos;- /* opt */- ZSTD_optimal_t* priceTable;- ZSTD_match_t* matchTable;- U32* matchLengthFreq;- U32* litLengthFreq;- U32* litFreq;- U32* offCodeFreq;- U32 matchLengthSum;- U32 matchSum;- U32 litLengthSum;- U32 litSum;- U32 offCodeSum;- U32 log2matchLengthSum;- U32 log2matchSum;- U32 log2litLengthSum;- U32 log2litSum;- U32 log2offCodeSum;- U32 factor;- U32 staticPrices;- U32 cachedPrice;- U32 cachedLitLength;- const BYTE* cachedLiterals; } seqStore_t; -const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx);-void ZSTD_seqToCodes(const seqStore_t* seqStorePtr);-int ZSTD_isSkipFrame(ZSTD_DCtx* dctx);+/**+ * Contains the compressed frame size and an upper-bound for the decompressed frame size.+ * Note: before using `compressedSize`, check for errors using ZSTD_isError().+ * similarly, before using `decompressedBound`, check for errors using:+ * `decompressedBound != ZSTD_CONTENTSIZE_ERROR`+ */+typedef struct {+ size_t compressedSize;+ unsigned long long decompressedBound;+} ZSTD_frameSizeInfo; /* decompress & legacy */ +const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx); /* compress & dictBuilder */+void ZSTD_seqToCodes(const seqStore_t* seqStorePtr); /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */+ /* custom memory allocation functions */-void* ZSTD_defaultAllocFunction(void* opaque, size_t size);-void ZSTD_defaultFreeFunction(void* opaque, void* address);-#ifndef ZSTD_DLL_IMPORT-static const ZSTD_customMem defaultCustomMem = { ZSTD_defaultAllocFunction, ZSTD_defaultFreeFunction, NULL };-#endif void* ZSTD_malloc(size_t size, ZSTD_customMem customMem);+void* ZSTD_calloc(size_t size, ZSTD_customMem customMem); void ZSTD_free(void* ptr, ZSTD_customMem customMem); -/*====== common function ======*/--MEM_STATIC U32 ZSTD_highbit32(U32 val)+MEM_STATIC U32 ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus */ {+ assert(val != 0);+ { # if defined(_MSC_VER) /* Visual */- unsigned long r=0;- _BitScanReverse(&r, val);- return (unsigned)r;+ unsigned long r=0;+ _BitScanReverse(&r, val);+ return (unsigned)r; # elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */- return 31 - __builtin_clz(val);+ return 31 - __builtin_clz(val); # else /* Software version */- static const int DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };- U32 v = val;- int r;- v |= v >> 1;- v |= v >> 2;- v |= v >> 4;- v |= v >> 8;- v |= v >> 16;- r = DeBruijnClz[(U32)(v * 0x07C4ACDDU) >> 27];- return r;+ static const U32 DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };+ U32 v = val;+ v |= v >> 1;+ v |= v >> 2;+ v |= v >> 4;+ v |= v >> 8;+ v |= v >> 16;+ return DeBruijnClz[(v * 0x07C4ACDDU) >> 27]; # endif+ } } ++/* ZSTD_invalidateRepCodes() :+ * ensures next compression will not use repcodes from previous block.+ * Note : only works with regular variant;+ * do not use with extDict variant ! */+void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx); /* zstdmt, adaptive_compression (shouldn't get this definition from here) */+++typedef struct {+ blockType_e blockType;+ U32 lastBlock;+ U32 origSize;+} blockProperties_t; /* declared here for decompress and fullbench */++/*! ZSTD_getcBlockSize() :+ * Provides the size of compressed block from block header `src` */+/* Used by: decompress, fullbench (does not get its definition from here) */+size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,+ blockProperties_t* bpPtr);++/*! ZSTD_decodeSeqHeaders() :+ * decode sequence header from src */+/* Used by: decompress, fullbench (does not get its definition from here) */+size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,+ const void* src, size_t srcSize);+++#if defined (__cplusplus)+}+#endif #endif /* ZSTD_CCOMMON_H_MODULE */
zstd/lib/compress/fse_compress.c view
@@ -1,6 +1,6 @@ /* ****************************************************************** FSE : Finite State Entropy encoder- Copyright (C) 2013-2015, Yann Collet.+ Copyright (C) 2013-present, Yann Collet. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -33,41 +33,24 @@ ****************************************************************** */ /* **************************************************************-* Compiler specifics-****************************************************************/-#ifdef _MSC_VER /* Visual Studio */-# define FORCE_INLINE static __forceinline-# include <intrin.h> /* For Visual 2005 */-# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */-# pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */-#else-# if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */-# ifdef __GNUC__-# define FORCE_INLINE static inline __attribute__((always_inline))-# else-# define FORCE_INLINE static inline-# endif-# else-# define FORCE_INLINE static-# endif /* __STDC_VERSION__ */-#endif---/* ************************************************************** * Includes ****************************************************************/ #include <stdlib.h> /* malloc, free, qsort */ #include <string.h> /* memcpy, memset */-#include <stdio.h> /* printf (debug) */+#include "compiler.h"+#include "mem.h" /* U32, U16, etc. */+#include "debug.h" /* assert, DEBUGLOG */+#include "hist.h" /* HIST_count_wksp */ #include "bitstream.h" #define FSE_STATIC_LINKING_ONLY #include "fse.h"+#include "error_private.h" /* ************************************************************** * Error Management ****************************************************************/-#define FSE_STATIC_ASSERT(c) { enum { FSE_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */+#define FSE_isError ERR_isError /* **************************************************************@@ -100,7 +83,9 @@ * wkspSize should be sized to handle worst case situation, which is `1<<max_tableLog * sizeof(FSE_FUNCTION_TYPE)` * workSpace must also be properly aligned with FSE_FUNCTION_TYPE requirements */-size_t FSE_buildCTable_wksp(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize)+size_t FSE_buildCTable_wksp(FSE_CTable* ct,+ const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog,+ void* workSpace, size_t wkspSize) { U32 const tableSize = 1 << tableLog; U32 const tableMask = tableSize - 1;@@ -118,14 +103,19 @@ if (((size_t)1 << tableLog) * sizeof(FSE_FUNCTION_TYPE) > wkspSize) return ERROR(tableLog_tooLarge); tableU16[-2] = (U16) tableLog; tableU16[-1] = (U16) maxSymbolValue;+ assert(tableLog < 16); /* required for threshold strategy to work */ /* For explanations on how to distribute symbol values over the table :- * http://fastcompression.blogspot.fr/2014/02/fse-distributing-symbol-values.html */+ * http://fastcompression.blogspot.fr/2014/02/fse-distributing-symbol-values.html */ + #ifdef __clang_analyzer__+ memset(tableSymbol, 0, sizeof(*tableSymbol) * tableSize); /* useless initialization, just to keep scan-build happy */+ #endif+ /* symbol start positions */ { U32 u; cumul[0] = 0;- for (u=1; u<=maxSymbolValue+1; u++) {+ for (u=1; u <= maxSymbolValue+1; u++) { if (normalizedCounter[u-1]==-1) { /* Low proba symbol */ cumul[u] = cumul[u-1] + 1; tableSymbol[highThreshold--] = (FSE_FUNCTION_TYPE)(u-1);@@ -139,14 +129,16 @@ { U32 position = 0; U32 symbol; for (symbol=0; symbol<=maxSymbolValue; symbol++) {- int nbOccurences;- for (nbOccurences=0; nbOccurences<normalizedCounter[symbol]; nbOccurences++) {+ int nbOccurrences;+ int const freq = normalizedCounter[symbol];+ for (nbOccurrences=0; nbOccurrences<freq; nbOccurrences++) { tableSymbol[position] = (FSE_FUNCTION_TYPE)symbol; position = (position + step) & tableMask;- while (position > highThreshold) position = (position + step) & tableMask; /* Low proba area */+ while (position > highThreshold)+ position = (position + step) & tableMask; /* Low proba area */ } } - if (position!=0) return ERROR(GENERIC); /* Must have gone through all positions */+ assert(position==0); /* Must have initialized all positions */ } /* Build table */@@ -161,7 +153,10 @@ for (s=0; s<=maxSymbolValue; s++) { switch (normalizedCounter[s]) {- case 0: break;+ case 0:+ /* filling nonetheless, for compatibility with FSE_getMaxNbBits() */+ symbolTT[s].deltaNbBits = ((tableLog+1) << 16) - (1<<tableLog);+ break; case -1: case 1:@@ -178,6 +173,18 @@ total += normalizedCounter[s]; } } } } +#if 0 /* debug : symbol costs */+ DEBUGLOG(5, "\n --- table statistics : ");+ { U32 symbol;+ for (symbol=0; symbol<=maxSymbolValue; symbol++) {+ DEBUGLOG(5, "%3u: w=%3i, maxBits=%u, fracBits=%.2f",+ symbol, normalizedCounter[symbol],+ FSE_getMaxNbBits(symbolTT, symbol),+ (double)FSE_bitCost(symbolTT, tableLog, symbol, 8) / 256);+ }+ }+#endif+ return 0; } @@ -192,8 +199,9 @@ #ifndef FSE_COMMONDEFS_ONLY + /*-**************************************************************-* FSE NCount encoding-decoding+* FSE NCount encoding ****************************************************************/ size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog) {@@ -201,11 +209,10 @@ return maxSymbolValue ? maxHeaderSize : FSE_NCOUNTBOUND; /* maxSymbolValue==0 ? use default */ } -static short FSE_abs(short a) { return (short)(a<0 ? -a : a); }--static size_t FSE_writeNCount_generic (void* header, size_t headerBufferSize,- const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog,- unsigned writeIsSafe)+static size_t+FSE_writeNCount_generic (void* header, size_t headerBufferSize,+ const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog,+ unsigned writeIsSafe) { BYTE* const ostart = (BYTE*) header; BYTE* out = ostart;@@ -214,13 +221,12 @@ const int tableSize = 1 << tableLog; int remaining; int threshold;- U32 bitStream;- int bitCount;- unsigned charnum = 0;- int previous0 = 0;+ U32 bitStream = 0;+ int bitCount = 0;+ unsigned symbol = 0;+ unsigned const alphabetSize = maxSymbolValue + 1;+ int previousIs0 = 0; - bitStream = 0;- bitCount = 0; /* Table Size */ bitStream += (tableLog-FSE_MIN_TABLELOG) << bitCount; bitCount += 4;@@ -230,48 +236,53 @@ threshold = tableSize; nbBits = tableLog+1; - while (remaining>1) { /* stops at 1 */- if (previous0) {- unsigned start = charnum;- while (!normalizedCounter[charnum]) charnum++;- while (charnum >= start+24) {+ while ((symbol < alphabetSize) && (remaining>1)) { /* stops at 1 */+ if (previousIs0) {+ unsigned start = symbol;+ while ((symbol < alphabetSize) && !normalizedCounter[symbol]) symbol++;+ if (symbol == alphabetSize) break; /* incorrect distribution */+ while (symbol >= start+24) { start+=24; bitStream += 0xFFFFU << bitCount;- if ((!writeIsSafe) && (out > oend-2)) return ERROR(dstSize_tooSmall); /* Buffer overflow */+ if ((!writeIsSafe) && (out > oend-2))+ return ERROR(dstSize_tooSmall); /* Buffer overflow */ out[0] = (BYTE) bitStream; out[1] = (BYTE)(bitStream>>8); out+=2; bitStream>>=16; }- while (charnum >= start+3) {+ while (symbol >= start+3) { start+=3; bitStream += 3 << bitCount; bitCount += 2; }- bitStream += (charnum-start) << bitCount;+ bitStream += (symbol-start) << bitCount; bitCount += 2; if (bitCount>16) {- if ((!writeIsSafe) && (out > oend - 2)) return ERROR(dstSize_tooSmall); /* Buffer overflow */+ if ((!writeIsSafe) && (out > oend - 2))+ return ERROR(dstSize_tooSmall); /* Buffer overflow */ out[0] = (BYTE)bitStream; out[1] = (BYTE)(bitStream>>8); out += 2; bitStream >>= 16; bitCount -= 16; } }- { short count = normalizedCounter[charnum++];- const short max = (short)((2*threshold-1)-remaining);- remaining -= FSE_abs(count);- if (remaining<1) return ERROR(GENERIC);+ { int count = normalizedCounter[symbol++];+ int const max = (2*threshold-1) - remaining;+ remaining -= count < 0 ? -count : count; count++; /* +1 for extra accuracy */- if (count>=threshold) count += max; /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */+ if (count>=threshold)+ count += max; /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */ bitStream += count << bitCount; bitCount += nbBits; bitCount -= (count<max);- previous0 = (count==1);- while (remaining<threshold) nbBits--, threshold>>=1;+ previousIs0 = (count==1);+ if (remaining<1) return ERROR(GENERIC);+ while (remaining<threshold) { nbBits--; threshold>>=1; } } if (bitCount>16) {- if ((!writeIsSafe) && (out > oend - 2)) return ERROR(dstSize_tooSmall); /* Buffer overflow */+ if ((!writeIsSafe) && (out > oend - 2))+ return ERROR(dstSize_tooSmall); /* Buffer overflow */ out[0] = (BYTE)bitStream; out[1] = (BYTE)(bitStream>>8); out += 2;@@ -279,191 +290,37 @@ bitCount -= 16; } } + if (remaining != 1)+ return ERROR(GENERIC); /* incorrect normalized distribution */+ assert(symbol <= alphabetSize);+ /* flush remaining bitStream */- if ((!writeIsSafe) && (out > oend - 2)) return ERROR(dstSize_tooSmall); /* Buffer overflow */+ if ((!writeIsSafe) && (out > oend - 2))+ return ERROR(dstSize_tooSmall); /* Buffer overflow */ out[0] = (BYTE)bitStream; out[1] = (BYTE)(bitStream>>8); out+= (bitCount+7) /8; - if (charnum > maxSymbolValue + 1) return ERROR(GENERIC);- return (out-ostart); } -size_t FSE_writeNCount (void* buffer, size_t bufferSize, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog)+size_t FSE_writeNCount (void* buffer, size_t bufferSize,+ const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog) {- if (tableLog > FSE_MAX_TABLELOG) return ERROR(GENERIC); /* Unsupported */+ if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge); /* Unsupported */ if (tableLog < FSE_MIN_TABLELOG) return ERROR(GENERIC); /* Unsupported */ if (bufferSize < FSE_NCountWriteBound(maxSymbolValue, tableLog)) return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 0); - return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 1);+ return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 1 /* write in buffer is safe */); } - /*-**************************************************************-* Counting histogram-****************************************************************/-/*! FSE_count_simple- This function counts byte values within `src`, and store the histogram into table `count`.- It doesn't use any additional memory.- But this function is unsafe : it doesn't check that all values within `src` can fit into `count`.- For this reason, prefer using a table `count` with 256 elements.- @return : count of most numerous element-*/-size_t FSE_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,- const void* src, size_t srcSize)-{- const BYTE* ip = (const BYTE*)src;- const BYTE* const end = ip + srcSize;- unsigned maxSymbolValue = *maxSymbolValuePtr;- unsigned max=0;-- memset(count, 0, (maxSymbolValue+1)*sizeof(*count));- if (srcSize==0) { *maxSymbolValuePtr = 0; return 0; }-- while (ip<end) count[*ip++]++;-- while (!count[maxSymbolValue]) maxSymbolValue--;- *maxSymbolValuePtr = maxSymbolValue;-- { U32 s; for (s=0; s<=maxSymbolValue; s++) if (count[s] > max) max = count[s]; }-- return (size_t)max;-}---/* FSE_count_parallel_wksp() :- * Same as FSE_count_parallel(), but using an externally provided scratch buffer.- * `workSpace` size must be a minimum of `1024 * sizeof(unsigned)`` */-static size_t FSE_count_parallel_wksp(- unsigned* count, unsigned* maxSymbolValuePtr,- const void* source, size_t sourceSize,- unsigned checkMax, unsigned* const workSpace)-{- const BYTE* ip = (const BYTE*)source;- const BYTE* const iend = ip+sourceSize;- unsigned maxSymbolValue = *maxSymbolValuePtr;- unsigned max=0;- U32* const Counting1 = workSpace;- U32* const Counting2 = Counting1 + 256;- U32* const Counting3 = Counting2 + 256;- U32* const Counting4 = Counting3 + 256;-- memset(Counting1, 0, 4*256*sizeof(unsigned));-- /* safety checks */- if (!sourceSize) {- memset(count, 0, maxSymbolValue + 1);- *maxSymbolValuePtr = 0;- return 0;- }- if (!maxSymbolValue) maxSymbolValue = 255; /* 0 == default */-- /* by stripes of 16 bytes */- { U32 cached = MEM_read32(ip); ip += 4;- while (ip < iend-15) {- U32 c = cached; cached = MEM_read32(ip); ip += 4;- Counting1[(BYTE) c ]++;- Counting2[(BYTE)(c>>8) ]++;- Counting3[(BYTE)(c>>16)]++;- Counting4[ c>>24 ]++;- c = cached; cached = MEM_read32(ip); ip += 4;- Counting1[(BYTE) c ]++;- Counting2[(BYTE)(c>>8) ]++;- Counting3[(BYTE)(c>>16)]++;- Counting4[ c>>24 ]++;- c = cached; cached = MEM_read32(ip); ip += 4;- Counting1[(BYTE) c ]++;- Counting2[(BYTE)(c>>8) ]++;- Counting3[(BYTE)(c>>16)]++;- Counting4[ c>>24 ]++;- c = cached; cached = MEM_read32(ip); ip += 4;- Counting1[(BYTE) c ]++;- Counting2[(BYTE)(c>>8) ]++;- Counting3[(BYTE)(c>>16)]++;- Counting4[ c>>24 ]++;- }- ip-=4;- }-- /* finish last symbols */- while (ip<iend) Counting1[*ip++]++;-- if (checkMax) { /* verify stats will fit into destination table */- U32 s; for (s=255; s>maxSymbolValue; s--) {- Counting1[s] += Counting2[s] + Counting3[s] + Counting4[s];- if (Counting1[s]) return ERROR(maxSymbolValue_tooSmall);- } }-- { U32 s; for (s=0; s<=maxSymbolValue; s++) {- count[s] = Counting1[s] + Counting2[s] + Counting3[s] + Counting4[s];- if (count[s] > max) max = count[s];- } }-- while (!count[maxSymbolValue]) maxSymbolValue--;- *maxSymbolValuePtr = maxSymbolValue;- return (size_t)max;-}--/* FSE_countFast_wksp() :- * Same as FSE_countFast(), but using an externally provided scratch buffer.- * `workSpace` size must be table of >= `1024` unsigned */-size_t FSE_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,- const void* source, size_t sourceSize, unsigned* workSpace)-{- if (sourceSize < 1500) return FSE_count_simple(count, maxSymbolValuePtr, source, sourceSize);- return FSE_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, 0, workSpace);-}--/* fast variant (unsafe : won't check if src contains values beyond count[] limit) */-size_t FSE_countFast(unsigned* count, unsigned* maxSymbolValuePtr,- const void* source, size_t sourceSize)-{- unsigned tmpCounters[1024];- return FSE_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, tmpCounters);-}--/* FSE_count_wksp() :- * Same as FSE_count(), but using an externally provided scratch buffer.- * `workSpace` size must be table of >= `1024` unsigned */-size_t FSE_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,- const void* source, size_t sourceSize, unsigned* workSpace)-{- if (*maxSymbolValuePtr < 255)- return FSE_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, 1, workSpace);- *maxSymbolValuePtr = 255;- return FSE_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace);-}--size_t FSE_count(unsigned* count, unsigned* maxSymbolValuePtr,- const void* src, size_t srcSize)-{- unsigned tmpCounters[1024];- return FSE_count_wksp(count, maxSymbolValuePtr, src, srcSize, tmpCounters);-}----/*-************************************************************** * FSE Compression Code ****************************************************************/-/*! FSE_sizeof_CTable() :- FSE_CTable is a variable size structure which contains :- `U16 tableLog;`- `U16 maxSymbolValue;`- `U16 nextStateNumber[1 << tableLog];` // This size is variable- `FSE_symbolCompressionTransform symbolTT[maxSymbolValue+1];` // This size is variable-Allocation is manual (C standard does not support variable-size structures).-*/-size_t FSE_sizeof_CTable (unsigned maxSymbolValue, unsigned tableLog)-{- if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge);- return FSE_CTABLE_SIZE_U32 (tableLog, maxSymbolValue) * sizeof(U32);-} FSE_CTable* FSE_createCTable (unsigned maxSymbolValue, unsigned tableLog) {@@ -478,20 +335,22 @@ /* provides the minimum logSize to safely represent a distribution */ static unsigned FSE_minTableLog(size_t srcSize, unsigned maxSymbolValue) {- U32 minBitsSrc = BIT_highbit32((U32)(srcSize - 1)) + 1;- U32 minBitsSymbols = BIT_highbit32(maxSymbolValue) + 2;- U32 minBits = minBitsSrc < minBitsSymbols ? minBitsSrc : minBitsSymbols;- return minBits;+ U32 minBitsSrc = BIT_highbit32((U32)(srcSize)) + 1;+ U32 minBitsSymbols = BIT_highbit32(maxSymbolValue) + 2;+ U32 minBits = minBitsSrc < minBitsSymbols ? minBitsSrc : minBitsSymbols;+ assert(srcSize > 1); /* Not supported, RLE should be used instead */+ return minBits; } unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus) {- U32 maxBitsSrc = BIT_highbit32((U32)(srcSize - 1)) - minus;+ U32 maxBitsSrc = BIT_highbit32((U32)(srcSize - 1)) - minus; U32 tableLog = maxTableLog;- U32 minBits = FSE_minTableLog(srcSize, maxSymbolValue);+ U32 minBits = FSE_minTableLog(srcSize, maxSymbolValue);+ assert(srcSize > 1); /* Not supported, RLE should be used instead */ if (tableLog==0) tableLog = FSE_DEFAULT_TABLELOG;- if (maxBitsSrc < tableLog) tableLog = maxBitsSrc; /* Accuracy can be reduced */- if (minBits > tableLog) tableLog = minBits; /* Need a minimum to safely represent all symbol values */+ if (maxBitsSrc < tableLog) tableLog = maxBitsSrc; /* Accuracy can be reduced */+ if (minBits > tableLog) tableLog = minBits; /* Need a minimum to safely represent all symbol values */ if (tableLog < FSE_MIN_TABLELOG) tableLog = FSE_MIN_TABLELOG; if (tableLog > FSE_MAX_TABLELOG) tableLog = FSE_MAX_TABLELOG; return tableLog;@@ -508,6 +367,7 @@ static size_t FSE_normalizeM2(short* norm, U32 tableLog, const unsigned* count, size_t total, U32 maxSymbolValue) {+ short const NOT_YET_ASSIGNED = -2; U32 s; U32 distributed = 0; U32 ToDistribute;@@ -533,15 +393,19 @@ total -= count[s]; continue; }- norm[s]=-2;++ norm[s]=NOT_YET_ASSIGNED; } ToDistribute = (1 << tableLog) - distributed; + if (ToDistribute == 0)+ return 0;+ if ((total / ToDistribute) > lowOne) { /* risk of rounding to zero */ lowOne = (U32)((total * 3) / (ToDistribute * 2)); for (s=0; s<=maxSymbolValue; s++) {- if ((norm[s] == -2) && (count[s] <= lowOne)) {+ if ((norm[s] == NOT_YET_ASSIGNED) && (count[s] <= lowOne)) { norm[s] = 1; distributed++; total -= count[s];@@ -556,17 +420,24 @@ find max, then give all remaining points to max */ U32 maxV = 0, maxC = 0; for (s=0; s<=maxSymbolValue; s++)- if (count[s] > maxC) maxV=s, maxC=count[s];+ if (count[s] > maxC) { maxV=s; maxC=count[s]; } norm[maxV] += (short)ToDistribute; return 0; } + if (total == 0) {+ /* all of the symbols were low enough for the lowOne or lowThreshold */+ for (s=0; ToDistribute > 0; s = (s+1)%(maxSymbolValue+1))+ if (norm[s] > 0) { ToDistribute--; norm[s]++; }+ return 0;+ }+ { U64 const vStepLog = 62 - tableLog; U64 const mid = (1ULL << (vStepLog-1)) - 1; U64 const rStep = ((((U64)1<<vStepLog) * ToDistribute) + mid) / total; /* scale on remaining */ U64 tmpTotal = mid; for (s=0; s<=maxSymbolValue; s++) {- if (norm[s]==-2) {+ if (norm[s]==NOT_YET_ASSIGNED) { U64 const end = tmpTotal + (count[s] * rStep); U32 const sStart = (U32)(tmpTotal >> vStepLog); U32 const sEnd = (U32)(end >> vStepLog);@@ -591,7 +462,7 @@ if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge); /* Unsupported size */ if (tableLog < FSE_minTableLog(total, maxSymbolValue)) return ERROR(GENERIC); /* Too small tableLog, compression potentially impossible */ - { U32 const rtbTable[] = { 0, 473195, 504333, 520860, 550000, 700000, 750000, 830000 };+ { static U32 const rtbTable[] = { 0, 473195, 504333, 520860, 550000, 700000, 750000, 830000 }; U64 const scale = 62 - tableLog; U64 const step = ((U64)1<<62) / total; /* <== here, one division ! */ U64 const vStep = 1ULL<<(scale-20);@@ -613,7 +484,7 @@ U64 restToBeat = vStep * rtbTable[proba]; proba += (count[s]*step) - ((U64)proba<<scale) > restToBeat; }- if (proba > largestP) largestP=proba, largest=s;+ if (proba > largestP) { largestP=proba; largest=s; } normalizedCounter[s] = proba; stillToDistribute -= proba; } }@@ -630,11 +501,11 @@ U32 s; U32 nTotal = 0; for (s=0; s<=maxSymbolValue; s++)- printf("%3i: %4i \n", s, normalizedCounter[s]);+ RAWLOG(2, "%3i: %4i \n", s, normalizedCounter[s]); for (s=0; s<=maxSymbolValue; s++) nTotal += abs(normalizedCounter[s]); if (nTotal != (1U<<tableLog))- printf("Warning !!! Total == %u != %u !!!", nTotal, 1U<<tableLog);+ RAWLOG(2, "Warning !!! Total == %u != %u !!!", nTotal, 1U<<tableLog); getchar(); } #endif@@ -774,7 +645,7 @@ size_t FSE_compressBound(size_t size) { return FSE_COMPRESSBOUND(size); } -#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return f+#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return e #define CHECK_F(f) { CHECK_V_F(_var_err__, f); } /* FSE_compress_wksp() :@@ -787,7 +658,7 @@ BYTE* op = ostart; BYTE* const oend = ostart + dstSize; - U32 count[FSE_MAX_SYMBOL_VALUE+1];+ unsigned count[FSE_MAX_SYMBOL_VALUE+1]; S16 norm[FSE_MAX_SYMBOL_VALUE+1]; FSE_CTable* CTable = (FSE_CTable*)workSpace; size_t const CTableSize = FSE_CTABLE_SIZE_U32(tableLog, maxSymbolValue);@@ -801,7 +672,7 @@ if (!tableLog) tableLog = FSE_DEFAULT_TABLELOG; /* Scan input and build symbol stats */- { CHECK_V_F(maxCount, FSE_count(count, &maxSymbolValue, src, srcSize) );+ { CHECK_V_F(maxCount, HIST_count_wksp(count, &maxSymbolValue, src, srcSize, scratchBuffer, scratchBufferSize) ); if (maxCount == srcSize) return 1; /* only a single symbol in src : rle */ if (maxCount == 1) return 0; /* each symbol present maximum once => not compressible */ if (maxCount < (srcSize >> 7)) return 0; /* Heuristic : not compressible enough */@@ -836,7 +707,7 @@ size_t FSE_compress2 (void* dst, size_t dstCapacity, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog) { fseWkspMax_t scratchBuffer;- FSE_STATIC_ASSERT(sizeof(scratchBuffer) >= FSE_WKSP_SIZE_U32(FSE_MAX_TABLELOG, FSE_MAX_SYMBOL_VALUE)); /* compilation failures here means scratchBuffer is not large enough */+ DEBUG_STATIC_ASSERT(sizeof(scratchBuffer) >= FSE_WKSP_SIZE_U32(FSE_MAX_TABLELOG, FSE_MAX_SYMBOL_VALUE)); /* compilation failures here means scratchBuffer is not large enough */ if (tableLog > FSE_MAX_TABLELOG) return ERROR(tableLog_tooLarge); return FSE_compress_wksp(dst, dstCapacity, src, srcSize, maxSymbolValue, tableLog, &scratchBuffer, sizeof(scratchBuffer)); }
+ zstd/lib/compress/hist.c view
@@ -0,0 +1,203 @@+/* ******************************************************************+ hist : Histogram functions+ part of Finite State Entropy project+ Copyright (C) 2013-present, Yann Collet.++ BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)++ Redistribution and use in source and binary forms, with or without+ modification, are permitted provided that the following conditions are+ met:++ * Redistributions of source code must retain the above copyright+ notice, this list of conditions and the following disclaimer.+ * Redistributions in binary form must reproduce the above+ copyright notice, this list of conditions and the following disclaimer+ in the documentation and/or other materials provided with the+ distribution.++ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS+ "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT+ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR+ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT+ OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,+ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT+ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,+ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY+ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.++ You can contact the author at :+ - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy+ - Public forum : https://groups.google.com/forum/#!forum/lz4c+****************************************************************** */++/* --- dependencies --- */+#include "mem.h" /* U32, BYTE, etc. */+#include "debug.h" /* assert, DEBUGLOG */+#include "error_private.h" /* ERROR */+#include "hist.h"+++/* --- Error management --- */+unsigned HIST_isError(size_t code) { return ERR_isError(code); }++/*-**************************************************************+ * Histogram functions+ ****************************************************************/+unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize)+{+ const BYTE* ip = (const BYTE*)src;+ const BYTE* const end = ip + srcSize;+ unsigned maxSymbolValue = *maxSymbolValuePtr;+ unsigned largestCount=0;++ memset(count, 0, (maxSymbolValue+1) * sizeof(*count));+ if (srcSize==0) { *maxSymbolValuePtr = 0; return 0; }++ while (ip<end) {+ assert(*ip <= maxSymbolValue);+ count[*ip++]++;+ }++ while (!count[maxSymbolValue]) maxSymbolValue--;+ *maxSymbolValuePtr = maxSymbolValue;++ { U32 s;+ for (s=0; s<=maxSymbolValue; s++)+ if (count[s] > largestCount) largestCount = count[s];+ }++ return largestCount;+}++typedef enum { trustInput, checkMaxSymbolValue } HIST_checkInput_e;++/* HIST_count_parallel_wksp() :+ * store histogram into 4 intermediate tables, recombined at the end.+ * this design makes better use of OoO cpus,+ * and is noticeably faster when some values are heavily repeated.+ * But it needs some additional workspace for intermediate tables.+ * `workSpace` size must be a table of size >= HIST_WKSP_SIZE_U32.+ * @return : largest histogram frequency,+ * or an error code (notably when histogram would be larger than *maxSymbolValuePtr). */+static size_t HIST_count_parallel_wksp(+ unsigned* count, unsigned* maxSymbolValuePtr,+ const void* source, size_t sourceSize,+ HIST_checkInput_e check,+ U32* const workSpace)+{+ const BYTE* ip = (const BYTE*)source;+ const BYTE* const iend = ip+sourceSize;+ unsigned maxSymbolValue = *maxSymbolValuePtr;+ unsigned max=0;+ U32* const Counting1 = workSpace;+ U32* const Counting2 = Counting1 + 256;+ U32* const Counting3 = Counting2 + 256;+ U32* const Counting4 = Counting3 + 256;++ memset(workSpace, 0, 4*256*sizeof(unsigned));++ /* safety checks */+ if (!sourceSize) {+ memset(count, 0, maxSymbolValue + 1);+ *maxSymbolValuePtr = 0;+ return 0;+ }+ if (!maxSymbolValue) maxSymbolValue = 255; /* 0 == default */++ /* by stripes of 16 bytes */+ { U32 cached = MEM_read32(ip); ip += 4;+ while (ip < iend-15) {+ U32 c = cached; cached = MEM_read32(ip); ip += 4;+ Counting1[(BYTE) c ]++;+ Counting2[(BYTE)(c>>8) ]++;+ Counting3[(BYTE)(c>>16)]++;+ Counting4[ c>>24 ]++;+ c = cached; cached = MEM_read32(ip); ip += 4;+ Counting1[(BYTE) c ]++;+ Counting2[(BYTE)(c>>8) ]++;+ Counting3[(BYTE)(c>>16)]++;+ Counting4[ c>>24 ]++;+ c = cached; cached = MEM_read32(ip); ip += 4;+ Counting1[(BYTE) c ]++;+ Counting2[(BYTE)(c>>8) ]++;+ Counting3[(BYTE)(c>>16)]++;+ Counting4[ c>>24 ]++;+ c = cached; cached = MEM_read32(ip); ip += 4;+ Counting1[(BYTE) c ]++;+ Counting2[(BYTE)(c>>8) ]++;+ Counting3[(BYTE)(c>>16)]++;+ Counting4[ c>>24 ]++;+ }+ ip-=4;+ }++ /* finish last symbols */+ while (ip<iend) Counting1[*ip++]++;++ if (check) { /* verify stats will fit into destination table */+ U32 s; for (s=255; s>maxSymbolValue; s--) {+ Counting1[s] += Counting2[s] + Counting3[s] + Counting4[s];+ if (Counting1[s]) return ERROR(maxSymbolValue_tooSmall);+ } }++ { U32 s;+ if (maxSymbolValue > 255) maxSymbolValue = 255;+ for (s=0; s<=maxSymbolValue; s++) {+ count[s] = Counting1[s] + Counting2[s] + Counting3[s] + Counting4[s];+ if (count[s] > max) max = count[s];+ } }++ while (!count[maxSymbolValue]) maxSymbolValue--;+ *maxSymbolValuePtr = maxSymbolValue;+ return (size_t)max;+}++/* HIST_countFast_wksp() :+ * Same as HIST_countFast(), but using an externally provided scratch buffer.+ * `workSpace` is a writable buffer which must be 4-bytes aligned,+ * `workSpaceSize` must be >= HIST_WKSP_SIZE+ */+size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* source, size_t sourceSize,+ void* workSpace, size_t workSpaceSize)+{+ if (sourceSize < 1500) /* heuristic threshold */+ return HIST_count_simple(count, maxSymbolValuePtr, source, sourceSize);+ if ((size_t)workSpace & 3) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */+ if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);+ return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, trustInput, (U32*)workSpace);+}++/* fast variant (unsafe : won't check if src contains values beyond count[] limit) */+size_t HIST_countFast(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* source, size_t sourceSize)+{+ unsigned tmpCounters[HIST_WKSP_SIZE_U32];+ return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, tmpCounters, sizeof(tmpCounters));+}++/* HIST_count_wksp() :+ * Same as HIST_count(), but using an externally provided scratch buffer.+ * `workSpace` size must be table of >= HIST_WKSP_SIZE_U32 unsigned */+size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* source, size_t sourceSize,+ void* workSpace, size_t workSpaceSize)+{+ if ((size_t)workSpace & 3) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */+ if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);+ if (*maxSymbolValuePtr < 255)+ return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, checkMaxSymbolValue, (U32*)workSpace);+ *maxSymbolValuePtr = 255;+ return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace, workSpaceSize);+}++size_t HIST_count(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize)+{+ unsigned tmpCounters[HIST_WKSP_SIZE_U32];+ return HIST_count_wksp(count, maxSymbolValuePtr, src, srcSize, tmpCounters, sizeof(tmpCounters));+}
+ zstd/lib/compress/hist.h view
@@ -0,0 +1,95 @@+/* ******************************************************************+ hist : Histogram functions+ part of Finite State Entropy project+ Copyright (C) 2013-present, Yann Collet.++ BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)++ Redistribution and use in source and binary forms, with or without+ modification, are permitted provided that the following conditions are+ met:++ * Redistributions of source code must retain the above copyright+ notice, this list of conditions and the following disclaimer.+ * Redistributions in binary form must reproduce the above+ copyright notice, this list of conditions and the following disclaimer+ in the documentation and/or other materials provided with the+ distribution.++ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS+ "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT+ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR+ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT+ OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,+ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT+ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,+ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY+ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.++ You can contact the author at :+ - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy+ - Public forum : https://groups.google.com/forum/#!forum/lz4c+****************************************************************** */++/* --- dependencies --- */+#include <stddef.h> /* size_t */+++/* --- simple histogram functions --- */++/*! HIST_count():+ * Provides the precise count of each byte within a table 'count'.+ * 'count' is a table of unsigned int, of minimum size (*maxSymbolValuePtr+1).+ * Updates *maxSymbolValuePtr with actual largest symbol value detected.+ * @return : count of the most frequent symbol (which isn't identified).+ * or an error code, which can be tested using HIST_isError().+ * note : if return == srcSize, there is only one symbol.+ */+size_t HIST_count(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize);++unsigned HIST_isError(size_t code); /**< tells if a return value is an error code */+++/* --- advanced histogram functions --- */++#define HIST_WKSP_SIZE_U32 1024+#define HIST_WKSP_SIZE (HIST_WKSP_SIZE_U32 * sizeof(unsigned))+/** HIST_count_wksp() :+ * Same as HIST_count(), but using an externally provided scratch buffer.+ * Benefit is this function will use very little stack space.+ * `workSpace` is a writable buffer which must be 4-bytes aligned,+ * `workSpaceSize` must be >= HIST_WKSP_SIZE+ */+size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize,+ void* workSpace, size_t workSpaceSize);++/** HIST_countFast() :+ * same as HIST_count(), but blindly trusts that all byte values within src are <= *maxSymbolValuePtr.+ * This function is unsafe, and will segfault if any value within `src` is `> *maxSymbolValuePtr`+ */+size_t HIST_countFast(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize);++/** HIST_countFast_wksp() :+ * Same as HIST_countFast(), but using an externally provided scratch buffer.+ * `workSpace` is a writable buffer which must be 4-bytes aligned,+ * `workSpaceSize` must be >= HIST_WKSP_SIZE+ */+size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize,+ void* workSpace, size_t workSpaceSize);++/*! HIST_count_simple() :+ * Same as HIST_countFast(), this function is unsafe,+ * and will segfault if any value within `src` is `> *maxSymbolValuePtr`.+ * It is also a bit slower for large inputs.+ * However, it does not need any additional memory (not even on stack).+ * @return : count of the most frequent symbol.+ * Note this function doesn't produce any error (i.e. it must succeed).+ */+unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,+ const void* src, size_t srcSize);
zstd/lib/compress/huf_compress.c view
@@ -45,18 +45,22 @@ ****************************************************************/ #include <string.h> /* memcpy, memset */ #include <stdio.h> /* printf (debug) */+#include "compiler.h" #include "bitstream.h"+#include "hist.h" #define FSE_STATIC_LINKING_ONLY /* FSE_optimalTableLog_internal */ #include "fse.h" /* header compression */ #define HUF_STATIC_LINKING_ONLY #include "huf.h"+#include "error_private.h" /* ************************************************************** * Error Management ****************************************************************/-#define HUF_STATIC_ASSERT(c) { enum { HUF_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */-#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return f+#define HUF_isError ERR_isError+#define HUF_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c) /* use only *after* variable declarations */+#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return e #define CHECK_F(f) { CHECK_V_F(_var_err__, f); } @@ -78,28 +82,28 @@ * Note : all elements within weightTable are supposed to be <= HUF_TABLELOG_MAX. */ #define MAX_FSE_TABLELOG_FOR_HUFF_HEADER 6-size_t HUF_compressWeights (void* dst, size_t dstSize, const void* weightTable, size_t wtSize)+static size_t HUF_compressWeights (void* dst, size_t dstSize, const void* weightTable, size_t wtSize) { BYTE* const ostart = (BYTE*) dst; BYTE* op = ostart; BYTE* const oend = ostart + dstSize; - U32 maxSymbolValue = HUF_TABLELOG_MAX;+ unsigned maxSymbolValue = HUF_TABLELOG_MAX; U32 tableLog = MAX_FSE_TABLELOG_FOR_HUFF_HEADER; FSE_CTable CTable[FSE_CTABLE_SIZE_U32(MAX_FSE_TABLELOG_FOR_HUFF_HEADER, HUF_TABLELOG_MAX)]; BYTE scratchBuffer[1<<MAX_FSE_TABLELOG_FOR_HUFF_HEADER]; - U32 count[HUF_TABLELOG_MAX+1];+ unsigned count[HUF_TABLELOG_MAX+1]; S16 norm[HUF_TABLELOG_MAX+1]; /* init conditions */ if (wtSize <= 1) return 0; /* Not compressible */ /* Scan input and build symbol stats */- { CHECK_V_F(maxCount, FSE_count_simple(count, &maxSymbolValue, weightTable, wtSize) );+ { unsigned const maxCount = HIST_count_simple(count, &maxSymbolValue, weightTable, wtSize); /* never fails */ if (maxCount == wtSize) return 1; /* only a single symbol in src : rle */- if (maxCount == 1) return 0; /* each symbol present maximum once => not compressible */+ if (maxCount == 1) return 0; /* each symbol present maximum once => not compressible */ } tableLog = FSE_optimalTableLog(tableLog, wtSize, maxSymbolValue);@@ -127,10 +131,10 @@ }; /* typedef'd to HUF_CElt within "huf.h" */ /*! HUF_writeCTable() :- `CTable` : huffman tree to save, using huf representation.+ `CTable` : Huffman tree to save, using huf representation. @return : size of saved CTable */ size_t HUF_writeCTable (void* dst, size_t maxDstSize,- const HUF_CElt* CTable, U32 maxSymbolValue, U32 huffLog)+ const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog) { BYTE bitsToWeight[HUF_TABLELOG_MAX + 1]; /* precomputed conversion table */ BYTE huffWeight[HUF_SYMBOLVALUE_MAX];@@ -165,7 +169,7 @@ } -size_t HUF_readCTable (HUF_CElt* CTable, U32 maxSymbolValue, const void* src, size_t srcSize)+size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize) { BYTE huffWeight[HUF_SYMBOLVALUE_MAX + 1]; /* init not required, even though some static analyzer may complain */ U32 rankVal[HUF_TABLELOG_ABSOLUTEMAX + 1]; /* large enough for values from 0 to 16 */@@ -177,7 +181,7 @@ /* check result */ if (tableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);- if (nbSymbols > maxSymbolValue+1) return ERROR(maxSymbolValue_tooSmall);+ if (nbSymbols > *maxSymbolValuePtr+1) return ERROR(maxSymbolValue_tooSmall); /* Prepare base value per rank */ { U32 n, nextRankStart = 0;@@ -206,13 +210,21 @@ min >>= 1; } } /* assign value within rank, symbol order */- { U32 n; for (n=0; n<=maxSymbolValue; n++) CTable[n].val = valPerRank[CTable[n].nbBits]++; }+ { U32 n; for (n=0; n<nbSymbols; n++) CTable[n].val = valPerRank[CTable[n].nbBits]++; } } + *maxSymbolValuePtr = nbSymbols - 1; return readSize; } +U32 HUF_getNbBits(const void* symbolTable, U32 symbolValue)+{+ const HUF_CElt* table = (const HUF_CElt*)symbolTable;+ assert(symbolValue <= HUF_SYMBOLVALUE_MAX);+ return table[symbolValue].nbBits;+} + typedef struct nodeElt_s { U32 count; U16 parent;@@ -266,7 +278,8 @@ if (highTotal <= lowTotal) break; } } /* only triggered when no more rank 1 symbol left => find closest one (note : there is necessarily at least one !) */- while ((nBitsToDecrease<=HUF_TABLELOG_MAX) && (rankLast[nBitsToDecrease] == noSymbol)) /* HUF_MAX_TABLELOG test just to please gcc 5+; but it should not be necessary */+ /* HUF_MAX_TABLELOG test just to please gcc 5+; but it should not be necessary */+ while ((nBitsToDecrease<=HUF_TABLELOG_MAX) && (rankLast[nBitsToDecrease] == noSymbol)) nBitsToDecrease ++; totalCost -= 1 << (nBitsToDecrease-1); if (rankLast[nBitsToDecrease-1] == noSymbol)@@ -302,7 +315,7 @@ U32 current; } rankPos; -static void HUF_sort(nodeElt* huffNode, const U32* count, U32 maxSymbolValue)+static void HUF_sort(nodeElt* huffNode, const unsigned* count, U32 maxSymbolValue) { rankPos rank[32]; U32 n;@@ -318,7 +331,10 @@ U32 const c = count[n]; U32 const r = BIT_highbit32(c+1) + 1; U32 pos = rank[r].current++;- while ((pos > rank[r].base) && (c > huffNode[pos-1].count)) huffNode[pos]=huffNode[pos-1], pos--;+ while ((pos > rank[r].base) && (c > huffNode[pos-1].count)) {+ huffNode[pos] = huffNode[pos-1];+ pos--;+ } huffNode[pos].count = c; huffNode[pos].byte = (BYTE)n; }@@ -327,11 +343,11 @@ /** HUF_buildCTable_wksp() : * Same as HUF_buildCTable(), but using externally allocated scratch buffer.- * `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as a table of 1024 unsigned.+ * `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as a table of HUF_CTABLE_WORKSPACE_SIZE_U32 unsigned. */ #define STARTNODE (HUF_SYMBOLVALUE_MAX+1)-typedef nodeElt huffNodeTable[2*HUF_SYMBOLVALUE_MAX+1 +1];-size_t HUF_buildCTable_wksp (HUF_CElt* tree, const U32* count, U32 maxSymbolValue, U32 maxNbBits, void* workSpace, size_t wkspSize)+typedef nodeElt huffNodeTable[HUF_CTABLE_WORKSPACE_SIZE_U32];+size_t HUF_buildCTable_wksp (HUF_CElt* tree, const unsigned* count, U32 maxSymbolValue, U32 maxNbBits, void* workSpace, size_t wkspSize) { nodeElt* const huffNode0 = (nodeElt*)workSpace; nodeElt* const huffNode = huffNode0+1;@@ -341,9 +357,10 @@ U32 nodeRoot; /* safety checks */- if (wkspSize < sizeof(huffNodeTable)) return ERROR(GENERIC); /* workSpace is not large enough */+ if (((size_t)workSpace & 3) != 0) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */+ if (wkspSize < sizeof(huffNodeTable)) return ERROR(workSpace_tooSmall); if (maxNbBits == 0) maxNbBits = HUF_TABLELOG_DEFAULT;- if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(GENERIC);+ if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge); memset(huffNode0, 0, sizeof(huffNodeTable)); /* sort, decreasing order */@@ -401,37 +418,60 @@ } /** HUF_buildCTable() :+ * @return : maxNbBits * Note : count is used before tree is written, so they can safely overlap */-size_t HUF_buildCTable (HUF_CElt* tree, const U32* count, U32 maxSymbolValue, U32 maxNbBits)+size_t HUF_buildCTable (HUF_CElt* tree, const unsigned* count, unsigned maxSymbolValue, unsigned maxNbBits) { huffNodeTable nodeTable; return HUF_buildCTable_wksp(tree, count, maxSymbolValue, maxNbBits, nodeTable, sizeof(nodeTable)); } -static void HUF_encodeSymbol(BIT_CStream_t* bitCPtr, U32 symbol, const HUF_CElt* CTable)+static size_t HUF_estimateCompressedSize(HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue) {- BIT_addBitsFast(bitCPtr, CTable[symbol].val, CTable[symbol].nbBits);+ size_t nbBits = 0;+ int s;+ for (s = 0; s <= (int)maxSymbolValue; ++s) {+ nbBits += CTable[s].nbBits * count[s];+ }+ return nbBits >> 3; } +static int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue) {+ int bad = 0;+ int s;+ for (s = 0; s <= (int)maxSymbolValue; ++s) {+ bad |= (count[s] != 0) & (CTable[s].nbBits == 0);+ }+ return !bad;+}+ size_t HUF_compressBound(size_t size) { return HUF_COMPRESSBOUND(size); } -#define HUF_FLUSHBITS(s) (fast ? BIT_flushBitsFast(s) : BIT_flushBits(s))+FORCE_INLINE_TEMPLATE void+HUF_encodeSymbol(BIT_CStream_t* bitCPtr, U32 symbol, const HUF_CElt* CTable)+{+ BIT_addBitsFast(bitCPtr, CTable[symbol].val, CTable[symbol].nbBits);+} +#define HUF_FLUSHBITS(s) BIT_flushBits(s)+ #define HUF_FLUSHBITS_1(stream) \ if (sizeof((stream)->bitContainer)*8 < HUF_TABLELOG_MAX*2+7) HUF_FLUSHBITS(stream) #define HUF_FLUSHBITS_2(stream) \ if (sizeof((stream)->bitContainer)*8 < HUF_TABLELOG_MAX*4+7) HUF_FLUSHBITS(stream) -size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable)+FORCE_INLINE_TEMPLATE size_t+HUF_compress1X_usingCTable_internal_body(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ const HUF_CElt* CTable) { const BYTE* ip = (const BYTE*) src; BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstSize; BYTE* op = ostart; size_t n;- const unsigned fast = (dstSize >= HUF_BLOCKBOUND(srcSize)); BIT_CStream_t bitC; /* init */@@ -444,12 +484,15 @@ { case 3 : HUF_encodeSymbol(&bitC, ip[n+ 2], CTable); HUF_FLUSHBITS_2(&bitC);+ /* fall-through */ case 2 : HUF_encodeSymbol(&bitC, ip[n+ 1], CTable); HUF_FLUSHBITS_1(&bitC);+ /* fall-through */ case 1 : HUF_encodeSymbol(&bitC, ip[n+ 0], CTable); HUF_FLUSHBITS(&bitC);- case 0 :- default: ;+ /* fall-through */+ case 0 : /* fall-through */+ default: break; } for (; n>0; n-=4) { /* note : n&3==0 at this stage */@@ -466,9 +509,59 @@ return BIT_closeCStream(&bitC); } +#if DYNAMIC_BMI2 -size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable)+static TARGET_ATTRIBUTE("bmi2") size_t+HUF_compress1X_usingCTable_internal_bmi2(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ const HUF_CElt* CTable) {+ return HUF_compress1X_usingCTable_internal_body(dst, dstSize, src, srcSize, CTable);+}++static size_t+HUF_compress1X_usingCTable_internal_default(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ const HUF_CElt* CTable)+{+ return HUF_compress1X_usingCTable_internal_body(dst, dstSize, src, srcSize, CTable);+}++static size_t+HUF_compress1X_usingCTable_internal(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ const HUF_CElt* CTable, const int bmi2)+{+ if (bmi2) {+ return HUF_compress1X_usingCTable_internal_bmi2(dst, dstSize, src, srcSize, CTable);+ }+ return HUF_compress1X_usingCTable_internal_default(dst, dstSize, src, srcSize, CTable);+}++#else++static size_t+HUF_compress1X_usingCTable_internal(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ const HUF_CElt* CTable, const int bmi2)+{+ (void)bmi2;+ return HUF_compress1X_usingCTable_internal_body(dst, dstSize, src, srcSize, CTable);+}++#endif++size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable)+{+ return HUF_compress1X_usingCTable_internal(dst, dstSize, src, srcSize, CTable, /* bmi2 */ 0);+}+++static size_t+HUF_compress4X_usingCTable_internal(void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ const HUF_CElt* CTable, int bmi2)+{ size_t const segmentSize = (srcSize+3)/4; /* first 3 segments */ const BYTE* ip = (const BYTE*) src; const BYTE* const iend = ip + srcSize;@@ -480,28 +573,31 @@ if (srcSize < 12) return 0; /* no saving possible : too small input */ op += 6; /* jumpTable */ - { CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend-op, ip, segmentSize, CTable) );+ { CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, oend-op, ip, segmentSize, CTable, bmi2) ); if (cSize==0) return 0;+ assert(cSize <= 65535); MEM_writeLE16(ostart, (U16)cSize); op += cSize; } ip += segmentSize;- { CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend-op, ip, segmentSize, CTable) );+ { CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, oend-op, ip, segmentSize, CTable, bmi2) ); if (cSize==0) return 0;+ assert(cSize <= 65535); MEM_writeLE16(ostart+2, (U16)cSize); op += cSize; } ip += segmentSize;- { CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend-op, ip, segmentSize, CTable) );+ { CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, oend-op, ip, segmentSize, CTable, bmi2) ); if (cSize==0) return 0;+ assert(cSize <= 65535); MEM_writeLE16(ostart+4, (U16)cSize); op += cSize; } ip += segmentSize;- { CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend-op, ip, iend-ip, CTable) );+ { CHECK_V_F(cSize, HUF_compress1X_usingCTable_internal(op, oend-op, ip, iend-ip, CTable, bmi2) ); if (cSize==0) return 0; op += cSize; }@@ -509,65 +605,122 @@ return op-ostart; } +size_t HUF_compress4X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable)+{+ return HUF_compress4X_usingCTable_internal(dst, dstSize, src, srcSize, CTable, /* bmi2 */ 0);+} -/* `workSpace` must a table of at least 1024 unsigned */-static size_t HUF_compress_internal (- void* dst, size_t dstSize,+typedef enum { HUF_singleStream, HUF_fourStreams } HUF_nbStreams_e;++static size_t HUF_compressCTable_internal(+ BYTE* const ostart, BYTE* op, BYTE* const oend, const void* src, size_t srcSize,- unsigned maxSymbolValue, unsigned huffLog,- unsigned singleStream,- void* workSpace, size_t wkspSize)+ HUF_nbStreams_e nbStreams, const HUF_CElt* CTable, const int bmi2) {+ size_t const cSize = (nbStreams==HUF_singleStream) ?+ HUF_compress1X_usingCTable_internal(op, oend - op, src, srcSize, CTable, bmi2) :+ HUF_compress4X_usingCTable_internal(op, oend - op, src, srcSize, CTable, bmi2);+ if (HUF_isError(cSize)) { return cSize; }+ if (cSize==0) { return 0; } /* uncompressible */+ op += cSize;+ /* check compressibility */+ if ((size_t)(op-ostart) >= srcSize-1) { return 0; }+ return op-ostart;+}++typedef struct {+ unsigned count[HUF_SYMBOLVALUE_MAX + 1];+ HUF_CElt CTable[HUF_SYMBOLVALUE_MAX + 1];+ huffNodeTable nodeTable;+} HUF_compress_tables_t;++/* HUF_compress_internal() :+ * `workSpace` must a table of at least HUF_WORKSPACE_SIZE_U32 unsigned */+static size_t+HUF_compress_internal (void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned huffLog,+ HUF_nbStreams_e nbStreams,+ void* workSpace, size_t wkspSize,+ HUF_CElt* oldHufTable, HUF_repeat* repeat, int preferRepeat,+ const int bmi2)+{+ HUF_compress_tables_t* const table = (HUF_compress_tables_t*)workSpace; BYTE* const ostart = (BYTE*)dst; BYTE* const oend = ostart + dstSize; BYTE* op = ostart; - union {- U32 count[HUF_SYMBOLVALUE_MAX+1];- HUF_CElt CTable[HUF_SYMBOLVALUE_MAX+1];- } table; /* `count` can overlap with `CTable`; saves 1 KB */- /* checks & inits */- if (wkspSize < sizeof(huffNodeTable)) return ERROR(GENERIC);- if (!srcSize) return 0; /* Uncompressed (note : 1 means rle, so first byte must be correct) */- if (!dstSize) return 0; /* cannot fit within dst budget */+ if (((size_t)workSpace & 3) != 0) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */+ if (wkspSize < HUF_WORKSPACE_SIZE) return ERROR(workSpace_tooSmall);+ if (!srcSize) return 0; /* Uncompressed */+ if (!dstSize) return 0; /* cannot fit anything within dst budget */ if (srcSize > HUF_BLOCKSIZE_MAX) return ERROR(srcSize_wrong); /* current block size limit */ if (huffLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);+ if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge); if (!maxSymbolValue) maxSymbolValue = HUF_SYMBOLVALUE_MAX; if (!huffLog) huffLog = HUF_TABLELOG_DEFAULT; + /* Heuristic : If old table is valid, use it for small inputs */+ if (preferRepeat && repeat && *repeat == HUF_repeat_valid) {+ return HUF_compressCTable_internal(ostart, op, oend,+ src, srcSize,+ nbStreams, oldHufTable, bmi2);+ }+ /* Scan input and build symbol stats */- { CHECK_V_F(largest, FSE_count_wksp (table.count, &maxSymbolValue, (const BYTE*)src, srcSize, (U32*)workSpace) );+ { CHECK_V_F(largest, HIST_count_wksp (table->count, &maxSymbolValue, (const BYTE*)src, srcSize, workSpace, wkspSize) ); if (largest == srcSize) { *ostart = ((const BYTE*)src)[0]; return 1; } /* single symbol, rle */- if (largest <= (srcSize >> 7)+1) return 0; /* Fast heuristic : not compressible enough */+ if (largest <= (srcSize >> 7)+4) return 0; /* heuristic : probably not compressible enough */ } + /* Check validity of previous table */+ if ( repeat+ && *repeat == HUF_repeat_check+ && !HUF_validateCTable(oldHufTable, table->count, maxSymbolValue)) {+ *repeat = HUF_repeat_none;+ }+ /* Heuristic : use existing table for small inputs */+ if (preferRepeat && repeat && *repeat != HUF_repeat_none) {+ return HUF_compressCTable_internal(ostart, op, oend,+ src, srcSize,+ nbStreams, oldHufTable, bmi2);+ }+ /* Build Huffman Tree */ huffLog = HUF_optimalTableLog(huffLog, srcSize, maxSymbolValue);- { CHECK_V_F(maxBits, HUF_buildCTable_wksp (table.CTable, table.count, maxSymbolValue, huffLog, workSpace, wkspSize) );+ { size_t const maxBits = HUF_buildCTable_wksp(table->CTable, table->count,+ maxSymbolValue, huffLog,+ table->nodeTable, sizeof(table->nodeTable));+ CHECK_F(maxBits); huffLog = (U32)maxBits;+ /* Zero unused symbols in CTable, so we can check it for validity */+ memset(table->CTable + (maxSymbolValue + 1), 0,+ sizeof(table->CTable) - ((maxSymbolValue + 1) * sizeof(HUF_CElt))); } /* Write table description header */- { CHECK_V_F(hSize, HUF_writeCTable (op, dstSize, table.CTable, maxSymbolValue, huffLog) );- if (hSize + 12 >= srcSize) return 0; /* not useful to try compression */- op += hSize;- }+ { CHECK_V_F(hSize, HUF_writeCTable (op, dstSize, table->CTable, maxSymbolValue, huffLog) );+ /* Check if using previous huffman table is beneficial */+ if (repeat && *repeat != HUF_repeat_none) {+ size_t const oldSize = HUF_estimateCompressedSize(oldHufTable, table->count, maxSymbolValue);+ size_t const newSize = HUF_estimateCompressedSize(table->CTable, table->count, maxSymbolValue);+ if (oldSize <= hSize + newSize || hSize + 12 >= srcSize) {+ return HUF_compressCTable_internal(ostart, op, oend,+ src, srcSize,+ nbStreams, oldHufTable, bmi2);+ } } - /* Compress */- { size_t const cSize = (singleStream) ?- HUF_compress1X_usingCTable(op, oend - op, src, srcSize, table.CTable) : /* single segment */- HUF_compress4X_usingCTable(op, oend - op, src, srcSize, table.CTable);- if (HUF_isError(cSize)) return cSize;- if (cSize==0) return 0; /* uncompressible */- op += cSize;+ /* Use the new huffman table */+ if (hSize + 12ul >= srcSize) { return 0; }+ op += hSize;+ if (repeat) { *repeat = HUF_repeat_none; }+ if (oldHufTable)+ memcpy(oldHufTable, table->CTable, sizeof(table->CTable)); /* Save new table */ }-- /* check compressibility */- if ((size_t)(op-ostart) >= srcSize-1)- return 0;-- return op-ostart;+ return HUF_compressCTable_internal(ostart, op, oend,+ src, srcSize,+ nbStreams, table->CTable, bmi2); } @@ -576,34 +729,70 @@ unsigned maxSymbolValue, unsigned huffLog, void* workSpace, size_t wkspSize) {- return HUF_compress_internal(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, 1 /* single stream */, workSpace, wkspSize);+ return HUF_compress_internal(dst, dstSize, src, srcSize,+ maxSymbolValue, huffLog, HUF_singleStream,+ workSpace, wkspSize,+ NULL, NULL, 0, 0 /*bmi2*/); } +size_t HUF_compress1X_repeat (void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned huffLog,+ void* workSpace, size_t wkspSize,+ HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2)+{+ return HUF_compress_internal(dst, dstSize, src, srcSize,+ maxSymbolValue, huffLog, HUF_singleStream,+ workSpace, wkspSize, hufTable,+ repeat, preferRepeat, bmi2);+}+ size_t HUF_compress1X (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog) {- unsigned workSpace[1024];+ unsigned workSpace[HUF_WORKSPACE_SIZE_U32]; return HUF_compress1X_wksp(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, workSpace, sizeof(workSpace)); } +/* HUF_compress4X_repeat():+ * compress input using 4 streams.+ * provide workspace to generate compression tables */ size_t HUF_compress4X_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog, void* workSpace, size_t wkspSize) {- return HUF_compress_internal(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, 0 /* 4 streams */, workSpace, wkspSize);+ return HUF_compress_internal(dst, dstSize, src, srcSize,+ maxSymbolValue, huffLog, HUF_fourStreams,+ workSpace, wkspSize,+ NULL, NULL, 0, 0 /*bmi2*/); } +/* HUF_compress4X_repeat():+ * compress input using 4 streams.+ * re-use an existing huffman compression table */+size_t HUF_compress4X_repeat (void* dst, size_t dstSize,+ const void* src, size_t srcSize,+ unsigned maxSymbolValue, unsigned huffLog,+ void* workSpace, size_t wkspSize,+ HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2)+{+ return HUF_compress_internal(dst, dstSize, src, srcSize,+ maxSymbolValue, huffLog, HUF_fourStreams,+ workSpace, wkspSize,+ hufTable, repeat, preferRepeat, bmi2);+}+ size_t HUF_compress2 (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog) {- unsigned workSpace[1024];+ unsigned workSpace[HUF_WORKSPACE_SIZE_U32]; return HUF_compress4X_wksp(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, workSpace, sizeof(workSpace)); } size_t HUF_compress (void* dst, size_t maxDstSize, const void* src, size_t srcSize) {- return HUF_compress2(dst, maxDstSize, src, (U32)srcSize, 255, HUF_TABLELOG_DEFAULT);+ return HUF_compress2(dst, maxDstSize, src, srcSize, 255, HUF_TABLELOG_DEFAULT); }
zstd/lib/compress/zstd_compress.c view
@@ -1,3306 +1,4417 @@-/**- * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.- * All rights reserved.- *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.- */---/*-*************************************-* Dependencies-***************************************/-#include <string.h> /* memset */-#include "mem.h"-#define XXH_STATIC_LINKING_ONLY /* XXH64_state_t */-#include "xxhash.h" /* XXH_reset, update, digest */-#define FSE_STATIC_LINKING_ONLY /* FSE_encodeSymbol */-#include "fse.h"-#define HUF_STATIC_LINKING_ONLY-#include "huf.h"-#include "zstd_internal.h" /* includes zstd.h */---/*-*************************************-* Constants-***************************************/-static const U32 g_searchStrength = 8; /* control skip over incompressible data */-#define HASH_READ_SIZE 8-typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;---/*-*************************************-* Helper functions-***************************************/-#define ZSTD_STATIC_ASSERT(c) { enum { ZSTD_static_assert = 1/(int)(!!(c)) }; }-size_t ZSTD_compressBound(size_t srcSize) { return FSE_compressBound(srcSize) + 12; }---/*-*************************************-* Sequence storage-***************************************/-static void ZSTD_resetSeqStore(seqStore_t* ssPtr)-{- ssPtr->lit = ssPtr->litStart;- ssPtr->sequences = ssPtr->sequencesStart;- ssPtr->longLengthID = 0;-}---/*-*************************************-* Context memory management-***************************************/-struct ZSTD_CCtx_s {- const BYTE* nextSrc; /* next block here to continue on current prefix */- const BYTE* base; /* All regular indexes relative to this position */- const BYTE* dictBase; /* extDict indexes relative to this position */- U32 dictLimit; /* below that point, need extDict */- U32 lowLimit; /* below that point, no more data */- U32 nextToUpdate; /* index from which to continue dictionary update */- U32 nextToUpdate3; /* index from which to continue dictionary update */- U32 hashLog3; /* dispatch table : larger == faster, more memory */- U32 loadedDictEnd;- ZSTD_compressionStage_e stage;- U32 rep[ZSTD_REP_NUM];- U32 savedRep[ZSTD_REP_NUM];- U32 dictID;- ZSTD_parameters params;- void* workSpace;- size_t workSpaceSize;- size_t blockSize;- U64 frameContentSize;- XXH64_state_t xxhState;- ZSTD_customMem customMem;-- seqStore_t seqStore; /* sequences storage ptrs */- U32* hashTable;- U32* hashTable3;- U32* chainTable;- HUF_CElt* hufTable;- U32 flagStaticTables;- FSE_CTable offcodeCTable [FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];- FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];- FSE_CTable litlengthCTable [FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];- unsigned tmpCounters[1024];-};--ZSTD_CCtx* ZSTD_createCCtx(void)-{- return ZSTD_createCCtx_advanced(defaultCustomMem);-}--ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem)-{- ZSTD_CCtx* cctx;-- if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;- if (!customMem.customAlloc || !customMem.customFree) return NULL;-- cctx = (ZSTD_CCtx*) ZSTD_malloc(sizeof(ZSTD_CCtx), customMem);- if (!cctx) return NULL;- memset(cctx, 0, sizeof(ZSTD_CCtx));- memcpy(&(cctx->customMem), &customMem, sizeof(customMem));- return cctx;-}--size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)-{- if (cctx==NULL) return 0; /* support free on NULL */- ZSTD_free(cctx->workSpace, cctx->customMem);- ZSTD_free(cctx, cctx->customMem);- return 0; /* reserved as a potential error code in the future */-}--size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx)-{- if (cctx==NULL) return 0; /* support sizeof on NULL */- return sizeof(*cctx) + cctx->workSpaceSize;-}--const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx) /* hidden interface */-{- return &(ctx->seqStore);-}--static ZSTD_parameters ZSTD_getParamsFromCCtx(const ZSTD_CCtx* cctx)-{- return cctx->params;-}---/** ZSTD_checkParams() :- ensure param values remain within authorized range.- @return : 0, or an error code if one value is beyond authorized range */-size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)-{-# define CLAMPCHECK(val,min,max) { if ((val<min) | (val>max)) return ERROR(compressionParameter_unsupported); }- CLAMPCHECK(cParams.windowLog, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX);- CLAMPCHECK(cParams.chainLog, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX);- CLAMPCHECK(cParams.hashLog, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX);- CLAMPCHECK(cParams.searchLog, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX);- { U32 const searchLengthMin = ((cParams.strategy == ZSTD_fast) | (cParams.strategy == ZSTD_greedy)) ? ZSTD_SEARCHLENGTH_MIN+1 : ZSTD_SEARCHLENGTH_MIN;- U32 const searchLengthMax = (cParams.strategy == ZSTD_fast) ? ZSTD_SEARCHLENGTH_MAX : ZSTD_SEARCHLENGTH_MAX-1;- CLAMPCHECK(cParams.searchLength, searchLengthMin, searchLengthMax); }- CLAMPCHECK(cParams.targetLength, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX);- if ((U32)(cParams.strategy) > (U32)ZSTD_btopt2) return ERROR(compressionParameter_unsupported);- return 0;-}---/** ZSTD_cycleLog() :- * condition for correct operation : hashLog > 1 */-static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)-{- U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);- return hashLog - btScale;-}--/** ZSTD_adjustCParams() :- optimize `cPar` for a given input (`srcSize` and `dictSize`).- mostly downsizing to reduce memory consumption and initialization.- Both `srcSize` and `dictSize` are optional (use 0 if unknown),- but if both are 0, no optimization can be done.- Note : cPar is considered validated at this stage. Use ZSTD_checkParams() to ensure that. */-ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize)-{- if (srcSize+dictSize == 0) return cPar; /* no size information available : no adjustment */-- /* resize params, to use less memory when necessary */- { U32 const minSrcSize = (srcSize==0) ? 500 : 0;- U64 const rSize = srcSize + dictSize + minSrcSize;- if (rSize < ((U64)1<<ZSTD_WINDOWLOG_MAX)) {- U32 const srcLog = MAX(ZSTD_HASHLOG_MIN, ZSTD_highbit32((U32)(rSize)-1) + 1);- if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;- } }- if (cPar.hashLog > cPar.windowLog) cPar.hashLog = cPar.windowLog;- { U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);- if (cycleLog > cPar.windowLog) cPar.chainLog -= (cycleLog - cPar.windowLog);- }-- if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN) cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* required for frame header */-- return cPar;-}---size_t ZSTD_estimateCCtxSize(ZSTD_compressionParameters cParams)-{- size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << cParams.windowLog);- U32 const divider = (cParams.searchLength==3) ? 3 : 4;- size_t const maxNbSeq = blockSize / divider;- size_t const tokenSpace = blockSize + 11*maxNbSeq;-- size_t const chainSize = (cParams.strategy == ZSTD_fast) ? 0 : (1 << cParams.chainLog);- size_t const hSize = ((size_t)1) << cParams.hashLog;- U32 const hashLog3 = (cParams.searchLength>3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, cParams.windowLog);- size_t const h3Size = ((size_t)1) << hashLog3;- size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);-- size_t const optSpace = ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits))*sizeof(U32)- + (ZSTD_OPT_NUM+1)*(sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));- size_t const neededSpace = tableSpace + (256*sizeof(U32)) /* huffTable */ + tokenSpace- + (((cParams.strategy == ZSTD_btopt) || (cParams.strategy == ZSTD_btopt2)) ? optSpace : 0);-- return sizeof(ZSTD_CCtx) + neededSpace;-}---static U32 ZSTD_equivalentParams(ZSTD_parameters param1, ZSTD_parameters param2)-{- return (param1.cParams.hashLog == param2.cParams.hashLog)- & (param1.cParams.chainLog == param2.cParams.chainLog)- & (param1.cParams.strategy == param2.cParams.strategy)- & ((param1.cParams.searchLength==3) == (param2.cParams.searchLength==3));-}--/*! ZSTD_continueCCtx() :- reuse CCtx without reset (note : requires no dictionary) */-static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_parameters params, U64 frameContentSize)-{- U32 const end = (U32)(cctx->nextSrc - cctx->base);- cctx->params = params;- cctx->frameContentSize = frameContentSize;- cctx->lowLimit = end;- cctx->dictLimit = end;- cctx->nextToUpdate = end+1;- cctx->stage = ZSTDcs_init;- cctx->dictID = 0;- cctx->loadedDictEnd = 0;- { int i; for (i=0; i<ZSTD_REP_NUM; i++) cctx->rep[i] = repStartValue[i]; }- cctx->seqStore.litLengthSum = 0; /* force reset of btopt stats */- XXH64_reset(&cctx->xxhState, 0);- return 0;-}--typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset, ZSTDcrp_fullReset } ZSTD_compResetPolicy_e;--/*! ZSTD_resetCCtx_advanced() :- note : 'params' must be validated */-static size_t ZSTD_resetCCtx_advanced (ZSTD_CCtx* zc,- ZSTD_parameters params, U64 frameContentSize,- ZSTD_compResetPolicy_e const crp)-{- if (crp == ZSTDcrp_continue)- if (ZSTD_equivalentParams(params, zc->params))- return ZSTD_continueCCtx(zc, params, frameContentSize);-- { size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << params.cParams.windowLog);- U32 const divider = (params.cParams.searchLength==3) ? 3 : 4;- size_t const maxNbSeq = blockSize / divider;- size_t const tokenSpace = blockSize + 11*maxNbSeq;- size_t const chainSize = (params.cParams.strategy == ZSTD_fast) ? 0 : (1 << params.cParams.chainLog);- size_t const hSize = ((size_t)1) << params.cParams.hashLog;- U32 const hashLog3 = (params.cParams.searchLength>3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, params.cParams.windowLog);- size_t const h3Size = ((size_t)1) << hashLog3;- size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);- void* ptr;-- /* Check if workSpace is large enough, alloc a new one if needed */- { size_t const optSpace = ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits))*sizeof(U32)- + (ZSTD_OPT_NUM+1)*(sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));- size_t const neededSpace = tableSpace + (256*sizeof(U32)) /* huffTable */ + tokenSpace- + (((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btopt2)) ? optSpace : 0);- if (zc->workSpaceSize < neededSpace) {- ZSTD_free(zc->workSpace, zc->customMem);- zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem);- if (zc->workSpace == NULL) return ERROR(memory_allocation);- zc->workSpaceSize = neededSpace;- } }-- if (crp!=ZSTDcrp_noMemset) memset(zc->workSpace, 0, tableSpace); /* reset tables only */- XXH64_reset(&zc->xxhState, 0);- zc->hashLog3 = hashLog3;- zc->hashTable = (U32*)(zc->workSpace);- zc->chainTable = zc->hashTable + hSize;- zc->hashTable3 = zc->chainTable + chainSize;- ptr = zc->hashTable3 + h3Size;- zc->hufTable = (HUF_CElt*)ptr;- zc->flagStaticTables = 0;- ptr = ((U32*)ptr) + 256; /* note : HUF_CElt* is incomplete type, size is simulated using U32 */-- zc->nextToUpdate = 1;- zc->nextSrc = NULL;- zc->base = NULL;- zc->dictBase = NULL;- zc->dictLimit = 0;- zc->lowLimit = 0;- zc->params = params;- zc->blockSize = blockSize;- zc->frameContentSize = frameContentSize;- { int i; for (i=0; i<ZSTD_REP_NUM; i++) zc->rep[i] = repStartValue[i]; }-- if ((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btopt2)) {- zc->seqStore.litFreq = (U32*)ptr;- zc->seqStore.litLengthFreq = zc->seqStore.litFreq + (1<<Litbits);- zc->seqStore.matchLengthFreq = zc->seqStore.litLengthFreq + (MaxLL+1);- zc->seqStore.offCodeFreq = zc->seqStore.matchLengthFreq + (MaxML+1);- ptr = zc->seqStore.offCodeFreq + (MaxOff+1);- zc->seqStore.matchTable = (ZSTD_match_t*)ptr;- ptr = zc->seqStore.matchTable + ZSTD_OPT_NUM+1;- zc->seqStore.priceTable = (ZSTD_optimal_t*)ptr;- ptr = zc->seqStore.priceTable + ZSTD_OPT_NUM+1;- zc->seqStore.litLengthSum = 0;- }- zc->seqStore.sequencesStart = (seqDef*)ptr;- ptr = zc->seqStore.sequencesStart + maxNbSeq;- zc->seqStore.llCode = (BYTE*) ptr;- zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq;- zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq;- zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq;-- zc->stage = ZSTDcs_init;- zc->dictID = 0;- zc->loadedDictEnd = 0;-- return 0;- }-}---/*! ZSTD_copyCCtx() :-* Duplicate an existing context `srcCCtx` into another one `dstCCtx`.-* Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).-* @return : 0, or an error code */-size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, unsigned long long pledgedSrcSize)-{- if (srcCCtx->stage!=ZSTDcs_init) return ERROR(stage_wrong);-- memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));- ZSTD_resetCCtx_advanced(dstCCtx, srcCCtx->params, pledgedSrcSize, ZSTDcrp_noMemset);-- /* copy tables */- { size_t const chainSize = (srcCCtx->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << srcCCtx->params.cParams.chainLog);- size_t const hSize = ((size_t)1) << srcCCtx->params.cParams.hashLog;- size_t const h3Size = (size_t)1 << srcCCtx->hashLog3;- size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);- memcpy(dstCCtx->workSpace, srcCCtx->workSpace, tableSpace);- }-- /* copy dictionary offsets */- dstCCtx->nextToUpdate = srcCCtx->nextToUpdate;- dstCCtx->nextToUpdate3= srcCCtx->nextToUpdate3;- dstCCtx->nextSrc = srcCCtx->nextSrc;- dstCCtx->base = srcCCtx->base;- dstCCtx->dictBase = srcCCtx->dictBase;- dstCCtx->dictLimit = srcCCtx->dictLimit;- dstCCtx->lowLimit = srcCCtx->lowLimit;- dstCCtx->loadedDictEnd= srcCCtx->loadedDictEnd;- dstCCtx->dictID = srcCCtx->dictID;-- /* copy entropy tables */- dstCCtx->flagStaticTables = srcCCtx->flagStaticTables;- if (srcCCtx->flagStaticTables) {- memcpy(dstCCtx->hufTable, srcCCtx->hufTable, 256*4);- memcpy(dstCCtx->litlengthCTable, srcCCtx->litlengthCTable, sizeof(dstCCtx->litlengthCTable));- memcpy(dstCCtx->matchlengthCTable, srcCCtx->matchlengthCTable, sizeof(dstCCtx->matchlengthCTable));- memcpy(dstCCtx->offcodeCTable, srcCCtx->offcodeCTable, sizeof(dstCCtx->offcodeCTable));- }-- return 0;-}---/*! ZSTD_reduceTable() :-* reduce table indexes by `reducerValue` */-static void ZSTD_reduceTable (U32* const table, U32 const size, U32 const reducerValue)-{- U32 u;- for (u=0 ; u < size ; u++) {- if (table[u] < reducerValue) table[u] = 0;- else table[u] -= reducerValue;- }-}--/*! ZSTD_reduceIndex() :-* rescale all indexes to avoid future overflow (indexes are U32) */-static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue)-{- { U32 const hSize = 1 << zc->params.cParams.hashLog;- ZSTD_reduceTable(zc->hashTable, hSize, reducerValue); }-- { U32 const chainSize = (zc->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << zc->params.cParams.chainLog);- ZSTD_reduceTable(zc->chainTable, chainSize, reducerValue); }-- { U32 const h3Size = (zc->hashLog3) ? 1 << zc->hashLog3 : 0;- ZSTD_reduceTable(zc->hashTable3, h3Size, reducerValue); }-}---/*-*******************************************************-* Block entropic compression-*********************************************************/--/* See doc/zstd_compression_format.md for detailed format description */--size_t ZSTD_noCompressBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- if (srcSize + ZSTD_blockHeaderSize > dstCapacity) return ERROR(dstSize_tooSmall);- memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize);- MEM_writeLE24(dst, (U32)(srcSize << 2) + (U32)bt_raw);- return ZSTD_blockHeaderSize+srcSize;-}---static size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- BYTE* const ostart = (BYTE* const)dst;- U32 const flSize = 1 + (srcSize>31) + (srcSize>4095);-- if (srcSize + flSize > dstCapacity) return ERROR(dstSize_tooSmall);-- switch(flSize)- {- case 1: /* 2 - 1 - 5 */- ostart[0] = (BYTE)((U32)set_basic + (srcSize<<3));- break;- case 2: /* 2 - 2 - 12 */- MEM_writeLE16(ostart, (U16)((U32)set_basic + (1<<2) + (srcSize<<4)));- break;- default: /*note : should not be necessary : flSize is within {1,2,3} */- case 3: /* 2 - 2 - 20 */- MEM_writeLE32(ostart, (U32)((U32)set_basic + (3<<2) + (srcSize<<4)));- break;- }-- memcpy(ostart + flSize, src, srcSize);- return srcSize + flSize;-}--static size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- BYTE* const ostart = (BYTE* const)dst;- U32 const flSize = 1 + (srcSize>31) + (srcSize>4095);-- (void)dstCapacity; /* dstCapacity already guaranteed to be >=4, hence large enough */-- switch(flSize)- {- case 1: /* 2 - 1 - 5 */- ostart[0] = (BYTE)((U32)set_rle + (srcSize<<3));- break;- case 2: /* 2 - 2 - 12 */- MEM_writeLE16(ostart, (U16)((U32)set_rle + (1<<2) + (srcSize<<4)));- break;- default: /*note : should not be necessary : flSize is necessarily within {1,2,3} */- case 3: /* 2 - 2 - 20 */- MEM_writeLE32(ostart, (U32)((U32)set_rle + (3<<2) + (srcSize<<4)));- break;- }-- ostart[flSize] = *(const BYTE*)src;- return flSize+1;-}---static size_t ZSTD_minGain(size_t srcSize) { return (srcSize >> 6) + 2; }--static size_t ZSTD_compressLiterals (ZSTD_CCtx* zc,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize)-{- size_t const minGain = ZSTD_minGain(srcSize);- size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);- BYTE* const ostart = (BYTE*)dst;- U32 singleStream = srcSize < 256;- symbolEncodingType_e hType = set_compressed;- size_t cLitSize;--- /* small ? don't even attempt compression (speed opt) */-# define LITERAL_NOENTROPY 63- { size_t const minLitSize = zc->flagStaticTables ? 6 : LITERAL_NOENTROPY;- if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);- }-- if (dstCapacity < lhSize+1) return ERROR(dstSize_tooSmall); /* not enough space for compression */- if (zc->flagStaticTables && (lhSize==3)) {- hType = set_repeat;- singleStream = 1;- cLitSize = HUF_compress1X_usingCTable(ostart+lhSize, dstCapacity-lhSize, src, srcSize, zc->hufTable);- } else {- cLitSize = singleStream ? HUF_compress1X_wksp(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11, zc->tmpCounters, sizeof(zc->tmpCounters))- : HUF_compress4X_wksp(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11, zc->tmpCounters, sizeof(zc->tmpCounters));- }-- if ((cLitSize==0) | (cLitSize >= srcSize - minGain))- return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);- if (cLitSize==1)- return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);-- /* Build header */- switch(lhSize)- {- case 3: /* 2 - 2 - 10 - 10 */- { U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14);- MEM_writeLE24(ostart, lhc);- break;- }- case 4: /* 2 - 2 - 14 - 14 */- { U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18);- MEM_writeLE32(ostart, lhc);- break;- }- default: /* should not be necessary, lhSize is only {3,4,5} */- case 5: /* 2 - 2 - 18 - 18 */- { U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22);- MEM_writeLE32(ostart, lhc);- ostart[4] = (BYTE)(cLitSize >> 10);- break;- }- }- return lhSize+cLitSize;-}--static const BYTE LL_Code[64] = { 0, 1, 2, 3, 4, 5, 6, 7,- 8, 9, 10, 11, 12, 13, 14, 15,- 16, 16, 17, 17, 18, 18, 19, 19,- 20, 20, 20, 20, 21, 21, 21, 21,- 22, 22, 22, 22, 22, 22, 22, 22,- 23, 23, 23, 23, 23, 23, 23, 23,- 24, 24, 24, 24, 24, 24, 24, 24,- 24, 24, 24, 24, 24, 24, 24, 24 };--static const BYTE ML_Code[128] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,- 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,- 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37,- 38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39,- 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,- 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41,- 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42,- 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 };---void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)-{- BYTE const LL_deltaCode = 19;- BYTE const ML_deltaCode = 36;- const seqDef* const sequences = seqStorePtr->sequencesStart;- BYTE* const llCodeTable = seqStorePtr->llCode;- BYTE* const ofCodeTable = seqStorePtr->ofCode;- BYTE* const mlCodeTable = seqStorePtr->mlCode;- U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);- U32 u;- for (u=0; u<nbSeq; u++) {- U32 const llv = sequences[u].litLength;- U32 const mlv = sequences[u].matchLength;- llCodeTable[u] = (llv> 63) ? (BYTE)ZSTD_highbit32(llv) + LL_deltaCode : LL_Code[llv];- ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset);- mlCodeTable[u] = (mlv>127) ? (BYTE)ZSTD_highbit32(mlv) + ML_deltaCode : ML_Code[mlv];- }- if (seqStorePtr->longLengthID==1)- llCodeTable[seqStorePtr->longLengthPos] = MaxLL;- if (seqStorePtr->longLengthID==2)- mlCodeTable[seqStorePtr->longLengthPos] = MaxML;-}---size_t ZSTD_compressSequences(ZSTD_CCtx* zc,- void* dst, size_t dstCapacity,- size_t srcSize)-{- const seqStore_t* seqStorePtr = &(zc->seqStore);- U32 count[MaxSeq+1];- S16 norm[MaxSeq+1];- FSE_CTable* CTable_LitLength = zc->litlengthCTable;- FSE_CTable* CTable_OffsetBits = zc->offcodeCTable;- FSE_CTable* CTable_MatchLength = zc->matchlengthCTable;- U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */- const seqDef* const sequences = seqStorePtr->sequencesStart;- const BYTE* const ofCodeTable = seqStorePtr->ofCode;- const BYTE* const llCodeTable = seqStorePtr->llCode;- const BYTE* const mlCodeTable = seqStorePtr->mlCode;- BYTE* const ostart = (BYTE*)dst;- BYTE* const oend = ostart + dstCapacity;- BYTE* op = ostart;- size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;- BYTE* seqHead;- BYTE scratchBuffer[1<<MAX(MLFSELog,LLFSELog)];-- /* Compress literals */- { const BYTE* const literals = seqStorePtr->litStart;- size_t const litSize = seqStorePtr->lit - literals;- size_t const cSize = ZSTD_compressLiterals(zc, op, dstCapacity, literals, litSize);- if (ZSTD_isError(cSize)) return cSize;- op += cSize;- }-- /* Sequences Header */- if ((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead */) return ERROR(dstSize_tooSmall);- if (nbSeq < 0x7F) *op++ = (BYTE)nbSeq;- else if (nbSeq < LONGNBSEQ) op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;- else op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;- if (nbSeq==0) goto _check_compressibility;-- /* seqHead : flags for FSE encoding type */- seqHead = op++;--#define MIN_SEQ_FOR_DYNAMIC_FSE 64-#define MAX_SEQ_FOR_STATIC_FSE 1000-- /* convert length/distances into codes */- ZSTD_seqToCodes(seqStorePtr);-- /* CTable for Literal Lengths */- { U32 max = MaxLL;- size_t const mostFrequent = FSE_countFast_wksp(count, &max, llCodeTable, nbSeq, zc->tmpCounters);- if ((mostFrequent == nbSeq) && (nbSeq > 2)) {- *op++ = llCodeTable[0];- FSE_buildCTable_rle(CTable_LitLength, (BYTE)max);- LLtype = set_rle;- } else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {- LLtype = set_repeat;- } else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (LL_defaultNormLog-1)))) {- FSE_buildCTable_wksp(CTable_LitLength, LL_defaultNorm, MaxLL, LL_defaultNormLog, scratchBuffer, sizeof(scratchBuffer));- LLtype = set_basic;- } else {- size_t nbSeq_1 = nbSeq;- const U32 tableLog = FSE_optimalTableLog(LLFSELog, nbSeq, max);- if (count[llCodeTable[nbSeq-1]]>1) { count[llCodeTable[nbSeq-1]]--; nbSeq_1--; }- FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);- { size_t const NCountSize = FSE_writeNCount(op, oend-op, norm, max, tableLog); /* overflow protected */- if (FSE_isError(NCountSize)) return ERROR(GENERIC);- op += NCountSize; }- FSE_buildCTable_wksp(CTable_LitLength, norm, max, tableLog, scratchBuffer, sizeof(scratchBuffer));- LLtype = set_compressed;- } }-- /* CTable for Offsets */- { U32 max = MaxOff;- size_t const mostFrequent = FSE_countFast_wksp(count, &max, ofCodeTable, nbSeq, zc->tmpCounters);- if ((mostFrequent == nbSeq) && (nbSeq > 2)) {- *op++ = ofCodeTable[0];- FSE_buildCTable_rle(CTable_OffsetBits, (BYTE)max);- Offtype = set_rle;- } else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {- Offtype = set_repeat;- } else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (OF_defaultNormLog-1)))) {- FSE_buildCTable_wksp(CTable_OffsetBits, OF_defaultNorm, MaxOff, OF_defaultNormLog, scratchBuffer, sizeof(scratchBuffer));- Offtype = set_basic;- } else {- size_t nbSeq_1 = nbSeq;- const U32 tableLog = FSE_optimalTableLog(OffFSELog, nbSeq, max);- if (count[ofCodeTable[nbSeq-1]]>1) { count[ofCodeTable[nbSeq-1]]--; nbSeq_1--; }- FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);- { size_t const NCountSize = FSE_writeNCount(op, oend-op, norm, max, tableLog); /* overflow protected */- if (FSE_isError(NCountSize)) return ERROR(GENERIC);- op += NCountSize; }- FSE_buildCTable_wksp(CTable_OffsetBits, norm, max, tableLog, scratchBuffer, sizeof(scratchBuffer));- Offtype = set_compressed;- } }-- /* CTable for MatchLengths */- { U32 max = MaxML;- size_t const mostFrequent = FSE_countFast_wksp(count, &max, mlCodeTable, nbSeq, zc->tmpCounters);- if ((mostFrequent == nbSeq) && (nbSeq > 2)) {- *op++ = *mlCodeTable;- FSE_buildCTable_rle(CTable_MatchLength, (BYTE)max);- MLtype = set_rle;- } else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {- MLtype = set_repeat;- } else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (ML_defaultNormLog-1)))) {- FSE_buildCTable_wksp(CTable_MatchLength, ML_defaultNorm, MaxML, ML_defaultNormLog, scratchBuffer, sizeof(scratchBuffer));- MLtype = set_basic;- } else {- size_t nbSeq_1 = nbSeq;- const U32 tableLog = FSE_optimalTableLog(MLFSELog, nbSeq, max);- if (count[mlCodeTable[nbSeq-1]]>1) { count[mlCodeTable[nbSeq-1]]--; nbSeq_1--; }- FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);- { size_t const NCountSize = FSE_writeNCount(op, oend-op, norm, max, tableLog); /* overflow protected */- if (FSE_isError(NCountSize)) return ERROR(GENERIC);- op += NCountSize; }- FSE_buildCTable_wksp(CTable_MatchLength, norm, max, tableLog, scratchBuffer, sizeof(scratchBuffer));- MLtype = set_compressed;- } }-- *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2));- zc->flagStaticTables = 0;-- /* Encoding Sequences */- { BIT_CStream_t blockStream;- FSE_CState_t stateMatchLength;- FSE_CState_t stateOffsetBits;- FSE_CState_t stateLitLength;-- CHECK_E(BIT_initCStream(&blockStream, op, oend-op), dstSize_tooSmall); /* not enough space remaining */-- /* first symbols */- FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]);- FSE_initCState2(&stateOffsetBits, CTable_OffsetBits, ofCodeTable[nbSeq-1]);- FSE_initCState2(&stateLitLength, CTable_LitLength, llCodeTable[nbSeq-1]);- BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]);- if (MEM_32bits()) BIT_flushBits(&blockStream);- BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]);- if (MEM_32bits()) BIT_flushBits(&blockStream);- BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]);- BIT_flushBits(&blockStream);-- { size_t n;- for (n=nbSeq-2 ; n<nbSeq ; n--) { /* intentional underflow */- BYTE const llCode = llCodeTable[n];- BYTE const ofCode = ofCodeTable[n];- BYTE const mlCode = mlCodeTable[n];- U32 const llBits = LL_bits[llCode];- U32 const ofBits = ofCode; /* 32b*/ /* 64b*/- U32 const mlBits = ML_bits[mlCode];- /* (7)*/ /* (7)*/- FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode); /* 15 */ /* 15 */- FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode); /* 24 */ /* 24 */- if (MEM_32bits()) BIT_flushBits(&blockStream); /* (7)*/- FSE_encodeSymbol(&blockStream, &stateLitLength, llCode); /* 16 */ /* 33 */- if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog)))- BIT_flushBits(&blockStream); /* (7)*/- BIT_addBits(&blockStream, sequences[n].litLength, llBits);- if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);- BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);- if (MEM_32bits()) BIT_flushBits(&blockStream); /* (7)*/- BIT_addBits(&blockStream, sequences[n].offset, ofBits); /* 31 */- BIT_flushBits(&blockStream); /* (7)*/- } }-- FSE_flushCState(&blockStream, &stateMatchLength);- FSE_flushCState(&blockStream, &stateOffsetBits);- FSE_flushCState(&blockStream, &stateLitLength);-- { size_t const streamSize = BIT_closeCStream(&blockStream);- if (streamSize==0) return ERROR(dstSize_tooSmall); /* not enough space */- op += streamSize;- } }-- /* check compressibility */-_check_compressibility:- { size_t const minGain = ZSTD_minGain(srcSize);- size_t const maxCSize = srcSize - minGain;- if ((size_t)(op-ostart) >= maxCSize) return 0; }-- /* confirm repcodes */- { int i; for (i=0; i<ZSTD_REP_NUM; i++) zc->rep[i] = zc->savedRep[i]; }-- return op - ostart;-}---/*! ZSTD_storeSeq() :- Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.- `offsetCode` : distance to match, or 0 == repCode.- `matchCode` : matchLength - MINMATCH-*/-MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t matchCode)-{-#if 0 /* for debug */- static const BYTE* g_start = NULL;- const U32 pos = (U32)((const BYTE*)literals - g_start);- if (g_start==NULL) g_start = (const BYTE*)literals;- //if ((pos > 1) && (pos < 50000))- printf("Cpos %6u :%5u literals & match %3u bytes at distance %6u \n",- pos, (U32)litLength, (U32)matchCode+MINMATCH, (U32)offsetCode);-#endif- /* copy Literals */- ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);- seqStorePtr->lit += litLength;-- /* literal Length */- if (litLength>0xFFFF) { seqStorePtr->longLengthID = 1; seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); }- seqStorePtr->sequences[0].litLength = (U16)litLength;-- /* match offset */- seqStorePtr->sequences[0].offset = offsetCode + 1;-- /* match Length */- if (matchCode>0xFFFF) { seqStorePtr->longLengthID = 2; seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); }- seqStorePtr->sequences[0].matchLength = (U16)matchCode;-- seqStorePtr->sequences++;-}---/*-*************************************-* Match length counter-***************************************/-static unsigned ZSTD_NbCommonBytes (register size_t val)-{- if (MEM_isLittleEndian()) {- if (MEM_64bits()) {-# if defined(_MSC_VER) && defined(_WIN64)- unsigned long r = 0;- _BitScanForward64( &r, (U64)val );- return (unsigned)(r>>3);-# elif defined(__GNUC__) && (__GNUC__ >= 3)- return (__builtin_ctzll((U64)val) >> 3);-# else- static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, 0, 3, 1, 3, 1, 4, 2, 7, 0, 2, 3, 6, 1, 5, 3, 5, 1, 3, 4, 4, 2, 5, 6, 7, 7, 0, 1, 2, 3, 3, 4, 6, 2, 6, 5, 5, 3, 4, 5, 6, 7, 1, 2, 4, 6, 4, 4, 5, 7, 2, 6, 5, 7, 6, 7, 7 };- return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];-# endif- } else { /* 32 bits */-# if defined(_MSC_VER)- unsigned long r=0;- _BitScanForward( &r, (U32)val );- return (unsigned)(r>>3);-# elif defined(__GNUC__) && (__GNUC__ >= 3)- return (__builtin_ctz((U32)val) >> 3);-# else- static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, 3, 2, 2, 1, 3, 2, 0, 1, 3, 3, 1, 2, 2, 2, 2, 0, 3, 1, 2, 0, 1, 0, 1, 1 };- return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];-# endif- }- } else { /* Big Endian CPU */- if (MEM_64bits()) {-# if defined(_MSC_VER) && defined(_WIN64)- unsigned long r = 0;- _BitScanReverse64( &r, val );- return (unsigned)(r>>3);-# elif defined(__GNUC__) && (__GNUC__ >= 3)- return (__builtin_clzll(val) >> 3);-# else- unsigned r;- const unsigned n32 = sizeof(size_t)*4; /* calculate this way due to compiler complaining in 32-bits mode */- if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; }- if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }- r += (!val);- return r;-# endif- } else { /* 32 bits */-# if defined(_MSC_VER)- unsigned long r = 0;- _BitScanReverse( &r, (unsigned long)val );- return (unsigned)(r>>3);-# elif defined(__GNUC__) && (__GNUC__ >= 3)- return (__builtin_clz((U32)val) >> 3);-# else- unsigned r;- if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }- r += (!val);- return r;-# endif- } }-}---static size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit)-{- const BYTE* const pStart = pIn;- const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1);-- while (pIn < pInLoopLimit) {- size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);- if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; }- pIn += ZSTD_NbCommonBytes(diff);- return (size_t)(pIn - pStart);- }- if (MEM_64bits()) if ((pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; }- if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; }- if ((pIn<pInLimit) && (*pMatch == *pIn)) pIn++;- return (size_t)(pIn - pStart);-}--/** ZSTD_count_2segments() :-* can count match length with `ip` & `match` in 2 different segments.-* convention : on reaching mEnd, match count continue starting from iStart-*/-static size_t ZSTD_count_2segments(const BYTE* ip, const BYTE* match, const BYTE* iEnd, const BYTE* mEnd, const BYTE* iStart)-{- const BYTE* const vEnd = MIN( ip + (mEnd - match), iEnd);- size_t const matchLength = ZSTD_count(ip, match, vEnd);- if (match + matchLength != mEnd) return matchLength;- return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd);-}---/*-*************************************-* Hashes-***************************************/-static const U32 prime3bytes = 506832829U;-static U32 ZSTD_hash3(U32 u, U32 h) { return ((u << (32-24)) * prime3bytes) >> (32-h) ; }-MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h); } /* only in zstd_opt.h */--static const U32 prime4bytes = 2654435761U;-static U32 ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32-h) ; }-static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_read32(ptr), h); }--static const U64 prime5bytes = 889523592379ULL;-static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u << (64-40)) * prime5bytes) >> (64-h)) ; }-static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h); }--static const U64 prime6bytes = 227718039650203ULL;-static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64-48)) * prime6bytes) >> (64-h)) ; }-static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }--static const U64 prime7bytes = 58295818150454627ULL;-static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u << (64-56)) * prime7bytes) >> (64-h)) ; }-static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h); }--static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;-static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }-static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }--static size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)-{- switch(mls)- {- default:- case 4: return ZSTD_hash4Ptr(p, hBits);- case 5: return ZSTD_hash5Ptr(p, hBits);- case 6: return ZSTD_hash6Ptr(p, hBits);- case 7: return ZSTD_hash7Ptr(p, hBits);- case 8: return ZSTD_hash8Ptr(p, hBits);- }-}---/*-*************************************-* Fast Scan-***************************************/-static void ZSTD_fillHashTable (ZSTD_CCtx* zc, const void* end, const U32 mls)-{- U32* const hashTable = zc->hashTable;- U32 const hBits = zc->params.cParams.hashLog;- const BYTE* const base = zc->base;- const BYTE* ip = base + zc->nextToUpdate;- const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;- const size_t fastHashFillStep = 3;-- while(ip <= iend) {- hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);- ip += fastHashFillStep;- }-}---FORCE_INLINE-void ZSTD_compressBlock_fast_generic(ZSTD_CCtx* cctx,- const void* src, size_t srcSize,- const U32 mls)-{- U32* const hashTable = cctx->hashTable;- U32 const hBits = cctx->params.cParams.hashLog;- seqStore_t* seqStorePtr = &(cctx->seqStore);- const BYTE* const base = cctx->base;- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const U32 lowestIndex = cctx->dictLimit;- const BYTE* const lowest = base + lowestIndex;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - HASH_READ_SIZE;- U32 offset_1=cctx->rep[0], offset_2=cctx->rep[1];- U32 offsetSaved = 0;-- /* init */- ip += (ip==lowest);- { U32 const maxRep = (U32)(ip-lowest);- if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;- if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;- }-- /* Main Search Loop */- while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */- size_t mLength;- size_t const h = ZSTD_hashPtr(ip, hBits, mls);- U32 const current = (U32)(ip-base);- U32 const matchIndex = hashTable[h];- const BYTE* match = base + matchIndex;- hashTable[h] = current; /* update hash table */-- if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {- mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;- ip++;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);- } else {- U32 offset;- if ( (matchIndex <= lowestIndex) || (MEM_read32(match) != MEM_read32(ip)) ) {- ip += ((ip-anchor) >> g_searchStrength) + 1;- continue;- }- mLength = ZSTD_count(ip+4, match+4, iend) + 4;- offset = (U32)(ip-match);- while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */- offset_2 = offset_1;- offset_1 = offset;-- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);- }-- /* match found */- ip += mLength;- anchor = ip;-- if (ip <= ilimit) {- /* Fill Table */- hashTable[ZSTD_hashPtr(base+current+2, hBits, mls)] = current+2; /* here because current+2 could be > iend-8 */- hashTable[ZSTD_hashPtr(ip-2, hBits, mls)] = (U32)(ip-2-base);- /* check immediate repcode */- while ( (ip <= ilimit)- && ( (offset_2>0)- & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {- /* store sequence */- size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;- { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */- hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip-base);- ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);- ip += rLength;- anchor = ip;- continue; /* faster when present ... (?) */- } } }-- /* save reps for next block */- cctx->savedRep[0] = offset_1 ? offset_1 : offsetSaved;- cctx->savedRep[1] = offset_2 ? offset_2 : offsetSaved;-- /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}---static void ZSTD_compressBlock_fast(ZSTD_CCtx* ctx,- const void* src, size_t srcSize)-{- const U32 mls = ctx->params.cParams.searchLength;- switch(mls)- {- default:- case 4 :- ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 4); return;- case 5 :- ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 5); return;- case 6 :- ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 6); return;- case 7 :- ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 7); return;- }-}---static void ZSTD_compressBlock_fast_extDict_generic(ZSTD_CCtx* ctx,- const void* src, size_t srcSize,- const U32 mls)-{- U32* hashTable = ctx->hashTable;- const U32 hBits = ctx->params.cParams.hashLog;- seqStore_t* seqStorePtr = &(ctx->seqStore);- const BYTE* const base = ctx->base;- const BYTE* const dictBase = ctx->dictBase;- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const U32 lowestIndex = ctx->lowLimit;- const BYTE* const dictStart = dictBase + lowestIndex;- const U32 dictLimit = ctx->dictLimit;- const BYTE* const lowPrefixPtr = base + dictLimit;- const BYTE* const dictEnd = dictBase + dictLimit;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - 8;- U32 offset_1=ctx->rep[0], offset_2=ctx->rep[1];-- /* Search Loop */- while (ip < ilimit) { /* < instead of <=, because (ip+1) */- const size_t h = ZSTD_hashPtr(ip, hBits, mls);- const U32 matchIndex = hashTable[h];- const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;- const BYTE* match = matchBase + matchIndex;- const U32 current = (U32)(ip-base);- const U32 repIndex = current + 1 - offset_1; /* offset_1 expected <= current +1 */- const BYTE* repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* repMatch = repBase + repIndex;- size_t mLength;- hashTable[h] = current; /* update hash table */-- if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))- && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {- const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;- mLength = ZSTD_count_2segments(ip+1+EQUAL_READ32, repMatch+EQUAL_READ32, iend, repMatchEnd, lowPrefixPtr) + EQUAL_READ32;- ip++;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);- } else {- if ( (matchIndex < lowestIndex) ||- (MEM_read32(match) != MEM_read32(ip)) ) {- ip += ((ip-anchor) >> g_searchStrength) + 1;- continue;- }- { const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;- const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;- U32 offset;- mLength = ZSTD_count_2segments(ip+EQUAL_READ32, match+EQUAL_READ32, iend, matchEnd, lowPrefixPtr) + EQUAL_READ32;- while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */- offset = current - matchIndex;- offset_2 = offset_1;- offset_1 = offset;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);- } }-- /* found a match : store it */- ip += mLength;- anchor = ip;-- if (ip <= ilimit) {- /* Fill Table */- hashTable[ZSTD_hashPtr(base+current+2, hBits, mls)] = current+2;- hashTable[ZSTD_hashPtr(ip-2, hBits, mls)] = (U32)(ip-2-base);- /* check immediate repcode */- while (ip <= ilimit) {- U32 const current2 = (U32)(ip-base);- U32 const repIndex2 = current2 - offset_2;- const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;- if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */- && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {- const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;- size_t repLength2 = ZSTD_count_2segments(ip+EQUAL_READ32, repMatch2+EQUAL_READ32, iend, repEnd2, lowPrefixPtr) + EQUAL_READ32;- U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */- ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);- hashTable[ZSTD_hashPtr(ip, hBits, mls)] = current2;- ip += repLength2;- anchor = ip;- continue;- }- break;- } } }-- /* save reps for next block */- ctx->savedRep[0] = offset_1; ctx->savedRep[1] = offset_2;-- /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}---static void ZSTD_compressBlock_fast_extDict(ZSTD_CCtx* ctx,- const void* src, size_t srcSize)-{- U32 const mls = ctx->params.cParams.searchLength;- switch(mls)- {- default:- case 4 :- ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 4); return;- case 5 :- ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 5); return;- case 6 :- ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 6); return;- case 7 :- ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 7); return;- }-}---/*-*************************************-* Double Fast-***************************************/-static void ZSTD_fillDoubleHashTable (ZSTD_CCtx* cctx, const void* end, const U32 mls)-{- U32* const hashLarge = cctx->hashTable;- U32 const hBitsL = cctx->params.cParams.hashLog;- U32* const hashSmall = cctx->chainTable;- U32 const hBitsS = cctx->params.cParams.chainLog;- const BYTE* const base = cctx->base;- const BYTE* ip = base + cctx->nextToUpdate;- const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;- const size_t fastHashFillStep = 3;-- while(ip <= iend) {- hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);- hashLarge[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);- ip += fastHashFillStep;- }-}---FORCE_INLINE-void ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx* cctx,- const void* src, size_t srcSize,- const U32 mls)-{- U32* const hashLong = cctx->hashTable;- const U32 hBitsL = cctx->params.cParams.hashLog;- U32* const hashSmall = cctx->chainTable;- const U32 hBitsS = cctx->params.cParams.chainLog;- seqStore_t* seqStorePtr = &(cctx->seqStore);- const BYTE* const base = cctx->base;- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const U32 lowestIndex = cctx->dictLimit;- const BYTE* const lowest = base + lowestIndex;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - HASH_READ_SIZE;- U32 offset_1=cctx->rep[0], offset_2=cctx->rep[1];- U32 offsetSaved = 0;-- /* init */- ip += (ip==lowest);- { U32 const maxRep = (U32)(ip-lowest);- if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;- if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;- }-- /* Main Search Loop */- while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */- size_t mLength;- size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);- size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);- U32 const current = (U32)(ip-base);- U32 const matchIndexL = hashLong[h2];- U32 const matchIndexS = hashSmall[h];- const BYTE* matchLong = base + matchIndexL;- const BYTE* match = base + matchIndexS;- hashLong[h2] = hashSmall[h] = current; /* update hash tables */-- if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) { /* note : by construction, offset_1 <= current */- mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;- ip++;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);- } else {- U32 offset;- if ( (matchIndexL > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip)) ) {- mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;- offset = (U32)(ip-matchLong);- while (((ip>anchor) & (matchLong>lowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */- } else if ( (matchIndexS > lowestIndex) && (MEM_read32(match) == MEM_read32(ip)) ) {- size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);- U32 const matchIndex3 = hashLong[h3];- const BYTE* match3 = base + matchIndex3;- hashLong[h3] = current + 1;- if ( (matchIndex3 > lowestIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {- mLength = ZSTD_count(ip+9, match3+8, iend) + 8;- ip++;- offset = (U32)(ip-match3);- while (((ip>anchor) & (match3>lowest)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */- } else {- mLength = ZSTD_count(ip+4, match+4, iend) + 4;- offset = (U32)(ip-match);- while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */- }- } else {- ip += ((ip-anchor) >> g_searchStrength) + 1;- continue;- }-- offset_2 = offset_1;- offset_1 = offset;-- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);- }-- /* match found */- ip += mLength;- anchor = ip;-- if (ip <= ilimit) {- /* Fill Table */- hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] =- hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2; /* here because current+2 could be > iend-8 */- hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] =- hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);-- /* check immediate repcode */- while ( (ip <= ilimit)- && ( (offset_2>0)- & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {- /* store sequence */- size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;- { U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */- hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);- hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);- ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);- ip += rLength;- anchor = ip;- continue; /* faster when present ... (?) */- } } }-- /* save reps for next block */- cctx->savedRep[0] = offset_1 ? offset_1 : offsetSaved;- cctx->savedRep[1] = offset_2 ? offset_2 : offsetSaved;-- /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}---static void ZSTD_compressBlock_doubleFast(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- const U32 mls = ctx->params.cParams.searchLength;- switch(mls)- {- default:- case 4 :- ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 4); return;- case 5 :- ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 5); return;- case 6 :- ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 6); return;- case 7 :- ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 7); return;- }-}---static void ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx* ctx,- const void* src, size_t srcSize,- const U32 mls)-{- U32* const hashLong = ctx->hashTable;- U32 const hBitsL = ctx->params.cParams.hashLog;- U32* const hashSmall = ctx->chainTable;- U32 const hBitsS = ctx->params.cParams.chainLog;- seqStore_t* seqStorePtr = &(ctx->seqStore);- const BYTE* const base = ctx->base;- const BYTE* const dictBase = ctx->dictBase;- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const U32 lowestIndex = ctx->lowLimit;- const BYTE* const dictStart = dictBase + lowestIndex;- const U32 dictLimit = ctx->dictLimit;- const BYTE* const lowPrefixPtr = base + dictLimit;- const BYTE* const dictEnd = dictBase + dictLimit;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - 8;- U32 offset_1=ctx->rep[0], offset_2=ctx->rep[1];-- /* Search Loop */- while (ip < ilimit) { /* < instead of <=, because (ip+1) */- const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);- const U32 matchIndex = hashSmall[hSmall];- const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;- const BYTE* match = matchBase + matchIndex;-- const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);- const U32 matchLongIndex = hashLong[hLong];- const BYTE* matchLongBase = matchLongIndex < dictLimit ? dictBase : base;- const BYTE* matchLong = matchLongBase + matchLongIndex;-- const U32 current = (U32)(ip-base);- const U32 repIndex = current + 1 - offset_1; /* offset_1 expected <= current +1 */- const BYTE* repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* repMatch = repBase + repIndex;- size_t mLength;- hashSmall[hSmall] = hashLong[hLong] = current; /* update hash table */-- if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))- && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {- const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;- mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, lowPrefixPtr) + 4;- ip++;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);- } else {- if ((matchLongIndex > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {- const BYTE* matchEnd = matchLongIndex < dictLimit ? dictEnd : iend;- const BYTE* lowMatchPtr = matchLongIndex < dictLimit ? dictStart : lowPrefixPtr;- U32 offset;- mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, lowPrefixPtr) + 8;- offset = current - matchLongIndex;- while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */- offset_2 = offset_1;- offset_1 = offset;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);-- } else if ((matchIndex > lowestIndex) && (MEM_read32(match) == MEM_read32(ip))) {- size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);- U32 const matchIndex3 = hashLong[h3];- const BYTE* const match3Base = matchIndex3 < dictLimit ? dictBase : base;- const BYTE* match3 = match3Base + matchIndex3;- U32 offset;- hashLong[h3] = current + 1;- if ( (matchIndex3 > lowestIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {- const BYTE* matchEnd = matchIndex3 < dictLimit ? dictEnd : iend;- const BYTE* lowMatchPtr = matchIndex3 < dictLimit ? dictStart : lowPrefixPtr;- mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, lowPrefixPtr) + 8;- ip++;- offset = current+1 - matchIndex3;- while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */- } else {- const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;- const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;- mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, lowPrefixPtr) + 4;- offset = current - matchIndex;- while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */- }- offset_2 = offset_1;- offset_1 = offset;- ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);-- } else {- ip += ((ip-anchor) >> g_searchStrength) + 1;- continue;- } }-- /* found a match : store it */- ip += mLength;- anchor = ip;-- if (ip <= ilimit) {- /* Fill Table */- hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;- hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2;- hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);- hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);- /* check immediate repcode */- while (ip <= ilimit) {- U32 const current2 = (U32)(ip-base);- U32 const repIndex2 = current2 - offset_2;- const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;- if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */- && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {- const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;- size_t const repLength2 = ZSTD_count_2segments(ip+EQUAL_READ32, repMatch2+EQUAL_READ32, iend, repEnd2, lowPrefixPtr) + EQUAL_READ32;- U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */- ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);- hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;- hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;- ip += repLength2;- anchor = ip;- continue;- }- break;- } } }-- /* save reps for next block */- ctx->savedRep[0] = offset_1; ctx->savedRep[1] = offset_2;-- /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}---static void ZSTD_compressBlock_doubleFast_extDict(ZSTD_CCtx* ctx,- const void* src, size_t srcSize)-{- U32 const mls = ctx->params.cParams.searchLength;- switch(mls)- {- default:- case 4 :- ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 4); return;- case 5 :- ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 5); return;- case 6 :- ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 6); return;- case 7 :- ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 7); return;- }-}---/*-*************************************-* Binary Tree search-***************************************/-/** ZSTD_insertBt1() : add one or multiple positions to tree.-* ip : assumed <= iend-8 .-* @return : nb of positions added */-static U32 ZSTD_insertBt1(ZSTD_CCtx* zc, const BYTE* const ip, const U32 mls, const BYTE* const iend, U32 nbCompares,- U32 extDict)-{- U32* const hashTable = zc->hashTable;- U32 const hashLog = zc->params.cParams.hashLog;- size_t const h = ZSTD_hashPtr(ip, hashLog, mls);- U32* const bt = zc->chainTable;- U32 const btLog = zc->params.cParams.chainLog - 1;- U32 const btMask = (1 << btLog) - 1;- U32 matchIndex = hashTable[h];- size_t commonLengthSmaller=0, commonLengthLarger=0;- const BYTE* const base = zc->base;- const BYTE* const dictBase = zc->dictBase;- const U32 dictLimit = zc->dictLimit;- const BYTE* const dictEnd = dictBase + dictLimit;- const BYTE* const prefixStart = base + dictLimit;- const BYTE* match;- const U32 current = (U32)(ip-base);- const U32 btLow = btMask >= current ? 0 : current - btMask;- U32* smallerPtr = bt + 2*(current&btMask);- U32* largerPtr = smallerPtr + 1;- U32 dummy32; /* to be nullified at the end */- U32 const windowLow = zc->lowLimit;- U32 matchEndIdx = current+8;- size_t bestLength = 8;-#ifdef ZSTD_C_PREDICT- U32 predictedSmall = *(bt + 2*((current-1)&btMask) + 0);- U32 predictedLarge = *(bt + 2*((current-1)&btMask) + 1);- predictedSmall += (predictedSmall>0);- predictedLarge += (predictedLarge>0);-#endif /* ZSTD_C_PREDICT */-- hashTable[h] = current; /* Update Hash Table */-- while (nbCompares-- && (matchIndex > windowLow)) {- U32* const nextPtr = bt + 2*(matchIndex & btMask);- size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */--#ifdef ZSTD_C_PREDICT /* note : can create issues when hlog small <= 11 */- const U32* predictPtr = bt + 2*((matchIndex-1) & btMask); /* written this way, as bt is a roll buffer */- if (matchIndex == predictedSmall) {- /* no need to check length, result known */- *smallerPtr = matchIndex;- if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */- smallerPtr = nextPtr+1; /* new "smaller" => larger of match */- matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */- predictedSmall = predictPtr[1] + (predictPtr[1]>0);- continue;- }- if (matchIndex == predictedLarge) {- *largerPtr = matchIndex;- if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */- largerPtr = nextPtr;- matchIndex = nextPtr[0];- predictedLarge = predictPtr[0] + (predictPtr[0]>0);- continue;- }-#endif- if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {- match = base + matchIndex;- if (match[matchLength] == ip[matchLength])- matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;- } else {- match = dictBase + matchIndex;- matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);- if (matchIndex+matchLength >= dictLimit)- match = base + matchIndex; /* to prepare for next usage of match[matchLength] */- }-- if (matchLength > bestLength) {- bestLength = matchLength;- if (matchLength > matchEndIdx - matchIndex)- matchEndIdx = matchIndex + (U32)matchLength;- }-- if (ip+matchLength == iend) /* equal : no way to know if inf or sup */- break; /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt the tree */-- if (match[matchLength] < ip[matchLength]) { /* necessarily within correct buffer */- /* match is smaller than current */- *smallerPtr = matchIndex; /* update smaller idx */- commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */- if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */- smallerPtr = nextPtr+1; /* new "smaller" => larger of match */- matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */- } else {- /* match is larger than current */- *largerPtr = matchIndex;- commonLengthLarger = matchLength;- if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */- largerPtr = nextPtr;- matchIndex = nextPtr[0];- } }-- *smallerPtr = *largerPtr = 0;- if (bestLength > 384) return MIN(192, (U32)(bestLength - 384)); /* speed optimization */- if (matchEndIdx > current + 8) return matchEndIdx - current - 8;- return 1;-}---static size_t ZSTD_insertBtAndFindBestMatch (- ZSTD_CCtx* zc,- const BYTE* const ip, const BYTE* const iend,- size_t* offsetPtr,- U32 nbCompares, const U32 mls,- U32 extDict)-{- U32* const hashTable = zc->hashTable;- U32 const hashLog = zc->params.cParams.hashLog;- size_t const h = ZSTD_hashPtr(ip, hashLog, mls);- U32* const bt = zc->chainTable;- U32 const btLog = zc->params.cParams.chainLog - 1;- U32 const btMask = (1 << btLog) - 1;- U32 matchIndex = hashTable[h];- size_t commonLengthSmaller=0, commonLengthLarger=0;- const BYTE* const base = zc->base;- const BYTE* const dictBase = zc->dictBase;- const U32 dictLimit = zc->dictLimit;- const BYTE* const dictEnd = dictBase + dictLimit;- const BYTE* const prefixStart = base + dictLimit;- const U32 current = (U32)(ip-base);- const U32 btLow = btMask >= current ? 0 : current - btMask;- const U32 windowLow = zc->lowLimit;- U32* smallerPtr = bt + 2*(current&btMask);- U32* largerPtr = bt + 2*(current&btMask) + 1;- U32 matchEndIdx = current+8;- U32 dummy32; /* to be nullified at the end */- size_t bestLength = 0;-- hashTable[h] = current; /* Update Hash Table */-- while (nbCompares-- && (matchIndex > windowLow)) {- U32* const nextPtr = bt + 2*(matchIndex & btMask);- size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */- const BYTE* match;-- if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {- match = base + matchIndex;- if (match[matchLength] == ip[matchLength])- matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;- } else {- match = dictBase + matchIndex;- matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);- if (matchIndex+matchLength >= dictLimit)- match = base + matchIndex; /* to prepare for next usage of match[matchLength] */- }-- if (matchLength > bestLength) {- if (matchLength > matchEndIdx - matchIndex)- matchEndIdx = matchIndex + (U32)matchLength;- if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )- bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;- if (ip+matchLength == iend) /* equal : no way to know if inf or sup */- break; /* drop, to guarantee consistency (miss a little bit of compression) */- }-- if (match[matchLength] < ip[matchLength]) {- /* match is smaller than current */- *smallerPtr = matchIndex; /* update smaller idx */- commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */- if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */- smallerPtr = nextPtr+1; /* new "smaller" => larger of match */- matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */- } else {- /* match is larger than current */- *largerPtr = matchIndex;- commonLengthLarger = matchLength;- if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */- largerPtr = nextPtr;- matchIndex = nextPtr[0];- } }-- *smallerPtr = *largerPtr = 0;-- zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;- return bestLength;-}---static void ZSTD_updateTree(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls)-{- const BYTE* const base = zc->base;- const U32 target = (U32)(ip - base);- U32 idx = zc->nextToUpdate;-- while(idx < target)- idx += ZSTD_insertBt1(zc, base+idx, mls, iend, nbCompares, 0);-}--/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */-static size_t ZSTD_BtFindBestMatch (- ZSTD_CCtx* zc,- const BYTE* const ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 mls)-{- if (ip < zc->base + zc->nextToUpdate) return 0; /* skipped area */- ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);- return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 0);-}---static size_t ZSTD_BtFindBestMatch_selectMLS (- ZSTD_CCtx* zc, /* Index table will be updated */- const BYTE* ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 matchLengthSearch)-{- switch(matchLengthSearch)- {- default :- case 4 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);- case 5 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);- case 6 : return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);- }-}---static void ZSTD_updateTree_extDict(ZSTD_CCtx* zc, const BYTE* const ip, const BYTE* const iend, const U32 nbCompares, const U32 mls)-{- const BYTE* const base = zc->base;- const U32 target = (U32)(ip - base);- U32 idx = zc->nextToUpdate;-- while (idx < target) idx += ZSTD_insertBt1(zc, base+idx, mls, iend, nbCompares, 1);-}---/** Tree updater, providing best match */-static size_t ZSTD_BtFindBestMatch_extDict (- ZSTD_CCtx* zc,- const BYTE* const ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 mls)-{- if (ip < zc->base + zc->nextToUpdate) return 0; /* skipped area */- ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);- return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 1);-}---static size_t ZSTD_BtFindBestMatch_selectMLS_extDict (- ZSTD_CCtx* zc, /* Index table will be updated */- const BYTE* ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 matchLengthSearch)-{- switch(matchLengthSearch)- {- default :- case 4 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);- case 5 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);- case 6 : return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);- }-}----/* *********************************-* Hash Chain-***********************************/-#define NEXT_IN_CHAIN(d, mask) chainTable[(d) & mask]--/* Update chains up to ip (excluded)- Assumption : always within prefix (ie. not within extDict) */-FORCE_INLINE-U32 ZSTD_insertAndFindFirstIndex (ZSTD_CCtx* zc, const BYTE* ip, U32 mls)-{- U32* const hashTable = zc->hashTable;- const U32 hashLog = zc->params.cParams.hashLog;- U32* const chainTable = zc->chainTable;- const U32 chainMask = (1 << zc->params.cParams.chainLog) - 1;- const BYTE* const base = zc->base;- const U32 target = (U32)(ip - base);- U32 idx = zc->nextToUpdate;-- while(idx < target) { /* catch up */- size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);- NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];- hashTable[h] = idx;- idx++;- }-- zc->nextToUpdate = target;- return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];-}----FORCE_INLINE /* inlining is important to hardwire a hot branch (template emulation) */-size_t ZSTD_HcFindBestMatch_generic (- ZSTD_CCtx* zc, /* Index table will be updated */- const BYTE* const ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 mls, const U32 extDict)-{- U32* const chainTable = zc->chainTable;- const U32 chainSize = (1 << zc->params.cParams.chainLog);- const U32 chainMask = chainSize-1;- const BYTE* const base = zc->base;- const BYTE* const dictBase = zc->dictBase;- const U32 dictLimit = zc->dictLimit;- const BYTE* const prefixStart = base + dictLimit;- const BYTE* const dictEnd = dictBase + dictLimit;- const U32 lowLimit = zc->lowLimit;- const U32 current = (U32)(ip-base);- const U32 minChain = current > chainSize ? current - chainSize : 0;- int nbAttempts=maxNbAttempts;- size_t ml=EQUAL_READ32-1;-- /* HC4 match finder */- U32 matchIndex = ZSTD_insertAndFindFirstIndex (zc, ip, mls);-- for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) {- const BYTE* match;- size_t currentMl=0;- if ((!extDict) || matchIndex >= dictLimit) {- match = base + matchIndex;- if (match[ml] == ip[ml]) /* potentially better */- currentMl = ZSTD_count(ip, match, iLimit);- } else {- match = dictBase + matchIndex;- if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */- currentMl = ZSTD_count_2segments(ip+EQUAL_READ32, match+EQUAL_READ32, iLimit, dictEnd, prefixStart) + EQUAL_READ32;- }-- /* save best solution */- if (currentMl > ml) { ml = currentMl; *offsetPtr = current - matchIndex + ZSTD_REP_MOVE; if (ip+currentMl == iLimit) break; /* best possible, and avoid read overflow*/ }-- if (matchIndex <= minChain) break;- matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);- }-- return ml;-}---FORCE_INLINE size_t ZSTD_HcFindBestMatch_selectMLS (- ZSTD_CCtx* zc,- const BYTE* ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 matchLengthSearch)-{- switch(matchLengthSearch)- {- default :- case 4 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 0);- case 5 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 0);- case 6 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 0);- }-}---FORCE_INLINE size_t ZSTD_HcFindBestMatch_extDict_selectMLS (- ZSTD_CCtx* zc,- const BYTE* ip, const BYTE* const iLimit,- size_t* offsetPtr,- const U32 maxNbAttempts, const U32 matchLengthSearch)-{- switch(matchLengthSearch)- {- default :- case 4 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 1);- case 5 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 1);- case 6 : return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 1);- }-}---/* *******************************-* Common parser - lazy strategy-*********************************/-FORCE_INLINE-void ZSTD_compressBlock_lazy_generic(ZSTD_CCtx* ctx,- const void* src, size_t srcSize,- const U32 searchMethod, const U32 depth)-{- seqStore_t* seqStorePtr = &(ctx->seqStore);- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - 8;- const BYTE* const base = ctx->base + ctx->dictLimit;-- U32 const maxSearches = 1 << ctx->params.cParams.searchLog;- U32 const mls = ctx->params.cParams.searchLength;-- typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,- size_t* offsetPtr,- U32 maxNbAttempts, U32 matchLengthSearch);- searchMax_f const searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS;- U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1], savedOffset=0;-- /* init */- ip += (ip==base);- ctx->nextToUpdate3 = ctx->nextToUpdate;- { U32 const maxRep = (U32)(ip-base);- if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;- if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0;- }-- /* Match Loop */- while (ip < ilimit) {- size_t matchLength=0;- size_t offset=0;- const BYTE* start=ip+1;-- /* check repCode */- if ((offset_1>0) & (MEM_read32(ip+1) == MEM_read32(ip+1 - offset_1))) {- /* repcode : we take it */- matchLength = ZSTD_count(ip+1+EQUAL_READ32, ip+1+EQUAL_READ32-offset_1, iend) + EQUAL_READ32;- if (depth==0) goto _storeSequence;- }-- /* first search (depth 0) */- { size_t offsetFound = 99999999;- size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);- if (ml2 > matchLength)- matchLength = ml2, start = ip, offset=offsetFound;- }-- if (matchLength < EQUAL_READ32) {- ip += ((ip-anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */- continue;- }-- /* let's try to find a better solution */- if (depth>=1)- while (ip<ilimit) {- ip ++;- if ((offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {- size_t const mlRep = ZSTD_count(ip+EQUAL_READ32, ip+EQUAL_READ32-offset_1, iend) + EQUAL_READ32;- int const gain2 = (int)(mlRep * 3);- int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);- if ((mlRep >= EQUAL_READ32) && (gain2 > gain1))- matchLength = mlRep, offset = 0, start = ip;- }- { size_t offset2=99999999;- size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);- int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */- int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);- if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {- matchLength = ml2, offset = offset2, start = ip;- continue; /* search a better one */- } }-- /* let's find an even better one */- if ((depth==2) && (ip<ilimit)) {- ip ++;- if ((offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {- size_t const ml2 = ZSTD_count(ip+EQUAL_READ32, ip+EQUAL_READ32-offset_1, iend) + EQUAL_READ32;- int const gain2 = (int)(ml2 * 4);- int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);- if ((ml2 >= EQUAL_READ32) && (gain2 > gain1))- matchLength = ml2, offset = 0, start = ip;- }- { size_t offset2=99999999;- size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);- int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */- int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);- if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {- matchLength = ml2, offset = offset2, start = ip;- continue;- } } }- break; /* nothing found : store previous solution */- }-- /* catch up */- if (offset) {- while ((start>anchor) && (start>base+offset-ZSTD_REP_MOVE) && (start[-1] == start[-1-offset+ZSTD_REP_MOVE])) /* only search for offset within prefix */- { start--; matchLength++; }- offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);- }-- /* store sequence */-_storeSequence:- { size_t const litLength = start - anchor;- ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength-MINMATCH);- anchor = ip = start + matchLength;- }-- /* check immediate repcode */- while ( (ip <= ilimit)- && ((offset_2>0)- & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {- /* store sequence */- matchLength = ZSTD_count(ip+EQUAL_READ32, ip+EQUAL_READ32-offset_2, iend) + EQUAL_READ32;- offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap repcodes */- ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength-MINMATCH);- ip += matchLength;- anchor = ip;- continue; /* faster when present ... (?) */- } }-- /* Save reps for next block */- ctx->savedRep[0] = offset_1 ? offset_1 : savedOffset;- ctx->savedRep[1] = offset_2 ? offset_2 : savedOffset;-- /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}---static void ZSTD_compressBlock_btlazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 1, 2);-}--static void ZSTD_compressBlock_lazy2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 2);-}--static void ZSTD_compressBlock_lazy(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 1);-}--static void ZSTD_compressBlock_greedy(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 0);-}---FORCE_INLINE-void ZSTD_compressBlock_lazy_extDict_generic(ZSTD_CCtx* ctx,- const void* src, size_t srcSize,- const U32 searchMethod, const U32 depth)-{- seqStore_t* seqStorePtr = &(ctx->seqStore);- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - 8;- const BYTE* const base = ctx->base;- const U32 dictLimit = ctx->dictLimit;- const U32 lowestIndex = ctx->lowLimit;- const BYTE* const prefixStart = base + dictLimit;- const BYTE* const dictBase = ctx->dictBase;- const BYTE* const dictEnd = dictBase + dictLimit;- const BYTE* const dictStart = dictBase + ctx->lowLimit;-- const U32 maxSearches = 1 << ctx->params.cParams.searchLog;- const U32 mls = ctx->params.cParams.searchLength;-- typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,- size_t* offsetPtr,- U32 maxNbAttempts, U32 matchLengthSearch);- searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS_extDict : ZSTD_HcFindBestMatch_extDict_selectMLS;-- U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];-- /* init */- ctx->nextToUpdate3 = ctx->nextToUpdate;- ip += (ip == prefixStart);-- /* Match Loop */- while (ip < ilimit) {- size_t matchLength=0;- size_t offset=0;- const BYTE* start=ip+1;- U32 current = (U32)(ip-base);-- /* check repCode */- { const U32 repIndex = (U32)(current+1 - offset_1);- const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* const repMatch = repBase + repIndex;- if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */- if (MEM_read32(ip+1) == MEM_read32(repMatch)) {- /* repcode detected we should take it */- const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;- matchLength = ZSTD_count_2segments(ip+1+EQUAL_READ32, repMatch+EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;- if (depth==0) goto _storeSequence;- } }-- /* first search (depth 0) */- { size_t offsetFound = 99999999;- size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);- if (ml2 > matchLength)- matchLength = ml2, start = ip, offset=offsetFound;- }-- if (matchLength < EQUAL_READ32) {- ip += ((ip-anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */- continue;- }-- /* let's try to find a better solution */- if (depth>=1)- while (ip<ilimit) {- ip ++;- current++;- /* check repCode */- if (offset) {- const U32 repIndex = (U32)(current - offset_1);- const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* const repMatch = repBase + repIndex;- if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */- if (MEM_read32(ip) == MEM_read32(repMatch)) {- /* repcode detected */- const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;- size_t const repLength = ZSTD_count_2segments(ip+EQUAL_READ32, repMatch+EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;- int const gain2 = (int)(repLength * 3);- int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);- if ((repLength >= EQUAL_READ32) && (gain2 > gain1))- matchLength = repLength, offset = 0, start = ip;- } }-- /* search match, depth 1 */- { size_t offset2=99999999;- size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);- int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */- int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);- if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {- matchLength = ml2, offset = offset2, start = ip;- continue; /* search a better one */- } }-- /* let's find an even better one */- if ((depth==2) && (ip<ilimit)) {- ip ++;- current++;- /* check repCode */- if (offset) {- const U32 repIndex = (U32)(current - offset_1);- const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* const repMatch = repBase + repIndex;- if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */- if (MEM_read32(ip) == MEM_read32(repMatch)) {- /* repcode detected */- const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;- size_t repLength = ZSTD_count_2segments(ip+EQUAL_READ32, repMatch+EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;- int gain2 = (int)(repLength * 4);- int gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);- if ((repLength >= EQUAL_READ32) && (gain2 > gain1))- matchLength = repLength, offset = 0, start = ip;- } }-- /* search match, depth 2 */- { size_t offset2=99999999;- size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);- int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */- int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);- if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {- matchLength = ml2, offset = offset2, start = ip;- continue;- } } }- break; /* nothing found : store previous solution */- }-- /* catch up */- if (offset) {- U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));- const BYTE* match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;- const BYTE* const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;- while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; } /* catch up */- offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);- }-- /* store sequence */-_storeSequence:- { size_t const litLength = start - anchor;- ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength-MINMATCH);- anchor = ip = start + matchLength;- }-- /* check immediate repcode */- while (ip <= ilimit) {- const U32 repIndex = (U32)((ip-base) - offset_2);- const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* const repMatch = repBase + repIndex;- if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */- if (MEM_read32(ip) == MEM_read32(repMatch)) {- /* repcode detected we should take it */- const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;- matchLength = ZSTD_count_2segments(ip+EQUAL_READ32, repMatch+EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;- offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap offset history */- ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength-MINMATCH);- ip += matchLength;- anchor = ip;- continue; /* faster when present ... (?) */- }- break;- } }-- /* Save reps for next block */- ctx->savedRep[0] = offset_1; ctx->savedRep[1] = offset_2;-- /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}---void ZSTD_compressBlock_greedy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 0);-}--static void ZSTD_compressBlock_lazy_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 1);-}--static void ZSTD_compressBlock_lazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 2);-}--static void ZSTD_compressBlock_btlazy2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{- ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 1, 2);-}---/* The optimal parser */-#include "zstd_opt.h"--static void ZSTD_compressBlock_btopt(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{-#ifdef ZSTD_OPT_H_91842398743- ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 0);-#else- (void)ctx; (void)src; (void)srcSize;- return;-#endif-}--static void ZSTD_compressBlock_btopt2(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{-#ifdef ZSTD_OPT_H_91842398743- ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 1);-#else- (void)ctx; (void)src; (void)srcSize;- return;-#endif-}--static void ZSTD_compressBlock_btopt_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{-#ifdef ZSTD_OPT_H_91842398743- ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 0);-#else- (void)ctx; (void)src; (void)srcSize;- return;-#endif-}--static void ZSTD_compressBlock_btopt2_extDict(ZSTD_CCtx* ctx, const void* src, size_t srcSize)-{-#ifdef ZSTD_OPT_H_91842398743- ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 1);-#else- (void)ctx; (void)src; (void)srcSize;- return;-#endif-}---typedef void (*ZSTD_blockCompressor) (ZSTD_CCtx* ctx, const void* src, size_t srcSize);--static ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict)-{- static const ZSTD_blockCompressor blockCompressor[2][8] = {- { ZSTD_compressBlock_fast, ZSTD_compressBlock_doubleFast, ZSTD_compressBlock_greedy, ZSTD_compressBlock_lazy, ZSTD_compressBlock_lazy2, ZSTD_compressBlock_btlazy2, ZSTD_compressBlock_btopt, ZSTD_compressBlock_btopt2 },- { ZSTD_compressBlock_fast_extDict, ZSTD_compressBlock_doubleFast_extDict, ZSTD_compressBlock_greedy_extDict, ZSTD_compressBlock_lazy_extDict,ZSTD_compressBlock_lazy2_extDict, ZSTD_compressBlock_btlazy2_extDict, ZSTD_compressBlock_btopt_extDict, ZSTD_compressBlock_btopt2_extDict }- };-- return blockCompressor[extDict][(U32)strat];-}---static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc, void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->params.cParams.strategy, zc->lowLimit < zc->dictLimit);- const BYTE* const base = zc->base;- const BYTE* const istart = (const BYTE*)src;- const U32 current = (U32)(istart-base);- if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) return 0; /* don't even attempt compression below a certain srcSize */- ZSTD_resetSeqStore(&(zc->seqStore));- if (current > zc->nextToUpdate + 384)- zc->nextToUpdate = current - MIN(192, (U32)(current - zc->nextToUpdate - 384)); /* update tree not updated after finding very long rep matches */- blockCompressor(zc, src, srcSize);- return ZSTD_compressSequences(zc, dst, dstCapacity, srcSize);-}---/*! ZSTD_compress_generic() :-* Compress a chunk of data into one or multiple blocks.-* All blocks will be terminated, all input will be consumed.-* Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.-* Frame is supposed already started (header already produced)-* @return : compressed size, or an error code-*/-static size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- U32 lastFrameChunk)-{- size_t blockSize = cctx->blockSize;- size_t remaining = srcSize;- const BYTE* ip = (const BYTE*)src;- BYTE* const ostart = (BYTE*)dst;- BYTE* op = ostart;- U32 const maxDist = 1 << cctx->params.cParams.windowLog;-- if (cctx->params.fParams.checksumFlag && srcSize)- XXH64_update(&cctx->xxhState, src, srcSize);-- while (remaining) {- U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);- size_t cSize;-- if (dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE) return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */- if (remaining < blockSize) blockSize = remaining;-- /* preemptive overflow correction */- if (cctx->lowLimit > (2U<<30)) {- U32 const cycleMask = (1 << ZSTD_cycleLog(cctx->params.cParams.hashLog, cctx->params.cParams.strategy)) - 1;- U32 const current = (U32)(ip - cctx->base);- U32 const newCurrent = (current & cycleMask) + (1 << cctx->params.cParams.windowLog);- U32 const correction = current - newCurrent;- ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_64 <= 30);- ZSTD_reduceIndex(cctx, correction);- cctx->base += correction;- cctx->dictBase += correction;- cctx->lowLimit -= correction;- cctx->dictLimit -= correction;- if (cctx->nextToUpdate < correction) cctx->nextToUpdate = 0;- else cctx->nextToUpdate -= correction;- }-- if ((U32)(ip+blockSize - cctx->base) > cctx->loadedDictEnd + maxDist) {- /* enforce maxDist */- U32 const newLowLimit = (U32)(ip+blockSize - cctx->base) - maxDist;- if (cctx->lowLimit < newLowLimit) cctx->lowLimit = newLowLimit;- if (cctx->dictLimit < cctx->lowLimit) cctx->dictLimit = cctx->lowLimit;- }-- cSize = ZSTD_compressBlock_internal(cctx, op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize, ip, blockSize);- if (ZSTD_isError(cSize)) return cSize;-- if (cSize == 0) { /* block is not compressible */- U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(blockSize << 3);- if (blockSize + ZSTD_blockHeaderSize > dstCapacity) return ERROR(dstSize_tooSmall);- MEM_writeLE32(op, cBlockHeader24); /* no pb, 4th byte will be overwritten */- memcpy(op + ZSTD_blockHeaderSize, ip, blockSize);- cSize = ZSTD_blockHeaderSize+blockSize;- } else {- U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);- MEM_writeLE24(op, cBlockHeader24);- cSize += ZSTD_blockHeaderSize;- }-- remaining -= blockSize;- dstCapacity -= cSize;- ip += blockSize;- op += cSize;- }-- if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending;- return op-ostart;-}---static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity,- ZSTD_parameters params, U64 pledgedSrcSize, U32 dictID)-{ BYTE* const op = (BYTE*)dst;- U32 const dictIDSizeCode = (dictID>0) + (dictID>=256) + (dictID>=65536); /* 0-3 */- U32 const checksumFlag = params.fParams.checksumFlag>0;- U32 const windowSize = 1U << params.cParams.windowLog;- U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize > (pledgedSrcSize-1));- BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);- U32 const fcsCode = params.fParams.contentSizeFlag ?- (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : /* 0-3 */- 0;- BYTE const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );- size_t pos;-- if (dstCapacity < ZSTD_frameHeaderSize_max) return ERROR(dstSize_tooSmall);-- MEM_writeLE32(dst, ZSTD_MAGICNUMBER);- op[4] = frameHeaderDecriptionByte; pos=5;- if (!singleSegment) op[pos++] = windowLogByte;- switch(dictIDSizeCode)- {- default: /* impossible */- case 0 : break;- case 1 : op[pos] = (BYTE)(dictID); pos++; break;- case 2 : MEM_writeLE16(op+pos, (U16)dictID); pos+=2; break;- case 3 : MEM_writeLE32(op+pos, dictID); pos+=4; break;- }- switch(fcsCode)- {- default: /* impossible */- case 0 : if (singleSegment) op[pos++] = (BYTE)(pledgedSrcSize); break;- case 1 : MEM_writeLE16(op+pos, (U16)(pledgedSrcSize-256)); pos+=2; break;- case 2 : MEM_writeLE32(op+pos, (U32)(pledgedSrcSize)); pos+=4; break;- case 3 : MEM_writeLE64(op+pos, (U64)(pledgedSrcSize)); pos+=8; break;- }- return pos;-}---static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- U32 frame, U32 lastFrameChunk)-{- const BYTE* const ip = (const BYTE*) src;- size_t fhSize = 0;-- if (cctx->stage==ZSTDcs_created) return ERROR(stage_wrong); /* missing init (ZSTD_compressBegin) */-- if (frame && (cctx->stage==ZSTDcs_init)) {- fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->params, cctx->frameContentSize, cctx->dictID);- if (ZSTD_isError(fhSize)) return fhSize;- dstCapacity -= fhSize;- dst = (char*)dst + fhSize;- cctx->stage = ZSTDcs_ongoing;- }-- /* Check if blocks follow each other */- if (src != cctx->nextSrc) {- /* not contiguous */- ptrdiff_t const delta = cctx->nextSrc - ip;- cctx->lowLimit = cctx->dictLimit;- cctx->dictLimit = (U32)(cctx->nextSrc - cctx->base);- cctx->dictBase = cctx->base;- cctx->base -= delta;- cctx->nextToUpdate = cctx->dictLimit;- if (cctx->dictLimit - cctx->lowLimit < HASH_READ_SIZE) cctx->lowLimit = cctx->dictLimit; /* too small extDict */- }-- /* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */- if ((ip+srcSize > cctx->dictBase + cctx->lowLimit) & (ip < cctx->dictBase + cctx->dictLimit)) {- ptrdiff_t const highInputIdx = (ip + srcSize) - cctx->dictBase;- U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)cctx->dictLimit) ? cctx->dictLimit : (U32)highInputIdx;- cctx->lowLimit = lowLimitMax;- }-- cctx->nextSrc = ip + srcSize;-- { size_t const cSize = frame ?- ZSTD_compress_generic (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) :- ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize);- if (ZSTD_isError(cSize)) return cSize;- return cSize + fhSize;- }-}---size_t ZSTD_compressContinue (ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize)-{- return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1, 0);-}---size_t ZSTD_getBlockSizeMax(ZSTD_CCtx* cctx)-{- return MIN (ZSTD_BLOCKSIZE_ABSOLUTEMAX, 1 << cctx->params.cParams.windowLog);-}--size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- size_t const blockSizeMax = ZSTD_getBlockSizeMax(cctx);- if (srcSize > blockSizeMax) return ERROR(srcSize_wrong);- return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0, 0);-}---static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t srcSize)-{- const BYTE* const ip = (const BYTE*) src;- const BYTE* const iend = ip + srcSize;-- /* input becomes current prefix */- zc->lowLimit = zc->dictLimit;- zc->dictLimit = (U32)(zc->nextSrc - zc->base);- zc->dictBase = zc->base;- zc->base += ip - zc->nextSrc;- zc->nextToUpdate = zc->dictLimit;- zc->loadedDictEnd = (U32)(iend - zc->base);-- zc->nextSrc = iend;- if (srcSize <= HASH_READ_SIZE) return 0;-- switch(zc->params.cParams.strategy)- {- case ZSTD_fast:- ZSTD_fillHashTable (zc, iend, zc->params.cParams.searchLength);- break;-- case ZSTD_dfast:- ZSTD_fillDoubleHashTable (zc, iend, zc->params.cParams.searchLength);- break;-- case ZSTD_greedy:- case ZSTD_lazy:- case ZSTD_lazy2:- ZSTD_insertAndFindFirstIndex (zc, iend-HASH_READ_SIZE, zc->params.cParams.searchLength);- break;-- case ZSTD_btlazy2:- case ZSTD_btopt:- case ZSTD_btopt2:- ZSTD_updateTree(zc, iend-HASH_READ_SIZE, iend, 1 << zc->params.cParams.searchLog, zc->params.cParams.searchLength);- break;-- default:- return ERROR(GENERIC); /* strategy doesn't exist; impossible */- }-- zc->nextToUpdate = zc->loadedDictEnd;- return 0;-}---/* Dictionaries that assign zero probability to symbols that show up causes problems- when FSE encoding. Refuse dictionaries that assign zero probability to symbols- that we may encounter during compression.- NOTE: This behavior is not standard and could be improved in the future. */-static size_t ZSTD_checkDictNCount(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue) {- U32 s;- if (dictMaxSymbolValue < maxSymbolValue) return ERROR(dictionary_corrupted);- for (s = 0; s <= maxSymbolValue; ++s) {- if (normalizedCounter[s] == 0) return ERROR(dictionary_corrupted);- }- return 0;-}---/* Dictionary format :- Magic == ZSTD_DICT_MAGIC (4 bytes)- HUF_writeCTable(256)- FSE_writeNCount(off)- FSE_writeNCount(ml)- FSE_writeNCount(ll)- RepOffsets- Dictionary content-*/-/*! ZSTD_loadDictEntropyStats() :- @return : size read from dictionary- note : magic number supposed already checked */-static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* cctx, const void* dict, size_t dictSize)-{- const BYTE* dictPtr = (const BYTE*)dict;- const BYTE* const dictEnd = dictPtr + dictSize;- short offcodeNCount[MaxOff+1];- unsigned offcodeMaxValue = MaxOff;- BYTE scratchBuffer[1<<MAX(MLFSELog,LLFSELog)];-- { size_t const hufHeaderSize = HUF_readCTable(cctx->hufTable, 255, dict, dictSize);- if (HUF_isError(hufHeaderSize)) return ERROR(dictionary_corrupted);- dictPtr += hufHeaderSize;- }-- { unsigned offcodeLog;- size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);- if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted);- if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted);- /* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */- CHECK_E (FSE_buildCTable_wksp(cctx->offcodeCTable, offcodeNCount, offcodeMaxValue, offcodeLog, scratchBuffer, sizeof(scratchBuffer)), dictionary_corrupted);- dictPtr += offcodeHeaderSize;- }-- { short matchlengthNCount[MaxML+1];- unsigned matchlengthMaxValue = MaxML, matchlengthLog;- size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);- if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted);- if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted);- /* Every match length code must have non-zero probability */- CHECK_F (ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));- CHECK_E (FSE_buildCTable_wksp(cctx->matchlengthCTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, scratchBuffer, sizeof(scratchBuffer)), dictionary_corrupted);- dictPtr += matchlengthHeaderSize;- }-- { short litlengthNCount[MaxLL+1];- unsigned litlengthMaxValue = MaxLL, litlengthLog;- size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);- if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted);- if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted);- /* Every literal length code must have non-zero probability */- CHECK_F (ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));- CHECK_E(FSE_buildCTable_wksp(cctx->litlengthCTable, litlengthNCount, litlengthMaxValue, litlengthLog, scratchBuffer, sizeof(scratchBuffer)), dictionary_corrupted);- dictPtr += litlengthHeaderSize;- }-- if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted);- cctx->rep[0] = MEM_readLE32(dictPtr+0); if (cctx->rep[0] == 0 || cctx->rep[0] >= dictSize) return ERROR(dictionary_corrupted);- cctx->rep[1] = MEM_readLE32(dictPtr+4); if (cctx->rep[1] == 0 || cctx->rep[1] >= dictSize) return ERROR(dictionary_corrupted);- cctx->rep[2] = MEM_readLE32(dictPtr+8); if (cctx->rep[2] == 0 || cctx->rep[2] >= dictSize) return ERROR(dictionary_corrupted);- dictPtr += 12;-- { U32 offcodeMax = MaxOff;- if ((size_t)(dictEnd - dictPtr) <= ((U32)-1) - 128 KB) {- U32 const maxOffset = (U32)(dictEnd - dictPtr) + 128 KB; /* The maximum offset that must be supported */- /* Calculate minimum offset code required to represent maxOffset */- offcodeMax = ZSTD_highbit32(maxOffset);- }- /* Every possible supported offset <= dictContentSize + 128 KB must be representable */- CHECK_F (ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));- }-- cctx->flagStaticTables = 1;- return dictPtr - (const BYTE*)dict;-}--/** ZSTD_compress_insertDictionary() :-* @return : 0, or an error code */-static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx* zc, const void* dict, size_t dictSize)-{- if ((dict==NULL) || (dictSize<=8)) return 0;-- /* default : dict is pure content */- if (MEM_readLE32(dict) != ZSTD_DICT_MAGIC) return ZSTD_loadDictionaryContent(zc, dict, dictSize);- zc->dictID = zc->params.fParams.noDictIDFlag ? 0 : MEM_readLE32((const char*)dict+4);-- /* known magic number : dict is parsed for entropy stats and content */- { size_t const loadError = ZSTD_loadDictEntropyStats(zc, (const char*)dict+8 /* skip dictHeader */, dictSize-8);- size_t const eSize = loadError + 8;- if (ZSTD_isError(loadError)) return loadError;- return ZSTD_loadDictionaryContent(zc, (const char*)dict+eSize, dictSize-eSize);- }-}---/*! ZSTD_compressBegin_internal() :-* @return : 0, or an error code */-static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx,- const void* dict, size_t dictSize,- ZSTD_parameters params, U64 pledgedSrcSize)-{- ZSTD_compResetPolicy_e const crp = dictSize ? ZSTDcrp_fullReset : ZSTDcrp_continue;- CHECK_F(ZSTD_resetCCtx_advanced(cctx, params, pledgedSrcSize, crp));- return ZSTD_compress_insertDictionary(cctx, dict, dictSize);-}---/*! ZSTD_compressBegin_advanced() :-* @return : 0, or an error code */-size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx,- const void* dict, size_t dictSize,- ZSTD_parameters params, unsigned long long pledgedSrcSize)-{- /* compression parameters verification and optimization */- CHECK_F(ZSTD_checkCParams(params.cParams));- return ZSTD_compressBegin_internal(cctx, dict, dictSize, params, pledgedSrcSize);-}---size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)-{- ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);- return ZSTD_compressBegin_internal(cctx, dict, dictSize, params, 0);-}---size_t ZSTD_compressBegin(ZSTD_CCtx* zc, int compressionLevel)-{- return ZSTD_compressBegin_usingDict(zc, NULL, 0, compressionLevel);-}---/*! ZSTD_writeEpilogue() :-* Ends a frame.-* @return : nb of bytes written into dst (or an error code) */-static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity)-{- BYTE* const ostart = (BYTE*)dst;- BYTE* op = ostart;- size_t fhSize = 0;-- if (cctx->stage == ZSTDcs_created) return ERROR(stage_wrong); /* init missing */-- /* special case : empty frame */- if (cctx->stage == ZSTDcs_init) {- fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->params, 0, 0);- if (ZSTD_isError(fhSize)) return fhSize;- dstCapacity -= fhSize;- op += fhSize;- cctx->stage = ZSTDcs_ongoing;- }-- if (cctx->stage != ZSTDcs_ending) {- /* write one last empty block, make it the "last" block */- U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0;- if (dstCapacity<4) return ERROR(dstSize_tooSmall);- MEM_writeLE32(op, cBlockHeader24);- op += ZSTD_blockHeaderSize;- dstCapacity -= ZSTD_blockHeaderSize;- }-- if (cctx->params.fParams.checksumFlag) {- U32 const checksum = (U32) XXH64_digest(&cctx->xxhState);- if (dstCapacity<4) return ERROR(dstSize_tooSmall);- MEM_writeLE32(op, checksum);- op += 4;- }-- cctx->stage = ZSTDcs_created; /* return to "created but no init" status */- return op-ostart;-}---size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize)-{- size_t endResult;- size_t const cSize = ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1, 1);- if (ZSTD_isError(cSize)) return cSize;- endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize);- if (ZSTD_isError(endResult)) return endResult;- return cSize + endResult;-}---static size_t ZSTD_compress_internal (ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const void* dict,size_t dictSize,- ZSTD_parameters params)-{- CHECK_F(ZSTD_compressBegin_internal(cctx, dict, dictSize, params, srcSize));- return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);-}--size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const void* dict,size_t dictSize,- ZSTD_parameters params)-{- CHECK_F(ZSTD_checkCParams(params.cParams));- return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params);-}--size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, const void* dict, size_t dictSize, int compressionLevel)-{- ZSTD_parameters params = ZSTD_getParams(compressionLevel, srcSize, dict ? dictSize : 0);- params.fParams.contentSizeFlag = 1;- return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params);-}--size_t ZSTD_compressCCtx (ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, int compressionLevel)-{- return ZSTD_compress_usingDict(ctx, dst, dstCapacity, src, srcSize, NULL, 0, compressionLevel);-}--size_t ZSTD_compress(void* dst, size_t dstCapacity, const void* src, size_t srcSize, int compressionLevel)-{- size_t result;- ZSTD_CCtx ctxBody;- memset(&ctxBody, 0, sizeof(ctxBody));- memcpy(&ctxBody.customMem, &defaultCustomMem, sizeof(ZSTD_customMem));- result = ZSTD_compressCCtx(&ctxBody, dst, dstCapacity, src, srcSize, compressionLevel);- ZSTD_free(ctxBody.workSpace, defaultCustomMem); /* can't free ctxBody itself, as it's on stack; free only heap content */- return result;-}---/* ===== Dictionary API ===== */--struct ZSTD_CDict_s {- void* dictBuffer;- const void* dictContent;- size_t dictContentSize;- ZSTD_CCtx* refContext;-}; /* typedef'd tp ZSTD_CDict within "zstd.h" */--size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict)-{- if (cdict==NULL) return 0; /* support sizeof on NULL */- return ZSTD_sizeof_CCtx(cdict->refContext) + (cdict->dictBuffer ? cdict->dictContentSize : 0) + sizeof(*cdict);-}--ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize, unsigned byReference,- ZSTD_parameters params, ZSTD_customMem customMem)-{- if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;- if (!customMem.customAlloc || !customMem.customFree) return NULL;-- { ZSTD_CDict* const cdict = (ZSTD_CDict*) ZSTD_malloc(sizeof(ZSTD_CDict), customMem);- ZSTD_CCtx* const cctx = ZSTD_createCCtx_advanced(customMem);-- if (!cdict || !cctx) {- ZSTD_free(cdict, customMem);- ZSTD_free(cctx, customMem);- return NULL;- }-- if ((byReference) || (!dictBuffer) || (!dictSize)) {- cdict->dictBuffer = NULL;- cdict->dictContent = dictBuffer;- } else {- void* const internalBuffer = ZSTD_malloc(dictSize, customMem);- if (!internalBuffer) { ZSTD_free(cctx, customMem); ZSTD_free(cdict, customMem); return NULL; }- memcpy(internalBuffer, dictBuffer, dictSize);- cdict->dictBuffer = internalBuffer;- cdict->dictContent = internalBuffer;- }-- { size_t const errorCode = ZSTD_compressBegin_advanced(cctx, cdict->dictContent, dictSize, params, 0);- if (ZSTD_isError(errorCode)) {- ZSTD_free(cdict->dictBuffer, customMem);- ZSTD_free(cctx, customMem);- ZSTD_free(cdict, customMem);- return NULL;- } }-- cdict->refContext = cctx;- cdict->dictContentSize = dictSize;- return cdict;- }-}--ZSTD_CDict* ZSTD_createCDict(const void* dict, size_t dictSize, int compressionLevel)-{- ZSTD_customMem const allocator = { NULL, NULL, NULL };- ZSTD_parameters params = ZSTD_getParams(compressionLevel, 0, dictSize);- params.fParams.contentSizeFlag = 1;- return ZSTD_createCDict_advanced(dict, dictSize, 0, params, allocator);-}--ZSTD_CDict* ZSTD_createCDict_byReference(const void* dict, size_t dictSize, int compressionLevel)-{- ZSTD_customMem const allocator = { NULL, NULL, NULL };- ZSTD_parameters params = ZSTD_getParams(compressionLevel, 0, dictSize);- params.fParams.contentSizeFlag = 1;- return ZSTD_createCDict_advanced(dict, dictSize, 1, params, allocator);-}--size_t ZSTD_freeCDict(ZSTD_CDict* cdict)-{- if (cdict==NULL) return 0; /* support free on NULL */- { ZSTD_customMem const cMem = cdict->refContext->customMem;- ZSTD_freeCCtx(cdict->refContext);- ZSTD_free(cdict->dictBuffer, cMem);- ZSTD_free(cdict, cMem);- return 0;- }-}--static ZSTD_parameters ZSTD_getParamsFromCDict(const ZSTD_CDict* cdict) {- return ZSTD_getParamsFromCCtx(cdict->refContext);-}--size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict, U64 pledgedSrcSize)-{- if (cdict->dictContentSize) CHECK_F(ZSTD_copyCCtx(cctx, cdict->refContext, pledgedSrcSize))- else CHECK_F(ZSTD_compressBegin_advanced(cctx, NULL, 0, cdict->refContext->params, pledgedSrcSize));- return 0;-}--/*! ZSTD_compress_usingCDict() :-* Compression using a digested Dictionary.-* Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.-* Note that compression level is decided during dictionary creation */-size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const ZSTD_CDict* cdict)-{- CHECK_F(ZSTD_compressBegin_usingCDict(cctx, cdict, srcSize));-- if (cdict->refContext->params.fParams.contentSizeFlag==1) {- cctx->params.fParams.contentSizeFlag = 1;- cctx->frameContentSize = srcSize;- }-- return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);-}----/* ******************************************************************-* Streaming-********************************************************************/--typedef enum { zcss_init, zcss_load, zcss_flush, zcss_final } ZSTD_cStreamStage;--struct ZSTD_CStream_s {- ZSTD_CCtx* cctx;- ZSTD_CDict* cdictLocal;- const ZSTD_CDict* cdict;- char* inBuff;- size_t inBuffSize;- size_t inToCompress;- size_t inBuffPos;- size_t inBuffTarget;- size_t blockSize;- char* outBuff;- size_t outBuffSize;- size_t outBuffContentSize;- size_t outBuffFlushedSize;- ZSTD_cStreamStage stage;- U32 checksum;- U32 frameEnded;- U64 pledgedSrcSize;- U64 inputProcessed;- ZSTD_parameters params;- ZSTD_customMem customMem;-}; /* typedef'd to ZSTD_CStream within "zstd.h" */--ZSTD_CStream* ZSTD_createCStream(void)-{- return ZSTD_createCStream_advanced(defaultCustomMem);-}--ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem)-{- ZSTD_CStream* zcs;-- if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;- if (!customMem.customAlloc || !customMem.customFree) return NULL;-- zcs = (ZSTD_CStream*)ZSTD_malloc(sizeof(ZSTD_CStream), customMem);- if (zcs==NULL) return NULL;- memset(zcs, 0, sizeof(ZSTD_CStream));- memcpy(&zcs->customMem, &customMem, sizeof(ZSTD_customMem));- zcs->cctx = ZSTD_createCCtx_advanced(customMem);- if (zcs->cctx == NULL) { ZSTD_freeCStream(zcs); return NULL; }- return zcs;-}--size_t ZSTD_freeCStream(ZSTD_CStream* zcs)-{- if (zcs==NULL) return 0; /* support free on NULL */- { ZSTD_customMem const cMem = zcs->customMem;- ZSTD_freeCCtx(zcs->cctx);- ZSTD_freeCDict(zcs->cdictLocal);- ZSTD_free(zcs->inBuff, cMem);- ZSTD_free(zcs->outBuff, cMem);- ZSTD_free(zcs, cMem);- return 0;- }-}---/*====== Initialization ======*/--size_t ZSTD_CStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }-size_t ZSTD_CStreamOutSize(void) { return ZSTD_compressBound(ZSTD_BLOCKSIZE_ABSOLUTEMAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */ ; }--size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)-{- if (zcs->inBuffSize==0) return ERROR(stage_wrong); /* zcs has not been init at least once => can't reset */-- if (zcs->cdict) CHECK_F(ZSTD_compressBegin_usingCDict(zcs->cctx, zcs->cdict, pledgedSrcSize))- else CHECK_F(ZSTD_compressBegin_advanced(zcs->cctx, NULL, 0, zcs->params, pledgedSrcSize));-- zcs->inToCompress = 0;- zcs->inBuffPos = 0;- zcs->inBuffTarget = zcs->blockSize;- zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;- zcs->stage = zcss_load;- zcs->frameEnded = 0;- zcs->pledgedSrcSize = pledgedSrcSize;- zcs->inputProcessed = 0;- return 0; /* ready to go */-}--size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,- const void* dict, size_t dictSize,- ZSTD_parameters params, unsigned long long pledgedSrcSize)-{- /* allocate buffers */- { size_t const neededInBuffSize = (size_t)1 << params.cParams.windowLog;- if (zcs->inBuffSize < neededInBuffSize) {- zcs->inBuffSize = neededInBuffSize;- ZSTD_free(zcs->inBuff, zcs->customMem);- zcs->inBuff = (char*) ZSTD_malloc(neededInBuffSize, zcs->customMem);- if (zcs->inBuff == NULL) return ERROR(memory_allocation);- }- zcs->blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, neededInBuffSize);- }- if (zcs->outBuffSize < ZSTD_compressBound(zcs->blockSize)+1) {- zcs->outBuffSize = ZSTD_compressBound(zcs->blockSize)+1;- ZSTD_free(zcs->outBuff, zcs->customMem);- zcs->outBuff = (char*) ZSTD_malloc(zcs->outBuffSize, zcs->customMem);- if (zcs->outBuff == NULL) return ERROR(memory_allocation);- }-- if (dict) {- ZSTD_freeCDict(zcs->cdictLocal);- zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize, 0, params, zcs->customMem);- if (zcs->cdictLocal == NULL) return ERROR(memory_allocation);- zcs->cdict = zcs->cdictLocal;- } else zcs->cdict = NULL;-- zcs->checksum = params.fParams.checksumFlag > 0;- zcs->params = params;-- return ZSTD_resetCStream(zcs, pledgedSrcSize);-}--/* note : cdict must outlive compression session */-size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict)-{- ZSTD_parameters const params = ZSTD_getParamsFromCDict(cdict);- size_t const initError = ZSTD_initCStream_advanced(zcs, NULL, 0, params, 0);- zcs->cdict = cdict;- return initError;-}--size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)-{- ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);- return ZSTD_initCStream_advanced(zcs, dict, dictSize, params, 0);-}--size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize)-{- ZSTD_parameters params = ZSTD_getParams(compressionLevel, pledgedSrcSize, 0);- if (pledgedSrcSize) params.fParams.contentSizeFlag = 1;- return ZSTD_initCStream_advanced(zcs, NULL, 0, params, pledgedSrcSize);-}--size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)-{- return ZSTD_initCStream_usingDict(zcs, NULL, 0, compressionLevel);-}--size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)-{- if (zcs==NULL) return 0; /* support sizeof on NULL */- return sizeof(zcs) + ZSTD_sizeof_CCtx(zcs->cctx) + ZSTD_sizeof_CDict(zcs->cdictLocal) + zcs->outBuffSize + zcs->inBuffSize;-}--/*====== Compression ======*/--typedef enum { zsf_gather, zsf_flush, zsf_end } ZSTD_flush_e;--MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- size_t const length = MIN(dstCapacity, srcSize);- memcpy(dst, src, length);- return length;-}--static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,- void* dst, size_t* dstCapacityPtr,- const void* src, size_t* srcSizePtr,- ZSTD_flush_e const flush)-{- U32 someMoreWork = 1;- const char* const istart = (const char*)src;- const char* const iend = istart + *srcSizePtr;- const char* ip = istart;- char* const ostart = (char*)dst;- char* const oend = ostart + *dstCapacityPtr;- char* op = ostart;-- while (someMoreWork) {- switch(zcs->stage)- {- case zcss_init: return ERROR(init_missing); /* call ZBUFF_compressInit() first ! */-- case zcss_load:- /* complete inBuffer */- { size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos;- size_t const loaded = ZSTD_limitCopy(zcs->inBuff + zcs->inBuffPos, toLoad, ip, iend-ip);- zcs->inBuffPos += loaded;- ip += loaded;- if ( (zcs->inBuffPos==zcs->inToCompress) || (!flush && (toLoad != loaded)) ) {- someMoreWork = 0; break; /* not enough input to get a full block : stop there, wait for more */- } }- /* compress current block (note : this stage cannot be stopped in the middle) */- { void* cDst;- size_t cSize;- size_t const iSize = zcs->inBuffPos - zcs->inToCompress;- size_t oSize = oend-op;- if (oSize >= ZSTD_compressBound(iSize))- cDst = op; /* compress directly into output buffer (avoid flush stage) */- else- cDst = zcs->outBuff, oSize = zcs->outBuffSize;- cSize = (flush == zsf_end) ?- ZSTD_compressEnd(zcs->cctx, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize) :- ZSTD_compressContinue(zcs->cctx, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize);- if (ZSTD_isError(cSize)) return cSize;- if (flush == zsf_end) zcs->frameEnded = 1;- /* prepare next block */- zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;- if (zcs->inBuffTarget > zcs->inBuffSize)- zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize; /* note : inBuffSize >= blockSize */- zcs->inToCompress = zcs->inBuffPos;- if (cDst == op) { op += cSize; break; } /* no need to flush */- zcs->outBuffContentSize = cSize;- zcs->outBuffFlushedSize = 0;- zcs->stage = zcss_flush; /* pass-through to flush stage */- }-- case zcss_flush:- { size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;- size_t const flushed = ZSTD_limitCopy(op, oend-op, zcs->outBuff + zcs->outBuffFlushedSize, toFlush);- op += flushed;- zcs->outBuffFlushedSize += flushed;- if (toFlush!=flushed) { someMoreWork = 0; break; } /* dst too small to store flushed data : stop there */- zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;- zcs->stage = zcss_load;- break;- }-- case zcss_final:- someMoreWork = 0; /* do nothing */- break;-- default:- return ERROR(GENERIC); /* impossible */- }- }-- *srcSizePtr = ip - istart;- *dstCapacityPtr = op - ostart;- zcs->inputProcessed += *srcSizePtr;- if (zcs->frameEnded) return 0;- { size_t hintInSize = zcs->inBuffTarget - zcs->inBuffPos;- if (hintInSize==0) hintInSize = zcs->blockSize;- return hintInSize;- }-}--size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input)-{- size_t sizeRead = input->size - input->pos;- size_t sizeWritten = output->size - output->pos;- size_t const result = ZSTD_compressStream_generic(zcs,- (char*)(output->dst) + output->pos, &sizeWritten,- (const char*)(input->src) + input->pos, &sizeRead, zsf_gather);- input->pos += sizeRead;- output->pos += sizeWritten;- return result;-}---/*====== Finalize ======*/--/*! ZSTD_flushStream() :-* @return : amount of data remaining to flush */-size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)-{- size_t srcSize = 0;- size_t sizeWritten = output->size - output->pos;- size_t const result = ZSTD_compressStream_generic(zcs,- (char*)(output->dst) + output->pos, &sizeWritten,- &srcSize, &srcSize, /* use a valid src address instead of NULL */- zsf_flush);- output->pos += sizeWritten;- if (ZSTD_isError(result)) return result;- return zcs->outBuffContentSize - zcs->outBuffFlushedSize; /* remaining to flush */-}---size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)-{- BYTE* const ostart = (BYTE*)(output->dst) + output->pos;- BYTE* const oend = (BYTE*)(output->dst) + output->size;- BYTE* op = ostart;-- if ((zcs->pledgedSrcSize) && (zcs->inputProcessed != zcs->pledgedSrcSize))- return ERROR(srcSize_wrong); /* pledgedSrcSize not respected */-- if (zcs->stage != zcss_final) {- /* flush whatever remains */- size_t srcSize = 0;- size_t sizeWritten = output->size - output->pos;- size_t const notEnded = ZSTD_compressStream_generic(zcs, ostart, &sizeWritten, &srcSize, &srcSize, zsf_end); /* use a valid src address instead of NULL */- size_t const remainingToFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;- op += sizeWritten;- if (remainingToFlush) {- output->pos += sizeWritten;- return remainingToFlush + ZSTD_BLOCKHEADERSIZE /* final empty block */ + (zcs->checksum * 4);- }- /* create epilogue */- zcs->stage = zcss_final;- zcs->outBuffContentSize = !notEnded ? 0 :- ZSTD_compressEnd(zcs->cctx, zcs->outBuff, zcs->outBuffSize, NULL, 0); /* write epilogue, including final empty block, into outBuff */- }-- /* flush epilogue */- { size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;- size_t const flushed = ZSTD_limitCopy(op, oend-op, zcs->outBuff + zcs->outBuffFlushedSize, toFlush);- op += flushed;- zcs->outBuffFlushedSize += flushed;- output->pos += op-ostart;- if (toFlush==flushed) zcs->stage = zcss_init; /* end reached */- return toFlush - flushed;- }-}----/*-===== Pre-defined compression levels =====-*/--#define ZSTD_DEFAULT_CLEVEL 1-#define ZSTD_MAX_CLEVEL 22-int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; }--static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = {-{ /* "default" */- /* W, C, H, S, L, TL, strat */- { 18, 12, 12, 1, 7, 16, ZSTD_fast }, /* level 0 - never used */- { 19, 13, 14, 1, 7, 16, ZSTD_fast }, /* level 1 */- { 19, 15, 16, 1, 6, 16, ZSTD_fast }, /* level 2 */- { 20, 16, 17, 1, 5, 16, ZSTD_dfast }, /* level 3.*/- { 20, 18, 18, 1, 5, 16, ZSTD_dfast }, /* level 4.*/- { 20, 15, 18, 3, 5, 16, ZSTD_greedy }, /* level 5 */- { 21, 16, 19, 2, 5, 16, ZSTD_lazy }, /* level 6 */- { 21, 17, 20, 3, 5, 16, ZSTD_lazy }, /* level 7 */- { 21, 18, 20, 3, 5, 16, ZSTD_lazy2 }, /* level 8 */- { 21, 20, 20, 3, 5, 16, ZSTD_lazy2 }, /* level 9 */- { 21, 19, 21, 4, 5, 16, ZSTD_lazy2 }, /* level 10 */- { 22, 20, 22, 4, 5, 16, ZSTD_lazy2 }, /* level 11 */- { 22, 20, 22, 5, 5, 16, ZSTD_lazy2 }, /* level 12 */- { 22, 21, 22, 5, 5, 16, ZSTD_lazy2 }, /* level 13 */- { 22, 21, 22, 6, 5, 16, ZSTD_lazy2 }, /* level 14 */- { 22, 21, 21, 5, 5, 16, ZSTD_btlazy2 }, /* level 15 */- { 23, 22, 22, 5, 5, 16, ZSTD_btlazy2 }, /* level 16 */- { 23, 21, 22, 4, 5, 24, ZSTD_btopt }, /* level 17 */- { 23, 23, 22, 6, 5, 32, ZSTD_btopt }, /* level 18 */- { 23, 23, 22, 6, 3, 48, ZSTD_btopt }, /* level 19 */- { 25, 25, 23, 7, 3, 64, ZSTD_btopt2 }, /* level 20 */- { 26, 26, 23, 7, 3,256, ZSTD_btopt2 }, /* level 21 */- { 27, 27, 25, 9, 3,512, ZSTD_btopt2 }, /* level 22 */-},-{ /* for srcSize <= 256 KB */- /* W, C, H, S, L, T, strat */- { 0, 0, 0, 0, 0, 0, ZSTD_fast }, /* level 0 - not used */- { 18, 13, 14, 1, 6, 8, ZSTD_fast }, /* level 1 */- { 18, 14, 13, 1, 5, 8, ZSTD_dfast }, /* level 2 */- { 18, 16, 15, 1, 5, 8, ZSTD_dfast }, /* level 3 */- { 18, 15, 17, 1, 5, 8, ZSTD_greedy }, /* level 4.*/- { 18, 16, 17, 4, 5, 8, ZSTD_greedy }, /* level 5.*/- { 18, 16, 17, 3, 5, 8, ZSTD_lazy }, /* level 6.*/- { 18, 17, 17, 4, 4, 8, ZSTD_lazy }, /* level 7 */- { 18, 17, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */- { 18, 17, 17, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */- { 18, 17, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */- { 18, 18, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 11.*/- { 18, 18, 17, 7, 4, 8, ZSTD_lazy2 }, /* level 12.*/- { 18, 19, 17, 6, 4, 8, ZSTD_btlazy2 }, /* level 13 */- { 18, 18, 18, 4, 4, 16, ZSTD_btopt }, /* level 14.*/- { 18, 18, 18, 4, 3, 16, ZSTD_btopt }, /* level 15.*/- { 18, 19, 18, 6, 3, 32, ZSTD_btopt }, /* level 16.*/- { 18, 19, 18, 8, 3, 64, ZSTD_btopt }, /* level 17.*/- { 18, 19, 18, 9, 3,128, ZSTD_btopt }, /* level 18.*/- { 18, 19, 18, 10, 3,256, ZSTD_btopt }, /* level 19.*/- { 18, 19, 18, 11, 3,512, ZSTD_btopt2 }, /* level 20.*/- { 18, 19, 18, 12, 3,512, ZSTD_btopt2 }, /* level 21.*/- { 18, 19, 18, 13, 3,512, ZSTD_btopt2 }, /* level 22.*/-},-{ /* for srcSize <= 128 KB */- /* W, C, H, S, L, T, strat */- { 17, 12, 12, 1, 7, 8, ZSTD_fast }, /* level 0 - not used */- { 17, 12, 13, 1, 6, 8, ZSTD_fast }, /* level 1 */- { 17, 13, 16, 1, 5, 8, ZSTD_fast }, /* level 2 */- { 17, 16, 16, 2, 5, 8, ZSTD_dfast }, /* level 3 */- { 17, 13, 15, 3, 4, 8, ZSTD_greedy }, /* level 4 */- { 17, 15, 17, 4, 4, 8, ZSTD_greedy }, /* level 5 */- { 17, 16, 17, 3, 4, 8, ZSTD_lazy }, /* level 6 */- { 17, 15, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 7 */- { 17, 17, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */- { 17, 17, 17, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */- { 17, 17, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */- { 17, 17, 17, 7, 4, 8, ZSTD_lazy2 }, /* level 11 */- { 17, 17, 17, 8, 4, 8, ZSTD_lazy2 }, /* level 12 */- { 17, 18, 17, 6, 4, 8, ZSTD_btlazy2 }, /* level 13.*/- { 17, 17, 17, 7, 3, 8, ZSTD_btopt }, /* level 14.*/- { 17, 17, 17, 7, 3, 16, ZSTD_btopt }, /* level 15.*/- { 17, 18, 17, 7, 3, 32, ZSTD_btopt }, /* level 16.*/- { 17, 18, 17, 7, 3, 64, ZSTD_btopt }, /* level 17.*/- { 17, 18, 17, 7, 3,256, ZSTD_btopt }, /* level 18.*/- { 17, 18, 17, 8, 3,256, ZSTD_btopt }, /* level 19.*/- { 17, 18, 17, 9, 3,256, ZSTD_btopt2 }, /* level 20.*/- { 17, 18, 17, 10, 3,256, ZSTD_btopt2 }, /* level 21.*/- { 17, 18, 17, 11, 3,512, ZSTD_btopt2 }, /* level 22.*/-},-{ /* for srcSize <= 16 KB */- /* W, C, H, S, L, T, strat */- { 14, 12, 12, 1, 7, 6, ZSTD_fast }, /* level 0 - not used */- { 14, 14, 14, 1, 6, 6, ZSTD_fast }, /* level 1 */- { 14, 14, 14, 1, 4, 6, ZSTD_fast }, /* level 2 */- { 14, 14, 14, 1, 4, 6, ZSTD_dfast }, /* level 3.*/- { 14, 14, 14, 4, 4, 6, ZSTD_greedy }, /* level 4.*/- { 14, 14, 14, 3, 4, 6, ZSTD_lazy }, /* level 5.*/- { 14, 14, 14, 4, 4, 6, ZSTD_lazy2 }, /* level 6 */- { 14, 14, 14, 5, 4, 6, ZSTD_lazy2 }, /* level 7 */- { 14, 14, 14, 6, 4, 6, ZSTD_lazy2 }, /* level 8.*/- { 14, 15, 14, 6, 4, 6, ZSTD_btlazy2 }, /* level 9.*/- { 14, 15, 14, 3, 3, 6, ZSTD_btopt }, /* level 10.*/- { 14, 15, 14, 6, 3, 8, ZSTD_btopt }, /* level 11.*/- { 14, 15, 14, 6, 3, 16, ZSTD_btopt }, /* level 12.*/- { 14, 15, 14, 6, 3, 24, ZSTD_btopt }, /* level 13.*/- { 14, 15, 15, 6, 3, 48, ZSTD_btopt }, /* level 14.*/- { 14, 15, 15, 6, 3, 64, ZSTD_btopt }, /* level 15.*/- { 14, 15, 15, 6, 3, 96, ZSTD_btopt }, /* level 16.*/- { 14, 15, 15, 6, 3,128, ZSTD_btopt }, /* level 17.*/- { 14, 15, 15, 6, 3,256, ZSTD_btopt }, /* level 18.*/- { 14, 15, 15, 7, 3,256, ZSTD_btopt }, /* level 19.*/- { 14, 15, 15, 8, 3,256, ZSTD_btopt2 }, /* level 20.*/- { 14, 15, 15, 9, 3,256, ZSTD_btopt2 }, /* level 21.*/- { 14, 15, 15, 10, 3,256, ZSTD_btopt2 }, /* level 22.*/-},-};--/*! ZSTD_getCParams() :-* @return ZSTD_compressionParameters structure for a selected compression level, `srcSize` and `dictSize`.-* Size values are optional, provide 0 if not known or unused */-ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSize, size_t dictSize)-{- ZSTD_compressionParameters cp;- size_t const addedSize = srcSize ? 0 : 500;- U64 const rSize = srcSize+dictSize ? srcSize+dictSize+addedSize : (U64)-1;- U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB); /* intentional underflow for srcSizeHint == 0 */- if (compressionLevel <= 0) compressionLevel = ZSTD_DEFAULT_CLEVEL; /* 0 == default; no negative compressionLevel yet */- if (compressionLevel > ZSTD_MAX_CLEVEL) compressionLevel = ZSTD_MAX_CLEVEL;- cp = ZSTD_defaultCParameters[tableID][compressionLevel];- if (MEM_32bits()) { /* auto-correction, for 32-bits mode */- if (cp.windowLog > ZSTD_WINDOWLOG_MAX) cp.windowLog = ZSTD_WINDOWLOG_MAX;- if (cp.chainLog > ZSTD_CHAINLOG_MAX) cp.chainLog = ZSTD_CHAINLOG_MAX;- if (cp.hashLog > ZSTD_HASHLOG_MAX) cp.hashLog = ZSTD_HASHLOG_MAX;- }- cp = ZSTD_adjustCParams(cp, srcSize, dictSize);- return cp;-}--/*! ZSTD_getParams() :-* same as ZSTD_getCParams(), but @return a `ZSTD_parameters` object (instead of `ZSTD_compressionParameters`).-* All fields of `ZSTD_frameParameters` are set to default (0) */-ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSize, size_t dictSize) {- ZSTD_parameters params;- ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSize, dictSize);- memset(¶ms, 0, sizeof(params));- params.cParams = cParams;+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++/*-*************************************+* Dependencies+***************************************/+#include <limits.h> /* INT_MAX */+#include <string.h> /* memset */+#include "cpu.h"+#include "mem.h"+#include "hist.h" /* HIST_countFast_wksp */+#define FSE_STATIC_LINKING_ONLY /* FSE_encodeSymbol */+#include "fse.h"+#define HUF_STATIC_LINKING_ONLY+#include "huf.h"+#include "zstd_compress_internal.h"+#include "zstd_fast.h"+#include "zstd_double_fast.h"+#include "zstd_lazy.h"+#include "zstd_opt.h"+#include "zstd_ldm.h"+++/*-*************************************+* Helper functions+***************************************/+size_t ZSTD_compressBound(size_t srcSize) {+ return ZSTD_COMPRESSBOUND(srcSize);+}+++/*-*************************************+* Context memory management+***************************************/+struct ZSTD_CDict_s {+ void* dictBuffer;+ const void* dictContent;+ size_t dictContentSize;+ void* workspace;+ size_t workspaceSize;+ ZSTD_matchState_t matchState;+ ZSTD_compressedBlockState_t cBlockState;+ ZSTD_customMem customMem;+ U32 dictID;+}; /* typedef'd to ZSTD_CDict within "zstd.h" */++ZSTD_CCtx* ZSTD_createCCtx(void)+{+ return ZSTD_createCCtx_advanced(ZSTD_defaultCMem);+}++static void ZSTD_initCCtx(ZSTD_CCtx* cctx, ZSTD_customMem memManager)+{+ assert(cctx != NULL);+ memset(cctx, 0, sizeof(*cctx));+ cctx->customMem = memManager;+ cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());+ { size_t const err = ZSTD_CCtx_reset(cctx, ZSTD_reset_parameters);+ assert(!ZSTD_isError(err));+ (void)err;+ }+}++ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem)+{+ ZSTD_STATIC_ASSERT(zcss_init==0);+ ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN==(0ULL - 1));+ if (!customMem.customAlloc ^ !customMem.customFree) return NULL;+ { ZSTD_CCtx* const cctx = (ZSTD_CCtx*)ZSTD_malloc(sizeof(ZSTD_CCtx), customMem);+ if (!cctx) return NULL;+ ZSTD_initCCtx(cctx, customMem);+ return cctx;+ }+}++ZSTD_CCtx* ZSTD_initStaticCCtx(void *workspace, size_t workspaceSize)+{+ ZSTD_CCtx* const cctx = (ZSTD_CCtx*) workspace;+ if (workspaceSize <= sizeof(ZSTD_CCtx)) return NULL; /* minimum size */+ if ((size_t)workspace & 7) return NULL; /* must be 8-aligned */+ memset(workspace, 0, workspaceSize); /* may be a bit generous, could memset be smaller ? */+ cctx->staticSize = workspaceSize;+ cctx->workSpace = (void*)(cctx+1);+ cctx->workSpaceSize = workspaceSize - sizeof(ZSTD_CCtx);++ /* statically sized space. entropyWorkspace never moves (but prev/next block swap places) */+ if (cctx->workSpaceSize < HUF_WORKSPACE_SIZE + 2 * sizeof(ZSTD_compressedBlockState_t)) return NULL;+ assert(((size_t)cctx->workSpace & (sizeof(void*)-1)) == 0); /* ensure correct alignment */+ cctx->blockState.prevCBlock = (ZSTD_compressedBlockState_t*)cctx->workSpace;+ cctx->blockState.nextCBlock = cctx->blockState.prevCBlock + 1;+ {+ void* const ptr = cctx->blockState.nextCBlock + 1;+ cctx->entropyWorkspace = (U32*)ptr;+ }+ cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());+ return cctx;+}++/**+ * Clears and frees all of the dictionaries in the CCtx.+ */+static void ZSTD_clearAllDicts(ZSTD_CCtx* cctx)+{+ ZSTD_free(cctx->localDict.dictBuffer, cctx->customMem);+ ZSTD_freeCDict(cctx->localDict.cdict);+ memset(&cctx->localDict, 0, sizeof(cctx->localDict));+ memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));+ cctx->cdict = NULL;+}++static size_t ZSTD_sizeof_localDict(ZSTD_localDict dict)+{+ size_t const bufferSize = dict.dictBuffer != NULL ? dict.dictSize : 0;+ size_t const cdictSize = ZSTD_sizeof_CDict(dict.cdict);+ return bufferSize + cdictSize;+}++static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx)+{+ assert(cctx != NULL);+ assert(cctx->staticSize == 0);+ ZSTD_free(cctx->workSpace, cctx->customMem); cctx->workSpace = NULL;+ ZSTD_clearAllDicts(cctx);+#ifdef ZSTD_MULTITHREAD+ ZSTDMT_freeCCtx(cctx->mtctx); cctx->mtctx = NULL;+#endif+}++size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)+{+ if (cctx==NULL) return 0; /* support free on NULL */+ RETURN_ERROR_IF(cctx->staticSize, memory_allocation,+ "not compatible with static CCtx");+ ZSTD_freeCCtxContent(cctx);+ ZSTD_free(cctx, cctx->customMem);+ return 0;+}+++static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx)+{+#ifdef ZSTD_MULTITHREAD+ return ZSTDMT_sizeof_CCtx(cctx->mtctx);+#else+ (void)cctx;+ return 0;+#endif+}+++size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx)+{+ if (cctx==NULL) return 0; /* support sizeof on NULL */+ return sizeof(*cctx) + cctx->workSpaceSize+ + ZSTD_sizeof_localDict(cctx->localDict)+ + ZSTD_sizeof_mtctx(cctx);+}++size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)+{+ return ZSTD_sizeof_CCtx(zcs); /* same object */+}++/* private API call, for dictBuilder only */+const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx) { return &(ctx->seqStore); }++static ZSTD_CCtx_params ZSTD_makeCCtxParamsFromCParams(+ ZSTD_compressionParameters cParams)+{+ ZSTD_CCtx_params cctxParams;+ memset(&cctxParams, 0, sizeof(cctxParams));+ cctxParams.cParams = cParams;+ cctxParams.compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */+ assert(!ZSTD_checkCParams(cParams));+ cctxParams.fParams.contentSizeFlag = 1;+ return cctxParams;+}++static ZSTD_CCtx_params* ZSTD_createCCtxParams_advanced(+ ZSTD_customMem customMem)+{+ ZSTD_CCtx_params* params;+ if (!customMem.customAlloc ^ !customMem.customFree) return NULL;+ params = (ZSTD_CCtx_params*)ZSTD_calloc(+ sizeof(ZSTD_CCtx_params), customMem);+ if (!params) { return NULL; }+ params->customMem = customMem;+ params->compressionLevel = ZSTD_CLEVEL_DEFAULT;+ params->fParams.contentSizeFlag = 1;+ return params;+}++ZSTD_CCtx_params* ZSTD_createCCtxParams(void)+{+ return ZSTD_createCCtxParams_advanced(ZSTD_defaultCMem);+}++size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params)+{+ if (params == NULL) { return 0; }+ ZSTD_free(params, params->customMem);+ return 0;+}++size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params)+{+ return ZSTD_CCtxParams_init(params, ZSTD_CLEVEL_DEFAULT);+}++size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel) {+ RETURN_ERROR_IF(!cctxParams, GENERIC);+ memset(cctxParams, 0, sizeof(*cctxParams));+ cctxParams->compressionLevel = compressionLevel;+ cctxParams->fParams.contentSizeFlag = 1;+ return 0;+}++size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params)+{+ RETURN_ERROR_IF(!cctxParams, GENERIC);+ FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );+ memset(cctxParams, 0, sizeof(*cctxParams));+ cctxParams->cParams = params.cParams;+ cctxParams->fParams = params.fParams;+ cctxParams->compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */+ assert(!ZSTD_checkCParams(params.cParams));+ return 0;+}++/* ZSTD_assignParamsToCCtxParams() :+ * params is presumed valid at this stage */+static ZSTD_CCtx_params ZSTD_assignParamsToCCtxParams(+ ZSTD_CCtx_params cctxParams, ZSTD_parameters params)+{+ ZSTD_CCtx_params ret = cctxParams;+ ret.cParams = params.cParams;+ ret.fParams = params.fParams;+ ret.compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */+ assert(!ZSTD_checkCParams(params.cParams));+ return ret;+}++ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter param)+{+ ZSTD_bounds bounds = { 0, 0, 0 };++ switch(param)+ {+ case ZSTD_c_compressionLevel:+ bounds.lowerBound = ZSTD_minCLevel();+ bounds.upperBound = ZSTD_maxCLevel();+ return bounds;++ case ZSTD_c_windowLog:+ bounds.lowerBound = ZSTD_WINDOWLOG_MIN;+ bounds.upperBound = ZSTD_WINDOWLOG_MAX;+ return bounds;++ case ZSTD_c_hashLog:+ bounds.lowerBound = ZSTD_HASHLOG_MIN;+ bounds.upperBound = ZSTD_HASHLOG_MAX;+ return bounds;++ case ZSTD_c_chainLog:+ bounds.lowerBound = ZSTD_CHAINLOG_MIN;+ bounds.upperBound = ZSTD_CHAINLOG_MAX;+ return bounds;++ case ZSTD_c_searchLog:+ bounds.lowerBound = ZSTD_SEARCHLOG_MIN;+ bounds.upperBound = ZSTD_SEARCHLOG_MAX;+ return bounds;++ case ZSTD_c_minMatch:+ bounds.lowerBound = ZSTD_MINMATCH_MIN;+ bounds.upperBound = ZSTD_MINMATCH_MAX;+ return bounds;++ case ZSTD_c_targetLength:+ bounds.lowerBound = ZSTD_TARGETLENGTH_MIN;+ bounds.upperBound = ZSTD_TARGETLENGTH_MAX;+ return bounds;++ case ZSTD_c_strategy:+ bounds.lowerBound = ZSTD_STRATEGY_MIN;+ bounds.upperBound = ZSTD_STRATEGY_MAX;+ return bounds;++ case ZSTD_c_contentSizeFlag:+ bounds.lowerBound = 0;+ bounds.upperBound = 1;+ return bounds;++ case ZSTD_c_checksumFlag:+ bounds.lowerBound = 0;+ bounds.upperBound = 1;+ return bounds;++ case ZSTD_c_dictIDFlag:+ bounds.lowerBound = 0;+ bounds.upperBound = 1;+ return bounds;++ case ZSTD_c_nbWorkers:+ bounds.lowerBound = 0;+#ifdef ZSTD_MULTITHREAD+ bounds.upperBound = ZSTDMT_NBWORKERS_MAX;+#else+ bounds.upperBound = 0;+#endif+ return bounds;++ case ZSTD_c_jobSize:+ bounds.lowerBound = 0;+#ifdef ZSTD_MULTITHREAD+ bounds.upperBound = ZSTDMT_JOBSIZE_MAX;+#else+ bounds.upperBound = 0;+#endif+ return bounds;++ case ZSTD_c_overlapLog:+ bounds.lowerBound = ZSTD_OVERLAPLOG_MIN;+ bounds.upperBound = ZSTD_OVERLAPLOG_MAX;+ return bounds;++ case ZSTD_c_enableLongDistanceMatching:+ bounds.lowerBound = 0;+ bounds.upperBound = 1;+ return bounds;++ case ZSTD_c_ldmHashLog:+ bounds.lowerBound = ZSTD_LDM_HASHLOG_MIN;+ bounds.upperBound = ZSTD_LDM_HASHLOG_MAX;+ return bounds;++ case ZSTD_c_ldmMinMatch:+ bounds.lowerBound = ZSTD_LDM_MINMATCH_MIN;+ bounds.upperBound = ZSTD_LDM_MINMATCH_MAX;+ return bounds;++ case ZSTD_c_ldmBucketSizeLog:+ bounds.lowerBound = ZSTD_LDM_BUCKETSIZELOG_MIN;+ bounds.upperBound = ZSTD_LDM_BUCKETSIZELOG_MAX;+ return bounds;++ case ZSTD_c_ldmHashRateLog:+ bounds.lowerBound = ZSTD_LDM_HASHRATELOG_MIN;+ bounds.upperBound = ZSTD_LDM_HASHRATELOG_MAX;+ return bounds;++ /* experimental parameters */+ case ZSTD_c_rsyncable:+ bounds.lowerBound = 0;+ bounds.upperBound = 1;+ return bounds;++ case ZSTD_c_forceMaxWindow :+ bounds.lowerBound = 0;+ bounds.upperBound = 1;+ return bounds;++ case ZSTD_c_format:+ ZSTD_STATIC_ASSERT(ZSTD_f_zstd1 < ZSTD_f_zstd1_magicless);+ bounds.lowerBound = ZSTD_f_zstd1;+ bounds.upperBound = ZSTD_f_zstd1_magicless; /* note : how to ensure at compile time that this is the highest value enum ? */+ return bounds;++ case ZSTD_c_forceAttachDict:+ ZSTD_STATIC_ASSERT(ZSTD_dictDefaultAttach < ZSTD_dictForceCopy);+ bounds.lowerBound = ZSTD_dictDefaultAttach;+ bounds.upperBound = ZSTD_dictForceCopy; /* note : how to ensure at compile time that this is the highest value enum ? */+ return bounds;++ case ZSTD_c_literalCompressionMode:+ ZSTD_STATIC_ASSERT(ZSTD_lcm_auto < ZSTD_lcm_huffman && ZSTD_lcm_huffman < ZSTD_lcm_uncompressed);+ bounds.lowerBound = ZSTD_lcm_auto;+ bounds.upperBound = ZSTD_lcm_uncompressed;+ return bounds;++ default:+ { ZSTD_bounds const boundError = { ERROR(parameter_unsupported), 0, 0 };+ return boundError;+ }+ }+}++/* ZSTD_cParam_withinBounds:+ * @return 1 if value is within cParam bounds,+ * 0 otherwise */+static int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value)+{+ ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);+ if (ZSTD_isError(bounds.error)) return 0;+ if (value < bounds.lowerBound) return 0;+ if (value > bounds.upperBound) return 0;+ return 1;+}++/* ZSTD_cParam_clampBounds:+ * Clamps the value into the bounded range.+ */+static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value)+{+ ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);+ if (ZSTD_isError(bounds.error)) return bounds.error;+ if (*value < bounds.lowerBound) *value = bounds.lowerBound;+ if (*value > bounds.upperBound) *value = bounds.upperBound;+ return 0;+}++#define BOUNDCHECK(cParam, val) { \+ RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \+ parameter_outOfBound); \+}+++static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param)+{+ switch(param)+ {+ case ZSTD_c_compressionLevel:+ case ZSTD_c_hashLog:+ case ZSTD_c_chainLog:+ case ZSTD_c_searchLog:+ case ZSTD_c_minMatch:+ case ZSTD_c_targetLength:+ case ZSTD_c_strategy:+ return 1;++ case ZSTD_c_format:+ case ZSTD_c_windowLog:+ case ZSTD_c_contentSizeFlag:+ case ZSTD_c_checksumFlag:+ case ZSTD_c_dictIDFlag:+ case ZSTD_c_forceMaxWindow :+ case ZSTD_c_nbWorkers:+ case ZSTD_c_jobSize:+ case ZSTD_c_overlapLog:+ case ZSTD_c_rsyncable:+ case ZSTD_c_enableLongDistanceMatching:+ case ZSTD_c_ldmHashLog:+ case ZSTD_c_ldmMinMatch:+ case ZSTD_c_ldmBucketSizeLog:+ case ZSTD_c_ldmHashRateLog:+ case ZSTD_c_forceAttachDict:+ case ZSTD_c_literalCompressionMode:+ default:+ return 0;+ }+}++size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value)+{+ DEBUGLOG(4, "ZSTD_CCtx_setParameter (%i, %i)", (int)param, value);+ if (cctx->streamStage != zcss_init) {+ if (ZSTD_isUpdateAuthorized(param)) {+ cctx->cParamsChanged = 1;+ } else {+ RETURN_ERROR(stage_wrong);+ } }++ switch(param)+ {+ case ZSTD_c_nbWorkers:+ RETURN_ERROR_IF((value!=0) && cctx->staticSize, parameter_unsupported,+ "MT not compatible with static alloc");+ break;++ case ZSTD_c_compressionLevel:+ case ZSTD_c_windowLog:+ case ZSTD_c_hashLog:+ case ZSTD_c_chainLog:+ case ZSTD_c_searchLog:+ case ZSTD_c_minMatch:+ case ZSTD_c_targetLength:+ case ZSTD_c_strategy:+ case ZSTD_c_ldmHashRateLog:+ case ZSTD_c_format:+ case ZSTD_c_contentSizeFlag:+ case ZSTD_c_checksumFlag:+ case ZSTD_c_dictIDFlag:+ case ZSTD_c_forceMaxWindow:+ case ZSTD_c_forceAttachDict:+ case ZSTD_c_literalCompressionMode:+ case ZSTD_c_jobSize:+ case ZSTD_c_overlapLog:+ case ZSTD_c_rsyncable:+ case ZSTD_c_enableLongDistanceMatching:+ case ZSTD_c_ldmHashLog:+ case ZSTD_c_ldmMinMatch:+ case ZSTD_c_ldmBucketSizeLog:+ break;++ default: RETURN_ERROR(parameter_unsupported);+ }+ return ZSTD_CCtxParams_setParameter(&cctx->requestedParams, param, value);+}++size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,+ ZSTD_cParameter param, int value)+{+ DEBUGLOG(4, "ZSTD_CCtxParams_setParameter (%i, %i)", (int)param, value);+ switch(param)+ {+ case ZSTD_c_format :+ BOUNDCHECK(ZSTD_c_format, value);+ CCtxParams->format = (ZSTD_format_e)value;+ return (size_t)CCtxParams->format;++ case ZSTD_c_compressionLevel : {+ FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));+ if (value) { /* 0 : does not change current level */+ CCtxParams->compressionLevel = value;+ }+ if (CCtxParams->compressionLevel >= 0) return CCtxParams->compressionLevel;+ return 0; /* return type (size_t) cannot represent negative values */+ }++ case ZSTD_c_windowLog :+ if (value!=0) /* 0 => use default */+ BOUNDCHECK(ZSTD_c_windowLog, value);+ CCtxParams->cParams.windowLog = value;+ return CCtxParams->cParams.windowLog;++ case ZSTD_c_hashLog :+ if (value!=0) /* 0 => use default */+ BOUNDCHECK(ZSTD_c_hashLog, value);+ CCtxParams->cParams.hashLog = value;+ return CCtxParams->cParams.hashLog;++ case ZSTD_c_chainLog :+ if (value!=0) /* 0 => use default */+ BOUNDCHECK(ZSTD_c_chainLog, value);+ CCtxParams->cParams.chainLog = value;+ return CCtxParams->cParams.chainLog;++ case ZSTD_c_searchLog :+ if (value!=0) /* 0 => use default */+ BOUNDCHECK(ZSTD_c_searchLog, value);+ CCtxParams->cParams.searchLog = value;+ return value;++ case ZSTD_c_minMatch :+ if (value!=0) /* 0 => use default */+ BOUNDCHECK(ZSTD_c_minMatch, value);+ CCtxParams->cParams.minMatch = value;+ return CCtxParams->cParams.minMatch;++ case ZSTD_c_targetLength :+ BOUNDCHECK(ZSTD_c_targetLength, value);+ CCtxParams->cParams.targetLength = value;+ return CCtxParams->cParams.targetLength;++ case ZSTD_c_strategy :+ if (value!=0) /* 0 => use default */+ BOUNDCHECK(ZSTD_c_strategy, value);+ CCtxParams->cParams.strategy = (ZSTD_strategy)value;+ return (size_t)CCtxParams->cParams.strategy;++ case ZSTD_c_contentSizeFlag :+ /* Content size written in frame header _when known_ (default:1) */+ DEBUGLOG(4, "set content size flag = %u", (value!=0));+ CCtxParams->fParams.contentSizeFlag = value != 0;+ return CCtxParams->fParams.contentSizeFlag;++ case ZSTD_c_checksumFlag :+ /* A 32-bits content checksum will be calculated and written at end of frame (default:0) */+ CCtxParams->fParams.checksumFlag = value != 0;+ return CCtxParams->fParams.checksumFlag;++ case ZSTD_c_dictIDFlag : /* When applicable, dictionary's dictID is provided in frame header (default:1) */+ DEBUGLOG(4, "set dictIDFlag = %u", (value!=0));+ CCtxParams->fParams.noDictIDFlag = !value;+ return !CCtxParams->fParams.noDictIDFlag;++ case ZSTD_c_forceMaxWindow :+ CCtxParams->forceWindow = (value != 0);+ return CCtxParams->forceWindow;++ case ZSTD_c_forceAttachDict : {+ const ZSTD_dictAttachPref_e pref = (ZSTD_dictAttachPref_e)value;+ BOUNDCHECK(ZSTD_c_forceAttachDict, pref);+ CCtxParams->attachDictPref = pref;+ return CCtxParams->attachDictPref;+ }++ case ZSTD_c_literalCompressionMode : {+ const ZSTD_literalCompressionMode_e lcm = (ZSTD_literalCompressionMode_e)value;+ BOUNDCHECK(ZSTD_c_literalCompressionMode, lcm);+ CCtxParams->literalCompressionMode = lcm;+ return CCtxParams->literalCompressionMode;+ }++ case ZSTD_c_nbWorkers :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");+ return 0;+#else+ FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));+ CCtxParams->nbWorkers = value;+ return CCtxParams->nbWorkers;+#endif++ case ZSTD_c_jobSize :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");+ return 0;+#else+ /* Adjust to the minimum non-default value. */+ if (value != 0 && value < ZSTDMT_JOBSIZE_MIN)+ value = ZSTDMT_JOBSIZE_MIN;+ FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));+ assert(value >= 0);+ CCtxParams->jobSize = value;+ return CCtxParams->jobSize;+#endif++ case ZSTD_c_overlapLog :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");+ return 0;+#else+ FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value));+ CCtxParams->overlapLog = value;+ return CCtxParams->overlapLog;+#endif++ case ZSTD_c_rsyncable :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");+ return 0;+#else+ FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value));+ CCtxParams->rsyncable = value;+ return CCtxParams->rsyncable;+#endif++ case ZSTD_c_enableLongDistanceMatching :+ CCtxParams->ldmParams.enableLdm = (value!=0);+ return CCtxParams->ldmParams.enableLdm;++ case ZSTD_c_ldmHashLog :+ if (value!=0) /* 0 ==> auto */+ BOUNDCHECK(ZSTD_c_ldmHashLog, value);+ CCtxParams->ldmParams.hashLog = value;+ return CCtxParams->ldmParams.hashLog;++ case ZSTD_c_ldmMinMatch :+ if (value!=0) /* 0 ==> default */+ BOUNDCHECK(ZSTD_c_ldmMinMatch, value);+ CCtxParams->ldmParams.minMatchLength = value;+ return CCtxParams->ldmParams.minMatchLength;++ case ZSTD_c_ldmBucketSizeLog :+ if (value!=0) /* 0 ==> default */+ BOUNDCHECK(ZSTD_c_ldmBucketSizeLog, value);+ CCtxParams->ldmParams.bucketSizeLog = value;+ return CCtxParams->ldmParams.bucketSizeLog;++ case ZSTD_c_ldmHashRateLog :+ RETURN_ERROR_IF(value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN,+ parameter_outOfBound);+ CCtxParams->ldmParams.hashRateLog = value;+ return CCtxParams->ldmParams.hashRateLog;++ default: RETURN_ERROR(parameter_unsupported, "unknown parameter");+ }+}++size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value)+{+ return ZSTD_CCtxParams_getParameter(&cctx->requestedParams, param, value);+}++size_t ZSTD_CCtxParams_getParameter(+ ZSTD_CCtx_params* CCtxParams, ZSTD_cParameter param, int* value)+{+ switch(param)+ {+ case ZSTD_c_format :+ *value = CCtxParams->format;+ break;+ case ZSTD_c_compressionLevel :+ *value = CCtxParams->compressionLevel;+ break;+ case ZSTD_c_windowLog :+ *value = CCtxParams->cParams.windowLog;+ break;+ case ZSTD_c_hashLog :+ *value = CCtxParams->cParams.hashLog;+ break;+ case ZSTD_c_chainLog :+ *value = CCtxParams->cParams.chainLog;+ break;+ case ZSTD_c_searchLog :+ *value = CCtxParams->cParams.searchLog;+ break;+ case ZSTD_c_minMatch :+ *value = CCtxParams->cParams.minMatch;+ break;+ case ZSTD_c_targetLength :+ *value = CCtxParams->cParams.targetLength;+ break;+ case ZSTD_c_strategy :+ *value = (unsigned)CCtxParams->cParams.strategy;+ break;+ case ZSTD_c_contentSizeFlag :+ *value = CCtxParams->fParams.contentSizeFlag;+ break;+ case ZSTD_c_checksumFlag :+ *value = CCtxParams->fParams.checksumFlag;+ break;+ case ZSTD_c_dictIDFlag :+ *value = !CCtxParams->fParams.noDictIDFlag;+ break;+ case ZSTD_c_forceMaxWindow :+ *value = CCtxParams->forceWindow;+ break;+ case ZSTD_c_forceAttachDict :+ *value = CCtxParams->attachDictPref;+ break;+ case ZSTD_c_literalCompressionMode :+ *value = CCtxParams->literalCompressionMode;+ break;+ case ZSTD_c_nbWorkers :+#ifndef ZSTD_MULTITHREAD+ assert(CCtxParams->nbWorkers == 0);+#endif+ *value = CCtxParams->nbWorkers;+ break;+ case ZSTD_c_jobSize :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");+#else+ assert(CCtxParams->jobSize <= INT_MAX);+ *value = (int)CCtxParams->jobSize;+ break;+#endif+ case ZSTD_c_overlapLog :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");+#else+ *value = CCtxParams->overlapLog;+ break;+#endif+ case ZSTD_c_rsyncable :+#ifndef ZSTD_MULTITHREAD+ RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");+#else+ *value = CCtxParams->rsyncable;+ break;+#endif+ case ZSTD_c_enableLongDistanceMatching :+ *value = CCtxParams->ldmParams.enableLdm;+ break;+ case ZSTD_c_ldmHashLog :+ *value = CCtxParams->ldmParams.hashLog;+ break;+ case ZSTD_c_ldmMinMatch :+ *value = CCtxParams->ldmParams.minMatchLength;+ break;+ case ZSTD_c_ldmBucketSizeLog :+ *value = CCtxParams->ldmParams.bucketSizeLog;+ break;+ case ZSTD_c_ldmHashRateLog :+ *value = CCtxParams->ldmParams.hashRateLog;+ break;+ default: RETURN_ERROR(parameter_unsupported, "unknown parameter");+ }+ return 0;+}++/** ZSTD_CCtx_setParametersUsingCCtxParams() :+ * just applies `params` into `cctx`+ * no action is performed, parameters are merely stored.+ * If ZSTDMT is enabled, parameters are pushed to cctx->mtctx.+ * This is possible even if a compression is ongoing.+ * In which case, new parameters will be applied on the fly, starting with next compression job.+ */+size_t ZSTD_CCtx_setParametersUsingCCtxParams(+ ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params)+{+ DEBUGLOG(4, "ZSTD_CCtx_setParametersUsingCCtxParams");+ RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);+ RETURN_ERROR_IF(cctx->cdict, stage_wrong);++ cctx->requestedParams = *params;+ return 0;+}++ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %u bytes", (U32)pledgedSrcSize);+ RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);+ cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;+ return 0;+}++/**+ * Initializes the local dict using the requested parameters.+ * NOTE: This does not use the pledged src size, because it may be used for more+ * than one compression.+ */+static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx)+{+ ZSTD_localDict* const dl = &cctx->localDict;+ ZSTD_compressionParameters const cParams = ZSTD_getCParamsFromCCtxParams(+ &cctx->requestedParams, 0, dl->dictSize);+ if (dl->dict == NULL) {+ /* No local dictionary. */+ assert(dl->dictBuffer == NULL);+ assert(dl->cdict == NULL);+ assert(dl->dictSize == 0);+ return 0;+ }+ if (dl->cdict != NULL) {+ assert(cctx->cdict == dl->cdict);+ /* Local dictionary already initialized. */+ return 0;+ }+ assert(dl->dictSize > 0);+ assert(cctx->cdict == NULL);+ assert(cctx->prefixDict.dict == NULL);++ dl->cdict = ZSTD_createCDict_advanced(+ dl->dict,+ dl->dictSize,+ ZSTD_dlm_byRef,+ dl->dictContentType,+ cParams,+ cctx->customMem);+ RETURN_ERROR_IF(!dl->cdict, memory_allocation);+ cctx->cdict = dl->cdict;+ return 0;+}++size_t ZSTD_CCtx_loadDictionary_advanced(+ ZSTD_CCtx* cctx, const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType)+{+ RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);+ RETURN_ERROR_IF(cctx->staticSize, memory_allocation,+ "no malloc for static CCtx");+ DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize);+ ZSTD_clearAllDicts(cctx); /* in case one already exists */+ if (dict == NULL || dictSize == 0) /* no dictionary mode */+ return 0;+ if (dictLoadMethod == ZSTD_dlm_byRef) {+ cctx->localDict.dict = dict;+ } else {+ void* dictBuffer = ZSTD_malloc(dictSize, cctx->customMem);+ RETURN_ERROR_IF(!dictBuffer, memory_allocation);+ memcpy(dictBuffer, dict, dictSize);+ cctx->localDict.dictBuffer = dictBuffer;+ cctx->localDict.dict = dictBuffer;+ }+ cctx->localDict.dictSize = dictSize;+ cctx->localDict.dictContentType = dictContentType;+ return 0;+}++ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(+ ZSTD_CCtx* cctx, const void* dict, size_t dictSize)+{+ return ZSTD_CCtx_loadDictionary_advanced(+ cctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto);+}++ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize)+{+ return ZSTD_CCtx_loadDictionary_advanced(+ cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto);+}+++size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)+{+ RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);+ /* Free the existing local cdict (if any) to save memory. */+ ZSTD_clearAllDicts(cctx);+ cctx->cdict = cdict;+ return 0;+}++size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize)+{+ return ZSTD_CCtx_refPrefix_advanced(cctx, prefix, prefixSize, ZSTD_dct_rawContent);+}++size_t ZSTD_CCtx_refPrefix_advanced(+ ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)+{+ RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);+ ZSTD_clearAllDicts(cctx);+ cctx->prefixDict.dict = prefix;+ cctx->prefixDict.dictSize = prefixSize;+ cctx->prefixDict.dictContentType = dictContentType;+ return 0;+}++/*! ZSTD_CCtx_reset() :+ * Also dumps dictionary */+size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset)+{+ if ( (reset == ZSTD_reset_session_only)+ || (reset == ZSTD_reset_session_and_parameters) ) {+ cctx->streamStage = zcss_init;+ cctx->pledgedSrcSizePlusOne = 0;+ }+ if ( (reset == ZSTD_reset_parameters)+ || (reset == ZSTD_reset_session_and_parameters) ) {+ RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);+ ZSTD_clearAllDicts(cctx);+ return ZSTD_CCtxParams_reset(&cctx->requestedParams);+ }+ return 0;+}+++/** ZSTD_checkCParams() :+ control CParam values remain within authorized range.+ @return : 0, or an error code if one value is beyond authorized range */+size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)+{+ BOUNDCHECK(ZSTD_c_windowLog, cParams.windowLog);+ BOUNDCHECK(ZSTD_c_chainLog, cParams.chainLog);+ BOUNDCHECK(ZSTD_c_hashLog, cParams.hashLog);+ BOUNDCHECK(ZSTD_c_searchLog, cParams.searchLog);+ BOUNDCHECK(ZSTD_c_minMatch, cParams.minMatch);+ BOUNDCHECK(ZSTD_c_targetLength,cParams.targetLength);+ BOUNDCHECK(ZSTD_c_strategy, cParams.strategy);+ return 0;+}++/** ZSTD_clampCParams() :+ * make CParam values within valid range.+ * @return : valid CParams */+static ZSTD_compressionParameters+ZSTD_clampCParams(ZSTD_compressionParameters cParams)+{+# define CLAMP_TYPE(cParam, val, type) { \+ ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); \+ if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound; \+ else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \+ }+# define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, int)+ CLAMP(ZSTD_c_windowLog, cParams.windowLog);+ CLAMP(ZSTD_c_chainLog, cParams.chainLog);+ CLAMP(ZSTD_c_hashLog, cParams.hashLog);+ CLAMP(ZSTD_c_searchLog, cParams.searchLog);+ CLAMP(ZSTD_c_minMatch, cParams.minMatch);+ CLAMP(ZSTD_c_targetLength,cParams.targetLength);+ CLAMP_TYPE(ZSTD_c_strategy,cParams.strategy, ZSTD_strategy);+ return cParams;+}++/** ZSTD_cycleLog() :+ * condition for correct operation : hashLog > 1 */+static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)+{+ U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);+ return hashLog - btScale;+}++/** ZSTD_adjustCParams_internal() :+ * optimize `cPar` for a specified input (`srcSize` and `dictSize`).+ * mostly downsize to reduce memory consumption and initialization latency.+ * `srcSize` can be ZSTD_CONTENTSIZE_UNKNOWN when not known.+ * note : for the time being, `srcSize==0` means "unknown" too, for compatibility with older convention.+ * condition : cPar is presumed validated (can be checked using ZSTD_checkCParams()). */+static ZSTD_compressionParameters+ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar,+ unsigned long long srcSize,+ size_t dictSize)+{+ static const U64 minSrcSize = 513; /* (1<<9) + 1 */+ static const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1);+ assert(ZSTD_checkCParams(cPar)==0);++ if (dictSize && (srcSize+1<2) /* ZSTD_CONTENTSIZE_UNKNOWN and 0 mean "unknown" */ )+ srcSize = minSrcSize; /* presumed small when there is a dictionary */+ else if (srcSize == 0)+ srcSize = ZSTD_CONTENTSIZE_UNKNOWN; /* 0 == unknown : presumed large */++ /* resize windowLog if input is small enough, to use less memory */+ if ( (srcSize < maxWindowResize)+ && (dictSize < maxWindowResize) ) {+ U32 const tSize = (U32)(srcSize + dictSize);+ static U32 const hashSizeMin = 1 << ZSTD_HASHLOG_MIN;+ U32 const srcLog = (tSize < hashSizeMin) ? ZSTD_HASHLOG_MIN :+ ZSTD_highbit32(tSize-1) + 1;+ if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;+ }+ if (cPar.hashLog > cPar.windowLog+1) cPar.hashLog = cPar.windowLog+1;+ { U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);+ if (cycleLog > cPar.windowLog)+ cPar.chainLog -= (cycleLog - cPar.windowLog);+ }++ if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)+ cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* minimum wlog required for valid frame header */++ return cPar;+}++ZSTD_compressionParameters+ZSTD_adjustCParams(ZSTD_compressionParameters cPar,+ unsigned long long srcSize,+ size_t dictSize)+{+ cPar = ZSTD_clampCParams(cPar); /* resulting cPar is necessarily valid (all parameters within range) */+ return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize);+}++ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(+ const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize)+{+ ZSTD_compressionParameters cParams = ZSTD_getCParams(CCtxParams->compressionLevel, srcSizeHint, dictSize);+ if (CCtxParams->ldmParams.enableLdm) cParams.windowLog = ZSTD_LDM_DEFAULT_WINDOW_LOG;+ if (CCtxParams->cParams.windowLog) cParams.windowLog = CCtxParams->cParams.windowLog;+ if (CCtxParams->cParams.hashLog) cParams.hashLog = CCtxParams->cParams.hashLog;+ if (CCtxParams->cParams.chainLog) cParams.chainLog = CCtxParams->cParams.chainLog;+ if (CCtxParams->cParams.searchLog) cParams.searchLog = CCtxParams->cParams.searchLog;+ if (CCtxParams->cParams.minMatch) cParams.minMatch = CCtxParams->cParams.minMatch;+ if (CCtxParams->cParams.targetLength) cParams.targetLength = CCtxParams->cParams.targetLength;+ if (CCtxParams->cParams.strategy) cParams.strategy = CCtxParams->cParams.strategy;+ assert(!ZSTD_checkCParams(cParams));+ return ZSTD_adjustCParams_internal(cParams, srcSizeHint, dictSize);+}++static size_t+ZSTD_sizeof_matchState(const ZSTD_compressionParameters* const cParams,+ const U32 forCCtx)+{+ size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog);+ size_t const hSize = ((size_t)1) << cParams->hashLog;+ U32 const hashLog3 = (forCCtx && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;+ size_t const h3Size = ((size_t)1) << hashLog3;+ size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);+ size_t const optPotentialSpace = ((MaxML+1) + (MaxLL+1) + (MaxOff+1) + (1<<Litbits)) * sizeof(U32)+ + (ZSTD_OPT_NUM+1) * (sizeof(ZSTD_match_t)+sizeof(ZSTD_optimal_t));+ size_t const optSpace = (forCCtx && (cParams->strategy >= ZSTD_btopt))+ ? optPotentialSpace+ : 0;+ DEBUGLOG(4, "chainSize: %u - hSize: %u - h3Size: %u",+ (U32)chainSize, (U32)hSize, (U32)h3Size);+ return tableSpace + optSpace;+}++size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params)+{+ RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");+ { ZSTD_compressionParameters const cParams =+ ZSTD_getCParamsFromCCtxParams(params, 0, 0);+ size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);+ U32 const divider = (cParams.minMatch==3) ? 3 : 4;+ size_t const maxNbSeq = blockSize / divider;+ size_t const tokenSpace = WILDCOPY_OVERLENGTH + blockSize + 11*maxNbSeq;+ size_t const entropySpace = HUF_WORKSPACE_SIZE;+ size_t const blockStateSpace = 2 * sizeof(ZSTD_compressedBlockState_t);+ size_t const matchStateSize = ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 1);++ size_t const ldmSpace = ZSTD_ldm_getTableSize(params->ldmParams);+ size_t const ldmSeqSpace = ZSTD_ldm_getMaxNbSeq(params->ldmParams, blockSize) * sizeof(rawSeq);++ size_t const neededSpace = entropySpace + blockStateSpace + tokenSpace ++ matchStateSize + ldmSpace + ldmSeqSpace;++ DEBUGLOG(5, "sizeof(ZSTD_CCtx) : %u", (U32)sizeof(ZSTD_CCtx));+ DEBUGLOG(5, "estimate workSpace : %u", (U32)neededSpace);+ return sizeof(ZSTD_CCtx) + neededSpace;+ }+}++size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams)+{+ ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams);+ return ZSTD_estimateCCtxSize_usingCCtxParams(¶ms);+}++static size_t ZSTD_estimateCCtxSize_internal(int compressionLevel)+{+ ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, 0);+ return ZSTD_estimateCCtxSize_usingCParams(cParams);+}++size_t ZSTD_estimateCCtxSize(int compressionLevel)+{+ int level;+ size_t memBudget = 0;+ for (level=MIN(compressionLevel, 1); level<=compressionLevel; level++) {+ size_t const newMB = ZSTD_estimateCCtxSize_internal(level);+ if (newMB > memBudget) memBudget = newMB;+ }+ return memBudget;+}++size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)+{+ RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");+ { ZSTD_compressionParameters const cParams =+ ZSTD_getCParamsFromCCtxParams(params, 0, 0);+ size_t const CCtxSize = ZSTD_estimateCCtxSize_usingCCtxParams(params);+ size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);+ size_t const inBuffSize = ((size_t)1 << cParams.windowLog) + blockSize;+ size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1;+ size_t const streamingSize = inBuffSize + outBuffSize;++ return CCtxSize + streamingSize;+ }+}++size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams)+{+ ZSTD_CCtx_params const params = ZSTD_makeCCtxParamsFromCParams(cParams);+ return ZSTD_estimateCStreamSize_usingCCtxParams(¶ms);+}++static size_t ZSTD_estimateCStreamSize_internal(int compressionLevel)+{+ ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, 0);+ return ZSTD_estimateCStreamSize_usingCParams(cParams);+}++size_t ZSTD_estimateCStreamSize(int compressionLevel)+{+ int level;+ size_t memBudget = 0;+ for (level=MIN(compressionLevel, 1); level<=compressionLevel; level++) {+ size_t const newMB = ZSTD_estimateCStreamSize_internal(level);+ if (newMB > memBudget) memBudget = newMB;+ }+ return memBudget;+}++/* ZSTD_getFrameProgression():+ * tells how much data has been consumed (input) and produced (output) for current frame.+ * able to count progression inside worker threads (non-blocking mode).+ */+ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx)+{+#ifdef ZSTD_MULTITHREAD+ if (cctx->appliedParams.nbWorkers > 0) {+ return ZSTDMT_getFrameProgression(cctx->mtctx);+ }+#endif+ { ZSTD_frameProgression fp;+ size_t const buffered = (cctx->inBuff == NULL) ? 0 :+ cctx->inBuffPos - cctx->inToCompress;+ if (buffered) assert(cctx->inBuffPos >= cctx->inToCompress);+ assert(buffered <= ZSTD_BLOCKSIZE_MAX);+ fp.ingested = cctx->consumedSrcSize + buffered;+ fp.consumed = cctx->consumedSrcSize;+ fp.produced = cctx->producedCSize;+ fp.flushed = cctx->producedCSize; /* simplified; some data might still be left within streaming output buffer */+ fp.currentJobID = 0;+ fp.nbActiveWorkers = 0;+ return fp;+} }++/*! ZSTD_toFlushNow()+ * Only useful for multithreading scenarios currently (nbWorkers >= 1).+ */+size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx)+{+#ifdef ZSTD_MULTITHREAD+ if (cctx->appliedParams.nbWorkers > 0) {+ return ZSTDMT_toFlushNow(cctx->mtctx);+ }+#endif+ (void)cctx;+ return 0; /* over-simplification; could also check if context is currently running in streaming mode, and in which case, report how many bytes are left to be flushed within output buffer */+}++++static U32 ZSTD_equivalentCParams(ZSTD_compressionParameters cParams1,+ ZSTD_compressionParameters cParams2)+{+ return (cParams1.hashLog == cParams2.hashLog)+ & (cParams1.chainLog == cParams2.chainLog)+ & (cParams1.strategy == cParams2.strategy) /* opt parser space */+ & ((cParams1.minMatch==3) == (cParams2.minMatch==3)); /* hashlog3 space */+}++static void ZSTD_assertEqualCParams(ZSTD_compressionParameters cParams1,+ ZSTD_compressionParameters cParams2)+{+ (void)cParams1;+ (void)cParams2;+ assert(cParams1.windowLog == cParams2.windowLog);+ assert(cParams1.chainLog == cParams2.chainLog);+ assert(cParams1.hashLog == cParams2.hashLog);+ assert(cParams1.searchLog == cParams2.searchLog);+ assert(cParams1.minMatch == cParams2.minMatch);+ assert(cParams1.targetLength == cParams2.targetLength);+ assert(cParams1.strategy == cParams2.strategy);+}++/** The parameters are equivalent if ldm is not enabled in both sets or+ * all the parameters are equivalent. */+static U32 ZSTD_equivalentLdmParams(ldmParams_t ldmParams1,+ ldmParams_t ldmParams2)+{+ return (!ldmParams1.enableLdm && !ldmParams2.enableLdm) ||+ (ldmParams1.enableLdm == ldmParams2.enableLdm &&+ ldmParams1.hashLog == ldmParams2.hashLog &&+ ldmParams1.bucketSizeLog == ldmParams2.bucketSizeLog &&+ ldmParams1.minMatchLength == ldmParams2.minMatchLength &&+ ldmParams1.hashRateLog == ldmParams2.hashRateLog);+}++typedef enum { ZSTDb_not_buffered, ZSTDb_buffered } ZSTD_buffered_policy_e;++/* ZSTD_sufficientBuff() :+ * check internal buffers exist for streaming if buffPol == ZSTDb_buffered .+ * Note : they are assumed to be correctly sized if ZSTD_equivalentCParams()==1 */+static U32 ZSTD_sufficientBuff(size_t bufferSize1, size_t maxNbSeq1,+ size_t maxNbLit1,+ ZSTD_buffered_policy_e buffPol2,+ ZSTD_compressionParameters cParams2,+ U64 pledgedSrcSize)+{+ size_t const windowSize2 = MAX(1, (size_t)MIN(((U64)1 << cParams2.windowLog), pledgedSrcSize));+ size_t const blockSize2 = MIN(ZSTD_BLOCKSIZE_MAX, windowSize2);+ size_t const maxNbSeq2 = blockSize2 / ((cParams2.minMatch == 3) ? 3 : 4);+ size_t const maxNbLit2 = blockSize2;+ size_t const neededBufferSize2 = (buffPol2==ZSTDb_buffered) ? windowSize2 + blockSize2 : 0;+ DEBUGLOG(4, "ZSTD_sufficientBuff: is neededBufferSize2=%u <= bufferSize1=%u",+ (U32)neededBufferSize2, (U32)bufferSize1);+ DEBUGLOG(4, "ZSTD_sufficientBuff: is maxNbSeq2=%u <= maxNbSeq1=%u",+ (U32)maxNbSeq2, (U32)maxNbSeq1);+ DEBUGLOG(4, "ZSTD_sufficientBuff: is maxNbLit2=%u <= maxNbLit1=%u",+ (U32)maxNbLit2, (U32)maxNbLit1);+ return (maxNbLit2 <= maxNbLit1)+ & (maxNbSeq2 <= maxNbSeq1)+ & (neededBufferSize2 <= bufferSize1);+}++/** Equivalence for resetCCtx purposes */+static U32 ZSTD_equivalentParams(ZSTD_CCtx_params params1,+ ZSTD_CCtx_params params2,+ size_t buffSize1,+ size_t maxNbSeq1, size_t maxNbLit1,+ ZSTD_buffered_policy_e buffPol2,+ U64 pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_equivalentParams: pledgedSrcSize=%u", (U32)pledgedSrcSize);+ if (!ZSTD_equivalentCParams(params1.cParams, params2.cParams)) {+ DEBUGLOG(4, "ZSTD_equivalentCParams() == 0");+ return 0;+ }+ if (!ZSTD_equivalentLdmParams(params1.ldmParams, params2.ldmParams)) {+ DEBUGLOG(4, "ZSTD_equivalentLdmParams() == 0");+ return 0;+ }+ if (!ZSTD_sufficientBuff(buffSize1, maxNbSeq1, maxNbLit1, buffPol2,+ params2.cParams, pledgedSrcSize)) {+ DEBUGLOG(4, "ZSTD_sufficientBuff() == 0");+ return 0;+ }+ return 1;+}++static void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs)+{+ int i;+ for (i = 0; i < ZSTD_REP_NUM; ++i)+ bs->rep[i] = repStartValue[i];+ bs->entropy.huf.repeatMode = HUF_repeat_none;+ bs->entropy.fse.offcode_repeatMode = FSE_repeat_none;+ bs->entropy.fse.matchlength_repeatMode = FSE_repeat_none;+ bs->entropy.fse.litlength_repeatMode = FSE_repeat_none;+}++/*! ZSTD_invalidateMatchState()+ * Invalidate all the matches in the match finder tables.+ * Requires nextSrc and base to be set (can be NULL).+ */+static void ZSTD_invalidateMatchState(ZSTD_matchState_t* ms)+{+ ZSTD_window_clear(&ms->window);++ ms->nextToUpdate = ms->window.dictLimit;+ ms->nextToUpdate3 = ms->window.dictLimit;+ ms->loadedDictEnd = 0;+ ms->opt.litLengthSum = 0; /* force reset of btopt stats */+ ms->dictMatchState = NULL;+}++/*! ZSTD_continueCCtx() :+ * reuse CCtx without reset (note : requires no dictionary) */+static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_CCtx_params params, U64 pledgedSrcSize)+{+ size_t const windowSize = MAX(1, (size_t)MIN(((U64)1 << params.cParams.windowLog), pledgedSrcSize));+ size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, windowSize);+ DEBUGLOG(4, "ZSTD_continueCCtx: re-use context in place");++ cctx->blockSize = blockSize; /* previous block size could be different even for same windowLog, due to pledgedSrcSize */+ cctx->appliedParams = params;+ cctx->blockState.matchState.cParams = params.cParams;+ cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;+ cctx->consumedSrcSize = 0;+ cctx->producedCSize = 0;+ if (pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN)+ cctx->appliedParams.fParams.contentSizeFlag = 0;+ DEBUGLOG(4, "pledged content size : %u ; flag : %u",+ (U32)pledgedSrcSize, cctx->appliedParams.fParams.contentSizeFlag);+ cctx->stage = ZSTDcs_init;+ cctx->dictID = 0;+ if (params.ldmParams.enableLdm)+ ZSTD_window_clear(&cctx->ldmState.window);+ ZSTD_referenceExternalSequences(cctx, NULL, 0);+ ZSTD_invalidateMatchState(&cctx->blockState.matchState);+ ZSTD_reset_compressedBlockState(cctx->blockState.prevCBlock);+ XXH64_reset(&cctx->xxhState, 0);+ return 0;+}++typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset } ZSTD_compResetPolicy_e;++static void*+ZSTD_reset_matchState(ZSTD_matchState_t* ms,+ void* ptr,+ const ZSTD_compressionParameters* cParams,+ ZSTD_compResetPolicy_e const crp, U32 const forCCtx)+{+ size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog);+ size_t const hSize = ((size_t)1) << cParams->hashLog;+ U32 const hashLog3 = (forCCtx && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;+ size_t const h3Size = ((size_t)1) << hashLog3;+ size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);++ assert(((size_t)ptr & 3) == 0);++ ms->hashLog3 = hashLog3;+ memset(&ms->window, 0, sizeof(ms->window));+ ms->window.dictLimit = 1; /* start from 1, so that 1st position is valid */+ ms->window.lowLimit = 1; /* it ensures first and later CCtx usages compress the same */+ ms->window.nextSrc = ms->window.base + 1; /* see issue #1241 */+ ZSTD_invalidateMatchState(ms);++ /* opt parser space */+ if (forCCtx && (cParams->strategy >= ZSTD_btopt)) {+ DEBUGLOG(4, "reserving optimal parser space");+ ms->opt.litFreq = (unsigned*)ptr;+ ms->opt.litLengthFreq = ms->opt.litFreq + (1<<Litbits);+ ms->opt.matchLengthFreq = ms->opt.litLengthFreq + (MaxLL+1);+ ms->opt.offCodeFreq = ms->opt.matchLengthFreq + (MaxML+1);+ ptr = ms->opt.offCodeFreq + (MaxOff+1);+ ms->opt.matchTable = (ZSTD_match_t*)ptr;+ ptr = ms->opt.matchTable + ZSTD_OPT_NUM+1;+ ms->opt.priceTable = (ZSTD_optimal_t*)ptr;+ ptr = ms->opt.priceTable + ZSTD_OPT_NUM+1;+ }++ /* table Space */+ DEBUGLOG(4, "reset table : %u", crp!=ZSTDcrp_noMemset);+ assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */+ if (crp!=ZSTDcrp_noMemset) memset(ptr, 0, tableSpace); /* reset tables only */+ ms->hashTable = (U32*)(ptr);+ ms->chainTable = ms->hashTable + hSize;+ ms->hashTable3 = ms->chainTable + chainSize;+ ptr = ms->hashTable3 + h3Size;++ ms->cParams = *cParams;++ assert(((size_t)ptr & 3) == 0);+ return ptr;+}++#define ZSTD_WORKSPACETOOLARGE_FACTOR 3 /* define "workspace is too large" as this number of times larger than needed */+#define ZSTD_WORKSPACETOOLARGE_MAXDURATION 128 /* when workspace is continuously too large+ * during at least this number of times,+ * context's memory usage is considered wasteful,+ * because it's sized to handle a worst case scenario which rarely happens.+ * In which case, resize it down to free some memory */++/*! ZSTD_resetCCtx_internal() :+ note : `params` are assumed fully validated at this stage */+static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,+ ZSTD_CCtx_params params,+ U64 pledgedSrcSize,+ ZSTD_compResetPolicy_e const crp,+ ZSTD_buffered_policy_e const zbuff)+{+ DEBUGLOG(4, "ZSTD_resetCCtx_internal: pledgedSrcSize=%u, wlog=%u",+ (U32)pledgedSrcSize, params.cParams.windowLog);+ assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));++ if (crp == ZSTDcrp_continue) {+ if (ZSTD_equivalentParams(zc->appliedParams, params,+ zc->inBuffSize,+ zc->seqStore.maxNbSeq, zc->seqStore.maxNbLit,+ zbuff, pledgedSrcSize)) {+ DEBUGLOG(4, "ZSTD_equivalentParams()==1 -> continue mode (wLog1=%u, blockSize1=%zu)",+ zc->appliedParams.cParams.windowLog, zc->blockSize);+ zc->workSpaceOversizedDuration += (zc->workSpaceOversizedDuration > 0); /* if it was too large, it still is */+ if (zc->workSpaceOversizedDuration <= ZSTD_WORKSPACETOOLARGE_MAXDURATION)+ return ZSTD_continueCCtx(zc, params, pledgedSrcSize);+ } }+ DEBUGLOG(4, "ZSTD_equivalentParams()==0 -> reset CCtx");++ if (params.ldmParams.enableLdm) {+ /* Adjust long distance matching parameters */+ ZSTD_ldm_adjustParameters(¶ms.ldmParams, ¶ms.cParams);+ assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog);+ assert(params.ldmParams.hashRateLog < 32);+ zc->ldmState.hashPower = ZSTD_rollingHash_primePower(params.ldmParams.minMatchLength);+ }++ { size_t const windowSize = MAX(1, (size_t)MIN(((U64)1 << params.cParams.windowLog), pledgedSrcSize));+ size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, windowSize);+ U32 const divider = (params.cParams.minMatch==3) ? 3 : 4;+ size_t const maxNbSeq = blockSize / divider;+ size_t const tokenSpace = WILDCOPY_OVERLENGTH + blockSize + 11*maxNbSeq;+ size_t const buffOutSize = (zbuff==ZSTDb_buffered) ? ZSTD_compressBound(blockSize)+1 : 0;+ size_t const buffInSize = (zbuff==ZSTDb_buffered) ? windowSize + blockSize : 0;+ size_t const matchStateSize = ZSTD_sizeof_matchState(¶ms.cParams, /* forCCtx */ 1);+ size_t const maxNbLdmSeq = ZSTD_ldm_getMaxNbSeq(params.ldmParams, blockSize);+ void* ptr; /* used to partition workSpace */++ /* Check if workSpace is large enough, alloc a new one if needed */+ { size_t const entropySpace = HUF_WORKSPACE_SIZE;+ size_t const blockStateSpace = 2 * sizeof(ZSTD_compressedBlockState_t);+ size_t const bufferSpace = buffInSize + buffOutSize;+ size_t const ldmSpace = ZSTD_ldm_getTableSize(params.ldmParams);+ size_t const ldmSeqSpace = maxNbLdmSeq * sizeof(rawSeq);++ size_t const neededSpace = entropySpace + blockStateSpace + ldmSpace ++ ldmSeqSpace + matchStateSize + tokenSpace ++ bufferSpace;++ int const workSpaceTooSmall = zc->workSpaceSize < neededSpace;+ int const workSpaceTooLarge = zc->workSpaceSize > ZSTD_WORKSPACETOOLARGE_FACTOR * neededSpace;+ int const workSpaceWasteful = workSpaceTooLarge && (zc->workSpaceOversizedDuration > ZSTD_WORKSPACETOOLARGE_MAXDURATION);+ zc->workSpaceOversizedDuration = workSpaceTooLarge ? zc->workSpaceOversizedDuration+1 : 0;++ DEBUGLOG(4, "Need %zuKB workspace, including %zuKB for match state, and %zuKB for buffers",+ neededSpace>>10, matchStateSize>>10, bufferSpace>>10);+ DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize);++ if (workSpaceTooSmall || workSpaceWasteful) {+ DEBUGLOG(4, "Need to resize workSpaceSize from %zuKB to %zuKB",+ zc->workSpaceSize >> 10,+ neededSpace >> 10);++ RETURN_ERROR_IF(zc->staticSize, memory_allocation, "static cctx : no resize");++ zc->workSpaceSize = 0;+ ZSTD_free(zc->workSpace, zc->customMem);+ zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem);+ RETURN_ERROR_IF(zc->workSpace == NULL, memory_allocation);+ zc->workSpaceSize = neededSpace;+ zc->workSpaceOversizedDuration = 0;++ /* Statically sized space.+ * entropyWorkspace never moves,+ * though prev/next block swap places */+ assert(((size_t)zc->workSpace & 3) == 0); /* ensure correct alignment */+ assert(zc->workSpaceSize >= 2 * sizeof(ZSTD_compressedBlockState_t));+ zc->blockState.prevCBlock = (ZSTD_compressedBlockState_t*)zc->workSpace;+ zc->blockState.nextCBlock = zc->blockState.prevCBlock + 1;+ ptr = zc->blockState.nextCBlock + 1;+ zc->entropyWorkspace = (U32*)ptr;+ } }++ /* init params */+ zc->appliedParams = params;+ zc->blockState.matchState.cParams = params.cParams;+ zc->pledgedSrcSizePlusOne = pledgedSrcSize+1;+ zc->consumedSrcSize = 0;+ zc->producedCSize = 0;+ if (pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN)+ zc->appliedParams.fParams.contentSizeFlag = 0;+ DEBUGLOG(4, "pledged content size : %u ; flag : %u",+ (unsigned)pledgedSrcSize, zc->appliedParams.fParams.contentSizeFlag);+ zc->blockSize = blockSize;++ XXH64_reset(&zc->xxhState, 0);+ zc->stage = ZSTDcs_init;+ zc->dictID = 0;++ ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock);++ ptr = zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32;++ /* ldm hash table */+ /* initialize bucketOffsets table later for pointer alignment */+ if (params.ldmParams.enableLdm) {+ size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog;+ memset(ptr, 0, ldmHSize * sizeof(ldmEntry_t));+ assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */+ zc->ldmState.hashTable = (ldmEntry_t*)ptr;+ ptr = zc->ldmState.hashTable + ldmHSize;+ zc->ldmSequences = (rawSeq*)ptr;+ ptr = zc->ldmSequences + maxNbLdmSeq;+ zc->maxNbLdmSequences = maxNbLdmSeq;++ memset(&zc->ldmState.window, 0, sizeof(zc->ldmState.window));+ }+ assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */++ ptr = ZSTD_reset_matchState(&zc->blockState.matchState, ptr, ¶ms.cParams, crp, /* forCCtx */ 1);++ /* sequences storage */+ zc->seqStore.maxNbSeq = maxNbSeq;+ zc->seqStore.sequencesStart = (seqDef*)ptr;+ ptr = zc->seqStore.sequencesStart + maxNbSeq;+ zc->seqStore.llCode = (BYTE*) ptr;+ zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq;+ zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq;+ zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq;+ /* ZSTD_wildcopy() is used to copy into the literals buffer,+ * so we have to oversize the buffer by WILDCOPY_OVERLENGTH bytes.+ */+ zc->seqStore.maxNbLit = blockSize;+ ptr = zc->seqStore.litStart + blockSize + WILDCOPY_OVERLENGTH;++ /* ldm bucketOffsets table */+ if (params.ldmParams.enableLdm) {+ size_t const ldmBucketSize =+ ((size_t)1) << (params.ldmParams.hashLog -+ params.ldmParams.bucketSizeLog);+ memset(ptr, 0, ldmBucketSize);+ zc->ldmState.bucketOffsets = (BYTE*)ptr;+ ptr = zc->ldmState.bucketOffsets + ldmBucketSize;+ ZSTD_window_clear(&zc->ldmState.window);+ }+ ZSTD_referenceExternalSequences(zc, NULL, 0);++ /* buffers */+ zc->inBuffSize = buffInSize;+ zc->inBuff = (char*)ptr;+ zc->outBuffSize = buffOutSize;+ zc->outBuff = zc->inBuff + buffInSize;++ return 0;+ }+}++/* ZSTD_invalidateRepCodes() :+ * ensures next compression will not use repcodes from previous block.+ * Note : only works with regular variant;+ * do not use with extDict variant ! */+void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx) {+ int i;+ for (i=0; i<ZSTD_REP_NUM; i++) cctx->blockState.prevCBlock->rep[i] = 0;+ assert(!ZSTD_window_hasExtDict(cctx->blockState.matchState.window));+}++/* These are the approximate sizes for each strategy past which copying the+ * dictionary tables into the working context is faster than using them+ * in-place.+ */+static const size_t attachDictSizeCutoffs[ZSTD_STRATEGY_MAX+1] = {+ 8 KB, /* unused */+ 8 KB, /* ZSTD_fast */+ 16 KB, /* ZSTD_dfast */+ 32 KB, /* ZSTD_greedy */+ 32 KB, /* ZSTD_lazy */+ 32 KB, /* ZSTD_lazy2 */+ 32 KB, /* ZSTD_btlazy2 */+ 32 KB, /* ZSTD_btopt */+ 8 KB, /* ZSTD_btultra */+ 8 KB /* ZSTD_btultra2 */+};++static int ZSTD_shouldAttachDict(const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params,+ U64 pledgedSrcSize)+{+ size_t cutoff = attachDictSizeCutoffs[cdict->matchState.cParams.strategy];+ return ( pledgedSrcSize <= cutoff+ || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN+ || params.attachDictPref == ZSTD_dictForceAttach )+ && params.attachDictPref != ZSTD_dictForceCopy+ && !params.forceWindow; /* dictMatchState isn't correctly+ * handled in _enforceMaxDist */+}++static size_t ZSTD_resetCCtx_byAttachingCDict(+ ZSTD_CCtx* cctx,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params,+ U64 pledgedSrcSize,+ ZSTD_buffered_policy_e zbuff)+{+ {+ const ZSTD_compressionParameters *cdict_cParams = &cdict->matchState.cParams;+ unsigned const windowLog = params.cParams.windowLog;+ assert(windowLog != 0);+ /* Resize working context table params for input only, since the dict+ * has its own tables. */+ params.cParams = ZSTD_adjustCParams_internal(*cdict_cParams, pledgedSrcSize, 0);+ params.cParams.windowLog = windowLog;+ ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,+ ZSTDcrp_continue, zbuff);+ assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy);+ }++ {+ const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc+ - cdict->matchState.window.base);+ const U32 cdictLen = cdictEnd - cdict->matchState.window.dictLimit;+ if (cdictLen == 0) {+ /* don't even attach dictionaries with no contents */+ DEBUGLOG(4, "skipping attaching empty dictionary");+ } else {+ DEBUGLOG(4, "attaching dictionary into context");+ cctx->blockState.matchState.dictMatchState = &cdict->matchState;++ /* prep working match state so dict matches never have negative indices+ * when they are translated to the working context's index space. */+ if (cctx->blockState.matchState.window.dictLimit < cdictEnd) {+ cctx->blockState.matchState.window.nextSrc =+ cctx->blockState.matchState.window.base + cdictEnd;+ ZSTD_window_clear(&cctx->blockState.matchState.window);+ }+ cctx->blockState.matchState.loadedDictEnd = cctx->blockState.matchState.window.dictLimit;+ }+ }++ cctx->dictID = cdict->dictID;++ /* copy block state */+ memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState));++ return 0;+}++static size_t ZSTD_resetCCtx_byCopyingCDict(ZSTD_CCtx* cctx,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params,+ U64 pledgedSrcSize,+ ZSTD_buffered_policy_e zbuff)+{+ const ZSTD_compressionParameters *cdict_cParams = &cdict->matchState.cParams;++ DEBUGLOG(4, "copying dictionary into context");++ { unsigned const windowLog = params.cParams.windowLog;+ assert(windowLog != 0);+ /* Copy only compression parameters related to tables. */+ params.cParams = *cdict_cParams;+ params.cParams.windowLog = windowLog;+ ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,+ ZSTDcrp_noMemset, zbuff);+ assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy);+ assert(cctx->appliedParams.cParams.hashLog == cdict_cParams->hashLog);+ assert(cctx->appliedParams.cParams.chainLog == cdict_cParams->chainLog);+ }++ /* copy tables */+ { size_t const chainSize = (cdict_cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cdict_cParams->chainLog);+ size_t const hSize = (size_t)1 << cdict_cParams->hashLog;+ size_t const tableSpace = (chainSize + hSize) * sizeof(U32);+ assert((U32*)cctx->blockState.matchState.chainTable == (U32*)cctx->blockState.matchState.hashTable + hSize); /* chainTable must follow hashTable */+ assert((U32*)cctx->blockState.matchState.hashTable3 == (U32*)cctx->blockState.matchState.chainTable + chainSize);+ assert((U32*)cdict->matchState.chainTable == (U32*)cdict->matchState.hashTable + hSize); /* chainTable must follow hashTable */+ assert((U32*)cdict->matchState.hashTable3 == (U32*)cdict->matchState.chainTable + chainSize);+ memcpy(cctx->blockState.matchState.hashTable, cdict->matchState.hashTable, tableSpace); /* presumes all tables follow each other */+ }++ /* Zero the hashTable3, since the cdict never fills it */+ { size_t const h3Size = (size_t)1 << cctx->blockState.matchState.hashLog3;+ assert(cdict->matchState.hashLog3 == 0);+ memset(cctx->blockState.matchState.hashTable3, 0, h3Size * sizeof(U32));+ }++ /* copy dictionary offsets */+ { ZSTD_matchState_t const* srcMatchState = &cdict->matchState;+ ZSTD_matchState_t* dstMatchState = &cctx->blockState.matchState;+ dstMatchState->window = srcMatchState->window;+ dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;+ dstMatchState->nextToUpdate3= srcMatchState->nextToUpdate3;+ dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;+ }++ cctx->dictID = cdict->dictID;++ /* copy block state */+ memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState));++ return 0;+}++/* We have a choice between copying the dictionary context into the working+ * context, or referencing the dictionary context from the working context+ * in-place. We decide here which strategy to use. */+static size_t ZSTD_resetCCtx_usingCDict(ZSTD_CCtx* cctx,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params,+ U64 pledgedSrcSize,+ ZSTD_buffered_policy_e zbuff)+{++ DEBUGLOG(4, "ZSTD_resetCCtx_usingCDict (pledgedSrcSize=%u)",+ (unsigned)pledgedSrcSize);++ if (ZSTD_shouldAttachDict(cdict, params, pledgedSrcSize)) {+ return ZSTD_resetCCtx_byAttachingCDict(+ cctx, cdict, params, pledgedSrcSize, zbuff);+ } else {+ return ZSTD_resetCCtx_byCopyingCDict(+ cctx, cdict, params, pledgedSrcSize, zbuff);+ }+}++/*! ZSTD_copyCCtx_internal() :+ * Duplicate an existing context `srcCCtx` into another one `dstCCtx`.+ * Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).+ * The "context", in this case, refers to the hash and chain tables,+ * entropy tables, and dictionary references.+ * `windowLog` value is enforced if != 0, otherwise value is copied from srcCCtx.+ * @return : 0, or an error code */+static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx,+ const ZSTD_CCtx* srcCCtx,+ ZSTD_frameParameters fParams,+ U64 pledgedSrcSize,+ ZSTD_buffered_policy_e zbuff)+{+ DEBUGLOG(5, "ZSTD_copyCCtx_internal");+ RETURN_ERROR_IF(srcCCtx->stage!=ZSTDcs_init, stage_wrong);++ memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));+ { ZSTD_CCtx_params params = dstCCtx->requestedParams;+ /* Copy only compression parameters related to tables. */+ params.cParams = srcCCtx->appliedParams.cParams;+ params.fParams = fParams;+ ZSTD_resetCCtx_internal(dstCCtx, params, pledgedSrcSize,+ ZSTDcrp_noMemset, zbuff);+ assert(dstCCtx->appliedParams.cParams.windowLog == srcCCtx->appliedParams.cParams.windowLog);+ assert(dstCCtx->appliedParams.cParams.strategy == srcCCtx->appliedParams.cParams.strategy);+ assert(dstCCtx->appliedParams.cParams.hashLog == srcCCtx->appliedParams.cParams.hashLog);+ assert(dstCCtx->appliedParams.cParams.chainLog == srcCCtx->appliedParams.cParams.chainLog);+ assert(dstCCtx->blockState.matchState.hashLog3 == srcCCtx->blockState.matchState.hashLog3);+ }++ /* copy tables */+ { size_t const chainSize = (srcCCtx->appliedParams.cParams.strategy == ZSTD_fast) ? 0 : ((size_t)1 << srcCCtx->appliedParams.cParams.chainLog);+ size_t const hSize = (size_t)1 << srcCCtx->appliedParams.cParams.hashLog;+ size_t const h3Size = (size_t)1 << srcCCtx->blockState.matchState.hashLog3;+ size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);+ assert((U32*)dstCCtx->blockState.matchState.chainTable == (U32*)dstCCtx->blockState.matchState.hashTable + hSize); /* chainTable must follow hashTable */+ assert((U32*)dstCCtx->blockState.matchState.hashTable3 == (U32*)dstCCtx->blockState.matchState.chainTable + chainSize);+ memcpy(dstCCtx->blockState.matchState.hashTable, srcCCtx->blockState.matchState.hashTable, tableSpace); /* presumes all tables follow each other */+ }++ /* copy dictionary offsets */+ {+ const ZSTD_matchState_t* srcMatchState = &srcCCtx->blockState.matchState;+ ZSTD_matchState_t* dstMatchState = &dstCCtx->blockState.matchState;+ dstMatchState->window = srcMatchState->window;+ dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;+ dstMatchState->nextToUpdate3= srcMatchState->nextToUpdate3;+ dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;+ }+ dstCCtx->dictID = srcCCtx->dictID;++ /* copy block state */+ memcpy(dstCCtx->blockState.prevCBlock, srcCCtx->blockState.prevCBlock, sizeof(*srcCCtx->blockState.prevCBlock));++ return 0;+}++/*! ZSTD_copyCCtx() :+ * Duplicate an existing context `srcCCtx` into another one `dstCCtx`.+ * Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).+ * pledgedSrcSize==0 means "unknown".+* @return : 0, or an error code */+size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, unsigned long long pledgedSrcSize)+{+ ZSTD_frameParameters fParams = { 1 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };+ ZSTD_buffered_policy_e const zbuff = (ZSTD_buffered_policy_e)(srcCCtx->inBuffSize>0);+ ZSTD_STATIC_ASSERT((U32)ZSTDb_buffered==1);+ if (pledgedSrcSize==0) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN;+ fParams.contentSizeFlag = (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN);++ return ZSTD_copyCCtx_internal(dstCCtx, srcCCtx,+ fParams, pledgedSrcSize,+ zbuff);+}+++#define ZSTD_ROWSIZE 16+/*! ZSTD_reduceTable() :+ * reduce table indexes by `reducerValue`, or squash to zero.+ * PreserveMark preserves "unsorted mark" for btlazy2 strategy.+ * It must be set to a clear 0/1 value, to remove branch during inlining.+ * Presume table size is a multiple of ZSTD_ROWSIZE+ * to help auto-vectorization */+FORCE_INLINE_TEMPLATE void+ZSTD_reduceTable_internal (U32* const table, U32 const size, U32 const reducerValue, int const preserveMark)+{+ int const nbRows = (int)size / ZSTD_ROWSIZE;+ int cellNb = 0;+ int rowNb;+ assert((size & (ZSTD_ROWSIZE-1)) == 0); /* multiple of ZSTD_ROWSIZE */+ assert(size < (1U<<31)); /* can be casted to int */+ for (rowNb=0 ; rowNb < nbRows ; rowNb++) {+ int column;+ for (column=0; column<ZSTD_ROWSIZE; column++) {+ if (preserveMark) {+ U32 const adder = (table[cellNb] == ZSTD_DUBT_UNSORTED_MARK) ? reducerValue : 0;+ table[cellNb] += adder;+ }+ if (table[cellNb] < reducerValue) table[cellNb] = 0;+ else table[cellNb] -= reducerValue;+ cellNb++;+ } }+}++static void ZSTD_reduceTable(U32* const table, U32 const size, U32 const reducerValue)+{+ ZSTD_reduceTable_internal(table, size, reducerValue, 0);+}++static void ZSTD_reduceTable_btlazy2(U32* const table, U32 const size, U32 const reducerValue)+{+ ZSTD_reduceTable_internal(table, size, reducerValue, 1);+}++/*! ZSTD_reduceIndex() :+* rescale all indexes to avoid future overflow (indexes are U32) */+static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue)+{+ ZSTD_matchState_t* const ms = &zc->blockState.matchState;+ { U32 const hSize = (U32)1 << zc->appliedParams.cParams.hashLog;+ ZSTD_reduceTable(ms->hashTable, hSize, reducerValue);+ }++ if (zc->appliedParams.cParams.strategy != ZSTD_fast) {+ U32 const chainSize = (U32)1 << zc->appliedParams.cParams.chainLog;+ if (zc->appliedParams.cParams.strategy == ZSTD_btlazy2)+ ZSTD_reduceTable_btlazy2(ms->chainTable, chainSize, reducerValue);+ else+ ZSTD_reduceTable(ms->chainTable, chainSize, reducerValue);+ }++ if (ms->hashLog3) {+ U32 const h3Size = (U32)1 << ms->hashLog3;+ ZSTD_reduceTable(ms->hashTable3, h3Size, reducerValue);+ }+}+++/*-*******************************************************+* Block entropic compression+*********************************************************/++/* See doc/zstd_compression_format.md for detailed format description */++static size_t ZSTD_noCompressBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize, U32 lastBlock)+{+ U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(srcSize << 3);+ RETURN_ERROR_IF(srcSize + ZSTD_blockHeaderSize > dstCapacity,+ dstSize_tooSmall);+ MEM_writeLE24(dst, cBlockHeader24);+ memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize);+ return ZSTD_blockHeaderSize + srcSize;+}++static size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+ BYTE* const ostart = (BYTE* const)dst;+ U32 const flSize = 1 + (srcSize>31) + (srcSize>4095);++ RETURN_ERROR_IF(srcSize + flSize > dstCapacity, dstSize_tooSmall);++ switch(flSize)+ {+ case 1: /* 2 - 1 - 5 */+ ostart[0] = (BYTE)((U32)set_basic + (srcSize<<3));+ break;+ case 2: /* 2 - 2 - 12 */+ MEM_writeLE16(ostart, (U16)((U32)set_basic + (1<<2) + (srcSize<<4)));+ break;+ case 3: /* 2 - 2 - 20 */+ MEM_writeLE32(ostart, (U32)((U32)set_basic + (3<<2) + (srcSize<<4)));+ break;+ default: /* not necessary : flSize is {1,2,3} */+ assert(0);+ }++ memcpy(ostart + flSize, src, srcSize);+ return srcSize + flSize;+}++static size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+ BYTE* const ostart = (BYTE* const)dst;+ U32 const flSize = 1 + (srcSize>31) + (srcSize>4095);++ (void)dstCapacity; /* dstCapacity already guaranteed to be >=4, hence large enough */++ switch(flSize)+ {+ case 1: /* 2 - 1 - 5 */+ ostart[0] = (BYTE)((U32)set_rle + (srcSize<<3));+ break;+ case 2: /* 2 - 2 - 12 */+ MEM_writeLE16(ostart, (U16)((U32)set_rle + (1<<2) + (srcSize<<4)));+ break;+ case 3: /* 2 - 2 - 20 */+ MEM_writeLE32(ostart, (U32)((U32)set_rle + (3<<2) + (srcSize<<4)));+ break;+ default: /* not necessary : flSize is {1,2,3} */+ assert(0);+ }++ ostart[flSize] = *(const BYTE*)src;+ return flSize+1;+}+++/* ZSTD_minGain() :+ * minimum compression required+ * to generate a compress block or a compressed literals section.+ * note : use same formula for both situations */+static size_t ZSTD_minGain(size_t srcSize, ZSTD_strategy strat)+{+ U32 const minlog = (strat>=ZSTD_btultra) ? (U32)(strat) - 1 : 6;+ ZSTD_STATIC_ASSERT(ZSTD_btultra == 8);+ assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat));+ return (srcSize >> minlog) + 2;+}++static size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,+ ZSTD_hufCTables_t* nextHuf,+ ZSTD_strategy strategy, int disableLiteralCompression,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ void* workspace, size_t wkspSize,+ const int bmi2)+{+ size_t const minGain = ZSTD_minGain(srcSize, strategy);+ size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);+ BYTE* const ostart = (BYTE*)dst;+ U32 singleStream = srcSize < 256;+ symbolEncodingType_e hType = set_compressed;+ size_t cLitSize;++ DEBUGLOG(5,"ZSTD_compressLiterals (disableLiteralCompression=%i)",+ disableLiteralCompression);++ /* Prepare nextEntropy assuming reusing the existing table */+ memcpy(nextHuf, prevHuf, sizeof(*prevHuf));++ if (disableLiteralCompression)+ return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);++ /* small ? don't even attempt compression (speed opt) */+# define COMPRESS_LITERALS_SIZE_MIN 63+ { size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;+ if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);+ }++ RETURN_ERROR_IF(dstCapacity < lhSize+1, dstSize_tooSmall, "not enough space for compression");+ { HUF_repeat repeat = prevHuf->repeatMode;+ int const preferRepeat = strategy < ZSTD_lazy ? srcSize <= 1024 : 0;+ if (repeat == HUF_repeat_valid && lhSize == 3) singleStream = 1;+ cLitSize = singleStream ? HUF_compress1X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11,+ workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2)+ : HUF_compress4X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11,+ workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2);+ if (repeat != HUF_repeat_none) {+ /* reused the existing table */+ hType = set_repeat;+ }+ }++ if ((cLitSize==0) | (cLitSize >= srcSize - minGain) | ERR_isError(cLitSize)) {+ memcpy(nextHuf, prevHuf, sizeof(*prevHuf));+ return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);+ }+ if (cLitSize==1) {+ memcpy(nextHuf, prevHuf, sizeof(*prevHuf));+ return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);+ }++ if (hType == set_compressed) {+ /* using a newly constructed table */+ nextHuf->repeatMode = HUF_repeat_check;+ }++ /* Build header */+ switch(lhSize)+ {+ case 3: /* 2 - 2 - 10 - 10 */+ { U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14);+ MEM_writeLE24(ostart, lhc);+ break;+ }+ case 4: /* 2 - 2 - 14 - 14 */+ { U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18);+ MEM_writeLE32(ostart, lhc);+ break;+ }+ case 5: /* 2 - 2 - 18 - 18 */+ { U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22);+ MEM_writeLE32(ostart, lhc);+ ostart[4] = (BYTE)(cLitSize >> 10);+ break;+ }+ default: /* not possible : lhSize is {3,4,5} */+ assert(0);+ }+ return lhSize+cLitSize;+}+++void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)+{+ const seqDef* const sequences = seqStorePtr->sequencesStart;+ BYTE* const llCodeTable = seqStorePtr->llCode;+ BYTE* const ofCodeTable = seqStorePtr->ofCode;+ BYTE* const mlCodeTable = seqStorePtr->mlCode;+ U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);+ U32 u;+ assert(nbSeq <= seqStorePtr->maxNbSeq);+ for (u=0; u<nbSeq; u++) {+ U32 const llv = sequences[u].litLength;+ U32 const mlv = sequences[u].matchLength;+ llCodeTable[u] = (BYTE)ZSTD_LLcode(llv);+ ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset);+ mlCodeTable[u] = (BYTE)ZSTD_MLcode(mlv);+ }+ if (seqStorePtr->longLengthID==1)+ llCodeTable[seqStorePtr->longLengthPos] = MaxLL;+ if (seqStorePtr->longLengthID==2)+ mlCodeTable[seqStorePtr->longLengthPos] = MaxML;+}+++/**+ * -log2(x / 256) lookup table for x in [0, 256).+ * If x == 0: Return 0+ * Else: Return floor(-log2(x / 256) * 256)+ */+static unsigned const kInverseProbabilityLog256[256] = {+ 0, 2048, 1792, 1642, 1536, 1453, 1386, 1329, 1280, 1236, 1197, 1162,+ 1130, 1100, 1073, 1047, 1024, 1001, 980, 960, 941, 923, 906, 889,+ 874, 859, 844, 830, 817, 804, 791, 779, 768, 756, 745, 734,+ 724, 714, 704, 694, 685, 676, 667, 658, 650, 642, 633, 626,+ 618, 610, 603, 595, 588, 581, 574, 567, 561, 554, 548, 542,+ 535, 529, 523, 517, 512, 506, 500, 495, 489, 484, 478, 473,+ 468, 463, 458, 453, 448, 443, 438, 434, 429, 424, 420, 415,+ 411, 407, 402, 398, 394, 390, 386, 382, 377, 373, 370, 366,+ 362, 358, 354, 350, 347, 343, 339, 336, 332, 329, 325, 322,+ 318, 315, 311, 308, 305, 302, 298, 295, 292, 289, 286, 282,+ 279, 276, 273, 270, 267, 264, 261, 258, 256, 253, 250, 247,+ 244, 241, 239, 236, 233, 230, 228, 225, 222, 220, 217, 215,+ 212, 209, 207, 204, 202, 199, 197, 194, 192, 190, 187, 185,+ 182, 180, 178, 175, 173, 171, 168, 166, 164, 162, 159, 157,+ 155, 153, 151, 149, 146, 144, 142, 140, 138, 136, 134, 132,+ 130, 128, 126, 123, 121, 119, 117, 115, 114, 112, 110, 108,+ 106, 104, 102, 100, 98, 96, 94, 93, 91, 89, 87, 85,+ 83, 82, 80, 78, 76, 74, 73, 71, 69, 67, 66, 64,+ 62, 61, 59, 57, 55, 54, 52, 50, 49, 47, 46, 44,+ 42, 41, 39, 37, 36, 34, 33, 31, 30, 28, 26, 25,+ 23, 22, 20, 19, 17, 16, 14, 13, 11, 10, 8, 7,+ 5, 4, 2, 1,+};+++/**+ * Returns the cost in bits of encoding the distribution described by count+ * using the entropy bound.+ */+static size_t ZSTD_entropyCost(unsigned const* count, unsigned const max, size_t const total)+{+ unsigned cost = 0;+ unsigned s;+ for (s = 0; s <= max; ++s) {+ unsigned norm = (unsigned)((256 * count[s]) / total);+ if (count[s] != 0 && norm == 0)+ norm = 1;+ assert(count[s] < total);+ cost += count[s] * kInverseProbabilityLog256[norm];+ }+ return cost >> 8;+}+++/**+ * Returns the cost in bits of encoding the distribution in count using the+ * table described by norm. The max symbol support by norm is assumed >= max.+ * norm must be valid for every symbol with non-zero probability in count.+ */+static size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog,+ unsigned const* count, unsigned const max)+{+ unsigned const shift = 8 - accuracyLog;+ size_t cost = 0;+ unsigned s;+ assert(accuracyLog <= 8);+ for (s = 0; s <= max; ++s) {+ unsigned const normAcc = norm[s] != -1 ? norm[s] : 1;+ unsigned const norm256 = normAcc << shift;+ assert(norm256 > 0);+ assert(norm256 < 256);+ cost += count[s] * kInverseProbabilityLog256[norm256];+ }+ return cost >> 8;+}+++static unsigned ZSTD_getFSEMaxSymbolValue(FSE_CTable const* ctable) {+ void const* ptr = ctable;+ U16 const* u16ptr = (U16 const*)ptr;+ U32 const maxSymbolValue = MEM_read16(u16ptr + 1);+ return maxSymbolValue;+}+++/**+ * Returns the cost in bits of encoding the distribution in count using ctable.+ * Returns an error if ctable cannot represent all the symbols in count.+ */+static size_t ZSTD_fseBitCost(+ FSE_CTable const* ctable,+ unsigned const* count,+ unsigned const max)+{+ unsigned const kAccuracyLog = 8;+ size_t cost = 0;+ unsigned s;+ FSE_CState_t cstate;+ FSE_initCState(&cstate, ctable);+ RETURN_ERROR_IF(ZSTD_getFSEMaxSymbolValue(ctable) < max, GENERIC,+ "Repeat FSE_CTable has maxSymbolValue %u < %u",+ ZSTD_getFSEMaxSymbolValue(ctable), max);+ for (s = 0; s <= max; ++s) {+ unsigned const tableLog = cstate.stateLog;+ unsigned const badCost = (tableLog + 1) << kAccuracyLog;+ unsigned const bitCost = FSE_bitCost(cstate.symbolTT, tableLog, s, kAccuracyLog);+ if (count[s] == 0)+ continue;+ RETURN_ERROR_IF(bitCost >= badCost, GENERIC,+ "Repeat FSE_CTable has Prob[%u] == 0", s);+ cost += count[s] * bitCost;+ }+ return cost >> kAccuracyLog;+}++/**+ * Returns the cost in bytes of encoding the normalized count header.+ * Returns an error if any of the helper functions return an error.+ */+static size_t ZSTD_NCountCost(unsigned const* count, unsigned const max,+ size_t const nbSeq, unsigned const FSELog)+{+ BYTE wksp[FSE_NCOUNTBOUND];+ S16 norm[MaxSeq + 1];+ const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);+ FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq, max));+ return FSE_writeNCount(wksp, sizeof(wksp), norm, max, tableLog);+}+++typedef enum {+ ZSTD_defaultDisallowed = 0,+ ZSTD_defaultAllowed = 1+} ZSTD_defaultPolicy_e;++MEM_STATIC symbolEncodingType_e+ZSTD_selectEncodingType(+ FSE_repeat* repeatMode, unsigned const* count, unsigned const max,+ size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,+ FSE_CTable const* prevCTable,+ short const* defaultNorm, U32 defaultNormLog,+ ZSTD_defaultPolicy_e const isDefaultAllowed,+ ZSTD_strategy const strategy)+{+ ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0);+ if (mostFrequent == nbSeq) {+ *repeatMode = FSE_repeat_none;+ if (isDefaultAllowed && nbSeq <= 2) {+ /* Prefer set_basic over set_rle when there are 2 or less symbols,+ * since RLE uses 1 byte, but set_basic uses 5-6 bits per symbol.+ * If basic encoding isn't possible, always choose RLE.+ */+ DEBUGLOG(5, "Selected set_basic");+ return set_basic;+ }+ DEBUGLOG(5, "Selected set_rle");+ return set_rle;+ }+ if (strategy < ZSTD_lazy) {+ if (isDefaultAllowed) {+ size_t const staticFse_nbSeq_max = 1000;+ size_t const mult = 10 - strategy;+ size_t const baseLog = 3;+ size_t const dynamicFse_nbSeq_min = (((size_t)1 << defaultNormLog) * mult) >> baseLog; /* 28-36 for offset, 56-72 for lengths */+ assert(defaultNormLog >= 5 && defaultNormLog <= 6); /* xx_DEFAULTNORMLOG */+ assert(mult <= 9 && mult >= 7);+ if ( (*repeatMode == FSE_repeat_valid)+ && (nbSeq < staticFse_nbSeq_max) ) {+ DEBUGLOG(5, "Selected set_repeat");+ return set_repeat;+ }+ if ( (nbSeq < dynamicFse_nbSeq_min)+ || (mostFrequent < (nbSeq >> (defaultNormLog-1))) ) {+ DEBUGLOG(5, "Selected set_basic");+ /* The format allows default tables to be repeated, but it isn't useful.+ * When using simple heuristics to select encoding type, we don't want+ * to confuse these tables with dictionaries. When running more careful+ * analysis, we don't need to waste time checking both repeating tables+ * and default tables.+ */+ *repeatMode = FSE_repeat_none;+ return set_basic;+ }+ }+ } else {+ size_t const basicCost = isDefaultAllowed ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, count, max) : ERROR(GENERIC);+ size_t const repeatCost = *repeatMode != FSE_repeat_none ? ZSTD_fseBitCost(prevCTable, count, max) : ERROR(GENERIC);+ size_t const NCountCost = ZSTD_NCountCost(count, max, nbSeq, FSELog);+ size_t const compressedCost = (NCountCost << 3) + ZSTD_entropyCost(count, max, nbSeq);++ if (isDefaultAllowed) {+ assert(!ZSTD_isError(basicCost));+ assert(!(*repeatMode == FSE_repeat_valid && ZSTD_isError(repeatCost)));+ }+ assert(!ZSTD_isError(NCountCost));+ assert(compressedCost < ERROR(maxCode));+ DEBUGLOG(5, "Estimated bit costs: basic=%u\trepeat=%u\tcompressed=%u",+ (unsigned)basicCost, (unsigned)repeatCost, (unsigned)compressedCost);+ if (basicCost <= repeatCost && basicCost <= compressedCost) {+ DEBUGLOG(5, "Selected set_basic");+ assert(isDefaultAllowed);+ *repeatMode = FSE_repeat_none;+ return set_basic;+ }+ if (repeatCost <= compressedCost) {+ DEBUGLOG(5, "Selected set_repeat");+ assert(!ZSTD_isError(repeatCost));+ return set_repeat;+ }+ assert(compressedCost < basicCost && compressedCost < repeatCost);+ }+ DEBUGLOG(5, "Selected set_compressed");+ *repeatMode = FSE_repeat_check;+ return set_compressed;+}++MEM_STATIC size_t+ZSTD_buildCTable(void* dst, size_t dstCapacity,+ FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,+ unsigned* count, U32 max,+ const BYTE* codeTable, size_t nbSeq,+ const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,+ const FSE_CTable* prevCTable, size_t prevCTableSize,+ void* workspace, size_t workspaceSize)+{+ BYTE* op = (BYTE*)dst;+ const BYTE* const oend = op + dstCapacity;+ DEBUGLOG(6, "ZSTD_buildCTable (dstCapacity=%u)", (unsigned)dstCapacity);++ switch (type) {+ case set_rle:+ FORWARD_IF_ERROR(FSE_buildCTable_rle(nextCTable, (BYTE)max));+ RETURN_ERROR_IF(dstCapacity==0, dstSize_tooSmall);+ *op = codeTable[0];+ return 1;+ case set_repeat:+ memcpy(nextCTable, prevCTable, prevCTableSize);+ return 0;+ case set_basic:+ FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, workspace, workspaceSize)); /* note : could be pre-calculated */+ return 0;+ case set_compressed: {+ S16 norm[MaxSeq + 1];+ size_t nbSeq_1 = nbSeq;+ const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);+ if (count[codeTable[nbSeq-1]] > 1) {+ count[codeTable[nbSeq-1]]--;+ nbSeq_1--;+ }+ assert(nbSeq_1 > 1);+ FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max));+ { size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */+ FORWARD_IF_ERROR(NCountSize);+ FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, norm, max, tableLog, workspace, workspaceSize));+ return NCountSize;+ }+ }+ default: assert(0); RETURN_ERROR(GENERIC);+ }+}++FORCE_INLINE_TEMPLATE size_t+ZSTD_encodeSequences_body(+ void* dst, size_t dstCapacity,+ FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,+ FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,+ FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,+ seqDef const* sequences, size_t nbSeq, int longOffsets)+{+ BIT_CStream_t blockStream;+ FSE_CState_t stateMatchLength;+ FSE_CState_t stateOffsetBits;+ FSE_CState_t stateLitLength;++ RETURN_ERROR_IF(+ ERR_isError(BIT_initCStream(&blockStream, dst, dstCapacity)),+ dstSize_tooSmall, "not enough space remaining");+ DEBUGLOG(6, "available space for bitstream : %i (dstCapacity=%u)",+ (int)(blockStream.endPtr - blockStream.startPtr),+ (unsigned)dstCapacity);++ /* first symbols */+ FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]);+ FSE_initCState2(&stateOffsetBits, CTable_OffsetBits, ofCodeTable[nbSeq-1]);+ FSE_initCState2(&stateLitLength, CTable_LitLength, llCodeTable[nbSeq-1]);+ BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]);+ if (MEM_32bits()) BIT_flushBits(&blockStream);+ BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]);+ if (MEM_32bits()) BIT_flushBits(&blockStream);+ if (longOffsets) {+ U32 const ofBits = ofCodeTable[nbSeq-1];+ int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);+ if (extraBits) {+ BIT_addBits(&blockStream, sequences[nbSeq-1].offset, extraBits);+ BIT_flushBits(&blockStream);+ }+ BIT_addBits(&blockStream, sequences[nbSeq-1].offset >> extraBits,+ ofBits - extraBits);+ } else {+ BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]);+ }+ BIT_flushBits(&blockStream);++ { size_t n;+ for (n=nbSeq-2 ; n<nbSeq ; n--) { /* intentional underflow */+ BYTE const llCode = llCodeTable[n];+ BYTE const ofCode = ofCodeTable[n];+ BYTE const mlCode = mlCodeTable[n];+ U32 const llBits = LL_bits[llCode];+ U32 const ofBits = ofCode;+ U32 const mlBits = ML_bits[mlCode];+ DEBUGLOG(6, "encoding: litlen:%2u - matchlen:%2u - offCode:%7u",+ (unsigned)sequences[n].litLength,+ (unsigned)sequences[n].matchLength + MINMATCH,+ (unsigned)sequences[n].offset);+ /* 32b*/ /* 64b*/+ /* (7)*/ /* (7)*/+ FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode); /* 15 */ /* 15 */+ FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode); /* 24 */ /* 24 */+ if (MEM_32bits()) BIT_flushBits(&blockStream); /* (7)*/+ FSE_encodeSymbol(&blockStream, &stateLitLength, llCode); /* 16 */ /* 33 */+ if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog)))+ BIT_flushBits(&blockStream); /* (7)*/+ BIT_addBits(&blockStream, sequences[n].litLength, llBits);+ if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);+ BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);+ if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream);+ if (longOffsets) {+ int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);+ if (extraBits) {+ BIT_addBits(&blockStream, sequences[n].offset, extraBits);+ BIT_flushBits(&blockStream); /* (7)*/+ }+ BIT_addBits(&blockStream, sequences[n].offset >> extraBits,+ ofBits - extraBits); /* 31 */+ } else {+ BIT_addBits(&blockStream, sequences[n].offset, ofBits); /* 31 */+ }+ BIT_flushBits(&blockStream); /* (7)*/+ DEBUGLOG(7, "remaining space : %i", (int)(blockStream.endPtr - blockStream.ptr));+ } }++ DEBUGLOG(6, "ZSTD_encodeSequences: flushing ML state with %u bits", stateMatchLength.stateLog);+ FSE_flushCState(&blockStream, &stateMatchLength);+ DEBUGLOG(6, "ZSTD_encodeSequences: flushing Off state with %u bits", stateOffsetBits.stateLog);+ FSE_flushCState(&blockStream, &stateOffsetBits);+ DEBUGLOG(6, "ZSTD_encodeSequences: flushing LL state with %u bits", stateLitLength.stateLog);+ FSE_flushCState(&blockStream, &stateLitLength);++ { size_t const streamSize = BIT_closeCStream(&blockStream);+ RETURN_ERROR_IF(streamSize==0, dstSize_tooSmall, "not enough space");+ return streamSize;+ }+}++static size_t+ZSTD_encodeSequences_default(+ void* dst, size_t dstCapacity,+ FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,+ FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,+ FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,+ seqDef const* sequences, size_t nbSeq, int longOffsets)+{+ return ZSTD_encodeSequences_body(dst, dstCapacity,+ CTable_MatchLength, mlCodeTable,+ CTable_OffsetBits, ofCodeTable,+ CTable_LitLength, llCodeTable,+ sequences, nbSeq, longOffsets);+}+++#if DYNAMIC_BMI2++static TARGET_ATTRIBUTE("bmi2") size_t+ZSTD_encodeSequences_bmi2(+ void* dst, size_t dstCapacity,+ FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,+ FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,+ FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,+ seqDef const* sequences, size_t nbSeq, int longOffsets)+{+ return ZSTD_encodeSequences_body(dst, dstCapacity,+ CTable_MatchLength, mlCodeTable,+ CTable_OffsetBits, ofCodeTable,+ CTable_LitLength, llCodeTable,+ sequences, nbSeq, longOffsets);+}++#endif++static size_t ZSTD_encodeSequences(+ void* dst, size_t dstCapacity,+ FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,+ FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,+ FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,+ seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2)+{+ DEBUGLOG(5, "ZSTD_encodeSequences: dstCapacity = %u", (unsigned)dstCapacity);+#if DYNAMIC_BMI2+ if (bmi2) {+ return ZSTD_encodeSequences_bmi2(dst, dstCapacity,+ CTable_MatchLength, mlCodeTable,+ CTable_OffsetBits, ofCodeTable,+ CTable_LitLength, llCodeTable,+ sequences, nbSeq, longOffsets);+ }+#endif+ (void)bmi2;+ return ZSTD_encodeSequences_default(dst, dstCapacity,+ CTable_MatchLength, mlCodeTable,+ CTable_OffsetBits, ofCodeTable,+ CTable_LitLength, llCodeTable,+ sequences, nbSeq, longOffsets);+}++static int ZSTD_disableLiteralsCompression(const ZSTD_CCtx_params* cctxParams)+{+ switch (cctxParams->literalCompressionMode) {+ case ZSTD_lcm_huffman:+ return 0;+ case ZSTD_lcm_uncompressed:+ return 1;+ default:+ assert(0 /* impossible: pre-validated */);+ /* fall-through */+ case ZSTD_lcm_auto:+ return (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0);+ }+}++/* ZSTD_compressSequences_internal():+ * actually compresses both literals and sequences */+MEM_STATIC size_t+ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,+ const ZSTD_entropyCTables_t* prevEntropy,+ ZSTD_entropyCTables_t* nextEntropy,+ const ZSTD_CCtx_params* cctxParams,+ void* dst, size_t dstCapacity,+ void* workspace, size_t wkspSize,+ const int bmi2)+{+ const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN;+ ZSTD_strategy const strategy = cctxParams->cParams.strategy;+ unsigned count[MaxSeq+1];+ FSE_CTable* CTable_LitLength = nextEntropy->fse.litlengthCTable;+ FSE_CTable* CTable_OffsetBits = nextEntropy->fse.offcodeCTable;+ FSE_CTable* CTable_MatchLength = nextEntropy->fse.matchlengthCTable;+ U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */+ const seqDef* const sequences = seqStorePtr->sequencesStart;+ const BYTE* const ofCodeTable = seqStorePtr->ofCode;+ const BYTE* const llCodeTable = seqStorePtr->llCode;+ const BYTE* const mlCodeTable = seqStorePtr->mlCode;+ BYTE* const ostart = (BYTE*)dst;+ BYTE* const oend = ostart + dstCapacity;+ BYTE* op = ostart;+ size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;+ BYTE* seqHead;+ BYTE* lastNCount = NULL;++ ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog)));+ DEBUGLOG(5, "ZSTD_compressSequences_internal");++ /* Compress literals */+ { const BYTE* const literals = seqStorePtr->litStart;+ size_t const litSize = seqStorePtr->lit - literals;+ size_t const cSize = ZSTD_compressLiterals(+ &prevEntropy->huf, &nextEntropy->huf,+ cctxParams->cParams.strategy,+ ZSTD_disableLiteralsCompression(cctxParams),+ op, dstCapacity,+ literals, litSize,+ workspace, wkspSize,+ bmi2);+ FORWARD_IF_ERROR(cSize);+ assert(cSize <= dstCapacity);+ op += cSize;+ }++ /* Sequences Header */+ RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,+ dstSize_tooSmall);+ if (nbSeq < 0x7F)+ *op++ = (BYTE)nbSeq;+ else if (nbSeq < LONGNBSEQ)+ op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;+ else+ op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;+ if (nbSeq==0) {+ /* Copy the old tables over as if we repeated them */+ memcpy(&nextEntropy->fse, &prevEntropy->fse, sizeof(prevEntropy->fse));+ return op - ostart;+ }++ /* seqHead : flags for FSE encoding type */+ seqHead = op++;++ /* convert length/distances into codes */+ ZSTD_seqToCodes(seqStorePtr);+ /* build CTable for Literal Lengths */+ { unsigned max = MaxLL;+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace, wkspSize); /* can't fail */+ DEBUGLOG(5, "Building LL table");+ nextEntropy->fse.litlength_repeatMode = prevEntropy->fse.litlength_repeatMode;+ LLtype = ZSTD_selectEncodingType(&nextEntropy->fse.litlength_repeatMode,+ count, max, mostFrequent, nbSeq,+ LLFSELog, prevEntropy->fse.litlengthCTable,+ LL_defaultNorm, LL_defaultNormLog,+ ZSTD_defaultAllowed, strategy);+ assert(set_basic < set_compressed && set_rle < set_compressed);+ assert(!(LLtype < set_compressed && nextEntropy->fse.litlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */+ { size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,+ count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL,+ prevEntropy->fse.litlengthCTable, sizeof(prevEntropy->fse.litlengthCTable),+ workspace, wkspSize);+ FORWARD_IF_ERROR(countSize);+ if (LLtype == set_compressed)+ lastNCount = op;+ op += countSize;+ } }+ /* build CTable for Offsets */+ { unsigned max = MaxOff;+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, ofCodeTable, nbSeq, workspace, wkspSize); /* can't fail */+ /* We can only use the basic table if max <= DefaultMaxOff, otherwise the offsets are too large */+ ZSTD_defaultPolicy_e const defaultPolicy = (max <= DefaultMaxOff) ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed;+ DEBUGLOG(5, "Building OF table");+ nextEntropy->fse.offcode_repeatMode = prevEntropy->fse.offcode_repeatMode;+ Offtype = ZSTD_selectEncodingType(&nextEntropy->fse.offcode_repeatMode,+ count, max, mostFrequent, nbSeq,+ OffFSELog, prevEntropy->fse.offcodeCTable,+ OF_defaultNorm, OF_defaultNormLog,+ defaultPolicy, strategy);+ assert(!(Offtype < set_compressed && nextEntropy->fse.offcode_repeatMode != FSE_repeat_none)); /* We don't copy tables */+ { size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,+ count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,+ prevEntropy->fse.offcodeCTable, sizeof(prevEntropy->fse.offcodeCTable),+ workspace, wkspSize);+ FORWARD_IF_ERROR(countSize);+ if (Offtype == set_compressed)+ lastNCount = op;+ op += countSize;+ } }+ /* build CTable for MatchLengths */+ { unsigned max = MaxML;+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, mlCodeTable, nbSeq, workspace, wkspSize); /* can't fail */+ DEBUGLOG(5, "Building ML table (remaining space : %i)", (int)(oend-op));+ nextEntropy->fse.matchlength_repeatMode = prevEntropy->fse.matchlength_repeatMode;+ MLtype = ZSTD_selectEncodingType(&nextEntropy->fse.matchlength_repeatMode,+ count, max, mostFrequent, nbSeq,+ MLFSELog, prevEntropy->fse.matchlengthCTable,+ ML_defaultNorm, ML_defaultNormLog,+ ZSTD_defaultAllowed, strategy);+ assert(!(MLtype < set_compressed && nextEntropy->fse.matchlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */+ { size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,+ count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML,+ prevEntropy->fse.matchlengthCTable, sizeof(prevEntropy->fse.matchlengthCTable),+ workspace, wkspSize);+ FORWARD_IF_ERROR(countSize);+ if (MLtype == set_compressed)+ lastNCount = op;+ op += countSize;+ } }++ *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2));++ { size_t const bitstreamSize = ZSTD_encodeSequences(+ op, oend - op,+ CTable_MatchLength, mlCodeTable,+ CTable_OffsetBits, ofCodeTable,+ CTable_LitLength, llCodeTable,+ sequences, nbSeq,+ longOffsets, bmi2);+ FORWARD_IF_ERROR(bitstreamSize);+ op += bitstreamSize;+ /* zstd versions <= 1.3.4 mistakenly report corruption when+ * FSE_readNCount() receives a buffer < 4 bytes.+ * Fixed by https://github.com/facebook/zstd/pull/1146.+ * This can happen when the last set_compressed table present is 2+ * bytes and the bitstream is only one byte.+ * In this exceedingly rare case, we will simply emit an uncompressed+ * block, since it isn't worth optimizing.+ */+ if (lastNCount && (op - lastNCount) < 4) {+ /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */+ assert(op - lastNCount == 3);+ DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by "+ "emitting an uncompressed block.");+ return 0;+ }+ }++ DEBUGLOG(5, "compressed block size : %u", (unsigned)(op - ostart));+ return op - ostart;+}++MEM_STATIC size_t+ZSTD_compressSequences(seqStore_t* seqStorePtr,+ const ZSTD_entropyCTables_t* prevEntropy,+ ZSTD_entropyCTables_t* nextEntropy,+ const ZSTD_CCtx_params* cctxParams,+ void* dst, size_t dstCapacity,+ size_t srcSize,+ void* workspace, size_t wkspSize,+ int bmi2)+{+ size_t const cSize = ZSTD_compressSequences_internal(+ seqStorePtr, prevEntropy, nextEntropy, cctxParams,+ dst, dstCapacity,+ workspace, wkspSize, bmi2);+ if (cSize == 0) return 0;+ /* When srcSize <= dstCapacity, there is enough space to write a raw uncompressed block.+ * Since we ran out of space, block must be not compressible, so fall back to raw uncompressed block.+ */+ if ((cSize == ERROR(dstSize_tooSmall)) & (srcSize <= dstCapacity))+ return 0; /* block not compressed */+ FORWARD_IF_ERROR(cSize);++ /* Check compressibility */+ { size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, cctxParams->cParams.strategy);+ if (cSize >= maxCSize) return 0; /* block not compressed */+ }++ return cSize;+}++/* ZSTD_selectBlockCompressor() :+ * Not static, but internal use only (used by long distance matcher)+ * assumption : strat is a valid strategy */+ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_dictMode_e dictMode)+{+ static const ZSTD_blockCompressor blockCompressor[3][ZSTD_STRATEGY_MAX+1] = {+ { ZSTD_compressBlock_fast /* default for 0 */,+ ZSTD_compressBlock_fast,+ ZSTD_compressBlock_doubleFast,+ ZSTD_compressBlock_greedy,+ ZSTD_compressBlock_lazy,+ ZSTD_compressBlock_lazy2,+ ZSTD_compressBlock_btlazy2,+ ZSTD_compressBlock_btopt,+ ZSTD_compressBlock_btultra,+ ZSTD_compressBlock_btultra2 },+ { ZSTD_compressBlock_fast_extDict /* default for 0 */,+ ZSTD_compressBlock_fast_extDict,+ ZSTD_compressBlock_doubleFast_extDict,+ ZSTD_compressBlock_greedy_extDict,+ ZSTD_compressBlock_lazy_extDict,+ ZSTD_compressBlock_lazy2_extDict,+ ZSTD_compressBlock_btlazy2_extDict,+ ZSTD_compressBlock_btopt_extDict,+ ZSTD_compressBlock_btultra_extDict,+ ZSTD_compressBlock_btultra_extDict },+ { ZSTD_compressBlock_fast_dictMatchState /* default for 0 */,+ ZSTD_compressBlock_fast_dictMatchState,+ ZSTD_compressBlock_doubleFast_dictMatchState,+ ZSTD_compressBlock_greedy_dictMatchState,+ ZSTD_compressBlock_lazy_dictMatchState,+ ZSTD_compressBlock_lazy2_dictMatchState,+ ZSTD_compressBlock_btlazy2_dictMatchState,+ ZSTD_compressBlock_btopt_dictMatchState,+ ZSTD_compressBlock_btultra_dictMatchState,+ ZSTD_compressBlock_btultra_dictMatchState }+ };+ ZSTD_blockCompressor selectedCompressor;+ ZSTD_STATIC_ASSERT((unsigned)ZSTD_fast == 1);++ assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat));+ selectedCompressor = blockCompressor[(int)dictMode][(int)strat];+ assert(selectedCompressor != NULL);+ return selectedCompressor;+}++static void ZSTD_storeLastLiterals(seqStore_t* seqStorePtr,+ const BYTE* anchor, size_t lastLLSize)+{+ memcpy(seqStorePtr->lit, anchor, lastLLSize);+ seqStorePtr->lit += lastLLSize;+}++void ZSTD_resetSeqStore(seqStore_t* ssPtr)+{+ ssPtr->lit = ssPtr->litStart;+ ssPtr->sequences = ssPtr->sequencesStart;+ ssPtr->longLengthID = 0;+}++static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ ZSTD_matchState_t* const ms = &zc->blockState.matchState;+ size_t cSize;+ DEBUGLOG(5, "ZSTD_compressBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",+ (unsigned)dstCapacity, (unsigned)ms->window.dictLimit, (unsigned)ms->nextToUpdate);+ assert(srcSize <= ZSTD_BLOCKSIZE_MAX);++ /* Assert that we have correctly flushed the ctx params into the ms's copy */+ ZSTD_assertEqualCParams(zc->appliedParams.cParams, ms->cParams);++ if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) {+ ZSTD_ldm_skipSequences(&zc->externSeqStore, srcSize, zc->appliedParams.cParams.minMatch);+ cSize = 0;+ goto out; /* don't even attempt compression below a certain srcSize */+ }+ ZSTD_resetSeqStore(&(zc->seqStore));+ /* required for optimal parser to read stats from dictionary */+ ms->opt.symbolCosts = &zc->blockState.prevCBlock->entropy;+ /* tell the optimal parser how we expect to compress literals */+ ms->opt.literalCompressionMode = zc->appliedParams.literalCompressionMode;++ /* a gap between an attached dict and the current window is not safe,+ * they must remain adjacent,+ * and when that stops being the case, the dict must be unset */+ assert(ms->dictMatchState == NULL || ms->loadedDictEnd == ms->window.dictLimit);++ /* limited update after a very long match */+ { const BYTE* const base = ms->window.base;+ const BYTE* const istart = (const BYTE*)src;+ const U32 current = (U32)(istart-base);+ if (sizeof(ptrdiff_t)==8) assert(istart - base < (ptrdiff_t)(U32)(-1)); /* ensure no overflow */+ if (current > ms->nextToUpdate + 384)+ ms->nextToUpdate = current - MIN(192, (U32)(current - ms->nextToUpdate - 384));+ }++ /* select and store sequences */+ { ZSTD_dictMode_e const dictMode = ZSTD_matchState_dictMode(ms);+ size_t lastLLSize;+ { int i;+ for (i = 0; i < ZSTD_REP_NUM; ++i)+ zc->blockState.nextCBlock->rep[i] = zc->blockState.prevCBlock->rep[i];+ }+ if (zc->externSeqStore.pos < zc->externSeqStore.size) {+ assert(!zc->appliedParams.ldmParams.enableLdm);+ /* Updates ldmSeqStore.pos */+ lastLLSize =+ ZSTD_ldm_blockCompress(&zc->externSeqStore,+ ms, &zc->seqStore,+ zc->blockState.nextCBlock->rep,+ src, srcSize);+ assert(zc->externSeqStore.pos <= zc->externSeqStore.size);+ } else if (zc->appliedParams.ldmParams.enableLdm) {+ rawSeqStore_t ldmSeqStore = {NULL, 0, 0, 0};++ ldmSeqStore.seq = zc->ldmSequences;+ ldmSeqStore.capacity = zc->maxNbLdmSequences;+ /* Updates ldmSeqStore.size */+ FORWARD_IF_ERROR(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore,+ &zc->appliedParams.ldmParams,+ src, srcSize));+ /* Updates ldmSeqStore.pos */+ lastLLSize =+ ZSTD_ldm_blockCompress(&ldmSeqStore,+ ms, &zc->seqStore,+ zc->blockState.nextCBlock->rep,+ src, srcSize);+ assert(ldmSeqStore.pos == ldmSeqStore.size);+ } else { /* not long range mode */+ ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, dictMode);+ lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize);+ }+ { const BYTE* const lastLiterals = (const BYTE*)src + srcSize - lastLLSize;+ ZSTD_storeLastLiterals(&zc->seqStore, lastLiterals, lastLLSize);+ } }++ /* encode sequences and literals */+ cSize = ZSTD_compressSequences(&zc->seqStore,+ &zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy,+ &zc->appliedParams,+ dst, dstCapacity,+ srcSize,+ zc->entropyWorkspace, HUF_WORKSPACE_SIZE /* statically allocated in resetCCtx */,+ zc->bmi2);++out:+ if (!ZSTD_isError(cSize) && cSize != 0) {+ /* confirm repcodes and entropy tables when emitting a compressed block */+ ZSTD_compressedBlockState_t* const tmp = zc->blockState.prevCBlock;+ zc->blockState.prevCBlock = zc->blockState.nextCBlock;+ zc->blockState.nextCBlock = tmp;+ }+ /* We check that dictionaries have offset codes available for the first+ * block. After the first block, the offcode table might not have large+ * enough codes to represent the offsets in the data.+ */+ if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid)+ zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check;++ return cSize;+}+++/*! ZSTD_compress_frameChunk() :+* Compress a chunk of data into one or multiple blocks.+* All blocks will be terminated, all input will be consumed.+* Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.+* Frame is supposed already started (header already produced)+* @return : compressed size, or an error code+*/+static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ U32 lastFrameChunk)+{+ size_t blockSize = cctx->blockSize;+ size_t remaining = srcSize;+ const BYTE* ip = (const BYTE*)src;+ BYTE* const ostart = (BYTE*)dst;+ BYTE* op = ostart;+ U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog;+ assert(cctx->appliedParams.cParams.windowLog <= 31);++ DEBUGLOG(5, "ZSTD_compress_frameChunk (blockSize=%u)", (unsigned)blockSize);+ if (cctx->appliedParams.fParams.checksumFlag && srcSize)+ XXH64_update(&cctx->xxhState, src, srcSize);++ while (remaining) {+ ZSTD_matchState_t* const ms = &cctx->blockState.matchState;+ U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);++ RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE,+ dstSize_tooSmall,+ "not enough space to store compressed block");+ if (remaining < blockSize) blockSize = remaining;++ if (ZSTD_window_needOverflowCorrection(ms->window, ip + blockSize)) {+ U32 const cycleLog = ZSTD_cycleLog(cctx->appliedParams.cParams.chainLog, cctx->appliedParams.cParams.strategy);+ U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, maxDist, ip);+ ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);+ ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);+ ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);+ ZSTD_reduceIndex(cctx, correction);+ if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;+ else ms->nextToUpdate -= correction;+ ms->loadedDictEnd = 0;+ ms->dictMatchState = NULL;+ }+ ZSTD_window_enforceMaxDist(&ms->window, ip + blockSize, maxDist, &ms->loadedDictEnd, &ms->dictMatchState);+ if (ms->nextToUpdate < ms->window.lowLimit) ms->nextToUpdate = ms->window.lowLimit;++ { size_t cSize = ZSTD_compressBlock_internal(cctx,+ op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize,+ ip, blockSize);+ FORWARD_IF_ERROR(cSize);++ if (cSize == 0) { /* block is not compressible */+ cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock);+ FORWARD_IF_ERROR(cSize);+ } else {+ U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);+ MEM_writeLE24(op, cBlockHeader24);+ cSize += ZSTD_blockHeaderSize;+ }++ ip += blockSize;+ assert(remaining >= blockSize);+ remaining -= blockSize;+ op += cSize;+ assert(dstCapacity >= cSize);+ dstCapacity -= cSize;+ DEBUGLOG(5, "ZSTD_compress_frameChunk: adding a block of size %u",+ (unsigned)cSize);+ } }++ if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending;+ return op-ostart;+}+++static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity,+ ZSTD_CCtx_params params, U64 pledgedSrcSize, U32 dictID)+{ BYTE* const op = (BYTE*)dst;+ U32 const dictIDSizeCodeLength = (dictID>0) + (dictID>=256) + (dictID>=65536); /* 0-3 */+ U32 const dictIDSizeCode = params.fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength; /* 0-3 */+ U32 const checksumFlag = params.fParams.checksumFlag>0;+ U32 const windowSize = (U32)1 << params.cParams.windowLog;+ U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);+ BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);+ U32 const fcsCode = params.fParams.contentSizeFlag ?+ (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : 0; /* 0-3 */+ BYTE const frameHeaderDescriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );+ size_t pos=0;++ assert(!(params.fParams.contentSizeFlag && pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN));+ RETURN_ERROR_IF(dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX, dstSize_tooSmall);+ DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u",+ !params.fParams.noDictIDFlag, (unsigned)dictID, (unsigned)dictIDSizeCode);++ if (params.format == ZSTD_f_zstd1) {+ MEM_writeLE32(dst, ZSTD_MAGICNUMBER);+ pos = 4;+ }+ op[pos++] = frameHeaderDescriptionByte;+ if (!singleSegment) op[pos++] = windowLogByte;+ switch(dictIDSizeCode)+ {+ default: assert(0); /* impossible */+ case 0 : break;+ case 1 : op[pos] = (BYTE)(dictID); pos++; break;+ case 2 : MEM_writeLE16(op+pos, (U16)dictID); pos+=2; break;+ case 3 : MEM_writeLE32(op+pos, dictID); pos+=4; break;+ }+ switch(fcsCode)+ {+ default: assert(0); /* impossible */+ case 0 : if (singleSegment) op[pos++] = (BYTE)(pledgedSrcSize); break;+ case 1 : MEM_writeLE16(op+pos, (U16)(pledgedSrcSize-256)); pos+=2; break;+ case 2 : MEM_writeLE32(op+pos, (U32)(pledgedSrcSize)); pos+=4; break;+ case 3 : MEM_writeLE64(op+pos, (U64)(pledgedSrcSize)); pos+=8; break;+ }+ return pos;+}++/* ZSTD_writeLastEmptyBlock() :+ * output an empty Block with end-of-frame mark to complete a frame+ * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))+ * or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)+ */+size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity)+{+ RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize, dstSize_tooSmall);+ { U32 const cBlockHeader24 = 1 /*lastBlock*/ + (((U32)bt_raw)<<1); /* 0 size */+ MEM_writeLE24(dst, cBlockHeader24);+ return ZSTD_blockHeaderSize;+ }+}++size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq)+{+ RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong);+ RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm,+ parameter_unsupported);+ cctx->externSeqStore.seq = seq;+ cctx->externSeqStore.size = nbSeq;+ cctx->externSeqStore.capacity = nbSeq;+ cctx->externSeqStore.pos = 0;+ return 0;+}+++static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ U32 frame, U32 lastFrameChunk)+{+ ZSTD_matchState_t* const ms = &cctx->blockState.matchState;+ size_t fhSize = 0;++ DEBUGLOG(5, "ZSTD_compressContinue_internal, stage: %u, srcSize: %u",+ cctx->stage, (unsigned)srcSize);+ RETURN_ERROR_IF(cctx->stage==ZSTDcs_created, stage_wrong,+ "missing init (ZSTD_compressBegin)");++ if (frame && (cctx->stage==ZSTDcs_init)) {+ fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams,+ cctx->pledgedSrcSizePlusOne-1, cctx->dictID);+ FORWARD_IF_ERROR(fhSize);+ dstCapacity -= fhSize;+ dst = (char*)dst + fhSize;+ cctx->stage = ZSTDcs_ongoing;+ }++ if (!srcSize) return fhSize; /* do not generate an empty block if no input */++ if (!ZSTD_window_update(&ms->window, src, srcSize)) {+ ms->nextToUpdate = ms->window.dictLimit;+ }+ if (cctx->appliedParams.ldmParams.enableLdm) {+ ZSTD_window_update(&cctx->ldmState.window, src, srcSize);+ }++ if (!frame) {+ /* overflow check and correction for block mode */+ if (ZSTD_window_needOverflowCorrection(ms->window, (const char*)src + srcSize)) {+ U32 const cycleLog = ZSTD_cycleLog(cctx->appliedParams.cParams.chainLog, cctx->appliedParams.cParams.strategy);+ U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, 1 << cctx->appliedParams.cParams.windowLog, src);+ ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);+ ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);+ ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);+ ZSTD_reduceIndex(cctx, correction);+ if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;+ else ms->nextToUpdate -= correction;+ ms->loadedDictEnd = 0;+ ms->dictMatchState = NULL;+ }+ }++ DEBUGLOG(5, "ZSTD_compressContinue_internal (blockSize=%u)", (unsigned)cctx->blockSize);+ { size_t const cSize = frame ?+ ZSTD_compress_frameChunk (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) :+ ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize);+ FORWARD_IF_ERROR(cSize);+ cctx->consumedSrcSize += srcSize;+ cctx->producedCSize += (cSize + fhSize);+ assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));+ if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */+ ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);+ RETURN_ERROR_IF(+ cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne,+ srcSize_wrong,+ "error : pledgedSrcSize = %u, while realSrcSize >= %u",+ (unsigned)cctx->pledgedSrcSizePlusOne-1,+ (unsigned)cctx->consumedSrcSize);+ }+ return cSize + fhSize;+ }+}++size_t ZSTD_compressContinue (ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ DEBUGLOG(5, "ZSTD_compressContinue (srcSize=%u)", (unsigned)srcSize);+ return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 0 /* last chunk */);+}+++size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx)+{+ ZSTD_compressionParameters const cParams = cctx->appliedParams.cParams;+ assert(!ZSTD_checkCParams(cParams));+ return MIN (ZSTD_BLOCKSIZE_MAX, (U32)1 << cParams.windowLog);+}++size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+ size_t const blockSizeMax = ZSTD_getBlockSize(cctx);+ RETURN_ERROR_IF(srcSize > blockSizeMax, srcSize_wrong);++ return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0 /* frame mode */, 0 /* last chunk */);+}++/*! ZSTD_loadDictionaryContent() :+ * @return : 0, or an error code+ */+static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,+ ZSTD_CCtx_params const* params,+ const void* src, size_t srcSize,+ ZSTD_dictTableLoadMethod_e dtlm)+{+ const BYTE* const ip = (const BYTE*) src;+ const BYTE* const iend = ip + srcSize;++ ZSTD_window_update(&ms->window, src, srcSize);+ ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base);++ /* Assert that we the ms params match the params we're being given */+ ZSTD_assertEqualCParams(params->cParams, ms->cParams);++ if (srcSize <= HASH_READ_SIZE) return 0;++ switch(params->cParams.strategy)+ {+ case ZSTD_fast:+ ZSTD_fillHashTable(ms, iend, dtlm);+ break;+ case ZSTD_dfast:+ ZSTD_fillDoubleHashTable(ms, iend, dtlm);+ break;++ case ZSTD_greedy:+ case ZSTD_lazy:+ case ZSTD_lazy2:+ if (srcSize >= HASH_READ_SIZE)+ ZSTD_insertAndFindFirstIndex(ms, iend-HASH_READ_SIZE);+ break;++ case ZSTD_btlazy2: /* we want the dictionary table fully sorted */+ case ZSTD_btopt:+ case ZSTD_btultra:+ case ZSTD_btultra2:+ if (srcSize >= HASH_READ_SIZE)+ ZSTD_updateTree(ms, iend-HASH_READ_SIZE, iend);+ break;++ default:+ assert(0); /* not possible : not a valid strategy id */+ }++ ms->nextToUpdate = (U32)(iend - ms->window.base);+ return 0;+}+++/* Dictionaries that assign zero probability to symbols that show up causes problems+ when FSE encoding. Refuse dictionaries that assign zero probability to symbols+ that we may encounter during compression.+ NOTE: This behavior is not standard and could be improved in the future. */+static size_t ZSTD_checkDictNCount(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue) {+ U32 s;+ RETURN_ERROR_IF(dictMaxSymbolValue < maxSymbolValue, dictionary_corrupted);+ for (s = 0; s <= maxSymbolValue; ++s) {+ RETURN_ERROR_IF(normalizedCounter[s] == 0, dictionary_corrupted);+ }+ return 0;+}+++/* Dictionary format :+ * See :+ * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format+ */+/*! ZSTD_loadZstdDictionary() :+ * @return : dictID, or an error code+ * assumptions : magic number supposed already checked+ * dictSize supposed > 8+ */+static size_t ZSTD_loadZstdDictionary(ZSTD_compressedBlockState_t* bs,+ ZSTD_matchState_t* ms,+ ZSTD_CCtx_params const* params,+ const void* dict, size_t dictSize,+ ZSTD_dictTableLoadMethod_e dtlm,+ void* workspace)+{+ const BYTE* dictPtr = (const BYTE*)dict;+ const BYTE* const dictEnd = dictPtr + dictSize;+ short offcodeNCount[MaxOff+1];+ unsigned offcodeMaxValue = MaxOff;+ size_t dictID;++ ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog)));+ assert(dictSize > 8);+ assert(MEM_readLE32(dictPtr) == ZSTD_MAGIC_DICTIONARY);++ dictPtr += 4; /* skip magic number */+ dictID = params->fParams.noDictIDFlag ? 0 : MEM_readLE32(dictPtr);+ dictPtr += 4;++ { unsigned maxSymbolValue = 255;+ size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)bs->entropy.huf.CTable, &maxSymbolValue, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted);+ dictPtr += hufHeaderSize;+ }++ { unsigned offcodeLog;+ size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted);+ /* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */+ /* fill all offset symbols to avoid garbage at end of table */+ RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(+ bs->entropy.fse.offcodeCTable,+ offcodeNCount, MaxOff, offcodeLog,+ workspace, HUF_WORKSPACE_SIZE)),+ dictionary_corrupted);+ dictPtr += offcodeHeaderSize;+ }++ { short matchlengthNCount[MaxML+1];+ unsigned matchlengthMaxValue = MaxML, matchlengthLog;+ size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted);+ /* Every match length code must have non-zero probability */+ FORWARD_IF_ERROR( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));+ RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(+ bs->entropy.fse.matchlengthCTable,+ matchlengthNCount, matchlengthMaxValue, matchlengthLog,+ workspace, HUF_WORKSPACE_SIZE)),+ dictionary_corrupted);+ dictPtr += matchlengthHeaderSize;+ }++ { short litlengthNCount[MaxLL+1];+ unsigned litlengthMaxValue = MaxLL, litlengthLog;+ size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted);+ /* Every literal length code must have non-zero probability */+ FORWARD_IF_ERROR( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));+ RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(+ bs->entropy.fse.litlengthCTable,+ litlengthNCount, litlengthMaxValue, litlengthLog,+ workspace, HUF_WORKSPACE_SIZE)),+ dictionary_corrupted);+ dictPtr += litlengthHeaderSize;+ }++ RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted);+ bs->rep[0] = MEM_readLE32(dictPtr+0);+ bs->rep[1] = MEM_readLE32(dictPtr+4);+ bs->rep[2] = MEM_readLE32(dictPtr+8);+ dictPtr += 12;++ { size_t const dictContentSize = (size_t)(dictEnd - dictPtr);+ U32 offcodeMax = MaxOff;+ if (dictContentSize <= ((U32)-1) - 128 KB) {+ U32 const maxOffset = (U32)dictContentSize + 128 KB; /* The maximum offset that must be supported */+ offcodeMax = ZSTD_highbit32(maxOffset); /* Calculate minimum offset code required to represent maxOffset */+ }+ /* All offset values <= dictContentSize + 128 KB must be representable */+ FORWARD_IF_ERROR(ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));+ /* All repCodes must be <= dictContentSize and != 0*/+ { U32 u;+ for (u=0; u<3; u++) {+ RETURN_ERROR_IF(bs->rep[u] == 0, dictionary_corrupted);+ RETURN_ERROR_IF(bs->rep[u] > dictContentSize, dictionary_corrupted);+ } }++ bs->entropy.huf.repeatMode = HUF_repeat_valid;+ bs->entropy.fse.offcode_repeatMode = FSE_repeat_valid;+ bs->entropy.fse.matchlength_repeatMode = FSE_repeat_valid;+ bs->entropy.fse.litlength_repeatMode = FSE_repeat_valid;+ FORWARD_IF_ERROR(ZSTD_loadDictionaryContent(ms, params, dictPtr, dictContentSize, dtlm));+ return dictID;+ }+}++/** ZSTD_compress_insertDictionary() :+* @return : dictID, or an error code */+static size_t+ZSTD_compress_insertDictionary(ZSTD_compressedBlockState_t* bs,+ ZSTD_matchState_t* ms,+ const ZSTD_CCtx_params* params,+ const void* dict, size_t dictSize,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_dictTableLoadMethod_e dtlm,+ void* workspace)+{+ DEBUGLOG(4, "ZSTD_compress_insertDictionary (dictSize=%u)", (U32)dictSize);+ if ((dict==NULL) || (dictSize<=8)) return 0;++ ZSTD_reset_compressedBlockState(bs);++ /* dict restricted modes */+ if (dictContentType == ZSTD_dct_rawContent)+ return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm);++ if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) {+ if (dictContentType == ZSTD_dct_auto) {+ DEBUGLOG(4, "raw content dictionary detected");+ return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm);+ }+ RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong);+ assert(0); /* impossible */+ }++ /* dict as full zstd dictionary */+ return ZSTD_loadZstdDictionary(bs, ms, params, dict, dictSize, dtlm, workspace);+}++/*! ZSTD_compressBegin_internal() :+ * @return : 0, or an error code */+static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx,+ const void* dict, size_t dictSize,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_dictTableLoadMethod_e dtlm,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params, U64 pledgedSrcSize,+ ZSTD_buffered_policy_e zbuff)+{+ DEBUGLOG(4, "ZSTD_compressBegin_internal: wlog=%u", params.cParams.windowLog);+ /* params are supposed to be fully validated at this point */+ assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));+ assert(!((dict) && (cdict))); /* either dict or cdict, not both */++ if (cdict && cdict->dictContentSize>0) {+ return ZSTD_resetCCtx_usingCDict(cctx, cdict, params, pledgedSrcSize, zbuff);+ }++ FORWARD_IF_ERROR( ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,+ ZSTDcrp_continue, zbuff) );+ {+ size_t const dictID = ZSTD_compress_insertDictionary(+ cctx->blockState.prevCBlock, &cctx->blockState.matchState,+ ¶ms, dict, dictSize, dictContentType, dtlm, cctx->entropyWorkspace);+ FORWARD_IF_ERROR(dictID);+ assert(dictID <= (size_t)(U32)-1);+ cctx->dictID = (U32)dictID;+ }+ return 0;+}++size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,+ const void* dict, size_t dictSize,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_dictTableLoadMethod_e dtlm,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params,+ unsigned long long pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_compressBegin_advanced_internal: wlog=%u", params.cParams.windowLog);+ /* compression parameters verification and optimization */+ FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );+ return ZSTD_compressBegin_internal(cctx,+ dict, dictSize, dictContentType, dtlm,+ cdict,+ params, pledgedSrcSize,+ ZSTDb_not_buffered);+}++/*! ZSTD_compressBegin_advanced() :+* @return : 0, or an error code */+size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx,+ const void* dict, size_t dictSize,+ ZSTD_parameters params, unsigned long long pledgedSrcSize)+{+ ZSTD_CCtx_params const cctxParams =+ ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);+ return ZSTD_compressBegin_advanced_internal(cctx,+ dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast,+ NULL /*cdict*/,+ cctxParams, pledgedSrcSize);+}++size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel)+{+ ZSTD_parameters const params = ZSTD_getParams(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, dictSize);+ ZSTD_CCtx_params const cctxParams =+ ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);+ DEBUGLOG(4, "ZSTD_compressBegin_usingDict (dictSize=%u)", (unsigned)dictSize);+ return ZSTD_compressBegin_internal(cctx, dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,+ cctxParams, ZSTD_CONTENTSIZE_UNKNOWN, ZSTDb_not_buffered);+}++size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel)+{+ return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel);+}+++/*! ZSTD_writeEpilogue() :+* Ends a frame.+* @return : nb of bytes written into dst (or an error code) */+static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity)+{+ BYTE* const ostart = (BYTE*)dst;+ BYTE* op = ostart;+ size_t fhSize = 0;++ DEBUGLOG(4, "ZSTD_writeEpilogue");+ RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing");++ /* special case : empty frame */+ if (cctx->stage == ZSTDcs_init) {+ fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, 0, 0);+ FORWARD_IF_ERROR(fhSize);+ dstCapacity -= fhSize;+ op += fhSize;+ cctx->stage = ZSTDcs_ongoing;+ }++ if (cctx->stage != ZSTDcs_ending) {+ /* write one last empty block, make it the "last" block */+ U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0;+ RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall);+ MEM_writeLE32(op, cBlockHeader24);+ op += ZSTD_blockHeaderSize;+ dstCapacity -= ZSTD_blockHeaderSize;+ }++ if (cctx->appliedParams.fParams.checksumFlag) {+ U32 const checksum = (U32) XXH64_digest(&cctx->xxhState);+ RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall);+ DEBUGLOG(4, "ZSTD_writeEpilogue: write checksum : %08X", (unsigned)checksum);+ MEM_writeLE32(op, checksum);+ op += 4;+ }++ cctx->stage = ZSTDcs_created; /* return to "created but no init" status */+ return op-ostart;+}++size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ size_t endResult;+ size_t const cSize = ZSTD_compressContinue_internal(cctx,+ dst, dstCapacity, src, srcSize,+ 1 /* frame mode */, 1 /* last chunk */);+ FORWARD_IF_ERROR(cSize);+ endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize);+ FORWARD_IF_ERROR(endResult);+ assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));+ if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */+ ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);+ DEBUGLOG(4, "end of frame : controlling src size");+ RETURN_ERROR_IF(+ cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1,+ srcSize_wrong,+ "error : pledgedSrcSize = %u, while realSrcSize = %u",+ (unsigned)cctx->pledgedSrcSizePlusOne-1,+ (unsigned)cctx->consumedSrcSize);+ }+ return cSize + endResult;+}+++static size_t ZSTD_compress_internal (ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize,+ ZSTD_parameters params)+{+ ZSTD_CCtx_params const cctxParams =+ ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);+ DEBUGLOG(4, "ZSTD_compress_internal");+ return ZSTD_compress_advanced_internal(cctx,+ dst, dstCapacity,+ src, srcSize,+ dict, dictSize,+ cctxParams);+}++size_t ZSTD_compress_advanced (ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize,+ ZSTD_parameters params)+{+ DEBUGLOG(4, "ZSTD_compress_advanced");+ FORWARD_IF_ERROR(ZSTD_checkCParams(params.cParams));+ return ZSTD_compress_internal(cctx,+ dst, dstCapacity,+ src, srcSize,+ dict, dictSize,+ params);+}++/* Internal */+size_t ZSTD_compress_advanced_internal(+ ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize,+ ZSTD_CCtx_params params)+{+ DEBUGLOG(4, "ZSTD_compress_advanced_internal (srcSize:%u)", (unsigned)srcSize);+ FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,+ dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,+ params, srcSize, ZSTDb_not_buffered) );+ return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);+}++size_t ZSTD_compress_usingDict(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict, size_t dictSize,+ int compressionLevel)+{+ ZSTD_parameters const params = ZSTD_getParams(compressionLevel, srcSize + (!srcSize), dict ? dictSize : 0);+ ZSTD_CCtx_params cctxParams = ZSTD_assignParamsToCCtxParams(cctx->requestedParams, params);+ assert(params.fParams.contentSizeFlag == 1);+ return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, dict, dictSize, cctxParams);+}++size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ int compressionLevel)+{+ DEBUGLOG(4, "ZSTD_compressCCtx (srcSize=%u)", (unsigned)srcSize);+ assert(cctx != NULL);+ return ZSTD_compress_usingDict(cctx, dst, dstCapacity, src, srcSize, NULL, 0, compressionLevel);+}++size_t ZSTD_compress(void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ int compressionLevel)+{+ size_t result;+ ZSTD_CCtx ctxBody;+ ZSTD_initCCtx(&ctxBody, ZSTD_defaultCMem);+ result = ZSTD_compressCCtx(&ctxBody, dst, dstCapacity, src, srcSize, compressionLevel);+ ZSTD_freeCCtxContent(&ctxBody); /* can't free ctxBody itself, as it's on stack; free only heap content */+ return result;+}+++/* ===== Dictionary API ===== */++/*! ZSTD_estimateCDictSize_advanced() :+ * Estimate amount of memory that will be needed to create a dictionary with following arguments */+size_t ZSTD_estimateCDictSize_advanced(+ size_t dictSize, ZSTD_compressionParameters cParams,+ ZSTD_dictLoadMethod_e dictLoadMethod)+{+ DEBUGLOG(5, "sizeof(ZSTD_CDict) : %u", (unsigned)sizeof(ZSTD_CDict));+ return sizeof(ZSTD_CDict) + HUF_WORKSPACE_SIZE + ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0)+ + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);+}++size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel)+{+ ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, 0, dictSize);+ return ZSTD_estimateCDictSize_advanced(dictSize, cParams, ZSTD_dlm_byCopy);+}++size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict)+{+ if (cdict==NULL) return 0; /* support sizeof on NULL */+ DEBUGLOG(5, "sizeof(*cdict) : %u", (unsigned)sizeof(*cdict));+ return cdict->workspaceSize + (cdict->dictBuffer ? cdict->dictContentSize : 0) + sizeof(*cdict);+}++static size_t ZSTD_initCDict_internal(+ ZSTD_CDict* cdict,+ const void* dictBuffer, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_compressionParameters cParams)+{+ DEBUGLOG(3, "ZSTD_initCDict_internal (dictContentType:%u)", (unsigned)dictContentType);+ assert(!ZSTD_checkCParams(cParams));+ cdict->matchState.cParams = cParams;+ if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dictBuffer) || (!dictSize)) {+ cdict->dictBuffer = NULL;+ cdict->dictContent = dictBuffer;+ } else {+ void* const internalBuffer = ZSTD_malloc(dictSize, cdict->customMem);+ cdict->dictBuffer = internalBuffer;+ cdict->dictContent = internalBuffer;+ RETURN_ERROR_IF(!internalBuffer, memory_allocation);+ memcpy(internalBuffer, dictBuffer, dictSize);+ }+ cdict->dictContentSize = dictSize;++ /* Reset the state to no dictionary */+ ZSTD_reset_compressedBlockState(&cdict->cBlockState);+ { void* const end = ZSTD_reset_matchState(+ &cdict->matchState,+ (U32*)cdict->workspace + HUF_WORKSPACE_SIZE_U32,+ &cParams, ZSTDcrp_continue, /* forCCtx */ 0);+ assert(end == (char*)cdict->workspace + cdict->workspaceSize);+ (void)end;+ }+ /* (Maybe) load the dictionary+ * Skips loading the dictionary if it is <= 8 bytes.+ */+ { ZSTD_CCtx_params params;+ memset(¶ms, 0, sizeof(params));+ params.compressionLevel = ZSTD_CLEVEL_DEFAULT;+ params.fParams.contentSizeFlag = 1;+ params.cParams = cParams;+ { size_t const dictID = ZSTD_compress_insertDictionary(+ &cdict->cBlockState, &cdict->matchState, ¶ms,+ cdict->dictContent, cdict->dictContentSize,+ dictContentType, ZSTD_dtlm_full, cdict->workspace);+ FORWARD_IF_ERROR(dictID);+ assert(dictID <= (size_t)(U32)-1);+ cdict->dictID = (U32)dictID;+ }+ }++ return 0;+}++ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_compressionParameters cParams, ZSTD_customMem customMem)+{+ DEBUGLOG(3, "ZSTD_createCDict_advanced, mode %u", (unsigned)dictContentType);+ if (!customMem.customAlloc ^ !customMem.customFree) return NULL;++ { ZSTD_CDict* const cdict = (ZSTD_CDict*)ZSTD_malloc(sizeof(ZSTD_CDict), customMem);+ size_t const workspaceSize = HUF_WORKSPACE_SIZE + ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0);+ void* const workspace = ZSTD_malloc(workspaceSize, customMem);++ if (!cdict || !workspace) {+ ZSTD_free(cdict, customMem);+ ZSTD_free(workspace, customMem);+ return NULL;+ }+ cdict->customMem = customMem;+ cdict->workspace = workspace;+ cdict->workspaceSize = workspaceSize;+ if (ZSTD_isError( ZSTD_initCDict_internal(cdict,+ dictBuffer, dictSize,+ dictLoadMethod, dictContentType,+ cParams) )) {+ ZSTD_freeCDict(cdict);+ return NULL;+ }++ return cdict;+ }+}++ZSTD_CDict* ZSTD_createCDict(const void* dict, size_t dictSize, int compressionLevel)+{+ ZSTD_compressionParameters cParams = ZSTD_getCParams(compressionLevel, 0, dictSize);+ return ZSTD_createCDict_advanced(dict, dictSize,+ ZSTD_dlm_byCopy, ZSTD_dct_auto,+ cParams, ZSTD_defaultCMem);+}++ZSTD_CDict* ZSTD_createCDict_byReference(const void* dict, size_t dictSize, int compressionLevel)+{+ ZSTD_compressionParameters cParams = ZSTD_getCParams(compressionLevel, 0, dictSize);+ return ZSTD_createCDict_advanced(dict, dictSize,+ ZSTD_dlm_byRef, ZSTD_dct_auto,+ cParams, ZSTD_defaultCMem);+}++size_t ZSTD_freeCDict(ZSTD_CDict* cdict)+{+ if (cdict==NULL) return 0; /* support free on NULL */+ { ZSTD_customMem const cMem = cdict->customMem;+ ZSTD_free(cdict->workspace, cMem);+ ZSTD_free(cdict->dictBuffer, cMem);+ ZSTD_free(cdict, cMem);+ return 0;+ }+}++/*! ZSTD_initStaticCDict_advanced() :+ * Generate a digested dictionary in provided memory area.+ * workspace: The memory area to emplace the dictionary into.+ * Provided pointer must 8-bytes aligned.+ * It must outlive dictionary usage.+ * workspaceSize: Use ZSTD_estimateCDictSize()+ * to determine how large workspace must be.+ * cParams : use ZSTD_getCParams() to transform a compression level+ * into its relevants cParams.+ * @return : pointer to ZSTD_CDict*, or NULL if error (size too small)+ * Note : there is no corresponding "free" function.+ * Since workspace was allocated externally, it must be freed externally.+ */+const ZSTD_CDict* ZSTD_initStaticCDict(+ void* workspace, size_t workspaceSize,+ const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_compressionParameters cParams)+{+ size_t const matchStateSize = ZSTD_sizeof_matchState(&cParams, /* forCCtx */ 0);+ size_t const neededSize = sizeof(ZSTD_CDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize)+ + HUF_WORKSPACE_SIZE + matchStateSize;+ ZSTD_CDict* const cdict = (ZSTD_CDict*) workspace;+ void* ptr;+ if ((size_t)workspace & 7) return NULL; /* 8-aligned */+ DEBUGLOG(4, "(workspaceSize < neededSize) : (%u < %u) => %u",+ (unsigned)workspaceSize, (unsigned)neededSize, (unsigned)(workspaceSize < neededSize));+ if (workspaceSize < neededSize) return NULL;++ if (dictLoadMethod == ZSTD_dlm_byCopy) {+ memcpy(cdict+1, dict, dictSize);+ dict = cdict+1;+ ptr = (char*)workspace + sizeof(ZSTD_CDict) + dictSize;+ } else {+ ptr = cdict+1;+ }+ cdict->workspace = ptr;+ cdict->workspaceSize = HUF_WORKSPACE_SIZE + matchStateSize;++ if (ZSTD_isError( ZSTD_initCDict_internal(cdict,+ dict, dictSize,+ ZSTD_dlm_byRef, dictContentType,+ cParams) ))+ return NULL;++ return cdict;+}++ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict)+{+ assert(cdict != NULL);+ return cdict->matchState.cParams;+}++/* ZSTD_compressBegin_usingCDict_advanced() :+ * cdict must be != NULL */+size_t ZSTD_compressBegin_usingCDict_advanced(+ ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict,+ ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_compressBegin_usingCDict_advanced");+ RETURN_ERROR_IF(cdict==NULL, dictionary_wrong);+ { ZSTD_CCtx_params params = cctx->requestedParams;+ params.cParams = ZSTD_getCParamsFromCDict(cdict);+ /* Increase window log to fit the entire dictionary and source if the+ * source size is known. Limit the increase to 19, which is the+ * window log for compression level 1 with the largest source size.+ */+ if (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN) {+ U32 const limitedSrcSize = (U32)MIN(pledgedSrcSize, 1U << 19);+ U32 const limitedSrcLog = limitedSrcSize > 1 ? ZSTD_highbit32(limitedSrcSize - 1) + 1 : 1;+ params.cParams.windowLog = MAX(params.cParams.windowLog, limitedSrcLog);+ }+ params.fParams = fParams;+ return ZSTD_compressBegin_internal(cctx,+ NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast,+ cdict,+ params, pledgedSrcSize,+ ZSTDb_not_buffered);+ }+}++/* ZSTD_compressBegin_usingCDict() :+ * pledgedSrcSize=0 means "unknown"+ * if pledgedSrcSize>0, it will enable contentSizeFlag */+size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)+{+ ZSTD_frameParameters const fParams = { 0 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };+ DEBUGLOG(4, "ZSTD_compressBegin_usingCDict : dictIDFlag == %u", !fParams.noDictIDFlag);+ return ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN);+}++size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict, ZSTD_frameParameters fParams)+{+ FORWARD_IF_ERROR(ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize)); /* will check if cdict != NULL */+ return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);+}++/*! ZSTD_compress_usingCDict() :+ * Compression using a digested Dictionary.+ * Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.+ * Note that compression parameters are decided at CDict creation time+ * while frame parameters are hardcoded */+size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict)+{+ ZSTD_frameParameters const fParams = { 1 /*content*/, 0 /*checksum*/, 0 /*noDictID*/ };+ return ZSTD_compress_usingCDict_advanced(cctx, dst, dstCapacity, src, srcSize, cdict, fParams);+}++++/* ******************************************************************+* Streaming+********************************************************************/++ZSTD_CStream* ZSTD_createCStream(void)+{+ DEBUGLOG(3, "ZSTD_createCStream");+ return ZSTD_createCStream_advanced(ZSTD_defaultCMem);+}++ZSTD_CStream* ZSTD_initStaticCStream(void *workspace, size_t workspaceSize)+{+ return ZSTD_initStaticCCtx(workspace, workspaceSize);+}++ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem)+{ /* CStream and CCtx are now same object */+ return ZSTD_createCCtx_advanced(customMem);+}++size_t ZSTD_freeCStream(ZSTD_CStream* zcs)+{+ return ZSTD_freeCCtx(zcs); /* same object */+}++++/*====== Initialization ======*/++size_t ZSTD_CStreamInSize(void) { return ZSTD_BLOCKSIZE_MAX; }++size_t ZSTD_CStreamOutSize(void)+{+ return ZSTD_compressBound(ZSTD_BLOCKSIZE_MAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */ ;+}++static size_t ZSTD_resetCStream_internal(ZSTD_CStream* cctx,+ const void* const dict, size_t const dictSize, ZSTD_dictContentType_e const dictContentType,+ const ZSTD_CDict* const cdict,+ ZSTD_CCtx_params params, unsigned long long const pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_resetCStream_internal");+ /* Finalize the compression parameters */+ params.cParams = ZSTD_getCParamsFromCCtxParams(¶ms, pledgedSrcSize, dictSize);+ /* params are supposed to be fully validated at this point */+ assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));+ assert(!((dict) && (cdict))); /* either dict or cdict, not both */++ FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,+ dict, dictSize, dictContentType, ZSTD_dtlm_fast,+ cdict,+ params, pledgedSrcSize,+ ZSTDb_buffered) );++ cctx->inToCompress = 0;+ cctx->inBuffPos = 0;+ cctx->inBuffTarget = cctx->blockSize+ + (cctx->blockSize == pledgedSrcSize); /* for small input: avoid automatic flush on reaching end of block, since it would require to add a 3-bytes null block to end frame */+ cctx->outBuffContentSize = cctx->outBuffFlushedSize = 0;+ cctx->streamStage = zcss_load;+ cctx->frameEnded = 0;+ return 0; /* ready to go */+}++/* ZSTD_resetCStream():+ * pledgedSrcSize == 0 means "unknown" */+size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pss)+{+ /* temporary : 0 interpreted as "unknown" during transition period.+ * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.+ * 0 will be interpreted as "empty" in the future.+ */+ U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;+ DEBUGLOG(4, "ZSTD_resetCStream: pledgedSrcSize = %u", (unsigned)pledgedSrcSize);+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );+ return 0;+}++/*! ZSTD_initCStream_internal() :+ * Note : for lib/compress only. Used by zstdmt_compress.c.+ * Assumption 1 : params are valid+ * Assumption 2 : either dict, or cdict, is defined, not both */+size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,+ const void* dict, size_t dictSize, const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params, unsigned long long pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_initCStream_internal");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );+ assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));+ zcs->requestedParams = params;+ assert(!((dict) && (cdict))); /* either dict or cdict, not both */+ if (dict) {+ FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );+ } else {+ /* Dictionary is cleared if !cdict */+ FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );+ }+ return 0;+}++/* ZSTD_initCStream_usingCDict_advanced() :+ * same as ZSTD_initCStream_usingCDict(), with control over frame parameters */+size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,+ const ZSTD_CDict* cdict,+ ZSTD_frameParameters fParams,+ unsigned long long pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTD_initCStream_usingCDict_advanced");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );+ zcs->requestedParams.fParams = fParams;+ FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );+ return 0;+}++/* note : cdict must outlive compression session */+size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict)+{+ DEBUGLOG(4, "ZSTD_initCStream_usingCDict");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );+ return 0;+}+++/* ZSTD_initCStream_advanced() :+ * pledgedSrcSize must be exact.+ * if srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN.+ * dict is loaded with default parameters ZSTD_dm_auto and ZSTD_dlm_byCopy. */+size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,+ const void* dict, size_t dictSize,+ ZSTD_parameters params, unsigned long long pss)+{+ /* for compatibility with older programs relying on this behavior.+ * Users should now specify ZSTD_CONTENTSIZE_UNKNOWN.+ * This line will be removed in the future.+ */+ U64 const pledgedSrcSize = (pss==0 && params.fParams.contentSizeFlag==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;+ DEBUGLOG(4, "ZSTD_initCStream_advanced");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );+ FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );+ zcs->requestedParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);+ FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );+ return 0;+}++size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)+{+ DEBUGLOG(4, "ZSTD_initCStream_usingDict");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) );+ FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );+ return 0;+}++size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pss)+{+ /* temporary : 0 interpreted as "unknown" during transition period.+ * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.+ * 0 will be interpreted as "empty" in the future.+ */+ U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;+ DEBUGLOG(4, "ZSTD_initCStream_srcSize");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );+ return 0;+}++size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)+{+ DEBUGLOG(4, "ZSTD_initCStream");+ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) );+ FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) );+ return 0;+}++/*====== Compression ======*/++static size_t ZSTD_nextInputSizeHint(const ZSTD_CCtx* cctx)+{+ size_t hintInSize = cctx->inBuffTarget - cctx->inBuffPos;+ if (hintInSize==0) hintInSize = cctx->blockSize;+ return hintInSize;+}++static size_t ZSTD_limitCopy(void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ size_t const length = MIN(dstCapacity, srcSize);+ if (length) memcpy(dst, src, length);+ return length;+}++/** ZSTD_compressStream_generic():+ * internal function for all *compressStream*() variants+ * non-static, because can be called from zstdmt_compress.c+ * @return : hint size for next input */+static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,+ ZSTD_outBuffer* output,+ ZSTD_inBuffer* input,+ ZSTD_EndDirective const flushMode)+{+ const char* const istart = (const char*)input->src;+ const char* const iend = istart + input->size;+ const char* ip = istart + input->pos;+ char* const ostart = (char*)output->dst;+ char* const oend = ostart + output->size;+ char* op = ostart + output->pos;+ U32 someMoreWork = 1;++ /* check expectations */+ DEBUGLOG(5, "ZSTD_compressStream_generic, flush=%u", (unsigned)flushMode);+ assert(zcs->inBuff != NULL);+ assert(zcs->inBuffSize > 0);+ assert(zcs->outBuff != NULL);+ assert(zcs->outBuffSize > 0);+ assert(output->pos <= output->size);+ assert(input->pos <= input->size);++ while (someMoreWork) {+ switch(zcs->streamStage)+ {+ case zcss_init:+ RETURN_ERROR(init_missing, "call ZSTD_initCStream() first!");++ case zcss_load:+ if ( (flushMode == ZSTD_e_end)+ && ((size_t)(oend-op) >= ZSTD_compressBound(iend-ip)) /* enough dstCapacity */+ && (zcs->inBuffPos == 0) ) {+ /* shortcut to compression pass directly into output buffer */+ size_t const cSize = ZSTD_compressEnd(zcs,+ op, oend-op, ip, iend-ip);+ DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize);+ FORWARD_IF_ERROR(cSize);+ ip = iend;+ op += cSize;+ zcs->frameEnded = 1;+ ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ someMoreWork = 0; break;+ }+ /* complete loading into inBuffer */+ { size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos;+ size_t const loaded = ZSTD_limitCopy(+ zcs->inBuff + zcs->inBuffPos, toLoad,+ ip, iend-ip);+ zcs->inBuffPos += loaded;+ ip += loaded;+ if ( (flushMode == ZSTD_e_continue)+ && (zcs->inBuffPos < zcs->inBuffTarget) ) {+ /* not enough input to fill full block : stop here */+ someMoreWork = 0; break;+ }+ if ( (flushMode == ZSTD_e_flush)+ && (zcs->inBuffPos == zcs->inToCompress) ) {+ /* empty */+ someMoreWork = 0; break;+ }+ }+ /* compress current block (note : this stage cannot be stopped in the middle) */+ DEBUGLOG(5, "stream compression stage (flushMode==%u)", flushMode);+ { void* cDst;+ size_t cSize;+ size_t const iSize = zcs->inBuffPos - zcs->inToCompress;+ size_t oSize = oend-op;+ unsigned const lastBlock = (flushMode == ZSTD_e_end) && (ip==iend);+ if (oSize >= ZSTD_compressBound(iSize))+ cDst = op; /* compress into output buffer, to skip flush stage */+ else+ cDst = zcs->outBuff, oSize = zcs->outBuffSize;+ cSize = lastBlock ?+ ZSTD_compressEnd(zcs, cDst, oSize,+ zcs->inBuff + zcs->inToCompress, iSize) :+ ZSTD_compressContinue(zcs, cDst, oSize,+ zcs->inBuff + zcs->inToCompress, iSize);+ FORWARD_IF_ERROR(cSize);+ zcs->frameEnded = lastBlock;+ /* prepare next block */+ zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;+ if (zcs->inBuffTarget > zcs->inBuffSize)+ zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize;+ DEBUGLOG(5, "inBuffTarget:%u / inBuffSize:%u",+ (unsigned)zcs->inBuffTarget, (unsigned)zcs->inBuffSize);+ if (!lastBlock)+ assert(zcs->inBuffTarget <= zcs->inBuffSize);+ zcs->inToCompress = zcs->inBuffPos;+ if (cDst == op) { /* no need to flush */+ op += cSize;+ if (zcs->frameEnded) {+ DEBUGLOG(5, "Frame completed directly in outBuffer");+ someMoreWork = 0;+ ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ }+ break;+ }+ zcs->outBuffContentSize = cSize;+ zcs->outBuffFlushedSize = 0;+ zcs->streamStage = zcss_flush; /* pass-through to flush stage */+ }+ /* fall-through */+ case zcss_flush:+ DEBUGLOG(5, "flush stage");+ { size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;+ size_t const flushed = ZSTD_limitCopy(op, oend-op,+ zcs->outBuff + zcs->outBuffFlushedSize, toFlush);+ DEBUGLOG(5, "toFlush: %u into %u ==> flushed: %u",+ (unsigned)toFlush, (unsigned)(oend-op), (unsigned)flushed);+ op += flushed;+ zcs->outBuffFlushedSize += flushed;+ if (toFlush!=flushed) {+ /* flush not fully completed, presumably because dst is too small */+ assert(op==oend);+ someMoreWork = 0;+ break;+ }+ zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;+ if (zcs->frameEnded) {+ DEBUGLOG(5, "Frame completed on flush");+ someMoreWork = 0;+ ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ break;+ }+ zcs->streamStage = zcss_load;+ break;+ }++ default: /* impossible */+ assert(0);+ }+ }++ input->pos = ip - istart;+ output->pos = op - ostart;+ if (zcs->frameEnded) return 0;+ return ZSTD_nextInputSizeHint(zcs);+}++static size_t ZSTD_nextInputSizeHint_MTorST(const ZSTD_CCtx* cctx)+{+#ifdef ZSTD_MULTITHREAD+ if (cctx->appliedParams.nbWorkers >= 1) {+ assert(cctx->mtctx != NULL);+ return ZSTDMT_nextInputSizeHint(cctx->mtctx);+ }+#endif+ return ZSTD_nextInputSizeHint(cctx);++}++size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input)+{+ FORWARD_IF_ERROR( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) );+ return ZSTD_nextInputSizeHint_MTorST(zcs);+}+++size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,+ ZSTD_outBuffer* output,+ ZSTD_inBuffer* input,+ ZSTD_EndDirective endOp)+{+ DEBUGLOG(5, "ZSTD_compressStream2, endOp=%u ", (unsigned)endOp);+ /* check conditions */+ RETURN_ERROR_IF(output->pos > output->size, GENERIC);+ RETURN_ERROR_IF(input->pos > input->size, GENERIC);+ assert(cctx!=NULL);++ /* transparent initialization stage */+ if (cctx->streamStage == zcss_init) {+ ZSTD_CCtx_params params = cctx->requestedParams;+ ZSTD_prefixDict const prefixDict = cctx->prefixDict;+ FORWARD_IF_ERROR( ZSTD_initLocalDict(cctx) ); /* Init the local dict if present. */+ memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); /* single usage */+ assert(prefixDict.dict==NULL || cctx->cdict==NULL); /* only one can be set */+ DEBUGLOG(4, "ZSTD_compressStream2 : transparent init stage");+ if (endOp == ZSTD_e_end) cctx->pledgedSrcSizePlusOne = input->size + 1; /* auto-fix pledgedSrcSize */+ params.cParams = ZSTD_getCParamsFromCCtxParams(+ &cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/);+++#ifdef ZSTD_MULTITHREAD+ if ((cctx->pledgedSrcSizePlusOne-1) <= ZSTDMT_JOBSIZE_MIN) {+ params.nbWorkers = 0; /* do not invoke multi-threading when src size is too small */+ }+ if (params.nbWorkers > 0) {+ /* mt context creation */+ if (cctx->mtctx == NULL) {+ DEBUGLOG(4, "ZSTD_compressStream2: creating new mtctx for nbWorkers=%u",+ params.nbWorkers);+ cctx->mtctx = ZSTDMT_createCCtx_advanced(params.nbWorkers, cctx->customMem);+ RETURN_ERROR_IF(cctx->mtctx == NULL, memory_allocation);+ }+ /* mt compression */+ DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbWorkers=%u", params.nbWorkers);+ FORWARD_IF_ERROR( ZSTDMT_initCStream_internal(+ cctx->mtctx,+ prefixDict.dict, prefixDict.dictSize, ZSTD_dct_rawContent,+ cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) );+ cctx->streamStage = zcss_load;+ cctx->appliedParams.nbWorkers = params.nbWorkers;+ } else+#endif+ { FORWARD_IF_ERROR( ZSTD_resetCStream_internal(cctx,+ prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType,+ cctx->cdict,+ params, cctx->pledgedSrcSizePlusOne-1) );+ assert(cctx->streamStage == zcss_load);+ assert(cctx->appliedParams.nbWorkers == 0);+ } }+ /* end of transparent initialization stage */++ /* compression stage */+#ifdef ZSTD_MULTITHREAD+ if (cctx->appliedParams.nbWorkers > 0) {+ int const forceMaxProgress = (endOp == ZSTD_e_flush || endOp == ZSTD_e_end);+ size_t flushMin;+ assert(forceMaxProgress || endOp == ZSTD_e_continue /* Protection for a new flush type */);+ if (cctx->cParamsChanged) {+ ZSTDMT_updateCParams_whileCompressing(cctx->mtctx, &cctx->requestedParams);+ cctx->cParamsChanged = 0;+ }+ do {+ flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp);+ if ( ZSTD_isError(flushMin)+ || (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */+ ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only);+ }+ FORWARD_IF_ERROR(flushMin);+ } while (forceMaxProgress && flushMin != 0 && output->pos < output->size);+ DEBUGLOG(5, "completed ZSTD_compressStream2 delegating to ZSTDMT_compressStream_generic");+ /* Either we don't require maximum forward progress, we've finished the+ * flush, or we are out of output space.+ */+ assert(!forceMaxProgress || flushMin == 0 || output->pos == output->size);+ return flushMin;+ }+#endif+ FORWARD_IF_ERROR( ZSTD_compressStream_generic(cctx, output, input, endOp) );+ DEBUGLOG(5, "completed ZSTD_compressStream2");+ return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */+}++size_t ZSTD_compressStream2_simpleArgs (+ ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity, size_t* dstPos,+ const void* src, size_t srcSize, size_t* srcPos,+ ZSTD_EndDirective endOp)+{+ ZSTD_outBuffer output = { dst, dstCapacity, *dstPos };+ ZSTD_inBuffer input = { src, srcSize, *srcPos };+ /* ZSTD_compressStream2() will check validity of dstPos and srcPos */+ size_t const cErr = ZSTD_compressStream2(cctx, &output, &input, endOp);+ *dstPos = output.pos;+ *srcPos = input.pos;+ return cErr;+}++size_t ZSTD_compress2(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only);+ { size_t oPos = 0;+ size_t iPos = 0;+ size_t const result = ZSTD_compressStream2_simpleArgs(cctx,+ dst, dstCapacity, &oPos,+ src, srcSize, &iPos,+ ZSTD_e_end);+ FORWARD_IF_ERROR(result);+ if (result != 0) { /* compression not completed, due to lack of output space */+ assert(oPos == dstCapacity);+ RETURN_ERROR(dstSize_tooSmall);+ }+ assert(iPos == srcSize); /* all input is expected consumed */+ return oPos;+ }+}++/*====== Finalize ======*/++/*! ZSTD_flushStream() :+ * @return : amount of data remaining to flush */+size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)+{+ ZSTD_inBuffer input = { NULL, 0, 0 };+ return ZSTD_compressStream2(zcs, output, &input, ZSTD_e_flush);+}+++size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)+{+ ZSTD_inBuffer input = { NULL, 0, 0 };+ size_t const remainingToFlush = ZSTD_compressStream2(zcs, output, &input, ZSTD_e_end);+ FORWARD_IF_ERROR( remainingToFlush );+ if (zcs->appliedParams.nbWorkers > 0) return remainingToFlush; /* minimal estimation */+ /* single thread mode : attempt to calculate remaining to flush more precisely */+ { size_t const lastBlockSize = zcs->frameEnded ? 0 : ZSTD_BLOCKHEADERSIZE;+ size_t const checksumSize = zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4;+ size_t const toFlush = remainingToFlush + lastBlockSize + checksumSize;+ DEBUGLOG(4, "ZSTD_endStream : remaining to flush : %u", (unsigned)toFlush);+ return toFlush;+ }+}+++/*-===== Pre-defined compression levels =====-*/++#define ZSTD_MAX_CLEVEL 22+int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; }+int ZSTD_minCLevel(void) { return (int)-ZSTD_TARGETLENGTH_MAX; }++static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = {+{ /* "default" - for any srcSize > 256 KB */+ /* W, C, H, S, L, TL, strat */+ { 19, 12, 13, 1, 6, 1, ZSTD_fast }, /* base for negative levels */+ { 19, 13, 14, 1, 7, 0, ZSTD_fast }, /* level 1 */+ { 20, 15, 16, 1, 6, 0, ZSTD_fast }, /* level 2 */+ { 21, 16, 17, 1, 5, 1, ZSTD_dfast }, /* level 3 */+ { 21, 18, 18, 1, 5, 1, ZSTD_dfast }, /* level 4 */+ { 21, 18, 19, 2, 5, 2, ZSTD_greedy }, /* level 5 */+ { 21, 19, 19, 3, 5, 4, ZSTD_greedy }, /* level 6 */+ { 21, 19, 19, 3, 5, 8, ZSTD_lazy }, /* level 7 */+ { 21, 19, 19, 3, 5, 16, ZSTD_lazy2 }, /* level 8 */+ { 21, 19, 20, 4, 5, 16, ZSTD_lazy2 }, /* level 9 */+ { 22, 20, 21, 4, 5, 16, ZSTD_lazy2 }, /* level 10 */+ { 22, 21, 22, 4, 5, 16, ZSTD_lazy2 }, /* level 11 */+ { 22, 21, 22, 5, 5, 16, ZSTD_lazy2 }, /* level 12 */+ { 22, 21, 22, 5, 5, 32, ZSTD_btlazy2 }, /* level 13 */+ { 22, 22, 23, 5, 5, 32, ZSTD_btlazy2 }, /* level 14 */+ { 22, 23, 23, 6, 5, 32, ZSTD_btlazy2 }, /* level 15 */+ { 22, 22, 22, 5, 5, 48, ZSTD_btopt }, /* level 16 */+ { 23, 23, 22, 5, 4, 64, ZSTD_btopt }, /* level 17 */+ { 23, 23, 22, 6, 3, 64, ZSTD_btultra }, /* level 18 */+ { 23, 24, 22, 7, 3,256, ZSTD_btultra2}, /* level 19 */+ { 25, 25, 23, 7, 3,256, ZSTD_btultra2}, /* level 20 */+ { 26, 26, 24, 7, 3,512, ZSTD_btultra2}, /* level 21 */+ { 27, 27, 25, 9, 3,999, ZSTD_btultra2}, /* level 22 */+},+{ /* for srcSize <= 256 KB */+ /* W, C, H, S, L, T, strat */+ { 18, 12, 13, 1, 5, 1, ZSTD_fast }, /* base for negative levels */+ { 18, 13, 14, 1, 6, 0, ZSTD_fast }, /* level 1 */+ { 18, 14, 14, 1, 5, 1, ZSTD_dfast }, /* level 2 */+ { 18, 16, 16, 1, 4, 1, ZSTD_dfast }, /* level 3 */+ { 18, 16, 17, 2, 5, 2, ZSTD_greedy }, /* level 4.*/+ { 18, 18, 18, 3, 5, 2, ZSTD_greedy }, /* level 5.*/+ { 18, 18, 19, 3, 5, 4, ZSTD_lazy }, /* level 6.*/+ { 18, 18, 19, 4, 4, 4, ZSTD_lazy }, /* level 7 */+ { 18, 18, 19, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */+ { 18, 18, 19, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */+ { 18, 18, 19, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */+ { 18, 18, 19, 5, 4, 12, ZSTD_btlazy2 }, /* level 11.*/+ { 18, 19, 19, 7, 4, 12, ZSTD_btlazy2 }, /* level 12.*/+ { 18, 18, 19, 4, 4, 16, ZSTD_btopt }, /* level 13 */+ { 18, 18, 19, 4, 3, 32, ZSTD_btopt }, /* level 14.*/+ { 18, 18, 19, 6, 3,128, ZSTD_btopt }, /* level 15.*/+ { 18, 19, 19, 6, 3,128, ZSTD_btultra }, /* level 16.*/+ { 18, 19, 19, 8, 3,256, ZSTD_btultra }, /* level 17.*/+ { 18, 19, 19, 6, 3,128, ZSTD_btultra2}, /* level 18.*/+ { 18, 19, 19, 8, 3,256, ZSTD_btultra2}, /* level 19.*/+ { 18, 19, 19, 10, 3,512, ZSTD_btultra2}, /* level 20.*/+ { 18, 19, 19, 12, 3,512, ZSTD_btultra2}, /* level 21.*/+ { 18, 19, 19, 13, 3,999, ZSTD_btultra2}, /* level 22.*/+},+{ /* for srcSize <= 128 KB */+ /* W, C, H, S, L, T, strat */+ { 17, 12, 12, 1, 5, 1, ZSTD_fast }, /* base for negative levels */+ { 17, 12, 13, 1, 6, 0, ZSTD_fast }, /* level 1 */+ { 17, 13, 15, 1, 5, 0, ZSTD_fast }, /* level 2 */+ { 17, 15, 16, 2, 5, 1, ZSTD_dfast }, /* level 3 */+ { 17, 17, 17, 2, 4, 1, ZSTD_dfast }, /* level 4 */+ { 17, 16, 17, 3, 4, 2, ZSTD_greedy }, /* level 5 */+ { 17, 17, 17, 3, 4, 4, ZSTD_lazy }, /* level 6 */+ { 17, 17, 17, 3, 4, 8, ZSTD_lazy2 }, /* level 7 */+ { 17, 17, 17, 4, 4, 8, ZSTD_lazy2 }, /* level 8 */+ { 17, 17, 17, 5, 4, 8, ZSTD_lazy2 }, /* level 9 */+ { 17, 17, 17, 6, 4, 8, ZSTD_lazy2 }, /* level 10 */+ { 17, 17, 17, 5, 4, 8, ZSTD_btlazy2 }, /* level 11 */+ { 17, 18, 17, 7, 4, 12, ZSTD_btlazy2 }, /* level 12 */+ { 17, 18, 17, 3, 4, 12, ZSTD_btopt }, /* level 13.*/+ { 17, 18, 17, 4, 3, 32, ZSTD_btopt }, /* level 14.*/+ { 17, 18, 17, 6, 3,256, ZSTD_btopt }, /* level 15.*/+ { 17, 18, 17, 6, 3,128, ZSTD_btultra }, /* level 16.*/+ { 17, 18, 17, 8, 3,256, ZSTD_btultra }, /* level 17.*/+ { 17, 18, 17, 10, 3,512, ZSTD_btultra }, /* level 18.*/+ { 17, 18, 17, 5, 3,256, ZSTD_btultra2}, /* level 19.*/+ { 17, 18, 17, 7, 3,512, ZSTD_btultra2}, /* level 20.*/+ { 17, 18, 17, 9, 3,512, ZSTD_btultra2}, /* level 21.*/+ { 17, 18, 17, 11, 3,999, ZSTD_btultra2}, /* level 22.*/+},+{ /* for srcSize <= 16 KB */+ /* W, C, H, S, L, T, strat */+ { 14, 12, 13, 1, 5, 1, ZSTD_fast }, /* base for negative levels */+ { 14, 14, 15, 1, 5, 0, ZSTD_fast }, /* level 1 */+ { 14, 14, 15, 1, 4, 0, ZSTD_fast }, /* level 2 */+ { 14, 14, 15, 2, 4, 1, ZSTD_dfast }, /* level 3 */+ { 14, 14, 14, 4, 4, 2, ZSTD_greedy }, /* level 4 */+ { 14, 14, 14, 3, 4, 4, ZSTD_lazy }, /* level 5.*/+ { 14, 14, 14, 4, 4, 8, ZSTD_lazy2 }, /* level 6 */+ { 14, 14, 14, 6, 4, 8, ZSTD_lazy2 }, /* level 7 */+ { 14, 14, 14, 8, 4, 8, ZSTD_lazy2 }, /* level 8.*/+ { 14, 15, 14, 5, 4, 8, ZSTD_btlazy2 }, /* level 9.*/+ { 14, 15, 14, 9, 4, 8, ZSTD_btlazy2 }, /* level 10.*/+ { 14, 15, 14, 3, 4, 12, ZSTD_btopt }, /* level 11.*/+ { 14, 15, 14, 4, 3, 24, ZSTD_btopt }, /* level 12.*/+ { 14, 15, 14, 5, 3, 32, ZSTD_btultra }, /* level 13.*/+ { 14, 15, 15, 6, 3, 64, ZSTD_btultra }, /* level 14.*/+ { 14, 15, 15, 7, 3,256, ZSTD_btultra }, /* level 15.*/+ { 14, 15, 15, 5, 3, 48, ZSTD_btultra2}, /* level 16.*/+ { 14, 15, 15, 6, 3,128, ZSTD_btultra2}, /* level 17.*/+ { 14, 15, 15, 7, 3,256, ZSTD_btultra2}, /* level 18.*/+ { 14, 15, 15, 8, 3,256, ZSTD_btultra2}, /* level 19.*/+ { 14, 15, 15, 8, 3,512, ZSTD_btultra2}, /* level 20.*/+ { 14, 15, 15, 9, 3,512, ZSTD_btultra2}, /* level 21.*/+ { 14, 15, 15, 10, 3,999, ZSTD_btultra2}, /* level 22.*/+},+};++/*! ZSTD_getCParams() :+ * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize.+ * Size values are optional, provide 0 if not known or unused */+ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize)+{+ size_t const addedSize = srcSizeHint ? 0 : 500;+ U64 const rSize = srcSizeHint+dictSize ? srcSizeHint+dictSize+addedSize : ZSTD_CONTENTSIZE_UNKNOWN; /* intentional overflow for srcSizeHint == ZSTD_CONTENTSIZE_UNKNOWN */+ U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB);+ int row = compressionLevel;+ DEBUGLOG(5, "ZSTD_getCParams (cLevel=%i)", compressionLevel);+ if (compressionLevel == 0) row = ZSTD_CLEVEL_DEFAULT; /* 0 == default */+ if (compressionLevel < 0) row = 0; /* entry 0 is baseline for fast mode */+ if (compressionLevel > ZSTD_MAX_CLEVEL) row = ZSTD_MAX_CLEVEL;+ { ZSTD_compressionParameters cp = ZSTD_defaultCParameters[tableID][row];+ if (compressionLevel < 0) cp.targetLength = (unsigned)(-compressionLevel); /* acceleration factor */+ return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize); /* refine parameters based on srcSize & dictSize */+ }+}++/*! ZSTD_getParams() :+ * same idea as ZSTD_getCParams()+ * @return a `ZSTD_parameters` structure (instead of `ZSTD_compressionParameters`).+ * Fields of `ZSTD_frameParameters` are set to default values */+ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) {+ ZSTD_parameters params;+ ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSizeHint, dictSize);+ DEBUGLOG(5, "ZSTD_getParams (cLevel=%i)", compressionLevel);+ memset(¶ms, 0, sizeof(params));+ params.cParams = cParams;+ params.fParams.contentSizeFlag = 1; return params; }
+ zstd/lib/compress/zstd_compress_internal.h view
@@ -0,0 +1,863 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++/* This header contains definitions+ * that shall **only** be used by modules within lib/compress.+ */++#ifndef ZSTD_COMPRESS_H+#define ZSTD_COMPRESS_H++/*-*************************************+* Dependencies+***************************************/+#include "zstd_internal.h"+#ifdef ZSTD_MULTITHREAD+# include "zstdmt_compress.h"+#endif++#if defined (__cplusplus)+extern "C" {+#endif+++/*-*************************************+* Constants+***************************************/+#define kSearchStrength 8+#define HASH_READ_SIZE 8+#define ZSTD_DUBT_UNSORTED_MARK 1 /* For btlazy2 strategy, index 1 now means "unsorted".+ It could be confused for a real successor at index "1", if sorted as larger than its predecessor.+ It's not a big deal though : candidate will just be sorted again.+ Additionally, candidate position 1 will be lost.+ But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss.+ The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy+ Constant required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */+++/*-*************************************+* Context memory management+***************************************/+typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;+typedef enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage;++typedef struct ZSTD_prefixDict_s {+ const void* dict;+ size_t dictSize;+ ZSTD_dictContentType_e dictContentType;+} ZSTD_prefixDict;++typedef struct {+ void* dictBuffer;+ void const* dict;+ size_t dictSize;+ ZSTD_dictContentType_e dictContentType;+ ZSTD_CDict* cdict;+} ZSTD_localDict;++typedef struct {+ U32 CTable[HUF_CTABLE_SIZE_U32(255)];+ HUF_repeat repeatMode;+} ZSTD_hufCTables_t;++typedef struct {+ FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];+ FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];+ FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];+ FSE_repeat offcode_repeatMode;+ FSE_repeat matchlength_repeatMode;+ FSE_repeat litlength_repeatMode;+} ZSTD_fseCTables_t;++typedef struct {+ ZSTD_hufCTables_t huf;+ ZSTD_fseCTables_t fse;+} ZSTD_entropyCTables_t;++typedef struct {+ U32 off;+ U32 len;+} ZSTD_match_t;++typedef struct {+ int price;+ U32 off;+ U32 mlen;+ U32 litlen;+ U32 rep[ZSTD_REP_NUM];+} ZSTD_optimal_t;++typedef enum { zop_dynamic=0, zop_predef } ZSTD_OptPrice_e;++typedef struct {+ /* All tables are allocated inside cctx->workspace by ZSTD_resetCCtx_internal() */+ unsigned* litFreq; /* table of literals statistics, of size 256 */+ unsigned* litLengthFreq; /* table of litLength statistics, of size (MaxLL+1) */+ unsigned* matchLengthFreq; /* table of matchLength statistics, of size (MaxML+1) */+ unsigned* offCodeFreq; /* table of offCode statistics, of size (MaxOff+1) */+ ZSTD_match_t* matchTable; /* list of found matches, of size ZSTD_OPT_NUM+1 */+ ZSTD_optimal_t* priceTable; /* All positions tracked by optimal parser, of size ZSTD_OPT_NUM+1 */++ U32 litSum; /* nb of literals */+ U32 litLengthSum; /* nb of litLength codes */+ U32 matchLengthSum; /* nb of matchLength codes */+ U32 offCodeSum; /* nb of offset codes */+ U32 litSumBasePrice; /* to compare to log2(litfreq) */+ U32 litLengthSumBasePrice; /* to compare to log2(llfreq) */+ U32 matchLengthSumBasePrice;/* to compare to log2(mlfreq) */+ U32 offCodeSumBasePrice; /* to compare to log2(offreq) */+ ZSTD_OptPrice_e priceType; /* prices can be determined dynamically, or follow a pre-defined cost structure */+ const ZSTD_entropyCTables_t* symbolCosts; /* pre-calculated dictionary statistics */+ ZSTD_literalCompressionMode_e literalCompressionMode;+} optState_t;++typedef struct {+ ZSTD_entropyCTables_t entropy;+ U32 rep[ZSTD_REP_NUM];+} ZSTD_compressedBlockState_t;++typedef struct {+ BYTE const* nextSrc; /* next block here to continue on current prefix */+ BYTE const* base; /* All regular indexes relative to this position */+ BYTE const* dictBase; /* extDict indexes relative to this position */+ U32 dictLimit; /* below that point, need extDict */+ U32 lowLimit; /* below that point, no more data */+} ZSTD_window_t;++typedef struct ZSTD_matchState_t ZSTD_matchState_t;+struct ZSTD_matchState_t {+ ZSTD_window_t window; /* State for window round buffer management */+ U32 loadedDictEnd; /* index of end of dictionary */+ U32 nextToUpdate; /* index from which to continue table update */+ U32 nextToUpdate3; /* index from which to continue table update */+ U32 hashLog3; /* dispatch table : larger == faster, more memory */+ U32* hashTable;+ U32* hashTable3;+ U32* chainTable;+ optState_t opt; /* optimal parser state */+ const ZSTD_matchState_t * dictMatchState;+ ZSTD_compressionParameters cParams;+};++typedef struct {+ ZSTD_compressedBlockState_t* prevCBlock;+ ZSTD_compressedBlockState_t* nextCBlock;+ ZSTD_matchState_t matchState;+} ZSTD_blockState_t;++typedef struct {+ U32 offset;+ U32 checksum;+} ldmEntry_t;++typedef struct {+ ZSTD_window_t window; /* State for the window round buffer management */+ ldmEntry_t* hashTable;+ BYTE* bucketOffsets; /* Next position in bucket to insert entry */+ U64 hashPower; /* Used to compute the rolling hash.+ * Depends on ldmParams.minMatchLength */+} ldmState_t;++typedef struct {+ U32 enableLdm; /* 1 if enable long distance matching */+ U32 hashLog; /* Log size of hashTable */+ U32 bucketSizeLog; /* Log bucket size for collision resolution, at most 8 */+ U32 minMatchLength; /* Minimum match length */+ U32 hashRateLog; /* Log number of entries to skip */+ U32 windowLog; /* Window log for the LDM */+} ldmParams_t;++typedef struct {+ U32 offset;+ U32 litLength;+ U32 matchLength;+} rawSeq;++typedef struct {+ rawSeq* seq; /* The start of the sequences */+ size_t pos; /* The position where reading stopped. <= size. */+ size_t size; /* The number of sequences. <= capacity. */+ size_t capacity; /* The capacity starting from `seq` pointer */+} rawSeqStore_t;++struct ZSTD_CCtx_params_s {+ ZSTD_format_e format;+ ZSTD_compressionParameters cParams;+ ZSTD_frameParameters fParams;++ int compressionLevel;+ int forceWindow; /* force back-references to respect limit of+ * 1<<wLog, even for dictionary */++ ZSTD_dictAttachPref_e attachDictPref;+ ZSTD_literalCompressionMode_e literalCompressionMode;++ /* Multithreading: used to pass parameters to mtctx */+ int nbWorkers;+ size_t jobSize;+ int overlapLog;+ int rsyncable;++ /* Long distance matching parameters */+ ldmParams_t ldmParams;++ /* Internal use, for createCCtxParams() and freeCCtxParams() only */+ ZSTD_customMem customMem;+}; /* typedef'd to ZSTD_CCtx_params within "zstd.h" */++struct ZSTD_CCtx_s {+ ZSTD_compressionStage_e stage;+ int cParamsChanged; /* == 1 if cParams(except wlog) or compression level are changed in requestedParams. Triggers transmission of new params to ZSTDMT (if available) then reset to 0. */+ int bmi2; /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */+ ZSTD_CCtx_params requestedParams;+ ZSTD_CCtx_params appliedParams;+ U32 dictID;++ int workSpaceOversizedDuration;+ void* workSpace;+ size_t workSpaceSize;+ size_t blockSize;+ unsigned long long pledgedSrcSizePlusOne; /* this way, 0 (default) == unknown */+ unsigned long long consumedSrcSize;+ unsigned long long producedCSize;+ XXH64_state_t xxhState;+ ZSTD_customMem customMem;+ size_t staticSize;++ seqStore_t seqStore; /* sequences storage ptrs */+ ldmState_t ldmState; /* long distance matching state */+ rawSeq* ldmSequences; /* Storage for the ldm output sequences */+ size_t maxNbLdmSequences;+ rawSeqStore_t externSeqStore; /* Mutable reference to external sequences */+ ZSTD_blockState_t blockState;+ U32* entropyWorkspace; /* entropy workspace of HUF_WORKSPACE_SIZE bytes */++ /* streaming */+ char* inBuff;+ size_t inBuffSize;+ size_t inToCompress;+ size_t inBuffPos;+ size_t inBuffTarget;+ char* outBuff;+ size_t outBuffSize;+ size_t outBuffContentSize;+ size_t outBuffFlushedSize;+ ZSTD_cStreamStage streamStage;+ U32 frameEnded;++ /* Dictionary */+ ZSTD_localDict localDict;+ const ZSTD_CDict* cdict;+ ZSTD_prefixDict prefixDict; /* single-usage dictionary */++ /* Multi-threading */+#ifdef ZSTD_MULTITHREAD+ ZSTDMT_CCtx* mtctx;+#endif+};++typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e;++typedef enum { ZSTD_noDict = 0, ZSTD_extDict = 1, ZSTD_dictMatchState = 2 } ZSTD_dictMode_e;+++typedef size_t (*ZSTD_blockCompressor) (+ ZSTD_matchState_t* bs, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_dictMode_e dictMode);+++MEM_STATIC U32 ZSTD_LLcode(U32 litLength)+{+ static const BYTE LL_Code[64] = { 0, 1, 2, 3, 4, 5, 6, 7,+ 8, 9, 10, 11, 12, 13, 14, 15,+ 16, 16, 17, 17, 18, 18, 19, 19,+ 20, 20, 20, 20, 21, 21, 21, 21,+ 22, 22, 22, 22, 22, 22, 22, 22,+ 23, 23, 23, 23, 23, 23, 23, 23,+ 24, 24, 24, 24, 24, 24, 24, 24,+ 24, 24, 24, 24, 24, 24, 24, 24 };+ static const U32 LL_deltaCode = 19;+ return (litLength > 63) ? ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];+}++/* ZSTD_MLcode() :+ * note : mlBase = matchLength - MINMATCH;+ * because it's the format it's stored in seqStore->sequences */+MEM_STATIC U32 ZSTD_MLcode(U32 mlBase)+{+ static const BYTE ML_Code[128] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,+ 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,+ 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37,+ 38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39,+ 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,+ 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41,+ 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42,+ 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 };+ static const U32 ML_deltaCode = 36;+ return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase];+}++/*! ZSTD_storeSeq() :+ * Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.+ * `offsetCode` : distance to match + 3 (values 1-3 are repCodes).+ * `mlBase` : matchLength - MINMATCH+*/+MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t mlBase)+{+#if defined(DEBUGLEVEL) && (DEBUGLEVEL >= 6)+ static const BYTE* g_start = NULL;+ if (g_start==NULL) g_start = (const BYTE*)literals; /* note : index only works for compression within a single segment */+ { U32 const pos = (U32)((const BYTE*)literals - g_start);+ DEBUGLOG(6, "Cpos%7u :%3u literals, match%4u bytes at offCode%7u",+ pos, (U32)litLength, (U32)mlBase+MINMATCH, (U32)offsetCode);+ }+#endif+ assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq);+ /* copy Literals */+ assert(seqStorePtr->maxNbLit <= 128 KB);+ assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit);+ ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);+ seqStorePtr->lit += litLength;++ /* literal Length */+ if (litLength>0xFFFF) {+ assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */+ seqStorePtr->longLengthID = 1;+ seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);+ }+ seqStorePtr->sequences[0].litLength = (U16)litLength;++ /* match offset */+ seqStorePtr->sequences[0].offset = offsetCode + 1;++ /* match Length */+ if (mlBase>0xFFFF) {+ assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */+ seqStorePtr->longLengthID = 2;+ seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);+ }+ seqStorePtr->sequences[0].matchLength = (U16)mlBase;++ seqStorePtr->sequences++;+}+++/*-*************************************+* Match length counter+***************************************/+static unsigned ZSTD_NbCommonBytes (size_t val)+{+ if (MEM_isLittleEndian()) {+ if (MEM_64bits()) {+# if defined(_MSC_VER) && defined(_WIN64)+ unsigned long r = 0;+ _BitScanForward64( &r, (U64)val );+ return (unsigned)(r>>3);+# elif defined(__GNUC__) && (__GNUC__ >= 4)+ return (__builtin_ctzll((U64)val) >> 3);+# else+ static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2,+ 0, 3, 1, 3, 1, 4, 2, 7,+ 0, 2, 3, 6, 1, 5, 3, 5,+ 1, 3, 4, 4, 2, 5, 6, 7,+ 7, 0, 1, 2, 3, 3, 4, 6,+ 2, 6, 5, 5, 3, 4, 5, 6,+ 7, 1, 2, 4, 6, 4, 4, 5,+ 7, 2, 6, 5, 7, 6, 7, 7 };+ return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58];+# endif+ } else { /* 32 bits */+# if defined(_MSC_VER)+ unsigned long r=0;+ _BitScanForward( &r, (U32)val );+ return (unsigned)(r>>3);+# elif defined(__GNUC__) && (__GNUC__ >= 3)+ return (__builtin_ctz((U32)val) >> 3);+# else+ static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0,+ 3, 2, 2, 1, 3, 2, 0, 1,+ 3, 3, 1, 2, 2, 2, 2, 0,+ 3, 1, 2, 0, 1, 0, 1, 1 };+ return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];+# endif+ }+ } else { /* Big Endian CPU */+ if (MEM_64bits()) {+# if defined(_MSC_VER) && defined(_WIN64)+ unsigned long r = 0;+ _BitScanReverse64( &r, val );+ return (unsigned)(r>>3);+# elif defined(__GNUC__) && (__GNUC__ >= 4)+ return (__builtin_clzll(val) >> 3);+# else+ unsigned r;+ const unsigned n32 = sizeof(size_t)*4; /* calculate this way due to compiler complaining in 32-bits mode */+ if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; }+ if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }+ r += (!val);+ return r;+# endif+ } else { /* 32 bits */+# if defined(_MSC_VER)+ unsigned long r = 0;+ _BitScanReverse( &r, (unsigned long)val );+ return (unsigned)(r>>3);+# elif defined(__GNUC__) && (__GNUC__ >= 3)+ return (__builtin_clz((U32)val) >> 3);+# else+ unsigned r;+ if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }+ r += (!val);+ return r;+# endif+ } }+}+++MEM_STATIC size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit)+{+ const BYTE* const pStart = pIn;+ const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1);++ if (pIn < pInLoopLimit) {+ { size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);+ if (diff) return ZSTD_NbCommonBytes(diff); }+ pIn+=sizeof(size_t); pMatch+=sizeof(size_t);+ while (pIn < pInLoopLimit) {+ size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);+ if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; }+ pIn += ZSTD_NbCommonBytes(diff);+ return (size_t)(pIn - pStart);+ } }+ if (MEM_64bits() && (pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; }+ if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; }+ if ((pIn<pInLimit) && (*pMatch == *pIn)) pIn++;+ return (size_t)(pIn - pStart);+}++/** ZSTD_count_2segments() :+ * can count match length with `ip` & `match` in 2 different segments.+ * convention : on reaching mEnd, match count continue starting from iStart+ */+MEM_STATIC size_t+ZSTD_count_2segments(const BYTE* ip, const BYTE* match,+ const BYTE* iEnd, const BYTE* mEnd, const BYTE* iStart)+{+ const BYTE* const vEnd = MIN( ip + (mEnd - match), iEnd);+ size_t const matchLength = ZSTD_count(ip, match, vEnd);+ if (match + matchLength != mEnd) return matchLength;+ DEBUGLOG(7, "ZSTD_count_2segments: found a 2-parts match (current length==%zu)", matchLength);+ DEBUGLOG(7, "distance from match beginning to end dictionary = %zi", mEnd - match);+ DEBUGLOG(7, "distance from current pos to end buffer = %zi", iEnd - ip);+ DEBUGLOG(7, "next byte : ip==%02X, istart==%02X", ip[matchLength], *iStart);+ DEBUGLOG(7, "final match length = %zu", matchLength + ZSTD_count(ip+matchLength, iStart, iEnd));+ return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd);+}+++/*-*************************************+ * Hashes+ ***************************************/+static const U32 prime3bytes = 506832829U;+static U32 ZSTD_hash3(U32 u, U32 h) { return ((u << (32-24)) * prime3bytes) >> (32-h) ; }+MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h); } /* only in zstd_opt.h */++static const U32 prime4bytes = 2654435761U;+static U32 ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32-h) ; }+static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_read32(ptr), h); }++static const U64 prime5bytes = 889523592379ULL;+static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u << (64-40)) * prime5bytes) >> (64-h)) ; }+static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h); }++static const U64 prime6bytes = 227718039650203ULL;+static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64-48)) * prime6bytes) >> (64-h)) ; }+static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }++static const U64 prime7bytes = 58295818150454627ULL;+static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u << (64-56)) * prime7bytes) >> (64-h)) ; }+static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h); }++static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;+static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }+static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }++MEM_STATIC size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls)+{+ switch(mls)+ {+ default:+ case 4: return ZSTD_hash4Ptr(p, hBits);+ case 5: return ZSTD_hash5Ptr(p, hBits);+ case 6: return ZSTD_hash6Ptr(p, hBits);+ case 7: return ZSTD_hash7Ptr(p, hBits);+ case 8: return ZSTD_hash8Ptr(p, hBits);+ }+}++/** ZSTD_ipow() :+ * Return base^exponent.+ */+static U64 ZSTD_ipow(U64 base, U64 exponent)+{+ U64 power = 1;+ while (exponent) {+ if (exponent & 1) power *= base;+ exponent >>= 1;+ base *= base;+ }+ return power;+}++#define ZSTD_ROLL_HASH_CHAR_OFFSET 10++/** ZSTD_rollingHash_append() :+ * Add the buffer to the hash value.+ */+static U64 ZSTD_rollingHash_append(U64 hash, void const* buf, size_t size)+{+ BYTE const* istart = (BYTE const*)buf;+ size_t pos;+ for (pos = 0; pos < size; ++pos) {+ hash *= prime8bytes;+ hash += istart[pos] + ZSTD_ROLL_HASH_CHAR_OFFSET;+ }+ return hash;+}++/** ZSTD_rollingHash_compute() :+ * Compute the rolling hash value of the buffer.+ */+MEM_STATIC U64 ZSTD_rollingHash_compute(void const* buf, size_t size)+{+ return ZSTD_rollingHash_append(0, buf, size);+}++/** ZSTD_rollingHash_primePower() :+ * Compute the primePower to be passed to ZSTD_rollingHash_rotate() for a hash+ * over a window of length bytes.+ */+MEM_STATIC U64 ZSTD_rollingHash_primePower(U32 length)+{+ return ZSTD_ipow(prime8bytes, length - 1);+}++/** ZSTD_rollingHash_rotate() :+ * Rotate the rolling hash by one byte.+ */+MEM_STATIC U64 ZSTD_rollingHash_rotate(U64 hash, BYTE toRemove, BYTE toAdd, U64 primePower)+{+ hash -= (toRemove + ZSTD_ROLL_HASH_CHAR_OFFSET) * primePower;+ hash *= prime8bytes;+ hash += toAdd + ZSTD_ROLL_HASH_CHAR_OFFSET;+ return hash;+}++/*-*************************************+* Round buffer management+***************************************/+/* Max current allowed */+#define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX))+/* Maximum chunk size before overflow correction needs to be called again */+#define ZSTD_CHUNKSIZE_MAX \+ ( ((U32)-1) /* Maximum ending current index */ \+ - ZSTD_CURRENT_MAX) /* Maximum beginning lowLimit */++/**+ * ZSTD_window_clear():+ * Clears the window containing the history by simply setting it to empty.+ */+MEM_STATIC void ZSTD_window_clear(ZSTD_window_t* window)+{+ size_t const endT = (size_t)(window->nextSrc - window->base);+ U32 const end = (U32)endT;++ window->lowLimit = end;+ window->dictLimit = end;+}++/**+ * ZSTD_window_hasExtDict():+ * Returns non-zero if the window has a non-empty extDict.+ */+MEM_STATIC U32 ZSTD_window_hasExtDict(ZSTD_window_t const window)+{+ return window.lowLimit < window.dictLimit;+}++/**+ * ZSTD_matchState_dictMode():+ * Inspects the provided matchState and figures out what dictMode should be+ * passed to the compressor.+ */+MEM_STATIC ZSTD_dictMode_e ZSTD_matchState_dictMode(const ZSTD_matchState_t *ms)+{+ return ZSTD_window_hasExtDict(ms->window) ?+ ZSTD_extDict :+ ms->dictMatchState != NULL ?+ ZSTD_dictMatchState :+ ZSTD_noDict;+}++/**+ * ZSTD_window_needOverflowCorrection():+ * Returns non-zero if the indices are getting too large and need overflow+ * protection.+ */+MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window,+ void const* srcEnd)+{+ U32 const current = (U32)((BYTE const*)srcEnd - window.base);+ return current > ZSTD_CURRENT_MAX;+}++/**+ * ZSTD_window_correctOverflow():+ * Reduces the indices to protect from index overflow.+ * Returns the correction made to the indices, which must be applied to every+ * stored index.+ *+ * The least significant cycleLog bits of the indices must remain the same,+ * which may be 0. Every index up to maxDist in the past must be valid.+ * NOTE: (maxDist & cycleMask) must be zero.+ */+MEM_STATIC U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog,+ U32 maxDist, void const* src)+{+ /* preemptive overflow correction:+ * 1. correction is large enough:+ * lowLimit > (3<<29) ==> current > 3<<29 + 1<<windowLog+ * 1<<windowLog <= newCurrent < 1<<chainLog + 1<<windowLog+ *+ * current - newCurrent+ * > (3<<29 + 1<<windowLog) - (1<<windowLog + 1<<chainLog)+ * > (3<<29) - (1<<chainLog)+ * > (3<<29) - (1<<30) (NOTE: chainLog <= 30)+ * > 1<<29+ *+ * 2. (ip+ZSTD_CHUNKSIZE_MAX - cctx->base) doesn't overflow:+ * After correction, current is less than (1<<chainLog + 1<<windowLog).+ * In 64-bit mode we are safe, because we have 64-bit ptrdiff_t.+ * In 32-bit mode we are safe, because (chainLog <= 29), so+ * ip+ZSTD_CHUNKSIZE_MAX - cctx->base < 1<<32.+ * 3. (cctx->lowLimit + 1<<windowLog) < 1<<32:+ * windowLog <= 31 ==> 3<<29 + 1<<windowLog < 7<<29 < 1<<32.+ */+ U32 const cycleMask = (1U << cycleLog) - 1;+ U32 const current = (U32)((BYTE const*)src - window->base);+ U32 const newCurrent = (current & cycleMask) + maxDist;+ U32 const correction = current - newCurrent;+ assert((maxDist & cycleMask) == 0);+ assert(current > newCurrent);+ /* Loose bound, should be around 1<<29 (see above) */+ assert(correction > 1<<28);++ window->base += correction;+ window->dictBase += correction;+ window->lowLimit -= correction;+ window->dictLimit -= correction;++ DEBUGLOG(4, "Correction of 0x%x bytes to lowLimit=0x%x", correction,+ window->lowLimit);+ return correction;+}++/**+ * ZSTD_window_enforceMaxDist():+ * Updates lowLimit so that:+ * (srcEnd - base) - lowLimit == maxDist + loadedDictEnd+ *+ * This allows a simple check that index >= lowLimit to see if index is valid.+ * This must be called before a block compression call, with srcEnd as the block+ * source end.+ *+ * If loadedDictEndPtr is not NULL, we set it to zero once we update lowLimit.+ * This is because dictionaries are allowed to be referenced as long as the last+ * byte of the dictionary is in the window, but once they are out of range,+ * they cannot be referenced. If loadedDictEndPtr is NULL, we use+ * loadedDictEnd == 0.+ *+ * In normal dict mode, the dict is between lowLimit and dictLimit. In+ * dictMatchState mode, lowLimit and dictLimit are the same, and the dictionary+ * is below them. forceWindow and dictMatchState are therefore incompatible.+ */+MEM_STATIC void+ZSTD_window_enforceMaxDist(ZSTD_window_t* window,+ void const* srcEnd,+ U32 maxDist,+ U32* loadedDictEndPtr,+ const ZSTD_matchState_t** dictMatchStatePtr)+{+ U32 const blockEndIdx = (U32)((BYTE const*)srcEnd - window->base);+ U32 loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0;+ DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u",+ (unsigned)blockEndIdx, (unsigned)maxDist);+ if (blockEndIdx > maxDist + loadedDictEnd) {+ U32 const newLowLimit = blockEndIdx - maxDist;+ if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit;+ if (window->dictLimit < window->lowLimit) {+ DEBUGLOG(5, "Update dictLimit to match lowLimit, from %u to %u",+ (unsigned)window->dictLimit, (unsigned)window->lowLimit);+ window->dictLimit = window->lowLimit;+ }+ if (loadedDictEndPtr)+ *loadedDictEndPtr = 0;+ if (dictMatchStatePtr)+ *dictMatchStatePtr = NULL;+ }+}++/**+ * ZSTD_window_update():+ * Updates the window by appending [src, src + srcSize) to the window.+ * If it is not contiguous, the current prefix becomes the extDict, and we+ * forget about the extDict. Handles overlap of the prefix and extDict.+ * Returns non-zero if the segment is contiguous.+ */+MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window,+ void const* src, size_t srcSize)+{+ BYTE const* const ip = (BYTE const*)src;+ U32 contiguous = 1;+ DEBUGLOG(5, "ZSTD_window_update");+ /* Check if blocks follow each other */+ if (src != window->nextSrc) {+ /* not contiguous */+ size_t const distanceFromBase = (size_t)(window->nextSrc - window->base);+ DEBUGLOG(5, "Non contiguous blocks, new segment starts at %u", window->dictLimit);+ window->lowLimit = window->dictLimit;+ assert(distanceFromBase == (size_t)(U32)distanceFromBase); /* should never overflow */+ window->dictLimit = (U32)distanceFromBase;+ window->dictBase = window->base;+ window->base = ip - distanceFromBase;+ // ms->nextToUpdate = window->dictLimit;+ if (window->dictLimit - window->lowLimit < HASH_READ_SIZE) window->lowLimit = window->dictLimit; /* too small extDict */+ contiguous = 0;+ }+ window->nextSrc = ip + srcSize;+ /* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */+ if ( (ip+srcSize > window->dictBase + window->lowLimit)+ & (ip < window->dictBase + window->dictLimit)) {+ ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase;+ U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx;+ window->lowLimit = lowLimitMax;+ DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit);+ }+ return contiguous;+}+++/* debug functions */+#if (DEBUGLEVEL>=2)++MEM_STATIC double ZSTD_fWeight(U32 rawStat)+{+ U32 const fp_accuracy = 8;+ U32 const fp_multiplier = (1 << fp_accuracy);+ U32 const newStat = rawStat + 1;+ U32 const hb = ZSTD_highbit32(newStat);+ U32 const BWeight = hb * fp_multiplier;+ U32 const FWeight = (newStat << fp_accuracy) >> hb;+ U32 const weight = BWeight + FWeight;+ assert(hb + fp_accuracy < 31);+ return (double)weight / fp_multiplier;+}++/* display a table content,+ * listing each element, its frequency, and its predicted bit cost */+MEM_STATIC void ZSTD_debugTable(const U32* table, U32 max)+{+ unsigned u, sum;+ for (u=0, sum=0; u<=max; u++) sum += table[u];+ DEBUGLOG(2, "total nb elts: %u", sum);+ for (u=0; u<=max; u++) {+ DEBUGLOG(2, "%2u: %5u (%.2f)",+ u, table[u], ZSTD_fWeight(sum) - ZSTD_fWeight(table[u]) );+ }+}++#endif+++#if defined (__cplusplus)+}+#endif+++/* ==============================================================+ * Private declarations+ * These prototypes shall only be called from within lib/compress+ * ============================================================== */++/* ZSTD_getCParamsFromCCtxParams() :+ * cParams are built depending on compressionLevel, src size hints,+ * LDM and manually set compression parameters.+ */+ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(+ const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize);++/*! ZSTD_initCStream_internal() :+ * Private use only. Init streaming operation.+ * expects params to be valid.+ * must receive dict, or cdict, or none, but not both.+ * @return : 0, or an error code */+size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,+ const void* dict, size_t dictSize,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params, unsigned long long pledgedSrcSize);++void ZSTD_resetSeqStore(seqStore_t* ssPtr);++/*! ZSTD_getCParamsFromCDict() :+ * as the name implies */+ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict);++/* ZSTD_compressBegin_advanced_internal() :+ * Private use only. To be called from zstdmt_compress.c. */+size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,+ const void* dict, size_t dictSize,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_dictTableLoadMethod_e dtlm,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params,+ unsigned long long pledgedSrcSize);++/* ZSTD_compress_advanced_internal() :+ * Private use only. To be called from zstdmt_compress.c. */+size_t ZSTD_compress_advanced_internal(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize,+ ZSTD_CCtx_params params);+++/* ZSTD_writeLastEmptyBlock() :+ * output an empty Block with end-of-frame mark to complete a frame+ * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))+ * or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)+ */+size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity);+++/* ZSTD_referenceExternalSequences() :+ * Must be called before starting a compression operation.+ * seqs must parse a prefix of the source.+ * This cannot be used when long range matching is enabled.+ * Zstd will use these sequences, and pass the literals to a secondary block+ * compressor.+ * @return : An error code on failure.+ * NOTE: seqs are not verified! Invalid sequences can cause out-of-bounds memory+ * access and data corruption.+ */+size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq);+++#endif /* ZSTD_COMPRESS_H */
+ zstd/lib/compress/zstd_double_fast.c view
@@ -0,0 +1,499 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#include "zstd_compress_internal.h"+#include "zstd_double_fast.h"+++void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,+ void const* end, ZSTD_dictTableLoadMethod_e dtlm)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashLarge = ms->hashTable;+ U32 const hBitsL = cParams->hashLog;+ U32 const mls = cParams->minMatch;+ U32* const hashSmall = ms->chainTable;+ U32 const hBitsS = cParams->chainLog;+ const BYTE* const base = ms->window.base;+ const BYTE* ip = base + ms->nextToUpdate;+ const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;+ const U32 fastHashFillStep = 3;++ /* Always insert every fastHashFillStep position into the hash tables.+ * Insert the other positions into the large hash table if their entry+ * is empty.+ */+ for (; ip + fastHashFillStep - 1 <= iend; ip += fastHashFillStep) {+ U32 const current = (U32)(ip - base);+ U32 i;+ for (i = 0; i < fastHashFillStep; ++i) {+ size_t const smHash = ZSTD_hashPtr(ip + i, hBitsS, mls);+ size_t const lgHash = ZSTD_hashPtr(ip + i, hBitsL, 8);+ if (i == 0)+ hashSmall[smHash] = current + i;+ if (i == 0 || hashLarge[lgHash] == 0)+ hashLarge[lgHash] = current + i;+ /* Only load extra positions for ZSTD_dtlm_full */+ if (dtlm == ZSTD_dtlm_fast)+ break;+ }+ }+}+++FORCE_INLINE_TEMPLATE+size_t ZSTD_compressBlock_doubleFast_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize,+ U32 const mls /* template */, ZSTD_dictMode_e const dictMode)+{+ ZSTD_compressionParameters const* cParams = &ms->cParams;+ U32* const hashLong = ms->hashTable;+ const U32 hBitsL = cParams->hashLog;+ U32* const hashSmall = ms->chainTable;+ const U32 hBitsS = cParams->chainLog;+ const BYTE* const base = ms->window.base;+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const U32 prefixLowestIndex = ms->window.dictLimit;+ const BYTE* const prefixLowest = base + prefixLowestIndex;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - HASH_READ_SIZE;+ U32 offset_1=rep[0], offset_2=rep[1];+ U32 offsetSaved = 0;++ const ZSTD_matchState_t* const dms = ms->dictMatchState;+ const ZSTD_compressionParameters* const dictCParams =+ dictMode == ZSTD_dictMatchState ?+ &dms->cParams : NULL;+ const U32* const dictHashLong = dictMode == ZSTD_dictMatchState ?+ dms->hashTable : NULL;+ const U32* const dictHashSmall = dictMode == ZSTD_dictMatchState ?+ dms->chainTable : NULL;+ const U32 dictStartIndex = dictMode == ZSTD_dictMatchState ?+ dms->window.dictLimit : 0;+ const BYTE* const dictBase = dictMode == ZSTD_dictMatchState ?+ dms->window.base : NULL;+ const BYTE* const dictStart = dictMode == ZSTD_dictMatchState ?+ dictBase + dictStartIndex : NULL;+ const BYTE* const dictEnd = dictMode == ZSTD_dictMatchState ?+ dms->window.nextSrc : NULL;+ const U32 dictIndexDelta = dictMode == ZSTD_dictMatchState ?+ prefixLowestIndex - (U32)(dictEnd - dictBase) :+ 0;+ const U32 dictHBitsL = dictMode == ZSTD_dictMatchState ?+ dictCParams->hashLog : hBitsL;+ const U32 dictHBitsS = dictMode == ZSTD_dictMatchState ?+ dictCParams->chainLog : hBitsS;+ const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictStart);++ assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);++ /* init */+ ip += (dictAndPrefixLength == 0);+ if (dictMode == ZSTD_noDict) {+ U32 const maxRep = (U32)(ip - prefixLowest);+ if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;+ if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;+ }+ if (dictMode == ZSTD_dictMatchState) {+ /* dictMatchState repCode checks don't currently handle repCode == 0+ * disabling. */+ assert(offset_1 <= dictAndPrefixLength);+ assert(offset_2 <= dictAndPrefixLength);+ }++ /* Main Search Loop */+ while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */+ size_t mLength;+ U32 offset;+ size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);+ size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);+ size_t const dictHL = ZSTD_hashPtr(ip, dictHBitsL, 8);+ size_t const dictHS = ZSTD_hashPtr(ip, dictHBitsS, mls);+ U32 const current = (U32)(ip-base);+ U32 const matchIndexL = hashLong[h2];+ U32 matchIndexS = hashSmall[h];+ const BYTE* matchLong = base + matchIndexL;+ const BYTE* match = base + matchIndexS;+ const U32 repIndex = current + 1 - offset_1;+ const BYTE* repMatch = (dictMode == ZSTD_dictMatchState+ && repIndex < prefixLowestIndex) ?+ dictBase + (repIndex - dictIndexDelta) :+ base + repIndex;+ hashLong[h2] = hashSmall[h] = current; /* update hash tables */++ /* check dictMatchState repcode */+ if (dictMode == ZSTD_dictMatchState+ && ((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)+ && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {+ const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;+ mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;+ ip++;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);+ goto _match_stored;+ }++ /* check noDict repcode */+ if ( dictMode == ZSTD_noDict+ && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {+ mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;+ ip++;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);+ goto _match_stored;+ }++ if (matchIndexL > prefixLowestIndex) {+ /* check prefix long match */+ if (MEM_read64(matchLong) == MEM_read64(ip)) {+ mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;+ offset = (U32)(ip-matchLong);+ while (((ip>anchor) & (matchLong>prefixLowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */+ goto _match_found;+ }+ } else if (dictMode == ZSTD_dictMatchState) {+ /* check dictMatchState long match */+ U32 const dictMatchIndexL = dictHashLong[dictHL];+ const BYTE* dictMatchL = dictBase + dictMatchIndexL;+ assert(dictMatchL < dictEnd);++ if (dictMatchL > dictStart && MEM_read64(dictMatchL) == MEM_read64(ip)) {+ mLength = ZSTD_count_2segments(ip+8, dictMatchL+8, iend, dictEnd, prefixLowest) + 8;+ offset = (U32)(current - dictMatchIndexL - dictIndexDelta);+ while (((ip>anchor) & (dictMatchL>dictStart)) && (ip[-1] == dictMatchL[-1])) { ip--; dictMatchL--; mLength++; } /* catch up */+ goto _match_found;+ }+ }++ if (matchIndexS > prefixLowestIndex) {+ /* check prefix short match */+ if (MEM_read32(match) == MEM_read32(ip)) {+ goto _search_next_long;+ }+ } else if (dictMode == ZSTD_dictMatchState) {+ /* check dictMatchState short match */+ U32 const dictMatchIndexS = dictHashSmall[dictHS];+ match = dictBase + dictMatchIndexS;+ matchIndexS = dictMatchIndexS + dictIndexDelta;++ if (match > dictStart && MEM_read32(match) == MEM_read32(ip)) {+ goto _search_next_long;+ }+ }++ ip += ((ip-anchor) >> kSearchStrength) + 1;+ continue;++_search_next_long:++ {+ size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);+ size_t const dictHLNext = ZSTD_hashPtr(ip+1, dictHBitsL, 8);+ U32 const matchIndexL3 = hashLong[hl3];+ const BYTE* matchL3 = base + matchIndexL3;+ hashLong[hl3] = current + 1;++ /* check prefix long +1 match */+ if (matchIndexL3 > prefixLowestIndex) {+ if (MEM_read64(matchL3) == MEM_read64(ip+1)) {+ mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;+ ip++;+ offset = (U32)(ip-matchL3);+ while (((ip>anchor) & (matchL3>prefixLowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */+ goto _match_found;+ }+ } else if (dictMode == ZSTD_dictMatchState) {+ /* check dict long +1 match */+ U32 const dictMatchIndexL3 = dictHashLong[dictHLNext];+ const BYTE* dictMatchL3 = dictBase + dictMatchIndexL3;+ assert(dictMatchL3 < dictEnd);+ if (dictMatchL3 > dictStart && MEM_read64(dictMatchL3) == MEM_read64(ip+1)) {+ mLength = ZSTD_count_2segments(ip+1+8, dictMatchL3+8, iend, dictEnd, prefixLowest) + 8;+ ip++;+ offset = (U32)(current + 1 - dictMatchIndexL3 - dictIndexDelta);+ while (((ip>anchor) & (dictMatchL3>dictStart)) && (ip[-1] == dictMatchL3[-1])) { ip--; dictMatchL3--; mLength++; } /* catch up */+ goto _match_found;+ }+ }+ }++ /* if no long +1 match, explore the short match we found */+ if (dictMode == ZSTD_dictMatchState && matchIndexS < prefixLowestIndex) {+ mLength = ZSTD_count_2segments(ip+4, match+4, iend, dictEnd, prefixLowest) + 4;+ offset = (U32)(current - matchIndexS);+ while (((ip>anchor) & (match>dictStart)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */+ } else {+ mLength = ZSTD_count(ip+4, match+4, iend) + 4;+ offset = (U32)(ip - match);+ while (((ip>anchor) & (match>prefixLowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */+ }++ /* fall-through */++_match_found:+ offset_2 = offset_1;+ offset_1 = offset;++ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);++_match_stored:+ /* match found */+ ip += mLength;+ anchor = ip;++ if (ip <= ilimit) {+ /* Fill Table */+ hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] =+ hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2; /* here because current+2 could be > iend-8 */+ hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] =+ hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);++ /* check immediate repcode */+ if (dictMode == ZSTD_dictMatchState) {+ while (ip <= ilimit) {+ U32 const current2 = (U32)(ip-base);+ U32 const repIndex2 = current2 - offset_2;+ const BYTE* repMatch2 = dictMode == ZSTD_dictMatchState+ && repIndex2 < prefixLowestIndex ?+ dictBase - dictIndexDelta + repIndex2 :+ base + repIndex2;+ if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)+ && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {+ const BYTE* const repEnd2 = repIndex2 < prefixLowestIndex ? dictEnd : iend;+ size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixLowest) + 4;+ U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);+ hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;+ hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;+ ip += repLength2;+ anchor = ip;+ continue;+ }+ break;+ }+ }++ if (dictMode == ZSTD_noDict) {+ while ( (ip <= ilimit)+ && ( (offset_2>0)+ & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {+ /* store sequence */+ size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;+ U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; /* swap offset_2 <=> offset_1 */+ hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);+ hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);+ ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);+ ip += rLength;+ anchor = ip;+ continue; /* faster when present ... (?) */+ } } } }++ /* save reps for next block */+ rep[0] = offset_1 ? offset_1 : offsetSaved;+ rep[1] = offset_2 ? offset_2 : offsetSaved;++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_doubleFast(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ const U32 mls = ms->cParams.minMatch;+ switch(mls)+ {+ default: /* includes case 3 */+ case 4 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_noDict);+ case 5 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_noDict);+ case 6 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_noDict);+ case 7 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_noDict);+ }+}+++size_t ZSTD_compressBlock_doubleFast_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ const U32 mls = ms->cParams.minMatch;+ switch(mls)+ {+ default: /* includes case 3 */+ case 4 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_dictMatchState);+ case 5 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_dictMatchState);+ case 6 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_dictMatchState);+ case 7 :+ return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_dictMatchState);+ }+}+++static size_t ZSTD_compressBlock_doubleFast_extDict_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize,+ U32 const mls /* template */)+{+ ZSTD_compressionParameters const* cParams = &ms->cParams;+ U32* const hashLong = ms->hashTable;+ U32 const hBitsL = cParams->hashLog;+ U32* const hashSmall = ms->chainTable;+ U32 const hBitsS = cParams->chainLog;+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - 8;+ const U32 prefixStartIndex = ms->window.dictLimit;+ const BYTE* const base = ms->window.base;+ const BYTE* const prefixStart = base + prefixStartIndex;+ const U32 dictStartIndex = ms->window.lowLimit;+ const BYTE* const dictBase = ms->window.dictBase;+ const BYTE* const dictStart = dictBase + dictStartIndex;+ const BYTE* const dictEnd = dictBase + prefixStartIndex;+ U32 offset_1=rep[0], offset_2=rep[1];++ DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_extDict_generic (srcSize=%zu)", srcSize);++ /* Search Loop */+ while (ip < ilimit) { /* < instead of <=, because (ip+1) */+ const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);+ const U32 matchIndex = hashSmall[hSmall];+ const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;+ const BYTE* match = matchBase + matchIndex;++ const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);+ const U32 matchLongIndex = hashLong[hLong];+ const BYTE* const matchLongBase = matchLongIndex < prefixStartIndex ? dictBase : base;+ const BYTE* matchLong = matchLongBase + matchLongIndex;++ const U32 current = (U32)(ip-base);+ const U32 repIndex = current + 1 - offset_1; /* offset_1 expected <= current +1 */+ const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;+ const BYTE* const repMatch = repBase + repIndex;+ size_t mLength;+ hashSmall[hSmall] = hashLong[hLong] = current; /* update hash table */++ if ((((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex doesn't overlap dict + prefix */+ & (repIndex > dictStartIndex))+ && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {+ const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;+ mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;+ ip++;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);+ } else {+ if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {+ const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;+ const BYTE* const lowMatchPtr = matchLongIndex < prefixStartIndex ? dictStart : prefixStart;+ U32 offset;+ mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, prefixStart) + 8;+ offset = current - matchLongIndex;+ while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */+ offset_2 = offset_1;+ offset_1 = offset;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);++ } else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {+ size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);+ U32 const matchIndex3 = hashLong[h3];+ const BYTE* const match3Base = matchIndex3 < prefixStartIndex ? dictBase : base;+ const BYTE* match3 = match3Base + matchIndex3;+ U32 offset;+ hashLong[h3] = current + 1;+ if ( (matchIndex3 > dictStartIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {+ const BYTE* const matchEnd = matchIndex3 < prefixStartIndex ? dictEnd : iend;+ const BYTE* const lowMatchPtr = matchIndex3 < prefixStartIndex ? dictStart : prefixStart;+ mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, prefixStart) + 8;+ ip++;+ offset = current+1 - matchIndex3;+ while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */+ } else {+ const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;+ const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;+ mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;+ offset = current - matchIndex;+ while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */+ }+ offset_2 = offset_1;+ offset_1 = offset;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);++ } else {+ ip += ((ip-anchor) >> kSearchStrength) + 1;+ continue;+ } }++ /* found a match : store it */+ ip += mLength;+ anchor = ip;++ if (ip <= ilimit) {+ /* Fill Table */+ hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;+ hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2;+ hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);+ hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);+ /* check immediate repcode */+ while (ip <= ilimit) {+ U32 const current2 = (U32)(ip-base);+ U32 const repIndex2 = current2 - offset_2;+ const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;+ if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) /* intentional overflow : ensure repIndex2 doesn't overlap dict + prefix */+ & (repIndex2 > dictStartIndex))+ && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {+ const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;+ size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;+ U32 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);+ hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;+ hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;+ ip += repLength2;+ anchor = ip;+ continue;+ }+ break;+ } } }++ /* save reps for next block */+ rep[0] = offset_1;+ rep[1] = offset_2;++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_doubleFast_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ U32 const mls = ms->cParams.minMatch;+ switch(mls)+ {+ default: /* includes case 3 */+ case 4 :+ return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 4);+ case 5 :+ return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 5);+ case 6 :+ return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 6);+ case 7 :+ return ZSTD_compressBlock_doubleFast_extDict_generic(ms, seqStore, rep, src, srcSize, 7);+ }+}
+ zstd/lib/compress/zstd_double_fast.h view
@@ -0,0 +1,38 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#ifndef ZSTD_DOUBLE_FAST_H+#define ZSTD_DOUBLE_FAST_H++#if defined (__cplusplus)+extern "C" {+#endif++#include "mem.h" /* U32 */+#include "zstd_compress_internal.h" /* ZSTD_CCtx, size_t */++void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,+ void const* end, ZSTD_dictTableLoadMethod_e dtlm);+size_t ZSTD_compressBlock_doubleFast(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_doubleFast_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_doubleFast_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+++#if defined (__cplusplus)+}+#endif++#endif /* ZSTD_DOUBLE_FAST_H */
+ zstd/lib/compress/zstd_fast.c view
@@ -0,0 +1,470 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#include "zstd_compress_internal.h"+#include "zstd_fast.h"+++void ZSTD_fillHashTable(ZSTD_matchState_t* ms,+ void const* end, ZSTD_dictTableLoadMethod_e dtlm)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hBits = cParams->hashLog;+ U32 const mls = cParams->minMatch;+ const BYTE* const base = ms->window.base;+ const BYTE* ip = base + ms->nextToUpdate;+ const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;+ const U32 fastHashFillStep = 3;++ /* Always insert every fastHashFillStep position into the hash table.+ * Insert the other positions if their hash entry is empty.+ */+ for ( ; ip + fastHashFillStep < iend + 2; ip += fastHashFillStep) {+ U32 const current = (U32)(ip - base);+ size_t const hash0 = ZSTD_hashPtr(ip, hBits, mls);+ hashTable[hash0] = current;+ if (dtlm == ZSTD_dtlm_fast) continue;+ /* Only load extra positions for ZSTD_dtlm_full */+ { U32 p;+ for (p = 1; p < fastHashFillStep; ++p) {+ size_t const hash = ZSTD_hashPtr(ip + p, hBits, mls);+ if (hashTable[hash] == 0) { /* not yet filled */+ hashTable[hash] = current + p;+ } } } }+}++FORCE_INLINE_TEMPLATE+size_t ZSTD_compressBlock_fast_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize,+ U32 const mls)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hlog = cParams->hashLog;+ /* support stepSize of 0 */+ size_t const stepSize = cParams->targetLength + !(cParams->targetLength) + 1;+ const BYTE* const base = ms->window.base;+ const BYTE* const istart = (const BYTE*)src;+ /* We check ip0 (ip + 0) and ip1 (ip + 1) each loop */+ const BYTE* ip0 = istart;+ const BYTE* ip1;+ const BYTE* anchor = istart;+ const U32 prefixStartIndex = ms->window.dictLimit;+ const BYTE* const prefixStart = base + prefixStartIndex;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - HASH_READ_SIZE;+ U32 offset_1=rep[0], offset_2=rep[1];+ U32 offsetSaved = 0;++ /* init */+ ip0 += (ip0 == prefixStart);+ ip1 = ip0 + 1;+ {+ U32 const maxRep = (U32)(ip0 - prefixStart);+ if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;+ if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;+ }++ /* Main Search Loop */+ while (ip1 < ilimit) { /* < instead of <=, because check at ip0+2 */+ size_t mLength;+ BYTE const* ip2 = ip0 + 2;+ size_t const h0 = ZSTD_hashPtr(ip0, hlog, mls);+ U32 const val0 = MEM_read32(ip0);+ size_t const h1 = ZSTD_hashPtr(ip1, hlog, mls);+ U32 const val1 = MEM_read32(ip1);+ U32 const current0 = (U32)(ip0-base);+ U32 const current1 = (U32)(ip1-base);+ U32 const matchIndex0 = hashTable[h0];+ U32 const matchIndex1 = hashTable[h1];+ BYTE const* repMatch = ip2-offset_1;+ const BYTE* match0 = base + matchIndex0;+ const BYTE* match1 = base + matchIndex1;+ U32 offcode;+ hashTable[h0] = current0; /* update hash table */+ hashTable[h1] = current1; /* update hash table */++ assert(ip0 + 1 == ip1);++ if ((offset_1 > 0) & (MEM_read32(repMatch) == MEM_read32(ip2))) {+ mLength = ip2[-1] == repMatch[-1] ? 1 : 0;+ ip0 = ip2 - mLength;+ match0 = repMatch - mLength;+ offcode = 0;+ goto _match;+ }+ if ((matchIndex0 > prefixStartIndex) && MEM_read32(match0) == val0) {+ /* found a regular match */+ goto _offset;+ }+ if ((matchIndex1 > prefixStartIndex) && MEM_read32(match1) == val1) {+ /* found a regular match after one literal */+ ip0 = ip1;+ match0 = match1;+ goto _offset;+ }+ {+ size_t const step = ((ip0-anchor) >> (kSearchStrength - 1)) + stepSize;+ assert(step >= 2);+ ip0 += step;+ ip1 += step;+ continue;+ }+_offset: /* Requires: ip0, match0 */+ /* Compute the offset code */+ offset_2 = offset_1;+ offset_1 = (U32)(ip0-match0);+ offcode = offset_1 + ZSTD_REP_MOVE;+ mLength = 0;+ /* Count the backwards match length */+ while (((ip0>anchor) & (match0>prefixStart))+ && (ip0[-1] == match0[-1])) { ip0--; match0--; mLength++; } /* catch up */++_match: /* Requires: ip0, match0, offcode */+ /* Count the forward length */+ mLength += ZSTD_count(ip0+mLength+4, match0+mLength+4, iend) + 4;+ ZSTD_storeSeq(seqStore, ip0-anchor, anchor, offcode, mLength-MINMATCH);+ /* match found */+ ip0 += mLength;+ anchor = ip0;+ ip1 = ip0 + 1;++ if (ip0 <= ilimit) {+ /* Fill Table */+ assert(base+current0+2 > istart); /* check base overflow */+ hashTable[ZSTD_hashPtr(base+current0+2, hlog, mls)] = current0+2; /* here because current+2 could be > iend-8 */+ hashTable[ZSTD_hashPtr(ip0-2, hlog, mls)] = (U32)(ip0-2-base);++ while ( (ip0 <= ilimit)+ && ( (offset_2>0)+ & (MEM_read32(ip0) == MEM_read32(ip0 - offset_2)) )) {+ /* store sequence */+ size_t const rLength = ZSTD_count(ip0+4, ip0+4-offset_2, iend) + 4;+ U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; /* swap offset_2 <=> offset_1 */+ hashTable[ZSTD_hashPtr(ip0, hlog, mls)] = (U32)(ip0-base);+ ip0 += rLength;+ ip1 = ip0 + 1;+ ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);+ anchor = ip0;+ continue; /* faster when present (confirmed on gcc-8) ... (?) */+ }+ }+ }++ /* save reps for next block */+ rep[0] = offset_1 ? offset_1 : offsetSaved;+ rep[1] = offset_2 ? offset_2 : offsetSaved;++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_fast(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ ZSTD_compressionParameters const* cParams = &ms->cParams;+ U32 const mls = cParams->minMatch;+ assert(ms->dictMatchState == NULL);+ switch(mls)+ {+ default: /* includes case 3 */+ case 4 :+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4);+ case 5 :+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5);+ case 6 :+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6);+ case 7 :+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7);+ }+}++FORCE_INLINE_TEMPLATE+size_t ZSTD_compressBlock_fast_dictMatchState_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize, U32 const mls)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hlog = cParams->hashLog;+ /* support stepSize of 0 */+ U32 const stepSize = cParams->targetLength + !(cParams->targetLength);+ const BYTE* const base = ms->window.base;+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const U32 prefixStartIndex = ms->window.dictLimit;+ const BYTE* const prefixStart = base + prefixStartIndex;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - HASH_READ_SIZE;+ U32 offset_1=rep[0], offset_2=rep[1];+ U32 offsetSaved = 0;++ const ZSTD_matchState_t* const dms = ms->dictMatchState;+ const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;+ const U32* const dictHashTable = dms->hashTable;+ const U32 dictStartIndex = dms->window.dictLimit;+ const BYTE* const dictBase = dms->window.base;+ const BYTE* const dictStart = dictBase + dictStartIndex;+ const BYTE* const dictEnd = dms->window.nextSrc;+ const U32 dictIndexDelta = prefixStartIndex - (U32)(dictEnd - dictBase);+ const U32 dictAndPrefixLength = (U32)(ip - prefixStart + dictEnd - dictStart);+ const U32 dictHLog = dictCParams->hashLog;++ /* otherwise, we would get index underflow when translating a dict index+ * into a local index */+ assert(prefixStartIndex >= (U32)(dictEnd - dictBase));++ /* init */+ ip += (dictAndPrefixLength == 0);+ /* dictMatchState repCode checks don't currently handle repCode == 0+ * disabling. */+ assert(offset_1 <= dictAndPrefixLength);+ assert(offset_2 <= dictAndPrefixLength);++ /* Main Search Loop */+ while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */+ size_t mLength;+ size_t const h = ZSTD_hashPtr(ip, hlog, mls);+ U32 const current = (U32)(ip-base);+ U32 const matchIndex = hashTable[h];+ const BYTE* match = base + matchIndex;+ const U32 repIndex = current + 1 - offset_1;+ const BYTE* repMatch = (repIndex < prefixStartIndex) ?+ dictBase + (repIndex - dictIndexDelta) :+ base + repIndex;+ hashTable[h] = current; /* update hash table */++ if ( ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */+ && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {+ const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;+ mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;+ ip++;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);+ } else if ( (matchIndex <= prefixStartIndex) ) {+ size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);+ U32 const dictMatchIndex = dictHashTable[dictHash];+ const BYTE* dictMatch = dictBase + dictMatchIndex;+ if (dictMatchIndex <= dictStartIndex ||+ MEM_read32(dictMatch) != MEM_read32(ip)) {+ assert(stepSize >= 1);+ ip += ((ip-anchor) >> kSearchStrength) + stepSize;+ continue;+ } else {+ /* found a dict match */+ U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);+ mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;+ while (((ip>anchor) & (dictMatch>dictStart))+ && (ip[-1] == dictMatch[-1])) {+ ip--; dictMatch--; mLength++;+ } /* catch up */+ offset_2 = offset_1;+ offset_1 = offset;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);+ }+ } else if (MEM_read32(match) != MEM_read32(ip)) {+ /* it's not a match, and we're not going to check the dictionary */+ assert(stepSize >= 1);+ ip += ((ip-anchor) >> kSearchStrength) + stepSize;+ continue;+ } else {+ /* found a regular match */+ U32 const offset = (U32)(ip-match);+ mLength = ZSTD_count(ip+4, match+4, iend) + 4;+ while (((ip>anchor) & (match>prefixStart))+ && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */+ offset_2 = offset_1;+ offset_1 = offset;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);+ }++ /* match found */+ ip += mLength;+ anchor = ip;++ if (ip <= ilimit) {+ /* Fill Table */+ assert(base+current+2 > istart); /* check base overflow */+ hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2; /* here because current+2 could be > iend-8 */+ hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);++ /* check immediate repcode */+ while (ip <= ilimit) {+ U32 const current2 = (U32)(ip-base);+ U32 const repIndex2 = current2 - offset_2;+ const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?+ dictBase - dictIndexDelta + repIndex2 :+ base + repIndex2;+ if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)+ && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {+ const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;+ size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;+ U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);+ hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;+ ip += repLength2;+ anchor = ip;+ continue;+ }+ break;+ }+ }+ }++ /* save reps for next block */+ rep[0] = offset_1 ? offset_1 : offsetSaved;+ rep[1] = offset_2 ? offset_2 : offsetSaved;++ /* Return the last literals size */+ return iend - anchor;+}++size_t ZSTD_compressBlock_fast_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ ZSTD_compressionParameters const* cParams = &ms->cParams;+ U32 const mls = cParams->minMatch;+ assert(ms->dictMatchState != NULL);+ switch(mls)+ {+ default: /* includes case 3 */+ case 4 :+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 4);+ case 5 :+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 5);+ case 6 :+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 6);+ case 7 :+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 7);+ }+}+++static size_t ZSTD_compressBlock_fast_extDict_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize, U32 const mls)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hlog = cParams->hashLog;+ /* support stepSize of 0 */+ U32 const stepSize = cParams->targetLength + !(cParams->targetLength);+ const BYTE* const base = ms->window.base;+ const BYTE* const dictBase = ms->window.dictBase;+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const U32 dictStartIndex = ms->window.lowLimit;+ const BYTE* const dictStart = dictBase + dictStartIndex;+ const U32 prefixStartIndex = ms->window.dictLimit;+ const BYTE* const prefixStart = base + prefixStartIndex;+ const BYTE* const dictEnd = dictBase + prefixStartIndex;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - 8;+ U32 offset_1=rep[0], offset_2=rep[1];++ /* Search Loop */+ while (ip < ilimit) { /* < instead of <=, because (ip+1) */+ const size_t h = ZSTD_hashPtr(ip, hlog, mls);+ const U32 matchIndex = hashTable[h];+ const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;+ const BYTE* match = matchBase + matchIndex;+ const U32 current = (U32)(ip-base);+ const U32 repIndex = current + 1 - offset_1;+ const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;+ const BYTE* const repMatch = repBase + repIndex;+ size_t mLength;+ hashTable[h] = current; /* update hash table */+ assert(offset_1 <= current +1); /* check repIndex */++ if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex))+ && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {+ const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;+ mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;+ ip++;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);+ } else {+ if ( (matchIndex < dictStartIndex) ||+ (MEM_read32(match) != MEM_read32(ip)) ) {+ assert(stepSize >= 1);+ ip += ((ip-anchor) >> kSearchStrength) + stepSize;+ continue;+ }+ { const BYTE* matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;+ const BYTE* lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;+ U32 offset;+ mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;+ while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */+ offset = current - matchIndex;+ offset_2 = offset_1;+ offset_1 = offset;+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);+ } }++ /* found a match : store it */+ ip += mLength;+ anchor = ip;++ if (ip <= ilimit) {+ /* Fill Table */+ hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2;+ hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);+ /* check immediate repcode */+ while (ip <= ilimit) {+ U32 const current2 = (U32)(ip-base);+ U32 const repIndex2 = current2 - offset_2;+ const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;+ if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (repIndex2 > dictStartIndex)) /* intentional overflow */+ && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {+ const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;+ size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;+ U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);+ hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;+ ip += repLength2;+ anchor = ip;+ continue;+ }+ break;+ } } }++ /* save reps for next block */+ rep[0] = offset_1;+ rep[1] = offset_2;++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_fast_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ ZSTD_compressionParameters const* cParams = &ms->cParams;+ U32 const mls = cParams->minMatch;+ switch(mls)+ {+ default: /* includes case 3 */+ case 4 :+ return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 4);+ case 5 :+ return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 5);+ case 6 :+ return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 6);+ case 7 :+ return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 7);+ }+}
+ zstd/lib/compress/zstd_fast.h view
@@ -0,0 +1,37 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#ifndef ZSTD_FAST_H+#define ZSTD_FAST_H++#if defined (__cplusplus)+extern "C" {+#endif++#include "mem.h" /* U32 */+#include "zstd_compress_internal.h"++void ZSTD_fillHashTable(ZSTD_matchState_t* ms,+ void const* end, ZSTD_dictTableLoadMethod_e dtlm);+size_t ZSTD_compressBlock_fast(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_fast_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_fast_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);++#if defined (__cplusplus)+}+#endif++#endif /* ZSTD_FAST_H */
+ zstd/lib/compress/zstd_lazy.c view
@@ -0,0 +1,1106 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#include "zstd_compress_internal.h"+#include "zstd_lazy.h"+++/*-*************************************+* Binary Tree search+***************************************/++static void+ZSTD_updateDUBT(ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* iend,+ U32 mls)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hashLog = cParams->hashLog;++ U32* const bt = ms->chainTable;+ U32 const btLog = cParams->chainLog - 1;+ U32 const btMask = (1 << btLog) - 1;++ const BYTE* const base = ms->window.base;+ U32 const target = (U32)(ip - base);+ U32 idx = ms->nextToUpdate;++ if (idx != target)+ DEBUGLOG(7, "ZSTD_updateDUBT, from %u to %u (dictLimit:%u)",+ idx, target, ms->window.dictLimit);+ assert(ip + 8 <= iend); /* condition for ZSTD_hashPtr */+ (void)iend;++ assert(idx >= ms->window.dictLimit); /* condition for valid base+idx */+ for ( ; idx < target ; idx++) {+ size_t const h = ZSTD_hashPtr(base + idx, hashLog, mls); /* assumption : ip + 8 <= iend */+ U32 const matchIndex = hashTable[h];++ U32* const nextCandidatePtr = bt + 2*(idx&btMask);+ U32* const sortMarkPtr = nextCandidatePtr + 1;++ DEBUGLOG(8, "ZSTD_updateDUBT: insert %u", idx);+ hashTable[h] = idx; /* Update Hash Table */+ *nextCandidatePtr = matchIndex; /* update BT like a chain */+ *sortMarkPtr = ZSTD_DUBT_UNSORTED_MARK;+ }+ ms->nextToUpdate = target;+}+++/** ZSTD_insertDUBT1() :+ * sort one already inserted but unsorted position+ * assumption : current >= btlow == (current - btmask)+ * doesn't fail */+static void+ZSTD_insertDUBT1(ZSTD_matchState_t* ms,+ U32 current, const BYTE* inputEnd,+ U32 nbCompares, U32 btLow,+ const ZSTD_dictMode_e dictMode)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const bt = ms->chainTable;+ U32 const btLog = cParams->chainLog - 1;+ U32 const btMask = (1 << btLog) - 1;+ size_t commonLengthSmaller=0, commonLengthLarger=0;+ const BYTE* const base = ms->window.base;+ const BYTE* const dictBase = ms->window.dictBase;+ const U32 dictLimit = ms->window.dictLimit;+ const BYTE* const ip = (current>=dictLimit) ? base + current : dictBase + current;+ const BYTE* const iend = (current>=dictLimit) ? inputEnd : dictBase + dictLimit;+ const BYTE* const dictEnd = dictBase + dictLimit;+ const BYTE* const prefixStart = base + dictLimit;+ const BYTE* match;+ U32* smallerPtr = bt + 2*(current&btMask);+ U32* largerPtr = smallerPtr + 1;+ U32 matchIndex = *smallerPtr; /* this candidate is unsorted : next sorted candidate is reached through *smallerPtr, while *largerPtr contains previous unsorted candidate (which is already saved and can be overwritten) */+ U32 dummy32; /* to be nullified at the end */+ U32 const windowLow = ms->window.lowLimit;++ DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)",+ current, dictLimit, windowLow);+ assert(current >= btLow);+ assert(ip < iend); /* condition for ZSTD_count */++ while (nbCompares-- && (matchIndex > windowLow)) {+ U32* const nextPtr = bt + 2*(matchIndex & btMask);+ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */+ assert(matchIndex < current);+ /* note : all candidates are now supposed sorted,+ * but it's still possible to have nextPtr[1] == ZSTD_DUBT_UNSORTED_MARK+ * when a real index has the same value as ZSTD_DUBT_UNSORTED_MARK */++ if ( (dictMode != ZSTD_extDict)+ || (matchIndex+matchLength >= dictLimit) /* both in current segment*/+ || (current < dictLimit) /* both in extDict */) {+ const BYTE* const mBase = ( (dictMode != ZSTD_extDict)+ || (matchIndex+matchLength >= dictLimit)) ?+ base : dictBase;+ assert( (matchIndex+matchLength >= dictLimit) /* might be wrong if extDict is incorrectly set to 0 */+ || (current < dictLimit) );+ match = mBase + matchIndex;+ matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);+ } else {+ match = dictBase + matchIndex;+ matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);+ if (matchIndex+matchLength >= dictLimit)+ match = base + matchIndex; /* preparation for next read of match[matchLength] */+ }++ DEBUGLOG(8, "ZSTD_insertDUBT1: comparing %u with %u : found %u common bytes ",+ current, matchIndex, (U32)matchLength);++ if (ip+matchLength == iend) { /* equal : no way to know if inf or sup */+ break; /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */+ }++ if (match[matchLength] < ip[matchLength]) { /* necessarily within buffer */+ /* match is smaller than current */+ *smallerPtr = matchIndex; /* update smaller idx */+ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */+ if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop searching */+ DEBUGLOG(8, "ZSTD_insertDUBT1: %u (>btLow=%u) is smaller : next => %u",+ matchIndex, btLow, nextPtr[1]);+ smallerPtr = nextPtr+1; /* new "candidate" => larger than match, which was smaller than target */+ matchIndex = nextPtr[1]; /* new matchIndex, larger than previous and closer to current */+ } else {+ /* match is larger than current */+ *largerPtr = matchIndex;+ commonLengthLarger = matchLength;+ if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop searching */+ DEBUGLOG(8, "ZSTD_insertDUBT1: %u (>btLow=%u) is larger => %u",+ matchIndex, btLow, nextPtr[0]);+ largerPtr = nextPtr;+ matchIndex = nextPtr[0];+ } }++ *smallerPtr = *largerPtr = 0;+}+++static size_t+ZSTD_DUBT_findBetterDictMatch (+ ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iend,+ size_t* offsetPtr,+ size_t bestLength,+ U32 nbCompares,+ U32 const mls,+ const ZSTD_dictMode_e dictMode)+{+ const ZSTD_matchState_t * const dms = ms->dictMatchState;+ const ZSTD_compressionParameters* const dmsCParams = &dms->cParams;+ const U32 * const dictHashTable = dms->hashTable;+ U32 const hashLog = dmsCParams->hashLog;+ size_t const h = ZSTD_hashPtr(ip, hashLog, mls);+ U32 dictMatchIndex = dictHashTable[h];++ const BYTE* const base = ms->window.base;+ const BYTE* const prefixStart = base + ms->window.dictLimit;+ U32 const current = (U32)(ip-base);+ const BYTE* const dictBase = dms->window.base;+ const BYTE* const dictEnd = dms->window.nextSrc;+ U32 const dictHighLimit = (U32)(dms->window.nextSrc - dms->window.base);+ U32 const dictLowLimit = dms->window.lowLimit;+ U32 const dictIndexDelta = ms->window.lowLimit - dictHighLimit;++ U32* const dictBt = dms->chainTable;+ U32 const btLog = dmsCParams->chainLog - 1;+ U32 const btMask = (1 << btLog) - 1;+ U32 const btLow = (btMask >= dictHighLimit - dictLowLimit) ? dictLowLimit : dictHighLimit - btMask;++ size_t commonLengthSmaller=0, commonLengthLarger=0;++ (void)dictMode;+ assert(dictMode == ZSTD_dictMatchState);++ while (nbCompares-- && (dictMatchIndex > dictLowLimit)) {+ U32* const nextPtr = dictBt + 2*(dictMatchIndex & btMask);+ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */+ const BYTE* match = dictBase + dictMatchIndex;+ matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);+ if (dictMatchIndex+matchLength >= dictHighLimit)+ match = base + dictMatchIndex + dictIndexDelta; /* to prepare for next usage of match[matchLength] */++ if (matchLength > bestLength) {+ U32 matchIndex = dictMatchIndex + dictIndexDelta;+ if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) ) {+ DEBUGLOG(9, "ZSTD_DUBT_findBetterDictMatch(%u) : found better match length %u -> %u and offsetCode %u -> %u (dictMatchIndex %u, matchIndex %u)",+ current, (U32)bestLength, (U32)matchLength, (U32)*offsetPtr, ZSTD_REP_MOVE + current - matchIndex, dictMatchIndex, matchIndex);+ bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;+ }+ if (ip+matchLength == iend) { /* reached end of input : ip[matchLength] is not valid, no way to know if it's larger or smaller than match */+ break; /* drop, to guarantee consistency (miss a little bit of compression) */+ }+ }++ if (match[matchLength] < ip[matchLength]) {+ if (dictMatchIndex <= btLow) { break; } /* beyond tree size, stop the search */+ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */+ dictMatchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */+ } else {+ /* match is larger than current */+ if (dictMatchIndex <= btLow) { break; } /* beyond tree size, stop the search */+ commonLengthLarger = matchLength;+ dictMatchIndex = nextPtr[0];+ }+ }++ if (bestLength >= MINMATCH) {+ U32 const mIndex = current - ((U32)*offsetPtr - ZSTD_REP_MOVE); (void)mIndex;+ DEBUGLOG(8, "ZSTD_DUBT_findBetterDictMatch(%u) : found match of length %u and offsetCode %u (pos %u)",+ current, (U32)bestLength, (U32)*offsetPtr, mIndex);+ }+ return bestLength;++}+++static size_t+ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iend,+ size_t* offsetPtr,+ U32 const mls,+ const ZSTD_dictMode_e dictMode)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hashLog = cParams->hashLog;+ size_t const h = ZSTD_hashPtr(ip, hashLog, mls);+ U32 matchIndex = hashTable[h];++ const BYTE* const base = ms->window.base;+ U32 const current = (U32)(ip-base);+ U32 const windowLow = ms->window.lowLimit;++ U32* const bt = ms->chainTable;+ U32 const btLog = cParams->chainLog - 1;+ U32 const btMask = (1 << btLog) - 1;+ U32 const btLow = (btMask >= current) ? 0 : current - btMask;+ U32 const unsortLimit = MAX(btLow, windowLow);++ U32* nextCandidate = bt + 2*(matchIndex&btMask);+ U32* unsortedMark = bt + 2*(matchIndex&btMask) + 1;+ U32 nbCompares = 1U << cParams->searchLog;+ U32 nbCandidates = nbCompares;+ U32 previousCandidate = 0;++ DEBUGLOG(7, "ZSTD_DUBT_findBestMatch (%u) ", current);+ assert(ip <= iend-8); /* required for h calculation */++ /* reach end of unsorted candidates list */+ while ( (matchIndex > unsortLimit)+ && (*unsortedMark == ZSTD_DUBT_UNSORTED_MARK)+ && (nbCandidates > 1) ) {+ DEBUGLOG(8, "ZSTD_DUBT_findBestMatch: candidate %u is unsorted",+ matchIndex);+ *unsortedMark = previousCandidate; /* the unsortedMark becomes a reversed chain, to move up back to original position */+ previousCandidate = matchIndex;+ matchIndex = *nextCandidate;+ nextCandidate = bt + 2*(matchIndex&btMask);+ unsortedMark = bt + 2*(matchIndex&btMask) + 1;+ nbCandidates --;+ }++ /* nullify last candidate if it's still unsorted+ * simplification, detrimental to compression ratio, beneficial for speed */+ if ( (matchIndex > unsortLimit)+ && (*unsortedMark==ZSTD_DUBT_UNSORTED_MARK) ) {+ DEBUGLOG(7, "ZSTD_DUBT_findBestMatch: nullify last unsorted candidate %u",+ matchIndex);+ *nextCandidate = *unsortedMark = 0;+ }++ /* batch sort stacked candidates */+ matchIndex = previousCandidate;+ while (matchIndex) { /* will end on matchIndex == 0 */+ U32* const nextCandidateIdxPtr = bt + 2*(matchIndex&btMask) + 1;+ U32 const nextCandidateIdx = *nextCandidateIdxPtr;+ ZSTD_insertDUBT1(ms, matchIndex, iend,+ nbCandidates, unsortLimit, dictMode);+ matchIndex = nextCandidateIdx;+ nbCandidates++;+ }++ /* find longest match */+ { size_t commonLengthSmaller = 0, commonLengthLarger = 0;+ const BYTE* const dictBase = ms->window.dictBase;+ const U32 dictLimit = ms->window.dictLimit;+ const BYTE* const dictEnd = dictBase + dictLimit;+ const BYTE* const prefixStart = base + dictLimit;+ U32* smallerPtr = bt + 2*(current&btMask);+ U32* largerPtr = bt + 2*(current&btMask) + 1;+ U32 matchEndIdx = current + 8 + 1;+ U32 dummy32; /* to be nullified at the end */+ size_t bestLength = 0;++ matchIndex = hashTable[h];+ hashTable[h] = current; /* Update Hash Table */++ while (nbCompares-- && (matchIndex > windowLow)) {+ U32* const nextPtr = bt + 2*(matchIndex & btMask);+ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */+ const BYTE* match;++ if ((dictMode != ZSTD_extDict) || (matchIndex+matchLength >= dictLimit)) {+ match = base + matchIndex;+ matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);+ } else {+ match = dictBase + matchIndex;+ matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);+ if (matchIndex+matchLength >= dictLimit)+ match = base + matchIndex; /* to prepare for next usage of match[matchLength] */+ }++ if (matchLength > bestLength) {+ if (matchLength > matchEndIdx - matchIndex)+ matchEndIdx = matchIndex + (U32)matchLength;+ if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )+ bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;+ if (ip+matchLength == iend) { /* equal : no way to know if inf or sup */+ if (dictMode == ZSTD_dictMatchState) {+ nbCompares = 0; /* in addition to avoiding checking any+ * further in this loop, make sure we+ * skip checking in the dictionary. */+ }+ break; /* drop, to guarantee consistency (miss a little bit of compression) */+ }+ }++ if (match[matchLength] < ip[matchLength]) {+ /* match is smaller than current */+ *smallerPtr = matchIndex; /* update smaller idx */+ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */+ if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */+ smallerPtr = nextPtr+1; /* new "smaller" => larger of match */+ matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */+ } else {+ /* match is larger than current */+ *largerPtr = matchIndex;+ commonLengthLarger = matchLength;+ if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */+ largerPtr = nextPtr;+ matchIndex = nextPtr[0];+ } }++ *smallerPtr = *largerPtr = 0;++ if (dictMode == ZSTD_dictMatchState && nbCompares) {+ bestLength = ZSTD_DUBT_findBetterDictMatch(+ ms, ip, iend,+ offsetPtr, bestLength, nbCompares,+ mls, dictMode);+ }++ assert(matchEndIdx > current+8); /* ensure nextToUpdate is increased */+ ms->nextToUpdate = matchEndIdx - 8; /* skip repetitive patterns */+ if (bestLength >= MINMATCH) {+ U32 const mIndex = current - ((U32)*offsetPtr - ZSTD_REP_MOVE); (void)mIndex;+ DEBUGLOG(8, "ZSTD_DUBT_findBestMatch(%u) : found match of length %u and offsetCode %u (pos %u)",+ current, (U32)bestLength, (U32)*offsetPtr, mIndex);+ }+ return bestLength;+ }+}+++/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */+FORCE_INLINE_TEMPLATE size_t+ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iLimit,+ size_t* offsetPtr,+ const U32 mls /* template */,+ const ZSTD_dictMode_e dictMode)+{+ DEBUGLOG(7, "ZSTD_BtFindBestMatch");+ if (ip < ms->window.base + ms->nextToUpdate) return 0; /* skipped area */+ ZSTD_updateDUBT(ms, ip, iLimit, mls);+ return ZSTD_DUBT_findBestMatch(ms, ip, iLimit, offsetPtr, mls, dictMode);+}+++static size_t+ZSTD_BtFindBestMatch_selectMLS ( ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iLimit,+ size_t* offsetPtr)+{+ switch(ms->cParams.minMatch)+ {+ default : /* includes case 3 */+ case 4 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 4, ZSTD_noDict);+ case 5 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 5, ZSTD_noDict);+ case 7 :+ case 6 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 6, ZSTD_noDict);+ }+}+++static size_t ZSTD_BtFindBestMatch_dictMatchState_selectMLS (+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iLimit,+ size_t* offsetPtr)+{+ switch(ms->cParams.minMatch)+ {+ default : /* includes case 3 */+ case 4 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 4, ZSTD_dictMatchState);+ case 5 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 5, ZSTD_dictMatchState);+ case 7 :+ case 6 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 6, ZSTD_dictMatchState);+ }+}+++static size_t ZSTD_BtFindBestMatch_extDict_selectMLS (+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iLimit,+ size_t* offsetPtr)+{+ switch(ms->cParams.minMatch)+ {+ default : /* includes case 3 */+ case 4 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 4, ZSTD_extDict);+ case 5 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 5, ZSTD_extDict);+ case 7 :+ case 6 : return ZSTD_BtFindBestMatch(ms, ip, iLimit, offsetPtr, 6, ZSTD_extDict);+ }+}++++/* *********************************+* Hash Chain+***********************************/+#define NEXT_IN_CHAIN(d, mask) chainTable[(d) & (mask)]++/* Update chains up to ip (excluded)+ Assumption : always within prefix (i.e. not within extDict) */+static U32 ZSTD_insertAndFindFirstIndex_internal(+ ZSTD_matchState_t* ms,+ const ZSTD_compressionParameters* const cParams,+ const BYTE* ip, U32 const mls)+{+ U32* const hashTable = ms->hashTable;+ const U32 hashLog = cParams->hashLog;+ U32* const chainTable = ms->chainTable;+ const U32 chainMask = (1 << cParams->chainLog) - 1;+ const BYTE* const base = ms->window.base;+ const U32 target = (U32)(ip - base);+ U32 idx = ms->nextToUpdate;++ while(idx < target) { /* catch up */+ size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);+ NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];+ hashTable[h] = idx;+ idx++;+ }++ ms->nextToUpdate = target;+ return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];+}++U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) {+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ return ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, ms->cParams.minMatch);+}+++/* inlining is important to hardwire a hot branch (template emulation) */+FORCE_INLINE_TEMPLATE+size_t ZSTD_HcFindBestMatch_generic (+ ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iLimit,+ size_t* offsetPtr,+ const U32 mls, const ZSTD_dictMode_e dictMode)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const chainTable = ms->chainTable;+ const U32 chainSize = (1 << cParams->chainLog);+ const U32 chainMask = chainSize-1;+ const BYTE* const base = ms->window.base;+ const BYTE* const dictBase = ms->window.dictBase;+ const U32 dictLimit = ms->window.dictLimit;+ const BYTE* const prefixStart = base + dictLimit;+ const BYTE* const dictEnd = dictBase + dictLimit;+ const U32 lowLimit = ms->window.lowLimit;+ const U32 current = (U32)(ip-base);+ const U32 minChain = current > chainSize ? current - chainSize : 0;+ U32 nbAttempts = 1U << cParams->searchLog;+ size_t ml=4-1;++ /* HC4 match finder */+ U32 matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls);++ for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) {+ size_t currentMl=0;+ if ((dictMode != ZSTD_extDict) || matchIndex >= dictLimit) {+ const BYTE* const match = base + matchIndex;+ assert(matchIndex >= dictLimit); /* ensures this is true if dictMode != ZSTD_extDict */+ if (match[ml] == ip[ml]) /* potentially better */+ currentMl = ZSTD_count(ip, match, iLimit);+ } else {+ const BYTE* const match = dictBase + matchIndex;+ assert(match+4 <= dictEnd);+ if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */+ currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4;+ }++ /* save best solution */+ if (currentMl > ml) {+ ml = currentMl;+ *offsetPtr = current - matchIndex + ZSTD_REP_MOVE;+ if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */+ }++ if (matchIndex <= minChain) break;+ matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);+ }++ if (dictMode == ZSTD_dictMatchState) {+ const ZSTD_matchState_t* const dms = ms->dictMatchState;+ const U32* const dmsChainTable = dms->chainTable;+ const U32 dmsChainSize = (1 << dms->cParams.chainLog);+ const U32 dmsChainMask = dmsChainSize - 1;+ const U32 dmsLowestIndex = dms->window.dictLimit;+ const BYTE* const dmsBase = dms->window.base;+ const BYTE* const dmsEnd = dms->window.nextSrc;+ const U32 dmsSize = (U32)(dmsEnd - dmsBase);+ const U32 dmsIndexDelta = dictLimit - dmsSize;+ const U32 dmsMinChain = dmsSize > dmsChainSize ? dmsSize - dmsChainSize : 0;++ matchIndex = dms->hashTable[ZSTD_hashPtr(ip, dms->cParams.hashLog, mls)];++ for ( ; (matchIndex>dmsLowestIndex) & (nbAttempts>0) ; nbAttempts--) {+ size_t currentMl=0;+ const BYTE* const match = dmsBase + matchIndex;+ assert(match+4 <= dmsEnd);+ if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */+ currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dmsEnd, prefixStart) + 4;++ /* save best solution */+ if (currentMl > ml) {+ ml = currentMl;+ *offsetPtr = current - (matchIndex + dmsIndexDelta) + ZSTD_REP_MOVE;+ if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */+ }++ if (matchIndex <= dmsMinChain) break;+ matchIndex = dmsChainTable[matchIndex & dmsChainMask];+ }+ }++ return ml;+}+++FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_selectMLS (+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iLimit,+ size_t* offsetPtr)+{+ switch(ms->cParams.minMatch)+ {+ default : /* includes case 3 */+ case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_noDict);+ case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_noDict);+ case 7 :+ case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_noDict);+ }+}+++static size_t ZSTD_HcFindBestMatch_dictMatchState_selectMLS (+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iLimit,+ size_t* offsetPtr)+{+ switch(ms->cParams.minMatch)+ {+ default : /* includes case 3 */+ case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_dictMatchState);+ case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_dictMatchState);+ case 7 :+ case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_dictMatchState);+ }+}+++FORCE_INLINE_TEMPLATE size_t ZSTD_HcFindBestMatch_extDict_selectMLS (+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iLimit,+ size_t* offsetPtr)+{+ switch(ms->cParams.minMatch)+ {+ default : /* includes case 3 */+ case 4 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 4, ZSTD_extDict);+ case 5 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 5, ZSTD_extDict);+ case 7 :+ case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_extDict);+ }+}+++/* *******************************+* Common parser - lazy strategy+*********************************/+FORCE_INLINE_TEMPLATE+size_t ZSTD_compressBlock_lazy_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore,+ U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize,+ const U32 searchMethod, const U32 depth,+ ZSTD_dictMode_e const dictMode)+{+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - 8;+ const BYTE* const base = ms->window.base;+ const U32 prefixLowestIndex = ms->window.dictLimit;+ const BYTE* const prefixLowest = base + prefixLowestIndex;++ typedef size_t (*searchMax_f)(+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);+ searchMax_f const searchMax = dictMode == ZSTD_dictMatchState ?+ (searchMethod ? ZSTD_BtFindBestMatch_dictMatchState_selectMLS : ZSTD_HcFindBestMatch_dictMatchState_selectMLS) :+ (searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS);+ U32 offset_1 = rep[0], offset_2 = rep[1], savedOffset=0;++ const ZSTD_matchState_t* const dms = ms->dictMatchState;+ const U32 dictLowestIndex = dictMode == ZSTD_dictMatchState ?+ dms->window.dictLimit : 0;+ const BYTE* const dictBase = dictMode == ZSTD_dictMatchState ?+ dms->window.base : NULL;+ const BYTE* const dictLowest = dictMode == ZSTD_dictMatchState ?+ dictBase + dictLowestIndex : NULL;+ const BYTE* const dictEnd = dictMode == ZSTD_dictMatchState ?+ dms->window.nextSrc : NULL;+ const U32 dictIndexDelta = dictMode == ZSTD_dictMatchState ?+ prefixLowestIndex - (U32)(dictEnd - dictBase) :+ 0;+ const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictLowest);++ /* init */+ ip += (dictAndPrefixLength == 0);+ ms->nextToUpdate3 = ms->nextToUpdate;+ if (dictMode == ZSTD_noDict) {+ U32 const maxRep = (U32)(ip - prefixLowest);+ if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;+ if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0;+ }+ if (dictMode == ZSTD_dictMatchState) {+ /* dictMatchState repCode checks don't currently handle repCode == 0+ * disabling. */+ assert(offset_1 <= dictAndPrefixLength);+ assert(offset_2 <= dictAndPrefixLength);+ }++ /* Match Loop */+ while (ip < ilimit) {+ size_t matchLength=0;+ size_t offset=0;+ const BYTE* start=ip+1;++ /* check repCode */+ if (dictMode == ZSTD_dictMatchState) {+ const U32 repIndex = (U32)(ip - base) + 1 - offset_1;+ const BYTE* repMatch = (dictMode == ZSTD_dictMatchState+ && repIndex < prefixLowestIndex) ?+ dictBase + (repIndex - dictIndexDelta) :+ base + repIndex;+ if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)+ && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {+ const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;+ matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;+ if (depth==0) goto _storeSequence;+ }+ }+ if ( dictMode == ZSTD_noDict+ && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {+ matchLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;+ if (depth==0) goto _storeSequence;+ }++ /* first search (depth 0) */+ { size_t offsetFound = 999999999;+ size_t const ml2 = searchMax(ms, ip, iend, &offsetFound);+ if (ml2 > matchLength)+ matchLength = ml2, start = ip, offset=offsetFound;+ }++ if (matchLength < 4) {+ ip += ((ip-anchor) >> kSearchStrength) + 1; /* jump faster over incompressible sections */+ continue;+ }++ /* let's try to find a better solution */+ if (depth>=1)+ while (ip<ilimit) {+ ip ++;+ if ( (dictMode == ZSTD_noDict)+ && (offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {+ size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;+ int const gain2 = (int)(mlRep * 3);+ int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);+ if ((mlRep >= 4) && (gain2 > gain1))+ matchLength = mlRep, offset = 0, start = ip;+ }+ if (dictMode == ZSTD_dictMatchState) {+ const U32 repIndex = (U32)(ip - base) - offset_1;+ const BYTE* repMatch = repIndex < prefixLowestIndex ?+ dictBase + (repIndex - dictIndexDelta) :+ base + repIndex;+ if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)+ && (MEM_read32(repMatch) == MEM_read32(ip)) ) {+ const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;+ size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;+ int const gain2 = (int)(mlRep * 3);+ int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);+ if ((mlRep >= 4) && (gain2 > gain1))+ matchLength = mlRep, offset = 0, start = ip;+ }+ }+ { size_t offset2=999999999;+ size_t const ml2 = searchMax(ms, ip, iend, &offset2);+ int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);+ if ((ml2 >= 4) && (gain2 > gain1)) {+ matchLength = ml2, offset = offset2, start = ip;+ continue; /* search a better one */+ } }++ /* let's find an even better one */+ if ((depth==2) && (ip<ilimit)) {+ ip ++;+ if ( (dictMode == ZSTD_noDict)+ && (offset) && ((offset_1>0) & (MEM_read32(ip) == MEM_read32(ip - offset_1)))) {+ size_t const mlRep = ZSTD_count(ip+4, ip+4-offset_1, iend) + 4;+ int const gain2 = (int)(mlRep * 4);+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);+ if ((mlRep >= 4) && (gain2 > gain1))+ matchLength = mlRep, offset = 0, start = ip;+ }+ if (dictMode == ZSTD_dictMatchState) {+ const U32 repIndex = (U32)(ip - base) - offset_1;+ const BYTE* repMatch = repIndex < prefixLowestIndex ?+ dictBase + (repIndex - dictIndexDelta) :+ base + repIndex;+ if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)+ && (MEM_read32(repMatch) == MEM_read32(ip)) ) {+ const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;+ size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;+ int const gain2 = (int)(mlRep * 4);+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);+ if ((mlRep >= 4) && (gain2 > gain1))+ matchLength = mlRep, offset = 0, start = ip;+ }+ }+ { size_t offset2=999999999;+ size_t const ml2 = searchMax(ms, ip, iend, &offset2);+ int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);+ if ((ml2 >= 4) && (gain2 > gain1)) {+ matchLength = ml2, offset = offset2, start = ip;+ continue;+ } } }+ break; /* nothing found : store previous solution */+ }++ /* NOTE:+ * start[-offset+ZSTD_REP_MOVE-1] is undefined behavior.+ * (-offset+ZSTD_REP_MOVE-1) is unsigned, and is added to start, which+ * overflows the pointer, which is undefined behavior.+ */+ /* catch up */+ if (offset) {+ if (dictMode == ZSTD_noDict) {+ while ( ((start > anchor) & (start - (offset-ZSTD_REP_MOVE) > prefixLowest))+ && (start[-1] == (start-(offset-ZSTD_REP_MOVE))[-1]) ) /* only search for offset within prefix */+ { start--; matchLength++; }+ }+ if (dictMode == ZSTD_dictMatchState) {+ U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));+ const BYTE* match = (matchIndex < prefixLowestIndex) ? dictBase + matchIndex - dictIndexDelta : base + matchIndex;+ const BYTE* const mStart = (matchIndex < prefixLowestIndex) ? dictLowest : prefixLowest;+ while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; } /* catch up */+ }+ offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);+ }+ /* store sequence */+_storeSequence:+ { size_t const litLength = start - anchor;+ ZSTD_storeSeq(seqStore, litLength, anchor, (U32)offset, matchLength-MINMATCH);+ anchor = ip = start + matchLength;+ }++ /* check immediate repcode */+ if (dictMode == ZSTD_dictMatchState) {+ while (ip <= ilimit) {+ U32 const current2 = (U32)(ip-base);+ U32 const repIndex = current2 - offset_2;+ const BYTE* repMatch = dictMode == ZSTD_dictMatchState+ && repIndex < prefixLowestIndex ?+ dictBase - dictIndexDelta + repIndex :+ base + repIndex;+ if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex) >= 3 /* intentional overflow */)+ && (MEM_read32(repMatch) == MEM_read32(ip)) ) {+ const BYTE* const repEnd2 = repIndex < prefixLowestIndex ? dictEnd : iend;+ matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd2, prefixLowest) + 4;+ offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap offset_2 <=> offset_1 */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, matchLength-MINMATCH);+ ip += matchLength;+ anchor = ip;+ continue;+ }+ break;+ }+ }++ if (dictMode == ZSTD_noDict) {+ while ( ((ip <= ilimit) & (offset_2>0))+ && (MEM_read32(ip) == MEM_read32(ip - offset_2)) ) {+ /* store sequence */+ matchLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;+ offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap repcodes */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, matchLength-MINMATCH);+ ip += matchLength;+ anchor = ip;+ continue; /* faster when present ... (?) */+ } } }++ /* Save reps for next block */+ rep[0] = offset_1 ? offset_1 : savedOffset;+ rep[1] = offset_2 ? offset_2 : savedOffset;++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_btlazy2(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 1, 2, ZSTD_noDict);+}++size_t ZSTD_compressBlock_lazy2(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 2, ZSTD_noDict);+}++size_t ZSTD_compressBlock_lazy(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 1, ZSTD_noDict);+}++size_t ZSTD_compressBlock_greedy(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 0, ZSTD_noDict);+}++size_t ZSTD_compressBlock_btlazy2_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 1, 2, ZSTD_dictMatchState);+}++size_t ZSTD_compressBlock_lazy2_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 2, ZSTD_dictMatchState);+}++size_t ZSTD_compressBlock_lazy_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 1, ZSTD_dictMatchState);+}++size_t ZSTD_compressBlock_greedy_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 0, ZSTD_dictMatchState);+}+++FORCE_INLINE_TEMPLATE+size_t ZSTD_compressBlock_lazy_extDict_generic(+ ZSTD_matchState_t* ms, seqStore_t* seqStore,+ U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize,+ const U32 searchMethod, const U32 depth)+{+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - 8;+ const BYTE* const base = ms->window.base;+ const U32 dictLimit = ms->window.dictLimit;+ const U32 lowestIndex = ms->window.lowLimit;+ const BYTE* const prefixStart = base + dictLimit;+ const BYTE* const dictBase = ms->window.dictBase;+ const BYTE* const dictEnd = dictBase + dictLimit;+ const BYTE* const dictStart = dictBase + lowestIndex;++ typedef size_t (*searchMax_f)(+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);+ searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_extDict_selectMLS : ZSTD_HcFindBestMatch_extDict_selectMLS;++ U32 offset_1 = rep[0], offset_2 = rep[1];++ /* init */+ ms->nextToUpdate3 = ms->nextToUpdate;+ ip += (ip == prefixStart);++ /* Match Loop */+ while (ip < ilimit) {+ size_t matchLength=0;+ size_t offset=0;+ const BYTE* start=ip+1;+ U32 current = (U32)(ip-base);++ /* check repCode */+ { const U32 repIndex = (U32)(current+1 - offset_1);+ const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;+ const BYTE* const repMatch = repBase + repIndex;+ if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */+ if (MEM_read32(ip+1) == MEM_read32(repMatch)) {+ /* repcode detected we should take it */+ const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;+ matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repEnd, prefixStart) + 4;+ if (depth==0) goto _storeSequence;+ } }++ /* first search (depth 0) */+ { size_t offsetFound = 999999999;+ size_t const ml2 = searchMax(ms, ip, iend, &offsetFound);+ if (ml2 > matchLength)+ matchLength = ml2, start = ip, offset=offsetFound;+ }++ if (matchLength < 4) {+ ip += ((ip-anchor) >> kSearchStrength) + 1; /* jump faster over incompressible sections */+ continue;+ }++ /* let's try to find a better solution */+ if (depth>=1)+ while (ip<ilimit) {+ ip ++;+ current++;+ /* check repCode */+ if (offset) {+ const U32 repIndex = (U32)(current - offset_1);+ const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;+ const BYTE* const repMatch = repBase + repIndex;+ if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */+ if (MEM_read32(ip) == MEM_read32(repMatch)) {+ /* repcode detected */+ const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;+ size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;+ int const gain2 = (int)(repLength * 3);+ int const gain1 = (int)(matchLength*3 - ZSTD_highbit32((U32)offset+1) + 1);+ if ((repLength >= 4) && (gain2 > gain1))+ matchLength = repLength, offset = 0, start = ip;+ } }++ /* search match, depth 1 */+ { size_t offset2=999999999;+ size_t const ml2 = searchMax(ms, ip, iend, &offset2);+ int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 4);+ if ((ml2 >= 4) && (gain2 > gain1)) {+ matchLength = ml2, offset = offset2, start = ip;+ continue; /* search a better one */+ } }++ /* let's find an even better one */+ if ((depth==2) && (ip<ilimit)) {+ ip ++;+ current++;+ /* check repCode */+ if (offset) {+ const U32 repIndex = (U32)(current - offset_1);+ const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;+ const BYTE* const repMatch = repBase + repIndex;+ if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */+ if (MEM_read32(ip) == MEM_read32(repMatch)) {+ /* repcode detected */+ const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;+ size_t const repLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;+ int const gain2 = (int)(repLength * 4);+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 1);+ if ((repLength >= 4) && (gain2 > gain1))+ matchLength = repLength, offset = 0, start = ip;+ } }++ /* search match, depth 2 */+ { size_t offset2=999999999;+ size_t const ml2 = searchMax(ms, ip, iend, &offset2);+ int const gain2 = (int)(ml2*4 - ZSTD_highbit32((U32)offset2+1)); /* raw approx */+ int const gain1 = (int)(matchLength*4 - ZSTD_highbit32((U32)offset+1) + 7);+ if ((ml2 >= 4) && (gain2 > gain1)) {+ matchLength = ml2, offset = offset2, start = ip;+ continue;+ } } }+ break; /* nothing found : store previous solution */+ }++ /* catch up */+ if (offset) {+ U32 const matchIndex = (U32)((start-base) - (offset - ZSTD_REP_MOVE));+ const BYTE* match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;+ const BYTE* const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;+ while ((start>anchor) && (match>mStart) && (start[-1] == match[-1])) { start--; match--; matchLength++; } /* catch up */+ offset_2 = offset_1; offset_1 = (U32)(offset - ZSTD_REP_MOVE);+ }++ /* store sequence */+_storeSequence:+ { size_t const litLength = start - anchor;+ ZSTD_storeSeq(seqStore, litLength, anchor, (U32)offset, matchLength-MINMATCH);+ anchor = ip = start + matchLength;+ }++ /* check immediate repcode */+ while (ip <= ilimit) {+ const U32 repIndex = (U32)((ip-base) - offset_2);+ const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;+ const BYTE* const repMatch = repBase + repIndex;+ if (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */+ if (MEM_read32(ip) == MEM_read32(repMatch)) {+ /* repcode detected we should take it */+ const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;+ matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;+ offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap offset history */+ ZSTD_storeSeq(seqStore, 0, anchor, 0, matchLength-MINMATCH);+ ip += matchLength;+ anchor = ip;+ continue; /* faster when present ... (?) */+ }+ break;+ } }++ /* Save reps for next block */+ rep[0] = offset_1;+ rep[1] = offset_2;++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_greedy_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 0);+}++size_t ZSTD_compressBlock_lazy_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)++{+ return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 1);+}++size_t ZSTD_compressBlock_lazy2_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)++{+ return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 2);+}++size_t ZSTD_compressBlock_btlazy2_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)++{+ return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 1, 2);+}
+ zstd/lib/compress/zstd_lazy.h view
@@ -0,0 +1,67 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#ifndef ZSTD_LAZY_H+#define ZSTD_LAZY_H++#if defined (__cplusplus)+extern "C" {+#endif++#include "zstd_compress_internal.h"++U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip);++void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue); /*! used in ZSTD_reduceIndex(). preemptively increase value of ZSTD_DUBT_UNSORTED_MARK */++size_t ZSTD_compressBlock_btlazy2(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_lazy2(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_lazy(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_greedy(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);++size_t ZSTD_compressBlock_btlazy2_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_lazy2_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_lazy_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_greedy_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);++size_t ZSTD_compressBlock_greedy_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_lazy_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_lazy2_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_btlazy2_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);++#if defined (__cplusplus)+}+#endif++#endif /* ZSTD_LAZY_H */
+ zstd/lib/compress/zstd_ldm.c view
@@ -0,0 +1,597 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ */++#include "zstd_ldm.h"++#include "debug.h"+#include "zstd_fast.h" /* ZSTD_fillHashTable() */+#include "zstd_double_fast.h" /* ZSTD_fillDoubleHashTable() */++#define LDM_BUCKET_SIZE_LOG 3+#define LDM_MIN_MATCH_LENGTH 64+#define LDM_HASH_RLOG 7+#define LDM_HASH_CHAR_OFFSET 10++void ZSTD_ldm_adjustParameters(ldmParams_t* params,+ ZSTD_compressionParameters const* cParams)+{+ params->windowLog = cParams->windowLog;+ ZSTD_STATIC_ASSERT(LDM_BUCKET_SIZE_LOG <= ZSTD_LDM_BUCKETSIZELOG_MAX);+ DEBUGLOG(4, "ZSTD_ldm_adjustParameters");+ if (!params->bucketSizeLog) params->bucketSizeLog = LDM_BUCKET_SIZE_LOG;+ if (!params->minMatchLength) params->minMatchLength = LDM_MIN_MATCH_LENGTH;+ if (cParams->strategy >= ZSTD_btopt) {+ /* Get out of the way of the optimal parser */+ U32 const minMatch = MAX(cParams->targetLength, params->minMatchLength);+ assert(minMatch >= ZSTD_LDM_MINMATCH_MIN);+ assert(minMatch <= ZSTD_LDM_MINMATCH_MAX);+ params->minMatchLength = minMatch;+ }+ if (params->hashLog == 0) {+ params->hashLog = MAX(ZSTD_HASHLOG_MIN, params->windowLog - LDM_HASH_RLOG);+ assert(params->hashLog <= ZSTD_HASHLOG_MAX);+ }+ if (params->hashRateLog == 0) {+ params->hashRateLog = params->windowLog < params->hashLog+ ? 0+ : params->windowLog - params->hashLog;+ }+ params->bucketSizeLog = MIN(params->bucketSizeLog, params->hashLog);+}++size_t ZSTD_ldm_getTableSize(ldmParams_t params)+{+ size_t const ldmHSize = ((size_t)1) << params.hashLog;+ size_t const ldmBucketSizeLog = MIN(params.bucketSizeLog, params.hashLog);+ size_t const ldmBucketSize =+ ((size_t)1) << (params.hashLog - ldmBucketSizeLog);+ size_t const totalSize = ldmBucketSize + ldmHSize * sizeof(ldmEntry_t);+ return params.enableLdm ? totalSize : 0;+}++size_t ZSTD_ldm_getMaxNbSeq(ldmParams_t params, size_t maxChunkSize)+{+ return params.enableLdm ? (maxChunkSize / params.minMatchLength) : 0;+}++/** ZSTD_ldm_getSmallHash() :+ * numBits should be <= 32+ * If numBits==0, returns 0.+ * @return : the most significant numBits of value. */+static U32 ZSTD_ldm_getSmallHash(U64 value, U32 numBits)+{+ assert(numBits <= 32);+ return numBits == 0 ? 0 : (U32)(value >> (64 - numBits));+}++/** ZSTD_ldm_getChecksum() :+ * numBitsToDiscard should be <= 32+ * @return : the next most significant 32 bits after numBitsToDiscard */+static U32 ZSTD_ldm_getChecksum(U64 hash, U32 numBitsToDiscard)+{+ assert(numBitsToDiscard <= 32);+ return (hash >> (64 - 32 - numBitsToDiscard)) & 0xFFFFFFFF;+}++/** ZSTD_ldm_getTag() ;+ * Given the hash, returns the most significant numTagBits bits+ * after (32 + hbits) bits.+ *+ * If there are not enough bits remaining, return the last+ * numTagBits bits. */+static U32 ZSTD_ldm_getTag(U64 hash, U32 hbits, U32 numTagBits)+{+ assert(numTagBits < 32 && hbits <= 32);+ if (32 - hbits < numTagBits) {+ return hash & (((U32)1 << numTagBits) - 1);+ } else {+ return (hash >> (32 - hbits - numTagBits)) & (((U32)1 << numTagBits) - 1);+ }+}++/** ZSTD_ldm_getBucket() :+ * Returns a pointer to the start of the bucket associated with hash. */+static ldmEntry_t* ZSTD_ldm_getBucket(+ ldmState_t* ldmState, size_t hash, ldmParams_t const ldmParams)+{+ return ldmState->hashTable + (hash << ldmParams.bucketSizeLog);+}++/** ZSTD_ldm_insertEntry() :+ * Insert the entry with corresponding hash into the hash table */+static void ZSTD_ldm_insertEntry(ldmState_t* ldmState,+ size_t const hash, const ldmEntry_t entry,+ ldmParams_t const ldmParams)+{+ BYTE* const bucketOffsets = ldmState->bucketOffsets;+ *(ZSTD_ldm_getBucket(ldmState, hash, ldmParams) + bucketOffsets[hash]) = entry;+ bucketOffsets[hash]++;+ bucketOffsets[hash] &= ((U32)1 << ldmParams.bucketSizeLog) - 1;+}++/** ZSTD_ldm_makeEntryAndInsertByTag() :+ *+ * Gets the small hash, checksum, and tag from the rollingHash.+ *+ * If the tag matches (1 << ldmParams.hashRateLog)-1, then+ * creates an ldmEntry from the offset, and inserts it into the hash table.+ *+ * hBits is the length of the small hash, which is the most significant hBits+ * of rollingHash. The checksum is the next 32 most significant bits, followed+ * by ldmParams.hashRateLog bits that make up the tag. */+static void ZSTD_ldm_makeEntryAndInsertByTag(ldmState_t* ldmState,+ U64 const rollingHash,+ U32 const hBits,+ U32 const offset,+ ldmParams_t const ldmParams)+{+ U32 const tag = ZSTD_ldm_getTag(rollingHash, hBits, ldmParams.hashRateLog);+ U32 const tagMask = ((U32)1 << ldmParams.hashRateLog) - 1;+ if (tag == tagMask) {+ U32 const hash = ZSTD_ldm_getSmallHash(rollingHash, hBits);+ U32 const checksum = ZSTD_ldm_getChecksum(rollingHash, hBits);+ ldmEntry_t entry;+ entry.offset = offset;+ entry.checksum = checksum;+ ZSTD_ldm_insertEntry(ldmState, hash, entry, ldmParams);+ }+}++/** ZSTD_ldm_countBackwardsMatch() :+ * Returns the number of bytes that match backwards before pIn and pMatch.+ *+ * We count only bytes where pMatch >= pBase and pIn >= pAnchor. */+static size_t ZSTD_ldm_countBackwardsMatch(+ const BYTE* pIn, const BYTE* pAnchor,+ const BYTE* pMatch, const BYTE* pBase)+{+ size_t matchLength = 0;+ while (pIn > pAnchor && pMatch > pBase && pIn[-1] == pMatch[-1]) {+ pIn--;+ pMatch--;+ matchLength++;+ }+ return matchLength;+}++/** ZSTD_ldm_fillFastTables() :+ *+ * Fills the relevant tables for the ZSTD_fast and ZSTD_dfast strategies.+ * This is similar to ZSTD_loadDictionaryContent.+ *+ * The tables for the other strategies are filled within their+ * block compressors. */+static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms,+ void const* end)+{+ const BYTE* const iend = (const BYTE*)end;++ switch(ms->cParams.strategy)+ {+ case ZSTD_fast:+ ZSTD_fillHashTable(ms, iend, ZSTD_dtlm_fast);+ break;++ case ZSTD_dfast:+ ZSTD_fillDoubleHashTable(ms, iend, ZSTD_dtlm_fast);+ break;++ case ZSTD_greedy:+ case ZSTD_lazy:+ case ZSTD_lazy2:+ case ZSTD_btlazy2:+ case ZSTD_btopt:+ case ZSTD_btultra:+ case ZSTD_btultra2:+ break;+ default:+ assert(0); /* not possible : not a valid strategy id */+ }++ return 0;+}++/** ZSTD_ldm_fillLdmHashTable() :+ *+ * Fills hashTable from (lastHashed + 1) to iend (non-inclusive).+ * lastHash is the rolling hash that corresponds to lastHashed.+ *+ * Returns the rolling hash corresponding to position iend-1. */+static U64 ZSTD_ldm_fillLdmHashTable(ldmState_t* state,+ U64 lastHash, const BYTE* lastHashed,+ const BYTE* iend, const BYTE* base,+ U32 hBits, ldmParams_t const ldmParams)+{+ U64 rollingHash = lastHash;+ const BYTE* cur = lastHashed + 1;++ while (cur < iend) {+ rollingHash = ZSTD_rollingHash_rotate(rollingHash, cur[-1],+ cur[ldmParams.minMatchLength-1],+ state->hashPower);+ ZSTD_ldm_makeEntryAndInsertByTag(state,+ rollingHash, hBits,+ (U32)(cur - base), ldmParams);+ ++cur;+ }+ return rollingHash;+}+++/** ZSTD_ldm_limitTableUpdate() :+ *+ * Sets cctx->nextToUpdate to a position corresponding closer to anchor+ * if it is far way+ * (after a long match, only update tables a limited amount). */+static void ZSTD_ldm_limitTableUpdate(ZSTD_matchState_t* ms, const BYTE* anchor)+{+ U32 const current = (U32)(anchor - ms->window.base);+ if (current > ms->nextToUpdate + 1024) {+ ms->nextToUpdate =+ current - MIN(512, current - ms->nextToUpdate - 1024);+ }+}++static size_t ZSTD_ldm_generateSequences_internal(+ ldmState_t* ldmState, rawSeqStore_t* rawSeqStore,+ ldmParams_t const* params, void const* src, size_t srcSize)+{+ /* LDM parameters */+ int const extDict = ZSTD_window_hasExtDict(ldmState->window);+ U32 const minMatchLength = params->minMatchLength;+ U64 const hashPower = ldmState->hashPower;+ U32 const hBits = params->hashLog - params->bucketSizeLog;+ U32 const ldmBucketSize = 1U << params->bucketSizeLog;+ U32 const hashRateLog = params->hashRateLog;+ U32 const ldmTagMask = (1U << params->hashRateLog) - 1;+ /* Prefix and extDict parameters */+ U32 const dictLimit = ldmState->window.dictLimit;+ U32 const lowestIndex = extDict ? ldmState->window.lowLimit : dictLimit;+ BYTE const* const base = ldmState->window.base;+ BYTE const* const dictBase = extDict ? ldmState->window.dictBase : NULL;+ BYTE const* const dictStart = extDict ? dictBase + lowestIndex : NULL;+ BYTE const* const dictEnd = extDict ? dictBase + dictLimit : NULL;+ BYTE const* const lowPrefixPtr = base + dictLimit;+ /* Input bounds */+ BYTE const* const istart = (BYTE const*)src;+ BYTE const* const iend = istart + srcSize;+ BYTE const* const ilimit = iend - MAX(minMatchLength, HASH_READ_SIZE);+ /* Input positions */+ BYTE const* anchor = istart;+ BYTE const* ip = istart;+ /* Rolling hash */+ BYTE const* lastHashed = NULL;+ U64 rollingHash = 0;++ while (ip <= ilimit) {+ size_t mLength;+ U32 const current = (U32)(ip - base);+ size_t forwardMatchLength = 0, backwardMatchLength = 0;+ ldmEntry_t* bestEntry = NULL;+ if (ip != istart) {+ rollingHash = ZSTD_rollingHash_rotate(rollingHash, lastHashed[0],+ lastHashed[minMatchLength],+ hashPower);+ } else {+ rollingHash = ZSTD_rollingHash_compute(ip, minMatchLength);+ }+ lastHashed = ip;++ /* Do not insert and do not look for a match */+ if (ZSTD_ldm_getTag(rollingHash, hBits, hashRateLog) != ldmTagMask) {+ ip++;+ continue;+ }++ /* Get the best entry and compute the match lengths */+ {+ ldmEntry_t* const bucket =+ ZSTD_ldm_getBucket(ldmState,+ ZSTD_ldm_getSmallHash(rollingHash, hBits),+ *params);+ ldmEntry_t* cur;+ size_t bestMatchLength = 0;+ U32 const checksum = ZSTD_ldm_getChecksum(rollingHash, hBits);++ for (cur = bucket; cur < bucket + ldmBucketSize; ++cur) {+ size_t curForwardMatchLength, curBackwardMatchLength,+ curTotalMatchLength;+ if (cur->checksum != checksum || cur->offset <= lowestIndex) {+ continue;+ }+ if (extDict) {+ BYTE const* const curMatchBase =+ cur->offset < dictLimit ? dictBase : base;+ BYTE const* const pMatch = curMatchBase + cur->offset;+ BYTE const* const matchEnd =+ cur->offset < dictLimit ? dictEnd : iend;+ BYTE const* const lowMatchPtr =+ cur->offset < dictLimit ? dictStart : lowPrefixPtr;++ curForwardMatchLength = ZSTD_count_2segments(+ ip, pMatch, iend,+ matchEnd, lowPrefixPtr);+ if (curForwardMatchLength < minMatchLength) {+ continue;+ }+ curBackwardMatchLength =+ ZSTD_ldm_countBackwardsMatch(ip, anchor, pMatch,+ lowMatchPtr);+ curTotalMatchLength = curForwardMatchLength ++ curBackwardMatchLength;+ } else { /* !extDict */+ BYTE const* const pMatch = base + cur->offset;+ curForwardMatchLength = ZSTD_count(ip, pMatch, iend);+ if (curForwardMatchLength < minMatchLength) {+ continue;+ }+ curBackwardMatchLength =+ ZSTD_ldm_countBackwardsMatch(ip, anchor, pMatch,+ lowPrefixPtr);+ curTotalMatchLength = curForwardMatchLength ++ curBackwardMatchLength;+ }++ if (curTotalMatchLength > bestMatchLength) {+ bestMatchLength = curTotalMatchLength;+ forwardMatchLength = curForwardMatchLength;+ backwardMatchLength = curBackwardMatchLength;+ bestEntry = cur;+ }+ }+ }++ /* No match found -- continue searching */+ if (bestEntry == NULL) {+ ZSTD_ldm_makeEntryAndInsertByTag(ldmState, rollingHash,+ hBits, current,+ *params);+ ip++;+ continue;+ }++ /* Match found */+ mLength = forwardMatchLength + backwardMatchLength;+ ip -= backwardMatchLength;++ {+ /* Store the sequence:+ * ip = current - backwardMatchLength+ * The match is at (bestEntry->offset - backwardMatchLength)+ */+ U32 const matchIndex = bestEntry->offset;+ U32 const offset = current - matchIndex;+ rawSeq* const seq = rawSeqStore->seq + rawSeqStore->size;++ /* Out of sequence storage */+ if (rawSeqStore->size == rawSeqStore->capacity)+ return ERROR(dstSize_tooSmall);+ seq->litLength = (U32)(ip - anchor);+ seq->matchLength = (U32)mLength;+ seq->offset = offset;+ rawSeqStore->size++;+ }++ /* Insert the current entry into the hash table */+ ZSTD_ldm_makeEntryAndInsertByTag(ldmState, rollingHash, hBits,+ (U32)(lastHashed - base),+ *params);++ assert(ip + backwardMatchLength == lastHashed);++ /* Fill the hash table from lastHashed+1 to ip+mLength*/+ /* Heuristic: don't need to fill the entire table at end of block */+ if (ip + mLength <= ilimit) {+ rollingHash = ZSTD_ldm_fillLdmHashTable(+ ldmState, rollingHash, lastHashed,+ ip + mLength, base, hBits, *params);+ lastHashed = ip + mLength - 1;+ }+ ip += mLength;+ anchor = ip;+ }+ return iend - anchor;+}++/*! ZSTD_ldm_reduceTable() :+ * reduce table indexes by `reducerValue` */+static void ZSTD_ldm_reduceTable(ldmEntry_t* const table, U32 const size,+ U32 const reducerValue)+{+ U32 u;+ for (u = 0; u < size; u++) {+ if (table[u].offset < reducerValue) table[u].offset = 0;+ else table[u].offset -= reducerValue;+ }+}++size_t ZSTD_ldm_generateSequences(+ ldmState_t* ldmState, rawSeqStore_t* sequences,+ ldmParams_t const* params, void const* src, size_t srcSize)+{+ U32 const maxDist = 1U << params->windowLog;+ BYTE const* const istart = (BYTE const*)src;+ BYTE const* const iend = istart + srcSize;+ size_t const kMaxChunkSize = 1 << 20;+ size_t const nbChunks = (srcSize / kMaxChunkSize) + ((srcSize % kMaxChunkSize) != 0);+ size_t chunk;+ size_t leftoverSize = 0;++ assert(ZSTD_CHUNKSIZE_MAX >= kMaxChunkSize);+ /* Check that ZSTD_window_update() has been called for this chunk prior+ * to passing it to this function.+ */+ assert(ldmState->window.nextSrc >= (BYTE const*)src + srcSize);+ /* The input could be very large (in zstdmt), so it must be broken up into+ * chunks to enforce the maximum distance and handle overflow correction.+ */+ assert(sequences->pos <= sequences->size);+ assert(sequences->size <= sequences->capacity);+ for (chunk = 0; chunk < nbChunks && sequences->size < sequences->capacity; ++chunk) {+ BYTE const* const chunkStart = istart + chunk * kMaxChunkSize;+ size_t const remaining = (size_t)(iend - chunkStart);+ BYTE const *const chunkEnd =+ (remaining < kMaxChunkSize) ? iend : chunkStart + kMaxChunkSize;+ size_t const chunkSize = chunkEnd - chunkStart;+ size_t newLeftoverSize;+ size_t const prevSize = sequences->size;++ assert(chunkStart < iend);+ /* 1. Perform overflow correction if necessary. */+ if (ZSTD_window_needOverflowCorrection(ldmState->window, chunkEnd)) {+ U32 const ldmHSize = 1U << params->hashLog;+ U32 const correction = ZSTD_window_correctOverflow(+ &ldmState->window, /* cycleLog */ 0, maxDist, src);+ ZSTD_ldm_reduceTable(ldmState->hashTable, ldmHSize, correction);+ }+ /* 2. We enforce the maximum offset allowed.+ *+ * kMaxChunkSize should be small enough that we don't lose too much of+ * the window through early invalidation.+ * TODO: * Test the chunk size.+ * * Try invalidation after the sequence generation and test the+ * the offset against maxDist directly.+ */+ ZSTD_window_enforceMaxDist(&ldmState->window, chunkEnd, maxDist, NULL, NULL);+ /* 3. Generate the sequences for the chunk, and get newLeftoverSize. */+ newLeftoverSize = ZSTD_ldm_generateSequences_internal(+ ldmState, sequences, params, chunkStart, chunkSize);+ if (ZSTD_isError(newLeftoverSize))+ return newLeftoverSize;+ /* 4. We add the leftover literals from previous iterations to the first+ * newly generated sequence, or add the `newLeftoverSize` if none are+ * generated.+ */+ /* Prepend the leftover literals from the last call */+ if (prevSize < sequences->size) {+ sequences->seq[prevSize].litLength += (U32)leftoverSize;+ leftoverSize = newLeftoverSize;+ } else {+ assert(newLeftoverSize == chunkSize);+ leftoverSize += chunkSize;+ }+ }+ return 0;+}++void ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize, U32 const minMatch) {+ while (srcSize > 0 && rawSeqStore->pos < rawSeqStore->size) {+ rawSeq* seq = rawSeqStore->seq + rawSeqStore->pos;+ if (srcSize <= seq->litLength) {+ /* Skip past srcSize literals */+ seq->litLength -= (U32)srcSize;+ return;+ }+ srcSize -= seq->litLength;+ seq->litLength = 0;+ if (srcSize < seq->matchLength) {+ /* Skip past the first srcSize of the match */+ seq->matchLength -= (U32)srcSize;+ if (seq->matchLength < minMatch) {+ /* The match is too short, omit it */+ if (rawSeqStore->pos + 1 < rawSeqStore->size) {+ seq[1].litLength += seq[0].matchLength;+ }+ rawSeqStore->pos++;+ }+ return;+ }+ srcSize -= seq->matchLength;+ seq->matchLength = 0;+ rawSeqStore->pos++;+ }+}++/**+ * If the sequence length is longer than remaining then the sequence is split+ * between this block and the next.+ *+ * Returns the current sequence to handle, or if the rest of the block should+ * be literals, it returns a sequence with offset == 0.+ */+static rawSeq maybeSplitSequence(rawSeqStore_t* rawSeqStore,+ U32 const remaining, U32 const minMatch)+{+ rawSeq sequence = rawSeqStore->seq[rawSeqStore->pos];+ assert(sequence.offset > 0);+ /* Likely: No partial sequence */+ if (remaining >= sequence.litLength + sequence.matchLength) {+ rawSeqStore->pos++;+ return sequence;+ }+ /* Cut the sequence short (offset == 0 ==> rest is literals). */+ if (remaining <= sequence.litLength) {+ sequence.offset = 0;+ } else if (remaining < sequence.litLength + sequence.matchLength) {+ sequence.matchLength = remaining - sequence.litLength;+ if (sequence.matchLength < minMatch) {+ sequence.offset = 0;+ }+ }+ /* Skip past `remaining` bytes for the future sequences. */+ ZSTD_ldm_skipSequences(rawSeqStore, remaining, minMatch);+ return sequence;+}++size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ unsigned const minMatch = cParams->minMatch;+ ZSTD_blockCompressor const blockCompressor =+ ZSTD_selectBlockCompressor(cParams->strategy, ZSTD_matchState_dictMode(ms));+ /* Input bounds */+ BYTE const* const istart = (BYTE const*)src;+ BYTE const* const iend = istart + srcSize;+ /* Input positions */+ BYTE const* ip = istart;++ DEBUGLOG(5, "ZSTD_ldm_blockCompress: srcSize=%zu", srcSize);+ assert(rawSeqStore->pos <= rawSeqStore->size);+ assert(rawSeqStore->size <= rawSeqStore->capacity);+ /* Loop through each sequence and apply the block compressor to the lits */+ while (rawSeqStore->pos < rawSeqStore->size && ip < iend) {+ /* maybeSplitSequence updates rawSeqStore->pos */+ rawSeq const sequence = maybeSplitSequence(rawSeqStore,+ (U32)(iend - ip), minMatch);+ int i;+ /* End signal */+ if (sequence.offset == 0)+ break;++ assert(sequence.offset <= (1U << cParams->windowLog));+ assert(ip + sequence.litLength + sequence.matchLength <= iend);++ /* Fill tables for block compressor */+ ZSTD_ldm_limitTableUpdate(ms, ip);+ ZSTD_ldm_fillFastTables(ms, ip);+ /* Run the block compressor */+ DEBUGLOG(5, "calling block compressor on segment of size %u", sequence.litLength);+ {+ size_t const newLitLength =+ blockCompressor(ms, seqStore, rep, ip, sequence.litLength);+ ip += sequence.litLength;+ /* Update the repcodes */+ for (i = ZSTD_REP_NUM - 1; i > 0; i--)+ rep[i] = rep[i-1];+ rep[0] = sequence.offset;+ /* Store the sequence */+ ZSTD_storeSeq(seqStore, newLitLength, ip - newLitLength,+ sequence.offset + ZSTD_REP_MOVE,+ sequence.matchLength - MINMATCH);+ ip += sequence.matchLength;+ }+ }+ /* Fill the tables for the block compressor */+ ZSTD_ldm_limitTableUpdate(ms, ip);+ ZSTD_ldm_fillFastTables(ms, ip);+ /* Compress the last literals */+ return blockCompressor(ms, seqStore, rep, ip, iend - ip);+}
+ zstd/lib/compress/zstd_ldm.h view
@@ -0,0 +1,105 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ */++#ifndef ZSTD_LDM_H+#define ZSTD_LDM_H++#if defined (__cplusplus)+extern "C" {+#endif++#include "zstd_compress_internal.h" /* ldmParams_t, U32 */+#include "zstd.h" /* ZSTD_CCtx, size_t */++/*-*************************************+* Long distance matching+***************************************/++#define ZSTD_LDM_DEFAULT_WINDOW_LOG ZSTD_WINDOWLOG_LIMIT_DEFAULT++/**+ * ZSTD_ldm_generateSequences():+ *+ * Generates the sequences using the long distance match finder.+ * Generates long range matching sequences in `sequences`, which parse a prefix+ * of the source. `sequences` must be large enough to store every sequence,+ * which can be checked with `ZSTD_ldm_getMaxNbSeq()`.+ * @returns 0 or an error code.+ *+ * NOTE: The user must have called ZSTD_window_update() for all of the input+ * they have, even if they pass it to ZSTD_ldm_generateSequences() in chunks.+ * NOTE: This function returns an error if it runs out of space to store+ * sequences.+ */+size_t ZSTD_ldm_generateSequences(+ ldmState_t* ldms, rawSeqStore_t* sequences,+ ldmParams_t const* params, void const* src, size_t srcSize);++/**+ * ZSTD_ldm_blockCompress():+ *+ * Compresses a block using the predefined sequences, along with a secondary+ * block compressor. The literals section of every sequence is passed to the+ * secondary block compressor, and those sequences are interspersed with the+ * predefined sequences. Returns the length of the last literals.+ * Updates `rawSeqStore.pos` to indicate how many sequences have been consumed.+ * `rawSeqStore.seq` may also be updated to split the last sequence between two+ * blocks.+ * @return The length of the last literals.+ *+ * NOTE: The source must be at most the maximum block size, but the predefined+ * sequences can be any size, and may be longer than the block. In the case that+ * they are longer than the block, the last sequences may need to be split into+ * two. We handle that case correctly, and update `rawSeqStore` appropriately.+ * NOTE: This function does not return any errors.+ */+size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);++/**+ * ZSTD_ldm_skipSequences():+ *+ * Skip past `srcSize` bytes worth of sequences in `rawSeqStore`.+ * Avoids emitting matches less than `minMatch` bytes.+ * Must be called for data with is not passed to ZSTD_ldm_blockCompress().+ */+void ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize,+ U32 const minMatch);+++/** ZSTD_ldm_getTableSize() :+ * Estimate the space needed for long distance matching tables or 0 if LDM is+ * disabled.+ */+size_t ZSTD_ldm_getTableSize(ldmParams_t params);++/** ZSTD_ldm_getSeqSpace() :+ * Return an upper bound on the number of sequences that can be produced by+ * the long distance matcher, or 0 if LDM is disabled.+ */+size_t ZSTD_ldm_getMaxNbSeq(ldmParams_t params, size_t maxChunkSize);++/** ZSTD_ldm_adjustParameters() :+ * If the params->hashRateLog is not set, set it to its default value based on+ * windowLog and params->hashLog.+ *+ * Ensures that params->bucketSizeLog is <= params->hashLog (setting it to+ * params->hashLog if it is not).+ *+ * Ensures that the minMatchLength >= targetLength during optimal parsing.+ */+void ZSTD_ldm_adjustParameters(ldmParams_t* params,+ ZSTD_compressionParameters const* cParams);++#if defined (__cplusplus)+}+#endif++#endif /* ZSTD_FAST_H */
+ zstd/lib/compress/zstd_opt.c view
@@ -0,0 +1,1233 @@+/*+ * Copyright (c) 2016-present, Przemyslaw Skibinski, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++#include "zstd_compress_internal.h"+#include "hist.h"+#include "zstd_opt.h"+++#define ZSTD_LITFREQ_ADD 2 /* scaling factor for litFreq, so that frequencies adapt faster to new stats */+#define ZSTD_FREQ_DIV 4 /* log factor when using previous stats to init next stats */+#define ZSTD_MAX_PRICE (1<<30)++#define ZSTD_PREDEF_THRESHOLD 1024 /* if srcSize < ZSTD_PREDEF_THRESHOLD, symbols' cost is assumed static, directly determined by pre-defined distributions */+++/*-*************************************+* Price functions for optimal parser+***************************************/++#if 0 /* approximation at bit level */+# define BITCOST_ACCURACY 0+# define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)+# define WEIGHT(stat) ((void)opt, ZSTD_bitWeight(stat))+#elif 0 /* fractional bit accuracy */+# define BITCOST_ACCURACY 8+# define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)+# define WEIGHT(stat,opt) ((void)opt, ZSTD_fracWeight(stat))+#else /* opt==approx, ultra==accurate */+# define BITCOST_ACCURACY 8+# define BITCOST_MULTIPLIER (1 << BITCOST_ACCURACY)+# define WEIGHT(stat,opt) (opt ? ZSTD_fracWeight(stat) : ZSTD_bitWeight(stat))+#endif++MEM_STATIC U32 ZSTD_bitWeight(U32 stat)+{+ return (ZSTD_highbit32(stat+1) * BITCOST_MULTIPLIER);+}++MEM_STATIC U32 ZSTD_fracWeight(U32 rawStat)+{+ U32 const stat = rawStat + 1;+ U32 const hb = ZSTD_highbit32(stat);+ U32 const BWeight = hb * BITCOST_MULTIPLIER;+ U32 const FWeight = (stat << BITCOST_ACCURACY) >> hb;+ U32 const weight = BWeight + FWeight;+ assert(hb + BITCOST_ACCURACY < 31);+ return weight;+}++#if (DEBUGLEVEL>=2)+/* debugging function,+ * @return price in bytes as fractional value+ * for debug messages only */+MEM_STATIC double ZSTD_fCost(U32 price)+{+ return (double)price / (BITCOST_MULTIPLIER*8);+}+#endif++static int ZSTD_compressedLiterals(optState_t const* const optPtr)+{+ return optPtr->literalCompressionMode != ZSTD_lcm_uncompressed;+}++static void ZSTD_setBasePrices(optState_t* optPtr, int optLevel)+{+ if (ZSTD_compressedLiterals(optPtr))+ optPtr->litSumBasePrice = WEIGHT(optPtr->litSum, optLevel);+ optPtr->litLengthSumBasePrice = WEIGHT(optPtr->litLengthSum, optLevel);+ optPtr->matchLengthSumBasePrice = WEIGHT(optPtr->matchLengthSum, optLevel);+ optPtr->offCodeSumBasePrice = WEIGHT(optPtr->offCodeSum, optLevel);+}+++/* ZSTD_downscaleStat() :+ * reduce all elements in table by a factor 2^(ZSTD_FREQ_DIV+malus)+ * return the resulting sum of elements */+static U32 ZSTD_downscaleStat(unsigned* table, U32 lastEltIndex, int malus)+{+ U32 s, sum=0;+ DEBUGLOG(5, "ZSTD_downscaleStat (nbElts=%u)", (unsigned)lastEltIndex+1);+ assert(ZSTD_FREQ_DIV+malus > 0 && ZSTD_FREQ_DIV+malus < 31);+ for (s=0; s<lastEltIndex+1; s++) {+ table[s] = 1 + (table[s] >> (ZSTD_FREQ_DIV+malus));+ sum += table[s];+ }+ return sum;+}++/* ZSTD_rescaleFreqs() :+ * if first block (detected by optPtr->litLengthSum == 0) : init statistics+ * take hints from dictionary if there is one+ * or init from zero, using src for literals stats, or flat 1 for match symbols+ * otherwise downscale existing stats, to be used as seed for next block.+ */+static void+ZSTD_rescaleFreqs(optState_t* const optPtr,+ const BYTE* const src, size_t const srcSize,+ int const optLevel)+{+ int const compressedLiterals = ZSTD_compressedLiterals(optPtr);+ DEBUGLOG(5, "ZSTD_rescaleFreqs (srcSize=%u)", (unsigned)srcSize);+ optPtr->priceType = zop_dynamic;++ if (optPtr->litLengthSum == 0) { /* first block : init */+ if (srcSize <= ZSTD_PREDEF_THRESHOLD) { /* heuristic */+ DEBUGLOG(5, "(srcSize <= ZSTD_PREDEF_THRESHOLD) => zop_predef");+ optPtr->priceType = zop_predef;+ }++ assert(optPtr->symbolCosts != NULL);+ if (optPtr->symbolCosts->huf.repeatMode == HUF_repeat_valid) {+ /* huffman table presumed generated by dictionary */+ optPtr->priceType = zop_dynamic;++ if (compressedLiterals) {+ unsigned lit;+ assert(optPtr->litFreq != NULL);+ optPtr->litSum = 0;+ for (lit=0; lit<=MaxLit; lit++) {+ U32 const scaleLog = 11; /* scale to 2K */+ U32 const bitCost = HUF_getNbBits(optPtr->symbolCosts->huf.CTable, lit);+ assert(bitCost <= scaleLog);+ optPtr->litFreq[lit] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;+ optPtr->litSum += optPtr->litFreq[lit];+ } }++ { unsigned ll;+ FSE_CState_t llstate;+ FSE_initCState(&llstate, optPtr->symbolCosts->fse.litlengthCTable);+ optPtr->litLengthSum = 0;+ for (ll=0; ll<=MaxLL; ll++) {+ U32 const scaleLog = 10; /* scale to 1K */+ U32 const bitCost = FSE_getMaxNbBits(llstate.symbolTT, ll);+ assert(bitCost < scaleLog);+ optPtr->litLengthFreq[ll] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;+ optPtr->litLengthSum += optPtr->litLengthFreq[ll];+ } }++ { unsigned ml;+ FSE_CState_t mlstate;+ FSE_initCState(&mlstate, optPtr->symbolCosts->fse.matchlengthCTable);+ optPtr->matchLengthSum = 0;+ for (ml=0; ml<=MaxML; ml++) {+ U32 const scaleLog = 10;+ U32 const bitCost = FSE_getMaxNbBits(mlstate.symbolTT, ml);+ assert(bitCost < scaleLog);+ optPtr->matchLengthFreq[ml] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;+ optPtr->matchLengthSum += optPtr->matchLengthFreq[ml];+ } }++ { unsigned of;+ FSE_CState_t ofstate;+ FSE_initCState(&ofstate, optPtr->symbolCosts->fse.offcodeCTable);+ optPtr->offCodeSum = 0;+ for (of=0; of<=MaxOff; of++) {+ U32 const scaleLog = 10;+ U32 const bitCost = FSE_getMaxNbBits(ofstate.symbolTT, of);+ assert(bitCost < scaleLog);+ optPtr->offCodeFreq[of] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;+ optPtr->offCodeSum += optPtr->offCodeFreq[of];+ } }++ } else { /* not a dictionary */++ assert(optPtr->litFreq != NULL);+ if (compressedLiterals) {+ unsigned lit = MaxLit;+ HIST_count_simple(optPtr->litFreq, &lit, src, srcSize); /* use raw first block to init statistics */+ optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);+ }++ { unsigned ll;+ for (ll=0; ll<=MaxLL; ll++)+ optPtr->litLengthFreq[ll] = 1;+ }+ optPtr->litLengthSum = MaxLL+1;++ { unsigned ml;+ for (ml=0; ml<=MaxML; ml++)+ optPtr->matchLengthFreq[ml] = 1;+ }+ optPtr->matchLengthSum = MaxML+1;++ { unsigned of;+ for (of=0; of<=MaxOff; of++)+ optPtr->offCodeFreq[of] = 1;+ }+ optPtr->offCodeSum = MaxOff+1;++ }++ } else { /* new block : re-use previous statistics, scaled down */++ if (compressedLiterals)+ optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);+ optPtr->litLengthSum = ZSTD_downscaleStat(optPtr->litLengthFreq, MaxLL, 0);+ optPtr->matchLengthSum = ZSTD_downscaleStat(optPtr->matchLengthFreq, MaxML, 0);+ optPtr->offCodeSum = ZSTD_downscaleStat(optPtr->offCodeFreq, MaxOff, 0);+ }++ ZSTD_setBasePrices(optPtr, optLevel);+}++/* ZSTD_rawLiteralsCost() :+ * price of literals (only) in specified segment (which length can be 0).+ * does not include price of literalLength symbol */+static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength,+ const optState_t* const optPtr,+ int optLevel)+{+ if (litLength == 0) return 0;++ if (!ZSTD_compressedLiterals(optPtr))+ return (litLength << 3) * BITCOST_MULTIPLIER; /* Uncompressed - 8 bytes per literal. */++ if (optPtr->priceType == zop_predef)+ return (litLength*6) * BITCOST_MULTIPLIER; /* 6 bit per literal - no statistic used */++ /* dynamic statistics */+ { U32 price = litLength * optPtr->litSumBasePrice;+ U32 u;+ for (u=0; u < litLength; u++) {+ assert(WEIGHT(optPtr->litFreq[literals[u]], optLevel) <= optPtr->litSumBasePrice); /* literal cost should never be negative */+ price -= WEIGHT(optPtr->litFreq[literals[u]], optLevel);+ }+ return price;+ }+}++/* ZSTD_litLengthPrice() :+ * cost of literalLength symbol */+static U32 ZSTD_litLengthPrice(U32 const litLength, const optState_t* const optPtr, int optLevel)+{+ if (optPtr->priceType == zop_predef) return WEIGHT(litLength, optLevel);++ /* dynamic statistics */+ { U32 const llCode = ZSTD_LLcode(litLength);+ return (LL_bits[llCode] * BITCOST_MULTIPLIER)+ + optPtr->litLengthSumBasePrice+ - WEIGHT(optPtr->litLengthFreq[llCode], optLevel);+ }+}++/* ZSTD_litLengthContribution() :+ * @return ( cost(litlength) - cost(0) )+ * this value can then be added to rawLiteralsCost()+ * to provide a cost which is directly comparable to a match ending at same position */+static int ZSTD_litLengthContribution(U32 const litLength, const optState_t* const optPtr, int optLevel)+{+ if (optPtr->priceType >= zop_predef) return WEIGHT(litLength, optLevel);++ /* dynamic statistics */+ { U32 const llCode = ZSTD_LLcode(litLength);+ int const contribution = (LL_bits[llCode] * BITCOST_MULTIPLIER)+ + WEIGHT(optPtr->litLengthFreq[0], optLevel) /* note: log2litLengthSum cancel out */+ - WEIGHT(optPtr->litLengthFreq[llCode], optLevel);+#if 1+ return contribution;+#else+ return MAX(0, contribution); /* sometimes better, sometimes not ... */+#endif+ }+}++/* ZSTD_literalsContribution() :+ * creates a fake cost for the literals part of a sequence+ * which can be compared to the ending cost of a match+ * should a new match start at this position */+static int ZSTD_literalsContribution(const BYTE* const literals, U32 const litLength,+ const optState_t* const optPtr,+ int optLevel)+{+ int const contribution = ZSTD_rawLiteralsCost(literals, litLength, optPtr, optLevel)+ + ZSTD_litLengthContribution(litLength, optPtr, optLevel);+ return contribution;+}++/* ZSTD_getMatchPrice() :+ * Provides the cost of the match part (offset + matchLength) of a sequence+ * Must be combined with ZSTD_fullLiteralsCost() to get the full cost of a sequence.+ * optLevel: when <2, favors small offset for decompression speed (improved cache efficiency) */+FORCE_INLINE_TEMPLATE U32+ZSTD_getMatchPrice(U32 const offset,+ U32 const matchLength,+ const optState_t* const optPtr,+ int const optLevel)+{+ U32 price;+ U32 const offCode = ZSTD_highbit32(offset+1);+ U32 const mlBase = matchLength - MINMATCH;+ assert(matchLength >= MINMATCH);++ if (optPtr->priceType == zop_predef) /* fixed scheme, do not use statistics */+ return WEIGHT(mlBase, optLevel) + ((16 + offCode) * BITCOST_MULTIPLIER);++ /* dynamic statistics */+ price = (offCode * BITCOST_MULTIPLIER) + (optPtr->offCodeSumBasePrice - WEIGHT(optPtr->offCodeFreq[offCode], optLevel));+ if ((optLevel<2) /*static*/ && offCode >= 20)+ price += (offCode-19)*2 * BITCOST_MULTIPLIER; /* handicap for long distance offsets, favor decompression speed */++ /* match Length */+ { U32 const mlCode = ZSTD_MLcode(mlBase);+ price += (ML_bits[mlCode] * BITCOST_MULTIPLIER) + (optPtr->matchLengthSumBasePrice - WEIGHT(optPtr->matchLengthFreq[mlCode], optLevel));+ }++ price += BITCOST_MULTIPLIER / 5; /* heuristic : make matches a bit more costly to favor less sequences -> faster decompression speed */++ DEBUGLOG(8, "ZSTD_getMatchPrice(ml:%u) = %u", matchLength, price);+ return price;+}++/* ZSTD_updateStats() :+ * assumption : literals + litLengtn <= iend */+static void ZSTD_updateStats(optState_t* const optPtr,+ U32 litLength, const BYTE* literals,+ U32 offsetCode, U32 matchLength)+{+ /* literals */+ if (ZSTD_compressedLiterals(optPtr)) {+ U32 u;+ for (u=0; u < litLength; u++)+ optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;+ optPtr->litSum += litLength*ZSTD_LITFREQ_ADD;+ }++ /* literal Length */+ { U32 const llCode = ZSTD_LLcode(litLength);+ optPtr->litLengthFreq[llCode]++;+ optPtr->litLengthSum++;+ }++ /* match offset code (0-2=>repCode; 3+=>offset+2) */+ { U32 const offCode = ZSTD_highbit32(offsetCode+1);+ assert(offCode <= MaxOff);+ optPtr->offCodeFreq[offCode]++;+ optPtr->offCodeSum++;+ }++ /* match Length */+ { U32 const mlBase = matchLength - MINMATCH;+ U32 const mlCode = ZSTD_MLcode(mlBase);+ optPtr->matchLengthFreq[mlCode]++;+ optPtr->matchLengthSum++;+ }+}+++/* ZSTD_readMINMATCH() :+ * function safe only for comparisons+ * assumption : memPtr must be at least 4 bytes before end of buffer */+MEM_STATIC U32 ZSTD_readMINMATCH(const void* memPtr, U32 length)+{+ switch (length)+ {+ default :+ case 4 : return MEM_read32(memPtr);+ case 3 : if (MEM_isLittleEndian())+ return MEM_read32(memPtr)<<8;+ else+ return MEM_read32(memPtr)>>8;+ }+}+++/* Update hashTable3 up to ip (excluded)+ Assumption : always within prefix (i.e. not within extDict) */+static U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_matchState_t* ms, const BYTE* const ip)+{+ U32* const hashTable3 = ms->hashTable3;+ U32 const hashLog3 = ms->hashLog3;+ const BYTE* const base = ms->window.base;+ U32 idx = ms->nextToUpdate3;+ U32 const target = ms->nextToUpdate3 = (U32)(ip - base);+ size_t const hash3 = ZSTD_hash3Ptr(ip, hashLog3);+ assert(hashLog3 > 0);++ while(idx < target) {+ hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx;+ idx++;+ }++ return hashTable3[hash3];+}+++/*-*************************************+* Binary Tree search+***************************************/+/** ZSTD_insertBt1() : add one or multiple positions to tree.+ * ip : assumed <= iend-8 .+ * @return : nb of positions added */+static U32 ZSTD_insertBt1(+ ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iend,+ U32 const mls, const int extDict)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32* const hashTable = ms->hashTable;+ U32 const hashLog = cParams->hashLog;+ size_t const h = ZSTD_hashPtr(ip, hashLog, mls);+ U32* const bt = ms->chainTable;+ U32 const btLog = cParams->chainLog - 1;+ U32 const btMask = (1 << btLog) - 1;+ U32 matchIndex = hashTable[h];+ size_t commonLengthSmaller=0, commonLengthLarger=0;+ const BYTE* const base = ms->window.base;+ const BYTE* const dictBase = ms->window.dictBase;+ const U32 dictLimit = ms->window.dictLimit;+ const BYTE* const dictEnd = dictBase + dictLimit;+ const BYTE* const prefixStart = base + dictLimit;+ const BYTE* match;+ const U32 current = (U32)(ip-base);+ const U32 btLow = btMask >= current ? 0 : current - btMask;+ U32* smallerPtr = bt + 2*(current&btMask);+ U32* largerPtr = smallerPtr + 1;+ U32 dummy32; /* to be nullified at the end */+ U32 const windowLow = ms->window.lowLimit;+ U32 matchEndIdx = current+8+1;+ size_t bestLength = 8;+ U32 nbCompares = 1U << cParams->searchLog;+#ifdef ZSTD_C_PREDICT+ U32 predictedSmall = *(bt + 2*((current-1)&btMask) + 0);+ U32 predictedLarge = *(bt + 2*((current-1)&btMask) + 1);+ predictedSmall += (predictedSmall>0);+ predictedLarge += (predictedLarge>0);+#endif /* ZSTD_C_PREDICT */++ DEBUGLOG(8, "ZSTD_insertBt1 (%u)", current);++ assert(ip <= iend-8); /* required for h calculation */+ hashTable[h] = current; /* Update Hash Table */++ assert(windowLow > 0);+ while (nbCompares-- && (matchIndex >= windowLow)) {+ U32* const nextPtr = bt + 2*(matchIndex & btMask);+ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */+ assert(matchIndex < current);++#ifdef ZSTD_C_PREDICT /* note : can create issues when hlog small <= 11 */+ const U32* predictPtr = bt + 2*((matchIndex-1) & btMask); /* written this way, as bt is a roll buffer */+ if (matchIndex == predictedSmall) {+ /* no need to check length, result known */+ *smallerPtr = matchIndex;+ if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */+ smallerPtr = nextPtr+1; /* new "smaller" => larger of match */+ matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */+ predictedSmall = predictPtr[1] + (predictPtr[1]>0);+ continue;+ }+ if (matchIndex == predictedLarge) {+ *largerPtr = matchIndex;+ if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */+ largerPtr = nextPtr;+ matchIndex = nextPtr[0];+ predictedLarge = predictPtr[0] + (predictPtr[0]>0);+ continue;+ }+#endif++ if (!extDict || (matchIndex+matchLength >= dictLimit)) {+ assert(matchIndex+matchLength >= dictLimit); /* might be wrong if actually extDict */+ match = base + matchIndex;+ matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);+ } else {+ match = dictBase + matchIndex;+ matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);+ if (matchIndex+matchLength >= dictLimit)+ match = base + matchIndex; /* to prepare for next usage of match[matchLength] */+ }++ if (matchLength > bestLength) {+ bestLength = matchLength;+ if (matchLength > matchEndIdx - matchIndex)+ matchEndIdx = matchIndex + (U32)matchLength;+ }++ if (ip+matchLength == iend) { /* equal : no way to know if inf or sup */+ break; /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */+ }++ if (match[matchLength] < ip[matchLength]) { /* necessarily within buffer */+ /* match is smaller than current */+ *smallerPtr = matchIndex; /* update smaller idx */+ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */+ if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop searching */+ smallerPtr = nextPtr+1; /* new "candidate" => larger than match, which was smaller than target */+ matchIndex = nextPtr[1]; /* new matchIndex, larger than previous and closer to current */+ } else {+ /* match is larger than current */+ *largerPtr = matchIndex;+ commonLengthLarger = matchLength;+ if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop searching */+ largerPtr = nextPtr;+ matchIndex = nextPtr[0];+ } }++ *smallerPtr = *largerPtr = 0;+ if (bestLength > 384) return MIN(192, (U32)(bestLength - 384)); /* speed optimization */+ assert(matchEndIdx > current + 8);+ return matchEndIdx - (current + 8);+}++FORCE_INLINE_TEMPLATE+void ZSTD_updateTree_internal(+ ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iend,+ const U32 mls, const ZSTD_dictMode_e dictMode)+{+ const BYTE* const base = ms->window.base;+ U32 const target = (U32)(ip - base);+ U32 idx = ms->nextToUpdate;+ DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)",+ idx, target, dictMode);++ while(idx < target)+ idx += ZSTD_insertBt1(ms, base+idx, iend, mls, dictMode == ZSTD_extDict);+ ms->nextToUpdate = target;+}++void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) {+ ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict);+}++FORCE_INLINE_TEMPLATE+U32 ZSTD_insertBtAndGetAllMatches (+ ZSTD_matchState_t* ms,+ const BYTE* const ip, const BYTE* const iLimit, const ZSTD_dictMode_e dictMode,+ U32 rep[ZSTD_REP_NUM],+ U32 const ll0, /* tells if associated literal length is 0 or not. This value must be 0 or 1 */+ ZSTD_match_t* matches,+ const U32 lengthToBeat,+ U32 const mls /* template */)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);+ const BYTE* const base = ms->window.base;+ U32 const current = (U32)(ip-base);+ U32 const hashLog = cParams->hashLog;+ U32 const minMatch = (mls==3) ? 3 : 4;+ U32* const hashTable = ms->hashTable;+ size_t const h = ZSTD_hashPtr(ip, hashLog, mls);+ U32 matchIndex = hashTable[h];+ U32* const bt = ms->chainTable;+ U32 const btLog = cParams->chainLog - 1;+ U32 const btMask= (1U << btLog) - 1;+ size_t commonLengthSmaller=0, commonLengthLarger=0;+ const BYTE* const dictBase = ms->window.dictBase;+ U32 const dictLimit = ms->window.dictLimit;+ const BYTE* const dictEnd = dictBase + dictLimit;+ const BYTE* const prefixStart = base + dictLimit;+ U32 const btLow = btMask >= current ? 0 : current - btMask;+ U32 const windowLow = ms->window.lowLimit;+ U32 const matchLow = windowLow ? windowLow : 1;+ U32* smallerPtr = bt + 2*(current&btMask);+ U32* largerPtr = bt + 2*(current&btMask) + 1;+ U32 matchEndIdx = current+8+1; /* farthest referenced position of any match => detects repetitive patterns */+ U32 dummy32; /* to be nullified at the end */+ U32 mnum = 0;+ U32 nbCompares = 1U << cParams->searchLog;++ const ZSTD_matchState_t* dms = dictMode == ZSTD_dictMatchState ? ms->dictMatchState : NULL;+ const ZSTD_compressionParameters* const dmsCParams =+ dictMode == ZSTD_dictMatchState ? &dms->cParams : NULL;+ const BYTE* const dmsBase = dictMode == ZSTD_dictMatchState ? dms->window.base : NULL;+ const BYTE* const dmsEnd = dictMode == ZSTD_dictMatchState ? dms->window.nextSrc : NULL;+ U32 const dmsHighLimit = dictMode == ZSTD_dictMatchState ? (U32)(dmsEnd - dmsBase) : 0;+ U32 const dmsLowLimit = dictMode == ZSTD_dictMatchState ? dms->window.lowLimit : 0;+ U32 const dmsIndexDelta = dictMode == ZSTD_dictMatchState ? windowLow - dmsHighLimit : 0;+ U32 const dmsHashLog = dictMode == ZSTD_dictMatchState ? dmsCParams->hashLog : hashLog;+ U32 const dmsBtLog = dictMode == ZSTD_dictMatchState ? dmsCParams->chainLog - 1 : btLog;+ U32 const dmsBtMask = dictMode == ZSTD_dictMatchState ? (1U << dmsBtLog) - 1 : 0;+ U32 const dmsBtLow = dictMode == ZSTD_dictMatchState && dmsBtMask < dmsHighLimit - dmsLowLimit ? dmsHighLimit - dmsBtMask : dmsLowLimit;++ size_t bestLength = lengthToBeat-1;+ DEBUGLOG(8, "ZSTD_insertBtAndGetAllMatches: current=%u", current);++ /* check repCode */+ assert(ll0 <= 1); /* necessarily 1 or 0 */+ { U32 const lastR = ZSTD_REP_NUM + ll0;+ U32 repCode;+ for (repCode = ll0; repCode < lastR; repCode++) {+ U32 const repOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode];+ U32 const repIndex = current - repOffset;+ U32 repLen = 0;+ assert(current >= dictLimit);+ if (repOffset-1 /* intentional overflow, discards 0 and -1 */ < current-dictLimit) { /* equivalent to `current > repIndex >= dictLimit` */+ if (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(ip - repOffset, minMatch)) {+ repLen = (U32)ZSTD_count(ip+minMatch, ip+minMatch-repOffset, iLimit) + minMatch;+ }+ } else { /* repIndex < dictLimit || repIndex >= current */+ const BYTE* const repMatch = dictMode == ZSTD_dictMatchState ?+ dmsBase + repIndex - dmsIndexDelta :+ dictBase + repIndex;+ assert(current >= windowLow);+ if ( dictMode == ZSTD_extDict+ && ( ((repOffset-1) /*intentional overflow*/ < current - windowLow) /* equivalent to `current > repIndex >= windowLow` */+ & (((U32)((dictLimit-1) - repIndex) >= 3) ) /* intentional overflow : do not test positions overlapping 2 memory segments */)+ && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {+ repLen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iLimit, dictEnd, prefixStart) + minMatch;+ }+ if (dictMode == ZSTD_dictMatchState+ && ( ((repOffset-1) /*intentional overflow*/ < current - (dmsLowLimit + dmsIndexDelta)) /* equivalent to `current > repIndex >= dmsLowLimit` */+ & ((U32)((dictLimit-1) - repIndex) >= 3) ) /* intentional overflow : do not test positions overlapping 2 memory segments */+ && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {+ repLen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iLimit, dmsEnd, prefixStart) + minMatch;+ } }+ /* save longer solution */+ if (repLen > bestLength) {+ DEBUGLOG(8, "found repCode %u (ll0:%u, offset:%u) of length %u",+ repCode, ll0, repOffset, repLen);+ bestLength = repLen;+ matches[mnum].off = repCode - ll0;+ matches[mnum].len = (U32)repLen;+ mnum++;+ if ( (repLen > sufficient_len)+ | (ip+repLen == iLimit) ) { /* best possible */+ return mnum;+ } } } }++ /* HC3 match finder */+ if ((mls == 3) /*static*/ && (bestLength < mls)) {+ U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(ms, ip);+ if ((matchIndex3 >= matchLow)+ & (current - matchIndex3 < (1<<18)) /*heuristic : longer distance likely too expensive*/ ) {+ size_t mlen;+ if ((dictMode == ZSTD_noDict) /*static*/ || (dictMode == ZSTD_dictMatchState) /*static*/ || (matchIndex3 >= dictLimit)) {+ const BYTE* const match = base + matchIndex3;+ mlen = ZSTD_count(ip, match, iLimit);+ } else {+ const BYTE* const match = dictBase + matchIndex3;+ mlen = ZSTD_count_2segments(ip, match, iLimit, dictEnd, prefixStart);+ }++ /* save best solution */+ if (mlen >= mls /* == 3 > bestLength */) {+ DEBUGLOG(8, "found small match with hlog3, of length %u",+ (U32)mlen);+ bestLength = mlen;+ assert(current > matchIndex3);+ assert(mnum==0); /* no prior solution */+ matches[0].off = (current - matchIndex3) + ZSTD_REP_MOVE;+ matches[0].len = (U32)mlen;+ mnum = 1;+ if ( (mlen > sufficient_len) |+ (ip+mlen == iLimit) ) { /* best possible length */+ ms->nextToUpdate = current+1; /* skip insertion */+ return 1;+ }+ }+ }+ /* no dictMatchState lookup: dicts don't have a populated HC3 table */+ }++ hashTable[h] = current; /* Update Hash Table */++ while (nbCompares-- && (matchIndex >= matchLow)) {+ U32* const nextPtr = bt + 2*(matchIndex & btMask);+ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */+ const BYTE* match;+ assert(current > matchIndex);++ if ((dictMode == ZSTD_noDict) || (dictMode == ZSTD_dictMatchState) || (matchIndex+matchLength >= dictLimit)) {+ assert(matchIndex+matchLength >= dictLimit); /* ensure the condition is correct when !extDict */+ match = base + matchIndex;+ matchLength += ZSTD_count(ip+matchLength, match+matchLength, iLimit);+ } else {+ match = dictBase + matchIndex;+ matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);+ if (matchIndex+matchLength >= dictLimit)+ match = base + matchIndex; /* prepare for match[matchLength] */+ }++ if (matchLength > bestLength) {+ DEBUGLOG(8, "found match of length %u at distance %u (offCode=%u)",+ (U32)matchLength, current - matchIndex, current - matchIndex + ZSTD_REP_MOVE);+ assert(matchEndIdx > matchIndex);+ if (matchLength > matchEndIdx - matchIndex)+ matchEndIdx = matchIndex + (U32)matchLength;+ bestLength = matchLength;+ matches[mnum].off = (current - matchIndex) + ZSTD_REP_MOVE;+ matches[mnum].len = (U32)matchLength;+ mnum++;+ if ( (matchLength > ZSTD_OPT_NUM)+ | (ip+matchLength == iLimit) /* equal : no way to know if inf or sup */) {+ if (dictMode == ZSTD_dictMatchState) nbCompares = 0; /* break should also skip searching dms */+ break; /* drop, to preserve bt consistency (miss a little bit of compression) */+ }+ }++ if (match[matchLength] < ip[matchLength]) {+ /* match smaller than current */+ *smallerPtr = matchIndex; /* update smaller idx */+ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */+ if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */+ smallerPtr = nextPtr+1; /* new candidate => larger than match, which was smaller than current */+ matchIndex = nextPtr[1]; /* new matchIndex, larger than previous, closer to current */+ } else {+ *largerPtr = matchIndex;+ commonLengthLarger = matchLength;+ if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */+ largerPtr = nextPtr;+ matchIndex = nextPtr[0];+ } }++ *smallerPtr = *largerPtr = 0;++ if (dictMode == ZSTD_dictMatchState && nbCompares) {+ size_t const dmsH = ZSTD_hashPtr(ip, dmsHashLog, mls);+ U32 dictMatchIndex = dms->hashTable[dmsH];+ const U32* const dmsBt = dms->chainTable;+ commonLengthSmaller = commonLengthLarger = 0;+ while (nbCompares-- && (dictMatchIndex > dmsLowLimit)) {+ const U32* const nextPtr = dmsBt + 2*(dictMatchIndex & dmsBtMask);+ size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */+ const BYTE* match = dmsBase + dictMatchIndex;+ matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dmsEnd, prefixStart);+ if (dictMatchIndex+matchLength >= dmsHighLimit)+ match = base + dictMatchIndex + dmsIndexDelta; /* to prepare for next usage of match[matchLength] */++ if (matchLength > bestLength) {+ matchIndex = dictMatchIndex + dmsIndexDelta;+ DEBUGLOG(8, "found dms match of length %u at distance %u (offCode=%u)",+ (U32)matchLength, current - matchIndex, current - matchIndex + ZSTD_REP_MOVE);+ if (matchLength > matchEndIdx - matchIndex)+ matchEndIdx = matchIndex + (U32)matchLength;+ bestLength = matchLength;+ matches[mnum].off = (current - matchIndex) + ZSTD_REP_MOVE;+ matches[mnum].len = (U32)matchLength;+ mnum++;+ if ( (matchLength > ZSTD_OPT_NUM)+ | (ip+matchLength == iLimit) /* equal : no way to know if inf or sup */) {+ break; /* drop, to guarantee consistency (miss a little bit of compression) */+ }+ }++ if (dictMatchIndex <= dmsBtLow) { break; } /* beyond tree size, stop the search */+ if (match[matchLength] < ip[matchLength]) {+ commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */+ dictMatchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */+ } else {+ /* match is larger than current */+ commonLengthLarger = matchLength;+ dictMatchIndex = nextPtr[0];+ }+ }+ }++ assert(matchEndIdx > current+8);+ ms->nextToUpdate = matchEndIdx - 8; /* skip repetitive patterns */+ return mnum;+}+++FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllMatches (+ ZSTD_matchState_t* ms,+ const BYTE* ip, const BYTE* const iHighLimit, const ZSTD_dictMode_e dictMode,+ U32 rep[ZSTD_REP_NUM], U32 const ll0,+ ZSTD_match_t* matches, U32 const lengthToBeat)+{+ const ZSTD_compressionParameters* const cParams = &ms->cParams;+ U32 const matchLengthSearch = cParams->minMatch;+ DEBUGLOG(8, "ZSTD_BtGetAllMatches");+ if (ip < ms->window.base + ms->nextToUpdate) return 0; /* skipped area */+ ZSTD_updateTree_internal(ms, ip, iHighLimit, matchLengthSearch, dictMode);+ switch(matchLengthSearch)+ {+ case 3 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 3);+ default :+ case 4 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 4);+ case 5 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 5);+ case 7 :+ case 6 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 6);+ }+}+++/*-*******************************+* Optimal parser+*********************************/+typedef struct repcodes_s {+ U32 rep[3];+} repcodes_t;++static repcodes_t ZSTD_updateRep(U32 const rep[3], U32 const offset, U32 const ll0)+{+ repcodes_t newReps;+ if (offset >= ZSTD_REP_NUM) { /* full offset */+ newReps.rep[2] = rep[1];+ newReps.rep[1] = rep[0];+ newReps.rep[0] = offset - ZSTD_REP_MOVE;+ } else { /* repcode */+ U32 const repCode = offset + ll0;+ if (repCode > 0) { /* note : if repCode==0, no change */+ U32 const currentOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode];+ newReps.rep[2] = (repCode >= 2) ? rep[1] : rep[2];+ newReps.rep[1] = rep[0];+ newReps.rep[0] = currentOffset;+ } else { /* repCode == 0 */+ memcpy(&newReps, rep, sizeof(newReps));+ }+ }+ return newReps;+}+++static U32 ZSTD_totalLen(ZSTD_optimal_t sol)+{+ return sol.litlen + sol.mlen;+}++#if 0 /* debug */++static void+listStats(const U32* table, int lastEltID)+{+ int const nbElts = lastEltID + 1;+ int enb;+ for (enb=0; enb < nbElts; enb++) {+ (void)table;+ //RAWLOG(2, "%3i:%3i, ", enb, table[enb]);+ RAWLOG(2, "%4i,", table[enb]);+ }+ RAWLOG(2, " \n");+}++#endif++FORCE_INLINE_TEMPLATE size_t+ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,+ seqStore_t* seqStore,+ U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize,+ const int optLevel,+ const ZSTD_dictMode_e dictMode)+{+ optState_t* const optStatePtr = &ms->opt;+ const BYTE* const istart = (const BYTE*)src;+ const BYTE* ip = istart;+ const BYTE* anchor = istart;+ const BYTE* const iend = istart + srcSize;+ const BYTE* const ilimit = iend - 8;+ const BYTE* const base = ms->window.base;+ const BYTE* const prefixStart = base + ms->window.dictLimit;+ const ZSTD_compressionParameters* const cParams = &ms->cParams;++ U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);+ U32 const minMatch = (cParams->minMatch == 3) ? 3 : 4;++ ZSTD_optimal_t* const opt = optStatePtr->priceTable;+ ZSTD_match_t* const matches = optStatePtr->matchTable;+ ZSTD_optimal_t lastSequence;++ /* init */+ DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u",+ (U32)(ip - base), ms->window.dictLimit, ms->nextToUpdate);+ assert(optLevel <= 2);+ ms->nextToUpdate3 = ms->nextToUpdate;+ ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize, optLevel);+ ip += (ip==prefixStart);++ /* Match Loop */+ while (ip < ilimit) {+ U32 cur, last_pos = 0;++ /* find first match */+ { U32 const litlen = (U32)(ip - anchor);+ U32 const ll0 = !litlen;+ U32 const nbMatches = ZSTD_BtGetAllMatches(ms, ip, iend, dictMode, rep, ll0, matches, minMatch);+ if (!nbMatches) { ip++; continue; }++ /* initialize opt[0] */+ { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }+ opt[0].mlen = 0; /* means is_a_literal */+ opt[0].litlen = litlen;+ opt[0].price = ZSTD_literalsContribution(anchor, litlen, optStatePtr, optLevel);++ /* large match -> immediate encoding */+ { U32 const maxML = matches[nbMatches-1].len;+ U32 const maxOffset = matches[nbMatches-1].off;+ DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new series",+ nbMatches, maxML, maxOffset, (U32)(ip-prefixStart));++ if (maxML > sufficient_len) {+ lastSequence.litlen = litlen;+ lastSequence.mlen = maxML;+ lastSequence.off = maxOffset;+ DEBUGLOG(6, "large match (%u>%u), immediate encoding",+ maxML, sufficient_len);+ cur = 0;+ last_pos = ZSTD_totalLen(lastSequence);+ goto _shortestPath;+ } }++ /* set prices for first matches starting position == 0 */+ { U32 const literalsPrice = opt[0].price + ZSTD_litLengthPrice(0, optStatePtr, optLevel);+ U32 pos;+ U32 matchNb;+ for (pos = 1; pos < minMatch; pos++) {+ opt[pos].price = ZSTD_MAX_PRICE; /* mlen, litlen and price will be fixed during forward scanning */+ }+ for (matchNb = 0; matchNb < nbMatches; matchNb++) {+ U32 const offset = matches[matchNb].off;+ U32 const end = matches[matchNb].len;+ repcodes_t const repHistory = ZSTD_updateRep(rep, offset, ll0);+ for ( ; pos <= end ; pos++ ) {+ U32 const matchPrice = ZSTD_getMatchPrice(offset, pos, optStatePtr, optLevel);+ U32 const sequencePrice = literalsPrice + matchPrice;+ DEBUGLOG(7, "rPos:%u => set initial price : %.2f",+ pos, ZSTD_fCost(sequencePrice));+ opt[pos].mlen = pos;+ opt[pos].off = offset;+ opt[pos].litlen = litlen;+ opt[pos].price = sequencePrice;+ ZSTD_STATIC_ASSERT(sizeof(opt[pos].rep) == sizeof(repHistory));+ memcpy(opt[pos].rep, &repHistory, sizeof(repHistory));+ } }+ last_pos = pos-1;+ }+ }++ /* check further positions */+ for (cur = 1; cur <= last_pos; cur++) {+ const BYTE* const inr = ip + cur;+ assert(cur < ZSTD_OPT_NUM);+ DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur)++ /* Fix current position with one literal if cheaper */+ { U32 const litlen = (opt[cur-1].mlen == 0) ? opt[cur-1].litlen + 1 : 1;+ int const price = opt[cur-1].price+ + ZSTD_rawLiteralsCost(ip+cur-1, 1, optStatePtr, optLevel)+ + ZSTD_litLengthPrice(litlen, optStatePtr, optLevel)+ - ZSTD_litLengthPrice(litlen-1, optStatePtr, optLevel);+ assert(price < 1000000000); /* overflow check */+ if (price <= opt[cur].price) {+ DEBUGLOG(7, "cPos:%zi==rPos:%u : better price (%.2f<=%.2f) using literal (ll==%u) (hist:%u,%u,%u)",+ inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), litlen,+ opt[cur-1].rep[0], opt[cur-1].rep[1], opt[cur-1].rep[2]);+ opt[cur].mlen = 0;+ opt[cur].off = 0;+ opt[cur].litlen = litlen;+ opt[cur].price = price;+ memcpy(opt[cur].rep, opt[cur-1].rep, sizeof(opt[cur].rep));+ } else {+ DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f) (hist:%u,%u,%u)",+ inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price),+ opt[cur].rep[0], opt[cur].rep[1], opt[cur].rep[2]);+ }+ }++ /* last match must start at a minimum distance of 8 from oend */+ if (inr > ilimit) continue;++ if (cur == last_pos) break;++ if ( (optLevel==0) /*static_test*/+ && (opt[cur+1].price <= opt[cur].price + (BITCOST_MULTIPLIER/2)) ) {+ DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1);+ continue; /* skip unpromising positions; about ~+6% speed, -0.01 ratio */+ }++ { U32 const ll0 = (opt[cur].mlen != 0);+ U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0;+ U32 const previousPrice = opt[cur].price;+ U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel);+ U32 const nbMatches = ZSTD_BtGetAllMatches(ms, inr, iend, dictMode, opt[cur].rep, ll0, matches, minMatch);+ U32 matchNb;+ if (!nbMatches) {+ DEBUGLOG(7, "rPos:%u : no match found", cur);+ continue;+ }++ { U32 const maxML = matches[nbMatches-1].len;+ DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u",+ inr-istart, cur, nbMatches, maxML);++ if ( (maxML > sufficient_len)+ || (cur + maxML >= ZSTD_OPT_NUM) ) {+ lastSequence.mlen = maxML;+ lastSequence.off = matches[nbMatches-1].off;+ lastSequence.litlen = litlen;+ cur -= (opt[cur].mlen==0) ? opt[cur].litlen : 0; /* last sequence is actually only literals, fix cur to last match - note : may underflow, in which case, it's first sequence, and it's okay */+ last_pos = cur + ZSTD_totalLen(lastSequence);+ if (cur > ZSTD_OPT_NUM) cur = 0; /* underflow => first match */+ goto _shortestPath;+ } }++ /* set prices using matches found at position == cur */+ for (matchNb = 0; matchNb < nbMatches; matchNb++) {+ U32 const offset = matches[matchNb].off;+ repcodes_t const repHistory = ZSTD_updateRep(opt[cur].rep, offset, ll0);+ U32 const lastML = matches[matchNb].len;+ U32 const startML = (matchNb>0) ? matches[matchNb-1].len+1 : minMatch;+ U32 mlen;++ DEBUGLOG(7, "testing match %u => offCode=%4u, mlen=%2u, llen=%2u",+ matchNb, matches[matchNb].off, lastML, litlen);++ for (mlen = lastML; mlen >= startML; mlen--) { /* scan downward */+ U32 const pos = cur + mlen;+ int const price = basePrice + ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel);++ if ((pos > last_pos) || (price < opt[pos].price)) {+ DEBUGLOG(7, "rPos:%u (ml=%2u) => new better price (%.2f<%.2f)",+ pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price));+ while (last_pos < pos) { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; } /* fill empty positions */+ opt[pos].mlen = mlen;+ opt[pos].off = offset;+ opt[pos].litlen = litlen;+ opt[pos].price = price;+ ZSTD_STATIC_ASSERT(sizeof(opt[pos].rep) == sizeof(repHistory));+ memcpy(opt[pos].rep, &repHistory, sizeof(repHistory));+ } else {+ DEBUGLOG(7, "rPos:%u (ml=%2u) => new price is worse (%.2f>=%.2f)",+ pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price));+ if (optLevel==0) break; /* early update abort; gets ~+10% speed for about -0.01 ratio loss */+ }+ } } }+ } /* for (cur = 1; cur <= last_pos; cur++) */++ lastSequence = opt[last_pos];+ cur = last_pos > ZSTD_totalLen(lastSequence) ? last_pos - ZSTD_totalLen(lastSequence) : 0; /* single sequence, and it starts before `ip` */+ assert(cur < ZSTD_OPT_NUM); /* control overflow*/++_shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */+ assert(opt[0].mlen == 0);++ { U32 const storeEnd = cur + 1;+ U32 storeStart = storeEnd;+ U32 seqPos = cur;++ DEBUGLOG(6, "start reverse traversal (last_pos:%u, cur:%u)",+ last_pos, cur); (void)last_pos;+ assert(storeEnd < ZSTD_OPT_NUM);+ DEBUGLOG(6, "last sequence copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",+ storeEnd, lastSequence.litlen, lastSequence.mlen, lastSequence.off);+ opt[storeEnd] = lastSequence;+ while (seqPos > 0) {+ U32 const backDist = ZSTD_totalLen(opt[seqPos]);+ storeStart--;+ DEBUGLOG(6, "sequence from rPos=%u copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",+ seqPos, storeStart, opt[seqPos].litlen, opt[seqPos].mlen, opt[seqPos].off);+ opt[storeStart] = opt[seqPos];+ seqPos = (seqPos > backDist) ? seqPos - backDist : 0;+ }++ /* save sequences */+ DEBUGLOG(6, "sending selected sequences into seqStore")+ { U32 storePos;+ for (storePos=storeStart; storePos <= storeEnd; storePos++) {+ U32 const llen = opt[storePos].litlen;+ U32 const mlen = opt[storePos].mlen;+ U32 const offCode = opt[storePos].off;+ U32 const advance = llen + mlen;+ DEBUGLOG(6, "considering seq starting at %zi, llen=%u, mlen=%u",+ anchor - istart, (unsigned)llen, (unsigned)mlen);++ if (mlen==0) { /* only literals => must be last "sequence", actually starting a new stream of sequences */+ assert(storePos == storeEnd); /* must be last sequence */+ ip = anchor + llen; /* last "sequence" is a bunch of literals => don't progress anchor */+ continue; /* will finish */+ }++ /* repcodes update : like ZSTD_updateRep(), but update in place */+ if (offCode >= ZSTD_REP_NUM) { /* full offset */+ rep[2] = rep[1];+ rep[1] = rep[0];+ rep[0] = offCode - ZSTD_REP_MOVE;+ } else { /* repcode */+ U32 const repCode = offCode + (llen==0);+ if (repCode) { /* note : if repCode==0, no change */+ U32 const currentOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode];+ if (repCode >= 2) rep[2] = rep[1];+ rep[1] = rep[0];+ rep[0] = currentOffset;+ } }++ assert(anchor + llen <= iend);+ ZSTD_updateStats(optStatePtr, llen, anchor, offCode, mlen);+ ZSTD_storeSeq(seqStore, llen, anchor, offCode, mlen-MINMATCH);+ anchor += advance;+ ip = anchor;+ } }+ ZSTD_setBasePrices(optStatePtr, optLevel);+ }++ } /* while (ip < ilimit) */++ /* Return the last literals size */+ return iend - anchor;+}+++size_t ZSTD_compressBlock_btopt(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ DEBUGLOG(5, "ZSTD_compressBlock_btopt");+ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /*optLevel*/, ZSTD_noDict);+}+++/* used in 2-pass strategy */+static U32 ZSTD_upscaleStat(unsigned* table, U32 lastEltIndex, int bonus)+{+ U32 s, sum=0;+ assert(ZSTD_FREQ_DIV+bonus >= 0);+ for (s=0; s<lastEltIndex+1; s++) {+ table[s] <<= ZSTD_FREQ_DIV+bonus;+ table[s]--;+ sum += table[s];+ }+ return sum;+}++/* used in 2-pass strategy */+MEM_STATIC void ZSTD_upscaleStats(optState_t* optPtr)+{+ if (ZSTD_compressedLiterals(optPtr))+ optPtr->litSum = ZSTD_upscaleStat(optPtr->litFreq, MaxLit, 0);+ optPtr->litLengthSum = ZSTD_upscaleStat(optPtr->litLengthFreq, MaxLL, 0);+ optPtr->matchLengthSum = ZSTD_upscaleStat(optPtr->matchLengthFreq, MaxML, 0);+ optPtr->offCodeSum = ZSTD_upscaleStat(optPtr->offCodeFreq, MaxOff, 0);+}++/* ZSTD_initStats_ultra():+ * make a first compression pass, just to seed stats with more accurate starting values.+ * only works on first block, with no dictionary and no ldm.+ * this function cannot error, hence its contract must be respected.+ */+static void+ZSTD_initStats_ultra(ZSTD_matchState_t* ms,+ seqStore_t* seqStore,+ U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ U32 tmpRep[ZSTD_REP_NUM]; /* updated rep codes will sink here */+ memcpy(tmpRep, rep, sizeof(tmpRep));++ DEBUGLOG(4, "ZSTD_initStats_ultra (srcSize=%zu)", srcSize);+ assert(ms->opt.litLengthSum == 0); /* first block */+ assert(seqStore->sequences == seqStore->sequencesStart); /* no ldm */+ assert(ms->window.dictLimit == ms->window.lowLimit); /* no dictionary */+ assert(ms->window.dictLimit - ms->nextToUpdate <= 1); /* no prefix (note: intentional overflow, defined as 2-complement) */++ ZSTD_compressBlock_opt_generic(ms, seqStore, tmpRep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict); /* generate stats into ms->opt*/++ /* invalidate first scan from history */+ ZSTD_resetSeqStore(seqStore);+ ms->window.base -= srcSize;+ ms->window.dictLimit += (U32)srcSize;+ ms->window.lowLimit = ms->window.dictLimit;+ ms->nextToUpdate = ms->window.dictLimit;+ ms->nextToUpdate3 = ms->window.dictLimit;++ /* re-inforce weight of collected statistics */+ ZSTD_upscaleStats(&ms->opt);+}++size_t ZSTD_compressBlock_btultra(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ DEBUGLOG(5, "ZSTD_compressBlock_btultra (srcSize=%zu)", srcSize);+ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict);+}++size_t ZSTD_compressBlock_btultra2(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ U32 const current = (U32)((const BYTE*)src - ms->window.base);+ DEBUGLOG(5, "ZSTD_compressBlock_btultra2 (srcSize=%zu)", srcSize);++ /* 2-pass strategy:+ * this strategy makes a first pass over first block to collect statistics+ * and seed next round's statistics with it.+ * After 1st pass, function forgets everything, and starts a new block.+ * Consequently, this can only work if no data has been previously loaded in tables,+ * aka, no dictionary, no prefix, no ldm preprocessing.+ * The compression ratio gain is generally small (~0.5% on first block),+ * the cost is 2x cpu time on first block. */+ assert(srcSize <= ZSTD_BLOCKSIZE_MAX);+ if ( (ms->opt.litLengthSum==0) /* first block */+ && (seqStore->sequences == seqStore->sequencesStart) /* no ldm */+ && (ms->window.dictLimit == ms->window.lowLimit) /* no dictionary */+ && (current == ms->window.dictLimit) /* start of frame, nothing already loaded nor skipped */+ && (srcSize > ZSTD_PREDEF_THRESHOLD)+ ) {+ ZSTD_initStats_ultra(ms, seqStore, rep, src, srcSize);+ }++ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict);+}++size_t ZSTD_compressBlock_btopt_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /*optLevel*/, ZSTD_dictMatchState);+}++size_t ZSTD_compressBlock_btultra_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_dictMatchState);+}++size_t ZSTD_compressBlock_btopt_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /*optLevel*/, ZSTD_extDict);+}++size_t ZSTD_compressBlock_btultra_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ const void* src, size_t srcSize)+{+ return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /*optLevel*/, ZSTD_extDict);+}++/* note : no btultra2 variant for extDict nor dictMatchState,+ * because btultra2 is not meant to work with dictionaries+ * and is only specific for the first block (no prefix) */
zstd/lib/compress/zstd_opt.h view
@@ -1,919 +1,56 @@-/**- * Copyright (c) 2016-present, Przemyslaw Skibinski, Yann Collet, Facebook, Inc.+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ --/* Note : this file is intended to be included within zstd_compress.c */---#ifndef ZSTD_OPT_H_91842398743-#define ZSTD_OPT_H_91842398743---#define ZSTD_LITFREQ_ADD 2-#define ZSTD_FREQ_DIV 4-#define ZSTD_MAX_PRICE (1<<30)--/*-*************************************-* Price functions for optimal parser-***************************************/-FORCE_INLINE void ZSTD_setLog2Prices(seqStore_t* ssPtr)-{- ssPtr->log2matchLengthSum = ZSTD_highbit32(ssPtr->matchLengthSum+1);- ssPtr->log2litLengthSum = ZSTD_highbit32(ssPtr->litLengthSum+1);- ssPtr->log2litSum = ZSTD_highbit32(ssPtr->litSum+1);- ssPtr->log2offCodeSum = ZSTD_highbit32(ssPtr->offCodeSum+1);- ssPtr->factor = 1 + ((ssPtr->litSum>>5) / ssPtr->litLengthSum) + ((ssPtr->litSum<<1) / (ssPtr->litSum + ssPtr->matchSum));-}---MEM_STATIC void ZSTD_rescaleFreqs(seqStore_t* ssPtr, const BYTE* src, size_t srcSize)-{- unsigned u;-- ssPtr->cachedLiterals = NULL;- ssPtr->cachedPrice = ssPtr->cachedLitLength = 0;- ssPtr->staticPrices = 0; -- if (ssPtr->litLengthSum == 0) {- if (srcSize <= 1024) ssPtr->staticPrices = 1;-- for (u=0; u<=MaxLit; u++)- ssPtr->litFreq[u] = 0;- for (u=0; u<srcSize; u++)- ssPtr->litFreq[src[u]]++;-- ssPtr->litSum = 0;- ssPtr->litLengthSum = MaxLL+1;- ssPtr->matchLengthSum = MaxML+1;- ssPtr->offCodeSum = (MaxOff+1);- ssPtr->matchSum = (ZSTD_LITFREQ_ADD<<Litbits);-- for (u=0; u<=MaxLit; u++) {- ssPtr->litFreq[u] = 1 + (ssPtr->litFreq[u]>>ZSTD_FREQ_DIV);- ssPtr->litSum += ssPtr->litFreq[u]; - }- for (u=0; u<=MaxLL; u++)- ssPtr->litLengthFreq[u] = 1;- for (u=0; u<=MaxML; u++)- ssPtr->matchLengthFreq[u] = 1;- for (u=0; u<=MaxOff; u++)- ssPtr->offCodeFreq[u] = 1;- } else {- ssPtr->matchLengthSum = 0;- ssPtr->litLengthSum = 0;- ssPtr->offCodeSum = 0;- ssPtr->matchSum = 0;- ssPtr->litSum = 0;-- for (u=0; u<=MaxLit; u++) {- ssPtr->litFreq[u] = 1 + (ssPtr->litFreq[u]>>(ZSTD_FREQ_DIV+1));- ssPtr->litSum += ssPtr->litFreq[u];- }- for (u=0; u<=MaxLL; u++) {- ssPtr->litLengthFreq[u] = 1 + (ssPtr->litLengthFreq[u]>>(ZSTD_FREQ_DIV+1));- ssPtr->litLengthSum += ssPtr->litLengthFreq[u];- }- for (u=0; u<=MaxML; u++) {- ssPtr->matchLengthFreq[u] = 1 + (ssPtr->matchLengthFreq[u]>>ZSTD_FREQ_DIV);- ssPtr->matchLengthSum += ssPtr->matchLengthFreq[u];- ssPtr->matchSum += ssPtr->matchLengthFreq[u] * (u + 3);- }- ssPtr->matchSum *= ZSTD_LITFREQ_ADD;- for (u=0; u<=MaxOff; u++) {- ssPtr->offCodeFreq[u] = 1 + (ssPtr->offCodeFreq[u]>>ZSTD_FREQ_DIV);- ssPtr->offCodeSum += ssPtr->offCodeFreq[u];- }- }-- ZSTD_setLog2Prices(ssPtr);-}---FORCE_INLINE U32 ZSTD_getLiteralPrice(seqStore_t* ssPtr, U32 litLength, const BYTE* literals)-{- U32 price, u;-- if (ssPtr->staticPrices)- return ZSTD_highbit32((U32)litLength+1) + (litLength*6);-- if (litLength == 0)- return ssPtr->log2litLengthSum - ZSTD_highbit32(ssPtr->litLengthFreq[0]+1);-- /* literals */- if (ssPtr->cachedLiterals == literals) {- U32 const additional = litLength - ssPtr->cachedLitLength;- const BYTE* literals2 = ssPtr->cachedLiterals + ssPtr->cachedLitLength;- price = ssPtr->cachedPrice + additional * ssPtr->log2litSum;- for (u=0; u < additional; u++)- price -= ZSTD_highbit32(ssPtr->litFreq[literals2[u]]+1);- ssPtr->cachedPrice = price;- ssPtr->cachedLitLength = litLength;- } else {- price = litLength * ssPtr->log2litSum;- for (u=0; u < litLength; u++)- price -= ZSTD_highbit32(ssPtr->litFreq[literals[u]]+1);-- if (litLength >= 12) {- ssPtr->cachedLiterals = literals;- ssPtr->cachedPrice = price;- ssPtr->cachedLitLength = litLength;- }- }-- /* literal Length */- { const BYTE LL_deltaCode = 19;- const BYTE llCode = (litLength>63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];- price += LL_bits[llCode] + ssPtr->log2litLengthSum - ZSTD_highbit32(ssPtr->litLengthFreq[llCode]+1);- }-- return price;-}---FORCE_INLINE U32 ZSTD_getPrice(seqStore_t* seqStorePtr, U32 litLength, const BYTE* literals, U32 offset, U32 matchLength, const int ultra)-{- /* offset */- U32 price;- BYTE const offCode = (BYTE)ZSTD_highbit32(offset+1);-- if (seqStorePtr->staticPrices)- return ZSTD_getLiteralPrice(seqStorePtr, litLength, literals) + ZSTD_highbit32((U32)matchLength+1) + 16 + offCode;-- price = offCode + seqStorePtr->log2offCodeSum - ZSTD_highbit32(seqStorePtr->offCodeFreq[offCode]+1);- if (!ultra && offCode >= 20) price += (offCode-19)*2;-- /* match Length */- { const BYTE ML_deltaCode = 36;- const BYTE mlCode = (matchLength>127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];- price += ML_bits[mlCode] + seqStorePtr->log2matchLengthSum - ZSTD_highbit32(seqStorePtr->matchLengthFreq[mlCode]+1);- }-- return price + ZSTD_getLiteralPrice(seqStorePtr, litLength, literals) + seqStorePtr->factor;-}---MEM_STATIC void ZSTD_updatePrice(seqStore_t* seqStorePtr, U32 litLength, const BYTE* literals, U32 offset, U32 matchLength)-{- U32 u;-- /* literals */- seqStorePtr->litSum += litLength*ZSTD_LITFREQ_ADD;- for (u=0; u < litLength; u++)- seqStorePtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;-- /* literal Length */- { const BYTE LL_deltaCode = 19;- const BYTE llCode = (litLength>63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];- seqStorePtr->litLengthFreq[llCode]++;- seqStorePtr->litLengthSum++;- }-- /* match offset */- { BYTE const offCode = (BYTE)ZSTD_highbit32(offset+1);- seqStorePtr->offCodeSum++;- seqStorePtr->offCodeFreq[offCode]++;- }-- /* match Length */- { const BYTE ML_deltaCode = 36;- const BYTE mlCode = (matchLength>127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];- seqStorePtr->matchLengthFreq[mlCode]++;- seqStorePtr->matchLengthSum++;- }-- ZSTD_setLog2Prices(seqStorePtr);-}---#define SET_PRICE(pos, mlen_, offset_, litlen_, price_) \- { \- while (last_pos < pos) { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; } \- opt[pos].mlen = mlen_; \- opt[pos].off = offset_; \- opt[pos].litlen = litlen_; \- opt[pos].price = price_; \- }----/* Update hashTable3 up to ip (excluded)- Assumption : always within prefix (ie. not within extDict) */-FORCE_INLINE-U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_CCtx* zc, const BYTE* ip)-{- U32* const hashTable3 = zc->hashTable3;- U32 const hashLog3 = zc->hashLog3;- const BYTE* const base = zc->base;- U32 idx = zc->nextToUpdate3;- const U32 target = zc->nextToUpdate3 = (U32)(ip - base);- const size_t hash3 = ZSTD_hash3Ptr(ip, hashLog3);-- while(idx < target) {- hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx;- idx++;- }-- return hashTable3[hash3];-}---/*-*************************************-* Binary Tree search-***************************************/-static U32 ZSTD_insertBtAndGetAllMatches (- ZSTD_CCtx* zc,- const BYTE* const ip, const BYTE* const iLimit,- U32 nbCompares, const U32 mls,- U32 extDict, ZSTD_match_t* matches, const U32 minMatchLen)-{- const BYTE* const base = zc->base;- const U32 current = (U32)(ip-base);- const U32 hashLog = zc->params.cParams.hashLog;- const size_t h = ZSTD_hashPtr(ip, hashLog, mls);- U32* const hashTable = zc->hashTable;- U32 matchIndex = hashTable[h];- U32* const bt = zc->chainTable;- const U32 btLog = zc->params.cParams.chainLog - 1;- const U32 btMask= (1U << btLog) - 1;- size_t commonLengthSmaller=0, commonLengthLarger=0;- const BYTE* const dictBase = zc->dictBase;- const U32 dictLimit = zc->dictLimit;- const BYTE* const dictEnd = dictBase + dictLimit;- const BYTE* const prefixStart = base + dictLimit;- const U32 btLow = btMask >= current ? 0 : current - btMask;- const U32 windowLow = zc->lowLimit;- U32* smallerPtr = bt + 2*(current&btMask);- U32* largerPtr = bt + 2*(current&btMask) + 1;- U32 matchEndIdx = current+8;- U32 dummy32; /* to be nullified at the end */- U32 mnum = 0;-- const U32 minMatch = (mls == 3) ? 3 : 4;- size_t bestLength = minMatchLen-1;-- if (minMatch == 3) { /* HC3 match finder */- U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3 (zc, ip);- if (matchIndex3>windowLow && (current - matchIndex3 < (1<<18))) {- const BYTE* match;- size_t currentMl=0;- if ((!extDict) || matchIndex3 >= dictLimit) {- match = base + matchIndex3;- if (match[bestLength] == ip[bestLength]) currentMl = ZSTD_count(ip, match, iLimit);- } else {- match = dictBase + matchIndex3;- if (MEM_readMINMATCH(match, MINMATCH) == MEM_readMINMATCH(ip, MINMATCH)) /* assumption : matchIndex3 <= dictLimit-4 (by table construction) */- currentMl = ZSTD_count_2segments(ip+MINMATCH, match+MINMATCH, iLimit, dictEnd, prefixStart) + MINMATCH;- }-- /* save best solution */- if (currentMl > bestLength) {- bestLength = currentMl;- matches[mnum].off = ZSTD_REP_MOVE_OPT + current - matchIndex3;- matches[mnum].len = (U32)currentMl;- mnum++;- if (currentMl > ZSTD_OPT_NUM) goto update;- if (ip+currentMl == iLimit) goto update; /* best possible, and avoid read overflow*/- }- }- }-- hashTable[h] = current; /* Update Hash Table */-- while (nbCompares-- && (matchIndex > windowLow)) {- U32* nextPtr = bt + 2*(matchIndex & btMask);- size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */- const BYTE* match;-- if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {- match = base + matchIndex;- if (match[matchLength] == ip[matchLength]) {- matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iLimit) +1;- }- } else {- match = dictBase + matchIndex;- matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);- if (matchIndex+matchLength >= dictLimit)- match = base + matchIndex; /* to prepare for next usage of match[matchLength] */- }-- if (matchLength > bestLength) {- if (matchLength > matchEndIdx - matchIndex) matchEndIdx = matchIndex + (U32)matchLength;- bestLength = matchLength;- matches[mnum].off = ZSTD_REP_MOVE_OPT + current - matchIndex;- matches[mnum].len = (U32)matchLength;- mnum++;- if (matchLength > ZSTD_OPT_NUM) break;- if (ip+matchLength == iLimit) /* equal : no way to know if inf or sup */- break; /* drop, to guarantee consistency (miss a little bit of compression) */- }-- if (match[matchLength] < ip[matchLength]) {- /* match is smaller than current */- *smallerPtr = matchIndex; /* update smaller idx */- commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */- if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */- smallerPtr = nextPtr+1; /* new "smaller" => larger of match */- matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */- } else {- /* match is larger than current */- *largerPtr = matchIndex;- commonLengthLarger = matchLength;- if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */- largerPtr = nextPtr;- matchIndex = nextPtr[0];- } }-- *smallerPtr = *largerPtr = 0;--update:- zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;- return mnum;-}---/** Tree updater, providing best match */-static U32 ZSTD_BtGetAllMatches (- ZSTD_CCtx* zc,- const BYTE* const ip, const BYTE* const iLimit,- const U32 maxNbAttempts, const U32 mls, ZSTD_match_t* matches, const U32 minMatchLen)-{- if (ip < zc->base + zc->nextToUpdate) return 0; /* skipped area */- ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);- return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 0, matches, minMatchLen);-}---static U32 ZSTD_BtGetAllMatches_selectMLS (- ZSTD_CCtx* zc, /* Index table will be updated */- const BYTE* ip, const BYTE* const iHighLimit,- const U32 maxNbAttempts, const U32 matchLengthSearch, ZSTD_match_t* matches, const U32 minMatchLen)-{- switch(matchLengthSearch)- {- case 3 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);- default :- case 4 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);- case 5 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);- case 6 : return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);- }-}--/** Tree updater, providing best match */-static U32 ZSTD_BtGetAllMatches_extDict (- ZSTD_CCtx* zc,- const BYTE* const ip, const BYTE* const iLimit,- const U32 maxNbAttempts, const U32 mls, ZSTD_match_t* matches, const U32 minMatchLen)-{- if (ip < zc->base + zc->nextToUpdate) return 0; /* skipped area */- ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);- return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 1, matches, minMatchLen);-}---static U32 ZSTD_BtGetAllMatches_selectMLS_extDict (- ZSTD_CCtx* zc, /* Index table will be updated */- const BYTE* ip, const BYTE* const iHighLimit,- const U32 maxNbAttempts, const U32 matchLengthSearch, ZSTD_match_t* matches, const U32 minMatchLen)-{- switch(matchLengthSearch)- {- case 3 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);- default :- case 4 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);- case 5 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);- case 6 : return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);- }-}---/*-*******************************-* Optimal parser-*********************************/-FORCE_INLINE-void ZSTD_compressBlock_opt_generic(ZSTD_CCtx* ctx,- const void* src, size_t srcSize, const int ultra)-{- seqStore_t* seqStorePtr = &(ctx->seqStore);- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - 8;- const BYTE* const base = ctx->base;- const BYTE* const prefixStart = base + ctx->dictLimit;-- const U32 maxSearches = 1U << ctx->params.cParams.searchLog;- const U32 sufficient_len = ctx->params.cParams.targetLength;- const U32 mls = ctx->params.cParams.searchLength;- const U32 minMatch = (ctx->params.cParams.searchLength == 3) ? 3 : 4;-- ZSTD_optimal_t* opt = seqStorePtr->priceTable;- ZSTD_match_t* matches = seqStorePtr->matchTable;- const BYTE* inr;- U32 offset, rep[ZSTD_REP_NUM];-- /* init */- ctx->nextToUpdate3 = ctx->nextToUpdate;- ZSTD_rescaleFreqs(seqStorePtr, (const BYTE*)src, srcSize);- ip += (ip==prefixStart);- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) rep[i]=ctx->rep[i]; }-- /* Match Loop */- while (ip < ilimit) {- U32 cur, match_num, last_pos, litlen, price;- U32 u, mlen, best_mlen, best_off, litLength;- memset(opt, 0, sizeof(ZSTD_optimal_t));- last_pos = 0;- litlen = (U32)(ip - anchor);-- /* check repCode */- { U32 i, last_i = ZSTD_REP_CHECK + (ip==anchor);- for (i=(ip == anchor); i<last_i; i++) {- const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];- if ( (repCur > 0) && (repCur < (S32)(ip-prefixStart))- && (MEM_readMINMATCH(ip, minMatch) == MEM_readMINMATCH(ip - repCur, minMatch))) {- mlen = (U32)ZSTD_count(ip+minMatch, ip+minMatch-repCur, iend) + minMatch;- if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {- best_mlen = mlen; best_off = i; cur = 0; last_pos = 1;- goto _storeSequence;- }- best_off = i - (ip == anchor);- do {- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);- if (mlen > last_pos || price < opt[mlen].price)- SET_PRICE(mlen, mlen, i, litlen, price); /* note : macro modifies last_pos */- mlen--;- } while (mlen >= minMatch);- } } }-- match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, ip, iend, maxSearches, mls, matches, minMatch);-- if (!last_pos && !match_num) { ip++; continue; }-- if (match_num && (matches[match_num-1].len > sufficient_len || matches[match_num-1].len >= ZSTD_OPT_NUM)) {- best_mlen = matches[match_num-1].len;- best_off = matches[match_num-1].off;- cur = 0;- last_pos = 1;- goto _storeSequence;- }-- /* set prices using matches at position = 0 */- best_mlen = (last_pos) ? last_pos : minMatch;- for (u = 0; u < match_num; u++) {- mlen = (u>0) ? matches[u-1].len+1 : best_mlen;- best_mlen = matches[u].len;- while (mlen <= best_mlen) {- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);- if (mlen > last_pos || price < opt[mlen].price)- SET_PRICE(mlen, mlen, matches[u].off, litlen, price); /* note : macro modifies last_pos */- mlen++;- } }-- if (last_pos < minMatch) { ip++; continue; }-- /* initialize opt[0] */- { U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }- opt[0].mlen = 1;- opt[0].litlen = litlen;-- /* check further positions */- for (cur = 1; cur <= last_pos; cur++) {- inr = ip + cur;-- if (opt[cur-1].mlen == 1) {- litlen = opt[cur-1].litlen + 1;- if (cur > litlen) {- price = opt[cur - litlen].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr-litlen);- } else- price = ZSTD_getLiteralPrice(seqStorePtr, litlen, anchor);- } else {- litlen = 1;- price = opt[cur - 1].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr-1);- }-- if (cur > last_pos || price <= opt[cur].price)- SET_PRICE(cur, 1, 0, litlen, price);-- if (cur == last_pos) break;-- if (inr > ilimit) /* last match must start at a minimum distance of 8 from oend */- continue;-- mlen = opt[cur].mlen;- if (opt[cur].off > ZSTD_REP_MOVE_OPT) {- opt[cur].rep[2] = opt[cur-mlen].rep[1];- opt[cur].rep[1] = opt[cur-mlen].rep[0];- opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;- } else {- opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur-mlen].rep[1] : opt[cur-mlen].rep[2];- opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur-mlen].rep[0] : opt[cur-mlen].rep[1];- opt[cur].rep[0] = ((opt[cur].off==ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur-mlen].rep[0] - 1) : (opt[cur-mlen].rep[opt[cur].off]);- }-- best_mlen = minMatch;- { U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);- for (i=(opt[cur].mlen != 1); i<last_i; i++) { /* check rep */- const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];- if ( (repCur > 0) && (repCur < (S32)(inr-prefixStart))- && (MEM_readMINMATCH(inr, minMatch) == MEM_readMINMATCH(inr - repCur, minMatch))) {- mlen = (U32)ZSTD_count(inr+minMatch, inr+minMatch - repCur, iend) + minMatch;-- if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {- best_mlen = mlen; best_off = i; last_pos = cur + 1;- goto _storeSequence;- }-- best_off = i - (opt[cur].mlen != 1);- if (mlen > best_mlen) best_mlen = mlen;-- do {- if (opt[cur].mlen == 1) {- litlen = opt[cur].litlen;- if (cur > litlen) {- price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, inr-litlen, best_off, mlen - MINMATCH, ultra);- } else- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);- } else {- litlen = 0;- price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);- }-- if (cur + mlen > last_pos || price <= opt[cur + mlen].price)- SET_PRICE(cur + mlen, mlen, i, litlen, price);- mlen--;- } while (mlen >= minMatch);- } } }-- match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, inr, iend, maxSearches, mls, matches, best_mlen);-- if (match_num > 0 && (matches[match_num-1].len > sufficient_len || cur + matches[match_num-1].len >= ZSTD_OPT_NUM)) {- best_mlen = matches[match_num-1].len;- best_off = matches[match_num-1].off;- last_pos = cur + 1;- goto _storeSequence;- }-- /* set prices using matches at position = cur */- for (u = 0; u < match_num; u++) {- mlen = (u>0) ? matches[u-1].len+1 : best_mlen;- best_mlen = matches[u].len;-- while (mlen <= best_mlen) {- if (opt[cur].mlen == 1) {- litlen = opt[cur].litlen;- if (cur > litlen)- price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, ip+cur-litlen, matches[u].off-1, mlen - MINMATCH, ultra);- else- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);- } else {- litlen = 0;- price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, matches[u].off-1, mlen - MINMATCH, ultra);- }-- if (cur + mlen > last_pos || (price < opt[cur + mlen].price))- SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);-- mlen++;- } } }-- best_mlen = opt[last_pos].mlen;- best_off = opt[last_pos].off;- cur = last_pos - best_mlen;-- /* store sequence */-_storeSequence: /* cur, last_pos, best_mlen, best_off have to be set */- opt[0].mlen = 1;-- while (1) {- mlen = opt[cur].mlen;- offset = opt[cur].off;- opt[cur].mlen = best_mlen;- opt[cur].off = best_off;- best_mlen = mlen;- best_off = offset;- if (mlen > cur) break;- cur -= mlen;- }-- for (u = 0; u <= last_pos;) {- u += opt[u].mlen;- }-- for (cur=0; cur < last_pos; ) {- mlen = opt[cur].mlen;- if (mlen == 1) { ip++; cur++; continue; }- offset = opt[cur].off;- cur += mlen;- litLength = (U32)(ip - anchor);+#ifndef ZSTD_OPT_H+#define ZSTD_OPT_H - if (offset > ZSTD_REP_MOVE_OPT) {- rep[2] = rep[1];- rep[1] = rep[0];- rep[0] = offset - ZSTD_REP_MOVE_OPT;- offset--;- } else {- if (offset != 0) {- best_off = (offset==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);- if (offset != 1) rep[2] = rep[1];- rep[1] = rep[0];- rep[0] = best_off;- }- if (litLength==0) offset--;- }+#if defined (__cplusplus)+extern "C" {+#endif - ZSTD_updatePrice(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);- ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);- anchor = ip = ip + mlen;- } } /* for (cur=0; cur < last_pos; ) */+#include "zstd_compress_internal.h" - /* Save reps for next block */- { int i; for (i=0; i<ZSTD_REP_NUM; i++) ctx->savedRep[i] = rep[i]; }+/* used in ZSTD_loadDictionaryContent() */+void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend); - /* Last Literals */- { size_t const lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }-}+size_t ZSTD_compressBlock_btopt(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_btultra(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_btultra2(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize); -FORCE_INLINE-void ZSTD_compressBlock_opt_extDict_generic(ZSTD_CCtx* ctx,- const void* src, size_t srcSize, const int ultra)-{- seqStore_t* seqStorePtr = &(ctx->seqStore);- const BYTE* const istart = (const BYTE*)src;- const BYTE* ip = istart;- const BYTE* anchor = istart;- const BYTE* const iend = istart + srcSize;- const BYTE* const ilimit = iend - 8;- const BYTE* const base = ctx->base;- const U32 lowestIndex = ctx->lowLimit;- const U32 dictLimit = ctx->dictLimit;- const BYTE* const prefixStart = base + dictLimit;- const BYTE* const dictBase = ctx->dictBase;- const BYTE* const dictEnd = dictBase + dictLimit;-- const U32 maxSearches = 1U << ctx->params.cParams.searchLog;- const U32 sufficient_len = ctx->params.cParams.targetLength;- const U32 mls = ctx->params.cParams.searchLength;- const U32 minMatch = (ctx->params.cParams.searchLength == 3) ? 3 : 4;-- ZSTD_optimal_t* opt = seqStorePtr->priceTable;- ZSTD_match_t* matches = seqStorePtr->matchTable;- const BYTE* inr;-- /* init */- U32 offset, rep[ZSTD_REP_NUM];- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) rep[i]=ctx->rep[i]; }-- ctx->nextToUpdate3 = ctx->nextToUpdate;- ZSTD_rescaleFreqs(seqStorePtr, (const BYTE*)src, srcSize);- ip += (ip==prefixStart);-- /* Match Loop */- while (ip < ilimit) {- U32 cur, match_num, last_pos, litlen, price;- U32 u, mlen, best_mlen, best_off, litLength;- U32 current = (U32)(ip-base);- memset(opt, 0, sizeof(ZSTD_optimal_t));- last_pos = 0;- opt[0].litlen = (U32)(ip - anchor);-- /* check repCode */- { U32 i, last_i = ZSTD_REP_CHECK + (ip==anchor);- for (i = (ip==anchor); i<last_i; i++) {- const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];- const U32 repIndex = (U32)(current - repCur);- const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* const repMatch = repBase + repIndex;- if ( (repCur > 0 && repCur <= (S32)current)- && (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex>lowestIndex)) /* intentional overflow */- && (MEM_readMINMATCH(ip, minMatch) == MEM_readMINMATCH(repMatch, minMatch)) ) {- /* repcode detected we should take it */- const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;- mlen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iend, repEnd, prefixStart) + minMatch;-- if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {- best_mlen = mlen; best_off = i; cur = 0; last_pos = 1;- goto _storeSequence;- }-- best_off = i - (ip==anchor);- litlen = opt[0].litlen;- do {- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);- if (mlen > last_pos || price < opt[mlen].price)- SET_PRICE(mlen, mlen, i, litlen, price); /* note : macro modifies last_pos */- mlen--;- } while (mlen >= minMatch);- } } }-- match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, ip, iend, maxSearches, mls, matches, minMatch); /* first search (depth 0) */-- if (!last_pos && !match_num) { ip++; continue; }-- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }- opt[0].mlen = 1;-- if (match_num && (matches[match_num-1].len > sufficient_len || matches[match_num-1].len >= ZSTD_OPT_NUM)) {- best_mlen = matches[match_num-1].len;- best_off = matches[match_num-1].off;- cur = 0;- last_pos = 1;- goto _storeSequence;- }-- best_mlen = (last_pos) ? last_pos : minMatch;-- /* set prices using matches at position = 0 */- for (u = 0; u < match_num; u++) {- mlen = (u>0) ? matches[u-1].len+1 : best_mlen;- best_mlen = matches[u].len;- litlen = opt[0].litlen;- while (mlen <= best_mlen) {- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);- if (mlen > last_pos || price < opt[mlen].price)- SET_PRICE(mlen, mlen, matches[u].off, litlen, price);- mlen++;- } }-- if (last_pos < minMatch) {- ip++; continue;- }-- /* check further positions */- for (cur = 1; cur <= last_pos; cur++) {- inr = ip + cur;-- if (opt[cur-1].mlen == 1) {- litlen = opt[cur-1].litlen + 1;- if (cur > litlen) {- price = opt[cur - litlen].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr-litlen);- } else- price = ZSTD_getLiteralPrice(seqStorePtr, litlen, anchor);- } else {- litlen = 1;- price = opt[cur - 1].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr-1);- }-- if (cur > last_pos || price <= opt[cur].price)- SET_PRICE(cur, 1, 0, litlen, price);-- if (cur == last_pos) break;-- if (inr > ilimit) /* last match must start at a minimum distance of 8 from oend */- continue;-- mlen = opt[cur].mlen;- if (opt[cur].off > ZSTD_REP_MOVE_OPT) {- opt[cur].rep[2] = opt[cur-mlen].rep[1];- opt[cur].rep[1] = opt[cur-mlen].rep[0];- opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;- } else {- opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur-mlen].rep[1] : opt[cur-mlen].rep[2];- opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur-mlen].rep[0] : opt[cur-mlen].rep[1];- opt[cur].rep[0] = ((opt[cur].off==ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur-mlen].rep[0] - 1) : (opt[cur-mlen].rep[opt[cur].off]);- }-- best_mlen = minMatch;- { U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);- for (i = (mlen != 1); i<last_i; i++) {- const S32 repCur = (i==ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];- const U32 repIndex = (U32)(current+cur - repCur);- const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;- const BYTE* const repMatch = repBase + repIndex;- if ( (repCur > 0 && repCur <= (S32)(current+cur))- && (((U32)((dictLimit-1) - repIndex) >= 3) & (repIndex>lowestIndex)) /* intentional overflow */- && (MEM_readMINMATCH(inr, minMatch) == MEM_readMINMATCH(repMatch, minMatch)) ) {- /* repcode detected */- const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;- mlen = (U32)ZSTD_count_2segments(inr+minMatch, repMatch+minMatch, iend, repEnd, prefixStart) + minMatch;-- if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {- best_mlen = mlen; best_off = i; last_pos = cur + 1;- goto _storeSequence;- }-- best_off = i - (opt[cur].mlen != 1);- if (mlen > best_mlen) best_mlen = mlen;-- do {- if (opt[cur].mlen == 1) {- litlen = opt[cur].litlen;- if (cur > litlen) {- price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, inr-litlen, best_off, mlen - MINMATCH, ultra);- } else- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);- } else {- litlen = 0;- price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);- }-- if (cur + mlen > last_pos || price <= opt[cur + mlen].price)- SET_PRICE(cur + mlen, mlen, i, litlen, price);- mlen--;- } while (mlen >= minMatch);- } } }-- match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, inr, iend, maxSearches, mls, matches, minMatch);-- if (match_num > 0 && matches[match_num-1].len > sufficient_len) {- best_mlen = matches[match_num-1].len;- best_off = matches[match_num-1].off;- last_pos = cur + 1;- goto _storeSequence;- }-- /* set prices using matches at position = cur */- for (u = 0; u < match_num; u++) {- mlen = (u>0) ? matches[u-1].len+1 : best_mlen;- best_mlen = (cur + matches[u].len < ZSTD_OPT_NUM) ? matches[u].len : ZSTD_OPT_NUM - cur;-- while (mlen <= best_mlen) {- if (opt[cur].mlen == 1) {- litlen = opt[cur].litlen;- if (cur > litlen)- price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, ip+cur-litlen, matches[u].off-1, mlen - MINMATCH, ultra);- else- price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off-1, mlen - MINMATCH, ultra);- } else {- litlen = 0;- price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, matches[u].off-1, mlen - MINMATCH, ultra);- }-- if (cur + mlen > last_pos || (price < opt[cur + mlen].price))- SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);-- mlen++;- } } } /* for (cur = 1; cur <= last_pos; cur++) */-- best_mlen = opt[last_pos].mlen;- best_off = opt[last_pos].off;- cur = last_pos - best_mlen;-- /* store sequence */-_storeSequence: /* cur, last_pos, best_mlen, best_off have to be set */- opt[0].mlen = 1;-- while (1) {- mlen = opt[cur].mlen;- offset = opt[cur].off;- opt[cur].mlen = best_mlen;- opt[cur].off = best_off;- best_mlen = mlen;- best_off = offset;- if (mlen > cur) break;- cur -= mlen;- }-- for (u = 0; u <= last_pos; ) {- u += opt[u].mlen;- }-- for (cur=0; cur < last_pos; ) {- mlen = opt[cur].mlen;- if (mlen == 1) { ip++; cur++; continue; }- offset = opt[cur].off;- cur += mlen;- litLength = (U32)(ip - anchor);-- if (offset > ZSTD_REP_MOVE_OPT) {- rep[2] = rep[1];- rep[1] = rep[0];- rep[0] = offset - ZSTD_REP_MOVE_OPT;- offset--;- } else {- if (offset != 0) {- best_off = (offset==ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);- if (offset != 1) rep[2] = rep[1];- rep[1] = rep[0];- rep[0] = best_off;- }-- if (litLength==0) offset--;- }+size_t ZSTD_compressBlock_btopt_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_btultra_dictMatchState(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize); - ZSTD_updatePrice(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);- ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);- anchor = ip = ip + mlen;- } } /* for (cur=0; cur < last_pos; ) */+size_t ZSTD_compressBlock_btopt_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize);+size_t ZSTD_compressBlock_btultra_extDict(+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],+ void const* src, size_t srcSize); - /* Save reps for next block */- { int i; for (i=0; i<ZSTD_REP_NUM; i++) ctx->savedRep[i] = rep[i]; }+ /* note : no btultra2 variant for extDict nor dictMatchState,+ * because btultra2 is not meant to work with dictionaries+ * and is only specific for the first block (no prefix) */ - /* Last Literals */- { size_t lastLLSize = iend - anchor;- memcpy(seqStorePtr->lit, anchor, lastLLSize);- seqStorePtr->lit += lastLLSize;- }+#if defined (__cplusplus) }+#endif -#endif /* ZSTD_OPT_H_91842398743 */+#endif /* ZSTD_OPT_H */
+ zstd/lib/compress/zstdmt_compress.c view
@@ -0,0 +1,2099 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+++/* ====== Compiler specifics ====== */+#if defined(_MSC_VER)+# pragma warning(disable : 4204) /* disable: C4204: non-constant aggregate initializer */+#endif+++/* ====== Constants ====== */+#define ZSTDMT_OVERLAPLOG_DEFAULT 0+++/* ====== Dependencies ====== */+#include <string.h> /* memcpy, memset */+#include <limits.h> /* INT_MAX, UINT_MAX */+#include "mem.h" /* MEM_STATIC */+#include "pool.h" /* threadpool */+#include "threading.h" /* mutex */+#include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */+#include "zstd_ldm.h"+#include "zstdmt_compress.h"++/* Guards code to support resizing the SeqPool.+ * We will want to resize the SeqPool to save memory in the future.+ * Until then, comment the code out since it is unused.+ */+#define ZSTD_RESIZE_SEQPOOL 0++/* ====== Debug ====== */+#if defined(DEBUGLEVEL) && (DEBUGLEVEL>=2) \+ && !defined(_MSC_VER) \+ && !defined(__MINGW32__)++# include <stdio.h>+# include <unistd.h>+# include <sys/times.h>++# define DEBUG_PRINTHEX(l,p,n) { \+ unsigned debug_u; \+ for (debug_u=0; debug_u<(n); debug_u++) \+ RAWLOG(l, "%02X ", ((const unsigned char*)(p))[debug_u]); \+ RAWLOG(l, " \n"); \+}++static unsigned long long GetCurrentClockTimeMicroseconds(void)+{+ static clock_t _ticksPerSecond = 0;+ if (_ticksPerSecond <= 0) _ticksPerSecond = sysconf(_SC_CLK_TCK);++ { struct tms junk; clock_t newTicks = (clock_t) times(&junk);+ return ((((unsigned long long)newTicks)*(1000000))/_ticksPerSecond);+} }++#define MUTEX_WAIT_TIME_DLEVEL 6+#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) { \+ if (DEBUGLEVEL >= MUTEX_WAIT_TIME_DLEVEL) { \+ unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds(); \+ ZSTD_pthread_mutex_lock(mutex); \+ { unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \+ unsigned long long const elapsedTime = (afterTime-beforeTime); \+ if (elapsedTime > 1000) { /* or whatever threshold you like; I'm using 1 millisecond here */ \+ DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, "Thread took %llu microseconds to acquire mutex %s \n", \+ elapsedTime, #mutex); \+ } } \+ } else { \+ ZSTD_pthread_mutex_lock(mutex); \+ } \+}++#else++# define ZSTD_PTHREAD_MUTEX_LOCK(m) ZSTD_pthread_mutex_lock(m)+# define DEBUG_PRINTHEX(l,p,n) {}++#endif+++/* ===== Buffer Pool ===== */+/* a single Buffer Pool can be invoked from multiple threads in parallel */++typedef struct buffer_s {+ void* start;+ size_t capacity;+} buffer_t;++static const buffer_t g_nullBuffer = { NULL, 0 };++typedef struct ZSTDMT_bufferPool_s {+ ZSTD_pthread_mutex_t poolMutex;+ size_t bufferSize;+ unsigned totalBuffers;+ unsigned nbBuffers;+ ZSTD_customMem cMem;+ buffer_t bTable[1]; /* variable size */+} ZSTDMT_bufferPool;++static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned nbWorkers, ZSTD_customMem cMem)+{+ unsigned const maxNbBuffers = 2*nbWorkers + 3;+ ZSTDMT_bufferPool* const bufPool = (ZSTDMT_bufferPool*)ZSTD_calloc(+ sizeof(ZSTDMT_bufferPool) + (maxNbBuffers-1) * sizeof(buffer_t), cMem);+ if (bufPool==NULL) return NULL;+ if (ZSTD_pthread_mutex_init(&bufPool->poolMutex, NULL)) {+ ZSTD_free(bufPool, cMem);+ return NULL;+ }+ bufPool->bufferSize = 64 KB;+ bufPool->totalBuffers = maxNbBuffers;+ bufPool->nbBuffers = 0;+ bufPool->cMem = cMem;+ return bufPool;+}++static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool)+{+ unsigned u;+ DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool);+ if (!bufPool) return; /* compatibility with free on NULL */+ for (u=0; u<bufPool->totalBuffers; u++) {+ DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->bTable[u].start);+ ZSTD_free(bufPool->bTable[u].start, bufPool->cMem);+ }+ ZSTD_pthread_mutex_destroy(&bufPool->poolMutex);+ ZSTD_free(bufPool, bufPool->cMem);+}++/* only works at initialization, not during compression */+static size_t ZSTDMT_sizeof_bufferPool(ZSTDMT_bufferPool* bufPool)+{+ size_t const poolSize = sizeof(*bufPool)+ + (bufPool->totalBuffers - 1) * sizeof(buffer_t);+ unsigned u;+ size_t totalBufferSize = 0;+ ZSTD_pthread_mutex_lock(&bufPool->poolMutex);+ for (u=0; u<bufPool->totalBuffers; u++)+ totalBufferSize += bufPool->bTable[u].capacity;+ ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);++ return poolSize + totalBufferSize;+}++/* ZSTDMT_setBufferSize() :+ * all future buffers provided by this buffer pool will have _at least_ this size+ * note : it's better for all buffers to have same size,+ * as they become freely interchangeable, reducing malloc/free usages and memory fragmentation */+static void ZSTDMT_setBufferSize(ZSTDMT_bufferPool* const bufPool, size_t const bSize)+{+ ZSTD_pthread_mutex_lock(&bufPool->poolMutex);+ DEBUGLOG(4, "ZSTDMT_setBufferSize: bSize = %u", (U32)bSize);+ bufPool->bufferSize = bSize;+ ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);+}+++static ZSTDMT_bufferPool* ZSTDMT_expandBufferPool(ZSTDMT_bufferPool* srcBufPool, U32 nbWorkers)+{+ unsigned const maxNbBuffers = 2*nbWorkers + 3;+ if (srcBufPool==NULL) return NULL;+ if (srcBufPool->totalBuffers >= maxNbBuffers) /* good enough */+ return srcBufPool;+ /* need a larger buffer pool */+ { ZSTD_customMem const cMem = srcBufPool->cMem;+ size_t const bSize = srcBufPool->bufferSize; /* forward parameters */+ ZSTDMT_bufferPool* newBufPool;+ ZSTDMT_freeBufferPool(srcBufPool);+ newBufPool = ZSTDMT_createBufferPool(nbWorkers, cMem);+ if (newBufPool==NULL) return newBufPool;+ ZSTDMT_setBufferSize(newBufPool, bSize);+ return newBufPool;+ }+}++/** ZSTDMT_getBuffer() :+ * assumption : bufPool must be valid+ * @return : a buffer, with start pointer and size+ * note: allocation may fail, in this case, start==NULL and size==0 */+static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)+{+ size_t const bSize = bufPool->bufferSize;+ DEBUGLOG(5, "ZSTDMT_getBuffer: bSize = %u", (U32)bufPool->bufferSize);+ ZSTD_pthread_mutex_lock(&bufPool->poolMutex);+ if (bufPool->nbBuffers) { /* try to use an existing buffer */+ buffer_t const buf = bufPool->bTable[--(bufPool->nbBuffers)];+ size_t const availBufferSize = buf.capacity;+ bufPool->bTable[bufPool->nbBuffers] = g_nullBuffer;+ if ((availBufferSize >= bSize) & ((availBufferSize>>3) <= bSize)) {+ /* large enough, but not too much */+ DEBUGLOG(5, "ZSTDMT_getBuffer: provide buffer %u of size %u",+ bufPool->nbBuffers, (U32)buf.capacity);+ ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);+ return buf;+ }+ /* size conditions not respected : scratch this buffer, create new one */+ DEBUGLOG(5, "ZSTDMT_getBuffer: existing buffer does not meet size conditions => freeing");+ ZSTD_free(buf.start, bufPool->cMem);+ }+ ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);+ /* create new buffer */+ DEBUGLOG(5, "ZSTDMT_getBuffer: create a new buffer");+ { buffer_t buffer;+ void* const start = ZSTD_malloc(bSize, bufPool->cMem);+ buffer.start = start; /* note : start can be NULL if malloc fails ! */+ buffer.capacity = (start==NULL) ? 0 : bSize;+ if (start==NULL) {+ DEBUGLOG(5, "ZSTDMT_getBuffer: buffer allocation failure !!");+ } else {+ DEBUGLOG(5, "ZSTDMT_getBuffer: created buffer of size %u", (U32)bSize);+ }+ return buffer;+ }+}++#if ZSTD_RESIZE_SEQPOOL+/** ZSTDMT_resizeBuffer() :+ * assumption : bufPool must be valid+ * @return : a buffer that is at least the buffer pool buffer size.+ * If a reallocation happens, the data in the input buffer is copied.+ */+static buffer_t ZSTDMT_resizeBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buffer)+{+ size_t const bSize = bufPool->bufferSize;+ if (buffer.capacity < bSize) {+ void* const start = ZSTD_malloc(bSize, bufPool->cMem);+ buffer_t newBuffer;+ newBuffer.start = start;+ newBuffer.capacity = start == NULL ? 0 : bSize;+ if (start != NULL) {+ assert(newBuffer.capacity >= buffer.capacity);+ memcpy(newBuffer.start, buffer.start, buffer.capacity);+ DEBUGLOG(5, "ZSTDMT_resizeBuffer: created buffer of size %u", (U32)bSize);+ return newBuffer;+ }+ DEBUGLOG(5, "ZSTDMT_resizeBuffer: buffer allocation failure !!");+ }+ return buffer;+}+#endif++/* store buffer for later re-use, up to pool capacity */+static void ZSTDMT_releaseBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buf)+{+ DEBUGLOG(5, "ZSTDMT_releaseBuffer");+ if (buf.start == NULL) return; /* compatible with release on NULL */+ ZSTD_pthread_mutex_lock(&bufPool->poolMutex);+ if (bufPool->nbBuffers < bufPool->totalBuffers) {+ bufPool->bTable[bufPool->nbBuffers++] = buf; /* stored for later use */+ DEBUGLOG(5, "ZSTDMT_releaseBuffer: stored buffer of size %u in slot %u",+ (U32)buf.capacity, (U32)(bufPool->nbBuffers-1));+ ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);+ return;+ }+ ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);+ /* Reached bufferPool capacity (should not happen) */+ DEBUGLOG(5, "ZSTDMT_releaseBuffer: pool capacity reached => freeing ");+ ZSTD_free(buf.start, bufPool->cMem);+}+++/* ===== Seq Pool Wrapper ====== */++static rawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0};++typedef ZSTDMT_bufferPool ZSTDMT_seqPool;++static size_t ZSTDMT_sizeof_seqPool(ZSTDMT_seqPool* seqPool)+{+ return ZSTDMT_sizeof_bufferPool(seqPool);+}++static rawSeqStore_t bufferToSeq(buffer_t buffer)+{+ rawSeqStore_t seq = {NULL, 0, 0, 0};+ seq.seq = (rawSeq*)buffer.start;+ seq.capacity = buffer.capacity / sizeof(rawSeq);+ return seq;+}++static buffer_t seqToBuffer(rawSeqStore_t seq)+{+ buffer_t buffer;+ buffer.start = seq.seq;+ buffer.capacity = seq.capacity * sizeof(rawSeq);+ return buffer;+}++static rawSeqStore_t ZSTDMT_getSeq(ZSTDMT_seqPool* seqPool)+{+ if (seqPool->bufferSize == 0) {+ return kNullRawSeqStore;+ }+ return bufferToSeq(ZSTDMT_getBuffer(seqPool));+}++#if ZSTD_RESIZE_SEQPOOL+static rawSeqStore_t ZSTDMT_resizeSeq(ZSTDMT_seqPool* seqPool, rawSeqStore_t seq)+{+ return bufferToSeq(ZSTDMT_resizeBuffer(seqPool, seqToBuffer(seq)));+}+#endif++static void ZSTDMT_releaseSeq(ZSTDMT_seqPool* seqPool, rawSeqStore_t seq)+{+ ZSTDMT_releaseBuffer(seqPool, seqToBuffer(seq));+}++static void ZSTDMT_setNbSeq(ZSTDMT_seqPool* const seqPool, size_t const nbSeq)+{+ ZSTDMT_setBufferSize(seqPool, nbSeq * sizeof(rawSeq));+}++static ZSTDMT_seqPool* ZSTDMT_createSeqPool(unsigned nbWorkers, ZSTD_customMem cMem)+{+ ZSTDMT_seqPool* const seqPool = ZSTDMT_createBufferPool(nbWorkers, cMem);+ if (seqPool == NULL) return NULL;+ ZSTDMT_setNbSeq(seqPool, 0);+ return seqPool;+}++static void ZSTDMT_freeSeqPool(ZSTDMT_seqPool* seqPool)+{+ ZSTDMT_freeBufferPool(seqPool);+}++static ZSTDMT_seqPool* ZSTDMT_expandSeqPool(ZSTDMT_seqPool* pool, U32 nbWorkers)+{+ return ZSTDMT_expandBufferPool(pool, nbWorkers);+}+++/* ===== CCtx Pool ===== */+/* a single CCtx Pool can be invoked from multiple threads in parallel */++typedef struct {+ ZSTD_pthread_mutex_t poolMutex;+ int totalCCtx;+ int availCCtx;+ ZSTD_customMem cMem;+ ZSTD_CCtx* cctx[1]; /* variable size */+} ZSTDMT_CCtxPool;++/* note : all CCtx borrowed from the pool should be released back to the pool _before_ freeing the pool */+static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool)+{+ int cid;+ for (cid=0; cid<pool->totalCCtx; cid++)+ ZSTD_freeCCtx(pool->cctx[cid]); /* note : compatible with free on NULL */+ ZSTD_pthread_mutex_destroy(&pool->poolMutex);+ ZSTD_free(pool, pool->cMem);+}++/* ZSTDMT_createCCtxPool() :+ * implies nbWorkers >= 1 , checked by caller ZSTDMT_createCCtx() */+static ZSTDMT_CCtxPool* ZSTDMT_createCCtxPool(int nbWorkers,+ ZSTD_customMem cMem)+{+ ZSTDMT_CCtxPool* const cctxPool = (ZSTDMT_CCtxPool*) ZSTD_calloc(+ sizeof(ZSTDMT_CCtxPool) + (nbWorkers-1)*sizeof(ZSTD_CCtx*), cMem);+ assert(nbWorkers > 0);+ if (!cctxPool) return NULL;+ if (ZSTD_pthread_mutex_init(&cctxPool->poolMutex, NULL)) {+ ZSTD_free(cctxPool, cMem);+ return NULL;+ }+ cctxPool->cMem = cMem;+ cctxPool->totalCCtx = nbWorkers;+ cctxPool->availCCtx = 1; /* at least one cctx for single-thread mode */+ cctxPool->cctx[0] = ZSTD_createCCtx_advanced(cMem);+ if (!cctxPool->cctx[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; }+ DEBUGLOG(3, "cctxPool created, with %u workers", nbWorkers);+ return cctxPool;+}++static ZSTDMT_CCtxPool* ZSTDMT_expandCCtxPool(ZSTDMT_CCtxPool* srcPool,+ int nbWorkers)+{+ if (srcPool==NULL) return NULL;+ if (nbWorkers <= srcPool->totalCCtx) return srcPool; /* good enough */+ /* need a larger cctx pool */+ { ZSTD_customMem const cMem = srcPool->cMem;+ ZSTDMT_freeCCtxPool(srcPool);+ return ZSTDMT_createCCtxPool(nbWorkers, cMem);+ }+}++/* only works during initialization phase, not during compression */+static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool)+{+ ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);+ { unsigned const nbWorkers = cctxPool->totalCCtx;+ size_t const poolSize = sizeof(*cctxPool)+ + (nbWorkers-1) * sizeof(ZSTD_CCtx*);+ unsigned u;+ size_t totalCCtxSize = 0;+ for (u=0; u<nbWorkers; u++) {+ totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctx[u]);+ }+ ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);+ assert(nbWorkers > 0);+ return poolSize + totalCCtxSize;+ }+}++static ZSTD_CCtx* ZSTDMT_getCCtx(ZSTDMT_CCtxPool* cctxPool)+{+ DEBUGLOG(5, "ZSTDMT_getCCtx");+ ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);+ if (cctxPool->availCCtx) {+ cctxPool->availCCtx--;+ { ZSTD_CCtx* const cctx = cctxPool->cctx[cctxPool->availCCtx];+ ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);+ return cctx;+ } }+ ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);+ DEBUGLOG(5, "create one more CCtx");+ return ZSTD_createCCtx_advanced(cctxPool->cMem); /* note : can be NULL, when creation fails ! */+}++static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx)+{+ if (cctx==NULL) return; /* compatibility with release on NULL */+ ZSTD_pthread_mutex_lock(&pool->poolMutex);+ if (pool->availCCtx < pool->totalCCtx)+ pool->cctx[pool->availCCtx++] = cctx;+ else {+ /* pool overflow : should not happen, since totalCCtx==nbWorkers */+ DEBUGLOG(4, "CCtx pool overflow : free cctx");+ ZSTD_freeCCtx(cctx);+ }+ ZSTD_pthread_mutex_unlock(&pool->poolMutex);+}++/* ==== Serial State ==== */++typedef struct {+ void const* start;+ size_t size;+} range_t;++typedef struct {+ /* All variables in the struct are protected by mutex. */+ ZSTD_pthread_mutex_t mutex;+ ZSTD_pthread_cond_t cond;+ ZSTD_CCtx_params params;+ ldmState_t ldmState;+ XXH64_state_t xxhState;+ unsigned nextJobID;+ /* Protects ldmWindow.+ * Must be acquired after the main mutex when acquiring both.+ */+ ZSTD_pthread_mutex_t ldmWindowMutex;+ ZSTD_pthread_cond_t ldmWindowCond; /* Signaled when ldmWindow is updated */+ ZSTD_window_t ldmWindow; /* A thread-safe copy of ldmState.window */+} serialState_t;++static int ZSTDMT_serialState_reset(serialState_t* serialState, ZSTDMT_seqPool* seqPool, ZSTD_CCtx_params params, size_t jobSize)+{+ /* Adjust parameters */+ if (params.ldmParams.enableLdm) {+ DEBUGLOG(4, "LDM window size = %u KB", (1U << params.cParams.windowLog) >> 10);+ ZSTD_ldm_adjustParameters(¶ms.ldmParams, ¶ms.cParams);+ assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog);+ assert(params.ldmParams.hashRateLog < 32);+ serialState->ldmState.hashPower =+ ZSTD_rollingHash_primePower(params.ldmParams.minMatchLength);+ } else {+ memset(¶ms.ldmParams, 0, sizeof(params.ldmParams));+ }+ serialState->nextJobID = 0;+ if (params.fParams.checksumFlag)+ XXH64_reset(&serialState->xxhState, 0);+ if (params.ldmParams.enableLdm) {+ ZSTD_customMem cMem = params.customMem;+ unsigned const hashLog = params.ldmParams.hashLog;+ size_t const hashSize = ((size_t)1 << hashLog) * sizeof(ldmEntry_t);+ unsigned const bucketLog =+ params.ldmParams.hashLog - params.ldmParams.bucketSizeLog;+ size_t const bucketSize = (size_t)1 << bucketLog;+ unsigned const prevBucketLog =+ serialState->params.ldmParams.hashLog -+ serialState->params.ldmParams.bucketSizeLog;+ /* Size the seq pool tables */+ ZSTDMT_setNbSeq(seqPool, ZSTD_ldm_getMaxNbSeq(params.ldmParams, jobSize));+ /* Reset the window */+ ZSTD_window_clear(&serialState->ldmState.window);+ serialState->ldmWindow = serialState->ldmState.window;+ /* Resize tables and output space if necessary. */+ if (serialState->ldmState.hashTable == NULL || serialState->params.ldmParams.hashLog < hashLog) {+ ZSTD_free(serialState->ldmState.hashTable, cMem);+ serialState->ldmState.hashTable = (ldmEntry_t*)ZSTD_malloc(hashSize, cMem);+ }+ if (serialState->ldmState.bucketOffsets == NULL || prevBucketLog < bucketLog) {+ ZSTD_free(serialState->ldmState.bucketOffsets, cMem);+ serialState->ldmState.bucketOffsets = (BYTE*)ZSTD_malloc(bucketSize, cMem);+ }+ if (!serialState->ldmState.hashTable || !serialState->ldmState.bucketOffsets)+ return 1;+ /* Zero the tables */+ memset(serialState->ldmState.hashTable, 0, hashSize);+ memset(serialState->ldmState.bucketOffsets, 0, bucketSize);+ }+ serialState->params = params;+ serialState->params.jobSize = (U32)jobSize;+ return 0;+}++static int ZSTDMT_serialState_init(serialState_t* serialState)+{+ int initError = 0;+ memset(serialState, 0, sizeof(*serialState));+ initError |= ZSTD_pthread_mutex_init(&serialState->mutex, NULL);+ initError |= ZSTD_pthread_cond_init(&serialState->cond, NULL);+ initError |= ZSTD_pthread_mutex_init(&serialState->ldmWindowMutex, NULL);+ initError |= ZSTD_pthread_cond_init(&serialState->ldmWindowCond, NULL);+ return initError;+}++static void ZSTDMT_serialState_free(serialState_t* serialState)+{+ ZSTD_customMem cMem = serialState->params.customMem;+ ZSTD_pthread_mutex_destroy(&serialState->mutex);+ ZSTD_pthread_cond_destroy(&serialState->cond);+ ZSTD_pthread_mutex_destroy(&serialState->ldmWindowMutex);+ ZSTD_pthread_cond_destroy(&serialState->ldmWindowCond);+ ZSTD_free(serialState->ldmState.hashTable, cMem);+ ZSTD_free(serialState->ldmState.bucketOffsets, cMem);+}++static void ZSTDMT_serialState_update(serialState_t* serialState,+ ZSTD_CCtx* jobCCtx, rawSeqStore_t seqStore,+ range_t src, unsigned jobID)+{+ /* Wait for our turn */+ ZSTD_PTHREAD_MUTEX_LOCK(&serialState->mutex);+ while (serialState->nextJobID < jobID) {+ DEBUGLOG(5, "wait for serialState->cond");+ ZSTD_pthread_cond_wait(&serialState->cond, &serialState->mutex);+ }+ /* A future job may error and skip our job */+ if (serialState->nextJobID == jobID) {+ /* It is now our turn, do any processing necessary */+ if (serialState->params.ldmParams.enableLdm) {+ size_t error;+ assert(seqStore.seq != NULL && seqStore.pos == 0 &&+ seqStore.size == 0 && seqStore.capacity > 0);+ assert(src.size <= serialState->params.jobSize);+ ZSTD_window_update(&serialState->ldmState.window, src.start, src.size);+ error = ZSTD_ldm_generateSequences(+ &serialState->ldmState, &seqStore,+ &serialState->params.ldmParams, src.start, src.size);+ /* We provide a large enough buffer to never fail. */+ assert(!ZSTD_isError(error)); (void)error;+ /* Update ldmWindow to match the ldmState.window and signal the main+ * thread if it is waiting for a buffer.+ */+ ZSTD_PTHREAD_MUTEX_LOCK(&serialState->ldmWindowMutex);+ serialState->ldmWindow = serialState->ldmState.window;+ ZSTD_pthread_cond_signal(&serialState->ldmWindowCond);+ ZSTD_pthread_mutex_unlock(&serialState->ldmWindowMutex);+ }+ if (serialState->params.fParams.checksumFlag && src.size > 0)+ XXH64_update(&serialState->xxhState, src.start, src.size);+ }+ /* Now it is the next jobs turn */+ serialState->nextJobID++;+ ZSTD_pthread_cond_broadcast(&serialState->cond);+ ZSTD_pthread_mutex_unlock(&serialState->mutex);++ if (seqStore.size > 0) {+ size_t const err = ZSTD_referenceExternalSequences(+ jobCCtx, seqStore.seq, seqStore.size);+ assert(serialState->params.ldmParams.enableLdm);+ assert(!ZSTD_isError(err));+ (void)err;+ }+}++static void ZSTDMT_serialState_ensureFinished(serialState_t* serialState,+ unsigned jobID, size_t cSize)+{+ ZSTD_PTHREAD_MUTEX_LOCK(&serialState->mutex);+ if (serialState->nextJobID <= jobID) {+ assert(ZSTD_isError(cSize)); (void)cSize;+ DEBUGLOG(5, "Skipping past job %u because of error", jobID);+ serialState->nextJobID = jobID + 1;+ ZSTD_pthread_cond_broadcast(&serialState->cond);++ ZSTD_PTHREAD_MUTEX_LOCK(&serialState->ldmWindowMutex);+ ZSTD_window_clear(&serialState->ldmWindow);+ ZSTD_pthread_cond_signal(&serialState->ldmWindowCond);+ ZSTD_pthread_mutex_unlock(&serialState->ldmWindowMutex);+ }+ ZSTD_pthread_mutex_unlock(&serialState->mutex);++}+++/* ------------------------------------------ */+/* ===== Worker thread ===== */+/* ------------------------------------------ */++static const range_t kNullRange = { NULL, 0 };++typedef struct {+ size_t consumed; /* SHARED - set0 by mtctx, then modified by worker AND read by mtctx */+ size_t cSize; /* SHARED - set0 by mtctx, then modified by worker AND read by mtctx, then set0 by mtctx */+ ZSTD_pthread_mutex_t job_mutex; /* Thread-safe - used by mtctx and worker */+ ZSTD_pthread_cond_t job_cond; /* Thread-safe - used by mtctx and worker */+ ZSTDMT_CCtxPool* cctxPool; /* Thread-safe - used by mtctx and (all) workers */+ ZSTDMT_bufferPool* bufPool; /* Thread-safe - used by mtctx and (all) workers */+ ZSTDMT_seqPool* seqPool; /* Thread-safe - used by mtctx and (all) workers */+ serialState_t* serial; /* Thread-safe - used by mtctx and (all) workers */+ buffer_t dstBuff; /* set by worker (or mtctx), then read by worker & mtctx, then modified by mtctx => no barrier */+ range_t prefix; /* set by mtctx, then read by worker & mtctx => no barrier */+ range_t src; /* set by mtctx, then read by worker & mtctx => no barrier */+ unsigned jobID; /* set by mtctx, then read by worker => no barrier */+ unsigned firstJob; /* set by mtctx, then read by worker => no barrier */+ unsigned lastJob; /* set by mtctx, then read by worker => no barrier */+ ZSTD_CCtx_params params; /* set by mtctx, then read by worker => no barrier */+ const ZSTD_CDict* cdict; /* set by mtctx, then read by worker => no barrier */+ unsigned long long fullFrameSize; /* set by mtctx, then read by worker => no barrier */+ size_t dstFlushed; /* used only by mtctx */+ unsigned frameChecksumNeeded; /* used only by mtctx */+} ZSTDMT_jobDescription;++#define JOB_ERROR(e) { \+ ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex); \+ job->cSize = e; \+ ZSTD_pthread_mutex_unlock(&job->job_mutex); \+ goto _endJob; \+}++/* ZSTDMT_compressionJob() is a POOL_function type */+static void ZSTDMT_compressionJob(void* jobDescription)+{+ ZSTDMT_jobDescription* const job = (ZSTDMT_jobDescription*)jobDescription;+ ZSTD_CCtx_params jobParams = job->params; /* do not modify job->params ! copy it, modify the copy */+ ZSTD_CCtx* const cctx = ZSTDMT_getCCtx(job->cctxPool);+ rawSeqStore_t rawSeqStore = ZSTDMT_getSeq(job->seqPool);+ buffer_t dstBuff = job->dstBuff;+ size_t lastCBlockSize = 0;++ /* resources */+ if (cctx==NULL) JOB_ERROR(ERROR(memory_allocation));+ if (dstBuff.start == NULL) { /* streaming job : doesn't provide a dstBuffer */+ dstBuff = ZSTDMT_getBuffer(job->bufPool);+ if (dstBuff.start==NULL) JOB_ERROR(ERROR(memory_allocation));+ job->dstBuff = dstBuff; /* this value can be read in ZSTDMT_flush, when it copies the whole job */+ }+ if (jobParams.ldmParams.enableLdm && rawSeqStore.seq == NULL)+ JOB_ERROR(ERROR(memory_allocation));++ /* Don't compute the checksum for chunks, since we compute it externally,+ * but write it in the header.+ */+ if (job->jobID != 0) jobParams.fParams.checksumFlag = 0;+ /* Don't run LDM for the chunks, since we handle it externally */+ jobParams.ldmParams.enableLdm = 0;+++ /* init */+ if (job->cdict) {+ size_t const initError = ZSTD_compressBegin_advanced_internal(cctx, NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast, job->cdict, jobParams, job->fullFrameSize);+ assert(job->firstJob); /* only allowed for first job */+ if (ZSTD_isError(initError)) JOB_ERROR(initError);+ } else { /* srcStart points at reloaded section */+ U64 const pledgedSrcSize = job->firstJob ? job->fullFrameSize : job->src.size;+ { size_t const forceWindowError = ZSTD_CCtxParams_setParameter(&jobParams, ZSTD_c_forceMaxWindow, !job->firstJob);+ if (ZSTD_isError(forceWindowError)) JOB_ERROR(forceWindowError);+ }+ { size_t const initError = ZSTD_compressBegin_advanced_internal(cctx,+ job->prefix.start, job->prefix.size, ZSTD_dct_rawContent, /* load dictionary in "content-only" mode (no header analysis) */+ ZSTD_dtlm_fast,+ NULL, /*cdict*/+ jobParams, pledgedSrcSize);+ if (ZSTD_isError(initError)) JOB_ERROR(initError);+ } }++ /* Perform serial step as early as possible, but after CCtx initialization */+ ZSTDMT_serialState_update(job->serial, cctx, rawSeqStore, job->src, job->jobID);++ if (!job->firstJob) { /* flush and overwrite frame header when it's not first job */+ size_t const hSize = ZSTD_compressContinue(cctx, dstBuff.start, dstBuff.capacity, job->src.start, 0);+ if (ZSTD_isError(hSize)) JOB_ERROR(hSize);+ DEBUGLOG(5, "ZSTDMT_compressionJob: flush and overwrite %u bytes of frame header (not first job)", (U32)hSize);+ ZSTD_invalidateRepCodes(cctx);+ }++ /* compress */+ { size_t const chunkSize = 4*ZSTD_BLOCKSIZE_MAX;+ int const nbChunks = (int)((job->src.size + (chunkSize-1)) / chunkSize);+ const BYTE* ip = (const BYTE*) job->src.start;+ BYTE* const ostart = (BYTE*)dstBuff.start;+ BYTE* op = ostart;+ BYTE* oend = op + dstBuff.capacity;+ int chunkNb;+ if (sizeof(size_t) > sizeof(int)) assert(job->src.size < ((size_t)INT_MAX) * chunkSize); /* check overflow */+ DEBUGLOG(5, "ZSTDMT_compressionJob: compress %u bytes in %i blocks", (U32)job->src.size, nbChunks);+ assert(job->cSize == 0);+ for (chunkNb = 1; chunkNb < nbChunks; chunkNb++) {+ size_t const cSize = ZSTD_compressContinue(cctx, op, oend-op, ip, chunkSize);+ if (ZSTD_isError(cSize)) JOB_ERROR(cSize);+ ip += chunkSize;+ op += cSize; assert(op < oend);+ /* stats */+ ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex);+ job->cSize += cSize;+ job->consumed = chunkSize * chunkNb;+ DEBUGLOG(5, "ZSTDMT_compressionJob: compress new block : cSize==%u bytes (total: %u)",+ (U32)cSize, (U32)job->cSize);+ ZSTD_pthread_cond_signal(&job->job_cond); /* warns some more data is ready to be flushed */+ ZSTD_pthread_mutex_unlock(&job->job_mutex);+ }+ /* last block */+ assert(chunkSize > 0);+ assert((chunkSize & (chunkSize - 1)) == 0); /* chunkSize must be power of 2 for mask==(chunkSize-1) to work */+ if ((nbChunks > 0) | job->lastJob /*must output a "last block" flag*/ ) {+ size_t const lastBlockSize1 = job->src.size & (chunkSize-1);+ size_t const lastBlockSize = ((lastBlockSize1==0) & (job->src.size>=chunkSize)) ? chunkSize : lastBlockSize1;+ size_t const cSize = (job->lastJob) ?+ ZSTD_compressEnd (cctx, op, oend-op, ip, lastBlockSize) :+ ZSTD_compressContinue(cctx, op, oend-op, ip, lastBlockSize);+ if (ZSTD_isError(cSize)) JOB_ERROR(cSize);+ lastCBlockSize = cSize;+ } }++_endJob:+ ZSTDMT_serialState_ensureFinished(job->serial, job->jobID, job->cSize);+ if (job->prefix.size > 0)+ DEBUGLOG(5, "Finished with prefix: %zx", (size_t)job->prefix.start);+ DEBUGLOG(5, "Finished with source: %zx", (size_t)job->src.start);+ /* release resources */+ ZSTDMT_releaseSeq(job->seqPool, rawSeqStore);+ ZSTDMT_releaseCCtx(job->cctxPool, cctx);+ /* report */+ ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex);+ if (ZSTD_isError(job->cSize)) assert(lastCBlockSize == 0);+ job->cSize += lastCBlockSize;+ job->consumed = job->src.size; /* when job->consumed == job->src.size , compression job is presumed completed */+ ZSTD_pthread_cond_signal(&job->job_cond);+ ZSTD_pthread_mutex_unlock(&job->job_mutex);+}+++/* ------------------------------------------ */+/* ===== Multi-threaded compression ===== */+/* ------------------------------------------ */++typedef struct {+ range_t prefix; /* read-only non-owned prefix buffer */+ buffer_t buffer;+ size_t filled;+} inBuff_t;++typedef struct {+ BYTE* buffer; /* The round input buffer. All jobs get references+ * to pieces of the buffer. ZSTDMT_tryGetInputRange()+ * handles handing out job input buffers, and makes+ * sure it doesn't overlap with any pieces still in use.+ */+ size_t capacity; /* The capacity of buffer. */+ size_t pos; /* The position of the current inBuff in the round+ * buffer. Updated past the end if the inBuff once+ * the inBuff is sent to the worker thread.+ * pos <= capacity.+ */+} roundBuff_t;++static const roundBuff_t kNullRoundBuff = {NULL, 0, 0};++#define RSYNC_LENGTH 32++typedef struct {+ U64 hash;+ U64 hitMask;+ U64 primePower;+} rsyncState_t;++struct ZSTDMT_CCtx_s {+ POOL_ctx* factory;+ ZSTDMT_jobDescription* jobs;+ ZSTDMT_bufferPool* bufPool;+ ZSTDMT_CCtxPool* cctxPool;+ ZSTDMT_seqPool* seqPool;+ ZSTD_CCtx_params params;+ size_t targetSectionSize;+ size_t targetPrefixSize;+ int jobReady; /* 1 => one job is already prepared, but pool has shortage of workers. Don't create a new job. */+ inBuff_t inBuff;+ roundBuff_t roundBuff;+ serialState_t serial;+ rsyncState_t rsync;+ unsigned singleBlockingThread;+ unsigned jobIDMask;+ unsigned doneJobID;+ unsigned nextJobID;+ unsigned frameEnded;+ unsigned allJobsCompleted;+ unsigned long long frameContentSize;+ unsigned long long consumed;+ unsigned long long produced;+ ZSTD_customMem cMem;+ ZSTD_CDict* cdictLocal;+ const ZSTD_CDict* cdict;+};++static void ZSTDMT_freeJobsTable(ZSTDMT_jobDescription* jobTable, U32 nbJobs, ZSTD_customMem cMem)+{+ U32 jobNb;+ if (jobTable == NULL) return;+ for (jobNb=0; jobNb<nbJobs; jobNb++) {+ ZSTD_pthread_mutex_destroy(&jobTable[jobNb].job_mutex);+ ZSTD_pthread_cond_destroy(&jobTable[jobNb].job_cond);+ }+ ZSTD_free(jobTable, cMem);+}++/* ZSTDMT_allocJobsTable()+ * allocate and init a job table.+ * update *nbJobsPtr to next power of 2 value, as size of table */+static ZSTDMT_jobDescription* ZSTDMT_createJobsTable(U32* nbJobsPtr, ZSTD_customMem cMem)+{+ U32 const nbJobsLog2 = ZSTD_highbit32(*nbJobsPtr) + 1;+ U32 const nbJobs = 1 << nbJobsLog2;+ U32 jobNb;+ ZSTDMT_jobDescription* const jobTable = (ZSTDMT_jobDescription*)+ ZSTD_calloc(nbJobs * sizeof(ZSTDMT_jobDescription), cMem);+ int initError = 0;+ if (jobTable==NULL) return NULL;+ *nbJobsPtr = nbJobs;+ for (jobNb=0; jobNb<nbJobs; jobNb++) {+ initError |= ZSTD_pthread_mutex_init(&jobTable[jobNb].job_mutex, NULL);+ initError |= ZSTD_pthread_cond_init(&jobTable[jobNb].job_cond, NULL);+ }+ if (initError != 0) {+ ZSTDMT_freeJobsTable(jobTable, nbJobs, cMem);+ return NULL;+ }+ return jobTable;+}++static size_t ZSTDMT_expandJobsTable (ZSTDMT_CCtx* mtctx, U32 nbWorkers) {+ U32 nbJobs = nbWorkers + 2;+ if (nbJobs > mtctx->jobIDMask+1) { /* need more job capacity */+ ZSTDMT_freeJobsTable(mtctx->jobs, mtctx->jobIDMask+1, mtctx->cMem);+ mtctx->jobIDMask = 0;+ mtctx->jobs = ZSTDMT_createJobsTable(&nbJobs, mtctx->cMem);+ if (mtctx->jobs==NULL) return ERROR(memory_allocation);+ assert((nbJobs != 0) && ((nbJobs & (nbJobs - 1)) == 0)); /* ensure nbJobs is a power of 2 */+ mtctx->jobIDMask = nbJobs - 1;+ }+ return 0;+}+++/* ZSTDMT_CCtxParam_setNbWorkers():+ * Internal use only */+size_t ZSTDMT_CCtxParam_setNbWorkers(ZSTD_CCtx_params* params, unsigned nbWorkers)+{+ return ZSTD_CCtxParams_setParameter(params, ZSTD_c_nbWorkers, (int)nbWorkers);+}++MEM_STATIC ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced_internal(unsigned nbWorkers, ZSTD_customMem cMem)+{+ ZSTDMT_CCtx* mtctx;+ U32 nbJobs = nbWorkers + 2;+ int initError;+ DEBUGLOG(3, "ZSTDMT_createCCtx_advanced (nbWorkers = %u)", nbWorkers);++ if (nbWorkers < 1) return NULL;+ nbWorkers = MIN(nbWorkers , ZSTDMT_NBWORKERS_MAX);+ if ((cMem.customAlloc!=NULL) ^ (cMem.customFree!=NULL))+ /* invalid custom allocator */+ return NULL;++ mtctx = (ZSTDMT_CCtx*) ZSTD_calloc(sizeof(ZSTDMT_CCtx), cMem);+ if (!mtctx) return NULL;+ ZSTDMT_CCtxParam_setNbWorkers(&mtctx->params, nbWorkers);+ mtctx->cMem = cMem;+ mtctx->allJobsCompleted = 1;+ mtctx->factory = POOL_create_advanced(nbWorkers, 0, cMem);+ mtctx->jobs = ZSTDMT_createJobsTable(&nbJobs, cMem);+ assert(nbJobs > 0); assert((nbJobs & (nbJobs - 1)) == 0); /* ensure nbJobs is a power of 2 */+ mtctx->jobIDMask = nbJobs - 1;+ mtctx->bufPool = ZSTDMT_createBufferPool(nbWorkers, cMem);+ mtctx->cctxPool = ZSTDMT_createCCtxPool(nbWorkers, cMem);+ mtctx->seqPool = ZSTDMT_createSeqPool(nbWorkers, cMem);+ initError = ZSTDMT_serialState_init(&mtctx->serial);+ mtctx->roundBuff = kNullRoundBuff;+ if (!mtctx->factory | !mtctx->jobs | !mtctx->bufPool | !mtctx->cctxPool | !mtctx->seqPool | initError) {+ ZSTDMT_freeCCtx(mtctx);+ return NULL;+ }+ DEBUGLOG(3, "mt_cctx created, for %u threads", nbWorkers);+ return mtctx;+}++ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers, ZSTD_customMem cMem)+{+#ifdef ZSTD_MULTITHREAD+ return ZSTDMT_createCCtx_advanced_internal(nbWorkers, cMem);+#else+ (void)nbWorkers;+ (void)cMem;+ return NULL;+#endif+}++ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers)+{+ return ZSTDMT_createCCtx_advanced(nbWorkers, ZSTD_defaultCMem);+}+++/* ZSTDMT_releaseAllJobResources() :+ * note : ensure all workers are killed first ! */+static void ZSTDMT_releaseAllJobResources(ZSTDMT_CCtx* mtctx)+{+ unsigned jobID;+ DEBUGLOG(3, "ZSTDMT_releaseAllJobResources");+ for (jobID=0; jobID <= mtctx->jobIDMask; jobID++) {+ DEBUGLOG(4, "job%02u: release dst address %08X", jobID, (U32)(size_t)mtctx->jobs[jobID].dstBuff.start);+ ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[jobID].dstBuff);+ mtctx->jobs[jobID].dstBuff = g_nullBuffer;+ mtctx->jobs[jobID].cSize = 0;+ }+ memset(mtctx->jobs, 0, (mtctx->jobIDMask+1)*sizeof(ZSTDMT_jobDescription));+ mtctx->inBuff.buffer = g_nullBuffer;+ mtctx->inBuff.filled = 0;+ mtctx->allJobsCompleted = 1;+}++static void ZSTDMT_waitForAllJobsCompleted(ZSTDMT_CCtx* mtctx)+{+ DEBUGLOG(4, "ZSTDMT_waitForAllJobsCompleted");+ while (mtctx->doneJobID < mtctx->nextJobID) {+ unsigned const jobID = mtctx->doneJobID & mtctx->jobIDMask;+ ZSTD_PTHREAD_MUTEX_LOCK(&mtctx->jobs[jobID].job_mutex);+ while (mtctx->jobs[jobID].consumed < mtctx->jobs[jobID].src.size) {+ DEBUGLOG(4, "waiting for jobCompleted signal from job %u", mtctx->doneJobID); /* we want to block when waiting for data to flush */+ ZSTD_pthread_cond_wait(&mtctx->jobs[jobID].job_cond, &mtctx->jobs[jobID].job_mutex);+ }+ ZSTD_pthread_mutex_unlock(&mtctx->jobs[jobID].job_mutex);+ mtctx->doneJobID++;+ }+}++size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx)+{+ if (mtctx==NULL) return 0; /* compatible with free on NULL */+ POOL_free(mtctx->factory); /* stop and free worker threads */+ ZSTDMT_releaseAllJobResources(mtctx); /* release job resources into pools first */+ ZSTDMT_freeJobsTable(mtctx->jobs, mtctx->jobIDMask+1, mtctx->cMem);+ ZSTDMT_freeBufferPool(mtctx->bufPool);+ ZSTDMT_freeCCtxPool(mtctx->cctxPool);+ ZSTDMT_freeSeqPool(mtctx->seqPool);+ ZSTDMT_serialState_free(&mtctx->serial);+ ZSTD_freeCDict(mtctx->cdictLocal);+ if (mtctx->roundBuff.buffer)+ ZSTD_free(mtctx->roundBuff.buffer, mtctx->cMem);+ ZSTD_free(mtctx, mtctx->cMem);+ return 0;+}++size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx)+{+ if (mtctx == NULL) return 0; /* supports sizeof NULL */+ return sizeof(*mtctx)+ + POOL_sizeof(mtctx->factory)+ + ZSTDMT_sizeof_bufferPool(mtctx->bufPool)+ + (mtctx->jobIDMask+1) * sizeof(ZSTDMT_jobDescription)+ + ZSTDMT_sizeof_CCtxPool(mtctx->cctxPool)+ + ZSTDMT_sizeof_seqPool(mtctx->seqPool)+ + ZSTD_sizeof_CDict(mtctx->cdictLocal)+ + mtctx->roundBuff.capacity;+}++/* Internal only */+size_t+ZSTDMT_CCtxParam_setMTCtxParameter(ZSTD_CCtx_params* params,+ ZSTDMT_parameter parameter,+ int value)+{+ DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter");+ switch(parameter)+ {+ case ZSTDMT_p_jobSize :+ DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter : set jobSize to %i", value);+ return ZSTD_CCtxParams_setParameter(params, ZSTD_c_jobSize, value);+ case ZSTDMT_p_overlapLog :+ DEBUGLOG(4, "ZSTDMT_p_overlapLog : %i", value);+ return ZSTD_CCtxParams_setParameter(params, ZSTD_c_overlapLog, value);+ case ZSTDMT_p_rsyncable :+ DEBUGLOG(4, "ZSTD_p_rsyncable : %i", value);+ return ZSTD_CCtxParams_setParameter(params, ZSTD_c_rsyncable, value);+ default :+ return ERROR(parameter_unsupported);+ }+}++size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value)+{+ DEBUGLOG(4, "ZSTDMT_setMTCtxParameter");+ return ZSTDMT_CCtxParam_setMTCtxParameter(&mtctx->params, parameter, value);+}++size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value)+{+ switch (parameter) {+ case ZSTDMT_p_jobSize:+ return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_jobSize, value);+ case ZSTDMT_p_overlapLog:+ return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_overlapLog, value);+ case ZSTDMT_p_rsyncable:+ return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_rsyncable, value);+ default:+ return ERROR(parameter_unsupported);+ }+}++/* Sets parameters relevant to the compression job,+ * initializing others to default values. */+static ZSTD_CCtx_params ZSTDMT_initJobCCtxParams(ZSTD_CCtx_params const params)+{+ ZSTD_CCtx_params jobParams = params;+ /* Clear parameters related to multithreading */+ jobParams.forceWindow = 0;+ jobParams.nbWorkers = 0;+ jobParams.jobSize = 0;+ jobParams.overlapLog = 0;+ jobParams.rsyncable = 0;+ memset(&jobParams.ldmParams, 0, sizeof(ldmParams_t));+ memset(&jobParams.customMem, 0, sizeof(ZSTD_customMem));+ return jobParams;+}+++/* ZSTDMT_resize() :+ * @return : error code if fails, 0 on success */+static size_t ZSTDMT_resize(ZSTDMT_CCtx* mtctx, unsigned nbWorkers)+{+ if (POOL_resize(mtctx->factory, nbWorkers)) return ERROR(memory_allocation);+ FORWARD_IF_ERROR( ZSTDMT_expandJobsTable(mtctx, nbWorkers) );+ mtctx->bufPool = ZSTDMT_expandBufferPool(mtctx->bufPool, nbWorkers);+ if (mtctx->bufPool == NULL) return ERROR(memory_allocation);+ mtctx->cctxPool = ZSTDMT_expandCCtxPool(mtctx->cctxPool, nbWorkers);+ if (mtctx->cctxPool == NULL) return ERROR(memory_allocation);+ mtctx->seqPool = ZSTDMT_expandSeqPool(mtctx->seqPool, nbWorkers);+ if (mtctx->seqPool == NULL) return ERROR(memory_allocation);+ ZSTDMT_CCtxParam_setNbWorkers(&mtctx->params, nbWorkers);+ return 0;+}+++/*! ZSTDMT_updateCParams_whileCompressing() :+ * Updates a selected set of compression parameters, remaining compatible with currently active frame.+ * New parameters will be applied to next compression job. */+void ZSTDMT_updateCParams_whileCompressing(ZSTDMT_CCtx* mtctx, const ZSTD_CCtx_params* cctxParams)+{+ U32 const saved_wlog = mtctx->params.cParams.windowLog; /* Do not modify windowLog while compressing */+ int const compressionLevel = cctxParams->compressionLevel;+ DEBUGLOG(5, "ZSTDMT_updateCParams_whileCompressing (level:%i)",+ compressionLevel);+ mtctx->params.compressionLevel = compressionLevel;+ { ZSTD_compressionParameters cParams = ZSTD_getCParamsFromCCtxParams(cctxParams, 0, 0);+ cParams.windowLog = saved_wlog;+ mtctx->params.cParams = cParams;+ }+}++/* ZSTDMT_getFrameProgression():+ * tells how much data has been consumed (input) and produced (output) for current frame.+ * able to count progression inside worker threads.+ * Note : mutex will be acquired during statistics collection inside workers. */+ZSTD_frameProgression ZSTDMT_getFrameProgression(ZSTDMT_CCtx* mtctx)+{+ ZSTD_frameProgression fps;+ DEBUGLOG(5, "ZSTDMT_getFrameProgression");+ fps.ingested = mtctx->consumed + mtctx->inBuff.filled;+ fps.consumed = mtctx->consumed;+ fps.produced = fps.flushed = mtctx->produced;+ fps.currentJobID = mtctx->nextJobID;+ fps.nbActiveWorkers = 0;+ { unsigned jobNb;+ unsigned lastJobNb = mtctx->nextJobID + mtctx->jobReady; assert(mtctx->jobReady <= 1);+ DEBUGLOG(6, "ZSTDMT_getFrameProgression: jobs: from %u to <%u (jobReady:%u)",+ mtctx->doneJobID, lastJobNb, mtctx->jobReady)+ for (jobNb = mtctx->doneJobID ; jobNb < lastJobNb ; jobNb++) {+ unsigned const wJobID = jobNb & mtctx->jobIDMask;+ ZSTDMT_jobDescription* jobPtr = &mtctx->jobs[wJobID];+ ZSTD_pthread_mutex_lock(&jobPtr->job_mutex);+ { size_t const cResult = jobPtr->cSize;+ size_t const produced = ZSTD_isError(cResult) ? 0 : cResult;+ size_t const flushed = ZSTD_isError(cResult) ? 0 : jobPtr->dstFlushed;+ assert(flushed <= produced);+ fps.ingested += jobPtr->src.size;+ fps.consumed += jobPtr->consumed;+ fps.produced += produced;+ fps.flushed += flushed;+ fps.nbActiveWorkers += (jobPtr->consumed < jobPtr->src.size);+ }+ ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex);+ }+ }+ return fps;+}+++size_t ZSTDMT_toFlushNow(ZSTDMT_CCtx* mtctx)+{+ size_t toFlush;+ unsigned const jobID = mtctx->doneJobID;+ assert(jobID <= mtctx->nextJobID);+ if (jobID == mtctx->nextJobID) return 0; /* no active job => nothing to flush */++ /* look into oldest non-fully-flushed job */+ { unsigned const wJobID = jobID & mtctx->jobIDMask;+ ZSTDMT_jobDescription* const jobPtr = &mtctx->jobs[wJobID];+ ZSTD_pthread_mutex_lock(&jobPtr->job_mutex);+ { size_t const cResult = jobPtr->cSize;+ size_t const produced = ZSTD_isError(cResult) ? 0 : cResult;+ size_t const flushed = ZSTD_isError(cResult) ? 0 : jobPtr->dstFlushed;+ assert(flushed <= produced);+ toFlush = produced - flushed;+ if (toFlush==0 && (jobPtr->consumed >= jobPtr->src.size)) {+ /* doneJobID is not-fully-flushed, but toFlush==0 : doneJobID should be compressing some more data */+ assert(jobPtr->consumed < jobPtr->src.size);+ }+ }+ ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex);+ }++ return toFlush;+}+++/* ------------------------------------------ */+/* ===== Multi-threaded compression ===== */+/* ------------------------------------------ */++static unsigned ZSTDMT_computeTargetJobLog(ZSTD_CCtx_params const params)+{+ if (params.ldmParams.enableLdm)+ /* In Long Range Mode, the windowLog is typically oversized.+ * In which case, it's preferable to determine the jobSize+ * based on chainLog instead. */+ return MAX(21, params.cParams.chainLog + 4);+ return MAX(20, params.cParams.windowLog + 2);+}++static int ZSTDMT_overlapLog_default(ZSTD_strategy strat)+{+ switch(strat)+ {+ case ZSTD_btultra2:+ return 9;+ case ZSTD_btultra:+ case ZSTD_btopt:+ return 8;+ case ZSTD_btlazy2:+ case ZSTD_lazy2:+ return 7;+ case ZSTD_lazy:+ case ZSTD_greedy:+ case ZSTD_dfast:+ case ZSTD_fast:+ default:;+ }+ return 6;+}++static int ZSTDMT_overlapLog(int ovlog, ZSTD_strategy strat)+{+ assert(0 <= ovlog && ovlog <= 9);+ if (ovlog == 0) return ZSTDMT_overlapLog_default(strat);+ return ovlog;+}++static size_t ZSTDMT_computeOverlapSize(ZSTD_CCtx_params const params)+{+ int const overlapRLog = 9 - ZSTDMT_overlapLog(params.overlapLog, params.cParams.strategy);+ int ovLog = (overlapRLog >= 8) ? 0 : (params.cParams.windowLog - overlapRLog);+ assert(0 <= overlapRLog && overlapRLog <= 8);+ if (params.ldmParams.enableLdm) {+ /* In Long Range Mode, the windowLog is typically oversized.+ * In which case, it's preferable to determine the jobSize+ * based on chainLog instead.+ * Then, ovLog becomes a fraction of the jobSize, rather than windowSize */+ ovLog = MIN(params.cParams.windowLog, ZSTDMT_computeTargetJobLog(params) - 2)+ - overlapRLog;+ }+ assert(0 <= ovLog && ovLog <= 30);+ DEBUGLOG(4, "overlapLog : %i", params.overlapLog);+ DEBUGLOG(4, "overlap size : %i", 1 << ovLog);+ return (ovLog==0) ? 0 : (size_t)1 << ovLog;+}++static unsigned+ZSTDMT_computeNbJobs(ZSTD_CCtx_params params, size_t srcSize, unsigned nbWorkers)+{+ assert(nbWorkers>0);+ { size_t const jobSizeTarget = (size_t)1 << ZSTDMT_computeTargetJobLog(params);+ size_t const jobMaxSize = jobSizeTarget << 2;+ size_t const passSizeMax = jobMaxSize * nbWorkers;+ unsigned const multiplier = (unsigned)(srcSize / passSizeMax) + 1;+ unsigned const nbJobsLarge = multiplier * nbWorkers;+ unsigned const nbJobsMax = (unsigned)(srcSize / jobSizeTarget) + 1;+ unsigned const nbJobsSmall = MIN(nbJobsMax, nbWorkers);+ return (multiplier>1) ? nbJobsLarge : nbJobsSmall;+} }++/* ZSTDMT_compress_advanced_internal() :+ * This is a blocking function : it will only give back control to caller after finishing its compression job.+ */+static size_t ZSTDMT_compress_advanced_internal(+ ZSTDMT_CCtx* mtctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params)+{+ ZSTD_CCtx_params const jobParams = ZSTDMT_initJobCCtxParams(params);+ size_t const overlapSize = ZSTDMT_computeOverlapSize(params);+ unsigned const nbJobs = ZSTDMT_computeNbJobs(params, srcSize, params.nbWorkers);+ size_t const proposedJobSize = (srcSize + (nbJobs-1)) / nbJobs;+ size_t const avgJobSize = (((proposedJobSize-1) & 0x1FFFF) < 0x7FFF) ? proposedJobSize + 0xFFFF : proposedJobSize; /* avoid too small last block */+ const char* const srcStart = (const char*)src;+ size_t remainingSrcSize = srcSize;+ unsigned const compressWithinDst = (dstCapacity >= ZSTD_compressBound(srcSize)) ? nbJobs : (unsigned)(dstCapacity / ZSTD_compressBound(avgJobSize)); /* presumes avgJobSize >= 256 KB, which should be the case */+ size_t frameStartPos = 0, dstBufferPos = 0;+ assert(jobParams.nbWorkers == 0);+ assert(mtctx->cctxPool->totalCCtx == params.nbWorkers);++ params.jobSize = (U32)avgJobSize;+ DEBUGLOG(4, "ZSTDMT_compress_advanced_internal: nbJobs=%2u (rawSize=%u bytes; fixedSize=%u) ",+ nbJobs, (U32)proposedJobSize, (U32)avgJobSize);++ if ((nbJobs==1) | (params.nbWorkers<=1)) { /* fallback to single-thread mode : this is a blocking invocation anyway */+ ZSTD_CCtx* const cctx = mtctx->cctxPool->cctx[0];+ DEBUGLOG(4, "ZSTDMT_compress_advanced_internal: fallback to single-thread mode");+ if (cdict) return ZSTD_compress_usingCDict_advanced(cctx, dst, dstCapacity, src, srcSize, cdict, jobParams.fParams);+ return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, NULL, 0, jobParams);+ }++ assert(avgJobSize >= 256 KB); /* condition for ZSTD_compressBound(A) + ZSTD_compressBound(B) <= ZSTD_compressBound(A+B), required to compress directly into Dst (no additional buffer) */+ ZSTDMT_setBufferSize(mtctx->bufPool, ZSTD_compressBound(avgJobSize) );+ if (ZSTDMT_serialState_reset(&mtctx->serial, mtctx->seqPool, params, avgJobSize))+ return ERROR(memory_allocation);++ FORWARD_IF_ERROR( ZSTDMT_expandJobsTable(mtctx, nbJobs) ); /* only expands if necessary */++ { unsigned u;+ for (u=0; u<nbJobs; u++) {+ size_t const jobSize = MIN(remainingSrcSize, avgJobSize);+ size_t const dstBufferCapacity = ZSTD_compressBound(jobSize);+ buffer_t const dstAsBuffer = { (char*)dst + dstBufferPos, dstBufferCapacity };+ buffer_t const dstBuffer = u < compressWithinDst ? dstAsBuffer : g_nullBuffer;+ size_t dictSize = u ? overlapSize : 0;++ mtctx->jobs[u].prefix.start = srcStart + frameStartPos - dictSize;+ mtctx->jobs[u].prefix.size = dictSize;+ mtctx->jobs[u].src.start = srcStart + frameStartPos;+ mtctx->jobs[u].src.size = jobSize; assert(jobSize > 0); /* avoid job.src.size == 0 */+ mtctx->jobs[u].consumed = 0;+ mtctx->jobs[u].cSize = 0;+ mtctx->jobs[u].cdict = (u==0) ? cdict : NULL;+ mtctx->jobs[u].fullFrameSize = srcSize;+ mtctx->jobs[u].params = jobParams;+ /* do not calculate checksum within sections, but write it in header for first section */+ mtctx->jobs[u].dstBuff = dstBuffer;+ mtctx->jobs[u].cctxPool = mtctx->cctxPool;+ mtctx->jobs[u].bufPool = mtctx->bufPool;+ mtctx->jobs[u].seqPool = mtctx->seqPool;+ mtctx->jobs[u].serial = &mtctx->serial;+ mtctx->jobs[u].jobID = u;+ mtctx->jobs[u].firstJob = (u==0);+ mtctx->jobs[u].lastJob = (u==nbJobs-1);++ DEBUGLOG(5, "ZSTDMT_compress_advanced_internal: posting job %u (%u bytes)", u, (U32)jobSize);+ DEBUG_PRINTHEX(6, mtctx->jobs[u].prefix.start, 12);+ POOL_add(mtctx->factory, ZSTDMT_compressionJob, &mtctx->jobs[u]);++ frameStartPos += jobSize;+ dstBufferPos += dstBufferCapacity;+ remainingSrcSize -= jobSize;+ } }++ /* collect result */+ { size_t error = 0, dstPos = 0;+ unsigned jobID;+ for (jobID=0; jobID<nbJobs; jobID++) {+ DEBUGLOG(5, "waiting for job %u ", jobID);+ ZSTD_PTHREAD_MUTEX_LOCK(&mtctx->jobs[jobID].job_mutex);+ while (mtctx->jobs[jobID].consumed < mtctx->jobs[jobID].src.size) {+ DEBUGLOG(5, "waiting for jobCompleted signal from job %u", jobID);+ ZSTD_pthread_cond_wait(&mtctx->jobs[jobID].job_cond, &mtctx->jobs[jobID].job_mutex);+ }+ ZSTD_pthread_mutex_unlock(&mtctx->jobs[jobID].job_mutex);+ DEBUGLOG(5, "ready to write job %u ", jobID);++ { size_t const cSize = mtctx->jobs[jobID].cSize;+ if (ZSTD_isError(cSize)) error = cSize;+ if ((!error) && (dstPos + cSize > dstCapacity)) error = ERROR(dstSize_tooSmall);+ if (jobID) { /* note : job 0 is written directly at dst, which is correct position */+ if (!error)+ memmove((char*)dst + dstPos, mtctx->jobs[jobID].dstBuff.start, cSize); /* may overlap when job compressed within dst */+ if (jobID >= compressWithinDst) { /* job compressed into its own buffer, which must be released */+ DEBUGLOG(5, "releasing buffer %u>=%u", jobID, compressWithinDst);+ ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[jobID].dstBuff);+ } }+ mtctx->jobs[jobID].dstBuff = g_nullBuffer;+ mtctx->jobs[jobID].cSize = 0;+ dstPos += cSize ;+ }+ } /* for (jobID=0; jobID<nbJobs; jobID++) */++ DEBUGLOG(4, "checksumFlag : %u ", params.fParams.checksumFlag);+ if (params.fParams.checksumFlag) {+ U32 const checksum = (U32)XXH64_digest(&mtctx->serial.xxhState);+ if (dstPos + 4 > dstCapacity) {+ error = ERROR(dstSize_tooSmall);+ } else {+ DEBUGLOG(4, "writing checksum : %08X \n", checksum);+ MEM_writeLE32((char*)dst + dstPos, checksum);+ dstPos += 4;+ } }++ if (!error) DEBUGLOG(4, "compressed size : %u ", (U32)dstPos);+ return error ? error : dstPos;+ }+}++size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict,+ ZSTD_parameters params,+ int overlapLog)+{+ ZSTD_CCtx_params cctxParams = mtctx->params;+ cctxParams.cParams = params.cParams;+ cctxParams.fParams = params.fParams;+ assert(ZSTD_OVERLAPLOG_MIN <= overlapLog && overlapLog <= ZSTD_OVERLAPLOG_MAX);+ cctxParams.overlapLog = overlapLog;+ return ZSTDMT_compress_advanced_internal(mtctx,+ dst, dstCapacity,+ src, srcSize,+ cdict, cctxParams);+}+++size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ int compressionLevel)+{+ ZSTD_parameters params = ZSTD_getParams(compressionLevel, srcSize, 0);+ int const overlapLog = ZSTDMT_overlapLog_default(params.cParams.strategy);+ params.fParams.contentSizeFlag = 1;+ return ZSTDMT_compress_advanced(mtctx, dst, dstCapacity, src, srcSize, NULL, params, overlapLog);+}+++/* ====================================== */+/* ======= Streaming API ======= */+/* ====================================== */++size_t ZSTDMT_initCStream_internal(+ ZSTDMT_CCtx* mtctx,+ const void* dict, size_t dictSize, ZSTD_dictContentType_e dictContentType,+ const ZSTD_CDict* cdict, ZSTD_CCtx_params params,+ unsigned long long pledgedSrcSize)+{+ DEBUGLOG(4, "ZSTDMT_initCStream_internal (pledgedSrcSize=%u, nbWorkers=%u, cctxPool=%u)",+ (U32)pledgedSrcSize, params.nbWorkers, mtctx->cctxPool->totalCCtx);++ /* params supposed partially fully validated at this point */+ assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));+ assert(!((dict) && (cdict))); /* either dict or cdict, not both */++ /* init */+ if (params.nbWorkers != mtctx->params.nbWorkers)+ FORWARD_IF_ERROR( ZSTDMT_resize(mtctx, params.nbWorkers) );++ if (params.jobSize != 0 && params.jobSize < ZSTDMT_JOBSIZE_MIN) params.jobSize = ZSTDMT_JOBSIZE_MIN;+ if (params.jobSize > (size_t)ZSTDMT_JOBSIZE_MAX) params.jobSize = ZSTDMT_JOBSIZE_MAX;++ mtctx->singleBlockingThread = (pledgedSrcSize <= ZSTDMT_JOBSIZE_MIN); /* do not trigger multi-threading when srcSize is too small */+ if (mtctx->singleBlockingThread) {+ ZSTD_CCtx_params const singleThreadParams = ZSTDMT_initJobCCtxParams(params);+ DEBUGLOG(5, "ZSTDMT_initCStream_internal: switch to single blocking thread mode");+ assert(singleThreadParams.nbWorkers == 0);+ return ZSTD_initCStream_internal(mtctx->cctxPool->cctx[0],+ dict, dictSize, cdict,+ singleThreadParams, pledgedSrcSize);+ }++ DEBUGLOG(4, "ZSTDMT_initCStream_internal: %u workers", params.nbWorkers);++ if (mtctx->allJobsCompleted == 0) { /* previous compression not correctly finished */+ ZSTDMT_waitForAllJobsCompleted(mtctx);+ ZSTDMT_releaseAllJobResources(mtctx);+ mtctx->allJobsCompleted = 1;+ }++ mtctx->params = params;+ mtctx->frameContentSize = pledgedSrcSize;+ if (dict) {+ ZSTD_freeCDict(mtctx->cdictLocal);+ mtctx->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,+ ZSTD_dlm_byCopy, dictContentType, /* note : a loadPrefix becomes an internal CDict */+ params.cParams, mtctx->cMem);+ mtctx->cdict = mtctx->cdictLocal;+ if (mtctx->cdictLocal == NULL) return ERROR(memory_allocation);+ } else {+ ZSTD_freeCDict(mtctx->cdictLocal);+ mtctx->cdictLocal = NULL;+ mtctx->cdict = cdict;+ }++ mtctx->targetPrefixSize = ZSTDMT_computeOverlapSize(params);+ DEBUGLOG(4, "overlapLog=%i => %u KB", params.overlapLog, (U32)(mtctx->targetPrefixSize>>10));+ mtctx->targetSectionSize = params.jobSize;+ if (mtctx->targetSectionSize == 0) {+ mtctx->targetSectionSize = 1ULL << ZSTDMT_computeTargetJobLog(params);+ }+ if (params.rsyncable) {+ /* Aim for the targetsectionSize as the average job size. */+ U32 const jobSizeMB = (U32)(mtctx->targetSectionSize >> 20);+ U32 const rsyncBits = ZSTD_highbit32(jobSizeMB) + 20;+ assert(jobSizeMB >= 1);+ DEBUGLOG(4, "rsyncLog = %u", rsyncBits);+ mtctx->rsync.hash = 0;+ mtctx->rsync.hitMask = (1ULL << rsyncBits) - 1;+ mtctx->rsync.primePower = ZSTD_rollingHash_primePower(RSYNC_LENGTH);+ }+ if (mtctx->targetSectionSize < mtctx->targetPrefixSize) mtctx->targetSectionSize = mtctx->targetPrefixSize; /* job size must be >= overlap size */+ DEBUGLOG(4, "Job Size : %u KB (note : set to %u)", (U32)(mtctx->targetSectionSize>>10), (U32)params.jobSize);+ DEBUGLOG(4, "inBuff Size : %u KB", (U32)(mtctx->targetSectionSize>>10));+ ZSTDMT_setBufferSize(mtctx->bufPool, ZSTD_compressBound(mtctx->targetSectionSize));+ {+ /* If ldm is enabled we need windowSize space. */+ size_t const windowSize = mtctx->params.ldmParams.enableLdm ? (1U << mtctx->params.cParams.windowLog) : 0;+ /* Two buffers of slack, plus extra space for the overlap+ * This is the minimum slack that LDM works with. One extra because+ * flush might waste up to targetSectionSize-1 bytes. Another extra+ * for the overlap (if > 0), then one to fill which doesn't overlap+ * with the LDM window.+ */+ size_t const nbSlackBuffers = 2 + (mtctx->targetPrefixSize > 0);+ size_t const slackSize = mtctx->targetSectionSize * nbSlackBuffers;+ /* Compute the total size, and always have enough slack */+ size_t const nbWorkers = MAX(mtctx->params.nbWorkers, 1);+ size_t const sectionsSize = mtctx->targetSectionSize * nbWorkers;+ size_t const capacity = MAX(windowSize, sectionsSize) + slackSize;+ if (mtctx->roundBuff.capacity < capacity) {+ if (mtctx->roundBuff.buffer)+ ZSTD_free(mtctx->roundBuff.buffer, mtctx->cMem);+ mtctx->roundBuff.buffer = (BYTE*)ZSTD_malloc(capacity, mtctx->cMem);+ if (mtctx->roundBuff.buffer == NULL) {+ mtctx->roundBuff.capacity = 0;+ return ERROR(memory_allocation);+ }+ mtctx->roundBuff.capacity = capacity;+ }+ }+ DEBUGLOG(4, "roundBuff capacity : %u KB", (U32)(mtctx->roundBuff.capacity>>10));+ mtctx->roundBuff.pos = 0;+ mtctx->inBuff.buffer = g_nullBuffer;+ mtctx->inBuff.filled = 0;+ mtctx->inBuff.prefix = kNullRange;+ mtctx->doneJobID = 0;+ mtctx->nextJobID = 0;+ mtctx->frameEnded = 0;+ mtctx->allJobsCompleted = 0;+ mtctx->consumed = 0;+ mtctx->produced = 0;+ if (ZSTDMT_serialState_reset(&mtctx->serial, mtctx->seqPool, params, mtctx->targetSectionSize))+ return ERROR(memory_allocation);+ return 0;+}++size_t ZSTDMT_initCStream_advanced(ZSTDMT_CCtx* mtctx,+ const void* dict, size_t dictSize,+ ZSTD_parameters params,+ unsigned long long pledgedSrcSize)+{+ ZSTD_CCtx_params cctxParams = mtctx->params; /* retrieve sticky params */+ DEBUGLOG(4, "ZSTDMT_initCStream_advanced (pledgedSrcSize=%u)", (U32)pledgedSrcSize);+ cctxParams.cParams = params.cParams;+ cctxParams.fParams = params.fParams;+ return ZSTDMT_initCStream_internal(mtctx, dict, dictSize, ZSTD_dct_auto, NULL,+ cctxParams, pledgedSrcSize);+}++size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,+ const ZSTD_CDict* cdict,+ ZSTD_frameParameters fParams,+ unsigned long long pledgedSrcSize)+{+ ZSTD_CCtx_params cctxParams = mtctx->params;+ if (cdict==NULL) return ERROR(dictionary_wrong); /* method incompatible with NULL cdict */+ cctxParams.cParams = ZSTD_getCParamsFromCDict(cdict);+ cctxParams.fParams = fParams;+ return ZSTDMT_initCStream_internal(mtctx, NULL, 0 /*dictSize*/, ZSTD_dct_auto, cdict,+ cctxParams, pledgedSrcSize);+}+++/* ZSTDMT_resetCStream() :+ * pledgedSrcSize can be zero == unknown (for the time being)+ * prefer using ZSTD_CONTENTSIZE_UNKNOWN,+ * as `0` might mean "empty" in the future */+size_t ZSTDMT_resetCStream(ZSTDMT_CCtx* mtctx, unsigned long long pledgedSrcSize)+{+ if (!pledgedSrcSize) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN;+ return ZSTDMT_initCStream_internal(mtctx, NULL, 0, ZSTD_dct_auto, 0, mtctx->params,+ pledgedSrcSize);+}++size_t ZSTDMT_initCStream(ZSTDMT_CCtx* mtctx, int compressionLevel) {+ ZSTD_parameters const params = ZSTD_getParams(compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN, 0);+ ZSTD_CCtx_params cctxParams = mtctx->params; /* retrieve sticky params */+ DEBUGLOG(4, "ZSTDMT_initCStream (cLevel=%i)", compressionLevel);+ cctxParams.cParams = params.cParams;+ cctxParams.fParams = params.fParams;+ return ZSTDMT_initCStream_internal(mtctx, NULL, 0, ZSTD_dct_auto, NULL, cctxParams, ZSTD_CONTENTSIZE_UNKNOWN);+}+++/* ZSTDMT_writeLastEmptyBlock()+ * Write a single empty block with an end-of-frame to finish a frame.+ * Job must be created from streaming variant.+ * This function is always successful if expected conditions are fulfilled.+ */+static void ZSTDMT_writeLastEmptyBlock(ZSTDMT_jobDescription* job)+{+ assert(job->lastJob == 1);+ assert(job->src.size == 0); /* last job is empty -> will be simplified into a last empty block */+ assert(job->firstJob == 0); /* cannot be first job, as it also needs to create frame header */+ assert(job->dstBuff.start == NULL); /* invoked from streaming variant only (otherwise, dstBuff might be user's output) */+ job->dstBuff = ZSTDMT_getBuffer(job->bufPool);+ if (job->dstBuff.start == NULL) {+ job->cSize = ERROR(memory_allocation);+ return;+ }+ assert(job->dstBuff.capacity >= ZSTD_blockHeaderSize); /* no buffer should ever be that small */+ job->src = kNullRange;+ job->cSize = ZSTD_writeLastEmptyBlock(job->dstBuff.start, job->dstBuff.capacity);+ assert(!ZSTD_isError(job->cSize));+ assert(job->consumed == 0);+}++static size_t ZSTDMT_createCompressionJob(ZSTDMT_CCtx* mtctx, size_t srcSize, ZSTD_EndDirective endOp)+{+ unsigned const jobID = mtctx->nextJobID & mtctx->jobIDMask;+ int const endFrame = (endOp == ZSTD_e_end);++ if (mtctx->nextJobID > mtctx->doneJobID + mtctx->jobIDMask) {+ DEBUGLOG(5, "ZSTDMT_createCompressionJob: will not create new job : table is full");+ assert((mtctx->nextJobID & mtctx->jobIDMask) == (mtctx->doneJobID & mtctx->jobIDMask));+ return 0;+ }++ if (!mtctx->jobReady) {+ BYTE const* src = (BYTE const*)mtctx->inBuff.buffer.start;+ DEBUGLOG(5, "ZSTDMT_createCompressionJob: preparing job %u to compress %u bytes with %u preload ",+ mtctx->nextJobID, (U32)srcSize, (U32)mtctx->inBuff.prefix.size);+ mtctx->jobs[jobID].src.start = src;+ mtctx->jobs[jobID].src.size = srcSize;+ assert(mtctx->inBuff.filled >= srcSize);+ mtctx->jobs[jobID].prefix = mtctx->inBuff.prefix;+ mtctx->jobs[jobID].consumed = 0;+ mtctx->jobs[jobID].cSize = 0;+ mtctx->jobs[jobID].params = mtctx->params;+ mtctx->jobs[jobID].cdict = mtctx->nextJobID==0 ? mtctx->cdict : NULL;+ mtctx->jobs[jobID].fullFrameSize = mtctx->frameContentSize;+ mtctx->jobs[jobID].dstBuff = g_nullBuffer;+ mtctx->jobs[jobID].cctxPool = mtctx->cctxPool;+ mtctx->jobs[jobID].bufPool = mtctx->bufPool;+ mtctx->jobs[jobID].seqPool = mtctx->seqPool;+ mtctx->jobs[jobID].serial = &mtctx->serial;+ mtctx->jobs[jobID].jobID = mtctx->nextJobID;+ mtctx->jobs[jobID].firstJob = (mtctx->nextJobID==0);+ mtctx->jobs[jobID].lastJob = endFrame;+ mtctx->jobs[jobID].frameChecksumNeeded = mtctx->params.fParams.checksumFlag && endFrame && (mtctx->nextJobID>0);+ mtctx->jobs[jobID].dstFlushed = 0;++ /* Update the round buffer pos and clear the input buffer to be reset */+ mtctx->roundBuff.pos += srcSize;+ mtctx->inBuff.buffer = g_nullBuffer;+ mtctx->inBuff.filled = 0;+ /* Set the prefix */+ if (!endFrame) {+ size_t const newPrefixSize = MIN(srcSize, mtctx->targetPrefixSize);+ mtctx->inBuff.prefix.start = src + srcSize - newPrefixSize;+ mtctx->inBuff.prefix.size = newPrefixSize;+ } else { /* endFrame==1 => no need for another input buffer */+ mtctx->inBuff.prefix = kNullRange;+ mtctx->frameEnded = endFrame;+ if (mtctx->nextJobID == 0) {+ /* single job exception : checksum is already calculated directly within worker thread */+ mtctx->params.fParams.checksumFlag = 0;+ } }++ if ( (srcSize == 0)+ && (mtctx->nextJobID>0)/*single job must also write frame header*/ ) {+ DEBUGLOG(5, "ZSTDMT_createCompressionJob: creating a last empty block to end frame");+ assert(endOp == ZSTD_e_end); /* only possible case : need to end the frame with an empty last block */+ ZSTDMT_writeLastEmptyBlock(mtctx->jobs + jobID);+ mtctx->nextJobID++;+ return 0;+ }+ }++ DEBUGLOG(5, "ZSTDMT_createCompressionJob: posting job %u : %u bytes (end:%u, jobNb == %u (mod:%u))",+ mtctx->nextJobID,+ (U32)mtctx->jobs[jobID].src.size,+ mtctx->jobs[jobID].lastJob,+ mtctx->nextJobID,+ jobID);+ if (POOL_tryAdd(mtctx->factory, ZSTDMT_compressionJob, &mtctx->jobs[jobID])) {+ mtctx->nextJobID++;+ mtctx->jobReady = 0;+ } else {+ DEBUGLOG(5, "ZSTDMT_createCompressionJob: no worker available for job %u", mtctx->nextJobID);+ mtctx->jobReady = 1;+ }+ return 0;+}+++/*! ZSTDMT_flushProduced() :+ * flush whatever data has been produced but not yet flushed in current job.+ * move to next job if current one is fully flushed.+ * `output` : `pos` will be updated with amount of data flushed .+ * `blockToFlush` : if >0, the function will block and wait if there is no data available to flush .+ * @return : amount of data remaining within internal buffer, 0 if no more, 1 if unknown but > 0, or an error code */+static size_t ZSTDMT_flushProduced(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, unsigned blockToFlush, ZSTD_EndDirective end)+{+ unsigned const wJobID = mtctx->doneJobID & mtctx->jobIDMask;+ DEBUGLOG(5, "ZSTDMT_flushProduced (blocking:%u , job %u <= %u)",+ blockToFlush, mtctx->doneJobID, mtctx->nextJobID);+ assert(output->size >= output->pos);++ ZSTD_PTHREAD_MUTEX_LOCK(&mtctx->jobs[wJobID].job_mutex);+ if ( blockToFlush+ && (mtctx->doneJobID < mtctx->nextJobID) ) {+ assert(mtctx->jobs[wJobID].dstFlushed <= mtctx->jobs[wJobID].cSize);+ while (mtctx->jobs[wJobID].dstFlushed == mtctx->jobs[wJobID].cSize) { /* nothing to flush */+ if (mtctx->jobs[wJobID].consumed == mtctx->jobs[wJobID].src.size) {+ DEBUGLOG(5, "job %u is completely consumed (%u == %u) => don't wait for cond, there will be none",+ mtctx->doneJobID, (U32)mtctx->jobs[wJobID].consumed, (U32)mtctx->jobs[wJobID].src.size);+ break;+ }+ DEBUGLOG(5, "waiting for something to flush from job %u (currently flushed: %u bytes)",+ mtctx->doneJobID, (U32)mtctx->jobs[wJobID].dstFlushed);+ ZSTD_pthread_cond_wait(&mtctx->jobs[wJobID].job_cond, &mtctx->jobs[wJobID].job_mutex); /* block when nothing to flush but some to come */+ } }++ /* try to flush something */+ { size_t cSize = mtctx->jobs[wJobID].cSize; /* shared */+ size_t const srcConsumed = mtctx->jobs[wJobID].consumed; /* shared */+ size_t const srcSize = mtctx->jobs[wJobID].src.size; /* read-only, could be done after mutex lock, but no-declaration-after-statement */+ ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex);+ if (ZSTD_isError(cSize)) {+ DEBUGLOG(5, "ZSTDMT_flushProduced: job %u : compression error detected : %s",+ mtctx->doneJobID, ZSTD_getErrorName(cSize));+ ZSTDMT_waitForAllJobsCompleted(mtctx);+ ZSTDMT_releaseAllJobResources(mtctx);+ return cSize;+ }+ /* add frame checksum if necessary (can only happen once) */+ assert(srcConsumed <= srcSize);+ if ( (srcConsumed == srcSize) /* job completed -> worker no longer active */+ && mtctx->jobs[wJobID].frameChecksumNeeded ) {+ U32 const checksum = (U32)XXH64_digest(&mtctx->serial.xxhState);+ DEBUGLOG(4, "ZSTDMT_flushProduced: writing checksum : %08X \n", checksum);+ MEM_writeLE32((char*)mtctx->jobs[wJobID].dstBuff.start + mtctx->jobs[wJobID].cSize, checksum);+ cSize += 4;+ mtctx->jobs[wJobID].cSize += 4; /* can write this shared value, as worker is no longer active */+ mtctx->jobs[wJobID].frameChecksumNeeded = 0;+ }++ if (cSize > 0) { /* compression is ongoing or completed */+ size_t const toFlush = MIN(cSize - mtctx->jobs[wJobID].dstFlushed, output->size - output->pos);+ DEBUGLOG(5, "ZSTDMT_flushProduced: Flushing %u bytes from job %u (completion:%u/%u, generated:%u)",+ (U32)toFlush, mtctx->doneJobID, (U32)srcConsumed, (U32)srcSize, (U32)cSize);+ assert(mtctx->doneJobID < mtctx->nextJobID);+ assert(cSize >= mtctx->jobs[wJobID].dstFlushed);+ assert(mtctx->jobs[wJobID].dstBuff.start != NULL);+ memcpy((char*)output->dst + output->pos,+ (const char*)mtctx->jobs[wJobID].dstBuff.start + mtctx->jobs[wJobID].dstFlushed,+ toFlush);+ output->pos += toFlush;+ mtctx->jobs[wJobID].dstFlushed += toFlush; /* can write : this value is only used by mtctx */++ if ( (srcConsumed == srcSize) /* job is completed */+ && (mtctx->jobs[wJobID].dstFlushed == cSize) ) { /* output buffer fully flushed => free this job position */+ DEBUGLOG(5, "Job %u completed (%u bytes), moving to next one",+ mtctx->doneJobID, (U32)mtctx->jobs[wJobID].dstFlushed);+ ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[wJobID].dstBuff);+ DEBUGLOG(5, "dstBuffer released");+ mtctx->jobs[wJobID].dstBuff = g_nullBuffer;+ mtctx->jobs[wJobID].cSize = 0; /* ensure this job slot is considered "not started" in future check */+ mtctx->consumed += srcSize;+ mtctx->produced += cSize;+ mtctx->doneJobID++;+ } }++ /* return value : how many bytes left in buffer ; fake it to 1 when unknown but >0 */+ if (cSize > mtctx->jobs[wJobID].dstFlushed) return (cSize - mtctx->jobs[wJobID].dstFlushed);+ if (srcSize > srcConsumed) return 1; /* current job not completely compressed */+ }+ if (mtctx->doneJobID < mtctx->nextJobID) return 1; /* some more jobs ongoing */+ if (mtctx->jobReady) return 1; /* one job is ready to push, just not yet in the list */+ if (mtctx->inBuff.filled > 0) return 1; /* input is not empty, and still needs to be converted into a job */+ mtctx->allJobsCompleted = mtctx->frameEnded; /* all jobs are entirely flushed => if this one is last one, frame is completed */+ if (end == ZSTD_e_end) return !mtctx->frameEnded; /* for ZSTD_e_end, question becomes : is frame completed ? instead of : are internal buffers fully flushed ? */+ return 0; /* internal buffers fully flushed */+}++/**+ * Returns the range of data used by the earliest job that is not yet complete.+ * If the data of the first job is broken up into two segments, we cover both+ * sections.+ */+static range_t ZSTDMT_getInputDataInUse(ZSTDMT_CCtx* mtctx)+{+ unsigned const firstJobID = mtctx->doneJobID;+ unsigned const lastJobID = mtctx->nextJobID;+ unsigned jobID;++ for (jobID = firstJobID; jobID < lastJobID; ++jobID) {+ unsigned const wJobID = jobID & mtctx->jobIDMask;+ size_t consumed;++ ZSTD_PTHREAD_MUTEX_LOCK(&mtctx->jobs[wJobID].job_mutex);+ consumed = mtctx->jobs[wJobID].consumed;+ ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex);++ if (consumed < mtctx->jobs[wJobID].src.size) {+ range_t range = mtctx->jobs[wJobID].prefix;+ if (range.size == 0) {+ /* Empty prefix */+ range = mtctx->jobs[wJobID].src;+ }+ /* Job source in multiple segments not supported yet */+ assert(range.start <= mtctx->jobs[wJobID].src.start);+ return range;+ }+ }+ return kNullRange;+}++/**+ * Returns non-zero iff buffer and range overlap.+ */+static int ZSTDMT_isOverlapped(buffer_t buffer, range_t range)+{+ BYTE const* const bufferStart = (BYTE const*)buffer.start;+ BYTE const* const bufferEnd = bufferStart + buffer.capacity;+ BYTE const* const rangeStart = (BYTE const*)range.start;+ BYTE const* const rangeEnd = rangeStart + range.size;++ if (rangeStart == NULL || bufferStart == NULL)+ return 0;+ /* Empty ranges cannot overlap */+ if (bufferStart == bufferEnd || rangeStart == rangeEnd)+ return 0;++ return bufferStart < rangeEnd && rangeStart < bufferEnd;+}++static int ZSTDMT_doesOverlapWindow(buffer_t buffer, ZSTD_window_t window)+{+ range_t extDict;+ range_t prefix;++ DEBUGLOG(5, "ZSTDMT_doesOverlapWindow");+ extDict.start = window.dictBase + window.lowLimit;+ extDict.size = window.dictLimit - window.lowLimit;++ prefix.start = window.base + window.dictLimit;+ prefix.size = window.nextSrc - (window.base + window.dictLimit);+ DEBUGLOG(5, "extDict [0x%zx, 0x%zx)",+ (size_t)extDict.start,+ (size_t)extDict.start + extDict.size);+ DEBUGLOG(5, "prefix [0x%zx, 0x%zx)",+ (size_t)prefix.start,+ (size_t)prefix.start + prefix.size);++ return ZSTDMT_isOverlapped(buffer, extDict)+ || ZSTDMT_isOverlapped(buffer, prefix);+}++static void ZSTDMT_waitForLdmComplete(ZSTDMT_CCtx* mtctx, buffer_t buffer)+{+ if (mtctx->params.ldmParams.enableLdm) {+ ZSTD_pthread_mutex_t* mutex = &mtctx->serial.ldmWindowMutex;+ DEBUGLOG(5, "ZSTDMT_waitForLdmComplete");+ DEBUGLOG(5, "source [0x%zx, 0x%zx)",+ (size_t)buffer.start,+ (size_t)buffer.start + buffer.capacity);+ ZSTD_PTHREAD_MUTEX_LOCK(mutex);+ while (ZSTDMT_doesOverlapWindow(buffer, mtctx->serial.ldmWindow)) {+ DEBUGLOG(5, "Waiting for LDM to finish...");+ ZSTD_pthread_cond_wait(&mtctx->serial.ldmWindowCond, mutex);+ }+ DEBUGLOG(6, "Done waiting for LDM to finish");+ ZSTD_pthread_mutex_unlock(mutex);+ }+}++/**+ * Attempts to set the inBuff to the next section to fill.+ * If any part of the new section is still in use we give up.+ * Returns non-zero if the buffer is filled.+ */+static int ZSTDMT_tryGetInputRange(ZSTDMT_CCtx* mtctx)+{+ range_t const inUse = ZSTDMT_getInputDataInUse(mtctx);+ size_t const spaceLeft = mtctx->roundBuff.capacity - mtctx->roundBuff.pos;+ size_t const target = mtctx->targetSectionSize;+ buffer_t buffer;++ DEBUGLOG(5, "ZSTDMT_tryGetInputRange");+ assert(mtctx->inBuff.buffer.start == NULL);+ assert(mtctx->roundBuff.capacity >= target);++ if (spaceLeft < target) {+ /* ZSTD_invalidateRepCodes() doesn't work for extDict variants.+ * Simply copy the prefix to the beginning in that case.+ */+ BYTE* const start = (BYTE*)mtctx->roundBuff.buffer;+ size_t const prefixSize = mtctx->inBuff.prefix.size;++ buffer.start = start;+ buffer.capacity = prefixSize;+ if (ZSTDMT_isOverlapped(buffer, inUse)) {+ DEBUGLOG(5, "Waiting for buffer...");+ return 0;+ }+ ZSTDMT_waitForLdmComplete(mtctx, buffer);+ memmove(start, mtctx->inBuff.prefix.start, prefixSize);+ mtctx->inBuff.prefix.start = start;+ mtctx->roundBuff.pos = prefixSize;+ }+ buffer.start = mtctx->roundBuff.buffer + mtctx->roundBuff.pos;+ buffer.capacity = target;++ if (ZSTDMT_isOverlapped(buffer, inUse)) {+ DEBUGLOG(5, "Waiting for buffer...");+ return 0;+ }+ assert(!ZSTDMT_isOverlapped(buffer, mtctx->inBuff.prefix));++ ZSTDMT_waitForLdmComplete(mtctx, buffer);++ DEBUGLOG(5, "Using prefix range [%zx, %zx)",+ (size_t)mtctx->inBuff.prefix.start,+ (size_t)mtctx->inBuff.prefix.start + mtctx->inBuff.prefix.size);+ DEBUGLOG(5, "Using source range [%zx, %zx)",+ (size_t)buffer.start,+ (size_t)buffer.start + buffer.capacity);+++ mtctx->inBuff.buffer = buffer;+ mtctx->inBuff.filled = 0;+ assert(mtctx->roundBuff.pos + buffer.capacity <= mtctx->roundBuff.capacity);+ return 1;+}++typedef struct {+ size_t toLoad; /* The number of bytes to load from the input. */+ int flush; /* Boolean declaring if we must flush because we found a synchronization point. */+} syncPoint_t;++/**+ * Searches through the input for a synchronization point. If one is found, we+ * will instruct the caller to flush, and return the number of bytes to load.+ * Otherwise, we will load as many bytes as possible and instruct the caller+ * to continue as normal.+ */+static syncPoint_t+findSynchronizationPoint(ZSTDMT_CCtx const* mtctx, ZSTD_inBuffer const input)+{+ BYTE const* const istart = (BYTE const*)input.src + input.pos;+ U64 const primePower = mtctx->rsync.primePower;+ U64 const hitMask = mtctx->rsync.hitMask;++ syncPoint_t syncPoint;+ U64 hash;+ BYTE const* prev;+ size_t pos;++ syncPoint.toLoad = MIN(input.size - input.pos, mtctx->targetSectionSize - mtctx->inBuff.filled);+ syncPoint.flush = 0;+ if (!mtctx->params.rsyncable)+ /* Rsync is disabled. */+ return syncPoint;+ if (mtctx->inBuff.filled + syncPoint.toLoad < RSYNC_LENGTH)+ /* Not enough to compute the hash.+ * We will miss any synchronization points in this RSYNC_LENGTH byte+ * window. However, since it depends only in the internal buffers, if the+ * state is already synchronized, we will remain synchronized.+ * Additionally, the probability that we miss a synchronization point is+ * low: RSYNC_LENGTH / targetSectionSize.+ */+ return syncPoint;+ /* Initialize the loop variables. */+ if (mtctx->inBuff.filled >= RSYNC_LENGTH) {+ /* We have enough bytes buffered to initialize the hash.+ * Start scanning at the beginning of the input.+ */+ pos = 0;+ prev = (BYTE const*)mtctx->inBuff.buffer.start + mtctx->inBuff.filled - RSYNC_LENGTH;+ hash = ZSTD_rollingHash_compute(prev, RSYNC_LENGTH);+ } else {+ /* We don't have enough bytes buffered to initialize the hash, but+ * we know we have at least RSYNC_LENGTH bytes total.+ * Start scanning after the first RSYNC_LENGTH bytes less the bytes+ * already buffered.+ */+ pos = RSYNC_LENGTH - mtctx->inBuff.filled;+ prev = (BYTE const*)mtctx->inBuff.buffer.start - pos;+ hash = ZSTD_rollingHash_compute(mtctx->inBuff.buffer.start, mtctx->inBuff.filled);+ hash = ZSTD_rollingHash_append(hash, istart, pos);+ }+ /* Starting with the hash of the previous RSYNC_LENGTH bytes, roll+ * through the input. If we hit a synchronization point, then cut the+ * job off, and tell the compressor to flush the job. Otherwise, load+ * all the bytes and continue as normal.+ * If we go too long without a synchronization point (targetSectionSize)+ * then a block will be emitted anyways, but this is okay, since if we+ * are already synchronized we will remain synchronized.+ */+ for (; pos < syncPoint.toLoad; ++pos) {+ BYTE const toRemove = pos < RSYNC_LENGTH ? prev[pos] : istart[pos - RSYNC_LENGTH];+ /* if (pos >= RSYNC_LENGTH) assert(ZSTD_rollingHash_compute(istart + pos - RSYNC_LENGTH, RSYNC_LENGTH) == hash); */+ hash = ZSTD_rollingHash_rotate(hash, toRemove, istart[pos], primePower);+ if ((hash & hitMask) == hitMask) {+ syncPoint.toLoad = pos + 1;+ syncPoint.flush = 1;+ break;+ }+ }+ return syncPoint;+}++size_t ZSTDMT_nextInputSizeHint(const ZSTDMT_CCtx* mtctx)+{+ size_t hintInSize = mtctx->targetSectionSize - mtctx->inBuff.filled;+ if (hintInSize==0) hintInSize = mtctx->targetSectionSize;+ return hintInSize;+}++/** ZSTDMT_compressStream_generic() :+ * internal use only - exposed to be invoked from zstd_compress.c+ * assumption : output and input are valid (pos <= size)+ * @return : minimum amount of data remaining to flush, 0 if none */+size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,+ ZSTD_outBuffer* output,+ ZSTD_inBuffer* input,+ ZSTD_EndDirective endOp)+{+ unsigned forwardInputProgress = 0;+ DEBUGLOG(5, "ZSTDMT_compressStream_generic (endOp=%u, srcSize=%u)",+ (U32)endOp, (U32)(input->size - input->pos));+ assert(output->pos <= output->size);+ assert(input->pos <= input->size);++ if (mtctx->singleBlockingThread) { /* delegate to single-thread (synchronous) */+ return ZSTD_compressStream2(mtctx->cctxPool->cctx[0], output, input, endOp);+ }++ if ((mtctx->frameEnded) && (endOp==ZSTD_e_continue)) {+ /* current frame being ended. Only flush/end are allowed */+ return ERROR(stage_wrong);+ }++ /* single-pass shortcut (note : synchronous-mode) */+ if ( (!mtctx->params.rsyncable) /* rsyncable mode is disabled */+ && (mtctx->nextJobID == 0) /* just started */+ && (mtctx->inBuff.filled == 0) /* nothing buffered */+ && (!mtctx->jobReady) /* no job already created */+ && (endOp == ZSTD_e_end) /* end order */+ && (output->size - output->pos >= ZSTD_compressBound(input->size - input->pos)) ) { /* enough space in dst */+ size_t const cSize = ZSTDMT_compress_advanced_internal(mtctx,+ (char*)output->dst + output->pos, output->size - output->pos,+ (const char*)input->src + input->pos, input->size - input->pos,+ mtctx->cdict, mtctx->params);+ if (ZSTD_isError(cSize)) return cSize;+ input->pos = input->size;+ output->pos += cSize;+ mtctx->allJobsCompleted = 1;+ mtctx->frameEnded = 1;+ return 0;+ }++ /* fill input buffer */+ if ( (!mtctx->jobReady)+ && (input->size > input->pos) ) { /* support NULL input */+ if (mtctx->inBuff.buffer.start == NULL) {+ assert(mtctx->inBuff.filled == 0); /* Can't fill an empty buffer */+ if (!ZSTDMT_tryGetInputRange(mtctx)) {+ /* It is only possible for this operation to fail if there are+ * still compression jobs ongoing.+ */+ DEBUGLOG(5, "ZSTDMT_tryGetInputRange failed");+ assert(mtctx->doneJobID != mtctx->nextJobID);+ } else+ DEBUGLOG(5, "ZSTDMT_tryGetInputRange completed successfully : mtctx->inBuff.buffer.start = %p", mtctx->inBuff.buffer.start);+ }+ if (mtctx->inBuff.buffer.start != NULL) {+ syncPoint_t const syncPoint = findSynchronizationPoint(mtctx, *input);+ if (syncPoint.flush && endOp == ZSTD_e_continue) {+ endOp = ZSTD_e_flush;+ }+ assert(mtctx->inBuff.buffer.capacity >= mtctx->targetSectionSize);+ DEBUGLOG(5, "ZSTDMT_compressStream_generic: adding %u bytes on top of %u to buffer of size %u",+ (U32)syncPoint.toLoad, (U32)mtctx->inBuff.filled, (U32)mtctx->targetSectionSize);+ memcpy((char*)mtctx->inBuff.buffer.start + mtctx->inBuff.filled, (const char*)input->src + input->pos, syncPoint.toLoad);+ input->pos += syncPoint.toLoad;+ mtctx->inBuff.filled += syncPoint.toLoad;+ forwardInputProgress = syncPoint.toLoad>0;+ }+ if ((input->pos < input->size) && (endOp == ZSTD_e_end))+ endOp = ZSTD_e_flush; /* can't end now : not all input consumed */+ }++ if ( (mtctx->jobReady)+ || (mtctx->inBuff.filled >= mtctx->targetSectionSize) /* filled enough : let's compress */+ || ((endOp != ZSTD_e_continue) && (mtctx->inBuff.filled > 0)) /* something to flush : let's go */+ || ((endOp == ZSTD_e_end) && (!mtctx->frameEnded)) ) { /* must finish the frame with a zero-size block */+ size_t const jobSize = mtctx->inBuff.filled;+ assert(mtctx->inBuff.filled <= mtctx->targetSectionSize);+ FORWARD_IF_ERROR( ZSTDMT_createCompressionJob(mtctx, jobSize, endOp) );+ }++ /* check for potential compressed data ready to be flushed */+ { size_t const remainingToFlush = ZSTDMT_flushProduced(mtctx, output, !forwardInputProgress, endOp); /* block if there was no forward input progress */+ if (input->pos < input->size) return MAX(remainingToFlush, 1); /* input not consumed : do not end flush yet */+ DEBUGLOG(5, "end of ZSTDMT_compressStream_generic: remainingToFlush = %u", (U32)remainingToFlush);+ return remainingToFlush;+ }+}+++size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input)+{+ FORWARD_IF_ERROR( ZSTDMT_compressStream_generic(mtctx, output, input, ZSTD_e_continue) );++ /* recommended next input size : fill current input buffer */+ return mtctx->targetSectionSize - mtctx->inBuff.filled; /* note : could be zero when input buffer is fully filled and no more availability to create new job */+}+++static size_t ZSTDMT_flushStream_internal(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_EndDirective endFrame)+{+ size_t const srcSize = mtctx->inBuff.filled;+ DEBUGLOG(5, "ZSTDMT_flushStream_internal");++ if ( mtctx->jobReady /* one job ready for a worker to pick up */+ || (srcSize > 0) /* still some data within input buffer */+ || ((endFrame==ZSTD_e_end) && !mtctx->frameEnded)) { /* need a last 0-size block to end frame */+ DEBUGLOG(5, "ZSTDMT_flushStream_internal : create a new job (%u bytes, end:%u)",+ (U32)srcSize, (U32)endFrame);+ FORWARD_IF_ERROR( ZSTDMT_createCompressionJob(mtctx, srcSize, endFrame) );+ }++ /* check if there is any data available to flush */+ return ZSTDMT_flushProduced(mtctx, output, 1 /* blockToFlush */, endFrame);+}+++size_t ZSTDMT_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output)+{+ DEBUGLOG(5, "ZSTDMT_flushStream");+ if (mtctx->singleBlockingThread)+ return ZSTD_flushStream(mtctx->cctxPool->cctx[0], output);+ return ZSTDMT_flushStream_internal(mtctx, output, ZSTD_e_flush);+}++size_t ZSTDMT_endStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output)+{+ DEBUGLOG(4, "ZSTDMT_endStream");+ if (mtctx->singleBlockingThread)+ return ZSTD_endStream(mtctx->cctxPool->cctx[0], output);+ return ZSTDMT_flushStream_internal(mtctx, output, ZSTD_e_end);+}
+ zstd/lib/compress/zstdmt_compress.h view
@@ -0,0 +1,191 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++ #ifndef ZSTDMT_COMPRESS_H+ #define ZSTDMT_COMPRESS_H++ #if defined (__cplusplus)+ extern "C" {+ #endif+++/* Note : This is an internal API.+ * These APIs used to be exposed with ZSTDLIB_API,+ * because it used to be the only way to invoke MT compression.+ * Now, it's recommended to use ZSTD_compress2 and ZSTD_compressStream2()+ * instead.+ *+ * If you depend on these APIs and can't switch, then define+ * ZSTD_LEGACY_MULTITHREADED_API when making the dynamic library.+ * However, we may completely remove these functions in a future+ * release, so please switch soon.+ *+ * This API requires ZSTD_MULTITHREAD to be defined during compilation,+ * otherwise ZSTDMT_createCCtx*() will fail.+ */++#ifdef ZSTD_LEGACY_MULTITHREADED_API+# define ZSTDMT_API ZSTDLIB_API+#else+# define ZSTDMT_API+#endif++/* === Dependencies === */+#include <stddef.h> /* size_t */+#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_parameters */+#include "zstd.h" /* ZSTD_inBuffer, ZSTD_outBuffer, ZSTDLIB_API */+++/* === Constants === */+#ifndef ZSTDMT_NBWORKERS_MAX+# define ZSTDMT_NBWORKERS_MAX 200+#endif+#ifndef ZSTDMT_JOBSIZE_MIN+# define ZSTDMT_JOBSIZE_MIN (1 MB)+#endif+#define ZSTDMT_JOBSIZE_MAX (MEM_32bits() ? (512 MB) : (1024 MB))+++/* === Memory management === */+typedef struct ZSTDMT_CCtx_s ZSTDMT_CCtx;+/* Requires ZSTD_MULTITHREAD to be defined during compilation, otherwise it will return NULL. */+ZSTDMT_API ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers);+/* Requires ZSTD_MULTITHREAD to be defined during compilation, otherwise it will return NULL. */+ZSTDMT_API ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers,+ ZSTD_customMem cMem);+ZSTDMT_API size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx);++ZSTDMT_API size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx);+++/* === Simple one-pass compression function === */++ZSTDMT_API size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ int compressionLevel);++++/* === Streaming functions === */++ZSTDMT_API size_t ZSTDMT_initCStream(ZSTDMT_CCtx* mtctx, int compressionLevel);+ZSTDMT_API size_t ZSTDMT_resetCStream(ZSTDMT_CCtx* mtctx, unsigned long long pledgedSrcSize); /**< if srcSize is not known at reset time, use ZSTD_CONTENTSIZE_UNKNOWN. Note: for compatibility with older programs, 0 means the same as ZSTD_CONTENTSIZE_UNKNOWN, but it will change in the future to mean "empty" */++ZSTDMT_API size_t ZSTDMT_nextInputSizeHint(const ZSTDMT_CCtx* mtctx);+ZSTDMT_API size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input);++ZSTDMT_API size_t ZSTDMT_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output); /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */+ZSTDMT_API size_t ZSTDMT_endStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output); /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */+++/* === Advanced functions and parameters === */++ZSTDMT_API size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict,+ ZSTD_parameters params,+ int overlapLog);++ZSTDMT_API size_t ZSTDMT_initCStream_advanced(ZSTDMT_CCtx* mtctx,+ const void* dict, size_t dictSize, /* dict can be released after init, a local copy is preserved within zcs */+ ZSTD_parameters params,+ unsigned long long pledgedSrcSize); /* pledgedSrcSize is optional and can be zero == unknown */++ZSTDMT_API size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,+ const ZSTD_CDict* cdict,+ ZSTD_frameParameters fparams,+ unsigned long long pledgedSrcSize); /* note : zero means empty */++/* ZSTDMT_parameter :+ * List of parameters that can be set using ZSTDMT_setMTCtxParameter() */+typedef enum {+ ZSTDMT_p_jobSize, /* Each job is compressed in parallel. By default, this value is dynamically determined depending on compression parameters. Can be set explicitly here. */+ ZSTDMT_p_overlapLog, /* Each job may reload a part of previous job to enhance compression ratio; 0 == no overlap, 6(default) == use 1/8th of window, >=9 == use full window. This is a "sticky" parameter : its value will be re-used on next compression job */+ ZSTDMT_p_rsyncable /* Enables rsyncable mode. */+} ZSTDMT_parameter;++/* ZSTDMT_setMTCtxParameter() :+ * allow setting individual parameters, one at a time, among a list of enums defined in ZSTDMT_parameter.+ * The function must be called typically after ZSTD_createCCtx() but __before ZSTDMT_init*() !__+ * Parameters not explicitly reset by ZSTDMT_init*() remain the same in consecutive compression sessions.+ * @return : 0, or an error code (which can be tested using ZSTD_isError()) */+ZSTDMT_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value);++/* ZSTDMT_getMTCtxParameter() :+ * Query the ZSTDMT_CCtx for a parameter value.+ * @return : 0, or an error code (which can be tested using ZSTD_isError()) */+ZSTDMT_API size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value);+++/*! ZSTDMT_compressStream_generic() :+ * Combines ZSTDMT_compressStream() with optional ZSTDMT_flushStream() or ZSTDMT_endStream()+ * depending on flush directive.+ * @return : minimum amount of data still to be flushed+ * 0 if fully flushed+ * or an error code+ * note : needs to be init using any ZSTD_initCStream*() variant */+ZSTDMT_API size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,+ ZSTD_outBuffer* output,+ ZSTD_inBuffer* input,+ ZSTD_EndDirective endOp);+++/* ========================================================+ * === Private interface, for use by ZSTD_compress.c ===+ * === Not exposed in libzstd. Never invoke directly ===+ * ======================================================== */++ /*! ZSTDMT_toFlushNow()+ * Tell how many bytes are ready to be flushed immediately.+ * Probe the oldest active job (not yet entirely flushed) and check its output buffer.+ * If return 0, it means there is no active job,+ * or, it means oldest job is still active, but everything produced has been flushed so far,+ * therefore flushing is limited by speed of oldest job. */+size_t ZSTDMT_toFlushNow(ZSTDMT_CCtx* mtctx);++/*! ZSTDMT_CCtxParam_setMTCtxParameter()+ * like ZSTDMT_setMTCtxParameter(), but into a ZSTD_CCtx_Params */+size_t ZSTDMT_CCtxParam_setMTCtxParameter(ZSTD_CCtx_params* params, ZSTDMT_parameter parameter, int value);++/*! ZSTDMT_CCtxParam_setNbWorkers()+ * Set nbWorkers, and clamp it.+ * Also reset jobSize and overlapLog */+size_t ZSTDMT_CCtxParam_setNbWorkers(ZSTD_CCtx_params* params, unsigned nbWorkers);++/*! ZSTDMT_updateCParams_whileCompressing() :+ * Updates only a selected set of compression parameters, to remain compatible with current frame.+ * New parameters will be applied to next compression job. */+void ZSTDMT_updateCParams_whileCompressing(ZSTDMT_CCtx* mtctx, const ZSTD_CCtx_params* cctxParams);++/*! ZSTDMT_getFrameProgression():+ * tells how much data has been consumed (input) and produced (output) for current frame.+ * able to count progression inside worker threads.+ */+ZSTD_frameProgression ZSTDMT_getFrameProgression(ZSTDMT_CCtx* mtctx);+++/*! ZSTDMT_initCStream_internal() :+ * Private use only. Init streaming operation.+ * expects params to be valid.+ * must receive dict, or cdict, or none, but not both.+ * @return : 0, or an error code */+size_t ZSTDMT_initCStream_internal(ZSTDMT_CCtx* zcs,+ const void* dict, size_t dictSize, ZSTD_dictContentType_e dictContentType,+ const ZSTD_CDict* cdict,+ ZSTD_CCtx_params params, unsigned long long pledgedSrcSize);+++#if defined (__cplusplus)+}+#endif++#endif /* ZSTDMT_COMPRESS_H */
zstd/lib/decompress/huf_decompress.c view
@@ -1,6 +1,7 @@ /* ******************************************************************- Huffman decoder, part of New Generation Entropy library- Copyright (C) 2013-2016, Yann Collet.+ huff0 huffman decoder,+ part of Finite State Entropy library+ Copyright (C) 2013-present, Yann Collet. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -29,45 +30,95 @@ You can contact the author at : - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy- - Public forum : https://groups.google.com/forum/#!forum/lz4c ****************************************************************** */ /* **************************************************************-* Compiler specifics-****************************************************************/-#if defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)-/* inline is defined */-#elif defined(_MSC_VER) || defined(__GNUC__)-# define inline __inline-#else-# define inline /* disable inline */-#endif--#ifdef _MSC_VER /* Visual Studio */-# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */-#endif---/* ************************************************************** * Dependencies ****************************************************************/ #include <string.h> /* memcpy, memset */+#include "compiler.h" #include "bitstream.h" /* BIT_* */-#include "fse.h" /* header compression */+#include "fse.h" /* to compress headers */ #define HUF_STATIC_LINKING_ONLY #include "huf.h"+#include "error_private.h" +/* **************************************************************+* Macros+****************************************************************/ +/* These two optional macros force the use one way or another of the two+ * Huffman decompression implementations. You can't force in both directions+ * at the same time.+ */+#if defined(HUF_FORCE_DECOMPRESS_X1) && \+ defined(HUF_FORCE_DECOMPRESS_X2)+#error "Cannot force the use of the X1 and X2 decoders at the same time!"+#endif++ /* ************************************************************** * Error Management ****************************************************************/-#define HUF_STATIC_ASSERT(c) { enum { HUF_static_assert = 1/(int)(!!(c)) }; } /* use only *after* variable declarations */+#define HUF_isError ERR_isError+#define CHECK_F(f) { size_t const err_ = (f); if (HUF_isError(err_)) return err_; } +/* **************************************************************+* Byte alignment for workSpace management+****************************************************************/+#define HUF_ALIGN(x, a) HUF_ALIGN_MASK((x), (a) - 1)+#define HUF_ALIGN_MASK(x, mask) (((x) + (mask)) & ~(mask))+++/* **************************************************************+* BMI2 Variant Wrappers+****************************************************************/+#if DYNAMIC_BMI2++#define HUF_DGEN(fn) \+ \+ static size_t fn##_default( \+ void* dst, size_t dstSize, \+ const void* cSrc, size_t cSrcSize, \+ const HUF_DTable* DTable) \+ { \+ return fn##_body(dst, dstSize, cSrc, cSrcSize, DTable); \+ } \+ \+ static TARGET_ATTRIBUTE("bmi2") size_t fn##_bmi2( \+ void* dst, size_t dstSize, \+ const void* cSrc, size_t cSrcSize, \+ const HUF_DTable* DTable) \+ { \+ return fn##_body(dst, dstSize, cSrc, cSrcSize, DTable); \+ } \+ \+ static size_t fn(void* dst, size_t dstSize, void const* cSrc, \+ size_t cSrcSize, HUF_DTable const* DTable, int bmi2) \+ { \+ if (bmi2) { \+ return fn##_bmi2(dst, dstSize, cSrc, cSrcSize, DTable); \+ } \+ return fn##_default(dst, dstSize, cSrc, cSrcSize, DTable); \+ }++#else++#define HUF_DGEN(fn) \+ static size_t fn(void* dst, size_t dstSize, void const* cSrc, \+ size_t cSrcSize, HUF_DTable const* DTable, int bmi2) \+ { \+ (void)bmi2; \+ return fn##_body(dst, dstSize, cSrc, cSrcSize, DTable); \+ }++#endif++ /*-***************************/ /* generic DTableDesc */ /*-***************************/- typedef struct { BYTE maxTableLog; BYTE tableType; BYTE tableLog; BYTE reserved; } DTableDesc; static DTableDesc HUF_getDTableDesc(const HUF_DTable* table)@@ -78,23 +129,33 @@ } +#ifndef HUF_FORCE_DECOMPRESS_X2+ /*-***************************/ /* single-symbol decoding */ /*-***************************/--typedef struct { BYTE byte; BYTE nbBits; } HUF_DEltX2; /* single-symbol decoding */+typedef struct { BYTE byte; BYTE nbBits; } HUF_DEltX1; /* single-symbol decoding */ -size_t HUF_readDTableX2 (HUF_DTable* DTable, const void* src, size_t srcSize)+size_t HUF_readDTableX1_wksp(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize) {- BYTE huffWeight[HUF_SYMBOLVALUE_MAX + 1];- U32 rankVal[HUF_TABLELOG_ABSOLUTEMAX + 1]; /* large enough for values from 0 to 16 */ U32 tableLog = 0; U32 nbSymbols = 0; size_t iSize; void* const dtPtr = DTable + 1;- HUF_DEltX2* const dt = (HUF_DEltX2*)dtPtr;+ HUF_DEltX1* const dt = (HUF_DEltX1*)dtPtr; - HUF_STATIC_ASSERT(sizeof(DTableDesc) == sizeof(HUF_DTable));+ U32* rankVal;+ BYTE* huffWeight;+ size_t spaceUsed32 = 0;++ rankVal = (U32 *)workSpace + spaceUsed32;+ spaceUsed32 += HUF_TABLELOG_ABSOLUTEMAX + 1;+ huffWeight = (BYTE *)((U32 *)workSpace + spaceUsed32);+ spaceUsed32 += HUF_ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2;++ if ((spaceUsed32 << 2) > wkspSize) return ERROR(tableLog_tooLarge);++ DEBUG_STATIC_ASSERT(sizeof(DTableDesc) == sizeof(HUF_DTable)); /* memset(huffWeight, 0, sizeof(huffWeight)); */ /* is not necessary, even though some analyzer complain ... */ iSize = HUF_readStats(huffWeight, HUF_SYMBOLVALUE_MAX + 1, rankVal, &nbSymbols, &tableLog, src, srcSize);@@ -102,16 +163,16 @@ /* Table header */ { DTableDesc dtd = HUF_getDTableDesc(DTable);- if (tableLog > (U32)(dtd.maxTableLog+1)) return ERROR(tableLog_tooLarge); /* DTable too small, huffman tree cannot fit in */+ if (tableLog > (U32)(dtd.maxTableLog+1)) return ERROR(tableLog_tooLarge); /* DTable too small, Huffman tree cannot fit in */ dtd.tableType = 0; dtd.tableLog = (BYTE)tableLog; memcpy(DTable, &dtd, sizeof(dtd)); } - /* Prepare ranks */+ /* Calculate starting value for each rank */ { U32 n, nextRankStart = 0; for (n=1; n<tableLog+1; n++) {- U32 current = nextRankStart;+ U32 const current = nextRankStart; nextRankStart += (rankVal[n] << (n-1)); rankVal[n] = current; } }@@ -121,19 +182,26 @@ for (n=0; n<nbSymbols; n++) { U32 const w = huffWeight[n]; U32 const length = (1 << w) >> 1;- U32 i;- HUF_DEltX2 D;+ U32 u;+ HUF_DEltX1 D; D.byte = (BYTE)n; D.nbBits = (BYTE)(tableLog + 1 - w);- for (i = rankVal[w]; i < rankVal[w] + length; i++)- dt[i] = D;+ for (u = rankVal[w]; u < rankVal[w] + length; u++)+ dt[u] = D; rankVal[w] += length; } } return iSize; } +size_t HUF_readDTableX1(HUF_DTable* DTable, const void* src, size_t srcSize)+{+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_readDTableX1_wksp(DTable, src, srcSize,+ workSpace, sizeof(workSpace));+} -static BYTE HUF_decodeSymbolX2(BIT_DStream_t* Dstream, const HUF_DEltX2* dt, const U32 dtLog)+FORCE_INLINE_TEMPLATE BYTE+HUF_decodeSymbolX1(BIT_DStream_t* Dstream, const HUF_DEltX1* dt, const U32 dtLog) { size_t const val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */ BYTE const c = dt[val].byte;@@ -141,41 +209,44 @@ return c; } -#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \- *ptr++ = HUF_decodeSymbolX2(DStreamPtr, dt, dtLog)+#define HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr) \+ *ptr++ = HUF_decodeSymbolX1(DStreamPtr, dt, dtLog) -#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \+#define HUF_DECODE_SYMBOLX1_1(ptr, DStreamPtr) \ if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \- HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)+ HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr) -#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \+#define HUF_DECODE_SYMBOLX1_2(ptr, DStreamPtr) \ if (MEM_64bits()) \- HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)+ HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr) -static inline size_t HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX2* const dt, const U32 dtLog)+HINT_INLINE size_t+HUF_decodeStreamX1(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX1* const dt, const U32 dtLog) { BYTE* const pStart = p; /* up to 4 symbols at a time */- while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-4)) {- HUF_DECODE_SYMBOLX2_2(p, bitDPtr);- HUF_DECODE_SYMBOLX2_1(p, bitDPtr);- HUF_DECODE_SYMBOLX2_2(p, bitDPtr);- HUF_DECODE_SYMBOLX2_0(p, bitDPtr);+ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd-3)) {+ HUF_DECODE_SYMBOLX1_2(p, bitDPtr);+ HUF_DECODE_SYMBOLX1_1(p, bitDPtr);+ HUF_DECODE_SYMBOLX1_2(p, bitDPtr);+ HUF_DECODE_SYMBOLX1_0(p, bitDPtr); } - /* closer to the end */- while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd))- HUF_DECODE_SYMBOLX2_0(p, bitDPtr);+ /* [0-3] symbols remaining */+ if (MEM_32bits())+ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd))+ HUF_DECODE_SYMBOLX1_0(p, bitDPtr); - /* no more data to retrieve from bitstream, hence no need to reload */+ /* no more data to retrieve from bitstream, no need to reload */ while (p < pEnd)- HUF_DECODE_SYMBOLX2_0(p, bitDPtr);+ HUF_DECODE_SYMBOLX1_0(p, bitDPtr); return pEnd-pStart; } -static size_t HUF_decompress1X2_usingDTable_internal(+FORCE_INLINE_TEMPLATE size_t+HUF_decompress1X1_usingDTable_internal_body( void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable)@@ -183,52 +254,22 @@ BYTE* op = (BYTE*)dst; BYTE* const oend = op + dstSize; const void* dtPtr = DTable + 1;- const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;+ const HUF_DEltX1* const dt = (const HUF_DEltX1*)dtPtr; BIT_DStream_t bitD; DTableDesc const dtd = HUF_getDTableDesc(DTable); U32 const dtLog = dtd.tableLog; - { size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);- if (HUF_isError(errorCode)) return errorCode; }+ CHECK_F( BIT_initDStream(&bitD, cSrc, cSrcSize) ); - HUF_decodeStreamX2(op, &bitD, oend, dt, dtLog);+ HUF_decodeStreamX1(op, &bitD, oend, dt, dtLog); - /* check */ if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected); return dstSize; } -size_t HUF_decompress1X2_usingDTable(- void* dst, size_t dstSize,- const void* cSrc, size_t cSrcSize,- const HUF_DTable* DTable)-{- DTableDesc dtd = HUF_getDTableDesc(DTable);- if (dtd.tableType != 0) return ERROR(GENERIC);- return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);-}--size_t HUF_decompress1X2_DCtx (HUF_DTable* DCtx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)-{- const BYTE* ip = (const BYTE*) cSrc;-- size_t const hSize = HUF_readDTableX2 (DCtx, cSrc, cSrcSize);- if (HUF_isError(hSize)) return hSize;- if (hSize >= cSrcSize) return ERROR(srcSize_wrong);- ip += hSize; cSrcSize -= hSize;-- return HUF_decompress1X2_usingDTable_internal (dst, dstSize, ip, cSrcSize, DCtx);-}--size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)-{- HUF_CREATE_STATIC_DTABLEX2(DTable, HUF_TABLELOG_MAX);- return HUF_decompress1X2_DCtx (DTable, dst, dstSize, cSrc, cSrcSize);-}---static size_t HUF_decompress4X2_usingDTable_internal(+FORCE_INLINE_TEMPLATE size_t+HUF_decompress4X1_usingDTable_internal_body( void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable)@@ -240,7 +281,7 @@ BYTE* const ostart = (BYTE*) dst; BYTE* const oend = ostart + dstSize; const void* const dtPtr = DTable + 1;- const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;+ const HUF_DEltX1* const dt = (const HUF_DEltX1*)dtPtr; /* Init */ BIT_DStream_t bitD1;@@ -263,57 +304,58 @@ BYTE* op2 = opStart2; BYTE* op3 = opStart3; BYTE* op4 = opStart4;- U32 endSignal;+ U32 endSignal = BIT_DStream_unfinished; DTableDesc const dtd = HUF_getDTableDesc(DTable); U32 const dtLog = dtd.tableLog; if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */- { size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);- if (HUF_isError(errorCode)) return errorCode; }- { size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);- if (HUF_isError(errorCode)) return errorCode; }- { size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);- if (HUF_isError(errorCode)) return errorCode; }- { size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);- if (HUF_isError(errorCode)) return errorCode; }+ CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );+ CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );+ CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );+ CHECK_F( BIT_initDStream(&bitD4, istart4, length4) ); - /* 16-32 symbols per loop (4-8 symbols per stream) */+ /* up to 16 symbols per loop (4 symbols per stream) in 64-bit mode */ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);- for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; ) {- HUF_DECODE_SYMBOLX2_2(op1, &bitD1);- HUF_DECODE_SYMBOLX2_2(op2, &bitD2);- HUF_DECODE_SYMBOLX2_2(op3, &bitD3);- HUF_DECODE_SYMBOLX2_2(op4, &bitD4);- HUF_DECODE_SYMBOLX2_1(op1, &bitD1);- HUF_DECODE_SYMBOLX2_1(op2, &bitD2);- HUF_DECODE_SYMBOLX2_1(op3, &bitD3);- HUF_DECODE_SYMBOLX2_1(op4, &bitD4);- HUF_DECODE_SYMBOLX2_2(op1, &bitD1);- HUF_DECODE_SYMBOLX2_2(op2, &bitD2);- HUF_DECODE_SYMBOLX2_2(op3, &bitD3);- HUF_DECODE_SYMBOLX2_2(op4, &bitD4);- HUF_DECODE_SYMBOLX2_0(op1, &bitD1);- HUF_DECODE_SYMBOLX2_0(op2, &bitD2);- HUF_DECODE_SYMBOLX2_0(op3, &bitD3);- HUF_DECODE_SYMBOLX2_0(op4, &bitD4);- endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);+ while ( (endSignal==BIT_DStream_unfinished) && (op4<(oend-3)) ) {+ HUF_DECODE_SYMBOLX1_2(op1, &bitD1);+ HUF_DECODE_SYMBOLX1_2(op2, &bitD2);+ HUF_DECODE_SYMBOLX1_2(op3, &bitD3);+ HUF_DECODE_SYMBOLX1_2(op4, &bitD4);+ HUF_DECODE_SYMBOLX1_1(op1, &bitD1);+ HUF_DECODE_SYMBOLX1_1(op2, &bitD2);+ HUF_DECODE_SYMBOLX1_1(op3, &bitD3);+ HUF_DECODE_SYMBOLX1_1(op4, &bitD4);+ HUF_DECODE_SYMBOLX1_2(op1, &bitD1);+ HUF_DECODE_SYMBOLX1_2(op2, &bitD2);+ HUF_DECODE_SYMBOLX1_2(op3, &bitD3);+ HUF_DECODE_SYMBOLX1_2(op4, &bitD4);+ HUF_DECODE_SYMBOLX1_0(op1, &bitD1);+ HUF_DECODE_SYMBOLX1_0(op2, &bitD2);+ HUF_DECODE_SYMBOLX1_0(op3, &bitD3);+ HUF_DECODE_SYMBOLX1_0(op4, &bitD4);+ BIT_reloadDStream(&bitD1);+ BIT_reloadDStream(&bitD2);+ BIT_reloadDStream(&bitD3);+ BIT_reloadDStream(&bitD4); } /* check corruption */+ /* note : should not be necessary : op# advance in lock step, and we control op4.+ * but curiously, binary generated by gcc 7.2 & 7.3 with -mbmi2 runs faster when >=1 test is present */ if (op1 > opStart2) return ERROR(corruption_detected); if (op2 > opStart3) return ERROR(corruption_detected); if (op3 > opStart4) return ERROR(corruption_detected); /* note : op4 supposed already verified within main loop */ /* finish bitStreams one by one */- HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog);- HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog);- HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog);- HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog);+ HUF_decodeStreamX1(op1, &bitD1, opStart2, dt, dtLog);+ HUF_decodeStreamX1(op2, &bitD2, opStart3, dt, dtLog);+ HUF_decodeStreamX1(op3, &bitD3, opStart4, dt, dtLog);+ HUF_decodeStreamX1(op4, &bitD4, oend, dt, dtLog); /* check */- endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);- if (!endSignal) return ERROR(corruption_detected);+ { U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);+ if (!endCheck) return ERROR(corruption_detected); } /* decoded size */ return dstSize;@@ -321,51 +363,123 @@ } -size_t HUF_decompress4X2_usingDTable(+typedef size_t (*HUF_decompress_usingDTable_t)(void *dst, size_t dstSize,+ const void *cSrc,+ size_t cSrcSize,+ const HUF_DTable *DTable);++HUF_DGEN(HUF_decompress1X1_usingDTable_internal)+HUF_DGEN(HUF_decompress4X1_usingDTable_internal)++++size_t HUF_decompress1X1_usingDTable( void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable) { DTableDesc dtd = HUF_getDTableDesc(DTable); if (dtd.tableType != 0) return ERROR(GENERIC);- return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);+ return HUF_decompress1X1_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0); } +size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* DCtx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize)+{+ const BYTE* ip = (const BYTE*) cSrc; -size_t HUF_decompress4X2_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+ size_t const hSize = HUF_readDTableX1_wksp(DCtx, cSrc, cSrcSize, workSpace, wkspSize);+ if (HUF_isError(hSize)) return hSize;+ if (hSize >= cSrcSize) return ERROR(srcSize_wrong);+ ip += hSize; cSrcSize -= hSize;++ return HUF_decompress1X1_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx, /* bmi2 */ 0);+}+++size_t HUF_decompress1X1_DCtx(HUF_DTable* DCtx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize) {+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_decompress1X1_DCtx_wksp(DCtx, dst, dstSize, cSrc, cSrcSize,+ workSpace, sizeof(workSpace));+}++size_t HUF_decompress1X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+{+ HUF_CREATE_STATIC_DTABLEX1(DTable, HUF_TABLELOG_MAX);+ return HUF_decompress1X1_DCtx (DTable, dst, dstSize, cSrc, cSrcSize);+}++size_t HUF_decompress4X1_usingDTable(+ void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ const HUF_DTable* DTable)+{+ DTableDesc dtd = HUF_getDTableDesc(DTable);+ if (dtd.tableType != 0) return ERROR(GENERIC);+ return HUF_decompress4X1_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+}++static size_t HUF_decompress4X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize, int bmi2)+{ const BYTE* ip = (const BYTE*) cSrc; - size_t const hSize = HUF_readDTableX2 (dctx, cSrc, cSrcSize);+ size_t const hSize = HUF_readDTableX1_wksp (dctx, cSrc, cSrcSize,+ workSpace, wkspSize); if (HUF_isError(hSize)) return hSize; if (hSize >= cSrcSize) return ERROR(srcSize_wrong); ip += hSize; cSrcSize -= hSize; - return HUF_decompress4X2_usingDTable_internal (dst, dstSize, ip, cSrcSize, dctx);+ return HUF_decompress4X1_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2); } -size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+size_t HUF_decompress4X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize) {- HUF_CREATE_STATIC_DTABLEX2(DTable, HUF_TABLELOG_MAX);- return HUF_decompress4X2_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);+ return HUF_decompress4X1_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, 0); } +size_t HUF_decompress4X1_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+{+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_decompress4X1_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize,+ workSpace, sizeof(workSpace));+}+size_t HUF_decompress4X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+{+ HUF_CREATE_STATIC_DTABLEX1(DTable, HUF_TABLELOG_MAX);+ return HUF_decompress4X1_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);+}++#endif /* HUF_FORCE_DECOMPRESS_X2 */+++#ifndef HUF_FORCE_DECOMPRESS_X1+ /* *************************/ /* double-symbols decoding */ /* *************************/-typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX4; /* double-symbols decoding */ +typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX2; /* double-symbols decoding */ typedef struct { BYTE symbol; BYTE weight; } sortedSymbol_t;+typedef U32 rankValCol_t[HUF_TABLELOG_MAX + 1];+typedef rankValCol_t rankVal_t[HUF_TABLELOG_MAX]; -/* HUF_fillDTableX4Level2() :++/* HUF_fillDTableX2Level2() : * `rankValOrigin` must be a table of at least (HUF_TABLELOG_MAX + 1) U32 */-static void HUF_fillDTableX4Level2(HUF_DEltX4* DTable, U32 sizeLog, const U32 consumed,+static void HUF_fillDTableX2Level2(HUF_DEltX2* DTable, U32 sizeLog, const U32 consumed, const U32* rankValOrigin, const int minWeight, const sortedSymbol_t* sortedSymbols, const U32 sortedListSize, U32 nbBitsBaseline, U16 baseSeq) {- HUF_DEltX4 DElt;+ HUF_DEltX2 DElt; U32 rankVal[HUF_TABLELOG_MAX + 1]; /* get pre-calculated rankVal */@@ -400,9 +514,8 @@ } } } -typedef U32 rankVal_t[HUF_TABLELOG_MAX][HUF_TABLELOG_MAX + 1]; -static void HUF_fillDTableX4(HUF_DEltX4* DTable, const U32 targetLog,+static void HUF_fillDTableX2(HUF_DEltX2* DTable, const U32 targetLog, const sortedSymbol_t* sortedList, const U32 sortedListSize, const U32* rankStart, rankVal_t rankValOrigin, const U32 maxWeight, const U32 nbBitsBaseline)@@ -427,12 +540,12 @@ int minWeight = nbBits + scaleLog; if (minWeight < 1) minWeight = 1; sortedRank = rankStart[minWeight];- HUF_fillDTableX4Level2(DTable+start, targetLog-nbBits, nbBits,+ HUF_fillDTableX2Level2(DTable+start, targetLog-nbBits, nbBits, rankValOrigin[nbBits], minWeight, sortedList+sortedRank, sortedListSize-sortedRank, nbBitsBaseline, symbol); } else {- HUF_DEltX4 DElt;+ HUF_DEltX2 DElt; MEM_writeLE16(&(DElt.sequence), symbol); DElt.nbBits = (BYTE)(nbBits); DElt.length = 1;@@ -444,22 +557,42 @@ } } -size_t HUF_readDTableX4 (HUF_DTable* DTable, const void* src, size_t srcSize)+size_t HUF_readDTableX2_wksp(HUF_DTable* DTable,+ const void* src, size_t srcSize,+ void* workSpace, size_t wkspSize) {- BYTE weightList[HUF_SYMBOLVALUE_MAX + 1];- sortedSymbol_t sortedSymbol[HUF_SYMBOLVALUE_MAX + 1];- U32 rankStats[HUF_TABLELOG_MAX + 1] = { 0 };- U32 rankStart0[HUF_TABLELOG_MAX + 2] = { 0 };- U32* const rankStart = rankStart0+1;- rankVal_t rankVal; U32 tableLog, maxW, sizeOfSort, nbSymbols; DTableDesc dtd = HUF_getDTableDesc(DTable); U32 const maxTableLog = dtd.maxTableLog; size_t iSize; void* dtPtr = DTable+1; /* force compiler to avoid strict-aliasing */- HUF_DEltX4* const dt = (HUF_DEltX4*)dtPtr;+ HUF_DEltX2* const dt = (HUF_DEltX2*)dtPtr;+ U32 *rankStart; - HUF_STATIC_ASSERT(sizeof(HUF_DEltX4) == sizeof(HUF_DTable)); /* if compilation fails here, assertion is false */+ rankValCol_t* rankVal;+ U32* rankStats;+ U32* rankStart0;+ sortedSymbol_t* sortedSymbol;+ BYTE* weightList;+ size_t spaceUsed32 = 0;++ rankVal = (rankValCol_t *)((U32 *)workSpace + spaceUsed32);+ spaceUsed32 += (sizeof(rankValCol_t) * HUF_TABLELOG_MAX) >> 2;+ rankStats = (U32 *)workSpace + spaceUsed32;+ spaceUsed32 += HUF_TABLELOG_MAX + 1;+ rankStart0 = (U32 *)workSpace + spaceUsed32;+ spaceUsed32 += HUF_TABLELOG_MAX + 2;+ sortedSymbol = (sortedSymbol_t *)workSpace + (spaceUsed32 * sizeof(U32)) / sizeof(sortedSymbol_t);+ spaceUsed32 += HUF_ALIGN(sizeof(sortedSymbol_t) * (HUF_SYMBOLVALUE_MAX + 1), sizeof(U32)) >> 2;+ weightList = (BYTE *)((U32 *)workSpace + spaceUsed32);+ spaceUsed32 += HUF_ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2;++ if ((spaceUsed32 << 2) > wkspSize) return ERROR(tableLog_tooLarge);++ rankStart = rankStart0 + 1;+ memset(rankStats, 0, sizeof(U32) * (2 * HUF_TABLELOG_MAX + 2 + 1));++ DEBUG_STATIC_ASSERT(sizeof(HUF_DEltX2) == sizeof(HUF_DTable)); /* if compiler fails here, assertion is wrong */ if (maxTableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge); /* memset(weightList, 0, sizeof(weightList)); */ /* is not necessary, even though some analyzer complain ... */ @@ -513,7 +646,7 @@ rankValPtr[w] = rankVal0[w] >> consumed; } } } } - HUF_fillDTableX4(dt, maxTableLog,+ HUF_fillDTableX2(dt, maxTableLog, sortedSymbol, sizeOfSort, rankStart0, rankVal, maxW, tableLog+1);@@ -524,8 +657,16 @@ return iSize; } +size_t HUF_readDTableX2(HUF_DTable* DTable, const void* src, size_t srcSize)+{+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_readDTableX2_wksp(DTable, src, srcSize,+ workSpace, sizeof(workSpace));+} -static U32 HUF_decodeSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)++FORCE_INLINE_TEMPLATE U32+HUF_decodeSymbolX2(void* op, BIT_DStream_t* DStream, const HUF_DEltX2* dt, const U32 dtLog) { size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */ memcpy(op, dt+val, 2);@@ -533,7 +674,8 @@ return dt[val].length; } -static U32 HUF_decodeLastSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)+FORCE_INLINE_TEMPLATE U32+HUF_decodeLastSymbolX2(void* op, BIT_DStream_t* DStream, const HUF_DEltX2* dt, const U32 dtLog) { size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */ memcpy(op, dt+val, 1);@@ -542,50 +684,52 @@ if (DStream->bitsConsumed < (sizeof(DStream->bitContainer)*8)) { BIT_skipBits(DStream, dt[val].nbBits); if (DStream->bitsConsumed > (sizeof(DStream->bitContainer)*8))- DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8); /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */+ /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */+ DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8); } } return 1; } --#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) \- ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)+#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \+ ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog) -#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr) \+#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \ if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \- ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)+ ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog) -#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \+#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \ if (MEM_64bits()) \- ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)+ ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog) -static inline size_t HUF_decodeStreamX4(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd, const HUF_DEltX4* const dt, const U32 dtLog)+HINT_INLINE size_t+HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd,+ const HUF_DEltX2* const dt, const U32 dtLog) { BYTE* const pStart = p; /* up to 8 symbols at a time */ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd-(sizeof(bitDPtr->bitContainer)-1))) {- HUF_DECODE_SYMBOLX4_2(p, bitDPtr);- HUF_DECODE_SYMBOLX4_1(p, bitDPtr);- HUF_DECODE_SYMBOLX4_2(p, bitDPtr);- HUF_DECODE_SYMBOLX4_0(p, bitDPtr);+ HUF_DECODE_SYMBOLX2_2(p, bitDPtr);+ HUF_DECODE_SYMBOLX2_1(p, bitDPtr);+ HUF_DECODE_SYMBOLX2_2(p, bitDPtr);+ HUF_DECODE_SYMBOLX2_0(p, bitDPtr); } /* closer to end : up to 2 symbols at a time */ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p <= pEnd-2))- HUF_DECODE_SYMBOLX4_0(p, bitDPtr);+ HUF_DECODE_SYMBOLX2_0(p, bitDPtr); while (p <= pEnd-2)- HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */+ HUF_DECODE_SYMBOLX2_0(p, bitDPtr); /* no need to reload : reached the end of DStream */ if (p < pEnd)- p += HUF_decodeLastSymbolX4(p, bitDPtr, dt, dtLog);+ p += HUF_decodeLastSymbolX2(p, bitDPtr, dt, dtLog); return p-pStart; } --static size_t HUF_decompress1X4_usingDTable_internal(+FORCE_INLINE_TEMPLATE size_t+HUF_decompress1X2_usingDTable_internal_body( void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable)@@ -593,17 +737,15 @@ BIT_DStream_t bitD; /* Init */- { size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);- if (HUF_isError(errorCode)) return errorCode;- }+ CHECK_F( BIT_initDStream(&bitD, cSrc, cSrcSize) ); /* decode */ { BYTE* const ostart = (BYTE*) dst; BYTE* const oend = ostart + dstSize; const void* const dtPtr = DTable+1; /* force compiler to not use strict-aliasing */- const HUF_DEltX4* const dt = (const HUF_DEltX4*)dtPtr;+ const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr; DTableDesc const dtd = HUF_getDTableDesc(DTable);- HUF_decodeStreamX4(ostart, &bitD, oend, dt, dtd.tableLog);+ HUF_decodeStreamX2(ostart, &bitD, oend, dt, dtd.tableLog); } /* check */@@ -613,35 +755,9 @@ return dstSize; } -size_t HUF_decompress1X4_usingDTable(- void* dst, size_t dstSize,- const void* cSrc, size_t cSrcSize,- const HUF_DTable* DTable)-{- DTableDesc dtd = HUF_getDTableDesc(DTable);- if (dtd.tableType != 1) return ERROR(GENERIC);- return HUF_decompress1X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);-} -size_t HUF_decompress1X4_DCtx (HUF_DTable* DCtx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)-{- const BYTE* ip = (const BYTE*) cSrc;-- size_t const hSize = HUF_readDTableX4 (DCtx, cSrc, cSrcSize);- if (HUF_isError(hSize)) return hSize;- if (hSize >= cSrcSize) return ERROR(srcSize_wrong);- ip += hSize; cSrcSize -= hSize;-- return HUF_decompress1X4_usingDTable_internal (dst, dstSize, ip, cSrcSize, DCtx);-}--size_t HUF_decompress1X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)-{- HUF_CREATE_STATIC_DTABLEX4(DTable, HUF_TABLELOG_MAX);- return HUF_decompress1X4_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);-}--static size_t HUF_decompress4X4_usingDTable_internal(+FORCE_INLINE_TEMPLATE size_t+HUF_decompress4X2_usingDTable_internal_body( void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable)@@ -652,7 +768,7 @@ BYTE* const ostart = (BYTE*) dst; BYTE* const oend = ostart + dstSize; const void* const dtPtr = DTable+1;- const HUF_DEltX4* const dt = (const HUF_DEltX4*)dtPtr;+ const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr; /* Init */ BIT_DStream_t bitD1;@@ -680,34 +796,30 @@ U32 const dtLog = dtd.tableLog; if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */- { size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);- if (HUF_isError(errorCode)) return errorCode; }- { size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);- if (HUF_isError(errorCode)) return errorCode; }- { size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);- if (HUF_isError(errorCode)) return errorCode; }- { size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);- if (HUF_isError(errorCode)) return errorCode; }+ CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );+ CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );+ CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );+ CHECK_F( BIT_initDStream(&bitD4, istart4, length4) ); /* 16-32 symbols per loop (4-8 symbols per stream) */ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); for ( ; (endSignal==BIT_DStream_unfinished) & (op4<(oend-(sizeof(bitD4.bitContainer)-1))) ; ) {- HUF_DECODE_SYMBOLX4_2(op1, &bitD1);- HUF_DECODE_SYMBOLX4_2(op2, &bitD2);- HUF_DECODE_SYMBOLX4_2(op3, &bitD3);- HUF_DECODE_SYMBOLX4_2(op4, &bitD4);- HUF_DECODE_SYMBOLX4_1(op1, &bitD1);- HUF_DECODE_SYMBOLX4_1(op2, &bitD2);- HUF_DECODE_SYMBOLX4_1(op3, &bitD3);- HUF_DECODE_SYMBOLX4_1(op4, &bitD4);- HUF_DECODE_SYMBOLX4_2(op1, &bitD1);- HUF_DECODE_SYMBOLX4_2(op2, &bitD2);- HUF_DECODE_SYMBOLX4_2(op3, &bitD3);- HUF_DECODE_SYMBOLX4_2(op4, &bitD4);- HUF_DECODE_SYMBOLX4_0(op1, &bitD1);- HUF_DECODE_SYMBOLX4_0(op2, &bitD2);- HUF_DECODE_SYMBOLX4_0(op3, &bitD3);- HUF_DECODE_SYMBOLX4_0(op4, &bitD4);+ HUF_DECODE_SYMBOLX2_2(op1, &bitD1);+ HUF_DECODE_SYMBOLX2_2(op2, &bitD2);+ HUF_DECODE_SYMBOLX2_2(op3, &bitD3);+ HUF_DECODE_SYMBOLX2_2(op4, &bitD4);+ HUF_DECODE_SYMBOLX2_1(op1, &bitD1);+ HUF_DECODE_SYMBOLX2_1(op2, &bitD2);+ HUF_DECODE_SYMBOLX2_1(op3, &bitD3);+ HUF_DECODE_SYMBOLX2_1(op4, &bitD4);+ HUF_DECODE_SYMBOLX2_2(op1, &bitD1);+ HUF_DECODE_SYMBOLX2_2(op2, &bitD2);+ HUF_DECODE_SYMBOLX2_2(op3, &bitD3);+ HUF_DECODE_SYMBOLX2_2(op4, &bitD4);+ HUF_DECODE_SYMBOLX2_0(op1, &bitD1);+ HUF_DECODE_SYMBOLX2_0(op2, &bitD2);+ HUF_DECODE_SYMBOLX2_0(op3, &bitD3);+ HUF_DECODE_SYMBOLX2_0(op4, &bitD4); endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4); }@@ -719,10 +831,10 @@ /* note : op4 already verified within main loop */ /* finish bitStreams one by one */- HUF_decodeStreamX4(op1, &bitD1, opStart2, dt, dtLog);- HUF_decodeStreamX4(op2, &bitD2, opStart3, dt, dtLog);- HUF_decodeStreamX4(op3, &bitD3, opStart4, dt, dtLog);- HUF_decodeStreamX4(op4, &bitD4, oend, dt, dtLog);+ HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog);+ HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog);+ HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog);+ HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog); /* check */ { U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);@@ -733,48 +845,120 @@ } } +HUF_DGEN(HUF_decompress1X2_usingDTable_internal)+HUF_DGEN(HUF_decompress4X2_usingDTable_internal) -size_t HUF_decompress4X4_usingDTable(+size_t HUF_decompress1X2_usingDTable( void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable) { DTableDesc dtd = HUF_getDTableDesc(DTable); if (dtd.tableType != 1) return ERROR(GENERIC);- return HUF_decompress4X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);+ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0); } +size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* DCtx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize)+{+ const BYTE* ip = (const BYTE*) cSrc; -size_t HUF_decompress4X4_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+ size_t const hSize = HUF_readDTableX2_wksp(DCtx, cSrc, cSrcSize,+ workSpace, wkspSize);+ if (HUF_isError(hSize)) return hSize;+ if (hSize >= cSrcSize) return ERROR(srcSize_wrong);+ ip += hSize; cSrcSize -= hSize;++ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx, /* bmi2 */ 0);+}+++size_t HUF_decompress1X2_DCtx(HUF_DTable* DCtx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize) {+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_decompress1X2_DCtx_wksp(DCtx, dst, dstSize, cSrc, cSrcSize,+ workSpace, sizeof(workSpace));+}++size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+{+ HUF_CREATE_STATIC_DTABLEX2(DTable, HUF_TABLELOG_MAX);+ return HUF_decompress1X2_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);+}++size_t HUF_decompress4X2_usingDTable(+ void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ const HUF_DTable* DTable)+{+ DTableDesc dtd = HUF_getDTableDesc(DTable);+ if (dtd.tableType != 1) return ERROR(GENERIC);+ return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+}++static size_t HUF_decompress4X2_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize, int bmi2)+{ const BYTE* ip = (const BYTE*) cSrc; - size_t hSize = HUF_readDTableX4 (dctx, cSrc, cSrcSize);+ size_t hSize = HUF_readDTableX2_wksp(dctx, cSrc, cSrcSize,+ workSpace, wkspSize); if (HUF_isError(hSize)) return hSize; if (hSize >= cSrcSize) return ERROR(srcSize_wrong); ip += hSize; cSrcSize -= hSize; - return HUF_decompress4X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx);+ return HUF_decompress4X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2); } -size_t HUF_decompress4X4 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize) {- HUF_CREATE_STATIC_DTABLEX4(DTable, HUF_TABLELOG_MAX);- return HUF_decompress4X4_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);+ return HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, /* bmi2 */ 0); } -/* ********************************/-/* Generic decompression selector */-/* ********************************/+size_t HUF_decompress4X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize)+{+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize,+ workSpace, sizeof(workSpace));+} +size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+{+ HUF_CREATE_STATIC_DTABLEX2(DTable, HUF_TABLELOG_MAX);+ return HUF_decompress4X2_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);+}++#endif /* HUF_FORCE_DECOMPRESS_X1 */+++/* ***********************************/+/* Universal decompression selectors */+/* ***********************************/+ size_t HUF_decompress1X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable) { DTableDesc const dtd = HUF_getDTableDesc(DTable);- return dtd.tableType ? HUF_decompress1X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable) :- HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)dtd;+ assert(dtd.tableType == 0);+ return HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)dtd;+ assert(dtd.tableType == 1);+ return HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+#else+ return dtd.tableType ? HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0) :+ HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+#endif } size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize,@@ -782,11 +966,22 @@ const HUF_DTable* DTable) { DTableDesc const dtd = HUF_getDTableDesc(DTable);- return dtd.tableType ? HUF_decompress4X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable) :- HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)dtd;+ assert(dtd.tableType == 0);+ return HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)dtd;+ assert(dtd.tableType == 1);+ return HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+#else+ return dtd.tableType ? HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0) :+ HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);+#endif } +#if !defined(HUF_FORCE_DECOMPRESS_X1) && !defined(HUF_FORCE_DECOMPRESS_X2) typedef struct { U32 tableTime; U32 decode256Time; } algo_time_t; static const algo_time_t algoTime[16 /* Quantization */][3 /* single, double, quad */] = {@@ -808,22 +1003,35 @@ {{1455,128}, {2422,124}, {4174,124}}, /* Q ==14 : 87-93% */ {{ 722,128}, {1891,145}, {1936,146}}, /* Q ==15 : 93-99% */ };+#endif /** HUF_selectDecoder() :-* Tells which decoder is likely to decode faster,-* based on a set of pre-determined metrics.-* @return : 0==HUF_decompress4X2, 1==HUF_decompress4X4 .-* Assumption : 0 < cSrcSize < dstSize <= 128 KB */+ * Tells which decoder is likely to decode faster,+ * based on a set of pre-computed metrics.+ * @return : 0==HUF_decompress4X1, 1==HUF_decompress4X2 .+ * Assumption : 0 < dstSize <= 128 KB */ U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize) {+ assert(dstSize > 0);+ assert(dstSize <= 128*1024);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)dstSize;+ (void)cSrcSize;+ return 0;+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)dstSize;+ (void)cSrcSize;+ return 1;+#else /* decoder timing evaluation */- U32 const Q = (U32)(cSrcSize * 16 / dstSize); /* Q < 16 since dstSize > cSrcSize */- U32 const D256 = (U32)(dstSize >> 8);- U32 const DTime0 = algoTime[Q][0].tableTime + (algoTime[Q][0].decode256Time * D256);- U32 DTime1 = algoTime[Q][1].tableTime + (algoTime[Q][1].decode256Time * D256);- DTime1 += DTime1 >> 3; /* advantage to algorithm using less memory, for cache eviction */-- return DTime1 < DTime0;+ { U32 const Q = (cSrcSize >= dstSize) ? 15 : (U32)(cSrcSize * 16 / dstSize); /* Q < 16 */+ U32 const D256 = (U32)(dstSize >> 8);+ U32 const DTime0 = algoTime[Q][0].tableTime + (algoTime[Q][0].decode256Time * D256);+ U32 DTime1 = algoTime[Q][1].tableTime + (algoTime[Q][1].decode256Time * D256);+ DTime1 += DTime1 >> 3; /* advantage to algorithm using less memory, to reduce cache eviction */+ return DTime1 < DTime0;+ }+#endif } @@ -831,7 +1039,9 @@ size_t HUF_decompress (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize) {- static const decompressionAlgo decompress[2] = { HUF_decompress4X2, HUF_decompress4X4 };+#if !defined(HUF_FORCE_DECOMPRESS_X1) && !defined(HUF_FORCE_DECOMPRESS_X2)+ static const decompressionAlgo decompress[2] = { HUF_decompress4X1, HUF_decompress4X2 };+#endif /* validation checks */ if (dstSize == 0) return ERROR(dstSize_tooSmall);@@ -840,7 +1050,17 @@ if (cSrcSize == 1) { memset(dst, *(const BYTE*)cSrc, dstSize); return dstSize; } /* RLE */ { U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)algoNb;+ assert(algoNb == 0);+ return HUF_decompress4X1(dst, dstSize, cSrc, cSrcSize);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)algoNb;+ assert(algoNb == 1);+ return HUF_decompress4X2(dst, dstSize, cSrc, cSrcSize);+#else return decompress[algoNb](dst, dstSize, cSrc, cSrcSize);+#endif } } @@ -853,24 +1073,58 @@ if (cSrcSize == 1) { memset(dst, *(const BYTE*)cSrc, dstSize); return dstSize; } /* RLE */ { U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);- return algoNb ? HUF_decompress4X4_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) :- HUF_decompress4X2_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) ;+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)algoNb;+ assert(algoNb == 0);+ return HUF_decompress4X1_DCtx(dctx, dst, dstSize, cSrc, cSrcSize);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)algoNb;+ assert(algoNb == 1);+ return HUF_decompress4X2_DCtx(dctx, dst, dstSize, cSrc, cSrcSize);+#else+ return algoNb ? HUF_decompress4X2_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) :+ HUF_decompress4X1_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) ;+#endif } } -size_t HUF_decompress4X_hufOnly (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+size_t HUF_decompress4X_hufOnly(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize) {+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_decompress4X_hufOnly_wksp(dctx, dst, dstSize, cSrc, cSrcSize,+ workSpace, sizeof(workSpace));+}+++size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst,+ size_t dstSize, const void* cSrc,+ size_t cSrcSize, void* workSpace,+ size_t wkspSize)+{ /* validation checks */ if (dstSize == 0) return ERROR(dstSize_tooSmall);- if ((cSrcSize >= dstSize) || (cSrcSize <= 1)) return ERROR(corruption_detected); /* invalid */+ if (cSrcSize == 0) return ERROR(corruption_detected); { U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);- return algoNb ? HUF_decompress4X4_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) :- HUF_decompress4X2_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) ;+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)algoNb;+ assert(algoNb == 0);+ return HUF_decompress4X1_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)algoNb;+ assert(algoNb == 1);+ return HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize);+#else+ return algoNb ? HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc,+ cSrcSize, workSpace, wkspSize):+ HUF_decompress4X1_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize);+#endif } } -size_t HUF_decompress1X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize)+size_t HUF_decompress1X_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize,+ void* workSpace, size_t wkspSize) { /* validation checks */ if (dstSize == 0) return ERROR(dstSize_tooSmall);@@ -879,7 +1133,100 @@ if (cSrcSize == 1) { memset(dst, *(const BYTE*)cSrc, dstSize); return dstSize; } /* RLE */ { U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);- return algoNb ? HUF_decompress1X4_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) :- HUF_decompress1X2_DCtx(dctx, dst, dstSize, cSrc, cSrcSize) ;+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)algoNb;+ assert(algoNb == 0);+ return HUF_decompress1X1_DCtx_wksp(dctx, dst, dstSize, cSrc,+ cSrcSize, workSpace, wkspSize);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)algoNb;+ assert(algoNb == 1);+ return HUF_decompress1X2_DCtx_wksp(dctx, dst, dstSize, cSrc,+ cSrcSize, workSpace, wkspSize);+#else+ return algoNb ? HUF_decompress1X2_DCtx_wksp(dctx, dst, dstSize, cSrc,+ cSrcSize, workSpace, wkspSize):+ HUF_decompress1X1_DCtx_wksp(dctx, dst, dstSize, cSrc,+ cSrcSize, workSpace, wkspSize);+#endif+ }+}++size_t HUF_decompress1X_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize,+ const void* cSrc, size_t cSrcSize)+{+ U32 workSpace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];+ return HUF_decompress1X_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize,+ workSpace, sizeof(workSpace));+}+++size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2)+{+ DTableDesc const dtd = HUF_getDTableDesc(DTable);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)dtd;+ assert(dtd.tableType == 0);+ return HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)dtd;+ assert(dtd.tableType == 1);+ return HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);+#else+ return dtd.tableType ? HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2) :+ HUF_decompress1X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);+#endif+}++#ifndef HUF_FORCE_DECOMPRESS_X2+size_t HUF_decompress1X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2)+{+ const BYTE* ip = (const BYTE*) cSrc;++ size_t const hSize = HUF_readDTableX1_wksp(dctx, cSrc, cSrcSize, workSpace, wkspSize);+ if (HUF_isError(hSize)) return hSize;+ if (hSize >= cSrcSize) return ERROR(srcSize_wrong);+ ip += hSize; cSrcSize -= hSize;++ return HUF_decompress1X1_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);+}+#endif++size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2)+{+ DTableDesc const dtd = HUF_getDTableDesc(DTable);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)dtd;+ assert(dtd.tableType == 0);+ return HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)dtd;+ assert(dtd.tableType == 1);+ return HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);+#else+ return dtd.tableType ? HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2) :+ HUF_decompress4X1_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);+#endif+}++size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2)+{+ /* validation checks */+ if (dstSize == 0) return ERROR(dstSize_tooSmall);+ if (cSrcSize == 0) return ERROR(corruption_detected);++ { U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);+#if defined(HUF_FORCE_DECOMPRESS_X1)+ (void)algoNb;+ assert(algoNb == 0);+ return HUF_decompress4X1_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);+#elif defined(HUF_FORCE_DECOMPRESS_X2)+ (void)algoNb;+ assert(algoNb == 1);+ return HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);+#else+ return algoNb ? HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2) :+ HUF_decompress4X1_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);+#endif } }
+ zstd/lib/decompress/zstd_ddict.c view
@@ -0,0 +1,240 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++/* zstd_ddict.c :+ * concentrates all logic that needs to know the internals of ZSTD_DDict object */++/*-*******************************************************+* Dependencies+*********************************************************/+#include <string.h> /* memcpy, memmove, memset */+#include "cpu.h" /* bmi2 */+#include "mem.h" /* low level memory routines */+#define FSE_STATIC_LINKING_ONLY+#include "fse.h"+#define HUF_STATIC_LINKING_ONLY+#include "huf.h"+#include "zstd_decompress_internal.h"+#include "zstd_ddict.h"++#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)+# include "zstd_legacy.h"+#endif++++/*-*******************************************************+* Types+*********************************************************/+struct ZSTD_DDict_s {+ void* dictBuffer;+ const void* dictContent;+ size_t dictSize;+ ZSTD_entropyDTables_t entropy;+ U32 dictID;+ U32 entropyPresent;+ ZSTD_customMem cMem;+}; /* typedef'd to ZSTD_DDict within "zstd.h" */++const void* ZSTD_DDict_dictContent(const ZSTD_DDict* ddict)+{+ assert(ddict != NULL);+ return ddict->dictContent;+}++size_t ZSTD_DDict_dictSize(const ZSTD_DDict* ddict)+{+ assert(ddict != NULL);+ return ddict->dictSize;+}++void ZSTD_copyDDictParameters(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)+{+ DEBUGLOG(4, "ZSTD_copyDDictParameters");+ assert(dctx != NULL);+ assert(ddict != NULL);+ dctx->dictID = ddict->dictID;+ dctx->prefixStart = ddict->dictContent;+ dctx->virtualStart = ddict->dictContent;+ dctx->dictEnd = (const BYTE*)ddict->dictContent + ddict->dictSize;+ dctx->previousDstEnd = dctx->dictEnd;+ if (ddict->entropyPresent) {+ dctx->litEntropy = 1;+ dctx->fseEntropy = 1;+ dctx->LLTptr = ddict->entropy.LLTable;+ dctx->MLTptr = ddict->entropy.MLTable;+ dctx->OFTptr = ddict->entropy.OFTable;+ dctx->HUFptr = ddict->entropy.hufTable;+ dctx->entropy.rep[0] = ddict->entropy.rep[0];+ dctx->entropy.rep[1] = ddict->entropy.rep[1];+ dctx->entropy.rep[2] = ddict->entropy.rep[2];+ } else {+ dctx->litEntropy = 0;+ dctx->fseEntropy = 0;+ }+}+++static size_t+ZSTD_loadEntropy_intoDDict(ZSTD_DDict* ddict,+ ZSTD_dictContentType_e dictContentType)+{+ ddict->dictID = 0;+ ddict->entropyPresent = 0;+ if (dictContentType == ZSTD_dct_rawContent) return 0;++ if (ddict->dictSize < 8) {+ if (dictContentType == ZSTD_dct_fullDict)+ return ERROR(dictionary_corrupted); /* only accept specified dictionaries */+ return 0; /* pure content mode */+ }+ { U32 const magic = MEM_readLE32(ddict->dictContent);+ if (magic != ZSTD_MAGIC_DICTIONARY) {+ if (dictContentType == ZSTD_dct_fullDict)+ return ERROR(dictionary_corrupted); /* only accept specified dictionaries */+ return 0; /* pure content mode */+ }+ }+ ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_FRAMEIDSIZE);++ /* load entropy tables */+ RETURN_ERROR_IF(ZSTD_isError(ZSTD_loadDEntropy(+ &ddict->entropy, ddict->dictContent, ddict->dictSize)),+ dictionary_corrupted);+ ddict->entropyPresent = 1;+ return 0;+}+++static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict,+ const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType)+{+ if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dict) || (!dictSize)) {+ ddict->dictBuffer = NULL;+ ddict->dictContent = dict;+ if (!dict) dictSize = 0;+ } else {+ void* const internalBuffer = ZSTD_malloc(dictSize, ddict->cMem);+ ddict->dictBuffer = internalBuffer;+ ddict->dictContent = internalBuffer;+ if (!internalBuffer) return ERROR(memory_allocation);+ memcpy(internalBuffer, dict, dictSize);+ }+ ddict->dictSize = dictSize;+ ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */++ /* parse dictionary content */+ FORWARD_IF_ERROR( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) );++ return 0;+}++ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_customMem customMem)+{+ if (!customMem.customAlloc ^ !customMem.customFree) return NULL;++ { ZSTD_DDict* const ddict = (ZSTD_DDict*) ZSTD_malloc(sizeof(ZSTD_DDict), customMem);+ if (ddict == NULL) return NULL;+ ddict->cMem = customMem;+ { size_t const initResult = ZSTD_initDDict_internal(ddict,+ dict, dictSize,+ dictLoadMethod, dictContentType);+ if (ZSTD_isError(initResult)) {+ ZSTD_freeDDict(ddict);+ return NULL;+ } }+ return ddict;+ }+}++/*! ZSTD_createDDict() :+* Create a digested dictionary, to start decompression without startup delay.+* `dict` content is copied inside DDict.+* Consequently, `dict` can be released after `ZSTD_DDict` creation */+ZSTD_DDict* ZSTD_createDDict(const void* dict, size_t dictSize)+{+ ZSTD_customMem const allocator = { NULL, NULL, NULL };+ return ZSTD_createDDict_advanced(dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto, allocator);+}++/*! ZSTD_createDDict_byReference() :+ * Create a digested dictionary, to start decompression without startup delay.+ * Dictionary content is simply referenced, it will be accessed during decompression.+ * Warning : dictBuffer must outlive DDict (DDict must be freed before dictBuffer) */+ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize)+{+ ZSTD_customMem const allocator = { NULL, NULL, NULL };+ return ZSTD_createDDict_advanced(dictBuffer, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto, allocator);+}+++const ZSTD_DDict* ZSTD_initStaticDDict(+ void* sBuffer, size_t sBufferSize,+ const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType)+{+ size_t const neededSpace = sizeof(ZSTD_DDict)+ + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);+ ZSTD_DDict* const ddict = (ZSTD_DDict*)sBuffer;+ assert(sBuffer != NULL);+ assert(dict != NULL);+ if ((size_t)sBuffer & 7) return NULL; /* 8-aligned */+ if (sBufferSize < neededSpace) return NULL;+ if (dictLoadMethod == ZSTD_dlm_byCopy) {+ memcpy(ddict+1, dict, dictSize); /* local copy */+ dict = ddict+1;+ }+ if (ZSTD_isError( ZSTD_initDDict_internal(ddict,+ dict, dictSize,+ ZSTD_dlm_byRef, dictContentType) ))+ return NULL;+ return ddict;+}+++size_t ZSTD_freeDDict(ZSTD_DDict* ddict)+{+ if (ddict==NULL) return 0; /* support free on NULL */+ { ZSTD_customMem const cMem = ddict->cMem;+ ZSTD_free(ddict->dictBuffer, cMem);+ ZSTD_free(ddict, cMem);+ return 0;+ }+}++/*! ZSTD_estimateDDictSize() :+ * Estimate amount of memory that will be needed to create a dictionary for decompression.+ * Note : dictionary created by reference using ZSTD_dlm_byRef are smaller */+size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod)+{+ return sizeof(ZSTD_DDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);+}++size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict)+{+ if (ddict==NULL) return 0; /* support sizeof on NULL */+ return sizeof(*ddict) + (ddict->dictBuffer ? ddict->dictSize : 0) ;+}++/*! ZSTD_getDictID_fromDDict() :+ * Provides the dictID of the dictionary loaded into `ddict`.+ * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.+ * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */+unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict)+{+ if (ddict==NULL) return 0;+ return ZSTD_getDictID_fromDict(ddict->dictContent, ddict->dictSize);+}
+ zstd/lib/decompress/zstd_ddict.h view
@@ -0,0 +1,44 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+++#ifndef ZSTD_DDICT_H+#define ZSTD_DDICT_H++/*-*******************************************************+ * Dependencies+ *********************************************************/+#include <stddef.h> /* size_t */+#include "zstd.h" /* ZSTD_DDict, and several public functions */+++/*-*******************************************************+ * Interface+ *********************************************************/++/* note: several prototypes are already published in `zstd.h` :+ * ZSTD_createDDict()+ * ZSTD_createDDict_byReference()+ * ZSTD_createDDict_advanced()+ * ZSTD_freeDDict()+ * ZSTD_initStaticDDict()+ * ZSTD_sizeof_DDict()+ * ZSTD_estimateDDictSize()+ * ZSTD_getDictID_fromDict()+ */++const void* ZSTD_DDict_dictContent(const ZSTD_DDict* ddict);+size_t ZSTD_DDict_dictSize(const ZSTD_DDict* ddict);++void ZSTD_copyDDictParameters(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);++++#endif /* ZSTD_DDICT_H */
zstd/lib/decompress/zstd_decompress.c view
@@ -1,2171 +1,1768 @@-/**- * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.- * All rights reserved.- *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.- */---/* ***************************************************************-* Tuning parameters-*****************************************************************/-/*!- * HEAPMODE :- * Select how default decompression function ZSTD_decompress() will allocate memory,- * in memory stack (0), or in memory heap (1, requires malloc())- */-#ifndef ZSTD_HEAPMODE-# define ZSTD_HEAPMODE 1-#endif--/*!-* LEGACY_SUPPORT :-* if set to 1, ZSTD_decompress() can decode older formats (v0.1+)-*/-#ifndef ZSTD_LEGACY_SUPPORT-# define ZSTD_LEGACY_SUPPORT 0-#endif--/*!-* MAXWINDOWSIZE_DEFAULT :-* maximum window size accepted by DStream, by default.-* Frames requiring more memory will be rejected.-*/-#ifndef ZSTD_MAXWINDOWSIZE_DEFAULT-# define ZSTD_MAXWINDOWSIZE_DEFAULT ((1 << ZSTD_WINDOWLOG_MAX) + 1) /* defined within zstd.h */-#endif---/*-*******************************************************-* Dependencies-*********************************************************/-#include <string.h> /* memcpy, memmove, memset */-#include "mem.h" /* low level memory routines */-#define XXH_STATIC_LINKING_ONLY /* XXH64_state_t */-#include "xxhash.h" /* XXH64_* */-#define FSE_STATIC_LINKING_ONLY-#include "fse.h"-#define HUF_STATIC_LINKING_ONLY-#include "huf.h"-#include "zstd_internal.h"--#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)-# include "zstd_legacy.h"-#endif---#if defined(_MSC_VER)-# include <mmintrin.h> /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */-# define ZSTD_PREFETCH(ptr) _mm_prefetch((const char*)ptr, _MM_HINT_T0)-#elif defined(__GNUC__)-# define ZSTD_PREFETCH(ptr) __builtin_prefetch(ptr, 0, 0)-#else-# define ZSTD_PREFETCH(ptr) /* disabled */-#endif--/*-*************************************-* Macros-***************************************/-#define ZSTD_isError ERR_isError /* for inlining */-#define FSE_isError ERR_isError-#define HUF_isError ERR_isError---/*_*******************************************************-* Memory operations-**********************************************************/-static void ZSTD_copy4(void* dst, const void* src) { memcpy(dst, src, 4); }---/*-*************************************************************-* Context management-***************************************************************/-typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader,- ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock,- ZSTDds_decompressLastBlock, ZSTDds_checkChecksum,- ZSTDds_decodeSkippableHeader, ZSTDds_skipFrame } ZSTD_dStage;--struct ZSTD_DCtx_s-{- const FSE_DTable* LLTptr;- const FSE_DTable* MLTptr;- const FSE_DTable* OFTptr;- const HUF_DTable* HUFptr;- FSE_DTable LLTable[FSE_DTABLE_SIZE_U32(LLFSELog)];- FSE_DTable OFTable[FSE_DTABLE_SIZE_U32(OffFSELog)];- FSE_DTable MLTable[FSE_DTABLE_SIZE_U32(MLFSELog)];- HUF_DTable hufTable[HUF_DTABLE_SIZE(HufLog)]; /* can accommodate HUF_decompress4X */- const void* previousDstEnd;- const void* base;- const void* vBase;- const void* dictEnd;- size_t expected;- U32 rep[ZSTD_REP_NUM];- ZSTD_frameParams fParams;- blockType_e bType; /* used in ZSTD_decompressContinue(), to transfer blockType between header decoding and block decoding stages */- ZSTD_dStage stage;- U32 litEntropy;- U32 fseEntropy;- XXH64_state_t xxhState;- size_t headerSize;- U32 dictID;- const BYTE* litPtr;- ZSTD_customMem customMem;- size_t litSize;- size_t rleSize;- BYTE litBuffer[ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH];- BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];-}; /* typedef'd to ZSTD_DCtx within "zstd.h" */--size_t ZSTD_sizeof_DCtx (const ZSTD_DCtx* dctx) { return (dctx==NULL) ? 0 : sizeof(ZSTD_DCtx); }--size_t ZSTD_estimateDCtxSize(void) { return sizeof(ZSTD_DCtx); }--size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)-{- dctx->expected = ZSTD_frameHeaderSize_prefix;- dctx->stage = ZSTDds_getFrameHeaderSize;- dctx->previousDstEnd = NULL;- dctx->base = NULL;- dctx->vBase = NULL;- dctx->dictEnd = NULL;- dctx->hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */- dctx->litEntropy = dctx->fseEntropy = 0;- dctx->dictID = 0;- MEM_STATIC_ASSERT(sizeof(dctx->rep) == sizeof(repStartValue));- memcpy(dctx->rep, repStartValue, sizeof(repStartValue)); /* initial repcodes */- dctx->LLTptr = dctx->LLTable;- dctx->MLTptr = dctx->MLTable;- dctx->OFTptr = dctx->OFTable;- dctx->HUFptr = dctx->hufTable;- return 0;-}--ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem)-{- ZSTD_DCtx* dctx;-- if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;- if (!customMem.customAlloc || !customMem.customFree) return NULL;-- dctx = (ZSTD_DCtx*)ZSTD_malloc(sizeof(ZSTD_DCtx), customMem);- if (!dctx) return NULL;- memcpy(&dctx->customMem, &customMem, sizeof(customMem));- ZSTD_decompressBegin(dctx);- return dctx;-}--ZSTD_DCtx* ZSTD_createDCtx(void)-{- return ZSTD_createDCtx_advanced(defaultCustomMem);-}--size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx)-{- if (dctx==NULL) return 0; /* support free on NULL */- ZSTD_free(dctx, dctx->customMem);- return 0; /* reserved as a potential error code in the future */-}--void ZSTD_copyDCtx(ZSTD_DCtx* dstDCtx, const ZSTD_DCtx* srcDCtx)-{- size_t const workSpaceSize = (ZSTD_BLOCKSIZE_ABSOLUTEMAX+WILDCOPY_OVERLENGTH) + ZSTD_frameHeaderSize_max;- memcpy(dstDCtx, srcDCtx, sizeof(ZSTD_DCtx) - workSpaceSize); /* no need to copy workspace */-}--static void ZSTD_refDCtx(ZSTD_DCtx* dstDCtx, const ZSTD_DCtx* srcDCtx)-{- ZSTD_decompressBegin(dstDCtx); /* init */- if (srcDCtx) { /* support refDCtx on NULL */- dstDCtx->dictEnd = srcDCtx->dictEnd;- dstDCtx->vBase = srcDCtx->vBase;- dstDCtx->base = srcDCtx->base;- dstDCtx->previousDstEnd = srcDCtx->previousDstEnd;- dstDCtx->dictID = srcDCtx->dictID;- dstDCtx->litEntropy = srcDCtx->litEntropy;- dstDCtx->fseEntropy = srcDCtx->fseEntropy;- dstDCtx->LLTptr = srcDCtx->LLTable;- dstDCtx->MLTptr = srcDCtx->MLTable;- dstDCtx->OFTptr = srcDCtx->OFTable;- dstDCtx->HUFptr = srcDCtx->hufTable;- dstDCtx->rep[0] = srcDCtx->rep[0];- dstDCtx->rep[1] = srcDCtx->rep[1];- dstDCtx->rep[2] = srcDCtx->rep[2];- }-}---/*-*************************************************************-* Decompression section-***************************************************************/--/*! ZSTD_isFrame() :- * Tells if the content of `buffer` starts with a valid Frame Identifier.- * Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.- * Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.- * Note 3 : Skippable Frame Identifiers are considered valid. */-unsigned ZSTD_isFrame(const void* buffer, size_t size)-{- if (size < 4) return 0;- { U32 const magic = MEM_readLE32(buffer);- if (magic == ZSTD_MAGICNUMBER) return 1;- if ((magic & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) return 1;- }-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)- if (ZSTD_isLegacy(buffer, size)) return 1;-#endif- return 0;-}---/** ZSTD_frameHeaderSize() :-* srcSize must be >= ZSTD_frameHeaderSize_prefix.-* @return : size of the Frame Header */-static size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize)-{- if (srcSize < ZSTD_frameHeaderSize_prefix) return ERROR(srcSize_wrong);- { BYTE const fhd = ((const BYTE*)src)[4];- U32 const dictID= fhd & 3;- U32 const singleSegment = (fhd >> 5) & 1;- U32 const fcsId = fhd >> 6;- return ZSTD_frameHeaderSize_prefix + !singleSegment + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId]- + (singleSegment && !fcsId);- }-}---/** ZSTD_getFrameParams() :-* decode Frame Header, or require larger `srcSize`.-* @return : 0, `fparamsPtr` is correctly filled,-* >0, `srcSize` is too small, result is expected `srcSize`,-* or an error code, which can be tested using ZSTD_isError() */-size_t ZSTD_getFrameParams(ZSTD_frameParams* fparamsPtr, const void* src, size_t srcSize)-{- const BYTE* ip = (const BYTE*)src;-- if (srcSize < ZSTD_frameHeaderSize_prefix) return ZSTD_frameHeaderSize_prefix;- if (MEM_readLE32(src) != ZSTD_MAGICNUMBER) {- if ((MEM_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {- if (srcSize < ZSTD_skippableHeaderSize) return ZSTD_skippableHeaderSize; /* magic number + skippable frame length */- memset(fparamsPtr, 0, sizeof(*fparamsPtr));- fparamsPtr->frameContentSize = MEM_readLE32((const char *)src + 4);- fparamsPtr->windowSize = 0; /* windowSize==0 means a frame is skippable */- return 0;- }- return ERROR(prefix_unknown);- }-- /* ensure there is enough `srcSize` to fully read/decode frame header */- { size_t const fhsize = ZSTD_frameHeaderSize(src, srcSize);- if (srcSize < fhsize) return fhsize; }-- { BYTE const fhdByte = ip[4];- size_t pos = 5;- U32 const dictIDSizeCode = fhdByte&3;- U32 const checksumFlag = (fhdByte>>2)&1;- U32 const singleSegment = (fhdByte>>5)&1;- U32 const fcsID = fhdByte>>6;- U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX;- U32 windowSize = 0;- U32 dictID = 0;- U64 frameContentSize = 0;- if ((fhdByte & 0x08) != 0) return ERROR(frameParameter_unsupported); /* reserved bits, which must be zero */- if (!singleSegment) {- BYTE const wlByte = ip[pos++];- U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;- if (windowLog > ZSTD_WINDOWLOG_MAX) return ERROR(frameParameter_windowTooLarge); /* avoids issue with 1 << windowLog */- windowSize = (1U << windowLog);- windowSize += (windowSize >> 3) * (wlByte&7);- }-- switch(dictIDSizeCode)- {- default: /* impossible */- case 0 : break;- case 1 : dictID = ip[pos]; pos++; break;- case 2 : dictID = MEM_readLE16(ip+pos); pos+=2; break;- case 3 : dictID = MEM_readLE32(ip+pos); pos+=4; break;- }- switch(fcsID)- {- default: /* impossible */- case 0 : if (singleSegment) frameContentSize = ip[pos]; break;- case 1 : frameContentSize = MEM_readLE16(ip+pos)+256; break;- case 2 : frameContentSize = MEM_readLE32(ip+pos); break;- case 3 : frameContentSize = MEM_readLE64(ip+pos); break;- }- if (!windowSize) windowSize = (U32)frameContentSize;- if (windowSize > windowSizeMax) return ERROR(frameParameter_windowTooLarge);- fparamsPtr->frameContentSize = frameContentSize;- fparamsPtr->windowSize = windowSize;- fparamsPtr->dictID = dictID;- fparamsPtr->checksumFlag = checksumFlag;- }- return 0;-}---/** ZSTD_getDecompressedSize() :-* compatible with legacy mode-* @return : decompressed size if known, 0 otherwise- note : 0 can mean any of the following :- - decompressed size is not present within frame header- - frame header unknown / not supported- - frame header not complete (`srcSize` too small) */-unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize)-{-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT==1)- if (ZSTD_isLegacy(src, srcSize)) return ZSTD_getDecompressedSize_legacy(src, srcSize);-#endif- { ZSTD_frameParams fparams;- size_t const frResult = ZSTD_getFrameParams(&fparams, src, srcSize);- if (frResult!=0) return 0;- return fparams.frameContentSize;- }-}---/** ZSTD_decodeFrameHeader() :-* `headerSize` must be the size provided by ZSTD_frameHeaderSize().-* @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */-static size_t ZSTD_decodeFrameHeader(ZSTD_DCtx* dctx, const void* src, size_t headerSize)-{- size_t const result = ZSTD_getFrameParams(&(dctx->fParams), src, headerSize);- if (ZSTD_isError(result)) return result; /* invalid header */- if (result>0) return ERROR(srcSize_wrong); /* headerSize too small */- if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID)) return ERROR(dictionary_wrong);- if (dctx->fParams.checksumFlag) XXH64_reset(&dctx->xxhState, 0);- return 0;-}---typedef struct-{- blockType_e blockType;- U32 lastBlock;- U32 origSize;-} blockProperties_t;--/*! ZSTD_getcBlockSize() :-* Provides the size of compressed block from block header `src` */-size_t ZSTD_getcBlockSize(const void* src, size_t srcSize, blockProperties_t* bpPtr)-{- if (srcSize < ZSTD_blockHeaderSize) return ERROR(srcSize_wrong);- { U32 const cBlockHeader = MEM_readLE24(src);- U32 const cSize = cBlockHeader >> 3;- bpPtr->lastBlock = cBlockHeader & 1;- bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);- bpPtr->origSize = cSize; /* only useful for RLE */- if (bpPtr->blockType == bt_rle) return 1;- if (bpPtr->blockType == bt_reserved) return ERROR(corruption_detected);- return cSize;- }-}---static size_t ZSTD_copyRawBlock(void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- if (srcSize > dstCapacity) return ERROR(dstSize_tooSmall);- memcpy(dst, src, srcSize);- return srcSize;-}---static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity, const void* src, size_t srcSize, size_t regenSize)-{- if (srcSize != 1) return ERROR(srcSize_wrong);- if (regenSize > dstCapacity) return ERROR(dstSize_tooSmall);- memset(dst, *(const BYTE*)src, regenSize);- return regenSize;-}--/*! ZSTD_decodeLiteralsBlock() :- @return : nb of bytes read from src (< srcSize ) */-size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,- const void* src, size_t srcSize) /* note : srcSize < BLOCKSIZE */-{- if (srcSize < MIN_CBLOCK_SIZE) return ERROR(corruption_detected);-- { const BYTE* const istart = (const BYTE*) src;- symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);-- switch(litEncType)- {- case set_repeat:- if (dctx->litEntropy==0) return ERROR(dictionary_corrupted);- /* fall-through */- case set_compressed:- if (srcSize < 5) return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3 */- { size_t lhSize, litSize, litCSize;- U32 singleStream=0;- U32 const lhlCode = (istart[0] >> 2) & 3;- U32 const lhc = MEM_readLE32(istart);- switch(lhlCode)- {- case 0: case 1: default: /* note : default is impossible, since lhlCode into [0..3] */- /* 2 - 2 - 10 - 10 */- singleStream = !lhlCode;- lhSize = 3;- litSize = (lhc >> 4) & 0x3FF;- litCSize = (lhc >> 14) & 0x3FF;- break;- case 2:- /* 2 - 2 - 14 - 14 */- lhSize = 4;- litSize = (lhc >> 4) & 0x3FFF;- litCSize = lhc >> 18;- break;- case 3:- /* 2 - 2 - 18 - 18 */- lhSize = 5;- litSize = (lhc >> 4) & 0x3FFFF;- litCSize = (lhc >> 22) + (istart[4] << 10);- break;- }- if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX) return ERROR(corruption_detected);- if (litCSize + lhSize > srcSize) return ERROR(corruption_detected);-- if (HUF_isError((litEncType==set_repeat) ?- ( singleStream ?- HUF_decompress1X_usingDTable(dctx->litBuffer, litSize, istart+lhSize, litCSize, dctx->HUFptr) :- HUF_decompress4X_usingDTable(dctx->litBuffer, litSize, istart+lhSize, litCSize, dctx->HUFptr) ) :- ( singleStream ?- HUF_decompress1X2_DCtx(dctx->hufTable, dctx->litBuffer, litSize, istart+lhSize, litCSize) :- HUF_decompress4X_hufOnly (dctx->hufTable, dctx->litBuffer, litSize, istart+lhSize, litCSize)) ))- return ERROR(corruption_detected);-- dctx->litPtr = dctx->litBuffer;- dctx->litSize = litSize;- dctx->litEntropy = 1;- if (litEncType==set_compressed) dctx->HUFptr = dctx->hufTable;- memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);- return litCSize + lhSize;- }-- case set_basic:- { size_t litSize, lhSize;- U32 const lhlCode = ((istart[0]) >> 2) & 3;- switch(lhlCode)- {- case 0: case 2: default: /* note : default is impossible, since lhlCode into [0..3] */- lhSize = 1;- litSize = istart[0] >> 3;- break;- case 1:- lhSize = 2;- litSize = MEM_readLE16(istart) >> 4;- break;- case 3:- lhSize = 3;- litSize = MEM_readLE24(istart) >> 4;- break;- }-- if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */- if (litSize+lhSize > srcSize) return ERROR(corruption_detected);- memcpy(dctx->litBuffer, istart+lhSize, litSize);- dctx->litPtr = dctx->litBuffer;- dctx->litSize = litSize;- memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);- return lhSize+litSize;- }- /* direct reference into compressed stream */- dctx->litPtr = istart+lhSize;- dctx->litSize = litSize;- return lhSize+litSize;- }-- case set_rle:- { U32 const lhlCode = ((istart[0]) >> 2) & 3;- size_t litSize, lhSize;- switch(lhlCode)- {- case 0: case 2: default: /* note : default is impossible, since lhlCode into [0..3] */- lhSize = 1;- litSize = istart[0] >> 3;- break;- case 1:- lhSize = 2;- litSize = MEM_readLE16(istart) >> 4;- break;- case 3:- lhSize = 3;- litSize = MEM_readLE24(istart) >> 4;- if (srcSize<4) return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4 */- break;- }- if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX) return ERROR(corruption_detected);- memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);- dctx->litPtr = dctx->litBuffer;- dctx->litSize = litSize;- return lhSize+1;- }- default:- return ERROR(corruption_detected); /* impossible */- }- }-}---typedef union {- FSE_decode_t realData;- U32 alignedBy4;-} FSE_decode_t4;--static const FSE_decode_t4 LL_defaultDTable[(1<<LL_DEFAULTNORMLOG)+1] = {- { { LL_DEFAULTNORMLOG, 1, 1 } }, /* header : tableLog, fastMode, fastMode */- { { 0, 0, 4 } }, /* 0 : base, symbol, bits */- { { 16, 0, 4 } },- { { 32, 1, 5 } },- { { 0, 3, 5 } },- { { 0, 4, 5 } },- { { 0, 6, 5 } },- { { 0, 7, 5 } },- { { 0, 9, 5 } },- { { 0, 10, 5 } },- { { 0, 12, 5 } },- { { 0, 14, 6 } },- { { 0, 16, 5 } },- { { 0, 18, 5 } },- { { 0, 19, 5 } },- { { 0, 21, 5 } },- { { 0, 22, 5 } },- { { 0, 24, 5 } },- { { 32, 25, 5 } },- { { 0, 26, 5 } },- { { 0, 27, 6 } },- { { 0, 29, 6 } },- { { 0, 31, 6 } },- { { 32, 0, 4 } },- { { 0, 1, 4 } },- { { 0, 2, 5 } },- { { 32, 4, 5 } },- { { 0, 5, 5 } },- { { 32, 7, 5 } },- { { 0, 8, 5 } },- { { 32, 10, 5 } },- { { 0, 11, 5 } },- { { 0, 13, 6 } },- { { 32, 16, 5 } },- { { 0, 17, 5 } },- { { 32, 19, 5 } },- { { 0, 20, 5 } },- { { 32, 22, 5 } },- { { 0, 23, 5 } },- { { 0, 25, 4 } },- { { 16, 25, 4 } },- { { 32, 26, 5 } },- { { 0, 28, 6 } },- { { 0, 30, 6 } },- { { 48, 0, 4 } },- { { 16, 1, 4 } },- { { 32, 2, 5 } },- { { 32, 3, 5 } },- { { 32, 5, 5 } },- { { 32, 6, 5 } },- { { 32, 8, 5 } },- { { 32, 9, 5 } },- { { 32, 11, 5 } },- { { 32, 12, 5 } },- { { 0, 15, 6 } },- { { 32, 17, 5 } },- { { 32, 18, 5 } },- { { 32, 20, 5 } },- { { 32, 21, 5 } },- { { 32, 23, 5 } },- { { 32, 24, 5 } },- { { 0, 35, 6 } },- { { 0, 34, 6 } },- { { 0, 33, 6 } },- { { 0, 32, 6 } },-}; /* LL_defaultDTable */--static const FSE_decode_t4 ML_defaultDTable[(1<<ML_DEFAULTNORMLOG)+1] = {- { { ML_DEFAULTNORMLOG, 1, 1 } }, /* header : tableLog, fastMode, fastMode */- { { 0, 0, 6 } }, /* 0 : base, symbol, bits */- { { 0, 1, 4 } },- { { 32, 2, 5 } },- { { 0, 3, 5 } },- { { 0, 5, 5 } },- { { 0, 6, 5 } },- { { 0, 8, 5 } },- { { 0, 10, 6 } },- { { 0, 13, 6 } },- { { 0, 16, 6 } },- { { 0, 19, 6 } },- { { 0, 22, 6 } },- { { 0, 25, 6 } },- { { 0, 28, 6 } },- { { 0, 31, 6 } },- { { 0, 33, 6 } },- { { 0, 35, 6 } },- { { 0, 37, 6 } },- { { 0, 39, 6 } },- { { 0, 41, 6 } },- { { 0, 43, 6 } },- { { 0, 45, 6 } },- { { 16, 1, 4 } },- { { 0, 2, 4 } },- { { 32, 3, 5 } },- { { 0, 4, 5 } },- { { 32, 6, 5 } },- { { 0, 7, 5 } },- { { 0, 9, 6 } },- { { 0, 12, 6 } },- { { 0, 15, 6 } },- { { 0, 18, 6 } },- { { 0, 21, 6 } },- { { 0, 24, 6 } },- { { 0, 27, 6 } },- { { 0, 30, 6 } },- { { 0, 32, 6 } },- { { 0, 34, 6 } },- { { 0, 36, 6 } },- { { 0, 38, 6 } },- { { 0, 40, 6 } },- { { 0, 42, 6 } },- { { 0, 44, 6 } },- { { 32, 1, 4 } },- { { 48, 1, 4 } },- { { 16, 2, 4 } },- { { 32, 4, 5 } },- { { 32, 5, 5 } },- { { 32, 7, 5 } },- { { 32, 8, 5 } },- { { 0, 11, 6 } },- { { 0, 14, 6 } },- { { 0, 17, 6 } },- { { 0, 20, 6 } },- { { 0, 23, 6 } },- { { 0, 26, 6 } },- { { 0, 29, 6 } },- { { 0, 52, 6 } },- { { 0, 51, 6 } },- { { 0, 50, 6 } },- { { 0, 49, 6 } },- { { 0, 48, 6 } },- { { 0, 47, 6 } },- { { 0, 46, 6 } },-}; /* ML_defaultDTable */--static const FSE_decode_t4 OF_defaultDTable[(1<<OF_DEFAULTNORMLOG)+1] = {- { { OF_DEFAULTNORMLOG, 1, 1 } }, /* header : tableLog, fastMode, fastMode */- { { 0, 0, 5 } }, /* 0 : base, symbol, bits */- { { 0, 6, 4 } },- { { 0, 9, 5 } },- { { 0, 15, 5 } },- { { 0, 21, 5 } },- { { 0, 3, 5 } },- { { 0, 7, 4 } },- { { 0, 12, 5 } },- { { 0, 18, 5 } },- { { 0, 23, 5 } },- { { 0, 5, 5 } },- { { 0, 8, 4 } },- { { 0, 14, 5 } },- { { 0, 20, 5 } },- { { 0, 2, 5 } },- { { 16, 7, 4 } },- { { 0, 11, 5 } },- { { 0, 17, 5 } },- { { 0, 22, 5 } },- { { 0, 4, 5 } },- { { 16, 8, 4 } },- { { 0, 13, 5 } },- { { 0, 19, 5 } },- { { 0, 1, 5 } },- { { 16, 6, 4 } },- { { 0, 10, 5 } },- { { 0, 16, 5 } },- { { 0, 28, 5 } },- { { 0, 27, 5 } },- { { 0, 26, 5 } },- { { 0, 25, 5 } },- { { 0, 24, 5 } },-}; /* OF_defaultDTable */--/*! ZSTD_buildSeqTable() :- @return : nb bytes read from src,- or an error code if it fails, testable with ZSTD_isError()-*/-static size_t ZSTD_buildSeqTable(FSE_DTable* DTableSpace, const FSE_DTable** DTablePtr,- symbolEncodingType_e type, U32 max, U32 maxLog,- const void* src, size_t srcSize,- const FSE_decode_t4* defaultTable, U32 flagRepeatTable)-{- const void* const tmpPtr = defaultTable; /* bypass strict aliasing */- switch(type)- {- case set_rle :- if (!srcSize) return ERROR(srcSize_wrong);- if ( (*(const BYTE*)src) > max) return ERROR(corruption_detected);- FSE_buildDTable_rle(DTableSpace, *(const BYTE*)src);- *DTablePtr = DTableSpace;- return 1;- case set_basic :- *DTablePtr = (const FSE_DTable*)tmpPtr;- return 0;- case set_repeat:- if (!flagRepeatTable) return ERROR(corruption_detected);- return 0;- default : /* impossible */- case set_compressed :- { U32 tableLog;- S16 norm[MaxSeq+1];- size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);- if (FSE_isError(headerSize)) return ERROR(corruption_detected);- if (tableLog > maxLog) return ERROR(corruption_detected);- FSE_buildDTable(DTableSpace, norm, max, tableLog);- *DTablePtr = DTableSpace;- return headerSize;- } }-}--size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,- const void* src, size_t srcSize)-{- const BYTE* const istart = (const BYTE* const)src;- const BYTE* const iend = istart + srcSize;- const BYTE* ip = istart;-- /* check */- if (srcSize < MIN_SEQUENCES_SIZE) return ERROR(srcSize_wrong);-- /* SeqHead */- { int nbSeq = *ip++;- if (!nbSeq) { *nbSeqPtr=0; return 1; }- if (nbSeq > 0x7F) {- if (nbSeq == 0xFF) {- if (ip+2 > iend) return ERROR(srcSize_wrong);- nbSeq = MEM_readLE16(ip) + LONGNBSEQ, ip+=2;- } else {- if (ip >= iend) return ERROR(srcSize_wrong);- nbSeq = ((nbSeq-0x80)<<8) + *ip++;- }- }- *nbSeqPtr = nbSeq;- }-- /* FSE table descriptors */- if (ip+4 > iend) return ERROR(srcSize_wrong); /* minimum possible size */- { symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);- symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);- symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);- ip++;-- /* Build DTables */- { size_t const llhSize = ZSTD_buildSeqTable(dctx->LLTable, &dctx->LLTptr,- LLtype, MaxLL, LLFSELog,- ip, iend-ip, LL_defaultDTable, dctx->fseEntropy);- if (ZSTD_isError(llhSize)) return ERROR(corruption_detected);- ip += llhSize;- }- { size_t const ofhSize = ZSTD_buildSeqTable(dctx->OFTable, &dctx->OFTptr,- OFtype, MaxOff, OffFSELog,- ip, iend-ip, OF_defaultDTable, dctx->fseEntropy);- if (ZSTD_isError(ofhSize)) return ERROR(corruption_detected);- ip += ofhSize;- }- { size_t const mlhSize = ZSTD_buildSeqTable(dctx->MLTable, &dctx->MLTptr,- MLtype, MaxML, MLFSELog,- ip, iend-ip, ML_defaultDTable, dctx->fseEntropy);- if (ZSTD_isError(mlhSize)) return ERROR(corruption_detected);- ip += mlhSize;- }- }-- return ip-istart;-}---typedef struct {- size_t litLength;- size_t matchLength;- size_t offset;- const BYTE* match;-} seq_t;--typedef struct {- BIT_DStream_t DStream;- FSE_DState_t stateLL;- FSE_DState_t stateOffb;- FSE_DState_t stateML;- size_t prevOffset[ZSTD_REP_NUM];- const BYTE* base;- size_t pos;- iPtrDiff gotoDict;-} seqState_t;---FORCE_NOINLINE-size_t ZSTD_execSequenceLast7(BYTE* op,- BYTE* const oend, seq_t sequence,- const BYTE** litPtr, const BYTE* const litLimit,- const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)-{- BYTE* const oLitEnd = op + sequence.litLength;- size_t const sequenceLength = sequence.litLength + sequence.matchLength;- BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */- BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;- const BYTE* const iLitEnd = *litPtr + sequence.litLength;- const BYTE* match = oLitEnd - sequence.offset;-- /* check */- if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */- if (iLitEnd > litLimit) return ERROR(corruption_detected); /* over-read beyond lit buffer */- if (oLitEnd <= oend_w) return ERROR(GENERIC); /* Precondition */-- /* copy literals */- if (op < oend_w) {- ZSTD_wildcopy(op, *litPtr, oend_w - op);- *litPtr += oend_w - op;- op = oend_w;- }- while (op < oLitEnd) *op++ = *(*litPtr)++;-- /* copy Match */- if (sequence.offset > (size_t)(oLitEnd - base)) {- /* offset beyond prefix */- if (sequence.offset > (size_t)(oLitEnd - vBase)) return ERROR(corruption_detected);- match = dictEnd - (base-match);- if (match + sequence.matchLength <= dictEnd) {- memmove(oLitEnd, match, sequence.matchLength);- return sequenceLength;- }- /* span extDict & currentPrefixSegment */- { size_t const length1 = dictEnd - match;- memmove(oLitEnd, match, length1);- op = oLitEnd + length1;- sequence.matchLength -= length1;- match = base;- } }- while (op < oMatchEnd) *op++ = *match++;- return sequenceLength;-}-----static seq_t ZSTD_decodeSequence(seqState_t* seqState)-{- seq_t seq;-- U32 const llCode = FSE_peekSymbol(&seqState->stateLL);- U32 const mlCode = FSE_peekSymbol(&seqState->stateML);- U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */-- U32 const llBits = LL_bits[llCode];- U32 const mlBits = ML_bits[mlCode];- U32 const ofBits = ofCode;- U32 const totalBits = llBits+mlBits+ofBits;-- static const U32 LL_base[MaxLL+1] = {- 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,- 16, 18, 20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000,- 0x2000, 0x4000, 0x8000, 0x10000 };-- static const U32 ML_base[MaxML+1] = {- 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,- 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,- 35, 37, 39, 41, 43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803,- 0x1003, 0x2003, 0x4003, 0x8003, 0x10003 };-- static const U32 OF_base[MaxOff+1] = {- 0, 1, 1, 5, 0xD, 0x1D, 0x3D, 0x7D,- 0xFD, 0x1FD, 0x3FD, 0x7FD, 0xFFD, 0x1FFD, 0x3FFD, 0x7FFD,- 0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD,- 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD };-- /* sequence */- { size_t offset;- if (!ofCode)- offset = 0;- else {- offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);- }-- if (ofCode <= 1) {- offset += (llCode==0);- if (offset) {- size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];- temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */- if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];- seqState->prevOffset[1] = seqState->prevOffset[0];- seqState->prevOffset[0] = offset = temp;- } else {- offset = seqState->prevOffset[0];- }- } else {- seqState->prevOffset[2] = seqState->prevOffset[1];- seqState->prevOffset[1] = seqState->prevOffset[0];- seqState->prevOffset[0] = offset;- }- seq.offset = offset;- }-- seq.matchLength = ML_base[mlCode] + ((mlCode>31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */- if (MEM_32bits() && (mlBits+llBits>24)) BIT_reloadDStream(&seqState->DStream);-- seq.litLength = LL_base[llCode] + ((llCode>15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */- if (MEM_32bits() ||- (totalBits > 64 - 7 - (LLFSELog+MLFSELog+OffFSELog)) ) BIT_reloadDStream(&seqState->DStream);-- /* ANS state update */- FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */- FSE_updateState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */- FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */-- return seq;-}---FORCE_INLINE-size_t ZSTD_execSequence(BYTE* op,- BYTE* const oend, seq_t sequence,- const BYTE** litPtr, const BYTE* const litLimit,- const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)-{- BYTE* const oLitEnd = op + sequence.litLength;- size_t const sequenceLength = sequence.litLength + sequence.matchLength;- BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */- BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;- const BYTE* const iLitEnd = *litPtr + sequence.litLength;- const BYTE* match = oLitEnd - sequence.offset;-- /* check */- if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */- if (iLitEnd > litLimit) return ERROR(corruption_detected); /* over-read beyond lit buffer */- if (oLitEnd>oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);-- /* copy Literals */- ZSTD_copy8(op, *litPtr);- if (sequence.litLength > 8)- ZSTD_wildcopy(op+8, (*litPtr)+8, sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */- op = oLitEnd;- *litPtr = iLitEnd; /* update for next sequence */-- /* copy Match */- if (sequence.offset > (size_t)(oLitEnd - base)) {- /* offset beyond prefix */- if (sequence.offset > (size_t)(oLitEnd - vBase)) return ERROR(corruption_detected);- match += (dictEnd-base);- if (match + sequence.matchLength <= dictEnd) {- memmove(oLitEnd, match, sequence.matchLength);- return sequenceLength;- }- /* span extDict & currentPrefixSegment */- { size_t const length1 = dictEnd - match;- memmove(oLitEnd, match, length1);- op = oLitEnd + length1;- sequence.matchLength -= length1;- match = base;- if (op > oend_w || sequence.matchLength < MINMATCH) {- U32 i;- for (i = 0; i < sequence.matchLength; ++i) op[i] = match[i];- return sequenceLength;- }- } }- /* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */-- /* match within prefix */- if (sequence.offset < 8) {- /* close range match, overlap */- static const U32 dec32table[] = { 0, 1, 2, 1, 4, 4, 4, 4 }; /* added */- static const int dec64table[] = { 8, 8, 8, 7, 8, 9,10,11 }; /* substracted */- int const sub2 = dec64table[sequence.offset];- op[0] = match[0];- op[1] = match[1];- op[2] = match[2];- op[3] = match[3];- match += dec32table[sequence.offset];- ZSTD_copy4(op+4, match);- match -= sub2;- } else {- ZSTD_copy8(op, match);- }- op += 8; match += 8;-- if (oMatchEnd > oend-(16-MINMATCH)) {- if (op < oend_w) {- ZSTD_wildcopy(op, match, oend_w - op);- match += oend_w - op;- op = oend_w;- }- while (op < oMatchEnd) *op++ = *match++;- } else {- ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */- }- return sequenceLength;-}---static size_t ZSTD_decompressSequences(- ZSTD_DCtx* dctx,- void* dst, size_t maxDstSize,- const void* seqStart, size_t seqSize)-{- const BYTE* ip = (const BYTE*)seqStart;- const BYTE* const iend = ip + seqSize;- BYTE* const ostart = (BYTE* const)dst;- BYTE* const oend = ostart + maxDstSize;- BYTE* op = ostart;- const BYTE* litPtr = dctx->litPtr;- const BYTE* const litEnd = litPtr + dctx->litSize;- const BYTE* const base = (const BYTE*) (dctx->base);- const BYTE* const vBase = (const BYTE*) (dctx->vBase);- const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);- int nbSeq;-- /* Build Decoding Tables */- { size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);- if (ZSTD_isError(seqHSize)) return seqHSize;- ip += seqHSize;- }-- /* Regen sequences */- if (nbSeq) {- seqState_t seqState;- dctx->fseEntropy = 1;- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->rep[i]; }- CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);- FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);- FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);- FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);-- for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {- nbSeq--;- { seq_t const sequence = ZSTD_decodeSequence(&seqState);- size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);- if (ZSTD_isError(oneSeqSize)) return oneSeqSize;- op += oneSeqSize;- } }-- /* check if reached exact end */- if (nbSeq) return ERROR(corruption_detected);- /* save reps for next block */- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->rep[i] = (U32)(seqState.prevOffset[i]); }- }-- /* last literal segment */- { size_t const lastLLSize = litEnd - litPtr;- if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);- memcpy(op, litPtr, lastLLSize);- op += lastLLSize;- }-- return op-ostart;-}---static seq_t ZSTD_decodeSequenceLong(seqState_t* seqState)-{- seq_t seq;-- U32 const llCode = FSE_peekSymbol(&seqState->stateLL);- U32 const mlCode = FSE_peekSymbol(&seqState->stateML);- U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */-- U32 const llBits = LL_bits[llCode];- U32 const mlBits = ML_bits[mlCode];- U32 const ofBits = ofCode;- U32 const totalBits = llBits+mlBits+ofBits;-- static const U32 LL_base[MaxLL+1] = {- 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,- 16, 18, 20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000,- 0x2000, 0x4000, 0x8000, 0x10000 };-- static const U32 ML_base[MaxML+1] = {- 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,- 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,- 35, 37, 39, 41, 43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803,- 0x1003, 0x2003, 0x4003, 0x8003, 0x10003 };-- static const U32 OF_base[MaxOff+1] = {- 0, 1, 1, 5, 0xD, 0x1D, 0x3D, 0x7D,- 0xFD, 0x1FD, 0x3FD, 0x7FD, 0xFFD, 0x1FFD, 0x3FFD, 0x7FFD,- 0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD,- 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD };-- /* sequence */- { size_t offset;- if (!ofCode)- offset = 0;- else {- offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);- }-- if (ofCode <= 1) {- offset += (llCode==0);- if (offset) {- size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];- temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */- if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];- seqState->prevOffset[1] = seqState->prevOffset[0];- seqState->prevOffset[0] = offset = temp;- } else {- offset = seqState->prevOffset[0];- }- } else {- seqState->prevOffset[2] = seqState->prevOffset[1];- seqState->prevOffset[1] = seqState->prevOffset[0];- seqState->prevOffset[0] = offset;- }- seq.offset = offset;- }-- seq.matchLength = ML_base[mlCode] + ((mlCode>31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */- if (MEM_32bits() && (mlBits+llBits>24)) BIT_reloadDStream(&seqState->DStream);-- seq.litLength = LL_base[llCode] + ((llCode>15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */- if (MEM_32bits() ||- (totalBits > 64 - 7 - (LLFSELog+MLFSELog+OffFSELog)) ) BIT_reloadDStream(&seqState->DStream);-- { size_t const pos = seqState->pos + seq.litLength;- seq.match = seqState->base + pos - seq.offset; /* single memory segment */- if (seq.offset > pos) seq.match += seqState->gotoDict; /* separate memory segment */- seqState->pos = pos + seq.matchLength;- }-- /* ANS state update */- FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */- FSE_updateState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */- FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */-- return seq;-}--FORCE_INLINE-size_t ZSTD_execSequenceLong(BYTE* op,- BYTE* const oend, seq_t sequence,- const BYTE** litPtr, const BYTE* const litLimit,- const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)-{- BYTE* const oLitEnd = op + sequence.litLength;- size_t const sequenceLength = sequence.litLength + sequence.matchLength;- BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */- BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;- const BYTE* const iLitEnd = *litPtr + sequence.litLength;- const BYTE* match = sequence.match;-- /* check */-#if 1- if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */- if (iLitEnd > litLimit) return ERROR(corruption_detected); /* over-read beyond lit buffer */- if (oLitEnd>oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);-#endif-- /* copy Literals */- ZSTD_copy8(op, *litPtr);- if (sequence.litLength > 8)- ZSTD_wildcopy(op+8, (*litPtr)+8, sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */- op = oLitEnd;- *litPtr = iLitEnd; /* update for next sequence */-- /* copy Match */-#if 1- if (sequence.offset > (size_t)(oLitEnd - base)) {- /* offset beyond prefix */- if (sequence.offset > (size_t)(oLitEnd - vBase)) return ERROR(corruption_detected);- if (match + sequence.matchLength <= dictEnd) {- memmove(oLitEnd, match, sequence.matchLength);- return sequenceLength;- }- /* span extDict & currentPrefixSegment */- { size_t const length1 = dictEnd - match;- memmove(oLitEnd, match, length1);- op = oLitEnd + length1;- sequence.matchLength -= length1;- match = base;- if (op > oend_w || sequence.matchLength < MINMATCH) {- U32 i;- for (i = 0; i < sequence.matchLength; ++i) op[i] = match[i];- return sequenceLength;- }- } }- /* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */-#endif-- /* match within prefix */- if (sequence.offset < 8) {- /* close range match, overlap */- static const U32 dec32table[] = { 0, 1, 2, 1, 4, 4, 4, 4 }; /* added */- static const int dec64table[] = { 8, 8, 8, 7, 8, 9,10,11 }; /* substracted */- int const sub2 = dec64table[sequence.offset];- op[0] = match[0];- op[1] = match[1];- op[2] = match[2];- op[3] = match[3];- match += dec32table[sequence.offset];- ZSTD_copy4(op+4, match);- match -= sub2;- } else {- ZSTD_copy8(op, match);- }- op += 8; match += 8;-- if (oMatchEnd > oend-(16-MINMATCH)) {- if (op < oend_w) {- ZSTD_wildcopy(op, match, oend_w - op);- match += oend_w - op;- op = oend_w;- }- while (op < oMatchEnd) *op++ = *match++;- } else {- ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */- }- return sequenceLength;-}--static size_t ZSTD_decompressSequencesLong(- ZSTD_DCtx* dctx,- void* dst, size_t maxDstSize,- const void* seqStart, size_t seqSize)-{- const BYTE* ip = (const BYTE*)seqStart;- const BYTE* const iend = ip + seqSize;- BYTE* const ostart = (BYTE* const)dst;- BYTE* const oend = ostart + maxDstSize;- BYTE* op = ostart;- const BYTE* litPtr = dctx->litPtr;- const BYTE* const litEnd = litPtr + dctx->litSize;- const BYTE* const base = (const BYTE*) (dctx->base);- const BYTE* const vBase = (const BYTE*) (dctx->vBase);- const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);- int nbSeq;-- /* Build Decoding Tables */- { size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);- if (ZSTD_isError(seqHSize)) return seqHSize;- ip += seqHSize;- }-- /* Regen sequences */- if (nbSeq) {-#define STORED_SEQS 4-#define STOSEQ_MASK (STORED_SEQS-1)-#define ADVANCED_SEQS 4- seq_t sequences[STORED_SEQS];- int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);- seqState_t seqState;- int seqNb;- dctx->fseEntropy = 1;- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->rep[i]; }- seqState.base = base;- seqState.pos = (size_t)(op-base);- seqState.gotoDict = (iPtrDiff)(dictEnd - base);- CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);- FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);- FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);- FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);-- /* prepare in advance */- for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && seqNb<seqAdvance; seqNb++) {- sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState);- }- if (seqNb<seqAdvance) return ERROR(corruption_detected);-- /* decode and decompress */- for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && seqNb<nbSeq ; seqNb++) {- seq_t const sequence = ZSTD_decodeSequenceLong(&seqState);- size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[(seqNb-ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);- if (ZSTD_isError(oneSeqSize)) return oneSeqSize;- ZSTD_PREFETCH(sequence.match);- sequences[seqNb&STOSEQ_MASK] = sequence;- op += oneSeqSize;- }- if (seqNb<nbSeq) return ERROR(corruption_detected);-- /* finish queue */- seqNb -= seqAdvance;- for ( ; seqNb<nbSeq ; seqNb++) {- size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb&STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);- if (ZSTD_isError(oneSeqSize)) return oneSeqSize;- op += oneSeqSize;- }-- /* save reps for next block */- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->rep[i] = (U32)(seqState.prevOffset[i]); }- }-- /* last literal segment */- { size_t const lastLLSize = litEnd - litPtr;- if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);- memcpy(op, litPtr, lastLLSize);- op += lastLLSize;- }-- return op-ostart;-}---static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize)-{ /* blockType == blockCompressed */- const BYTE* ip = (const BYTE*)src;-- if (srcSize >= ZSTD_BLOCKSIZE_ABSOLUTEMAX) return ERROR(srcSize_wrong);-- /* Decode literals sub-block */- { size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);- if (ZSTD_isError(litCSize)) return litCSize;- ip += litCSize;- srcSize -= litCSize;- }- if (dctx->fParams.windowSize > (1<<23)) return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize);- return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize);-}---static void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst)-{- if (dst != dctx->previousDstEnd) { /* not contiguous */- dctx->dictEnd = dctx->previousDstEnd;- dctx->vBase = (const char*)dst - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->base));- dctx->base = dst;- dctx->previousDstEnd = dst;- }-}--size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize)-{- size_t dSize;- ZSTD_checkContinuity(dctx, dst);- dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize);- dctx->previousDstEnd = (char*)dst + dSize;- return dSize;-}---/** ZSTD_insertBlock() :- insert `src` block into `dctx` history. Useful to track uncompressed blocks. */-ZSTDLIB_API size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize)-{- ZSTD_checkContinuity(dctx, blockStart);- dctx->previousDstEnd = (const char*)blockStart + blockSize;- return blockSize;-}---size_t ZSTD_generateNxBytes(void* dst, size_t dstCapacity, BYTE byte, size_t length)-{- if (length > dstCapacity) return ERROR(dstSize_tooSmall);- memset(dst, byte, length);- return length;-}---/*! ZSTD_decompressFrame() :-* `dctx` must be properly initialized */-static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize)-{- const BYTE* ip = (const BYTE*)src;- BYTE* const ostart = (BYTE* const)dst;- BYTE* const oend = ostart + dstCapacity;- BYTE* op = ostart;- size_t remainingSize = srcSize;-- /* check */- if (srcSize < ZSTD_frameHeaderSize_min+ZSTD_blockHeaderSize) return ERROR(srcSize_wrong);-- /* Frame Header */- { size_t const frameHeaderSize = ZSTD_frameHeaderSize(src, ZSTD_frameHeaderSize_prefix);- if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize;- if (srcSize < frameHeaderSize+ZSTD_blockHeaderSize) return ERROR(srcSize_wrong);- CHECK_F(ZSTD_decodeFrameHeader(dctx, src, frameHeaderSize));- ip += frameHeaderSize; remainingSize -= frameHeaderSize;- }-- /* Loop on each block */- while (1) {- size_t decodedSize;- blockProperties_t blockProperties;- size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);- if (ZSTD_isError(cBlockSize)) return cBlockSize;-- ip += ZSTD_blockHeaderSize;- remainingSize -= ZSTD_blockHeaderSize;- if (cBlockSize > remainingSize) return ERROR(srcSize_wrong);-- switch(blockProperties.blockType)- {- case bt_compressed:- decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend-op, ip, cBlockSize);- break;- case bt_raw :- decodedSize = ZSTD_copyRawBlock(op, oend-op, ip, cBlockSize);- break;- case bt_rle :- decodedSize = ZSTD_generateNxBytes(op, oend-op, *ip, blockProperties.origSize);- break;- case bt_reserved :- default:- return ERROR(corruption_detected);- }-- if (ZSTD_isError(decodedSize)) return decodedSize;- if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, op, decodedSize);- op += decodedSize;- ip += cBlockSize;- remainingSize -= cBlockSize;- if (blockProperties.lastBlock) break;- }-- if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */- U32 const checkCalc = (U32)XXH64_digest(&dctx->xxhState);- U32 checkRead;- if (remainingSize<4) return ERROR(checksum_wrong);- checkRead = MEM_readLE32(ip);- if (checkRead != checkCalc) return ERROR(checksum_wrong);- remainingSize -= 4;- }-- if (remainingSize) return ERROR(srcSize_wrong);- return op-ostart;-}---size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const void* dict, size_t dictSize)-{-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT==1)- if (ZSTD_isLegacy(src, srcSize)) return ZSTD_decompressLegacy(dst, dstCapacity, src, srcSize, dict, dictSize);-#endif- ZSTD_decompressBegin_usingDict(dctx, dict, dictSize);- ZSTD_checkContinuity(dctx, dst);- return ZSTD_decompressFrame(dctx, dst, dstCapacity, src, srcSize);-}---size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- return ZSTD_decompress_usingDict(dctx, dst, dstCapacity, src, srcSize, NULL, 0);-}---size_t ZSTD_decompress(void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{-#if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE==1)- size_t regenSize;- ZSTD_DCtx* const dctx = ZSTD_createDCtx();- if (dctx==NULL) return ERROR(memory_allocation);- regenSize = ZSTD_decompressDCtx(dctx, dst, dstCapacity, src, srcSize);- ZSTD_freeDCtx(dctx);- return regenSize;-#else /* stack mode */- ZSTD_DCtx dctx;- return ZSTD_decompressDCtx(&dctx, dst, dstCapacity, src, srcSize);-#endif-}---/*-**************************************-* Advanced Streaming Decompression API-* Bufferless and synchronous-****************************************/-size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx) { return dctx->expected; }--ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx) {- switch(dctx->stage)- {- default: /* should not happen */- case ZSTDds_getFrameHeaderSize:- case ZSTDds_decodeFrameHeader:- return ZSTDnit_frameHeader;- case ZSTDds_decodeBlockHeader:- return ZSTDnit_blockHeader;- case ZSTDds_decompressBlock:- return ZSTDnit_block;- case ZSTDds_decompressLastBlock:- return ZSTDnit_lastBlock;- case ZSTDds_checkChecksum:- return ZSTDnit_checksum;- case ZSTDds_decodeSkippableHeader:- case ZSTDds_skipFrame:- return ZSTDnit_skippableFrame;- }-}--int ZSTD_isSkipFrame(ZSTD_DCtx* dctx) { return dctx->stage == ZSTDds_skipFrame; } /* for zbuff */--/** ZSTD_decompressContinue() :-* @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity)-* or an error code, which can be tested using ZSTD_isError() */-size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- /* Sanity check */- if (srcSize != dctx->expected) return ERROR(srcSize_wrong);- if (dstCapacity) ZSTD_checkContinuity(dctx, dst);-- switch (dctx->stage)- {- case ZSTDds_getFrameHeaderSize :- if (srcSize != ZSTD_frameHeaderSize_prefix) return ERROR(srcSize_wrong); /* impossible */- if ((MEM_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */- memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);- dctx->expected = ZSTD_skippableHeaderSize - ZSTD_frameHeaderSize_prefix; /* magic number + skippable frame length */- dctx->stage = ZSTDds_decodeSkippableHeader;- return 0;- }- dctx->headerSize = ZSTD_frameHeaderSize(src, ZSTD_frameHeaderSize_prefix);- if (ZSTD_isError(dctx->headerSize)) return dctx->headerSize;- memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);- if (dctx->headerSize > ZSTD_frameHeaderSize_prefix) {- dctx->expected = dctx->headerSize - ZSTD_frameHeaderSize_prefix;- dctx->stage = ZSTDds_decodeFrameHeader;- return 0;- }- dctx->expected = 0; /* not necessary to copy more */-- case ZSTDds_decodeFrameHeader:- memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);- CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));- dctx->expected = ZSTD_blockHeaderSize;- dctx->stage = ZSTDds_decodeBlockHeader;- return 0;-- case ZSTDds_decodeBlockHeader:- { blockProperties_t bp;- size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);- if (ZSTD_isError(cBlockSize)) return cBlockSize;- dctx->expected = cBlockSize;- dctx->bType = bp.blockType;- dctx->rleSize = bp.origSize;- if (cBlockSize) {- dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock;- return 0;- }- /* empty block */- if (bp.lastBlock) {- if (dctx->fParams.checksumFlag) {- dctx->expected = 4;- dctx->stage = ZSTDds_checkChecksum;- } else {- dctx->expected = 0; /* end of frame */- dctx->stage = ZSTDds_getFrameHeaderSize;- }- } else {- dctx->expected = 3; /* go directly to next header */- dctx->stage = ZSTDds_decodeBlockHeader;- }- return 0;- }- case ZSTDds_decompressLastBlock:- case ZSTDds_decompressBlock:- { size_t rSize;- switch(dctx->bType)- {- case bt_compressed:- rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize);- break;- case bt_raw :- rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize);- break;- case bt_rle :- rSize = ZSTD_setRleBlock(dst, dstCapacity, src, srcSize, dctx->rleSize);- break;- case bt_reserved : /* should never happen */- default:- return ERROR(corruption_detected);- }- if (ZSTD_isError(rSize)) return rSize;- if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, dst, rSize);-- if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */- if (dctx->fParams.checksumFlag) { /* another round for frame checksum */- dctx->expected = 4;- dctx->stage = ZSTDds_checkChecksum;- } else {- dctx->expected = 0; /* ends here */- dctx->stage = ZSTDds_getFrameHeaderSize;- }- } else {- dctx->stage = ZSTDds_decodeBlockHeader;- dctx->expected = ZSTD_blockHeaderSize;- dctx->previousDstEnd = (char*)dst + rSize;- }- return rSize;- }- case ZSTDds_checkChecksum:- { U32 const h32 = (U32)XXH64_digest(&dctx->xxhState);- U32 const check32 = MEM_readLE32(src); /* srcSize == 4, guaranteed by dctx->expected */- if (check32 != h32) return ERROR(checksum_wrong);- dctx->expected = 0;- dctx->stage = ZSTDds_getFrameHeaderSize;- return 0;- }- case ZSTDds_decodeSkippableHeader:- { memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);- dctx->expected = MEM_readLE32(dctx->headerBuffer + 4);- dctx->stage = ZSTDds_skipFrame;- return 0;- }- case ZSTDds_skipFrame:- { dctx->expected = 0;- dctx->stage = ZSTDds_getFrameHeaderSize;- return 0;- }- default:- return ERROR(GENERIC); /* impossible */- }-}---static size_t ZSTD_refDictContent(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)-{- dctx->dictEnd = dctx->previousDstEnd;- dctx->vBase = (const char*)dict - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->base));- dctx->base = dict;- dctx->previousDstEnd = (const char*)dict + dictSize;- return 0;-}--static size_t ZSTD_loadEntropy(ZSTD_DCtx* dctx, const void* const dict, size_t const dictSize)-{- const BYTE* dictPtr = (const BYTE*)dict;- const BYTE* const dictEnd = dictPtr + dictSize;-- { size_t const hSize = HUF_readDTableX4(dctx->hufTable, dict, dictSize);- if (HUF_isError(hSize)) return ERROR(dictionary_corrupted);- dictPtr += hSize;- }-- { short offcodeNCount[MaxOff+1];- U32 offcodeMaxValue=MaxOff, offcodeLog;- size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);- if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted);- if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted);- CHECK_E(FSE_buildDTable(dctx->OFTable, offcodeNCount, offcodeMaxValue, offcodeLog), dictionary_corrupted);- dictPtr += offcodeHeaderSize;- }-- { short matchlengthNCount[MaxML+1];- unsigned matchlengthMaxValue = MaxML, matchlengthLog;- size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);- if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted);- if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted);- CHECK_E(FSE_buildDTable(dctx->MLTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog), dictionary_corrupted);- dictPtr += matchlengthHeaderSize;- }-- { short litlengthNCount[MaxLL+1];- unsigned litlengthMaxValue = MaxLL, litlengthLog;- size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);- if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted);- if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted);- CHECK_E(FSE_buildDTable(dctx->LLTable, litlengthNCount, litlengthMaxValue, litlengthLog), dictionary_corrupted);- dictPtr += litlengthHeaderSize;- }-- if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted);- dctx->rep[0] = MEM_readLE32(dictPtr+0); if (dctx->rep[0] == 0 || dctx->rep[0] >= dictSize) return ERROR(dictionary_corrupted);- dctx->rep[1] = MEM_readLE32(dictPtr+4); if (dctx->rep[1] == 0 || dctx->rep[1] >= dictSize) return ERROR(dictionary_corrupted);- dctx->rep[2] = MEM_readLE32(dictPtr+8); if (dctx->rep[2] == 0 || dctx->rep[2] >= dictSize) return ERROR(dictionary_corrupted);- dictPtr += 12;-- dctx->litEntropy = dctx->fseEntropy = 1;- return dictPtr - (const BYTE*)dict;-}--static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)-{- if (dictSize < 8) return ZSTD_refDictContent(dctx, dict, dictSize);- { U32 const magic = MEM_readLE32(dict);- if (magic != ZSTD_DICT_MAGIC) {- return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */- } }- dctx->dictID = MEM_readLE32((const char*)dict + 4);-- /* load entropy tables */- dict = (const char*)dict + 8;- dictSize -= 8;- { size_t const eSize = ZSTD_loadEntropy(dctx, dict, dictSize);- if (ZSTD_isError(eSize)) return ERROR(dictionary_corrupted);- dict = (const char*)dict + eSize;- dictSize -= eSize;- }-- /* reference dictionary content */- return ZSTD_refDictContent(dctx, dict, dictSize);-}--size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)-{- CHECK_F(ZSTD_decompressBegin(dctx));- if (dict && dictSize) CHECK_E(ZSTD_decompress_insertDictionary(dctx, dict, dictSize), dictionary_corrupted);- return 0;-}---/* ====== ZSTD_DDict ====== */--struct ZSTD_DDict_s {- void* dictBuffer;- const void* dictContent;- size_t dictSize;- ZSTD_DCtx* refContext;-}; /* typedef'd to ZSTD_DDict within "zstd.h" */--ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize, unsigned byReference, ZSTD_customMem customMem)-{- if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;- if (!customMem.customAlloc || !customMem.customFree) return NULL;-- { ZSTD_DDict* const ddict = (ZSTD_DDict*) ZSTD_malloc(sizeof(ZSTD_DDict), customMem);- ZSTD_DCtx* const dctx = ZSTD_createDCtx_advanced(customMem);-- if (!ddict || !dctx) {- ZSTD_free(ddict, customMem);- ZSTD_free(dctx, customMem);- return NULL;- }-- if ((byReference) || (!dict) || (!dictSize)) {- ddict->dictBuffer = NULL;- ddict->dictContent = dict;- } else {- void* const internalBuffer = ZSTD_malloc(dictSize, customMem);- if (!internalBuffer) { ZSTD_free(dctx, customMem); ZSTD_free(ddict, customMem); return NULL; }- memcpy(internalBuffer, dict, dictSize);- ddict->dictBuffer = internalBuffer;- ddict->dictContent = internalBuffer;- }- { size_t const errorCode = ZSTD_decompressBegin_usingDict(dctx, ddict->dictContent, dictSize);- if (ZSTD_isError(errorCode)) {- ZSTD_free(ddict->dictBuffer, customMem);- ZSTD_free(ddict, customMem);- ZSTD_free(dctx, customMem);- return NULL;- } }-- ddict->dictSize = dictSize;- ddict->refContext = dctx;- return ddict;- }-}--/*! ZSTD_createDDict() :-* Create a digested dictionary, ready to start decompression without startup delay.-* `dict` can be released after `ZSTD_DDict` creation */-ZSTD_DDict* ZSTD_createDDict(const void* dict, size_t dictSize)-{- ZSTD_customMem const allocator = { NULL, NULL, NULL };- return ZSTD_createDDict_advanced(dict, dictSize, 0, allocator);-}---/*! ZSTD_createDDict_byReference() :- * Create a digested dictionary, ready to start decompression operation without startup delay.- * Dictionary content is simply referenced, and therefore stays in dictBuffer.- * It is important that dictBuffer outlives DDict, it must remain read accessible throughout the lifetime of DDict */-ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize)-{- ZSTD_customMem const allocator = { NULL, NULL, NULL };- return ZSTD_createDDict_advanced(dictBuffer, dictSize, 1, allocator);-}---size_t ZSTD_freeDDict(ZSTD_DDict* ddict)-{- if (ddict==NULL) return 0; /* support free on NULL */- { ZSTD_customMem const cMem = ddict->refContext->customMem;- ZSTD_freeDCtx(ddict->refContext);- ZSTD_free(ddict->dictBuffer, cMem);- ZSTD_free(ddict, cMem);- return 0;- }-}--size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict)-{- if (ddict==NULL) return 0; /* support sizeof on NULL */- return sizeof(*ddict) + ZSTD_sizeof_DCtx(ddict->refContext) + (ddict->dictBuffer ? ddict->dictSize : 0) ;-}--/*! ZSTD_getDictID_fromDict() :- * Provides the dictID stored within dictionary.- * if @return == 0, the dictionary is not conformant with Zstandard specification.- * It can still be loaded, but as a content-only dictionary. */-unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize)-{- if (dictSize < 8) return 0;- if (MEM_readLE32(dict) != ZSTD_DICT_MAGIC) return 0;- return MEM_readLE32((const char*)dict + 4);-}--/*! ZSTD_getDictID_fromDDict() :- * Provides the dictID of the dictionary loaded into `ddict`.- * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.- * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */-unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict)-{- if (ddict==NULL) return 0;- return ZSTD_getDictID_fromDict(ddict->dictContent, ddict->dictSize);-}--/*! ZSTD_getDictID_fromFrame() :- * Provides the dictID required to decompressed the frame stored within `src`.- * If @return == 0, the dictID could not be decoded.- * This could for one of the following reasons :- * - The frame does not require a dictionary to be decoded (most common case).- * - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.- * Note : this use case also happens when using a non-conformant dictionary.- * - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).- * - This is not a Zstandard frame.- * When identifying the exact failure cause, it's possible to used ZSTD_getFrameParams(), which will provide a more precise error code. */-unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize)-{- ZSTD_frameParams zfp = { 0 , 0 , 0 , 0 };- size_t const hError = ZSTD_getFrameParams(&zfp, src, srcSize);- if (ZSTD_isError(hError)) return 0;- return zfp.dictID;-}---/*! ZSTD_decompress_usingDDict() :-* Decompression using a pre-digested Dictionary-* Use dictionary without significant overhead. */-size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const ZSTD_DDict* ddict)-{-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT==1)- if (ZSTD_isLegacy(src, srcSize)) return ZSTD_decompressLegacy(dst, dstCapacity, src, srcSize, ddict->dictContent, ddict->dictSize);-#endif- ZSTD_refDCtx(dctx, ddict->refContext);- ZSTD_checkContinuity(dctx, dst);- return ZSTD_decompressFrame(dctx, dst, dstCapacity, src, srcSize);-}---/*=====================================-* Streaming decompression-*====================================*/--typedef enum { zdss_init, zdss_loadHeader,- zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;--/* *** Resource management *** */-struct ZSTD_DStream_s {- ZSTD_DCtx* dctx;- ZSTD_DDict* ddictLocal;- const ZSTD_DDict* ddict;- ZSTD_frameParams fParams;- ZSTD_dStreamStage stage;- char* inBuff;- size_t inBuffSize;- size_t inPos;- size_t maxWindowSize;- char* outBuff;- size_t outBuffSize;- size_t outStart;- size_t outEnd;- size_t blockSize;- BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX]; /* tmp buffer to store frame header */- size_t lhSize;- ZSTD_customMem customMem;- void* legacyContext;- U32 previousLegacyVersion;- U32 legacyVersion;- U32 hostageByte;-}; /* typedef'd to ZSTD_DStream within "zstd.h" */---ZSTD_DStream* ZSTD_createDStream(void)-{- return ZSTD_createDStream_advanced(defaultCustomMem);-}--ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem)-{- ZSTD_DStream* zds;-- if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;- if (!customMem.customAlloc || !customMem.customFree) return NULL;-- zds = (ZSTD_DStream*) ZSTD_malloc(sizeof(ZSTD_DStream), customMem);- if (zds==NULL) return NULL;- memset(zds, 0, sizeof(ZSTD_DStream));- memcpy(&zds->customMem, &customMem, sizeof(ZSTD_customMem));- zds->dctx = ZSTD_createDCtx_advanced(customMem);- if (zds->dctx == NULL) { ZSTD_freeDStream(zds); return NULL; }- zds->stage = zdss_init;- zds->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;- return zds;-}--size_t ZSTD_freeDStream(ZSTD_DStream* zds)-{- if (zds==NULL) return 0; /* support free on null */- { ZSTD_customMem const cMem = zds->customMem;- ZSTD_freeDCtx(zds->dctx);- ZSTD_freeDDict(zds->ddictLocal);- ZSTD_free(zds->inBuff, cMem);- ZSTD_free(zds->outBuff, cMem);-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)- if (zds->legacyContext)- ZSTD_freeLegacyStreamContext(zds->legacyContext, zds->previousLegacyVersion);-#endif- ZSTD_free(zds, cMem);- return 0;- }-}---/* *** Initialization *** */--size_t ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX + ZSTD_blockHeaderSize; }-size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }--size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize)-{- zds->stage = zdss_loadHeader;- zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;- ZSTD_freeDDict(zds->ddictLocal);- if (dict) {- zds->ddictLocal = ZSTD_createDDict(dict, dictSize);- if (zds->ddictLocal == NULL) return ERROR(memory_allocation);- } else zds->ddictLocal = NULL;- zds->ddict = zds->ddictLocal;- zds->legacyVersion = 0;- zds->hostageByte = 0;- return ZSTD_frameHeaderSize_prefix;-}--size_t ZSTD_initDStream(ZSTD_DStream* zds)-{- return ZSTD_initDStream_usingDict(zds, NULL, 0);-}--size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict) /**< note : ddict will just be referenced, and must outlive decompression session */-{- size_t const initResult = ZSTD_initDStream(zds);- zds->ddict = ddict;- return initResult;-}--size_t ZSTD_resetDStream(ZSTD_DStream* zds)-{- zds->stage = zdss_loadHeader;- zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;- zds->legacyVersion = 0;- zds->hostageByte = 0;- return ZSTD_frameHeaderSize_prefix;-}--size_t ZSTD_setDStreamParameter(ZSTD_DStream* zds,- ZSTD_DStreamParameter_e paramType, unsigned paramValue)-{- switch(paramType)- {- default : return ERROR(parameter_unknown);- case ZSTDdsp_maxWindowSize : zds->maxWindowSize = paramValue ? paramValue : (U32)(-1); break;- }- return 0;-}---size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds)-{- if (zds==NULL) return 0; /* support sizeof on NULL */- return sizeof(*zds) + ZSTD_sizeof_DCtx(zds->dctx) + ZSTD_sizeof_DDict(zds->ddictLocal) + zds->inBuffSize + zds->outBuffSize;-}---/* ***** Decompression ***** */--MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize)-{- size_t const length = MIN(dstCapacity, srcSize);- memcpy(dst, src, length);- return length;-}---size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input)-{- const char* const istart = (const char*)(input->src) + input->pos;- const char* const iend = (const char*)(input->src) + input->size;- const char* ip = istart;- char* const ostart = (char*)(output->dst) + output->pos;- char* const oend = (char*)(output->dst) + output->size;- char* op = ostart;- U32 someMoreWork = 1;--#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)- if (zds->legacyVersion)- return ZSTD_decompressLegacyStream(zds->legacyContext, zds->legacyVersion, output, input);-#endif-- while (someMoreWork) {- switch(zds->stage)- {- case zdss_init :- ZSTD_resetDStream(zds); /* transparent reset on starting decoding a new frame */- /* fall-through */-- case zdss_loadHeader :- { size_t const hSize = ZSTD_getFrameParams(&zds->fParams, zds->headerBuffer, zds->lhSize);- if (ZSTD_isError(hSize))-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)- { U32 const legacyVersion = ZSTD_isLegacy(istart, iend-istart);- if (legacyVersion) {- const void* const dict = zds->ddict ? zds->ddict->dictContent : NULL;- size_t const dictSize = zds->ddict ? zds->ddict->dictSize : 0;- CHECK_F(ZSTD_initLegacyStream(&zds->legacyContext, zds->previousLegacyVersion, legacyVersion,- dict, dictSize));- zds->legacyVersion = zds->previousLegacyVersion = legacyVersion;- return ZSTD_decompressLegacyStream(zds->legacyContext, zds->legacyVersion, output, input);- } else {- return hSize; /* error */- } }-#else- return hSize;-#endif- if (hSize != 0) { /* need more input */- size_t const toLoad = hSize - zds->lhSize; /* if hSize!=0, hSize > zds->lhSize */- if (toLoad > (size_t)(iend-ip)) { /* not enough input to load full header */- memcpy(zds->headerBuffer + zds->lhSize, ip, iend-ip);- zds->lhSize += iend-ip;- input->pos = input->size;- return (MAX(ZSTD_frameHeaderSize_min, hSize) - zds->lhSize) + ZSTD_blockHeaderSize; /* remaining header bytes + next block header */- }- memcpy(zds->headerBuffer + zds->lhSize, ip, toLoad); zds->lhSize = hSize; ip += toLoad;- break;- } }-- /* Consume header */- { const ZSTD_DCtx* refContext = zds->ddict ? zds->ddict->refContext : NULL;- ZSTD_refDCtx(zds->dctx, refContext);- }- { size_t const h1Size = ZSTD_nextSrcSizeToDecompress(zds->dctx); /* == ZSTD_frameHeaderSize_prefix */- CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer, h1Size));- { size_t const h2Size = ZSTD_nextSrcSizeToDecompress(zds->dctx);- CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer+h1Size, h2Size));- } }-- zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);- if (zds->fParams.windowSize > zds->maxWindowSize) return ERROR(frameParameter_windowTooLarge);-- /* Adapt buffer sizes to frame header instructions */- { size_t const blockSize = MIN(zds->fParams.windowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);- size_t const neededOutSize = zds->fParams.windowSize + blockSize;- zds->blockSize = blockSize;- if (zds->inBuffSize < blockSize) {- ZSTD_free(zds->inBuff, zds->customMem);- zds->inBuffSize = blockSize;- zds->inBuff = (char*)ZSTD_malloc(blockSize, zds->customMem);- if (zds->inBuff == NULL) return ERROR(memory_allocation);- }- if (zds->outBuffSize < neededOutSize) {- ZSTD_free(zds->outBuff, zds->customMem);- zds->outBuffSize = neededOutSize;- zds->outBuff = (char*)ZSTD_malloc(neededOutSize, zds->customMem);- if (zds->outBuff == NULL) return ERROR(memory_allocation);- } }- zds->stage = zdss_read;- /* pass-through */-- case zdss_read:- { size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx);- if (neededInSize==0) { /* end of frame */- zds->stage = zdss_init;- someMoreWork = 0;- break;- }- if ((size_t)(iend-ip) >= neededInSize) { /* decode directly from src */- const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx);- size_t const decodedSize = ZSTD_decompressContinue(zds->dctx,- zds->outBuff + zds->outStart, (isSkipFrame ? 0 : zds->outBuffSize - zds->outStart),- ip, neededInSize);- if (ZSTD_isError(decodedSize)) return decodedSize;- ip += neededInSize;- if (!decodedSize && !isSkipFrame) break; /* this was just a header */- zds->outEnd = zds->outStart + decodedSize;- zds->stage = zdss_flush;- break;- }- if (ip==iend) { someMoreWork = 0; break; } /* no more input */- zds->stage = zdss_load;- /* pass-through */- }-- case zdss_load:- { size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx);- size_t const toLoad = neededInSize - zds->inPos; /* should always be <= remaining space within inBuff */- size_t loadedSize;- if (toLoad > zds->inBuffSize - zds->inPos) return ERROR(corruption_detected); /* should never happen */- loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend-ip);- ip += loadedSize;- zds->inPos += loadedSize;- if (loadedSize < toLoad) { someMoreWork = 0; break; } /* not enough input, wait for more */-- /* decode loaded input */- { const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx);- size_t const decodedSize = ZSTD_decompressContinue(zds->dctx,- zds->outBuff + zds->outStart, zds->outBuffSize - zds->outStart,- zds->inBuff, neededInSize);- if (ZSTD_isError(decodedSize)) return decodedSize;- zds->inPos = 0; /* input is consumed */- if (!decodedSize && !isSkipFrame) { zds->stage = zdss_read; break; } /* this was just a header */- zds->outEnd = zds->outStart + decodedSize;- zds->stage = zdss_flush;- /* pass-through */- } }-- case zdss_flush:- { size_t const toFlushSize = zds->outEnd - zds->outStart;- size_t const flushedSize = ZSTD_limitCopy(op, oend-op, zds->outBuff + zds->outStart, toFlushSize);- op += flushedSize;- zds->outStart += flushedSize;- if (flushedSize == toFlushSize) { /* flush completed */- zds->stage = zdss_read;- if (zds->outStart + zds->blockSize > zds->outBuffSize)- zds->outStart = zds->outEnd = 0;- break;- }- /* cannot complete flush */- someMoreWork = 0;- break;- }- default: return ERROR(GENERIC); /* impossible */- } }-- /* result */- input->pos += (size_t)(ip-istart);- output->pos += (size_t)(op-ostart);- { size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds->dctx);- if (!nextSrcSizeHint) { /* frame fully decoded */- if (zds->outEnd == zds->outStart) { /* output fully flushed */- if (zds->hostageByte) {- if (input->pos >= input->size) { zds->stage = zdss_read; return 1; } /* can't release hostage (not present) */- input->pos++; /* release hostage */- }- return 0;- }- if (!zds->hostageByte) { /* output not fully flushed; keep last byte as hostage; will be released when all output is flushed */- input->pos--; /* note : pos > 0, otherwise, impossible to finish reading last block */- zds->hostageByte=1;- }- return 1;- }- nextSrcSizeHint += ZSTD_blockHeaderSize * (ZSTD_nextInputType(zds->dctx) == ZSTDnit_block); /* preload header of next block */- if (zds->inPos > nextSrcSizeHint) return ERROR(GENERIC); /* should never happen */- nextSrcSizeHint -= zds->inPos; /* already loaded*/- return nextSrcSizeHint;- }+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+++/* ***************************************************************+* Tuning parameters+*****************************************************************/+/*!+ * HEAPMODE :+ * Select how default decompression function ZSTD_decompress() allocates its context,+ * on stack (0), or into heap (1, default; requires malloc()).+ * Note that functions with explicit context such as ZSTD_decompressDCtx() are unaffected.+ */+#ifndef ZSTD_HEAPMODE+# define ZSTD_HEAPMODE 1+#endif++/*!+* LEGACY_SUPPORT :+* if set to 1+, ZSTD_decompress() can decode older formats (v0.1+)+*/+#ifndef ZSTD_LEGACY_SUPPORT+# define ZSTD_LEGACY_SUPPORT 0+#endif++/*!+ * MAXWINDOWSIZE_DEFAULT :+ * maximum window size accepted by DStream __by default__.+ * Frames requiring more memory will be rejected.+ * It's possible to set a different limit using ZSTD_DCtx_setMaxWindowSize().+ */+#ifndef ZSTD_MAXWINDOWSIZE_DEFAULT+# define ZSTD_MAXWINDOWSIZE_DEFAULT (((U32)1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT) + 1)+#endif++/*!+ * NO_FORWARD_PROGRESS_MAX :+ * maximum allowed nb of calls to ZSTD_decompressStream()+ * without any forward progress+ * (defined as: no byte read from input, and no byte flushed to output)+ * before triggering an error.+ */+#ifndef ZSTD_NO_FORWARD_PROGRESS_MAX+# define ZSTD_NO_FORWARD_PROGRESS_MAX 16+#endif+++/*-*******************************************************+* Dependencies+*********************************************************/+#include <string.h> /* memcpy, memmove, memset */+#include "cpu.h" /* bmi2 */+#include "mem.h" /* low level memory routines */+#define FSE_STATIC_LINKING_ONLY+#include "fse.h"+#define HUF_STATIC_LINKING_ONLY+#include "huf.h"+#include "zstd_internal.h" /* blockProperties_t */+#include "zstd_decompress_internal.h" /* ZSTD_DCtx */+#include "zstd_ddict.h" /* ZSTD_DDictDictContent */+#include "zstd_decompress_block.h" /* ZSTD_decompressBlock_internal */++#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)+# include "zstd_legacy.h"+#endif+++/*-*************************************************************+* Context management+***************************************************************/+size_t ZSTD_sizeof_DCtx (const ZSTD_DCtx* dctx)+{+ if (dctx==NULL) return 0; /* support sizeof NULL */+ return sizeof(*dctx)+ + ZSTD_sizeof_DDict(dctx->ddictLocal)+ + dctx->inBuffSize + dctx->outBuffSize;+}++size_t ZSTD_estimateDCtxSize(void) { return sizeof(ZSTD_DCtx); }+++static size_t ZSTD_startingInputLength(ZSTD_format_e format)+{+ size_t const startingInputLength = (format==ZSTD_f_zstd1_magicless) ?+ ZSTD_FRAMEHEADERSIZE_PREFIX - ZSTD_FRAMEIDSIZE :+ ZSTD_FRAMEHEADERSIZE_PREFIX;+ ZSTD_STATIC_ASSERT(ZSTD_FRAMEHEADERSIZE_PREFIX >= ZSTD_FRAMEIDSIZE);+ /* only supports formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless */+ assert( (format == ZSTD_f_zstd1) || (format == ZSTD_f_zstd1_magicless) );+ return startingInputLength;+}++static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)+{+ dctx->format = ZSTD_f_zstd1; /* ZSTD_decompressBegin() invokes ZSTD_startingInputLength() with argument dctx->format */+ dctx->staticSize = 0;+ dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;+ dctx->ddict = NULL;+ dctx->ddictLocal = NULL;+ dctx->dictEnd = NULL;+ dctx->ddictIsCold = 0;+ dctx->dictUses = ZSTD_dont_use;+ dctx->inBuff = NULL;+ dctx->inBuffSize = 0;+ dctx->outBuffSize = 0;+ dctx->streamStage = zdss_init;+ dctx->legacyContext = NULL;+ dctx->previousLegacyVersion = 0;+ dctx->noForwardProgress = 0;+ dctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());+}++ZSTD_DCtx* ZSTD_initStaticDCtx(void *workspace, size_t workspaceSize)+{+ ZSTD_DCtx* const dctx = (ZSTD_DCtx*) workspace;++ if ((size_t)workspace & 7) return NULL; /* 8-aligned */+ if (workspaceSize < sizeof(ZSTD_DCtx)) return NULL; /* minimum size */++ ZSTD_initDCtx_internal(dctx);+ dctx->staticSize = workspaceSize;+ dctx->inBuff = (char*)(dctx+1);+ return dctx;+}++ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem)+{+ if (!customMem.customAlloc ^ !customMem.customFree) return NULL;++ { ZSTD_DCtx* const dctx = (ZSTD_DCtx*)ZSTD_malloc(sizeof(*dctx), customMem);+ if (!dctx) return NULL;+ dctx->customMem = customMem;+ ZSTD_initDCtx_internal(dctx);+ return dctx;+ }+}++ZSTD_DCtx* ZSTD_createDCtx(void)+{+ DEBUGLOG(3, "ZSTD_createDCtx");+ return ZSTD_createDCtx_advanced(ZSTD_defaultCMem);+}++static void ZSTD_clearDict(ZSTD_DCtx* dctx)+{+ ZSTD_freeDDict(dctx->ddictLocal);+ dctx->ddictLocal = NULL;+ dctx->ddict = NULL;+ dctx->dictUses = ZSTD_dont_use;+}++size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx)+{+ if (dctx==NULL) return 0; /* support free on NULL */+ RETURN_ERROR_IF(dctx->staticSize, memory_allocation, "not compatible with static DCtx");+ { ZSTD_customMem const cMem = dctx->customMem;+ ZSTD_clearDict(dctx);+ ZSTD_free(dctx->inBuff, cMem);+ dctx->inBuff = NULL;+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)+ if (dctx->legacyContext)+ ZSTD_freeLegacyStreamContext(dctx->legacyContext, dctx->previousLegacyVersion);+#endif+ ZSTD_free(dctx, cMem);+ return 0;+ }+}++/* no longer useful */+void ZSTD_copyDCtx(ZSTD_DCtx* dstDCtx, const ZSTD_DCtx* srcDCtx)+{+ size_t const toCopy = (size_t)((char*)(&dstDCtx->inBuff) - (char*)dstDCtx);+ memcpy(dstDCtx, srcDCtx, toCopy); /* no need to copy workspace */+}+++/*-*************************************************************+ * Frame header decoding+ ***************************************************************/++/*! ZSTD_isFrame() :+ * Tells if the content of `buffer` starts with a valid Frame Identifier.+ * Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.+ * Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.+ * Note 3 : Skippable Frame Identifiers are considered valid. */+unsigned ZSTD_isFrame(const void* buffer, size_t size)+{+ if (size < ZSTD_FRAMEIDSIZE) return 0;+ { U32 const magic = MEM_readLE32(buffer);+ if (magic == ZSTD_MAGICNUMBER) return 1;+ if ((magic & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) return 1;+ }+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)+ if (ZSTD_isLegacy(buffer, size)) return 1;+#endif+ return 0;+}++/** ZSTD_frameHeaderSize_internal() :+ * srcSize must be large enough to reach header size fields.+ * note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless.+ * @return : size of the Frame Header+ * or an error code, which can be tested with ZSTD_isError() */+static size_t ZSTD_frameHeaderSize_internal(const void* src, size_t srcSize, ZSTD_format_e format)+{+ size_t const minInputSize = ZSTD_startingInputLength(format);+ RETURN_ERROR_IF(srcSize < minInputSize, srcSize_wrong);++ { BYTE const fhd = ((const BYTE*)src)[minInputSize-1];+ U32 const dictID= fhd & 3;+ U32 const singleSegment = (fhd >> 5) & 1;+ U32 const fcsId = fhd >> 6;+ return minInputSize + !singleSegment+ + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId]+ + (singleSegment && !fcsId);+ }+}++/** ZSTD_frameHeaderSize() :+ * srcSize must be >= ZSTD_frameHeaderSize_prefix.+ * @return : size of the Frame Header,+ * or an error code (if srcSize is too small) */+size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize)+{+ return ZSTD_frameHeaderSize_internal(src, srcSize, ZSTD_f_zstd1);+}+++/** ZSTD_getFrameHeader_advanced() :+ * decode Frame Header, or require larger `srcSize`.+ * note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless+ * @return : 0, `zfhPtr` is correctly filled,+ * >0, `srcSize` is too small, value is wanted `srcSize` amount,+ * or an error code, which can be tested using ZSTD_isError() */+size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format)+{+ const BYTE* ip = (const BYTE*)src;+ size_t const minInputSize = ZSTD_startingInputLength(format);++ memset(zfhPtr, 0, sizeof(*zfhPtr)); /* not strictly necessary, but static analyzer do not understand that zfhPtr is only going to be read only if return value is zero, since they are 2 different signals */+ if (srcSize < minInputSize) return minInputSize;+ RETURN_ERROR_IF(src==NULL, GENERIC, "invalid parameter");++ if ( (format != ZSTD_f_zstd1_magicless)+ && (MEM_readLE32(src) != ZSTD_MAGICNUMBER) ) {+ if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {+ /* skippable frame */+ if (srcSize < ZSTD_SKIPPABLEHEADERSIZE)+ return ZSTD_SKIPPABLEHEADERSIZE; /* magic number + frame length */+ memset(zfhPtr, 0, sizeof(*zfhPtr));+ zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_FRAMEIDSIZE);+ zfhPtr->frameType = ZSTD_skippableFrame;+ return 0;+ }+ RETURN_ERROR(prefix_unknown);+ }++ /* ensure there is enough `srcSize` to fully read/decode frame header */+ { size_t const fhsize = ZSTD_frameHeaderSize_internal(src, srcSize, format);+ if (srcSize < fhsize) return fhsize;+ zfhPtr->headerSize = (U32)fhsize;+ }++ { BYTE const fhdByte = ip[minInputSize-1];+ size_t pos = minInputSize;+ U32 const dictIDSizeCode = fhdByte&3;+ U32 const checksumFlag = (fhdByte>>2)&1;+ U32 const singleSegment = (fhdByte>>5)&1;+ U32 const fcsID = fhdByte>>6;+ U64 windowSize = 0;+ U32 dictID = 0;+ U64 frameContentSize = ZSTD_CONTENTSIZE_UNKNOWN;+ RETURN_ERROR_IF((fhdByte & 0x08) != 0, frameParameter_unsupported,+ "reserved bits, must be zero");++ if (!singleSegment) {+ BYTE const wlByte = ip[pos++];+ U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;+ RETURN_ERROR_IF(windowLog > ZSTD_WINDOWLOG_MAX, frameParameter_windowTooLarge);+ windowSize = (1ULL << windowLog);+ windowSize += (windowSize >> 3) * (wlByte&7);+ }+ switch(dictIDSizeCode)+ {+ default: assert(0); /* impossible */+ case 0 : break;+ case 1 : dictID = ip[pos]; pos++; break;+ case 2 : dictID = MEM_readLE16(ip+pos); pos+=2; break;+ case 3 : dictID = MEM_readLE32(ip+pos); pos+=4; break;+ }+ switch(fcsID)+ {+ default: assert(0); /* impossible */+ case 0 : if (singleSegment) frameContentSize = ip[pos]; break;+ case 1 : frameContentSize = MEM_readLE16(ip+pos)+256; break;+ case 2 : frameContentSize = MEM_readLE32(ip+pos); break;+ case 3 : frameContentSize = MEM_readLE64(ip+pos); break;+ }+ if (singleSegment) windowSize = frameContentSize;++ zfhPtr->frameType = ZSTD_frame;+ zfhPtr->frameContentSize = frameContentSize;+ zfhPtr->windowSize = windowSize;+ zfhPtr->blockSizeMax = (unsigned) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);+ zfhPtr->dictID = dictID;+ zfhPtr->checksumFlag = checksumFlag;+ }+ return 0;+}++/** ZSTD_getFrameHeader() :+ * decode Frame Header, or require larger `srcSize`.+ * note : this function does not consume input, it only reads it.+ * @return : 0, `zfhPtr` is correctly filled,+ * >0, `srcSize` is too small, value is wanted `srcSize` amount,+ * or an error code, which can be tested using ZSTD_isError() */+size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize)+{+ return ZSTD_getFrameHeader_advanced(zfhPtr, src, srcSize, ZSTD_f_zstd1);+}+++/** ZSTD_getFrameContentSize() :+ * compatible with legacy mode+ * @return : decompressed size of the single frame pointed to be `src` if known, otherwise+ * - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined+ * - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small) */+unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize)+{+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)+ if (ZSTD_isLegacy(src, srcSize)) {+ unsigned long long const ret = ZSTD_getDecompressedSize_legacy(src, srcSize);+ return ret == 0 ? ZSTD_CONTENTSIZE_UNKNOWN : ret;+ }+#endif+ { ZSTD_frameHeader zfh;+ if (ZSTD_getFrameHeader(&zfh, src, srcSize) != 0)+ return ZSTD_CONTENTSIZE_ERROR;+ if (zfh.frameType == ZSTD_skippableFrame) {+ return 0;+ } else {+ return zfh.frameContentSize;+ } }+}++static size_t readSkippableFrameSize(void const* src, size_t srcSize)+{+ size_t const skippableHeaderSize = ZSTD_SKIPPABLEHEADERSIZE;+ U32 sizeU32;++ RETURN_ERROR_IF(srcSize < ZSTD_SKIPPABLEHEADERSIZE, srcSize_wrong);++ sizeU32 = MEM_readLE32((BYTE const*)src + ZSTD_FRAMEIDSIZE);+ RETURN_ERROR_IF((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32,+ frameParameter_unsupported);+ {+ size_t const skippableSize = skippableHeaderSize + sizeU32;+ RETURN_ERROR_IF(skippableSize > srcSize, srcSize_wrong);+ return skippableSize;+ }+}++/** ZSTD_findDecompressedSize() :+ * compatible with legacy mode+ * `srcSize` must be the exact length of some number of ZSTD compressed and/or+ * skippable frames+ * @return : decompressed size of the frames contained */+unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)+{+ unsigned long long totalDstSize = 0;++ while (srcSize >= ZSTD_FRAMEHEADERSIZE_PREFIX) {+ U32 const magicNumber = MEM_readLE32(src);++ if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {+ size_t const skippableSize = readSkippableFrameSize(src, srcSize);+ if (ZSTD_isError(skippableSize)) {+ return ZSTD_CONTENTSIZE_ERROR;+ }+ assert(skippableSize <= srcSize);++ src = (const BYTE *)src + skippableSize;+ srcSize -= skippableSize;+ continue;+ }++ { unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);+ if (ret >= ZSTD_CONTENTSIZE_ERROR) return ret;++ /* check for overflow */+ if (totalDstSize + ret < totalDstSize) return ZSTD_CONTENTSIZE_ERROR;+ totalDstSize += ret;+ }+ { size_t const frameSrcSize = ZSTD_findFrameCompressedSize(src, srcSize);+ if (ZSTD_isError(frameSrcSize)) {+ return ZSTD_CONTENTSIZE_ERROR;+ }++ src = (const BYTE *)src + frameSrcSize;+ srcSize -= frameSrcSize;+ }+ } /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */++ if (srcSize) return ZSTD_CONTENTSIZE_ERROR;++ return totalDstSize;+}++/** ZSTD_getDecompressedSize() :+ * compatible with legacy mode+ * @return : decompressed size if known, 0 otherwise+ note : 0 can mean any of the following :+ - frame content is empty+ - decompressed size field is not present in frame header+ - frame header unknown / not supported+ - frame header not complete (`srcSize` too small) */+unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize)+{+ unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);+ ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_ERROR < ZSTD_CONTENTSIZE_UNKNOWN);+ return (ret >= ZSTD_CONTENTSIZE_ERROR) ? 0 : ret;+}+++/** ZSTD_decodeFrameHeader() :+ * `headerSize` must be the size provided by ZSTD_frameHeaderSize().+ * @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */+static size_t ZSTD_decodeFrameHeader(ZSTD_DCtx* dctx, const void* src, size_t headerSize)+{+ size_t const result = ZSTD_getFrameHeader_advanced(&(dctx->fParams), src, headerSize, dctx->format);+ if (ZSTD_isError(result)) return result; /* invalid header */+ RETURN_ERROR_IF(result>0, srcSize_wrong, "headerSize too small");+#ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION+ /* Skip the dictID check in fuzzing mode, because it makes the search+ * harder.+ */+ RETURN_ERROR_IF(dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID),+ dictionary_wrong);+#endif+ if (dctx->fParams.checksumFlag) XXH64_reset(&dctx->xxhState, 0);+ return 0;+}++static ZSTD_frameSizeInfo ZSTD_errorFrameSizeInfo(size_t ret)+{+ ZSTD_frameSizeInfo frameSizeInfo;+ frameSizeInfo.compressedSize = ret;+ frameSizeInfo.decompressedBound = ZSTD_CONTENTSIZE_ERROR;+ return frameSizeInfo;+}++static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize)+{+ ZSTD_frameSizeInfo frameSizeInfo;+ memset(&frameSizeInfo, 0, sizeof(ZSTD_frameSizeInfo));++#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)+ if (ZSTD_isLegacy(src, srcSize))+ return ZSTD_findFrameSizeInfoLegacy(src, srcSize);+#endif++ if ((srcSize >= ZSTD_SKIPPABLEHEADERSIZE)+ && (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {+ frameSizeInfo.compressedSize = readSkippableFrameSize(src, srcSize);+ assert(ZSTD_isError(frameSizeInfo.compressedSize) ||+ frameSizeInfo.compressedSize <= srcSize);+ return frameSizeInfo;+ } else {+ const BYTE* ip = (const BYTE*)src;+ const BYTE* const ipstart = ip;+ size_t remainingSize = srcSize;+ size_t nbBlocks = 0;+ ZSTD_frameHeader zfh;++ /* Extract Frame Header */+ { size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);+ if (ZSTD_isError(ret))+ return ZSTD_errorFrameSizeInfo(ret);+ if (ret > 0)+ return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));+ }++ ip += zfh.headerSize;+ remainingSize -= zfh.headerSize;++ /* Iterate over each block */+ while (1) {+ blockProperties_t blockProperties;+ size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);+ if (ZSTD_isError(cBlockSize))+ return ZSTD_errorFrameSizeInfo(cBlockSize);++ if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)+ return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));++ ip += ZSTD_blockHeaderSize + cBlockSize;+ remainingSize -= ZSTD_blockHeaderSize + cBlockSize;+ nbBlocks++;++ if (blockProperties.lastBlock) break;+ }++ /* Final frame content checksum */+ if (zfh.checksumFlag) {+ if (remainingSize < 4)+ return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));+ ip += 4;+ }++ frameSizeInfo.compressedSize = ip - ipstart;+ frameSizeInfo.decompressedBound = (zfh.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN)+ ? zfh.frameContentSize+ : nbBlocks * zfh.blockSizeMax;+ return frameSizeInfo;+ }+}++/** ZSTD_findFrameCompressedSize() :+ * compatible with legacy mode+ * `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame+ * `srcSize` must be at least as large as the frame contained+ * @return : the compressed size of the frame starting at `src` */+size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)+{+ ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);+ return frameSizeInfo.compressedSize;+}++/** ZSTD_decompressBound() :+ * compatible with legacy mode+ * `src` must point to the start of a ZSTD frame or a skippeable frame+ * `srcSize` must be at least as large as the frame contained+ * @return : the maximum decompressed size of the compressed source+ */+unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize)+{+ unsigned long long bound = 0;+ /* Iterate over each frame */+ while (srcSize > 0) {+ ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);+ size_t const compressedSize = frameSizeInfo.compressedSize;+ unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;+ if (ZSTD_isError(compressedSize) || decompressedBound == ZSTD_CONTENTSIZE_ERROR)+ return ZSTD_CONTENTSIZE_ERROR;+ assert(srcSize >= compressedSize);+ src = (const BYTE*)src + compressedSize;+ srcSize -= compressedSize;+ bound += decompressedBound;+ }+ return bound;+}+++/*-*************************************************************+ * Frame decoding+ ***************************************************************/+++void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst)+{+ if (dst != dctx->previousDstEnd) { /* not contiguous */+ dctx->dictEnd = dctx->previousDstEnd;+ dctx->virtualStart = (const char*)dst - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart));+ dctx->prefixStart = dst;+ dctx->previousDstEnd = dst;+ }+}++/** ZSTD_insertBlock() :+ insert `src` block into `dctx` history. Useful to track uncompressed blocks. */+size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize)+{+ ZSTD_checkContinuity(dctx, blockStart);+ dctx->previousDstEnd = (const char*)blockStart + blockSize;+ return blockSize;+}+++static size_t ZSTD_copyRawBlock(void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ DEBUGLOG(5, "ZSTD_copyRawBlock");+ if (dst == NULL) {+ if (srcSize == 0) return 0;+ RETURN_ERROR(dstBuffer_null);+ }+ RETURN_ERROR_IF(srcSize > dstCapacity, dstSize_tooSmall);+ memcpy(dst, src, srcSize);+ return srcSize;+}++static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity,+ BYTE b,+ size_t regenSize)+{+ if (dst == NULL) {+ if (regenSize == 0) return 0;+ RETURN_ERROR(dstBuffer_null);+ }+ RETURN_ERROR_IF(regenSize > dstCapacity, dstSize_tooSmall);+ memset(dst, b, regenSize);+ return regenSize;+}+++/*! ZSTD_decompressFrame() :+ * @dctx must be properly initialized+ * will update *srcPtr and *srcSizePtr,+ * to make *srcPtr progress by one frame. */+static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void** srcPtr, size_t *srcSizePtr)+{+ const BYTE* ip = (const BYTE*)(*srcPtr);+ BYTE* const ostart = (BYTE* const)dst;+ BYTE* const oend = ostart + dstCapacity;+ BYTE* op = ostart;+ size_t remainingSrcSize = *srcSizePtr;++ DEBUGLOG(4, "ZSTD_decompressFrame (srcSize:%i)", (int)*srcSizePtr);++ /* check */+ RETURN_ERROR_IF(+ remainingSrcSize < ZSTD_FRAMEHEADERSIZE_MIN+ZSTD_blockHeaderSize,+ srcSize_wrong);++ /* Frame Header */+ { size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_FRAMEHEADERSIZE_PREFIX);+ if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize;+ RETURN_ERROR_IF(remainingSrcSize < frameHeaderSize+ZSTD_blockHeaderSize,+ srcSize_wrong);+ FORWARD_IF_ERROR( ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize) );+ ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize;+ }++ /* Loop on each block */+ while (1) {+ size_t decodedSize;+ blockProperties_t blockProperties;+ size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSrcSize, &blockProperties);+ if (ZSTD_isError(cBlockSize)) return cBlockSize;++ ip += ZSTD_blockHeaderSize;+ remainingSrcSize -= ZSTD_blockHeaderSize;+ RETURN_ERROR_IF(cBlockSize > remainingSrcSize, srcSize_wrong);++ switch(blockProperties.blockType)+ {+ case bt_compressed:+ decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend-op, ip, cBlockSize, /* frame */ 1);+ break;+ case bt_raw :+ decodedSize = ZSTD_copyRawBlock(op, oend-op, ip, cBlockSize);+ break;+ case bt_rle :+ decodedSize = ZSTD_setRleBlock(op, oend-op, *ip, blockProperties.origSize);+ break;+ case bt_reserved :+ default:+ RETURN_ERROR(corruption_detected);+ }++ if (ZSTD_isError(decodedSize)) return decodedSize;+ if (dctx->fParams.checksumFlag)+ XXH64_update(&dctx->xxhState, op, decodedSize);+ op += decodedSize;+ ip += cBlockSize;+ remainingSrcSize -= cBlockSize;+ if (blockProperties.lastBlock) break;+ }++ if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {+ RETURN_ERROR_IF((U64)(op-ostart) != dctx->fParams.frameContentSize,+ corruption_detected);+ }+ if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */+ U32 const checkCalc = (U32)XXH64_digest(&dctx->xxhState);+ U32 checkRead;+ RETURN_ERROR_IF(remainingSrcSize<4, checksum_wrong);+ checkRead = MEM_readLE32(ip);+ RETURN_ERROR_IF(checkRead != checkCalc, checksum_wrong);+ ip += 4;+ remainingSrcSize -= 4;+ }++ /* Allow caller to get size read */+ *srcPtr = ip;+ *srcSizePtr = remainingSrcSize;+ return op-ostart;+}++static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict, size_t dictSize,+ const ZSTD_DDict* ddict)+{+ void* const dststart = dst;+ int moreThan1Frame = 0;++ DEBUGLOG(5, "ZSTD_decompressMultiFrame");+ assert(dict==NULL || ddict==NULL); /* either dict or ddict set, not both */++ if (ddict) {+ dict = ZSTD_DDict_dictContent(ddict);+ dictSize = ZSTD_DDict_dictSize(ddict);+ }++ while (srcSize >= ZSTD_FRAMEHEADERSIZE_PREFIX) {++#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)+ if (ZSTD_isLegacy(src, srcSize)) {+ size_t decodedSize;+ size_t const frameSize = ZSTD_findFrameCompressedSizeLegacy(src, srcSize);+ if (ZSTD_isError(frameSize)) return frameSize;+ RETURN_ERROR_IF(dctx->staticSize, memory_allocation,+ "legacy support is not compatible with static dctx");++ decodedSize = ZSTD_decompressLegacy(dst, dstCapacity, src, frameSize, dict, dictSize);+ if (ZSTD_isError(decodedSize)) return decodedSize;++ assert(decodedSize <=- dstCapacity);+ dst = (BYTE*)dst + decodedSize;+ dstCapacity -= decodedSize;++ src = (const BYTE*)src + frameSize;+ srcSize -= frameSize;++ continue;+ }+#endif++ { U32 const magicNumber = MEM_readLE32(src);+ DEBUGLOG(4, "reading magic number %08X (expecting %08X)",+ (unsigned)magicNumber, ZSTD_MAGICNUMBER);+ if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {+ size_t const skippableSize = readSkippableFrameSize(src, srcSize);+ FORWARD_IF_ERROR(skippableSize);+ assert(skippableSize <= srcSize);++ src = (const BYTE *)src + skippableSize;+ srcSize -= skippableSize;+ continue;+ } }++ if (ddict) {+ /* we were called from ZSTD_decompress_usingDDict */+ FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(dctx, ddict));+ } else {+ /* this will initialize correctly with no dict if dict == NULL, so+ * use this in all cases but ddict */+ FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize));+ }+ ZSTD_checkContinuity(dctx, dst);++ { const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity,+ &src, &srcSize);+ RETURN_ERROR_IF(+ (ZSTD_getErrorCode(res) == ZSTD_error_prefix_unknown)+ && (moreThan1Frame==1),+ srcSize_wrong,+ "at least one frame successfully completed, but following "+ "bytes are garbage: it's more likely to be a srcSize error, "+ "specifying more bytes than compressed size of frame(s). This "+ "error message replaces ERROR(prefix_unknown), which would be "+ "confusing, as the first header is actually correct. Note that "+ "one could be unlucky, it might be a corruption error instead, "+ "happening right at the place where we expect zstd magic "+ "bytes. But this is _much_ less likely than a srcSize field "+ "error.");+ if (ZSTD_isError(res)) return res;+ assert(res <= dstCapacity);+ dst = (BYTE*)dst + res;+ dstCapacity -= res;+ }+ moreThan1Frame = 1;+ } /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */++ RETURN_ERROR_IF(srcSize, srcSize_wrong, "input not entirely consumed");++ return (BYTE*)dst - (BYTE*)dststart;+}++size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict, size_t dictSize)+{+ return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, dict, dictSize, NULL);+}+++static ZSTD_DDict const* ZSTD_getDDict(ZSTD_DCtx* dctx)+{+ switch (dctx->dictUses) {+ default:+ assert(0 /* Impossible */);+ /* fall-through */+ case ZSTD_dont_use:+ ZSTD_clearDict(dctx);+ return NULL;+ case ZSTD_use_indefinitely:+ return dctx->ddict;+ case ZSTD_use_once:+ dctx->dictUses = ZSTD_dont_use;+ return dctx->ddict;+ }+}++size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+ return ZSTD_decompress_usingDDict(dctx, dst, dstCapacity, src, srcSize, ZSTD_getDDict(dctx));+}+++size_t ZSTD_decompress(void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+#if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE>=1)+ size_t regenSize;+ ZSTD_DCtx* const dctx = ZSTD_createDCtx();+ RETURN_ERROR_IF(dctx==NULL, memory_allocation);+ regenSize = ZSTD_decompressDCtx(dctx, dst, dstCapacity, src, srcSize);+ ZSTD_freeDCtx(dctx);+ return regenSize;+#else /* stack mode */+ ZSTD_DCtx dctx;+ ZSTD_initDCtx_internal(&dctx);+ return ZSTD_decompressDCtx(&dctx, dst, dstCapacity, src, srcSize);+#endif+}+++/*-**************************************+* Advanced Streaming Decompression API+* Bufferless and synchronous+****************************************/+size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx) { return dctx->expected; }++ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx) {+ switch(dctx->stage)+ {+ default: /* should not happen */+ assert(0);+ case ZSTDds_getFrameHeaderSize:+ case ZSTDds_decodeFrameHeader:+ return ZSTDnit_frameHeader;+ case ZSTDds_decodeBlockHeader:+ return ZSTDnit_blockHeader;+ case ZSTDds_decompressBlock:+ return ZSTDnit_block;+ case ZSTDds_decompressLastBlock:+ return ZSTDnit_lastBlock;+ case ZSTDds_checkChecksum:+ return ZSTDnit_checksum;+ case ZSTDds_decodeSkippableHeader:+ case ZSTDds_skipFrame:+ return ZSTDnit_skippableFrame;+ }+}++static int ZSTD_isSkipFrame(ZSTD_DCtx* dctx) { return dctx->stage == ZSTDds_skipFrame; }++/** ZSTD_decompressContinue() :+ * srcSize : must be the exact nb of bytes expected (see ZSTD_nextSrcSizeToDecompress())+ * @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity)+ * or an error code, which can be tested using ZSTD_isError() */+size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+ DEBUGLOG(5, "ZSTD_decompressContinue (srcSize:%u)", (unsigned)srcSize);+ /* Sanity check */+ RETURN_ERROR_IF(srcSize != dctx->expected, srcSize_wrong, "not allowed");+ if (dstCapacity) ZSTD_checkContinuity(dctx, dst);++ switch (dctx->stage)+ {+ case ZSTDds_getFrameHeaderSize :+ assert(src != NULL);+ if (dctx->format == ZSTD_f_zstd1) { /* allows header */+ assert(srcSize >= ZSTD_FRAMEIDSIZE); /* to read skippable magic number */+ if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */+ memcpy(dctx->headerBuffer, src, srcSize);+ dctx->expected = ZSTD_SKIPPABLEHEADERSIZE - srcSize; /* remaining to load to get full skippable frame header */+ dctx->stage = ZSTDds_decodeSkippableHeader;+ return 0;+ } }+ dctx->headerSize = ZSTD_frameHeaderSize_internal(src, srcSize, dctx->format);+ if (ZSTD_isError(dctx->headerSize)) return dctx->headerSize;+ memcpy(dctx->headerBuffer, src, srcSize);+ dctx->expected = dctx->headerSize - srcSize;+ dctx->stage = ZSTDds_decodeFrameHeader;+ return 0;++ case ZSTDds_decodeFrameHeader:+ assert(src != NULL);+ memcpy(dctx->headerBuffer + (dctx->headerSize - srcSize), src, srcSize);+ FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));+ dctx->expected = ZSTD_blockHeaderSize;+ dctx->stage = ZSTDds_decodeBlockHeader;+ return 0;++ case ZSTDds_decodeBlockHeader:+ { blockProperties_t bp;+ size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);+ if (ZSTD_isError(cBlockSize)) return cBlockSize;+ dctx->expected = cBlockSize;+ dctx->bType = bp.blockType;+ dctx->rleSize = bp.origSize;+ if (cBlockSize) {+ dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock;+ return 0;+ }+ /* empty block */+ if (bp.lastBlock) {+ if (dctx->fParams.checksumFlag) {+ dctx->expected = 4;+ dctx->stage = ZSTDds_checkChecksum;+ } else {+ dctx->expected = 0; /* end of frame */+ dctx->stage = ZSTDds_getFrameHeaderSize;+ }+ } else {+ dctx->expected = ZSTD_blockHeaderSize; /* jump to next header */+ dctx->stage = ZSTDds_decodeBlockHeader;+ }+ return 0;+ }++ case ZSTDds_decompressLastBlock:+ case ZSTDds_decompressBlock:+ DEBUGLOG(5, "ZSTD_decompressContinue: case ZSTDds_decompressBlock");+ { size_t rSize;+ switch(dctx->bType)+ {+ case bt_compressed:+ DEBUGLOG(5, "ZSTD_decompressContinue: case bt_compressed");+ rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 1);+ break;+ case bt_raw :+ rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize);+ break;+ case bt_rle :+ rSize = ZSTD_setRleBlock(dst, dstCapacity, *(const BYTE*)src, dctx->rleSize);+ break;+ case bt_reserved : /* should never happen */+ default:+ RETURN_ERROR(corruption_detected);+ }+ if (ZSTD_isError(rSize)) return rSize;+ DEBUGLOG(5, "ZSTD_decompressContinue: decoded size from block : %u", (unsigned)rSize);+ dctx->decodedSize += rSize;+ if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, dst, rSize);++ if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */+ DEBUGLOG(4, "ZSTD_decompressContinue: decoded size from frame : %u", (unsigned)dctx->decodedSize);+ RETURN_ERROR_IF(+ dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN+ && dctx->decodedSize != dctx->fParams.frameContentSize,+ corruption_detected);+ if (dctx->fParams.checksumFlag) { /* another round for frame checksum */+ dctx->expected = 4;+ dctx->stage = ZSTDds_checkChecksum;+ } else {+ dctx->expected = 0; /* ends here */+ dctx->stage = ZSTDds_getFrameHeaderSize;+ }+ } else {+ dctx->stage = ZSTDds_decodeBlockHeader;+ dctx->expected = ZSTD_blockHeaderSize;+ dctx->previousDstEnd = (char*)dst + rSize;+ }+ return rSize;+ }++ case ZSTDds_checkChecksum:+ assert(srcSize == 4); /* guaranteed by dctx->expected */+ { U32 const h32 = (U32)XXH64_digest(&dctx->xxhState);+ U32 const check32 = MEM_readLE32(src);+ DEBUGLOG(4, "ZSTD_decompressContinue: checksum : calculated %08X :: %08X read", (unsigned)h32, (unsigned)check32);+ RETURN_ERROR_IF(check32 != h32, checksum_wrong);+ dctx->expected = 0;+ dctx->stage = ZSTDds_getFrameHeaderSize;+ return 0;+ }++ case ZSTDds_decodeSkippableHeader:+ assert(src != NULL);+ assert(srcSize <= ZSTD_SKIPPABLEHEADERSIZE);+ memcpy(dctx->headerBuffer + (ZSTD_SKIPPABLEHEADERSIZE - srcSize), src, srcSize); /* complete skippable header */+ dctx->expected = MEM_readLE32(dctx->headerBuffer + ZSTD_FRAMEIDSIZE); /* note : dctx->expected can grow seriously large, beyond local buffer size */+ dctx->stage = ZSTDds_skipFrame;+ return 0;++ case ZSTDds_skipFrame:+ dctx->expected = 0;+ dctx->stage = ZSTDds_getFrameHeaderSize;+ return 0;++ default:+ assert(0); /* impossible */+ RETURN_ERROR(GENERIC); /* some compiler require default to do something */+ }+}+++static size_t ZSTD_refDictContent(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)+{+ dctx->dictEnd = dctx->previousDstEnd;+ dctx->virtualStart = (const char*)dict - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart));+ dctx->prefixStart = dict;+ dctx->previousDstEnd = (const char*)dict + dictSize;+ return 0;+}++/*! ZSTD_loadDEntropy() :+ * dict : must point at beginning of a valid zstd dictionary.+ * @return : size of entropy tables read */+size_t+ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy,+ const void* const dict, size_t const dictSize)+{+ const BYTE* dictPtr = (const BYTE*)dict;+ const BYTE* const dictEnd = dictPtr + dictSize;++ RETURN_ERROR_IF(dictSize <= 8, dictionary_corrupted);+ assert(MEM_readLE32(dict) == ZSTD_MAGIC_DICTIONARY); /* dict must be valid */+ dictPtr += 8; /* skip header = magic + dictID */++ ZSTD_STATIC_ASSERT(offsetof(ZSTD_entropyDTables_t, OFTable) == offsetof(ZSTD_entropyDTables_t, LLTable) + sizeof(entropy->LLTable));+ ZSTD_STATIC_ASSERT(offsetof(ZSTD_entropyDTables_t, MLTable) == offsetof(ZSTD_entropyDTables_t, OFTable) + sizeof(entropy->OFTable));+ ZSTD_STATIC_ASSERT(sizeof(entropy->LLTable) + sizeof(entropy->OFTable) + sizeof(entropy->MLTable) >= HUF_DECOMPRESS_WORKSPACE_SIZE);+ { void* const workspace = &entropy->LLTable; /* use fse tables as temporary workspace; implies fse tables are grouped together */+ size_t const workspaceSize = sizeof(entropy->LLTable) + sizeof(entropy->OFTable) + sizeof(entropy->MLTable);+#ifdef HUF_FORCE_DECOMPRESS_X1+ /* in minimal huffman, we always use X1 variants */+ size_t const hSize = HUF_readDTableX1_wksp(entropy->hufTable,+ dictPtr, dictEnd - dictPtr,+ workspace, workspaceSize);+#else+ size_t const hSize = HUF_readDTableX2_wksp(entropy->hufTable,+ dictPtr, dictEnd - dictPtr,+ workspace, workspaceSize);+#endif+ RETURN_ERROR_IF(HUF_isError(hSize), dictionary_corrupted);+ dictPtr += hSize;+ }++ { short offcodeNCount[MaxOff+1];+ unsigned offcodeMaxValue = MaxOff, offcodeLog;+ size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(offcodeMaxValue > MaxOff, dictionary_corrupted);+ RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted);+ ZSTD_buildFSETable( entropy->OFTable,+ offcodeNCount, offcodeMaxValue,+ OF_base, OF_bits,+ offcodeLog);+ dictPtr += offcodeHeaderSize;+ }++ { short matchlengthNCount[MaxML+1];+ unsigned matchlengthMaxValue = MaxML, matchlengthLog;+ size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(matchlengthMaxValue > MaxML, dictionary_corrupted);+ RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted);+ ZSTD_buildFSETable( entropy->MLTable,+ matchlengthNCount, matchlengthMaxValue,+ ML_base, ML_bits,+ matchlengthLog);+ dictPtr += matchlengthHeaderSize;+ }++ { short litlengthNCount[MaxLL+1];+ unsigned litlengthMaxValue = MaxLL, litlengthLog;+ size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);+ RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted);+ RETURN_ERROR_IF(litlengthMaxValue > MaxLL, dictionary_corrupted);+ RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted);+ ZSTD_buildFSETable( entropy->LLTable,+ litlengthNCount, litlengthMaxValue,+ LL_base, LL_bits,+ litlengthLog);+ dictPtr += litlengthHeaderSize;+ }++ RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted);+ { int i;+ size_t const dictContentSize = (size_t)(dictEnd - (dictPtr+12));+ for (i=0; i<3; i++) {+ U32 const rep = MEM_readLE32(dictPtr); dictPtr += 4;+ RETURN_ERROR_IF(rep==0 || rep >= dictContentSize,+ dictionary_corrupted);+ entropy->rep[i] = rep;+ } }++ return dictPtr - (const BYTE*)dict;+}++static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)+{+ if (dictSize < 8) return ZSTD_refDictContent(dctx, dict, dictSize);+ { U32 const magic = MEM_readLE32(dict);+ if (magic != ZSTD_MAGIC_DICTIONARY) {+ return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */+ } }+ dctx->dictID = MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE);++ /* load entropy tables */+ { size_t const eSize = ZSTD_loadDEntropy(&dctx->entropy, dict, dictSize);+ RETURN_ERROR_IF(ZSTD_isError(eSize), dictionary_corrupted);+ dict = (const char*)dict + eSize;+ dictSize -= eSize;+ }+ dctx->litEntropy = dctx->fseEntropy = 1;++ /* reference dictionary content */+ return ZSTD_refDictContent(dctx, dict, dictSize);+}++size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)+{+ assert(dctx != NULL);+ dctx->expected = ZSTD_startingInputLength(dctx->format); /* dctx->format must be properly set */+ dctx->stage = ZSTDds_getFrameHeaderSize;+ dctx->decodedSize = 0;+ dctx->previousDstEnd = NULL;+ dctx->prefixStart = NULL;+ dctx->virtualStart = NULL;+ dctx->dictEnd = NULL;+ dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */+ dctx->litEntropy = dctx->fseEntropy = 0;+ dctx->dictID = 0;+ ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));+ memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue)); /* initial repcodes */+ dctx->LLTptr = dctx->entropy.LLTable;+ dctx->MLTptr = dctx->entropy.MLTable;+ dctx->OFTptr = dctx->entropy.OFTable;+ dctx->HUFptr = dctx->entropy.hufTable;+ return 0;+}++size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)+{+ FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) );+ if (dict && dictSize)+ RETURN_ERROR_IF(+ ZSTD_isError(ZSTD_decompress_insertDictionary(dctx, dict, dictSize)),+ dictionary_corrupted);+ return 0;+}+++/* ====== ZSTD_DDict ====== */++size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)+{+ DEBUGLOG(4, "ZSTD_decompressBegin_usingDDict");+ assert(dctx != NULL);+ if (ddict) {+ const char* const dictStart = (const char*)ZSTD_DDict_dictContent(ddict);+ size_t const dictSize = ZSTD_DDict_dictSize(ddict);+ const void* const dictEnd = dictStart + dictSize;+ dctx->ddictIsCold = (dctx->dictEnd != dictEnd);+ DEBUGLOG(4, "DDict is %s",+ dctx->ddictIsCold ? "~cold~" : "hot!");+ }+ FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) );+ if (ddict) { /* NULL ddict is equivalent to no dictionary */+ ZSTD_copyDDictParameters(dctx, ddict);+ }+ return 0;+}++/*! ZSTD_getDictID_fromDict() :+ * Provides the dictID stored within dictionary.+ * if @return == 0, the dictionary is not conformant with Zstandard specification.+ * It can still be loaded, but as a content-only dictionary. */+unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize)+{+ if (dictSize < 8) return 0;+ if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) return 0;+ return MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE);+}++/*! ZSTD_getDictID_fromFrame() :+ * Provides the dictID required to decompress frame stored within `src`.+ * If @return == 0, the dictID could not be decoded.+ * This could for one of the following reasons :+ * - The frame does not require a dictionary (most common case).+ * - The frame was built with dictID intentionally removed.+ * Needed dictionary is a hidden information.+ * Note : this use case also happens when using a non-conformant dictionary.+ * - `srcSize` is too small, and as a result, frame header could not be decoded.+ * Note : possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`.+ * - This is not a Zstandard frame.+ * When identifying the exact failure cause, it's possible to use+ * ZSTD_getFrameHeader(), which will provide a more precise error code. */+unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize)+{+ ZSTD_frameHeader zfp = { 0, 0, 0, ZSTD_frame, 0, 0, 0 };+ size_t const hError = ZSTD_getFrameHeader(&zfp, src, srcSize);+ if (ZSTD_isError(hError)) return 0;+ return zfp.dictID;+}+++/*! ZSTD_decompress_usingDDict() :+* Decompression using a pre-digested Dictionary+* Use dictionary without significant overhead. */+size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_DDict* ddict)+{+ /* pass content and size in case legacy frames are encountered */+ return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize,+ NULL, 0,+ ddict);+}+++/*=====================================+* Streaming decompression+*====================================*/++ZSTD_DStream* ZSTD_createDStream(void)+{+ DEBUGLOG(3, "ZSTD_createDStream");+ return ZSTD_createDStream_advanced(ZSTD_defaultCMem);+}++ZSTD_DStream* ZSTD_initStaticDStream(void *workspace, size_t workspaceSize)+{+ return ZSTD_initStaticDCtx(workspace, workspaceSize);+}++ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem)+{+ return ZSTD_createDCtx_advanced(customMem);+}++size_t ZSTD_freeDStream(ZSTD_DStream* zds)+{+ return ZSTD_freeDCtx(zds);+}+++/* *** Initialization *** */++size_t ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_MAX + ZSTD_blockHeaderSize; }+size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_MAX; }++size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx,+ const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType)+{+ RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);+ ZSTD_clearDict(dctx);+ if (dict && dictSize >= 8) {+ dctx->ddictLocal = ZSTD_createDDict_advanced(dict, dictSize, dictLoadMethod, dictContentType, dctx->customMem);+ RETURN_ERROR_IF(dctx->ddictLocal == NULL, memory_allocation);+ dctx->ddict = dctx->ddictLocal;+ dctx->dictUses = ZSTD_use_indefinitely;+ }+ return 0;+}++size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)+{+ return ZSTD_DCtx_loadDictionary_advanced(dctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto);+}++size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)+{+ return ZSTD_DCtx_loadDictionary_advanced(dctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto);+}++size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)+{+ FORWARD_IF_ERROR(ZSTD_DCtx_loadDictionary_advanced(dctx, prefix, prefixSize, ZSTD_dlm_byRef, dictContentType));+ dctx->dictUses = ZSTD_use_once;+ return 0;+}++size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize)+{+ return ZSTD_DCtx_refPrefix_advanced(dctx, prefix, prefixSize, ZSTD_dct_rawContent);+}+++/* ZSTD_initDStream_usingDict() :+ * return : expected size, aka ZSTD_FRAMEHEADERSIZE_PREFIX.+ * this function cannot fail */+size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize)+{+ DEBUGLOG(4, "ZSTD_initDStream_usingDict");+ FORWARD_IF_ERROR( ZSTD_DCtx_reset(zds, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_DCtx_loadDictionary(zds, dict, dictSize) );+ return ZSTD_FRAMEHEADERSIZE_PREFIX;+}++/* note : this variant can't fail */+size_t ZSTD_initDStream(ZSTD_DStream* zds)+{+ DEBUGLOG(4, "ZSTD_initDStream");+ return ZSTD_initDStream_usingDDict(zds, NULL);+}++/* ZSTD_initDStream_usingDDict() :+ * ddict will just be referenced, and must outlive decompression session+ * this function cannot fail */+size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* dctx, const ZSTD_DDict* ddict)+{+ FORWARD_IF_ERROR( ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only) );+ FORWARD_IF_ERROR( ZSTD_DCtx_refDDict(dctx, ddict) );+ return ZSTD_FRAMEHEADERSIZE_PREFIX;+}++/* ZSTD_resetDStream() :+ * return : expected size, aka ZSTD_FRAMEHEADERSIZE_PREFIX.+ * this function cannot fail */+size_t ZSTD_resetDStream(ZSTD_DStream* dctx)+{+ FORWARD_IF_ERROR(ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only));+ return ZSTD_FRAMEHEADERSIZE_PREFIX;+}+++size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)+{+ RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);+ ZSTD_clearDict(dctx);+ if (ddict) {+ dctx->ddict = ddict;+ dctx->dictUses = ZSTD_use_indefinitely;+ }+ return 0;+}++/* ZSTD_DCtx_setMaxWindowSize() :+ * note : no direct equivalence in ZSTD_DCtx_setParameter,+ * since this version sets windowSize, and the other sets windowLog */+size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize)+{+ ZSTD_bounds const bounds = ZSTD_dParam_getBounds(ZSTD_d_windowLogMax);+ size_t const min = (size_t)1 << bounds.lowerBound;+ size_t const max = (size_t)1 << bounds.upperBound;+ RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);+ RETURN_ERROR_IF(maxWindowSize < min, parameter_outOfBound);+ RETURN_ERROR_IF(maxWindowSize > max, parameter_outOfBound);+ dctx->maxWindowSize = maxWindowSize;+ return 0;+}++size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format)+{+ return ZSTD_DCtx_setParameter(dctx, ZSTD_d_format, format);+}++ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam)+{+ ZSTD_bounds bounds = { 0, 0, 0 };+ switch(dParam) {+ case ZSTD_d_windowLogMax:+ bounds.lowerBound = ZSTD_WINDOWLOG_ABSOLUTEMIN;+ bounds.upperBound = ZSTD_WINDOWLOG_MAX;+ return bounds;+ case ZSTD_d_format:+ bounds.lowerBound = (int)ZSTD_f_zstd1;+ bounds.upperBound = (int)ZSTD_f_zstd1_magicless;+ ZSTD_STATIC_ASSERT(ZSTD_f_zstd1 < ZSTD_f_zstd1_magicless);+ return bounds;+ default:;+ }+ bounds.error = ERROR(parameter_unsupported);+ return bounds;+}++/* ZSTD_dParam_withinBounds:+ * @return 1 if value is within dParam bounds,+ * 0 otherwise */+static int ZSTD_dParam_withinBounds(ZSTD_dParameter dParam, int value)+{+ ZSTD_bounds const bounds = ZSTD_dParam_getBounds(dParam);+ if (ZSTD_isError(bounds.error)) return 0;+ if (value < bounds.lowerBound) return 0;+ if (value > bounds.upperBound) return 0;+ return 1;+}++#define CHECK_DBOUNDS(p,v) { \+ RETURN_ERROR_IF(!ZSTD_dParam_withinBounds(p, v), parameter_outOfBound); \+}++size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value)+{+ RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);+ switch(dParam) {+ case ZSTD_d_windowLogMax:+ if (value == 0) value = ZSTD_WINDOWLOG_LIMIT_DEFAULT;+ CHECK_DBOUNDS(ZSTD_d_windowLogMax, value);+ dctx->maxWindowSize = ((size_t)1) << value;+ return 0;+ case ZSTD_d_format:+ CHECK_DBOUNDS(ZSTD_d_format, value);+ dctx->format = (ZSTD_format_e)value;+ return 0;+ default:;+ }+ RETURN_ERROR(parameter_unsupported);+}++size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset)+{+ if ( (reset == ZSTD_reset_session_only)+ || (reset == ZSTD_reset_session_and_parameters) ) {+ dctx->streamStage = zdss_init;+ dctx->noForwardProgress = 0;+ }+ if ( (reset == ZSTD_reset_parameters)+ || (reset == ZSTD_reset_session_and_parameters) ) {+ RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);+ ZSTD_clearDict(dctx);+ dctx->format = ZSTD_f_zstd1;+ dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;+ }+ return 0;+}+++size_t ZSTD_sizeof_DStream(const ZSTD_DStream* dctx)+{+ return ZSTD_sizeof_DCtx(dctx);+}++size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)+{+ size_t const blockSize = (size_t) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);+ unsigned long long const neededRBSize = windowSize + blockSize + (WILDCOPY_OVERLENGTH * 2);+ unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);+ size_t const minRBSize = (size_t) neededSize;+ RETURN_ERROR_IF((unsigned long long)minRBSize != neededSize,+ frameParameter_windowTooLarge);+ return minRBSize;+}++size_t ZSTD_estimateDStreamSize(size_t windowSize)+{+ size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX);+ size_t const inBuffSize = blockSize; /* no block can be larger */+ size_t const outBuffSize = ZSTD_decodingBufferSize_min(windowSize, ZSTD_CONTENTSIZE_UNKNOWN);+ return ZSTD_estimateDCtxSize() + inBuffSize + outBuffSize;+}++size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize)+{+ U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX; /* note : should be user-selectable, but requires an additional parameter (or a dctx) */+ ZSTD_frameHeader zfh;+ size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize);+ if (ZSTD_isError(err)) return err;+ RETURN_ERROR_IF(err>0, srcSize_wrong);+ RETURN_ERROR_IF(zfh.windowSize > windowSizeMax,+ frameParameter_windowTooLarge);+ return ZSTD_estimateDStreamSize((size_t)zfh.windowSize);+}+++/* ***** Decompression ***** */++MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize)+{+ size_t const length = MIN(dstCapacity, srcSize);+ memcpy(dst, src, length);+ return length;+}+++size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input)+{+ const char* const istart = (const char*)(input->src) + input->pos;+ const char* const iend = (const char*)(input->src) + input->size;+ const char* ip = istart;+ char* const ostart = (char*)(output->dst) + output->pos;+ char* const oend = (char*)(output->dst) + output->size;+ char* op = ostart;+ U32 someMoreWork = 1;++ DEBUGLOG(5, "ZSTD_decompressStream");+ RETURN_ERROR_IF(+ input->pos > input->size,+ srcSize_wrong,+ "forbidden. in: pos: %u vs size: %u",+ (U32)input->pos, (U32)input->size);+ RETURN_ERROR_IF(+ output->pos > output->size,+ dstSize_tooSmall,+ "forbidden. out: pos: %u vs size: %u",+ (U32)output->pos, (U32)output->size);+ DEBUGLOG(5, "input size : %u", (U32)(input->size - input->pos));++ while (someMoreWork) {+ switch(zds->streamStage)+ {+ case zdss_init :+ DEBUGLOG(5, "stage zdss_init => transparent reset ");+ zds->streamStage = zdss_loadHeader;+ zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;+ zds->legacyVersion = 0;+ zds->hostageByte = 0;+ /* fall-through */++ case zdss_loadHeader :+ DEBUGLOG(5, "stage zdss_loadHeader (srcSize : %u)", (U32)(iend - ip));+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)+ if (zds->legacyVersion) {+ RETURN_ERROR_IF(zds->staticSize, memory_allocation,+ "legacy support is incompatible with static dctx");+ { size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, zds->legacyVersion, output, input);+ if (hint==0) zds->streamStage = zdss_init;+ return hint;+ } }+#endif+ { size_t const hSize = ZSTD_getFrameHeader_advanced(&zds->fParams, zds->headerBuffer, zds->lhSize, zds->format);+ DEBUGLOG(5, "header size : %u", (U32)hSize);+ if (ZSTD_isError(hSize)) {+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)+ U32 const legacyVersion = ZSTD_isLegacy(istart, iend-istart);+ if (legacyVersion) {+ ZSTD_DDict const* const ddict = ZSTD_getDDict(zds);+ const void* const dict = ddict ? ZSTD_DDict_dictContent(ddict) : NULL;+ size_t const dictSize = ddict ? ZSTD_DDict_dictSize(ddict) : 0;+ DEBUGLOG(5, "ZSTD_decompressStream: detected legacy version v0.%u", legacyVersion);+ RETURN_ERROR_IF(zds->staticSize, memory_allocation,+ "legacy support is incompatible with static dctx");+ FORWARD_IF_ERROR(ZSTD_initLegacyStream(&zds->legacyContext,+ zds->previousLegacyVersion, legacyVersion,+ dict, dictSize));+ zds->legacyVersion = zds->previousLegacyVersion = legacyVersion;+ { size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, legacyVersion, output, input);+ if (hint==0) zds->streamStage = zdss_init; /* or stay in stage zdss_loadHeader */+ return hint;+ } }+#endif+ return hSize; /* error */+ }+ if (hSize != 0) { /* need more input */+ size_t const toLoad = hSize - zds->lhSize; /* if hSize!=0, hSize > zds->lhSize */+ size_t const remainingInput = (size_t)(iend-ip);+ assert(iend >= ip);+ if (toLoad > remainingInput) { /* not enough input to load full header */+ if (remainingInput > 0) {+ memcpy(zds->headerBuffer + zds->lhSize, ip, remainingInput);+ zds->lhSize += remainingInput;+ }+ input->pos = input->size;+ return (MAX(ZSTD_FRAMEHEADERSIZE_MIN, hSize) - zds->lhSize) + ZSTD_blockHeaderSize; /* remaining header bytes + next block header */+ }+ assert(ip != NULL);+ memcpy(zds->headerBuffer + zds->lhSize, ip, toLoad); zds->lhSize = hSize; ip += toLoad;+ break;+ } }++ /* check for single-pass mode opportunity */+ if (zds->fParams.frameContentSize && zds->fParams.windowSize /* skippable frame if == 0 */+ && (U64)(size_t)(oend-op) >= zds->fParams.frameContentSize) {+ size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend-istart);+ if (cSize <= (size_t)(iend-istart)) {+ /* shortcut : using single-pass mode */+ size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, oend-op, istart, cSize, ZSTD_getDDict(zds));+ if (ZSTD_isError(decompressedSize)) return decompressedSize;+ DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()")+ ip = istart + cSize;+ op += decompressedSize;+ zds->expected = 0;+ zds->streamStage = zdss_init;+ someMoreWork = 0;+ break;+ } }++ /* Consume header (see ZSTDds_decodeFrameHeader) */+ DEBUGLOG(4, "Consume header");+ FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(zds, ZSTD_getDDict(zds)));++ if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */+ zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE);+ zds->stage = ZSTDds_skipFrame;+ } else {+ FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize));+ zds->expected = ZSTD_blockHeaderSize;+ zds->stage = ZSTDds_decodeBlockHeader;+ }++ /* control buffer memory usage */+ DEBUGLOG(4, "Control max memory usage (%u KB <= max %u KB)",+ (U32)(zds->fParams.windowSize >>10),+ (U32)(zds->maxWindowSize >> 10) );+ zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);+ RETURN_ERROR_IF(zds->fParams.windowSize > zds->maxWindowSize,+ frameParameter_windowTooLarge);++ /* Adapt buffer sizes to frame header instructions */+ { size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);+ size_t const neededOutBuffSize = ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize);+ if ((zds->inBuffSize < neededInBuffSize) || (zds->outBuffSize < neededOutBuffSize)) {+ size_t const bufferSize = neededInBuffSize + neededOutBuffSize;+ DEBUGLOG(4, "inBuff : from %u to %u",+ (U32)zds->inBuffSize, (U32)neededInBuffSize);+ DEBUGLOG(4, "outBuff : from %u to %u",+ (U32)zds->outBuffSize, (U32)neededOutBuffSize);+ if (zds->staticSize) { /* static DCtx */+ DEBUGLOG(4, "staticSize : %u", (U32)zds->staticSize);+ assert(zds->staticSize >= sizeof(ZSTD_DCtx)); /* controlled at init */+ RETURN_ERROR_IF(+ bufferSize > zds->staticSize - sizeof(ZSTD_DCtx),+ memory_allocation);+ } else {+ ZSTD_free(zds->inBuff, zds->customMem);+ zds->inBuffSize = 0;+ zds->outBuffSize = 0;+ zds->inBuff = (char*)ZSTD_malloc(bufferSize, zds->customMem);+ RETURN_ERROR_IF(zds->inBuff == NULL, memory_allocation);+ }+ zds->inBuffSize = neededInBuffSize;+ zds->outBuff = zds->inBuff + zds->inBuffSize;+ zds->outBuffSize = neededOutBuffSize;+ } }+ zds->streamStage = zdss_read;+ /* fall-through */++ case zdss_read:+ DEBUGLOG(5, "stage zdss_read");+ { size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds);+ DEBUGLOG(5, "neededInSize = %u", (U32)neededInSize);+ if (neededInSize==0) { /* end of frame */+ zds->streamStage = zdss_init;+ someMoreWork = 0;+ break;+ }+ if ((size_t)(iend-ip) >= neededInSize) { /* decode directly from src */+ int const isSkipFrame = ZSTD_isSkipFrame(zds);+ size_t const decodedSize = ZSTD_decompressContinue(zds,+ zds->outBuff + zds->outStart, (isSkipFrame ? 0 : zds->outBuffSize - zds->outStart),+ ip, neededInSize);+ if (ZSTD_isError(decodedSize)) return decodedSize;+ ip += neededInSize;+ if (!decodedSize && !isSkipFrame) break; /* this was just a header */+ zds->outEnd = zds->outStart + decodedSize;+ zds->streamStage = zdss_flush;+ break;+ } }+ if (ip==iend) { someMoreWork = 0; break; } /* no more input */+ zds->streamStage = zdss_load;+ /* fall-through */++ case zdss_load:+ { size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds);+ size_t const toLoad = neededInSize - zds->inPos;+ int const isSkipFrame = ZSTD_isSkipFrame(zds);+ size_t loadedSize;+ if (isSkipFrame) {+ loadedSize = MIN(toLoad, (size_t)(iend-ip));+ } else {+ RETURN_ERROR_IF(toLoad > zds->inBuffSize - zds->inPos,+ corruption_detected,+ "should never happen");+ loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend-ip);+ }+ ip += loadedSize;+ zds->inPos += loadedSize;+ if (loadedSize < toLoad) { someMoreWork = 0; break; } /* not enough input, wait for more */++ /* decode loaded input */+ { size_t const decodedSize = ZSTD_decompressContinue(zds,+ zds->outBuff + zds->outStart, zds->outBuffSize - zds->outStart,+ zds->inBuff, neededInSize);+ if (ZSTD_isError(decodedSize)) return decodedSize;+ zds->inPos = 0; /* input is consumed */+ if (!decodedSize && !isSkipFrame) { zds->streamStage = zdss_read; break; } /* this was just a header */+ zds->outEnd = zds->outStart + decodedSize;+ } }+ zds->streamStage = zdss_flush;+ /* fall-through */++ case zdss_flush:+ { size_t const toFlushSize = zds->outEnd - zds->outStart;+ size_t const flushedSize = ZSTD_limitCopy(op, oend-op, zds->outBuff + zds->outStart, toFlushSize);+ op += flushedSize;+ zds->outStart += flushedSize;+ if (flushedSize == toFlushSize) { /* flush completed */+ zds->streamStage = zdss_read;+ if ( (zds->outBuffSize < zds->fParams.frameContentSize)+ && (zds->outStart + zds->fParams.blockSizeMax > zds->outBuffSize) ) {+ DEBUGLOG(5, "restart filling outBuff from beginning (left:%i, needed:%u)",+ (int)(zds->outBuffSize - zds->outStart),+ (U32)zds->fParams.blockSizeMax);+ zds->outStart = zds->outEnd = 0;+ }+ break;+ } }+ /* cannot complete flush */+ someMoreWork = 0;+ break;++ default:+ assert(0); /* impossible */+ RETURN_ERROR(GENERIC); /* some compiler require default to do something */+ } }++ /* result */+ input->pos = (size_t)(ip - (const char*)(input->src));+ output->pos = (size_t)(op - (char*)(output->dst));+ if ((ip==istart) && (op==ostart)) { /* no forward progress */+ zds->noForwardProgress ++;+ if (zds->noForwardProgress >= ZSTD_NO_FORWARD_PROGRESS_MAX) {+ RETURN_ERROR_IF(op==oend, dstSize_tooSmall);+ RETURN_ERROR_IF(ip==iend, srcSize_wrong);+ assert(0);+ }+ } else {+ zds->noForwardProgress = 0;+ }+ { size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds);+ if (!nextSrcSizeHint) { /* frame fully decoded */+ if (zds->outEnd == zds->outStart) { /* output fully flushed */+ if (zds->hostageByte) {+ if (input->pos >= input->size) {+ /* can't release hostage (not present) */+ zds->streamStage = zdss_read;+ return 1;+ }+ input->pos++; /* release hostage */+ } /* zds->hostageByte */+ return 0;+ } /* zds->outEnd == zds->outStart */+ if (!zds->hostageByte) { /* output not fully flushed; keep last byte as hostage; will be released when all output is flushed */+ input->pos--; /* note : pos > 0, otherwise, impossible to finish reading last block */+ zds->hostageByte=1;+ }+ return 1;+ } /* nextSrcSizeHint==0 */+ nextSrcSizeHint += ZSTD_blockHeaderSize * (ZSTD_nextInputType(zds) == ZSTDnit_block); /* preload header of next block */+ assert(zds->inPos <= nextSrcSizeHint);+ nextSrcSizeHint -= zds->inPos; /* part already loaded*/+ return nextSrcSizeHint;+ }+}++size_t ZSTD_decompressStream_simpleArgs (+ ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity, size_t* dstPos,+ const void* src, size_t srcSize, size_t* srcPos)+{+ ZSTD_outBuffer output = { dst, dstCapacity, *dstPos };+ ZSTD_inBuffer input = { src, srcSize, *srcPos };+ /* ZSTD_compress_generic() will check validity of dstPos and srcPos */+ size_t const cErr = ZSTD_decompressStream(dctx, &output, &input);+ *dstPos = output.pos;+ *srcPos = input.pos;+ return cErr; }
+ zstd/lib/decompress/zstd_decompress_block.c view
@@ -0,0 +1,1311 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */++/* zstd_decompress_block :+ * this module takes care of decompressing _compressed_ block */++/*-*******************************************************+* Dependencies+*********************************************************/+#include <string.h> /* memcpy, memmove, memset */+#include "compiler.h" /* prefetch */+#include "cpu.h" /* bmi2 */+#include "mem.h" /* low level memory routines */+#define FSE_STATIC_LINKING_ONLY+#include "fse.h"+#define HUF_STATIC_LINKING_ONLY+#include "huf.h"+#include "zstd_internal.h"+#include "zstd_decompress_internal.h" /* ZSTD_DCtx */+#include "zstd_ddict.h" /* ZSTD_DDictDictContent */+#include "zstd_decompress_block.h"++/*_*******************************************************+* Macros+**********************************************************/++/* These two optional macros force the use one way or another of the two+ * ZSTD_decompressSequences implementations. You can't force in both directions+ * at the same time.+ */+#if defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \+ defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)+#error "Cannot force the use of the short and the long ZSTD_decompressSequences variants!"+#endif+++/*_*******************************************************+* Memory operations+**********************************************************/+static void ZSTD_copy4(void* dst, const void* src) { memcpy(dst, src, 4); }+++/*-*************************************************************+ * Block decoding+ ***************************************************************/++/*! ZSTD_getcBlockSize() :+ * Provides the size of compressed block from block header `src` */+size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,+ blockProperties_t* bpPtr)+{+ RETURN_ERROR_IF(srcSize < ZSTD_blockHeaderSize, srcSize_wrong);++ { U32 const cBlockHeader = MEM_readLE24(src);+ U32 const cSize = cBlockHeader >> 3;+ bpPtr->lastBlock = cBlockHeader & 1;+ bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);+ bpPtr->origSize = cSize; /* only useful for RLE */+ if (bpPtr->blockType == bt_rle) return 1;+ RETURN_ERROR_IF(bpPtr->blockType == bt_reserved, corruption_detected);+ return cSize;+ }+}+++/* Hidden declaration for fullbench */+size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,+ const void* src, size_t srcSize);+/*! ZSTD_decodeLiteralsBlock() :+ * @return : nb of bytes read from src (< srcSize )+ * note : symbol not declared but exposed for fullbench */+size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,+ const void* src, size_t srcSize) /* note : srcSize < BLOCKSIZE */+{+ RETURN_ERROR_IF(srcSize < MIN_CBLOCK_SIZE, corruption_detected);++ { const BYTE* const istart = (const BYTE*) src;+ symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);++ switch(litEncType)+ {+ case set_repeat:+ RETURN_ERROR_IF(dctx->litEntropy==0, dictionary_corrupted);+ /* fall-through */++ case set_compressed:+ RETURN_ERROR_IF(srcSize < 5, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3");+ { size_t lhSize, litSize, litCSize;+ U32 singleStream=0;+ U32 const lhlCode = (istart[0] >> 2) & 3;+ U32 const lhc = MEM_readLE32(istart);+ size_t hufSuccess;+ switch(lhlCode)+ {+ case 0: case 1: default: /* note : default is impossible, since lhlCode into [0..3] */+ /* 2 - 2 - 10 - 10 */+ singleStream = !lhlCode;+ lhSize = 3;+ litSize = (lhc >> 4) & 0x3FF;+ litCSize = (lhc >> 14) & 0x3FF;+ break;+ case 2:+ /* 2 - 2 - 14 - 14 */+ lhSize = 4;+ litSize = (lhc >> 4) & 0x3FFF;+ litCSize = lhc >> 18;+ break;+ case 3:+ /* 2 - 2 - 18 - 18 */+ lhSize = 5;+ litSize = (lhc >> 4) & 0x3FFFF;+ litCSize = (lhc >> 22) + (istart[4] << 10);+ break;+ }+ RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected);+ RETURN_ERROR_IF(litCSize + lhSize > srcSize, corruption_detected);++ /* prefetch huffman table if cold */+ if (dctx->ddictIsCold && (litSize > 768 /* heuristic */)) {+ PREFETCH_AREA(dctx->HUFptr, sizeof(dctx->entropy.hufTable));+ }++ if (litEncType==set_repeat) {+ if (singleStream) {+ hufSuccess = HUF_decompress1X_usingDTable_bmi2(+ dctx->litBuffer, litSize, istart+lhSize, litCSize,+ dctx->HUFptr, dctx->bmi2);+ } else {+ hufSuccess = HUF_decompress4X_usingDTable_bmi2(+ dctx->litBuffer, litSize, istart+lhSize, litCSize,+ dctx->HUFptr, dctx->bmi2);+ }+ } else {+ if (singleStream) {+#if defined(HUF_FORCE_DECOMPRESS_X2)+ hufSuccess = HUF_decompress1X_DCtx_wksp(+ dctx->entropy.hufTable, dctx->litBuffer, litSize,+ istart+lhSize, litCSize, dctx->workspace,+ sizeof(dctx->workspace));+#else+ hufSuccess = HUF_decompress1X1_DCtx_wksp_bmi2(+ dctx->entropy.hufTable, dctx->litBuffer, litSize,+ istart+lhSize, litCSize, dctx->workspace,+ sizeof(dctx->workspace), dctx->bmi2);+#endif+ } else {+ hufSuccess = HUF_decompress4X_hufOnly_wksp_bmi2(+ dctx->entropy.hufTable, dctx->litBuffer, litSize,+ istart+lhSize, litCSize, dctx->workspace,+ sizeof(dctx->workspace), dctx->bmi2);+ }+ }++ RETURN_ERROR_IF(HUF_isError(hufSuccess), corruption_detected);++ dctx->litPtr = dctx->litBuffer;+ dctx->litSize = litSize;+ dctx->litEntropy = 1;+ if (litEncType==set_compressed) dctx->HUFptr = dctx->entropy.hufTable;+ memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);+ return litCSize + lhSize;+ }++ case set_basic:+ { size_t litSize, lhSize;+ U32 const lhlCode = ((istart[0]) >> 2) & 3;+ switch(lhlCode)+ {+ case 0: case 2: default: /* note : default is impossible, since lhlCode into [0..3] */+ lhSize = 1;+ litSize = istart[0] >> 3;+ break;+ case 1:+ lhSize = 2;+ litSize = MEM_readLE16(istart) >> 4;+ break;+ case 3:+ lhSize = 3;+ litSize = MEM_readLE24(istart) >> 4;+ break;+ }++ if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */+ RETURN_ERROR_IF(litSize+lhSize > srcSize, corruption_detected);+ memcpy(dctx->litBuffer, istart+lhSize, litSize);+ dctx->litPtr = dctx->litBuffer;+ dctx->litSize = litSize;+ memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);+ return lhSize+litSize;+ }+ /* direct reference into compressed stream */+ dctx->litPtr = istart+lhSize;+ dctx->litSize = litSize;+ return lhSize+litSize;+ }++ case set_rle:+ { U32 const lhlCode = ((istart[0]) >> 2) & 3;+ size_t litSize, lhSize;+ switch(lhlCode)+ {+ case 0: case 2: default: /* note : default is impossible, since lhlCode into [0..3] */+ lhSize = 1;+ litSize = istart[0] >> 3;+ break;+ case 1:+ lhSize = 2;+ litSize = MEM_readLE16(istart) >> 4;+ break;+ case 3:+ lhSize = 3;+ litSize = MEM_readLE24(istart) >> 4;+ RETURN_ERROR_IF(srcSize<4, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4");+ break;+ }+ RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected);+ memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);+ dctx->litPtr = dctx->litBuffer;+ dctx->litSize = litSize;+ return lhSize+1;+ }+ default:+ RETURN_ERROR(corruption_detected, "impossible");+ }+ }+}++/* Default FSE distribution tables.+ * These are pre-calculated FSE decoding tables using default distributions as defined in specification :+ * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#default-distributions+ * They were generated programmatically with following method :+ * - start from default distributions, present in /lib/common/zstd_internal.h+ * - generate tables normally, using ZSTD_buildFSETable()+ * - printout the content of tables+ * - pretify output, report below, test with fuzzer to ensure it's correct */++/* Default FSE distribution table for Literal Lengths */+static const ZSTD_seqSymbol LL_defaultDTable[(1<<LL_DEFAULTNORMLOG)+1] = {+ { 1, 1, 1, LL_DEFAULTNORMLOG}, /* header : fastMode, tableLog */+ /* nextState, nbAddBits, nbBits, baseVal */+ { 0, 0, 4, 0}, { 16, 0, 4, 0},+ { 32, 0, 5, 1}, { 0, 0, 5, 3},+ { 0, 0, 5, 4}, { 0, 0, 5, 6},+ { 0, 0, 5, 7}, { 0, 0, 5, 9},+ { 0, 0, 5, 10}, { 0, 0, 5, 12},+ { 0, 0, 6, 14}, { 0, 1, 5, 16},+ { 0, 1, 5, 20}, { 0, 1, 5, 22},+ { 0, 2, 5, 28}, { 0, 3, 5, 32},+ { 0, 4, 5, 48}, { 32, 6, 5, 64},+ { 0, 7, 5, 128}, { 0, 8, 6, 256},+ { 0, 10, 6, 1024}, { 0, 12, 6, 4096},+ { 32, 0, 4, 0}, { 0, 0, 4, 1},+ { 0, 0, 5, 2}, { 32, 0, 5, 4},+ { 0, 0, 5, 5}, { 32, 0, 5, 7},+ { 0, 0, 5, 8}, { 32, 0, 5, 10},+ { 0, 0, 5, 11}, { 0, 0, 6, 13},+ { 32, 1, 5, 16}, { 0, 1, 5, 18},+ { 32, 1, 5, 22}, { 0, 2, 5, 24},+ { 32, 3, 5, 32}, { 0, 3, 5, 40},+ { 0, 6, 4, 64}, { 16, 6, 4, 64},+ { 32, 7, 5, 128}, { 0, 9, 6, 512},+ { 0, 11, 6, 2048}, { 48, 0, 4, 0},+ { 16, 0, 4, 1}, { 32, 0, 5, 2},+ { 32, 0, 5, 3}, { 32, 0, 5, 5},+ { 32, 0, 5, 6}, { 32, 0, 5, 8},+ { 32, 0, 5, 9}, { 32, 0, 5, 11},+ { 32, 0, 5, 12}, { 0, 0, 6, 15},+ { 32, 1, 5, 18}, { 32, 1, 5, 20},+ { 32, 2, 5, 24}, { 32, 2, 5, 28},+ { 32, 3, 5, 40}, { 32, 4, 5, 48},+ { 0, 16, 6,65536}, { 0, 15, 6,32768},+ { 0, 14, 6,16384}, { 0, 13, 6, 8192},+}; /* LL_defaultDTable */++/* Default FSE distribution table for Offset Codes */+static const ZSTD_seqSymbol OF_defaultDTable[(1<<OF_DEFAULTNORMLOG)+1] = {+ { 1, 1, 1, OF_DEFAULTNORMLOG}, /* header : fastMode, tableLog */+ /* nextState, nbAddBits, nbBits, baseVal */+ { 0, 0, 5, 0}, { 0, 6, 4, 61},+ { 0, 9, 5, 509}, { 0, 15, 5,32765},+ { 0, 21, 5,2097149}, { 0, 3, 5, 5},+ { 0, 7, 4, 125}, { 0, 12, 5, 4093},+ { 0, 18, 5,262141}, { 0, 23, 5,8388605},+ { 0, 5, 5, 29}, { 0, 8, 4, 253},+ { 0, 14, 5,16381}, { 0, 20, 5,1048573},+ { 0, 2, 5, 1}, { 16, 7, 4, 125},+ { 0, 11, 5, 2045}, { 0, 17, 5,131069},+ { 0, 22, 5,4194301}, { 0, 4, 5, 13},+ { 16, 8, 4, 253}, { 0, 13, 5, 8189},+ { 0, 19, 5,524285}, { 0, 1, 5, 1},+ { 16, 6, 4, 61}, { 0, 10, 5, 1021},+ { 0, 16, 5,65533}, { 0, 28, 5,268435453},+ { 0, 27, 5,134217725}, { 0, 26, 5,67108861},+ { 0, 25, 5,33554429}, { 0, 24, 5,16777213},+}; /* OF_defaultDTable */+++/* Default FSE distribution table for Match Lengths */+static const ZSTD_seqSymbol ML_defaultDTable[(1<<ML_DEFAULTNORMLOG)+1] = {+ { 1, 1, 1, ML_DEFAULTNORMLOG}, /* header : fastMode, tableLog */+ /* nextState, nbAddBits, nbBits, baseVal */+ { 0, 0, 6, 3}, { 0, 0, 4, 4},+ { 32, 0, 5, 5}, { 0, 0, 5, 6},+ { 0, 0, 5, 8}, { 0, 0, 5, 9},+ { 0, 0, 5, 11}, { 0, 0, 6, 13},+ { 0, 0, 6, 16}, { 0, 0, 6, 19},+ { 0, 0, 6, 22}, { 0, 0, 6, 25},+ { 0, 0, 6, 28}, { 0, 0, 6, 31},+ { 0, 0, 6, 34}, { 0, 1, 6, 37},+ { 0, 1, 6, 41}, { 0, 2, 6, 47},+ { 0, 3, 6, 59}, { 0, 4, 6, 83},+ { 0, 7, 6, 131}, { 0, 9, 6, 515},+ { 16, 0, 4, 4}, { 0, 0, 4, 5},+ { 32, 0, 5, 6}, { 0, 0, 5, 7},+ { 32, 0, 5, 9}, { 0, 0, 5, 10},+ { 0, 0, 6, 12}, { 0, 0, 6, 15},+ { 0, 0, 6, 18}, { 0, 0, 6, 21},+ { 0, 0, 6, 24}, { 0, 0, 6, 27},+ { 0, 0, 6, 30}, { 0, 0, 6, 33},+ { 0, 1, 6, 35}, { 0, 1, 6, 39},+ { 0, 2, 6, 43}, { 0, 3, 6, 51},+ { 0, 4, 6, 67}, { 0, 5, 6, 99},+ { 0, 8, 6, 259}, { 32, 0, 4, 4},+ { 48, 0, 4, 4}, { 16, 0, 4, 5},+ { 32, 0, 5, 7}, { 32, 0, 5, 8},+ { 32, 0, 5, 10}, { 32, 0, 5, 11},+ { 0, 0, 6, 14}, { 0, 0, 6, 17},+ { 0, 0, 6, 20}, { 0, 0, 6, 23},+ { 0, 0, 6, 26}, { 0, 0, 6, 29},+ { 0, 0, 6, 32}, { 0, 16, 6,65539},+ { 0, 15, 6,32771}, { 0, 14, 6,16387},+ { 0, 13, 6, 8195}, { 0, 12, 6, 4099},+ { 0, 11, 6, 2051}, { 0, 10, 6, 1027},+}; /* ML_defaultDTable */+++static void ZSTD_buildSeqTable_rle(ZSTD_seqSymbol* dt, U32 baseValue, U32 nbAddBits)+{+ void* ptr = dt;+ ZSTD_seqSymbol_header* const DTableH = (ZSTD_seqSymbol_header*)ptr;+ ZSTD_seqSymbol* const cell = dt + 1;++ DTableH->tableLog = 0;+ DTableH->fastMode = 0;++ cell->nbBits = 0;+ cell->nextState = 0;+ assert(nbAddBits < 255);+ cell->nbAdditionalBits = (BYTE)nbAddBits;+ cell->baseValue = baseValue;+}+++/* ZSTD_buildFSETable() :+ * generate FSE decoding table for one symbol (ll, ml or off)+ * cannot fail if input is valid =>+ * all inputs are presumed validated at this stage */+void+ZSTD_buildFSETable(ZSTD_seqSymbol* dt,+ const short* normalizedCounter, unsigned maxSymbolValue,+ const U32* baseValue, const U32* nbAdditionalBits,+ unsigned tableLog)+{+ ZSTD_seqSymbol* const tableDecode = dt+1;+ U16 symbolNext[MaxSeq+1];++ U32 const maxSV1 = maxSymbolValue + 1;+ U32 const tableSize = 1 << tableLog;+ U32 highThreshold = tableSize-1;++ /* Sanity Checks */+ assert(maxSymbolValue <= MaxSeq);+ assert(tableLog <= MaxFSELog);++ /* Init, lay down lowprob symbols */+ { ZSTD_seqSymbol_header DTableH;+ DTableH.tableLog = tableLog;+ DTableH.fastMode = 1;+ { S16 const largeLimit= (S16)(1 << (tableLog-1));+ U32 s;+ for (s=0; s<maxSV1; s++) {+ if (normalizedCounter[s]==-1) {+ tableDecode[highThreshold--].baseValue = s;+ symbolNext[s] = 1;+ } else {+ if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;+ symbolNext[s] = normalizedCounter[s];+ } } }+ memcpy(dt, &DTableH, sizeof(DTableH));+ }++ /* Spread symbols */+ { U32 const tableMask = tableSize-1;+ U32 const step = FSE_TABLESTEP(tableSize);+ U32 s, position = 0;+ for (s=0; s<maxSV1; s++) {+ int i;+ for (i=0; i<normalizedCounter[s]; i++) {+ tableDecode[position].baseValue = s;+ position = (position + step) & tableMask;+ while (position > highThreshold) position = (position + step) & tableMask; /* lowprob area */+ } }+ assert(position == 0); /* position must reach all cells once, otherwise normalizedCounter is incorrect */+ }++ /* Build Decoding table */+ { U32 u;+ for (u=0; u<tableSize; u++) {+ U32 const symbol = tableDecode[u].baseValue;+ U32 const nextState = symbolNext[symbol]++;+ tableDecode[u].nbBits = (BYTE) (tableLog - BIT_highbit32(nextState) );+ tableDecode[u].nextState = (U16) ( (nextState << tableDecode[u].nbBits) - tableSize);+ assert(nbAdditionalBits[symbol] < 255);+ tableDecode[u].nbAdditionalBits = (BYTE)nbAdditionalBits[symbol];+ tableDecode[u].baseValue = baseValue[symbol];+ } }+}+++/*! ZSTD_buildSeqTable() :+ * @return : nb bytes read from src,+ * or an error code if it fails */+static size_t ZSTD_buildSeqTable(ZSTD_seqSymbol* DTableSpace, const ZSTD_seqSymbol** DTablePtr,+ symbolEncodingType_e type, unsigned max, U32 maxLog,+ const void* src, size_t srcSize,+ const U32* baseValue, const U32* nbAdditionalBits,+ const ZSTD_seqSymbol* defaultTable, U32 flagRepeatTable,+ int ddictIsCold, int nbSeq)+{+ switch(type)+ {+ case set_rle :+ RETURN_ERROR_IF(!srcSize, srcSize_wrong);+ RETURN_ERROR_IF((*(const BYTE*)src) > max, corruption_detected);+ { U32 const symbol = *(const BYTE*)src;+ U32 const baseline = baseValue[symbol];+ U32 const nbBits = nbAdditionalBits[symbol];+ ZSTD_buildSeqTable_rle(DTableSpace, baseline, nbBits);+ }+ *DTablePtr = DTableSpace;+ return 1;+ case set_basic :+ *DTablePtr = defaultTable;+ return 0;+ case set_repeat:+ RETURN_ERROR_IF(!flagRepeatTable, corruption_detected);+ /* prefetch FSE table if used */+ if (ddictIsCold && (nbSeq > 24 /* heuristic */)) {+ const void* const pStart = *DTablePtr;+ size_t const pSize = sizeof(ZSTD_seqSymbol) * (SEQSYMBOL_TABLE_SIZE(maxLog));+ PREFETCH_AREA(pStart, pSize);+ }+ return 0;+ case set_compressed :+ { unsigned tableLog;+ S16 norm[MaxSeq+1];+ size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);+ RETURN_ERROR_IF(FSE_isError(headerSize), corruption_detected);+ RETURN_ERROR_IF(tableLog > maxLog, corruption_detected);+ ZSTD_buildFSETable(DTableSpace, norm, max, baseValue, nbAdditionalBits, tableLog);+ *DTablePtr = DTableSpace;+ return headerSize;+ }+ default :+ assert(0);+ RETURN_ERROR(GENERIC, "impossible");+ }+}++size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,+ const void* src, size_t srcSize)+{+ const BYTE* const istart = (const BYTE* const)src;+ const BYTE* const iend = istart + srcSize;+ const BYTE* ip = istart;+ int nbSeq;+ DEBUGLOG(5, "ZSTD_decodeSeqHeaders");++ /* check */+ RETURN_ERROR_IF(srcSize < MIN_SEQUENCES_SIZE, srcSize_wrong);++ /* SeqHead */+ nbSeq = *ip++;+ if (!nbSeq) {+ *nbSeqPtr=0;+ RETURN_ERROR_IF(srcSize != 1, srcSize_wrong);+ return 1;+ }+ if (nbSeq > 0x7F) {+ if (nbSeq == 0xFF) {+ RETURN_ERROR_IF(ip+2 > iend, srcSize_wrong);+ nbSeq = MEM_readLE16(ip) + LONGNBSEQ, ip+=2;+ } else {+ RETURN_ERROR_IF(ip >= iend, srcSize_wrong);+ nbSeq = ((nbSeq-0x80)<<8) + *ip++;+ }+ }+ *nbSeqPtr = nbSeq;++ /* FSE table descriptors */+ RETURN_ERROR_IF(ip+4 > iend, srcSize_wrong); /* minimum possible size */+ { symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);+ symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);+ symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);+ ip++;++ /* Build DTables */+ { size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr,+ LLtype, MaxLL, LLFSELog,+ ip, iend-ip,+ LL_base, LL_bits,+ LL_defaultDTable, dctx->fseEntropy,+ dctx->ddictIsCold, nbSeq);+ RETURN_ERROR_IF(ZSTD_isError(llhSize), corruption_detected);+ ip += llhSize;+ }++ { size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr,+ OFtype, MaxOff, OffFSELog,+ ip, iend-ip,+ OF_base, OF_bits,+ OF_defaultDTable, dctx->fseEntropy,+ dctx->ddictIsCold, nbSeq);+ RETURN_ERROR_IF(ZSTD_isError(ofhSize), corruption_detected);+ ip += ofhSize;+ }++ { size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr,+ MLtype, MaxML, MLFSELog,+ ip, iend-ip,+ ML_base, ML_bits,+ ML_defaultDTable, dctx->fseEntropy,+ dctx->ddictIsCold, nbSeq);+ RETURN_ERROR_IF(ZSTD_isError(mlhSize), corruption_detected);+ ip += mlhSize;+ }+ }++ return ip-istart;+}+++typedef struct {+ size_t litLength;+ size_t matchLength;+ size_t offset;+ const BYTE* match;+} seq_t;++typedef struct {+ size_t state;+ const ZSTD_seqSymbol* table;+} ZSTD_fseState;++typedef struct {+ BIT_DStream_t DStream;+ ZSTD_fseState stateLL;+ ZSTD_fseState stateOffb;+ ZSTD_fseState stateML;+ size_t prevOffset[ZSTD_REP_NUM];+ const BYTE* prefixStart;+ const BYTE* dictEnd;+ size_t pos;+} seqState_t;+++/* ZSTD_execSequenceLast7():+ * exceptional case : decompress a match starting within last 7 bytes of output buffer.+ * requires more careful checks, to ensure there is no overflow.+ * performance does not matter though.+ * note : this case is supposed to be never generated "naturally" by reference encoder,+ * since in most cases it needs at least 8 bytes to look for a match.+ * but it's allowed by the specification. */+FORCE_NOINLINE+size_t ZSTD_execSequenceLast7(BYTE* op,+ BYTE* const oend, seq_t sequence,+ const BYTE** litPtr, const BYTE* const litLimit,+ const BYTE* const base, const BYTE* const vBase, const BYTE* const dictEnd)+{+ BYTE* const oLitEnd = op + sequence.litLength;+ size_t const sequenceLength = sequence.litLength + sequence.matchLength;+ BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */+ const BYTE* const iLitEnd = *litPtr + sequence.litLength;+ const BYTE* match = oLitEnd - sequence.offset;++ /* check */+ RETURN_ERROR_IF(oMatchEnd>oend, dstSize_tooSmall, "last match must fit within dstBuffer");+ RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "try to read beyond literal buffer");++ /* copy literals */+ while (op < oLitEnd) *op++ = *(*litPtr)++;++ /* copy Match */+ if (sequence.offset > (size_t)(oLitEnd - base)) {+ /* offset beyond prefix */+ RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - vBase),corruption_detected);+ match = dictEnd - (base-match);+ if (match + sequence.matchLength <= dictEnd) {+ memmove(oLitEnd, match, sequence.matchLength);+ return sequenceLength;+ }+ /* span extDict & currentPrefixSegment */+ { size_t const length1 = dictEnd - match;+ memmove(oLitEnd, match, length1);+ op = oLitEnd + length1;+ sequence.matchLength -= length1;+ match = base;+ } }+ while (op < oMatchEnd) *op++ = *match++;+ return sequenceLength;+}+++HINT_INLINE+size_t ZSTD_execSequence(BYTE* op,+ BYTE* const oend, seq_t sequence,+ const BYTE** litPtr, const BYTE* const litLimit,+ const BYTE* const prefixStart, const BYTE* const virtualStart, const BYTE* const dictEnd)+{+ BYTE* const oLitEnd = op + sequence.litLength;+ size_t const sequenceLength = sequence.litLength + sequence.matchLength;+ BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */+ BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;+ const BYTE* const iLitEnd = *litPtr + sequence.litLength;+ const BYTE* match = oLitEnd - sequence.offset;++ /* check */+ RETURN_ERROR_IF(oMatchEnd>oend, dstSize_tooSmall, "last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend");+ RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "over-read beyond lit buffer");+ if (oLitEnd>oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, prefixStart, virtualStart, dictEnd);++ /* copy Literals */+ ZSTD_copy8(op, *litPtr);+ if (sequence.litLength > 8)+ ZSTD_wildcopy(op+8, (*litPtr)+8, sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */+ op = oLitEnd;+ *litPtr = iLitEnd; /* update for next sequence */++ /* copy Match */+ if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {+ /* offset beyond prefix -> go into extDict */+ RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - virtualStart), corruption_detected);+ match = dictEnd + (match - prefixStart);+ if (match + sequence.matchLength <= dictEnd) {+ memmove(oLitEnd, match, sequence.matchLength);+ return sequenceLength;+ }+ /* span extDict & currentPrefixSegment */+ { size_t const length1 = dictEnd - match;+ memmove(oLitEnd, match, length1);+ op = oLitEnd + length1;+ sequence.matchLength -= length1;+ match = prefixStart;+ if (op > oend_w || sequence.matchLength < MINMATCH) {+ U32 i;+ for (i = 0; i < sequence.matchLength; ++i) op[i] = match[i];+ return sequenceLength;+ }+ } }+ /* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */++ /* match within prefix */+ if (sequence.offset < 8) {+ /* close range match, overlap */+ static const U32 dec32table[] = { 0, 1, 2, 1, 4, 4, 4, 4 }; /* added */+ static const int dec64table[] = { 8, 8, 8, 7, 8, 9,10,11 }; /* subtracted */+ int const sub2 = dec64table[sequence.offset];+ op[0] = match[0];+ op[1] = match[1];+ op[2] = match[2];+ op[3] = match[3];+ match += dec32table[sequence.offset];+ ZSTD_copy4(op+4, match);+ match -= sub2;+ } else {+ ZSTD_copy8(op, match);+ }+ op += 8; match += 8;++ if (oMatchEnd > oend-(16-MINMATCH)) {+ if (op < oend_w) {+ ZSTD_wildcopy(op, match, oend_w - op);+ match += oend_w - op;+ op = oend_w;+ }+ while (op < oMatchEnd) *op++ = *match++;+ } else {+ ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */+ }+ return sequenceLength;+}+++HINT_INLINE+size_t ZSTD_execSequenceLong(BYTE* op,+ BYTE* const oend, seq_t sequence,+ const BYTE** litPtr, const BYTE* const litLimit,+ const BYTE* const prefixStart, const BYTE* const dictStart, const BYTE* const dictEnd)+{+ BYTE* const oLitEnd = op + sequence.litLength;+ size_t const sequenceLength = sequence.litLength + sequence.matchLength;+ BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */+ BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;+ const BYTE* const iLitEnd = *litPtr + sequence.litLength;+ const BYTE* match = sequence.match;++ /* check */+ RETURN_ERROR_IF(oMatchEnd > oend, dstSize_tooSmall, "last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend");+ RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "over-read beyond lit buffer");+ if (oLitEnd > oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, prefixStart, dictStart, dictEnd);++ /* copy Literals */+ ZSTD_copy8(op, *litPtr); /* note : op <= oLitEnd <= oend_w == oend - 8 */+ if (sequence.litLength > 8)+ ZSTD_wildcopy(op+8, (*litPtr)+8, sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */+ op = oLitEnd;+ *litPtr = iLitEnd; /* update for next sequence */++ /* copy Match */+ if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {+ /* offset beyond prefix */+ RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - dictStart), corruption_detected);+ if (match + sequence.matchLength <= dictEnd) {+ memmove(oLitEnd, match, sequence.matchLength);+ return sequenceLength;+ }+ /* span extDict & currentPrefixSegment */+ { size_t const length1 = dictEnd - match;+ memmove(oLitEnd, match, length1);+ op = oLitEnd + length1;+ sequence.matchLength -= length1;+ match = prefixStart;+ if (op > oend_w || sequence.matchLength < MINMATCH) {+ U32 i;+ for (i = 0; i < sequence.matchLength; ++i) op[i] = match[i];+ return sequenceLength;+ }+ } }+ assert(op <= oend_w);+ assert(sequence.matchLength >= MINMATCH);++ /* match within prefix */+ if (sequence.offset < 8) {+ /* close range match, overlap */+ static const U32 dec32table[] = { 0, 1, 2, 1, 4, 4, 4, 4 }; /* added */+ static const int dec64table[] = { 8, 8, 8, 7, 8, 9,10,11 }; /* subtracted */+ int const sub2 = dec64table[sequence.offset];+ op[0] = match[0];+ op[1] = match[1];+ op[2] = match[2];+ op[3] = match[3];+ match += dec32table[sequence.offset];+ ZSTD_copy4(op+4, match);+ match -= sub2;+ } else {+ ZSTD_copy8(op, match);+ }+ op += 8; match += 8;++ if (oMatchEnd > oend-(16-MINMATCH)) {+ if (op < oend_w) {+ ZSTD_wildcopy(op, match, oend_w - op);+ match += oend_w - op;+ op = oend_w;+ }+ while (op < oMatchEnd) *op++ = *match++;+ } else {+ ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */+ }+ return sequenceLength;+}++static void+ZSTD_initFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, const ZSTD_seqSymbol* dt)+{+ const void* ptr = dt;+ const ZSTD_seqSymbol_header* const DTableH = (const ZSTD_seqSymbol_header*)ptr;+ DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);+ DEBUGLOG(6, "ZSTD_initFseState : val=%u using %u bits",+ (U32)DStatePtr->state, DTableH->tableLog);+ BIT_reloadDStream(bitD);+ DStatePtr->table = dt + 1;+}++FORCE_INLINE_TEMPLATE void+ZSTD_updateFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD)+{+ ZSTD_seqSymbol const DInfo = DStatePtr->table[DStatePtr->state];+ U32 const nbBits = DInfo.nbBits;+ size_t const lowBits = BIT_readBits(bitD, nbBits);+ DStatePtr->state = DInfo.nextState + lowBits;+}++/* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum+ * offset bits. But we can only read at most (STREAM_ACCUMULATOR_MIN_32 - 1)+ * bits before reloading. This value is the maximum number of bytes we read+ * after reloading when we are decoding long offsets.+ */+#define LONG_OFFSETS_MAX_EXTRA_BITS_32 \+ (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32 \+ ? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32 \+ : 0)++typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG+FORCE_INLINE_TEMPLATE seq_t+ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)+{+ seq_t seq;+ U32 const llBits = seqState->stateLL.table[seqState->stateLL.state].nbAdditionalBits;+ U32 const mlBits = seqState->stateML.table[seqState->stateML.state].nbAdditionalBits;+ U32 const ofBits = seqState->stateOffb.table[seqState->stateOffb.state].nbAdditionalBits;+ U32 const totalBits = llBits+mlBits+ofBits;+ U32 const llBase = seqState->stateLL.table[seqState->stateLL.state].baseValue;+ U32 const mlBase = seqState->stateML.table[seqState->stateML.state].baseValue;+ U32 const ofBase = seqState->stateOffb.table[seqState->stateOffb.state].baseValue;++ /* sequence */+ { size_t offset;+ if (!ofBits)+ offset = 0;+ else {+ ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);+ ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);+ assert(ofBits <= MaxOff);+ if (MEM_32bits() && longOffsets && (ofBits >= STREAM_ACCUMULATOR_MIN_32)) {+ U32 const extraBits = ofBits - MIN(ofBits, 32 - seqState->DStream.bitsConsumed);+ offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);+ BIT_reloadDStream(&seqState->DStream);+ if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);+ assert(extraBits <= LONG_OFFSETS_MAX_EXTRA_BITS_32); /* to avoid another reload */+ } else {+ offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits/*>0*/); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);+ }+ }++ if (ofBits <= 1) {+ offset += (llBase==0);+ if (offset) {+ size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];+ temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */+ if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];+ seqState->prevOffset[1] = seqState->prevOffset[0];+ seqState->prevOffset[0] = offset = temp;+ } else { /* offset == 0 */+ offset = seqState->prevOffset[0];+ }+ } else {+ seqState->prevOffset[2] = seqState->prevOffset[1];+ seqState->prevOffset[1] = seqState->prevOffset[0];+ seqState->prevOffset[0] = offset;+ }+ seq.offset = offset;+ }++ seq.matchLength = mlBase+ + ((mlBits>0) ? BIT_readBitsFast(&seqState->DStream, mlBits/*>0*/) : 0); /* <= 16 bits */+ if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))+ BIT_reloadDStream(&seqState->DStream);+ if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))+ BIT_reloadDStream(&seqState->DStream);+ /* Ensure there are enough bits to read the rest of data in 64-bit mode. */+ ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);++ seq.litLength = llBase+ + ((llBits>0) ? BIT_readBitsFast(&seqState->DStream, llBits/*>0*/) : 0); /* <= 16 bits */+ if (MEM_32bits())+ BIT_reloadDStream(&seqState->DStream);++ DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",+ (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);++ /* ANS state update */+ ZSTD_updateFseState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */+ ZSTD_updateFseState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */+ ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */++ return seq;+}++FORCE_INLINE_TEMPLATE size_t+ZSTD_decompressSequences_body( ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ const BYTE* ip = (const BYTE*)seqStart;+ const BYTE* const iend = ip + seqSize;+ BYTE* const ostart = (BYTE* const)dst;+ BYTE* const oend = ostart + maxDstSize;+ BYTE* op = ostart;+ const BYTE* litPtr = dctx->litPtr;+ const BYTE* const litEnd = litPtr + dctx->litSize;+ const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);+ const BYTE* const vBase = (const BYTE*) (dctx->virtualStart);+ const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);+ DEBUGLOG(5, "ZSTD_decompressSequences_body");++ /* Regen sequences */+ if (nbSeq) {+ seqState_t seqState;+ dctx->fseEntropy = 1;+ { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }+ RETURN_ERROR_IF(+ ERR_isError(BIT_initDStream(&seqState.DStream, ip, iend-ip)),+ corruption_detected);+ ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);+ ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);+ ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);++ for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {+ nbSeq--;+ { seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);+ size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, prefixStart, vBase, dictEnd);+ DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);+ if (ZSTD_isError(oneSeqSize)) return oneSeqSize;+ op += oneSeqSize;+ } }++ /* check if reached exact end */+ DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq);+ RETURN_ERROR_IF(nbSeq, corruption_detected);+ /* save reps for next block */+ { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }+ }++ /* last literal segment */+ { size_t const lastLLSize = litEnd - litPtr;+ RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall);+ memcpy(op, litPtr, lastLLSize);+ op += lastLLSize;+ }++ return op-ostart;+}++static size_t+ZSTD_decompressSequences_default(ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+}+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */++++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT+FORCE_INLINE_TEMPLATE seq_t+ZSTD_decodeSequenceLong(seqState_t* seqState, ZSTD_longOffset_e const longOffsets)+{+ seq_t seq;+ U32 const llBits = seqState->stateLL.table[seqState->stateLL.state].nbAdditionalBits;+ U32 const mlBits = seqState->stateML.table[seqState->stateML.state].nbAdditionalBits;+ U32 const ofBits = seqState->stateOffb.table[seqState->stateOffb.state].nbAdditionalBits;+ U32 const totalBits = llBits+mlBits+ofBits;+ U32 const llBase = seqState->stateLL.table[seqState->stateLL.state].baseValue;+ U32 const mlBase = seqState->stateML.table[seqState->stateML.state].baseValue;+ U32 const ofBase = seqState->stateOffb.table[seqState->stateOffb.state].baseValue;++ /* sequence */+ { size_t offset;+ if (!ofBits)+ offset = 0;+ else {+ ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);+ ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);+ assert(ofBits <= MaxOff);+ if (MEM_32bits() && longOffsets) {+ U32 const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN_32-1);+ offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);+ if (MEM_32bits() || extraBits) BIT_reloadDStream(&seqState->DStream);+ if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);+ } else {+ offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);+ }+ }++ if (ofBits <= 1) {+ offset += (llBase==0);+ if (offset) {+ size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];+ temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */+ if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];+ seqState->prevOffset[1] = seqState->prevOffset[0];+ seqState->prevOffset[0] = offset = temp;+ } else {+ offset = seqState->prevOffset[0];+ }+ } else {+ seqState->prevOffset[2] = seqState->prevOffset[1];+ seqState->prevOffset[1] = seqState->prevOffset[0];+ seqState->prevOffset[0] = offset;+ }+ seq.offset = offset;+ }++ seq.matchLength = mlBase + ((mlBits>0) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */+ if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))+ BIT_reloadDStream(&seqState->DStream);+ if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))+ BIT_reloadDStream(&seqState->DStream);+ /* Verify that there is enough bits to read the rest of the data in 64-bit mode. */+ ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);++ seq.litLength = llBase + ((llBits>0) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */+ if (MEM_32bits())+ BIT_reloadDStream(&seqState->DStream);++ { size_t const pos = seqState->pos + seq.litLength;+ const BYTE* const matchBase = (seq.offset > pos) ? seqState->dictEnd : seqState->prefixStart;+ seq.match = matchBase + pos - seq.offset; /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.+ * No consequence though : no memory access will occur, overly large offset will be detected in ZSTD_execSequenceLong() */+ seqState->pos = pos + seq.matchLength;+ }++ /* ANS state update */+ ZSTD_updateFseState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */+ ZSTD_updateFseState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */+ ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */++ return seq;+}++FORCE_INLINE_TEMPLATE size_t+ZSTD_decompressSequencesLong_body(+ ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ const BYTE* ip = (const BYTE*)seqStart;+ const BYTE* const iend = ip + seqSize;+ BYTE* const ostart = (BYTE* const)dst;+ BYTE* const oend = ostart + maxDstSize;+ BYTE* op = ostart;+ const BYTE* litPtr = dctx->litPtr;+ const BYTE* const litEnd = litPtr + dctx->litSize;+ const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);+ const BYTE* const dictStart = (const BYTE*) (dctx->virtualStart);+ const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);++ /* Regen sequences */+ if (nbSeq) {+#define STORED_SEQS 4+#define STORED_SEQS_MASK (STORED_SEQS-1)+#define ADVANCED_SEQS 4+ seq_t sequences[STORED_SEQS];+ int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);+ seqState_t seqState;+ int seqNb;+ dctx->fseEntropy = 1;+ { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }+ seqState.prefixStart = prefixStart;+ seqState.pos = (size_t)(op-prefixStart);+ seqState.dictEnd = dictEnd;+ assert(iend >= ip);+ RETURN_ERROR_IF(+ ERR_isError(BIT_initDStream(&seqState.DStream, ip, iend-ip)),+ corruption_detected);+ ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);+ ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);+ ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);++ /* prepare in advance */+ for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {+ sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, isLongOffset);+ PREFETCH_L1(sequences[seqNb].match); PREFETCH_L1(sequences[seqNb].match + sequences[seqNb].matchLength - 1); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */+ }+ RETURN_ERROR_IF(seqNb<seqAdvance, corruption_detected);++ /* decode and decompress */+ for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) {+ seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, isLongOffset);+ size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[(seqNb-ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);+ if (ZSTD_isError(oneSeqSize)) return oneSeqSize;+ PREFETCH_L1(sequence.match); PREFETCH_L1(sequence.match + sequence.matchLength - 1); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */+ sequences[seqNb & STORED_SEQS_MASK] = sequence;+ op += oneSeqSize;+ }+ RETURN_ERROR_IF(seqNb<nbSeq, corruption_detected);++ /* finish queue */+ seqNb -= seqAdvance;+ for ( ; seqNb<nbSeq ; seqNb++) {+ size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb&STORED_SEQS_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);+ if (ZSTD_isError(oneSeqSize)) return oneSeqSize;+ op += oneSeqSize;+ }++ /* save reps for next block */+ { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }+ }++ /* last literal segment */+ { size_t const lastLLSize = litEnd - litPtr;+ RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall);+ memcpy(op, litPtr, lastLLSize);+ op += lastLLSize;+ }++ return op-ostart;+}++static size_t+ZSTD_decompressSequencesLong_default(ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+}+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */++++#if DYNAMIC_BMI2++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG+static TARGET_ATTRIBUTE("bmi2") size_t+ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+}+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT+static TARGET_ATTRIBUTE("bmi2") size_t+ZSTD_decompressSequencesLong_bmi2(ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+}+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */++#endif /* DYNAMIC_BMI2 */++typedef size_t (*ZSTD_decompressSequences_t)(+ ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset);++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG+static size_t+ZSTD_decompressSequences(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ DEBUGLOG(5, "ZSTD_decompressSequences");+#if DYNAMIC_BMI2+ if (dctx->bmi2) {+ return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+ }+#endif+ return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+}+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */+++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT+/* ZSTD_decompressSequencesLong() :+ * decompression function triggered when a minimum share of offsets is considered "long",+ * aka out of cache.+ * note : "long" definition seems overloaded here, sometimes meaning "wider than bitstream register", and sometimes meaning "farther than memory cache distance".+ * This function will try to mitigate main memory latency through the use of prefetching */+static size_t+ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,+ void* dst, size_t maxDstSize,+ const void* seqStart, size_t seqSize, int nbSeq,+ const ZSTD_longOffset_e isLongOffset)+{+ DEBUGLOG(5, "ZSTD_decompressSequencesLong");+#if DYNAMIC_BMI2+ if (dctx->bmi2) {+ return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+ }+#endif+ return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);+}+#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */++++#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \+ !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)+/* ZSTD_getLongOffsetsShare() :+ * condition : offTable must be valid+ * @return : "share" of long offsets (arbitrarily defined as > (1<<23))+ * compared to maximum possible of (1<<OffFSELog) */+static unsigned+ZSTD_getLongOffsetsShare(const ZSTD_seqSymbol* offTable)+{+ const void* ptr = offTable;+ U32 const tableLog = ((const ZSTD_seqSymbol_header*)ptr)[0].tableLog;+ const ZSTD_seqSymbol* table = offTable + 1;+ U32 const max = 1 << tableLog;+ U32 u, total = 0;+ DEBUGLOG(5, "ZSTD_getLongOffsetsShare: (tableLog=%u)", tableLog);++ assert(max <= (1 << OffFSELog)); /* max not too large */+ for (u=0; u<max; u++) {+ if (table[u].nbAdditionalBits > 22) total += 1;+ }++ assert(tableLog <= OffFSELog);+ total <<= (OffFSELog - tableLog); /* scale to OffFSELog */++ return total;+}+#endif+++size_t+ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize, const int frame)+{ /* blockType == blockCompressed */+ const BYTE* ip = (const BYTE*)src;+ /* isLongOffset must be true if there are long offsets.+ * Offsets are long if they are larger than 2^STREAM_ACCUMULATOR_MIN.+ * We don't expect that to be the case in 64-bit mode.+ * In block mode, window size is not known, so we have to be conservative.+ * (note: but it could be evaluated from current-lowLimit)+ */+ ZSTD_longOffset_e const isLongOffset = (ZSTD_longOffset_e)(MEM_32bits() && (!frame || (dctx->fParams.windowSize > (1ULL << STREAM_ACCUMULATOR_MIN))));+ DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);++ RETURN_ERROR_IF(srcSize >= ZSTD_BLOCKSIZE_MAX, srcSize_wrong);++ /* Decode literals section */+ { size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);+ DEBUGLOG(5, "ZSTD_decodeLiteralsBlock : %u", (U32)litCSize);+ if (ZSTD_isError(litCSize)) return litCSize;+ ip += litCSize;+ srcSize -= litCSize;+ }++ /* Build Decoding Tables */+ {+ /* These macros control at build-time which decompressor implementation+ * we use. If neither is defined, we do some inspection and dispatch at+ * runtime.+ */+#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \+ !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)+ int usePrefetchDecoder = dctx->ddictIsCold;+#endif+ int nbSeq;+ size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, srcSize);+ if (ZSTD_isError(seqHSize)) return seqHSize;+ ip += seqHSize;+ srcSize -= seqHSize;++#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \+ !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)+ if ( !usePrefetchDecoder+ && (!frame || (dctx->fParams.windowSize > (1<<24)))+ && (nbSeq>ADVANCED_SEQS) ) { /* could probably use a larger nbSeq limit */+ U32 const shareLongOffsets = ZSTD_getLongOffsetsShare(dctx->OFTptr);+ U32 const minShare = MEM_64bits() ? 7 : 20; /* heuristic values, correspond to 2.73% and 7.81% */+ usePrefetchDecoder = (shareLongOffsets >= minShare);+ }+#endif++ dctx->ddictIsCold = 0;++#if !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT) && \+ !defined(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG)+ if (usePrefetchDecoder)+#endif+#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT+ return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);+#endif++#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG+ /* else */+ return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);+#endif+ }+}+++size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize)+{+ size_t dSize;+ ZSTD_checkContinuity(dctx, dst);+ dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 0);+ dctx->previousDstEnd = (char*)dst + dSize;+ return dSize;+}
+ zstd/lib/decompress/zstd_decompress_block.h view
@@ -0,0 +1,59 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+++#ifndef ZSTD_DEC_BLOCK_H+#define ZSTD_DEC_BLOCK_H++/*-*******************************************************+ * Dependencies+ *********************************************************/+#include <stddef.h> /* size_t */+#include "zstd.h" /* DCtx, and some public functions */+#include "zstd_internal.h" /* blockProperties_t, and some public functions */+#include "zstd_decompress_internal.h" /* ZSTD_seqSymbol */+++/* === Prototypes === */++/* note: prototypes already published within `zstd.h` :+ * ZSTD_decompressBlock()+ */++/* note: prototypes already published within `zstd_internal.h` :+ * ZSTD_getcBlockSize()+ * ZSTD_decodeSeqHeaders()+ */+++/* ZSTD_decompressBlock_internal() :+ * decompress block, starting at `src`,+ * into destination buffer `dst`.+ * @return : decompressed block size,+ * or an error code (which can be tested using ZSTD_isError())+ */+size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize, const int frame);++/* ZSTD_buildFSETable() :+ * generate FSE decoding table for one symbol (ll, ml or off)+ * this function must be called with valid parameters only+ * (dt is large enough, normalizedCounter distribution total is a power of 2, max is within range, etc.)+ * in which case it cannot fail.+ * Internal use only.+ */+void ZSTD_buildFSETable(ZSTD_seqSymbol* dt,+ const short* normalizedCounter, unsigned maxSymbolValue,+ const U32* baseValue, const U32* nbAdditionalBits,+ unsigned tableLog);+++#endif /* ZSTD_DEC_BLOCK_H */
+ zstd/lib/decompress/zstd_decompress_internal.h view
@@ -0,0 +1,175 @@+/*+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.+ * All rights reserved.+ *+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+++/* zstd_decompress_internal:+ * objects and definitions shared within lib/decompress modules */++ #ifndef ZSTD_DECOMPRESS_INTERNAL_H+ #define ZSTD_DECOMPRESS_INTERNAL_H+++/*-*******************************************************+ * Dependencies+ *********************************************************/+#include "mem.h" /* BYTE, U16, U32 */+#include "zstd_internal.h" /* ZSTD_seqSymbol */++++/*-*******************************************************+ * Constants+ *********************************************************/+static const U32 LL_base[MaxLL+1] = {+ 0, 1, 2, 3, 4, 5, 6, 7,+ 8, 9, 10, 11, 12, 13, 14, 15,+ 16, 18, 20, 22, 24, 28, 32, 40,+ 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000,+ 0x2000, 0x4000, 0x8000, 0x10000 };++static const U32 OF_base[MaxOff+1] = {+ 0, 1, 1, 5, 0xD, 0x1D, 0x3D, 0x7D,+ 0xFD, 0x1FD, 0x3FD, 0x7FD, 0xFFD, 0x1FFD, 0x3FFD, 0x7FFD,+ 0xFFFD, 0x1FFFD, 0x3FFFD, 0x7FFFD, 0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD,+ 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD, 0x1FFFFFFD, 0x3FFFFFFD, 0x7FFFFFFD };++static const U32 OF_bits[MaxOff+1] = {+ 0, 1, 2, 3, 4, 5, 6, 7,+ 8, 9, 10, 11, 12, 13, 14, 15,+ 16, 17, 18, 19, 20, 21, 22, 23,+ 24, 25, 26, 27, 28, 29, 30, 31 };++static const U32 ML_base[MaxML+1] = {+ 3, 4, 5, 6, 7, 8, 9, 10,+ 11, 12, 13, 14, 15, 16, 17, 18,+ 19, 20, 21, 22, 23, 24, 25, 26,+ 27, 28, 29, 30, 31, 32, 33, 34,+ 35, 37, 39, 41, 43, 47, 51, 59,+ 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803,+ 0x1003, 0x2003, 0x4003, 0x8003, 0x10003 };+++/*-*******************************************************+ * Decompression types+ *********************************************************/+ typedef struct {+ U32 fastMode;+ U32 tableLog;+ } ZSTD_seqSymbol_header;++ typedef struct {+ U16 nextState;+ BYTE nbAdditionalBits;+ BYTE nbBits;+ U32 baseValue;+ } ZSTD_seqSymbol;++ #define SEQSYMBOL_TABLE_SIZE(log) (1 + (1 << (log)))++typedef struct {+ ZSTD_seqSymbol LLTable[SEQSYMBOL_TABLE_SIZE(LLFSELog)]; /* Note : Space reserved for FSE Tables */+ ZSTD_seqSymbol OFTable[SEQSYMBOL_TABLE_SIZE(OffFSELog)]; /* is also used as temporary workspace while building hufTable during DDict creation */+ ZSTD_seqSymbol MLTable[SEQSYMBOL_TABLE_SIZE(MLFSELog)]; /* and therefore must be at least HUF_DECOMPRESS_WORKSPACE_SIZE large */+ HUF_DTable hufTable[HUF_DTABLE_SIZE(HufLog)]; /* can accommodate HUF_decompress4X */+ U32 rep[ZSTD_REP_NUM];+} ZSTD_entropyDTables_t;++typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader,+ ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock,+ ZSTDds_decompressLastBlock, ZSTDds_checkChecksum,+ ZSTDds_decodeSkippableHeader, ZSTDds_skipFrame } ZSTD_dStage;++typedef enum { zdss_init=0, zdss_loadHeader,+ zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;++typedef enum {+ ZSTD_use_indefinitely = -1, /* Use the dictionary indefinitely */+ ZSTD_dont_use = 0, /* Do not use the dictionary (if one exists free it) */+ ZSTD_use_once = 1 /* Use the dictionary once and set to ZSTD_dont_use */+} ZSTD_dictUses_e;++struct ZSTD_DCtx_s+{+ const ZSTD_seqSymbol* LLTptr;+ const ZSTD_seqSymbol* MLTptr;+ const ZSTD_seqSymbol* OFTptr;+ const HUF_DTable* HUFptr;+ ZSTD_entropyDTables_t entropy;+ U32 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32]; /* space needed when building huffman tables */+ const void* previousDstEnd; /* detect continuity */+ const void* prefixStart; /* start of current segment */+ const void* virtualStart; /* virtual start of previous segment if it was just before current one */+ const void* dictEnd; /* end of previous segment */+ size_t expected;+ ZSTD_frameHeader fParams;+ U64 decodedSize;+ blockType_e bType; /* used in ZSTD_decompressContinue(), store blockType between block header decoding and block decompression stages */+ ZSTD_dStage stage;+ U32 litEntropy;+ U32 fseEntropy;+ XXH64_state_t xxhState;+ size_t headerSize;+ ZSTD_format_e format;+ const BYTE* litPtr;+ ZSTD_customMem customMem;+ size_t litSize;+ size_t rleSize;+ size_t staticSize;+ int bmi2; /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */++ /* dictionary */+ ZSTD_DDict* ddictLocal;+ const ZSTD_DDict* ddict; /* set by ZSTD_initDStream_usingDDict(), or ZSTD_DCtx_refDDict() */+ U32 dictID;+ int ddictIsCold; /* if == 1 : dictionary is "new" for working context, and presumed "cold" (not in cpu cache) */+ ZSTD_dictUses_e dictUses;++ /* streaming */+ ZSTD_dStreamStage streamStage;+ char* inBuff;+ size_t inBuffSize;+ size_t inPos;+ size_t maxWindowSize;+ char* outBuff;+ size_t outBuffSize;+ size_t outStart;+ size_t outEnd;+ size_t lhSize;+ void* legacyContext;+ U32 previousLegacyVersion;+ U32 legacyVersion;+ U32 hostageByte;+ int noForwardProgress;++ /* workspace */+ BYTE litBuffer[ZSTD_BLOCKSIZE_MAX + WILDCOPY_OVERLENGTH];+ BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];+}; /* typedef'd to ZSTD_DCtx within "zstd.h" */+++/*-*******************************************************+ * Shared internal functions+ *********************************************************/++/*! ZSTD_loadDEntropy() :+ * dict : must point at beginning of a valid zstd dictionary.+ * @return : size of entropy tables read */+size_t ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy,+ const void* const dict, size_t const dictSize);++/*! ZSTD_checkContinuity() :+ * check if next `dst` follows previous position, where decompression ended.+ * If yes, do nothing (continue on current segment).+ * If not, classify previous segment as "external dictionary", and start a new segment.+ * This function cannot fail. */+void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst);+++#endif /* ZSTD_DECOMPRESS_INTERNAL_H */
zstd/lib/dictBuilder/cover.c view
@@ -1,12 +1,23 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ +/* *****************************************************************************+ * Constructs a dictionary using a heuristic based on the following paper:+ *+ * Liao, Petri, Moffat, Wirth+ * Effective Construction of Relative Lempel-Ziv Dictionaries+ * Published in WWW 2016.+ *+ * Adapted from code originally written by @ot (Giuseppe Ottaviano).+ ******************************************************************************/+ /*-************************************* * Dependencies ***************************************/@@ -14,11 +25,11 @@ #include <stdlib.h> /* malloc, free, qsort */ #include <string.h> /* memset */ #include <time.h> /* clock */-#ifdef ZSTD_PTHREAD-#include "threading.h"-#endif -#include "mem.h" /* read */+#include "mem.h" /* read */+#include "pool.h"+#include "threading.h"+#include "cover.h" #include "zstd_internal.h" /* includes zstd.h */ #ifndef ZDICT_STATIC_LINKING_ONLY #define ZDICT_STATIC_LINKING_ONLY@@ -28,7 +39,8 @@ /*-************************************* * Constants ***************************************/-#define COVER_MAX_SAMPLES_SIZE ((U32)-1)+#define COVER_MAX_SAMPLES_SIZE (sizeof(size_t) == 8 ? ((unsigned)-1) : ((unsigned)1 GB))+#define DEFAULT_SPLITPOINT 1.0 /*-************************************* * Console display@@ -50,8 +62,6 @@ if ((clock() - g_time > refreshRate) || (displayLevel >= 4)) { \ g_time = clock(); \ DISPLAY(__VA_ARGS__); \- if (displayLevel >= 4) \- fflush(stdout); \ } \ } #define DISPLAYUPDATE(l, ...) LOCALDISPLAYUPDATE(g_displayLevel, l, __VA_ARGS__)@@ -176,7 +186,7 @@ } /**- * Destroyes a map that is inited with COVER_map_init().+ * Destroys a map that is inited with COVER_map_init(). */ static void COVER_map_destroy(COVER_map_t *map) { if (map->data) {@@ -195,6 +205,8 @@ size_t *offsets; const size_t *samplesSizes; size_t nbSamples;+ size_t nbTrainSamples;+ size_t nbTestSamples; U32 *suffix; size_t suffixSize; U32 *freqs;@@ -212,9 +224,9 @@ /** * Returns the sum of the sample sizes. */-static size_t COVER_sum(const size_t *samplesSizes, unsigned nbSamples) {+size_t COVER_sum(const size_t *samplesSizes, unsigned nbSamples) { size_t sum = 0;- size_t i;+ unsigned i; for (i = 0; i < nbSamples; ++i) { sum += samplesSizes[i]; }@@ -227,10 +239,22 @@ * Returns 1 if the dmer at lp is greater than the dmer at rp. */ static int COVER_cmp(COVER_ctx_t *ctx, const void *lp, const void *rp) {- const U32 lhs = *(const U32 *)lp;- const U32 rhs = *(const U32 *)rp;+ U32 const lhs = *(U32 const *)lp;+ U32 const rhs = *(U32 const *)rp; return memcmp(ctx->samples + lhs, ctx->samples + rhs, ctx->d); }+/**+ * Faster version for d <= 8.+ */+static int COVER_cmp8(COVER_ctx_t *ctx, const void *lp, const void *rp) {+ U64 const mask = (ctx->d == 8) ? (U64)-1 : (((U64)1 << (8 * ctx->d)) - 1);+ U64 const lhs = MEM_readLE64(ctx->samples + *(U32 const *)lp) & mask;+ U64 const rhs = MEM_readLE64(ctx->samples + *(U32 const *)rp) & mask;+ if (lhs < rhs) {+ return -1;+ }+ return (lhs > rhs);+} /** * Same as COVER_cmp() except ties are broken by pointer value@@ -244,6 +268,16 @@ } return result; }+/**+ * Faster version for d <= 8.+ */+static int COVER_strict_cmp8(const void *lp, const void *rp) {+ int result = COVER_cmp8(g_ctx, lp, rp);+ if (result == 0) {+ result = lp < rp ? -1 : 1;+ }+ return result;+} /** * Returns the first pointer in [first, last) whose element does not compare@@ -347,14 +381,6 @@ ctx->suffix[dmerId] = freq; } -/**- * A segment is a range in the source as well as the score of the segment.- */-typedef struct {- U32 begin;- U32 end;- double score;-} COVER_segment_t; /** * Selects the best segment in an epoch.@@ -365,11 +391,12 @@ * * Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1}) *- * Once the dmer d is in the dictionay we set F(d) = 0.+ * Once the dmer d is in the dictionary we set F(d) = 0. */ static COVER_segment_t COVER_selectSegment(const COVER_ctx_t *ctx, U32 *freqs, COVER_map_t *activeDmers, U32 begin,- U32 end, COVER_params_t parameters) {+ U32 end,+ ZDICT_cover_params_t parameters) { /* Constants */ const U32 k = parameters.k; const U32 d = parameters.d;@@ -408,7 +435,7 @@ U32 *delDmerOcc = COVER_map_at(activeDmers, delDmer); activeSegment.begin += 1; *delDmerOcc -= 1;- /* If this is the last occurence of the dmer, subtract its score */+ /* If this is the last occurrence of the dmer, subtract its score */ if (*delDmerOcc == 0) { COVER_map_remove(activeDmers, delDmer); activeSegment.score -= freqs[delDmer];@@ -449,15 +476,24 @@ * Check the validity of the parameters. * Returns non-zero if the parameters are valid and 0 otherwise. */-static int COVER_checkParameters(COVER_params_t parameters) {+static int COVER_checkParameters(ZDICT_cover_params_t parameters,+ size_t maxDictSize) { /* k and d are required parameters */ if (parameters.d == 0 || parameters.k == 0) { return 0; }+ /* k <= maxDictSize */+ if (parameters.k > maxDictSize) {+ return 0;+ } /* d <= k */ if (parameters.d > parameters.k) { return 0; }+ /* 0 < splitPoint <= 1 */+ if (parameters.splitPoint <= 0 || parameters.splitPoint > 1){+ return 0;+ } return 1; } @@ -495,36 +531,58 @@ */ static int COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,- unsigned d) {+ unsigned d, double splitPoint) { const BYTE *const samples = (const BYTE *)samplesBuffer; const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);+ /* Split samples into testing and training sets */+ const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples;+ const unsigned nbTestSamples = splitPoint < 1.0 ? nbSamples - nbTrainSamples : nbSamples;+ const size_t trainingSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes, nbTrainSamples) : totalSamplesSize;+ const size_t testSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize; /* Checks */- if (totalSamplesSize < d ||- totalSamplesSize > (size_t)COVER_MAX_SAMPLES_SIZE) {+ if (totalSamplesSize < MAX(d, sizeof(U64)) ||+ totalSamplesSize >= (size_t)COVER_MAX_SAMPLES_SIZE) {+ DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",+ (unsigned)(totalSamplesSize>>20), (COVER_MAX_SAMPLES_SIZE >> 20)); return 0; }+ /* Check if there are at least 5 training samples */+ if (nbTrainSamples < 5) {+ DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid.", nbTrainSamples);+ return 0;+ }+ /* Check if there's testing sample */+ if (nbTestSamples < 1) {+ DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.", nbTestSamples);+ return 0;+ } /* Zero the context */ memset(ctx, 0, sizeof(*ctx));- DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbSamples,- (U32)totalSamplesSize);+ DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples,+ (unsigned)trainingSamplesSize);+ DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples,+ (unsigned)testSamplesSize); ctx->samples = samples; ctx->samplesSizes = samplesSizes; ctx->nbSamples = nbSamples;+ ctx->nbTrainSamples = nbTrainSamples;+ ctx->nbTestSamples = nbTestSamples; /* Partial suffix array */- ctx->suffixSize = totalSamplesSize - d + 1;+ ctx->suffixSize = trainingSamplesSize - MAX(d, sizeof(U64)) + 1; ctx->suffix = (U32 *)malloc(ctx->suffixSize * sizeof(U32)); /* Maps index to the dmerID */ ctx->dmerAt = (U32 *)malloc(ctx->suffixSize * sizeof(U32)); /* The offsets of each file */ ctx->offsets = (size_t *)malloc((nbSamples + 1) * sizeof(size_t)); if (!ctx->suffix || !ctx->dmerAt || !ctx->offsets) {+ DISPLAYLEVEL(1, "Failed to allocate scratch buffers\n"); COVER_ctx_destroy(ctx); return 0; } ctx->freqs = NULL; ctx->d = d; - /* Fill offsets from the samlesSizes */+ /* Fill offsets from the samplesSizes */ { U32 i; ctx->offsets[0] = 0;@@ -542,9 +600,17 @@ for (i = 0; i < ctx->suffixSize; ++i) { ctx->suffix[i] = i; }- /* qsort doesn't take an opaque pointer, so pass as a global */+ /* qsort doesn't take an opaque pointer, so pass as a global.+ * On OpenBSD qsort() is not guaranteed to be stable, their mergesort() is.+ */ g_ctx = ctx;- qsort(ctx->suffix, ctx->suffixSize, sizeof(U32), &COVER_strict_cmp);+#if defined(__OpenBSD__)+ mergesort(ctx->suffix, ctx->suffixSize, sizeof(U32),+ (ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));+#else+ qsort(ctx->suffix, ctx->suffixSize, sizeof(U32),+ (ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));+#endif } DISPLAYLEVEL(2, "Computing frequencies\n"); /* For each dmer group (group of positions with the same first d bytes):@@ -554,43 +620,86 @@ * 2. We calculate how many samples the dmer occurs in and save it in * freqs[dmerId]. */- COVER_groupBy(ctx->suffix, ctx->suffixSize, sizeof(U32), ctx, &COVER_cmp,- &COVER_group);+ COVER_groupBy(ctx->suffix, ctx->suffixSize, sizeof(U32), ctx,+ (ctx->d <= 8 ? &COVER_cmp8 : &COVER_cmp), &COVER_group); ctx->freqs = ctx->suffix; ctx->suffix = NULL; return 1; } +void COVER_warnOnSmallCorpus(size_t maxDictSize, size_t nbDmers, int displayLevel)+{+ const double ratio = (double)nbDmers / maxDictSize;+ if (ratio >= 10) {+ return;+ }+ LOCALDISPLAYLEVEL(displayLevel, 1,+ "WARNING: The maximum dictionary size %u is too large "+ "compared to the source size %u! "+ "size(source)/size(dictionary) = %f, but it should be >= "+ "10! This may lead to a subpar dictionary! We recommend "+ "training on sources at least 10x, and up to 100x the "+ "size of the dictionary!\n", (U32)maxDictSize,+ (U32)nbDmers, ratio);+}++COVER_epoch_info_t COVER_computeEpochs(U32 maxDictSize,+ U32 nbDmers, U32 k, U32 passes)+{+ const U32 minEpochSize = k * 10;+ COVER_epoch_info_t epochs;+ epochs.num = MAX(1, maxDictSize / k / passes);+ epochs.size = nbDmers / epochs.num;+ if (epochs.size >= minEpochSize) {+ assert(epochs.size * epochs.num <= nbDmers);+ return epochs;+ }+ epochs.size = MIN(minEpochSize, nbDmers);+ epochs.num = nbDmers / epochs.size;+ assert(epochs.size * epochs.num <= nbDmers);+ return epochs;+}+ /** * Given the prepared context build the dictionary. */ static size_t COVER_buildDictionary(const COVER_ctx_t *ctx, U32 *freqs, COVER_map_t *activeDmers, void *dictBuffer, size_t dictBufferCapacity,- COVER_params_t parameters) {+ ZDICT_cover_params_t parameters) { BYTE *const dict = (BYTE *)dictBuffer; size_t tail = dictBufferCapacity;- /* Divide the data up into epochs of equal size.- * We will select at least one segment from each epoch.- */- const U32 epochs = (U32)(dictBufferCapacity / parameters.k);- const U32 epochSize = (U32)(ctx->suffixSize / epochs);+ /* Divide the data into epochs. We will select one segment from each epoch. */+ const COVER_epoch_info_t epochs = COVER_computeEpochs(+ (U32)dictBufferCapacity, (U32)ctx->suffixSize, parameters.k, 4);+ const size_t maxZeroScoreRun = MAX(10, MIN(100, epochs.num >> 3));+ size_t zeroScoreRun = 0; size_t epoch;- DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n", epochs,- epochSize);+ DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n",+ (U32)epochs.num, (U32)epochs.size); /* Loop through the epochs until there are no more segments or the dictionary * is full. */- for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) {- const U32 epochBegin = (U32)(epoch * epochSize);- const U32 epochEnd = epochBegin + epochSize;+ for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs.num) {+ const U32 epochBegin = (U32)(epoch * epochs.size);+ const U32 epochEnd = epochBegin + epochs.size; size_t segmentSize; /* Select a segment */ COVER_segment_t segment = COVER_selectSegment( ctx, freqs, activeDmers, epochBegin, epochEnd, parameters);- /* Trim the segment if necessary and if it is empty then we are done */+ /* If the segment covers no dmers, then we are out of content.+ * There may be new content in other epochs, for continue for some time.+ */+ if (segment.score == 0) {+ if (++zeroScoreRun >= maxZeroScoreRun) {+ break;+ }+ continue;+ }+ zeroScoreRun = 0;+ /* Trim the segment if necessary and if it is too small then we are done */ segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);- if (segmentSize == 0) {+ if (segmentSize < parameters.d) { break; } /* We fill the dictionary from the back to allow the best segments to be@@ -600,40 +709,25 @@ memcpy(dict + tail, ctx->samples + segment.begin, segmentSize); DISPLAYUPDATE( 2, "\r%u%% ",- (U32)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity));+ (unsigned)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity)); } DISPLAYLEVEL(2, "\r%79s\r", ""); return tail; } -/**- * Translate from COVER_params_t to ZDICT_params_t required for finalizing the- * dictionary.- */-static ZDICT_params_t COVER_translateParams(COVER_params_t parameters) {- ZDICT_params_t zdictParams;- memset(&zdictParams, 0, sizeof(zdictParams));- zdictParams.notificationLevel = 1;- zdictParams.dictID = parameters.dictID;- zdictParams.compressionLevel = parameters.compressionLevel;- return zdictParams;-}--/**- * Constructs a dictionary using a heuristic based on the following paper:- *- * Liao, Petri, Moffat, Wirth- * Effective Construction of Relative Lempel-Ziv Dictionaries- * Published in WWW 2016.- */-ZDICTLIB_API size_t COVER_trainFromBuffer(- void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer,- const size_t *samplesSizes, unsigned nbSamples, COVER_params_t parameters) {- BYTE *const dict = (BYTE *)dictBuffer;+ZDICTLIB_API size_t ZDICT_trainFromBuffer_cover(+ void *dictBuffer, size_t dictBufferCapacity,+ const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,+ ZDICT_cover_params_t parameters)+{+ BYTE* const dict = (BYTE*)dictBuffer; COVER_ctx_t ctx; COVER_map_t activeDmers;+ parameters.splitPoint = 1.0;+ /* Initialize global data */+ g_displayLevel = parameters.zParams.notificationLevel; /* Checks */- if (!COVER_checkParameters(parameters)) {+ if (!COVER_checkParameters(parameters, dictBufferCapacity)) { DISPLAYLEVEL(1, "Cover parameters incorrect\n"); return ERROR(GENERIC); }@@ -646,14 +740,12 @@ ZDICT_DICTSIZE_MIN); return ERROR(dstSize_tooSmall); }- /* Initialize global data */- g_displayLevel = parameters.notificationLevel; /* Initialize context and activeDmers */ if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,- parameters.d)) {- DISPLAYLEVEL(1, "Failed to initialize context\n");+ parameters.d, parameters.splitPoint)) { return ERROR(GENERIC); }+ COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.suffixSize, g_displayLevel); if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) { DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n"); COVER_ctx_destroy(&ctx);@@ -665,13 +757,12 @@ const size_t tail = COVER_buildDictionary(&ctx, ctx.freqs, &activeDmers, dictBuffer, dictBufferCapacity, parameters);- ZDICT_params_t zdictParams = COVER_translateParams(parameters); const size_t dictionarySize = ZDICT_finalizeDictionary( dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,- samplesBuffer, samplesSizes, nbSamples, zdictParams);+ samplesBuffer, samplesSizes, nbSamples, parameters.zParams); if (!ZSTD_isError(dictionarySize)) { DISPLAYLEVEL(2, "Constructed dictionary of size %u\n",- (U32)dictionarySize);+ (unsigned)dictionarySize); } COVER_ctx_destroy(&ctx); COVER_map_destroy(&activeDmers);@@ -679,38 +770,69 @@ } } -/**- * COVER_best_t is used for two purposes:- * 1. Synchronizing threads.- * 2. Saving the best parameters and dictionary.- *- * All of the methods except COVER_best_init() are thread safe if zstd is- * compiled with multithreaded support.- */-typedef struct COVER_best_s {-#ifdef ZSTD_PTHREAD- pthread_mutex_t mutex;- pthread_cond_t cond;- size_t liveJobs;-#endif- void *dict;- size_t dictSize;- COVER_params_t parameters;- size_t compressedSize;-} COVER_best_t; ++size_t COVER_checkTotalCompressedSize(const ZDICT_cover_params_t parameters,+ const size_t *samplesSizes, const BYTE *samples,+ size_t *offsets,+ size_t nbTrainSamples, size_t nbSamples,+ BYTE *const dict, size_t dictBufferCapacity) {+ size_t totalCompressedSize = ERROR(GENERIC);+ /* Pointers */+ ZSTD_CCtx *cctx;+ ZSTD_CDict *cdict;+ void *dst;+ /* Local variables */+ size_t dstCapacity;+ size_t i;+ /* Allocate dst with enough space to compress the maximum sized sample */+ {+ size_t maxSampleSize = 0;+ i = parameters.splitPoint < 1.0 ? nbTrainSamples : 0;+ for (; i < nbSamples; ++i) {+ maxSampleSize = MAX(samplesSizes[i], maxSampleSize);+ }+ dstCapacity = ZSTD_compressBound(maxSampleSize);+ dst = malloc(dstCapacity);+ }+ /* Create the cctx and cdict */+ cctx = ZSTD_createCCtx();+ cdict = ZSTD_createCDict(dict, dictBufferCapacity,+ parameters.zParams.compressionLevel);+ if (!dst || !cctx || !cdict) {+ goto _compressCleanup;+ }+ /* Compress each sample and sum their sizes (or error) */+ totalCompressedSize = dictBufferCapacity;+ i = parameters.splitPoint < 1.0 ? nbTrainSamples : 0;+ for (; i < nbSamples; ++i) {+ const size_t size = ZSTD_compress_usingCDict(+ cctx, dst, dstCapacity, samples + offsets[i],+ samplesSizes[i], cdict);+ if (ZSTD_isError(size)) {+ totalCompressedSize = ERROR(GENERIC);+ goto _compressCleanup;+ }+ totalCompressedSize += size;+ }+_compressCleanup:+ ZSTD_freeCCtx(cctx);+ ZSTD_freeCDict(cdict);+ if (dst) {+ free(dst);+ }+ return totalCompressedSize;+}++ /** * Initialize the `COVER_best_t`. */-static void COVER_best_init(COVER_best_t *best) {- if (!best) {- return;- }-#ifdef ZSTD_PTHREAD- pthread_mutex_init(&best->mutex, NULL);- pthread_cond_init(&best->cond, NULL);+void COVER_best_init(COVER_best_t *best) {+ if (best==NULL) return; /* compatible with init on NULL */+ (void)ZSTD_pthread_mutex_init(&best->mutex, NULL);+ (void)ZSTD_pthread_cond_init(&best->cond, NULL); best->liveJobs = 0;-#endif best->dict = NULL; best->dictSize = 0; best->compressedSize = (size_t)-1;@@ -720,23 +842,21 @@ /** * Wait until liveJobs == 0. */-static void COVER_best_wait(COVER_best_t *best) {+void COVER_best_wait(COVER_best_t *best) { if (!best) { return; }-#ifdef ZSTD_PTHREAD- pthread_mutex_lock(&best->mutex);+ ZSTD_pthread_mutex_lock(&best->mutex); while (best->liveJobs != 0) {- pthread_cond_wait(&best->cond, &best->mutex);+ ZSTD_pthread_cond_wait(&best->cond, &best->mutex); }- pthread_mutex_unlock(&best->mutex);-#endif+ ZSTD_pthread_mutex_unlock(&best->mutex); } /** * Call COVER_best_wait() and then destroy the COVER_best_t. */-static void COVER_best_destroy(COVER_best_t *best) {+void COVER_best_destroy(COVER_best_t *best) { if (!best) { return; }@@ -744,25 +864,21 @@ if (best->dict) { free(best->dict); }-#ifdef ZSTD_PTHREAD- pthread_mutex_destroy(&best->mutex);- pthread_cond_destroy(&best->cond);-#endif+ ZSTD_pthread_mutex_destroy(&best->mutex);+ ZSTD_pthread_cond_destroy(&best->cond); } /** * Called when a thread is about to be launched. * Increments liveJobs. */-static void COVER_best_start(COVER_best_t *best) {+void COVER_best_start(COVER_best_t *best) { if (!best) { return; }-#ifdef ZSTD_PTHREAD- pthread_mutex_lock(&best->mutex);+ ZSTD_pthread_mutex_lock(&best->mutex); ++best->liveJobs;- pthread_mutex_unlock(&best->mutex);-#endif+ ZSTD_pthread_mutex_unlock(&best->mutex); } /**@@ -770,19 +886,17 @@ * Decrements liveJobs and signals any waiting threads if liveJobs == 0. * If this dictionary is the best so far save it and its parameters. */-static void COVER_best_finish(COVER_best_t *best, size_t compressedSize,- COVER_params_t parameters, void *dict,+void COVER_best_finish(COVER_best_t *best, size_t compressedSize,+ ZDICT_cover_params_t parameters, void *dict, size_t dictSize) { if (!best) { return; } {-#ifdef ZSTD_PTHREAD size_t liveJobs;- pthread_mutex_lock(&best->mutex);+ ZSTD_pthread_mutex_lock(&best->mutex); --best->liveJobs; liveJobs = best->liveJobs;-#endif /* If the new dictionary is better */ if (compressedSize < best->compressedSize) { /* Allocate space if necessary */@@ -794,6 +908,8 @@ if (!best->dict) { best->compressedSize = ERROR(GENERIC); best->dictSize = 0;+ ZSTD_pthread_cond_signal(&best->cond);+ ZSTD_pthread_mutex_unlock(&best->mutex); return; } }@@ -803,12 +919,10 @@ best->parameters = parameters; best->compressedSize = compressedSize; }-#ifdef ZSTD_PTHREAD- pthread_mutex_unlock(&best->mutex); if (liveJobs == 0) {- pthread_cond_broadcast(&best->cond);+ ZSTD_pthread_cond_broadcast(&best->cond); }-#endif+ ZSTD_pthread_mutex_unlock(&best->mutex); } } @@ -819,11 +933,11 @@ const COVER_ctx_t *ctx; COVER_best_t *best; size_t dictBufferCapacity;- COVER_params_t parameters;+ ZDICT_cover_params_t parameters; } COVER_tryParameters_data_t; /**- * Tries a set of parameters and upates the COVER_best_t with the results.+ * Tries a set of parameters and updates the COVER_best_t with the results. * This function is thread safe if zstd is compiled with multithreaded support. * It takes its parameters as an *OWNING* opaque pointer to support threading. */@@ -831,7 +945,7 @@ /* Save parameters as local variables */ COVER_tryParameters_data_t *const data = (COVER_tryParameters_data_t *)opaque; const COVER_ctx_t *const ctx = data->ctx;- const COVER_params_t parameters = data->parameters;+ const ZDICT_cover_params_t parameters = data->parameters; size_t dictBufferCapacity = data->dictBufferCapacity; size_t totalCompressedSize = ERROR(GENERIC); /* Allocate space for hash table, dict, and freqs */@@ -852,59 +966,20 @@ { const size_t tail = COVER_buildDictionary(ctx, freqs, &activeDmers, dict, dictBufferCapacity, parameters);- const ZDICT_params_t zdictParams = COVER_translateParams(parameters); dictBufferCapacity = ZDICT_finalizeDictionary( dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,- ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbSamples, zdictParams);+ ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbTrainSamples,+ parameters.zParams); if (ZDICT_isError(dictBufferCapacity)) { DISPLAYLEVEL(1, "Failed to finalize dictionary\n"); goto _cleanup; } } /* Check total compressed size */- {- /* Pointers */- ZSTD_CCtx *cctx;- ZSTD_CDict *cdict;- void *dst;- /* Local variables */- size_t dstCapacity;- size_t i;- /* Allocate dst with enough space to compress the maximum sized sample */- {- size_t maxSampleSize = 0;- for (i = 0; i < ctx->nbSamples; ++i) {- maxSampleSize = MAX(ctx->samplesSizes[i], maxSampleSize);- }- dstCapacity = ZSTD_compressBound(maxSampleSize);- dst = malloc(dstCapacity);- }- /* Create the cctx and cdict */- cctx = ZSTD_createCCtx();- cdict =- ZSTD_createCDict(dict, dictBufferCapacity, parameters.compressionLevel);- if (!dst || !cctx || !cdict) {- goto _compressCleanup;- }- /* Compress each sample and sum their sizes (or error) */- totalCompressedSize = 0;- for (i = 0; i < ctx->nbSamples; ++i) {- const size_t size = ZSTD_compress_usingCDict(- cctx, dst, dstCapacity, ctx->samples + ctx->offsets[i],- ctx->samplesSizes[i], cdict);- if (ZSTD_isError(size)) {- totalCompressedSize = ERROR(GENERIC);- goto _compressCleanup;- }- totalCompressedSize += size;- }- _compressCleanup:- ZSTD_freeCCtx(cctx);- ZSTD_freeCDict(cdict);- if (dst) {- free(dst);- }- }+ totalCompressedSize = COVER_checkTotalCompressedSize(parameters, ctx->samplesSizes,+ ctx->samples, ctx->offsets,+ ctx->nbTrainSamples, ctx->nbSamples,+ dict, dictBufferCapacity); _cleanup: COVER_best_finish(data->best, totalCompressedSize, parameters, dict,@@ -919,28 +994,36 @@ } } -ZDICTLIB_API size_t COVER_optimizeTrainFromBuffer(void *dictBuffer,- size_t dictBufferCapacity,- const void *samplesBuffer,- const size_t *samplesSizes,- unsigned nbSamples,- COVER_params_t *parameters) {+ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(+ void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer,+ const size_t *samplesSizes, unsigned nbSamples,+ ZDICT_cover_params_t *parameters) { /* constants */+ const unsigned nbThreads = parameters->nbThreads;+ const double splitPoint =+ parameters->splitPoint <= 0.0 ? DEFAULT_SPLITPOINT : parameters->splitPoint; const unsigned kMinD = parameters->d == 0 ? 6 : parameters->d;- const unsigned kMaxD = parameters->d == 0 ? 16 : parameters->d;- const unsigned kMinK = parameters->k == 0 ? kMaxD : parameters->k;- const unsigned kMaxK = parameters->k == 0 ? 2048 : parameters->k;- const unsigned kSteps = parameters->steps == 0 ? 256 : parameters->steps;+ const unsigned kMaxD = parameters->d == 0 ? 8 : parameters->d;+ const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k;+ const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k;+ const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps; const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1); const unsigned kIterations = (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize); /* Local variables */- const int displayLevel = parameters->notificationLevel;+ const int displayLevel = parameters->zParams.notificationLevel; unsigned iteration = 1; unsigned d; unsigned k; COVER_best_t best;+ POOL_ctx *pool = NULL;+ int warned = 0;+ /* Checks */+ if (splitPoint <= 0 || splitPoint > 1) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");+ return ERROR(GENERIC);+ } if (kMinK < kMaxD || kMaxK < kMinK) { LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n"); return ERROR(GENERIC);@@ -954,10 +1037,16 @@ ZDICT_DICTSIZE_MIN); return ERROR(dstSize_tooSmall); }+ if (nbThreads > 1) {+ pool = POOL_create(nbThreads, 1);+ if (!pool) {+ return ERROR(memory_allocation);+ }+ } /* Initialization */ COVER_best_init(&best); /* Turn down global display level to clean up display at level 2 and below */- g_displayLevel = parameters->notificationLevel - 1;+ g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1; /* Loop through d first because each new value needs a new context */ LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n", kIterations);@@ -965,11 +1054,16 @@ /* Initialize the context for this value of d */ COVER_ctx_t ctx; LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);- if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d)) {+ if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint)) { LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n"); COVER_best_destroy(&best);+ POOL_free(pool); return ERROR(GENERIC); }+ if (!warned) {+ COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.suffixSize, displayLevel);+ warned = 1;+ } /* Loop through k reusing the same context */ for (k = kMinK; k <= kMaxK; k += kStepSize) { /* Prepare the arguments */@@ -980,6 +1074,7 @@ LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n"); COVER_best_destroy(&best); COVER_ctx_destroy(&ctx);+ POOL_free(pool); return ERROR(GENERIC); } data->ctx = &ctx;@@ -988,18 +1083,25 @@ data->parameters = *parameters; data->parameters.k = k; data->parameters.d = d;+ data->parameters.splitPoint = splitPoint; data->parameters.steps = kSteps;+ data->parameters.zParams.notificationLevel = g_displayLevel; /* Check the parameters */- if (!COVER_checkParameters(data->parameters)) {+ if (!COVER_checkParameters(data->parameters, dictBufferCapacity)) { DISPLAYLEVEL(1, "Cover parameters incorrect\n");+ free(data); continue; } /* Call the function and pass ownership of data to it */ COVER_best_start(&best);- COVER_tryParameters(data);+ if (pool) {+ POOL_add(pool, &COVER_tryParameters, data);+ } else {+ COVER_tryParameters(data);+ } /* Print status */ LOCALDISPLAYUPDATE(displayLevel, 2, "\r%u%% ",- (U32)((iteration * 100) / kIterations));+ (unsigned)((iteration * 100) / kIterations)); ++iteration; } COVER_best_wait(&best);@@ -1010,12 +1112,15 @@ { const size_t dictSize = best.dictSize; if (ZSTD_isError(best.compressedSize)) {+ const size_t compressedSize = best.compressedSize; COVER_best_destroy(&best);- return best.compressedSize;+ POOL_free(pool);+ return compressedSize; } *parameters = best.parameters; memcpy(dictBuffer, best.dict, dictSize); COVER_best_destroy(&best);+ POOL_free(pool); return dictSize; } }
+ zstd/lib/dictBuilder/cover.h view
@@ -0,0 +1,112 @@+#include <stdio.h> /* fprintf */+#include <stdlib.h> /* malloc, free, qsort */+#include <string.h> /* memset */+#include <time.h> /* clock */+#include "mem.h" /* read */+#include "pool.h"+#include "threading.h"+#include "zstd_internal.h" /* includes zstd.h */+#ifndef ZDICT_STATIC_LINKING_ONLY+#define ZDICT_STATIC_LINKING_ONLY+#endif+#include "zdict.h"++/**+ * COVER_best_t is used for two purposes:+ * 1. Synchronizing threads.+ * 2. Saving the best parameters and dictionary.+ *+ * All of the methods except COVER_best_init() are thread safe if zstd is+ * compiled with multithreaded support.+ */+typedef struct COVER_best_s {+ ZSTD_pthread_mutex_t mutex;+ ZSTD_pthread_cond_t cond;+ size_t liveJobs;+ void *dict;+ size_t dictSize;+ ZDICT_cover_params_t parameters;+ size_t compressedSize;+} COVER_best_t;++/**+ * A segment is a range in the source as well as the score of the segment.+ */+typedef struct {+ U32 begin;+ U32 end;+ U32 score;+} COVER_segment_t;++/**+ *Number of epochs and size of each epoch.+ */+typedef struct {+ U32 num;+ U32 size;+} COVER_epoch_info_t;++/**+ * Computes the number of epochs and the size of each epoch.+ * We will make sure that each epoch gets at least 10 * k bytes.+ *+ * The COVER algorithms divide the data up into epochs of equal size and+ * select one segment from each epoch.+ *+ * @param maxDictSize The maximum allowed dictionary size.+ * @param nbDmers The number of dmers we are training on.+ * @param k The parameter k (segment size).+ * @param passes The target number of passes over the dmer corpus.+ * More passes means a better dictionary.+ */+COVER_epoch_info_t COVER_computeEpochs(U32 maxDictSize, U32 nbDmers,+ U32 k, U32 passes);++/**+ * Warns the user when their corpus is too small.+ */+void COVER_warnOnSmallCorpus(size_t maxDictSize, size_t nbDmers, int displayLevel);++/**+ * Checks total compressed size of a dictionary+ */+size_t COVER_checkTotalCompressedSize(const ZDICT_cover_params_t parameters,+ const size_t *samplesSizes, const BYTE *samples,+ size_t *offsets,+ size_t nbTrainSamples, size_t nbSamples,+ BYTE *const dict, size_t dictBufferCapacity);++/**+ * Returns the sum of the sample sizes.+ */+size_t COVER_sum(const size_t *samplesSizes, unsigned nbSamples) ;++/**+ * Initialize the `COVER_best_t`.+ */+void COVER_best_init(COVER_best_t *best);++/**+ * Wait until liveJobs == 0.+ */+void COVER_best_wait(COVER_best_t *best);++/**+ * Call COVER_best_wait() and then destroy the COVER_best_t.+ */+void COVER_best_destroy(COVER_best_t *best);++/**+ * Called when a thread is about to be launched.+ * Increments liveJobs.+ */+void COVER_best_start(COVER_best_t *best);++/**+ * Called when a thread finishes executing, both on error or success.+ * Decrements liveJobs and signals any waiting threads if liveJobs == 0.+ * If this dictionary is the best so far save it and its parameters.+ */+void COVER_best_finish(COVER_best_t *best, size_t compressedSize,+ ZDICT_cover_params_t parameters, void *dict,+ size_t dictSize);
zstd/lib/dictBuilder/divsufsort.c view
@@ -1637,7 +1637,7 @@ if(0 <= c2) { BUCKET_B(c2, c1) = k - SA; } k = SA + BUCKET_B(c2 = c0, c1); }- assert(k < j);+ assert(k < j); assert(k != NULL); *k-- = s; } else { assert(((s == 0) && (T[s] == c1)) || (s < 0));@@ -1701,7 +1701,7 @@ if(0 <= c2) { BUCKET_B(c2, c1) = k - SA; } k = SA + BUCKET_B(c2 = c0, c1); }- assert(k < j);+ assert(k < j); assert(k != NULL); *k-- = s; } else if(s != 0) { *j = ~s;@@ -1785,7 +1785,7 @@ if(0 <= c2) { BUCKET_B(c2, c1) = k - SA; } k = SA + BUCKET_B(c2 = c0, c1); }- assert(k < j);+ assert(k < j); assert(k != NULL); *k-- = s; } else if(s != 0) { *j = ~s;
+ zstd/lib/dictBuilder/fastcover.c view
@@ -0,0 +1,740 @@+/*-*************************************+* Dependencies+***************************************/+#include <stdio.h> /* fprintf */+#include <stdlib.h> /* malloc, free, qsort */+#include <string.h> /* memset */+#include <time.h> /* clock */++#include "mem.h" /* read */+#include "pool.h"+#include "threading.h"+#include "cover.h"+#include "zstd_internal.h" /* includes zstd.h */+#ifndef ZDICT_STATIC_LINKING_ONLY+#define ZDICT_STATIC_LINKING_ONLY+#endif+#include "zdict.h"+++/*-*************************************+* Constants+***************************************/+#define FASTCOVER_MAX_SAMPLES_SIZE (sizeof(size_t) == 8 ? ((unsigned)-1) : ((unsigned)1 GB))+#define FASTCOVER_MAX_F 31+#define FASTCOVER_MAX_ACCEL 10+#define DEFAULT_SPLITPOINT 0.75+#define DEFAULT_F 20+#define DEFAULT_ACCEL 1+++/*-*************************************+* Console display+***************************************/+static int g_displayLevel = 2;+#define DISPLAY(...) \+ { \+ fprintf(stderr, __VA_ARGS__); \+ fflush(stderr); \+ }+#define LOCALDISPLAYLEVEL(displayLevel, l, ...) \+ if (displayLevel >= l) { \+ DISPLAY(__VA_ARGS__); \+ } /* 0 : no display; 1: errors; 2: default; 3: details; 4: debug */+#define DISPLAYLEVEL(l, ...) LOCALDISPLAYLEVEL(g_displayLevel, l, __VA_ARGS__)++#define LOCALDISPLAYUPDATE(displayLevel, l, ...) \+ if (displayLevel >= l) { \+ if ((clock() - g_time > refreshRate) || (displayLevel >= 4)) { \+ g_time = clock(); \+ DISPLAY(__VA_ARGS__); \+ } \+ }+#define DISPLAYUPDATE(l, ...) LOCALDISPLAYUPDATE(g_displayLevel, l, __VA_ARGS__)+static const clock_t refreshRate = CLOCKS_PER_SEC * 15 / 100;+static clock_t g_time = 0;+++/*-*************************************+* Hash Functions+***************************************/+static const U64 prime6bytes = 227718039650203ULL;+static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64-48)) * prime6bytes) >> (64-h)) ; }+static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }++static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;+static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }+static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }+++/**+ * Hash the d-byte value pointed to by p and mod 2^f+ */+static size_t FASTCOVER_hashPtrToIndex(const void* p, U32 h, unsigned d) {+ if (d == 6) {+ return ZSTD_hash6Ptr(p, h) & ((1 << h) - 1);+ }+ return ZSTD_hash8Ptr(p, h) & ((1 << h) - 1);+}+++/*-*************************************+* Acceleration+***************************************/+typedef struct {+ unsigned finalize; /* Percentage of training samples used for ZDICT_finalizeDictionary */+ unsigned skip; /* Number of dmer skipped between each dmer counted in computeFrequency */+} FASTCOVER_accel_t;+++static const FASTCOVER_accel_t FASTCOVER_defaultAccelParameters[FASTCOVER_MAX_ACCEL+1] = {+ { 100, 0 }, /* accel = 0, should not happen because accel = 0 defaults to accel = 1 */+ { 100, 0 }, /* accel = 1 */+ { 50, 1 }, /* accel = 2 */+ { 34, 2 }, /* accel = 3 */+ { 25, 3 }, /* accel = 4 */+ { 20, 4 }, /* accel = 5 */+ { 17, 5 }, /* accel = 6 */+ { 14, 6 }, /* accel = 7 */+ { 13, 7 }, /* accel = 8 */+ { 11, 8 }, /* accel = 9 */+ { 10, 9 }, /* accel = 10 */+};+++/*-*************************************+* Context+***************************************/+typedef struct {+ const BYTE *samples;+ size_t *offsets;+ const size_t *samplesSizes;+ size_t nbSamples;+ size_t nbTrainSamples;+ size_t nbTestSamples;+ size_t nbDmers;+ U32 *freqs;+ unsigned d;+ unsigned f;+ FASTCOVER_accel_t accelParams;+} FASTCOVER_ctx_t;+++/*-*************************************+* Helper functions+***************************************/+/**+ * Selects the best segment in an epoch.+ * Segments of are scored according to the function:+ *+ * Let F(d) be the frequency of all dmers with hash value d.+ * Let S_i be hash value of the dmer at position i of segment S which has length k.+ *+ * Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1})+ *+ * Once the dmer with hash value d is in the dictionary we set F(d) = 0.+ */+static COVER_segment_t FASTCOVER_selectSegment(const FASTCOVER_ctx_t *ctx,+ U32 *freqs, U32 begin, U32 end,+ ZDICT_cover_params_t parameters,+ U16* segmentFreqs) {+ /* Constants */+ const U32 k = parameters.k;+ const U32 d = parameters.d;+ const U32 f = ctx->f;+ const U32 dmersInK = k - d + 1;++ /* Try each segment (activeSegment) and save the best (bestSegment) */+ COVER_segment_t bestSegment = {0, 0, 0};+ COVER_segment_t activeSegment;++ /* Reset the activeDmers in the segment */+ /* The activeSegment starts at the beginning of the epoch. */+ activeSegment.begin = begin;+ activeSegment.end = begin;+ activeSegment.score = 0;++ /* Slide the activeSegment through the whole epoch.+ * Save the best segment in bestSegment.+ */+ while (activeSegment.end < end) {+ /* Get hash value of current dmer */+ const size_t idx = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.end, f, d);++ /* Add frequency of this index to score if this is the first occurrence of index in active segment */+ if (segmentFreqs[idx] == 0) {+ activeSegment.score += freqs[idx];+ }+ /* Increment end of segment and segmentFreqs*/+ activeSegment.end += 1;+ segmentFreqs[idx] += 1;+ /* If the window is now too large, drop the first position */+ if (activeSegment.end - activeSegment.begin == dmersInK + 1) {+ /* Get hash value of the dmer to be eliminated from active segment */+ const size_t delIndex = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.begin, f, d);+ segmentFreqs[delIndex] -= 1;+ /* Subtract frequency of this index from score if this is the last occurrence of this index in active segment */+ if (segmentFreqs[delIndex] == 0) {+ activeSegment.score -= freqs[delIndex];+ }+ /* Increment start of segment */+ activeSegment.begin += 1;+ }++ /* If this segment is the best so far save it */+ if (activeSegment.score > bestSegment.score) {+ bestSegment = activeSegment;+ }+ }++ /* Zero out rest of segmentFreqs array */+ while (activeSegment.begin < end) {+ const size_t delIndex = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.begin, f, d);+ segmentFreqs[delIndex] -= 1;+ activeSegment.begin += 1;+ }++ {+ /* Zero the frequency of hash value of each dmer covered by the chosen segment. */+ U32 pos;+ for (pos = bestSegment.begin; pos != bestSegment.end; ++pos) {+ const size_t i = FASTCOVER_hashPtrToIndex(ctx->samples + pos, f, d);+ freqs[i] = 0;+ }+ }++ return bestSegment;+}+++static int FASTCOVER_checkParameters(ZDICT_cover_params_t parameters,+ size_t maxDictSize, unsigned f,+ unsigned accel) {+ /* k, d, and f are required parameters */+ if (parameters.d == 0 || parameters.k == 0) {+ return 0;+ }+ /* d has to be 6 or 8 */+ if (parameters.d != 6 && parameters.d != 8) {+ return 0;+ }+ /* k <= maxDictSize */+ if (parameters.k > maxDictSize) {+ return 0;+ }+ /* d <= k */+ if (parameters.d > parameters.k) {+ return 0;+ }+ /* 0 < f <= FASTCOVER_MAX_F*/+ if (f > FASTCOVER_MAX_F || f == 0) {+ return 0;+ }+ /* 0 < splitPoint <= 1 */+ if (parameters.splitPoint <= 0 || parameters.splitPoint > 1) {+ return 0;+ }+ /* 0 < accel <= 10 */+ if (accel > 10 || accel == 0) {+ return 0;+ }+ return 1;+}+++/**+ * Clean up a context initialized with `FASTCOVER_ctx_init()`.+ */+static void+FASTCOVER_ctx_destroy(FASTCOVER_ctx_t* ctx)+{+ if (!ctx) return;++ free(ctx->freqs);+ ctx->freqs = NULL;++ free(ctx->offsets);+ ctx->offsets = NULL;+}+++/**+ * Calculate for frequency of hash value of each dmer in ctx->samples+ */+static void+FASTCOVER_computeFrequency(U32* freqs, const FASTCOVER_ctx_t* ctx)+{+ const unsigned f = ctx->f;+ const unsigned d = ctx->d;+ const unsigned skip = ctx->accelParams.skip;+ const unsigned readLength = MAX(d, 8);+ size_t i;+ assert(ctx->nbTrainSamples >= 5);+ assert(ctx->nbTrainSamples <= ctx->nbSamples);+ for (i = 0; i < ctx->nbTrainSamples; i++) {+ size_t start = ctx->offsets[i]; /* start of current dmer */+ size_t const currSampleEnd = ctx->offsets[i+1];+ while (start + readLength <= currSampleEnd) {+ const size_t dmerIndex = FASTCOVER_hashPtrToIndex(ctx->samples + start, f, d);+ freqs[dmerIndex]++;+ start = start + skip + 1;+ }+ }+}+++/**+ * Prepare a context for dictionary building.+ * The context is only dependent on the parameter `d` and can used multiple+ * times.+ * Returns 1 on success or zero on error.+ * The context must be destroyed with `FASTCOVER_ctx_destroy()`.+ */+static int+FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx,+ const void* samplesBuffer,+ const size_t* samplesSizes, unsigned nbSamples,+ unsigned d, double splitPoint, unsigned f,+ FASTCOVER_accel_t accelParams)+{+ const BYTE* const samples = (const BYTE*)samplesBuffer;+ const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);+ /* Split samples into testing and training sets */+ const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples;+ const unsigned nbTestSamples = splitPoint < 1.0 ? nbSamples - nbTrainSamples : nbSamples;+ const size_t trainingSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes, nbTrainSamples) : totalSamplesSize;+ const size_t testSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize;++ /* Checks */+ if (totalSamplesSize < MAX(d, sizeof(U64)) ||+ totalSamplesSize >= (size_t)FASTCOVER_MAX_SAMPLES_SIZE) {+ DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",+ (unsigned)(totalSamplesSize >> 20), (FASTCOVER_MAX_SAMPLES_SIZE >> 20));+ return 0;+ }++ /* Check if there are at least 5 training samples */+ if (nbTrainSamples < 5) {+ DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid\n", nbTrainSamples);+ return 0;+ }++ /* Check if there's testing sample */+ if (nbTestSamples < 1) {+ DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.\n", nbTestSamples);+ return 0;+ }++ /* Zero the context */+ memset(ctx, 0, sizeof(*ctx));+ DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples,+ (unsigned)trainingSamplesSize);+ DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples,+ (unsigned)testSamplesSize);++ ctx->samples = samples;+ ctx->samplesSizes = samplesSizes;+ ctx->nbSamples = nbSamples;+ ctx->nbTrainSamples = nbTrainSamples;+ ctx->nbTestSamples = nbTestSamples;+ ctx->nbDmers = trainingSamplesSize - MAX(d, sizeof(U64)) + 1;+ ctx->d = d;+ ctx->f = f;+ ctx->accelParams = accelParams;++ /* The offsets of each file */+ ctx->offsets = (size_t*)calloc((nbSamples + 1), sizeof(size_t));+ if (ctx->offsets == NULL) {+ DISPLAYLEVEL(1, "Failed to allocate scratch buffers \n");+ FASTCOVER_ctx_destroy(ctx);+ return 0;+ }++ /* Fill offsets from the samplesSizes */+ { U32 i;+ ctx->offsets[0] = 0;+ assert(nbSamples >= 5);+ for (i = 1; i <= nbSamples; ++i) {+ ctx->offsets[i] = ctx->offsets[i - 1] + samplesSizes[i - 1];+ }+ }++ /* Initialize frequency array of size 2^f */+ ctx->freqs = (U32*)calloc(((U64)1 << f), sizeof(U32));+ if (ctx->freqs == NULL) {+ DISPLAYLEVEL(1, "Failed to allocate frequency table \n");+ FASTCOVER_ctx_destroy(ctx);+ return 0;+ }++ DISPLAYLEVEL(2, "Computing frequencies\n");+ FASTCOVER_computeFrequency(ctx->freqs, ctx);++ return 1;+}+++/**+ * Given the prepared context build the dictionary.+ */+static size_t+FASTCOVER_buildDictionary(const FASTCOVER_ctx_t* ctx,+ U32* freqs,+ void* dictBuffer, size_t dictBufferCapacity,+ ZDICT_cover_params_t parameters,+ U16* segmentFreqs)+{+ BYTE *const dict = (BYTE *)dictBuffer;+ size_t tail = dictBufferCapacity;+ /* Divide the data into epochs. We will select one segment from each epoch. */+ const COVER_epoch_info_t epochs = COVER_computeEpochs(+ (U32)dictBufferCapacity, (U32)ctx->nbDmers, parameters.k, 1);+ const size_t maxZeroScoreRun = 10;+ size_t zeroScoreRun = 0;+ size_t epoch;+ DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n",+ (U32)epochs.num, (U32)epochs.size);+ /* Loop through the epochs until there are no more segments or the dictionary+ * is full.+ */+ for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs.num) {+ const U32 epochBegin = (U32)(epoch * epochs.size);+ const U32 epochEnd = epochBegin + epochs.size;+ size_t segmentSize;+ /* Select a segment */+ COVER_segment_t segment = FASTCOVER_selectSegment(+ ctx, freqs, epochBegin, epochEnd, parameters, segmentFreqs);++ /* If the segment covers no dmers, then we are out of content.+ * There may be new content in other epochs, for continue for some time.+ */+ if (segment.score == 0) {+ if (++zeroScoreRun >= maxZeroScoreRun) {+ break;+ }+ continue;+ }+ zeroScoreRun = 0;++ /* Trim the segment if necessary and if it is too small then we are done */+ segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);+ if (segmentSize < parameters.d) {+ break;+ }++ /* We fill the dictionary from the back to allow the best segments to be+ * referenced with the smallest offsets.+ */+ tail -= segmentSize;+ memcpy(dict + tail, ctx->samples + segment.begin, segmentSize);+ DISPLAYUPDATE(+ 2, "\r%u%% ",+ (unsigned)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity));+ }+ DISPLAYLEVEL(2, "\r%79s\r", "");+ return tail;+}+++/**+ * Parameters for FASTCOVER_tryParameters().+ */+typedef struct FASTCOVER_tryParameters_data_s {+ const FASTCOVER_ctx_t* ctx;+ COVER_best_t* best;+ size_t dictBufferCapacity;+ ZDICT_cover_params_t parameters;+} FASTCOVER_tryParameters_data_t;+++/**+ * Tries a set of parameters and updates the COVER_best_t with the results.+ * This function is thread safe if zstd is compiled with multithreaded support.+ * It takes its parameters as an *OWNING* opaque pointer to support threading.+ */+static void FASTCOVER_tryParameters(void *opaque)+{+ /* Save parameters as local variables */+ FASTCOVER_tryParameters_data_t *const data = (FASTCOVER_tryParameters_data_t *)opaque;+ const FASTCOVER_ctx_t *const ctx = data->ctx;+ const ZDICT_cover_params_t parameters = data->parameters;+ size_t dictBufferCapacity = data->dictBufferCapacity;+ size_t totalCompressedSize = ERROR(GENERIC);+ /* Initialize array to keep track of frequency of dmer within activeSegment */+ U16* segmentFreqs = (U16 *)calloc(((U64)1 << ctx->f), sizeof(U16));+ /* Allocate space for hash table, dict, and freqs */+ BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);+ U32 *freqs = (U32*) malloc(((U64)1 << ctx->f) * sizeof(U32));+ if (!segmentFreqs || !dict || !freqs) {+ DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n");+ goto _cleanup;+ }+ /* Copy the frequencies because we need to modify them */+ memcpy(freqs, ctx->freqs, ((U64)1 << ctx->f) * sizeof(U32));+ /* Build the dictionary */+ { const size_t tail = FASTCOVER_buildDictionary(ctx, freqs, dict, dictBufferCapacity,+ parameters, segmentFreqs);+ const unsigned nbFinalizeSamples = (unsigned)(ctx->nbTrainSamples * ctx->accelParams.finalize / 100);+ dictBufferCapacity = ZDICT_finalizeDictionary(+ dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,+ ctx->samples, ctx->samplesSizes, nbFinalizeSamples, parameters.zParams);+ if (ZDICT_isError(dictBufferCapacity)) {+ DISPLAYLEVEL(1, "Failed to finalize dictionary\n");+ goto _cleanup;+ }+ }+ /* Check total compressed size */+ totalCompressedSize = COVER_checkTotalCompressedSize(parameters, ctx->samplesSizes,+ ctx->samples, ctx->offsets,+ ctx->nbTrainSamples, ctx->nbSamples,+ dict, dictBufferCapacity);+_cleanup:+ COVER_best_finish(data->best, totalCompressedSize, parameters, dict,+ dictBufferCapacity);+ free(data);+ free(segmentFreqs);+ free(dict);+ free(freqs);+}+++static void+FASTCOVER_convertToCoverParams(ZDICT_fastCover_params_t fastCoverParams,+ ZDICT_cover_params_t* coverParams)+{+ coverParams->k = fastCoverParams.k;+ coverParams->d = fastCoverParams.d;+ coverParams->steps = fastCoverParams.steps;+ coverParams->nbThreads = fastCoverParams.nbThreads;+ coverParams->splitPoint = fastCoverParams.splitPoint;+ coverParams->zParams = fastCoverParams.zParams;+}+++static void+FASTCOVER_convertToFastCoverParams(ZDICT_cover_params_t coverParams,+ ZDICT_fastCover_params_t* fastCoverParams,+ unsigned f, unsigned accel)+{+ fastCoverParams->k = coverParams.k;+ fastCoverParams->d = coverParams.d;+ fastCoverParams->steps = coverParams.steps;+ fastCoverParams->nbThreads = coverParams.nbThreads;+ fastCoverParams->splitPoint = coverParams.splitPoint;+ fastCoverParams->f = f;+ fastCoverParams->accel = accel;+ fastCoverParams->zParams = coverParams.zParams;+}+++ZDICTLIB_API size_t+ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity,+ const void* samplesBuffer,+ const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_fastCover_params_t parameters)+{+ BYTE* const dict = (BYTE*)dictBuffer;+ FASTCOVER_ctx_t ctx;+ ZDICT_cover_params_t coverParams;+ FASTCOVER_accel_t accelParams;+ /* Initialize global data */+ g_displayLevel = parameters.zParams.notificationLevel;+ /* Assign splitPoint and f if not provided */+ parameters.splitPoint = 1.0;+ parameters.f = parameters.f == 0 ? DEFAULT_F : parameters.f;+ parameters.accel = parameters.accel == 0 ? DEFAULT_ACCEL : parameters.accel;+ /* Convert to cover parameter */+ memset(&coverParams, 0 , sizeof(coverParams));+ FASTCOVER_convertToCoverParams(parameters, &coverParams);+ /* Checks */+ if (!FASTCOVER_checkParameters(coverParams, dictBufferCapacity, parameters.f,+ parameters.accel)) {+ DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n");+ return ERROR(GENERIC);+ }+ if (nbSamples == 0) {+ DISPLAYLEVEL(1, "FASTCOVER must have at least one input file\n");+ return ERROR(GENERIC);+ }+ if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {+ DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",+ ZDICT_DICTSIZE_MIN);+ return ERROR(dstSize_tooSmall);+ }+ /* Assign corresponding FASTCOVER_accel_t to accelParams*/+ accelParams = FASTCOVER_defaultAccelParameters[parameters.accel];+ /* Initialize context */+ if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,+ coverParams.d, parameters.splitPoint, parameters.f,+ accelParams)) {+ DISPLAYLEVEL(1, "Failed to initialize context\n");+ return ERROR(GENERIC);+ }+ COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.nbDmers, g_displayLevel);+ /* Build the dictionary */+ DISPLAYLEVEL(2, "Building dictionary\n");+ {+ /* Initialize array to keep track of frequency of dmer within activeSegment */+ U16* segmentFreqs = (U16 *)calloc(((U64)1 << parameters.f), sizeof(U16));+ const size_t tail = FASTCOVER_buildDictionary(&ctx, ctx.freqs, dictBuffer,+ dictBufferCapacity, coverParams, segmentFreqs);+ const unsigned nbFinalizeSamples = (unsigned)(ctx.nbTrainSamples * ctx.accelParams.finalize / 100);+ const size_t dictionarySize = ZDICT_finalizeDictionary(+ dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,+ samplesBuffer, samplesSizes, nbFinalizeSamples, coverParams.zParams);+ if (!ZSTD_isError(dictionarySize)) {+ DISPLAYLEVEL(2, "Constructed dictionary of size %u\n",+ (unsigned)dictionarySize);+ }+ FASTCOVER_ctx_destroy(&ctx);+ free(segmentFreqs);+ return dictionarySize;+ }+}+++ZDICTLIB_API size_t+ZDICT_optimizeTrainFromBuffer_fastCover(+ void* dictBuffer, size_t dictBufferCapacity,+ const void* samplesBuffer,+ const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_fastCover_params_t* parameters)+{+ ZDICT_cover_params_t coverParams;+ FASTCOVER_accel_t accelParams;+ /* constants */+ const unsigned nbThreads = parameters->nbThreads;+ const double splitPoint =+ parameters->splitPoint <= 0.0 ? DEFAULT_SPLITPOINT : parameters->splitPoint;+ const unsigned kMinD = parameters->d == 0 ? 6 : parameters->d;+ const unsigned kMaxD = parameters->d == 0 ? 8 : parameters->d;+ const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k;+ const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k;+ const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps;+ const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1);+ const unsigned kIterations =+ (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);+ const unsigned f = parameters->f == 0 ? DEFAULT_F : parameters->f;+ const unsigned accel = parameters->accel == 0 ? DEFAULT_ACCEL : parameters->accel;+ /* Local variables */+ const int displayLevel = parameters->zParams.notificationLevel;+ unsigned iteration = 1;+ unsigned d;+ unsigned k;+ COVER_best_t best;+ POOL_ctx *pool = NULL;+ int warned = 0;+ /* Checks */+ if (splitPoint <= 0 || splitPoint > 1) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect splitPoint\n");+ return ERROR(GENERIC);+ }+ if (accel == 0 || accel > FASTCOVER_MAX_ACCEL) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect accel\n");+ return ERROR(GENERIC);+ }+ if (kMinK < kMaxD || kMaxK < kMinK) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect k\n");+ return ERROR(GENERIC);+ }+ if (nbSamples == 0) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "FASTCOVER must have at least one input file\n");+ return ERROR(GENERIC);+ }+ if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "dictBufferCapacity must be at least %u\n",+ ZDICT_DICTSIZE_MIN);+ return ERROR(dstSize_tooSmall);+ }+ if (nbThreads > 1) {+ pool = POOL_create(nbThreads, 1);+ if (!pool) {+ return ERROR(memory_allocation);+ }+ }+ /* Initialization */+ COVER_best_init(&best);+ memset(&coverParams, 0 , sizeof(coverParams));+ FASTCOVER_convertToCoverParams(*parameters, &coverParams);+ accelParams = FASTCOVER_defaultAccelParameters[accel];+ /* Turn down global display level to clean up display at level 2 and below */+ g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1;+ /* Loop through d first because each new value needs a new context */+ LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n",+ kIterations);+ for (d = kMinD; d <= kMaxD; d += 2) {+ /* Initialize the context for this value of d */+ FASTCOVER_ctx_t ctx;+ LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);+ if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint, f, accelParams)) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");+ COVER_best_destroy(&best);+ POOL_free(pool);+ return ERROR(GENERIC);+ }+ if (!warned) {+ COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.nbDmers, displayLevel);+ warned = 1;+ }+ /* Loop through k reusing the same context */+ for (k = kMinK; k <= kMaxK; k += kStepSize) {+ /* Prepare the arguments */+ FASTCOVER_tryParameters_data_t *data = (FASTCOVER_tryParameters_data_t *)malloc(+ sizeof(FASTCOVER_tryParameters_data_t));+ LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k);+ if (!data) {+ LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n");+ COVER_best_destroy(&best);+ FASTCOVER_ctx_destroy(&ctx);+ POOL_free(pool);+ return ERROR(GENERIC);+ }+ data->ctx = &ctx;+ data->best = &best;+ data->dictBufferCapacity = dictBufferCapacity;+ data->parameters = coverParams;+ data->parameters.k = k;+ data->parameters.d = d;+ data->parameters.splitPoint = splitPoint;+ data->parameters.steps = kSteps;+ data->parameters.zParams.notificationLevel = g_displayLevel;+ /* Check the parameters */+ if (!FASTCOVER_checkParameters(data->parameters, dictBufferCapacity,+ data->ctx->f, accel)) {+ DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n");+ free(data);+ continue;+ }+ /* Call the function and pass ownership of data to it */+ COVER_best_start(&best);+ if (pool) {+ POOL_add(pool, &FASTCOVER_tryParameters, data);+ } else {+ FASTCOVER_tryParameters(data);+ }+ /* Print status */+ LOCALDISPLAYUPDATE(displayLevel, 2, "\r%u%% ",+ (unsigned)((iteration * 100) / kIterations));+ ++iteration;+ }+ COVER_best_wait(&best);+ FASTCOVER_ctx_destroy(&ctx);+ }+ LOCALDISPLAYLEVEL(displayLevel, 2, "\r%79s\r", "");+ /* Fill the output buffer and parameters with output of the best parameters */+ {+ const size_t dictSize = best.dictSize;+ if (ZSTD_isError(best.compressedSize)) {+ const size_t compressedSize = best.compressedSize;+ COVER_best_destroy(&best);+ POOL_free(pool);+ return compressedSize;+ }+ FASTCOVER_convertToFastCoverParams(best.parameters, parameters, f, accel);+ memcpy(dictBuffer, best.dict, dictSize);+ COVER_best_destroy(&best);+ POOL_free(pool);+ return dictSize;+ }++}
zstd/lib/dictBuilder/zdict.c view
@@ -1,18 +1,20 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ /*-************************************** * Tuning parameters ****************************************/+#define MINRATIO 4 /* minimum nb of apparition to be selected in dictionary */ #define ZDICT_MAX_SAMPLES_SIZE (2000U << 20)-#define ZDICT_MIN_SAMPLES_SIZE 512+#define ZDICT_MIN_SAMPLES_SIZE (ZDICT_CONTENTSIZE_MIN * MINRATIO) /*-**************************************@@ -59,11 +61,8 @@ #define NOISELENGTH 32 -#define MINRATIO 4-static const int g_compressionLevel_default = 6;+static const int g_compressionLevel_default = 3; static const U32 g_selectivity_default = 9;-static const size_t g_provision_entropySize = 200;-static const size_t g_min_fast_dictContent = 192; /*-*************************************@@ -96,7 +95,7 @@ unsigned ZDICT_getDictID(const void* dictBuffer, size_t dictSize) { if (dictSize < 8) return 0;- if (MEM_readLE32(dictBuffer) != ZSTD_DICT_MAGIC) return 0;+ if (MEM_readLE32(dictBuffer) != ZSTD_MAGIC_DICTIONARY) return 0; return MEM_readLE32((const char*)dictBuffer + 4); } @@ -104,7 +103,7 @@ /*-******************************************************** * Dictionary training functions **********************************************************/-static unsigned ZDICT_NbCommonBytes (register size_t val)+static unsigned ZDICT_NbCommonBytes (size_t val) { if (MEM_isLittleEndian()) { if (MEM_64bits()) {@@ -208,7 +207,6 @@ U32 cumulLength[LLIMIT] = {0}; U32 savings[LLIMIT] = {0}; const BYTE* b = (const BYTE*)buffer;- size_t length; size_t maxLength = LLIMIT; size_t pos = suffix[start]; U32 end = start;@@ -223,26 +221,30 @@ ||(MEM_read16(b+pos+1) == MEM_read16(b+pos+3)) ||(MEM_read16(b+pos+2) == MEM_read16(b+pos+4)) ) { /* skip and mark segment */- U16 u16 = MEM_read16(b+pos+4);- U32 u, e = 6;- while (MEM_read16(b+pos+e) == u16) e+=2 ;- if (b[pos+e] == b[pos+e-1]) e++;- for (u=1; u<e; u++)+ U16 const pattern16 = MEM_read16(b+pos+4);+ U32 u, patternEnd = 6;+ while (MEM_read16(b+pos+patternEnd) == pattern16) patternEnd+=2 ;+ if (b[pos+patternEnd] == b[pos+patternEnd-1]) patternEnd++;+ for (u=1; u<patternEnd; u++) doneMarks[pos+u] = 1; return solution; } /* look forward */- do {- end++;- length = ZDICT_count(b + pos, b + suffix[end]);- } while (length >=MINMATCHLENGTH);+ { size_t length;+ do {+ end++;+ length = ZDICT_count(b + pos, b + suffix[end]);+ } while (length >= MINMATCHLENGTH);+ } /* look backward */- do {- length = ZDICT_count(b + pos, b + *(suffix+start-1));- if (length >=MINMATCHLENGTH) start--;- } while(length >= MINMATCHLENGTH);+ { size_t length;+ do {+ length = ZDICT_count(b + pos, b + *(suffix+start-1));+ if (length >=MINMATCHLENGTH) start--;+ } while(length >= MINMATCHLENGTH);+ } /* exit if not found a minimum nb of repetitions */ if (end-start < minRatio) {@@ -253,15 +255,15 @@ } { int i;- U32 searchLength;+ U32 mml; U32 refinedStart = start; U32 refinedEnd = end; DISPLAYLEVEL(4, "\n");- DISPLAYLEVEL(4, "found %3u matches of length >= %i at pos %7u ", (U32)(end-start), MINMATCHLENGTH, (U32)pos);+ DISPLAYLEVEL(4, "found %3u matches of length >= %i at pos %7u ", (unsigned)(end-start), MINMATCHLENGTH, (unsigned)pos); DISPLAYLEVEL(4, "\n"); - for (searchLength = MINMATCHLENGTH ; ; searchLength++) {+ for (mml = MINMATCHLENGTH ; ; mml++) { BYTE currentChar = 0; U32 currentCount = 0; U32 currentID = refinedStart;@@ -269,13 +271,13 @@ U32 selectedCount = 0; U32 selectedID = currentID; for (id =refinedStart; id < refinedEnd; id++) {- if (b[ suffix[id] + searchLength] != currentChar) {+ if (b[suffix[id] + mml] != currentChar) { if (currentCount > selectedCount) { selectedCount = currentCount; selectedID = currentID; } currentID = id;- currentChar = b[ suffix[id] + searchLength];+ currentChar = b[ suffix[id] + mml]; currentCount = 0; } currentCount ++;@@ -291,27 +293,30 @@ refinedEnd = refinedStart + selectedCount; } - /* evaluate gain based on new ref */+ /* evaluate gain based on new dict */ start = refinedStart; pos = suffix[refinedStart]; end = start; memset(lengthList, 0, sizeof(lengthList)); /* look forward */- do {- end++;- length = ZDICT_count(b + pos, b + suffix[end]);- if (length >= LLIMIT) length = LLIMIT-1;- lengthList[length]++;- } while (length >=MINMATCHLENGTH);+ { size_t length;+ do {+ end++;+ length = ZDICT_count(b + pos, b + suffix[end]);+ if (length >= LLIMIT) length = LLIMIT-1;+ lengthList[length]++;+ } while (length >=MINMATCHLENGTH);+ } /* look backward */- length = MINMATCHLENGTH;- while ((length >= MINMATCHLENGTH) & (start > 0)) {- length = ZDICT_count(b + pos, b + suffix[start - 1]);- if (length >= LLIMIT) length = LLIMIT - 1;- lengthList[length]++;- if (length >= MINMATCHLENGTH) start--;+ { size_t length = MINMATCHLENGTH;+ while ((length >= MINMATCHLENGTH) & (start > 0)) {+ length = ZDICT_count(b + pos, b + suffix[start - 1]);+ if (length >= LLIMIT) length = LLIMIT - 1;+ lengthList[length]++;+ if (length >= MINMATCHLENGTH) start--;+ } } /* largest useful length */@@ -336,8 +341,8 @@ for (i=MINMATCHLENGTH; i<=(int)maxLength; i++) savings[i] = savings[i-1] + (lengthList[i] * (i-3)); - DISPLAYLEVEL(4, "Selected ref at position %u, of length %u : saves %u (ratio: %.2f) \n",- (U32)pos, (U32)maxLength, savings[maxLength], (double)savings[maxLength] / maxLength);+ DISPLAYLEVEL(4, "Selected dict at position %u, of length %u : saves %u (ratio: %.2f) \n",+ (unsigned)pos, (unsigned)maxLength, (unsigned)savings[maxLength], (double)savings[maxLength] / maxLength); solution.pos = (U32)pos; solution.length = (U32)maxLength;@@ -346,12 +351,12 @@ /* mark positions done */ { U32 id; for (id=start; id<end; id++) {- U32 p, pEnd;+ U32 p, pEnd, length; U32 const testedPos = suffix[id]; if (testedPos == pos) length = solution.length; else {- length = ZDICT_count(b+pos, b+testedPos);+ length = (U32)ZDICT_count(b+pos, b+testedPos); if (length > solution.length) length = solution.length; } pEnd = (U32)(testedPos + length);@@ -363,21 +368,35 @@ } -/*! ZDICT_checkMerge+static int isIncluded(const void* in, const void* container, size_t length)+{+ const char* const ip = (const char*) in;+ const char* const into = (const char*) container;+ size_t u;++ for (u=0; u<length; u++) { /* works because end of buffer is a noisy guard band */+ if (ip[u] != into[u]) break;+ }++ return u==length;+}++/*! ZDICT_tryMerge() : check if dictItem can be merged, do it if possible @return : id of destination elt, 0 if not merged */-static U32 ZDICT_checkMerge(dictItem* table, dictItem elt, U32 eltNbToSkip)+static U32 ZDICT_tryMerge(dictItem* table, dictItem elt, U32 eltNbToSkip, const void* buffer) { const U32 tableSize = table->pos; const U32 eltEnd = elt.pos + elt.length;+ const char* const buf = (const char*) buffer; /* tail overlap */ U32 u; for (u=1; u<tableSize; u++) { if (u==eltNbToSkip) continue; if ((table[u].pos > elt.pos) && (table[u].pos <= eltEnd)) { /* overlap, existing > new */ /* append */- U32 addedLength = table[u].pos - elt.pos;+ U32 const addedLength = table[u].pos - elt.pos; table[u].length += addedLength; table[u].pos = elt.pos; table[u].savings += elt.savings * addedLength / elt.length; /* rough approx */@@ -393,9 +412,10 @@ /* front overlap */ for (u=1; u<tableSize; u++) { if (u==eltNbToSkip) continue;+ if ((table[u].pos + table[u].length >= elt.pos) && (table[u].pos < elt.pos)) { /* overlap, existing < new */ /* append */- int addedLength = (int)eltEnd - (table[u].pos + table[u].length);+ int const addedLength = (int)eltEnd - (table[u].pos + table[u].length); table[u].savings += elt.length / 8; /* rough approx bonus */ if (addedLength > 0) { /* otherwise, elt fully included into existing */ table[u].length += addedLength;@@ -407,16 +427,27 @@ table[u] = table[u-1], u--; table[u] = elt; return u;- } }+ } + if (MEM_read64(buf + table[u].pos) == MEM_read64(buf + elt.pos + 1)) {+ if (isIncluded(buf + table[u].pos, buf + elt.pos + 1, table[u].length)) {+ size_t const addedLength = MAX( (int)elt.length - (int)table[u].length , 1 );+ table[u].pos = elt.pos;+ table[u].savings += (U32)(elt.savings * addedLength / elt.length);+ table[u].length = MIN(elt.length, table[u].length + 1);+ return u;+ }+ }+ }+ return 0; } static void ZDICT_removeDictItem(dictItem* table, U32 id) {- /* convention : first element is nb of elts */- U32 const max = table->pos;+ /* convention : table[0].pos stores nb of elts */+ U32 const max = table[0].pos; U32 u; if (!id) return; /* protection, should never happen */ for (u=id; u<max-1; u++)@@ -425,14 +456,14 @@ } -static void ZDICT_insertDictItem(dictItem* table, U32 maxSize, dictItem elt)+static void ZDICT_insertDictItem(dictItem* table, U32 maxSize, dictItem elt, const void* buffer) { /* merge if possible */- U32 mergeId = ZDICT_checkMerge(table, elt, 0);+ U32 mergeId = ZDICT_tryMerge(table, elt, 0, buffer); if (mergeId) { U32 newMerge = 1; while (newMerge) {- newMerge = ZDICT_checkMerge(table, table[mergeId], mergeId);+ newMerge = ZDICT_tryMerge(table, table[mergeId], mergeId, buffer); if (newMerge) ZDICT_removeDictItem(table, mergeId); mergeId = newMerge; }@@ -463,10 +494,10 @@ } -static size_t ZDICT_trainBuffer(dictItem* dictList, U32 dictListSize,+static size_t ZDICT_trainBuffer_legacy(dictItem* dictList, U32 dictListSize, const void* const buffer, size_t bufferSize, /* buffer must end with noisy guard band */ const size_t* fileSizes, unsigned nbFiles,- U32 minRatio, U32 notificationLevel)+ unsigned minRatio, U32 notificationLevel) { int* const suffix0 = (int*)malloc((bufferSize+2)*sizeof(*suffix0)); int* const suffix = suffix0+1;@@ -480,7 +511,7 @@ # define DISPLAYUPDATE(l, ...) if (notificationLevel>=l) { \ if (ZDICT_clockSpan(displayClock) > refreshRate) \ { displayClock = clock(); DISPLAY(__VA_ARGS__); \- if (notificationLevel>=4) fflush(stdout); } }+ if (notificationLevel>=4) fflush(stderr); } } /* init */ DISPLAYLEVEL(2, "\r%70s\r", ""); /* clean display line */@@ -492,11 +523,11 @@ memset(doneMarks, 0, bufferSize+16); /* limit sample set size (divsufsort limitation)*/- if (bufferSize > ZDICT_MAX_SAMPLES_SIZE) DISPLAYLEVEL(3, "sample set too large : reduced to %u MB ...\n", (U32)(ZDICT_MAX_SAMPLES_SIZE>>20));+ if (bufferSize > ZDICT_MAX_SAMPLES_SIZE) DISPLAYLEVEL(3, "sample set too large : reduced to %u MB ...\n", (unsigned)(ZDICT_MAX_SAMPLES_SIZE>>20)); while (bufferSize > ZDICT_MAX_SAMPLES_SIZE) bufferSize -= fileSizes[--nbFiles]; /* sort */- DISPLAYLEVEL(2, "sorting %u files of total size %u MB ...\n", nbFiles, (U32)(bufferSize>>20));+ DISPLAYLEVEL(2, "sorting %u files of total size %u MB ...\n", nbFiles, (unsigned)(bufferSize>>20)); { int const divSuftSortResult = divsufsort((const unsigned char*)buffer, suffix, (int)bufferSize, 0); if (divSuftSortResult != 0) { result = ERROR(GENERIC); goto _cleanup; } }@@ -521,7 +552,7 @@ if (doneMarks[cursor]) { cursor++; continue; } solution = ZDICT_analyzePos(doneMarks, suffix, reverseSuffix[cursor], buffer, minRatio, notificationLevel); if (solution.length==0) { cursor++; continue; }- ZDICT_insertDictItem(dictList, dictListSize, solution);+ ZDICT_insertDictItem(dictList, dictListSize, solution, buffer); cursor += solution.length; DISPLAYUPDATE(2, "\r%4.2f %% \r", (double)cursor / bufferSize * 100); } }@@ -550,29 +581,31 @@ typedef struct {- ZSTD_CCtx* ref;- ZSTD_CCtx* zc;- void* workPlace; /* must be ZSTD_BLOCKSIZE_ABSOLUTEMAX allocated */+ ZSTD_CDict* dict; /* dictionary */+ ZSTD_CCtx* zc; /* working context */+ void* workPlace; /* must be ZSTD_BLOCKSIZE_MAX allocated */ } EStats_ress_t; #define MAXREPOFFSET 1024 static void ZDICT_countEStats(EStats_ress_t esr, ZSTD_parameters params,- U32* countLit, U32* offsetcodeCount, U32* matchlengthCount, U32* litlengthCount, U32* repOffsets,- const void* src, size_t srcSize, U32 notificationLevel)+ unsigned* countLit, unsigned* offsetcodeCount, unsigned* matchlengthCount, unsigned* litlengthCount, U32* repOffsets,+ const void* src, size_t srcSize,+ U32 notificationLevel) {- size_t const blockSizeMax = MIN (ZSTD_BLOCKSIZE_ABSOLUTEMAX, 1 << params.cParams.windowLog);+ size_t const blockSizeMax = MIN (ZSTD_BLOCKSIZE_MAX, 1 << params.cParams.windowLog); size_t cSize; if (srcSize > blockSizeMax) srcSize = blockSizeMax; /* protection vs large samples */- { size_t const errorCode = ZSTD_copyCCtx(esr.zc, esr.ref, 0);- if (ZSTD_isError(errorCode)) { DISPLAYLEVEL(1, "warning : ZSTD_copyCCtx failed \n"); return; }+ { size_t const errorCode = ZSTD_compressBegin_usingCDict(esr.zc, esr.dict);+ if (ZSTD_isError(errorCode)) { DISPLAYLEVEL(1, "warning : ZSTD_compressBegin_usingCDict failed \n"); return; }+ }- cSize = ZSTD_compressBlock(esr.zc, esr.workPlace, ZSTD_BLOCKSIZE_ABSOLUTEMAX, src, srcSize);- if (ZSTD_isError(cSize)) { DISPLAYLEVEL(3, "warning : could not compress sample size %u \n", (U32)srcSize); return; }+ cSize = ZSTD_compressBlock(esr.zc, esr.workPlace, ZSTD_BLOCKSIZE_MAX, src, srcSize);+ if (ZSTD_isError(cSize)) { DISPLAYLEVEL(3, "warning : could not compress sample size %u \n", (unsigned)srcSize); return; } if (cSize) { /* if == 0; block is not compressible */- const seqStore_t* seqStorePtr = ZSTD_getSeqStore(esr.zc);+ const seqStore_t* const seqStorePtr = ZSTD_getSeqStore(esr.zc); /* literals stats */ { const BYTE* bytePtr;@@ -610,17 +643,6 @@ } } } } -/*-static size_t ZDICT_maxSampleSize(const size_t* fileSizes, unsigned nbFiles)-{- unsigned u;- size_t max=0;- for (u=0; u<nbFiles; u++)- if (max < fileSizes[u]) max = fileSizes[u];- return max;-}-*/- static size_t ZDICT_totalSampleSize(const size_t* fileSizes, unsigned nbFiles) { size_t total=0;@@ -645,6 +667,18 @@ } } +/* ZDICT_flatLit() :+ * rewrite `countLit` to contain a mostly flat but still compressible distribution of literals.+ * necessary to avoid generating a non-compressible distribution that HUF_writeCTable() cannot encode.+ */+static void ZDICT_flatLit(unsigned* countLit)+{+ int u;+ for (u=1; u<256; u++) countLit[u] = 2;+ countLit[0] = 4;+ countLit[253] = 1;+ countLit[254] = 1;+} #define OFFCODE_MAX 30 /* only applicable to first block */ static size_t ZDICT_analyzeEntropy(void* dstBuffer, size_t maxDstSize,@@ -653,18 +687,18 @@ const void* dictBuffer, size_t dictBufferSize, unsigned notificationLevel) {- U32 countLit[256];+ unsigned countLit[256]; HUF_CREATE_STATIC_CTABLE(hufTable, 255);- U32 offcodeCount[OFFCODE_MAX+1];+ unsigned offcodeCount[OFFCODE_MAX+1]; short offcodeNCount[OFFCODE_MAX+1]; U32 offcodeMax = ZSTD_highbit32((U32)(dictBufferSize + 128 KB));- U32 matchLengthCount[MaxML+1];+ unsigned matchLengthCount[MaxML+1]; short matchLengthNCount[MaxML+1];- U32 litLengthCount[MaxLL+1];+ unsigned litLengthCount[MaxLL+1]; short litLengthNCount[MaxLL+1]; U32 repOffset[MAXREPOFFSET]; offsetCount_t bestRepOffset[ZSTD_REP_NUM+1];- EStats_ress_t esr;+ EStats_ress_t esr = { NULL, NULL, NULL }; ZSTD_parameters params; U32 u, huffLog = 11, Offlog = OffFSELog, mlLog = MLFSELog, llLog = LLFSELog, total; size_t pos = 0, errorCode;@@ -674,32 +708,28 @@ BYTE* dstPtr = (BYTE*)dstBuffer; /* init */- esr.ref = ZSTD_createCCtx();+ DEBUGLOG(4, "ZDICT_analyzeEntropy");+ if (offcodeMax>OFFCODE_MAX) { eSize = ERROR(dictionaryCreation_failed); goto _cleanup; } /* too large dictionary */+ for (u=0; u<256; u++) countLit[u] = 1; /* any character must be described */+ for (u=0; u<=offcodeMax; u++) offcodeCount[u] = 1;+ for (u=0; u<=MaxML; u++) matchLengthCount[u] = 1;+ for (u=0; u<=MaxLL; u++) litLengthCount[u] = 1;+ memset(repOffset, 0, sizeof(repOffset));+ repOffset[1] = repOffset[4] = repOffset[8] = 1;+ memset(bestRepOffset, 0, sizeof(bestRepOffset));+ if (compressionLevel==0) compressionLevel = g_compressionLevel_default;+ params = ZSTD_getParams(compressionLevel, averageSampleSize, dictBufferSize);++ esr.dict = ZSTD_createCDict_advanced(dictBuffer, dictBufferSize, ZSTD_dlm_byRef, ZSTD_dct_rawContent, params.cParams, ZSTD_defaultCMem); esr.zc = ZSTD_createCCtx();- esr.workPlace = malloc(ZSTD_BLOCKSIZE_ABSOLUTEMAX);- if (!esr.ref || !esr.zc || !esr.workPlace) {+ esr.workPlace = malloc(ZSTD_BLOCKSIZE_MAX);+ if (!esr.dict || !esr.zc || !esr.workPlace) { eSize = ERROR(memory_allocation); DISPLAYLEVEL(1, "Not enough memory \n"); goto _cleanup; }- if (offcodeMax>OFFCODE_MAX) { eSize = ERROR(dictionary_wrong); goto _cleanup; } /* too large dictionary */- for (u=0; u<256; u++) countLit[u]=1; /* any character must be described */- for (u=0; u<=offcodeMax; u++) offcodeCount[u]=1;- for (u=0; u<=MaxML; u++) matchLengthCount[u]=1;- for (u=0; u<=MaxLL; u++) litLengthCount[u]=1;- memset(repOffset, 0, sizeof(repOffset));- repOffset[1] = repOffset[4] = repOffset[8] = 1;- memset(bestRepOffset, 0, sizeof(bestRepOffset));- if (compressionLevel==0) compressionLevel=g_compressionLevel_default;- params = ZSTD_getParams(compressionLevel, averageSampleSize, dictBufferSize);- { size_t const beginResult = ZSTD_compressBegin_advanced(esr.ref, dictBuffer, dictBufferSize, params, 0);- if (ZSTD_isError(beginResult)) {- eSize = ERROR(GENERIC);- DISPLAYLEVEL(1, "error : ZSTD_compressBegin_advanced failed \n");- goto _cleanup;- } } - /* collect stats on all files */+ /* collect stats on all samples */ for (u=0; u<nbFiles; u++) { ZDICT_countEStats(esr, params, countLit, offcodeCount, matchLengthCount, litLengthCount, repOffset,@@ -708,14 +738,21 @@ pos += fileSizes[u]; } - /* analyze */- errorCode = HUF_buildCTable (hufTable, countLit, 255, huffLog);- if (HUF_isError(errorCode)) {- eSize = ERROR(GENERIC);- DISPLAYLEVEL(1, "HUF_buildCTable error \n");- goto _cleanup;+ /* analyze, build stats, starting with literals */+ { size_t maxNbBits = HUF_buildCTable (hufTable, countLit, 255, huffLog);+ if (HUF_isError(maxNbBits)) {+ eSize = ERROR(GENERIC);+ DISPLAYLEVEL(1, " HUF_buildCTable error \n");+ goto _cleanup;+ }+ if (maxNbBits==8) { /* not compressible : will fail on HUF_writeCTable() */+ DISPLAYLEVEL(2, "warning : pathological dataset : literals are not compressible : samples are noisy or too regular \n");+ ZDICT_flatLit(countLit); /* replace distribution by a fake "mostly flat but still compressible" distribution, that HUF_writeCTable() can encode */+ maxNbBits = HUF_buildCTable (hufTable, countLit, 255, huffLog);+ assert(maxNbBits==9);+ }+ huffLog = (U32)maxNbBits; }- huffLog = (U32)errorCode; /* looking for most common first offsets */ { U32 offset;@@ -812,11 +849,10 @@ MEM_writeLE32(dstPtr+4, repStartValue[1]); MEM_writeLE32(dstPtr+8, repStartValue[2]); #endif- //dstPtr += 12; eSize += 12; _cleanup:- ZSTD_freeCCtx(esr.ref);+ ZSTD_freeCDict(esr.dict); ZSTD_freeCCtx(esr.zc); free(esr.workPlace); @@ -827,22 +863,23 @@ size_t ZDICT_finalizeDictionary(void* dictBuffer, size_t dictBufferCapacity, const void* customDictContent, size_t dictContentSize,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,- ZDICT_params_t params)+ const void* samplesBuffer, const size_t* samplesSizes,+ unsigned nbSamples, ZDICT_params_t params) { size_t hSize;-#define HBUFFSIZE 256+#define HBUFFSIZE 256 /* should prove large enough for all entropy headers */ BYTE header[HBUFFSIZE];- int const compressionLevel = (params.compressionLevel <= 0) ? g_compressionLevel_default : params.compressionLevel;+ int const compressionLevel = (params.compressionLevel == 0) ? g_compressionLevel_default : params.compressionLevel; U32 const notificationLevel = params.notificationLevel; /* check conditions */+ DEBUGLOG(4, "ZDICT_finalizeDictionary"); if (dictBufferCapacity < dictContentSize) return ERROR(dstSize_tooSmall); if (dictContentSize < ZDICT_CONTENTSIZE_MIN) return ERROR(srcSize_wrong); if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) return ERROR(dstSize_tooSmall); /* dictionary header */- MEM_writeLE32(header, ZSTD_DICT_MAGIC);+ MEM_writeLE32(header, ZSTD_MAGIC_DICTIONARY); { U64 const randomID = XXH64(customDictContent, dictContentSize, 0); U32 const compliantID = (randomID % ((1U<<31)-32768)) + 32768; U32 const dictID = params.dictID ? params.dictID : compliantID;@@ -873,24 +910,16 @@ } -size_t ZDICT_addEntropyTablesFromBuffer_advanced(void* dictBuffer, size_t dictContentSize, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,- ZDICT_params_t params)+static size_t ZDICT_addEntropyTablesFromBuffer_advanced(+ void* dictBuffer, size_t dictContentSize, size_t dictBufferCapacity,+ const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_params_t params) {- size_t hSize;- int const compressionLevel = (params.compressionLevel <= 0) ? g_compressionLevel_default : params.compressionLevel;+ int const compressionLevel = (params.compressionLevel == 0) ? g_compressionLevel_default : params.compressionLevel; U32 const notificationLevel = params.notificationLevel;-- /* dictionary header */- MEM_writeLE32(dictBuffer, ZSTD_DICT_MAGIC);- { U64 const randomID = XXH64((char*)dictBuffer + dictBufferCapacity - dictContentSize, dictContentSize, 0);- U32 const compliantID = (randomID % ((1U<<31)-32768)) + 32768;- U32 const dictID = params.dictID ? params.dictID : compliantID;- MEM_writeLE32((char*)dictBuffer+4, dictID);- }- hSize = 8;+ size_t hSize = 8; - /* entropy tables */+ /* calculate entropy tables */ DISPLAYLEVEL(2, "\r%70s\r", ""); /* clean display line */ DISPLAYLEVEL(2, "statistics ... \n"); { size_t const eSize = ZDICT_analyzeEntropy((char*)dictBuffer+hSize, dictBufferCapacity-hSize,@@ -902,21 +931,32 @@ hSize += eSize; } + /* add dictionary header (after entropy tables) */+ MEM_writeLE32(dictBuffer, ZSTD_MAGIC_DICTIONARY);+ { U64 const randomID = XXH64((char*)dictBuffer + dictBufferCapacity - dictContentSize, dictContentSize, 0);+ U32 const compliantID = (randomID % ((1U<<31)-32768)) + 32768;+ U32 const dictID = params.dictID ? params.dictID : compliantID;+ MEM_writeLE32((char*)dictBuffer+4, dictID);+ } if (hSize + dictContentSize < dictBufferCapacity) memmove((char*)dictBuffer + hSize, (char*)dictBuffer + dictBufferCapacity - dictContentSize, dictContentSize); return MIN(dictBufferCapacity, hSize+dictContentSize); } --/*! ZDICT_trainFromBuffer_unsafe() :+/* Hidden declaration for dbio.c */+size_t ZDICT_trainFromBuffer_unsafe_legacy(+ void* dictBuffer, size_t maxDictSize,+ const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_legacy_params_t params);+/*! ZDICT_trainFromBuffer_unsafe_legacy() : * Warning : `samplesBuffer` must be followed by noisy guard band. * @return : size of dictionary, or an error code which can be tested with ZDICT_isError() */-size_t ZDICT_trainFromBuffer_unsafe(+size_t ZDICT_trainFromBuffer_unsafe_legacy( void* dictBuffer, size_t maxDictSize, const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,- ZDICT_params_t params)+ ZDICT_legacy_params_t params) { U32 const dictListSize = MAX(MAX(DICTLISTSIZE_DEFAULT, nbSamples), (U32)(maxDictSize/16)); dictItem* const dictList = (dictItem*)malloc(dictListSize * sizeof(*dictList));@@ -925,60 +965,63 @@ size_t const targetDictSize = maxDictSize; size_t const samplesBuffSize = ZDICT_totalSampleSize(samplesSizes, nbSamples); size_t dictSize = 0;- U32 const notificationLevel = params.notificationLevel;+ U32 const notificationLevel = params.zParams.notificationLevel; /* checks */ if (!dictList) return ERROR(memory_allocation);- if (maxDictSize <= g_provision_entropySize + g_min_fast_dictContent) { free(dictList); return ERROR(dstSize_tooSmall); }- if (samplesBuffSize < ZDICT_MIN_SAMPLES_SIZE) { free(dictList); return 0; } /* not enough source to create dictionary */+ if (maxDictSize < ZDICT_DICTSIZE_MIN) { free(dictList); return ERROR(dstSize_tooSmall); } /* requested dictionary size is too small */+ if (samplesBuffSize < ZDICT_MIN_SAMPLES_SIZE) { free(dictList); return ERROR(dictionaryCreation_failed); } /* not enough source to create dictionary */ /* init */ ZDICT_initDictItem(dictList); /* build dictionary */- ZDICT_trainBuffer(dictList, dictListSize,- samplesBuffer, samplesBuffSize,- samplesSizes, nbSamples,- minRep, notificationLevel);+ ZDICT_trainBuffer_legacy(dictList, dictListSize,+ samplesBuffer, samplesBuffSize,+ samplesSizes, nbSamples,+ minRep, notificationLevel); /* display best matches */- if (params.notificationLevel>= 3) {- U32 const nb = MIN(25, dictList[0].pos);- U32 const dictContentSize = ZDICT_dictSize(dictList);- U32 u;- DISPLAYLEVEL(3, "\n %u segments found, of total size %u \n", dictList[0].pos-1, dictContentSize);+ if (params.zParams.notificationLevel>= 3) {+ unsigned const nb = MIN(25, dictList[0].pos);+ unsigned const dictContentSize = ZDICT_dictSize(dictList);+ unsigned u;+ DISPLAYLEVEL(3, "\n %u segments found, of total size %u \n", (unsigned)dictList[0].pos-1, dictContentSize); DISPLAYLEVEL(3, "list %u best segments \n", nb-1); for (u=1; u<nb; u++) {- U32 const pos = dictList[u].pos;- U32 const length = dictList[u].length;+ unsigned const pos = dictList[u].pos;+ unsigned const length = dictList[u].length; U32 const printedLength = MIN(40, length);- if ((pos > samplesBuffSize) || ((pos + length) > samplesBuffSize))+ if ((pos > samplesBuffSize) || ((pos + length) > samplesBuffSize)) {+ free(dictList); return ERROR(GENERIC); /* should never happen */+ } DISPLAYLEVEL(3, "%3u:%3u bytes at pos %8u, savings %7u bytes |",- u, length, pos, dictList[u].savings);+ u, length, pos, (unsigned)dictList[u].savings); ZDICT_printHex((const char*)samplesBuffer+pos, printedLength); DISPLAYLEVEL(3, "| \n"); } } /* create dictionary */- { U32 dictContentSize = ZDICT_dictSize(dictList);- if (dictContentSize < targetDictSize/3) {- DISPLAYLEVEL(2, "! warning : selected content significantly smaller than requested (%u < %u) \n", dictContentSize, (U32)maxDictSize);+ { unsigned dictContentSize = ZDICT_dictSize(dictList);+ if (dictContentSize < ZDICT_CONTENTSIZE_MIN) { free(dictList); return ERROR(dictionaryCreation_failed); } /* dictionary content too small */+ if (dictContentSize < targetDictSize/4) {+ DISPLAYLEVEL(2, "! warning : selected content significantly smaller than requested (%u < %u) \n", dictContentSize, (unsigned)maxDictSize);+ if (samplesBuffSize < 10 * targetDictSize)+ DISPLAYLEVEL(2, "! consider increasing the number of samples (total size : %u MB)\n", (unsigned)(samplesBuffSize>>20)); if (minRep > MINRATIO) { DISPLAYLEVEL(2, "! consider increasing selectivity to produce larger dictionary (-s%u) \n", selectivity+1); DISPLAYLEVEL(2, "! note : larger dictionaries are not necessarily better, test its efficiency on samples \n"); }- if (samplesBuffSize < 10 * targetDictSize)- DISPLAYLEVEL(2, "! consider increasing the number of samples (total size : %u MB)\n", (U32)(samplesBuffSize>>20)); } if ((dictContentSize > targetDictSize*3) && (nbSamples > 2*MINRATIO) && (selectivity>1)) {- U32 proposedSelectivity = selectivity-1;+ unsigned proposedSelectivity = selectivity-1; while ((nbSamples >> proposedSelectivity) <= MINRATIO) { proposedSelectivity--; }- DISPLAYLEVEL(2, "! note : calculated dictionary significantly larger than requested (%u > %u) \n", dictContentSize, (U32)maxDictSize);+ DISPLAYLEVEL(2, "! note : calculated dictionary significantly larger than requested (%u > %u) \n", dictContentSize, (unsigned)maxDictSize); DISPLAYLEVEL(2, "! consider increasing dictionary size, or produce denser dictionary (-s%u) \n", proposedSelectivity);- DISPLAYLEVEL(2, "! always test dictionary efficiency on samples \n");+ DISPLAYLEVEL(2, "! always test dictionary efficiency on real samples \n"); } /* limit dictionary size */@@ -1004,7 +1047,7 @@ dictSize = ZDICT_addEntropyTablesFromBuffer_advanced(dictBuffer, dictContentSize, maxDictSize, samplesBuffer, samplesSizes, nbSamples,- params);+ params.zParams); } /* clean up */@@ -1013,11 +1056,12 @@ } -/* issue : samplesBuffer need to be followed by a noisy guard band.-* work around : duplicate the buffer, and add the noise */-size_t ZDICT_trainFromBuffer_advanced(void* dictBuffer, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,- ZDICT_params_t params)+/* ZDICT_trainFromBuffer_legacy() :+ * issue : samplesBuffer need to be followed by a noisy guard band.+ * work around : duplicate the buffer, and add the noise */+size_t ZDICT_trainFromBuffer_legacy(void* dictBuffer, size_t dictBufferCapacity,+ const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_legacy_params_t params) { size_t result; void* newBuff;@@ -1030,10 +1074,9 @@ memcpy(newBuff, samplesBuffer, sBuffSize); ZDICT_fillNoise((char*)newBuff + sBuffSize, NOISELENGTH); /* guard band, for end of buffer condition */ - result = ZDICT_trainFromBuffer_unsafe(- dictBuffer, dictBufferCapacity,- newBuff, samplesSizes, nbSamples,- params);+ result =+ ZDICT_trainFromBuffer_unsafe_legacy(dictBuffer, dictBufferCapacity, newBuff,+ samplesSizes, nbSamples, params); free(newBuff); return result; }@@ -1042,15 +1085,23 @@ size_t ZDICT_trainFromBuffer(void* dictBuffer, size_t dictBufferCapacity, const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples) {- ZDICT_params_t params;+ ZDICT_fastCover_params_t params;+ DEBUGLOG(3, "ZDICT_trainFromBuffer"); memset(¶ms, 0, sizeof(params));- return ZDICT_trainFromBuffer_advanced(dictBuffer, dictBufferCapacity,- samplesBuffer, samplesSizes, nbSamples,- params);+ params.d = 8;+ params.steps = 4;+ /* Default to level 6 since no compression level information is available */+ params.zParams.compressionLevel = 3;+#if defined(DEBUGLEVEL) && (DEBUGLEVEL>=1)+ params.zParams.notificationLevel = DEBUGLEVEL;+#endif+ return ZDICT_optimizeTrainFromBuffer_fastCover(dictBuffer, dictBufferCapacity,+ samplesBuffer, samplesSizes, nbSamples,+ ¶ms); } size_t ZDICT_addEntropyTablesFromBuffer(void* dictBuffer, size_t dictContentSize, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples)+ const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples) { ZDICT_params_t params; memset(¶ms, 0, sizeof(params));
zstd/lib/dictBuilder/zdict.h view
@@ -1,10 +1,11 @@-/**+/* * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses. */ #ifndef DICTBUILDER_H_001@@ -20,10 +21,12 @@ /* ===== ZDICTLIB_API : control library symbols visibility ===== */-#if defined(__GNUC__) && (__GNUC__ >= 4)-# define ZDICTLIB_VISIBILITY __attribute__ ((visibility ("default")))-#else-# define ZDICTLIB_VISIBILITY+#ifndef ZDICTLIB_VISIBILITY+# if defined(__GNUC__) && (__GNUC__ >= 4)+# define ZDICTLIB_VISIBILITY __attribute__ ((visibility ("default")))+# else+# define ZDICTLIB_VISIBILITY+# endif #endif #if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1) # define ZDICTLIB_API __declspec(dllexport) ZDICTLIB_VISIBILITY@@ -34,20 +37,29 @@ #endif -/*! ZDICT_trainFromBuffer() :- Train a dictionary from an array of samples.- Samples must be stored concatenated in a single flat buffer `samplesBuffer`,- supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.- The resulting dictionary will be saved into `dictBuffer`.- @return : size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)- or an error code, which can be tested with ZDICT_isError().- Tips : In general, a reasonable dictionary has a size of ~ 100 KB.- It's obviously possible to target smaller or larger ones, just by specifying different `dictBufferCapacity`.- In general, it's recommended to provide a few thousands samples, but this can vary a lot.- It's recommended that total size of all samples be about ~x100 times the target size of dictionary.-*/+/*! ZDICT_trainFromBuffer():+ * Train a dictionary from an array of samples.+ * Redirect towards ZDICT_optimizeTrainFromBuffer_fastCover() single-threaded, with d=8, steps=4,+ * f=20, and accel=1.+ * Samples must be stored concatenated in a single flat buffer `samplesBuffer`,+ * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.+ * The resulting dictionary will be saved into `dictBuffer`.+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)+ * or an error code, which can be tested with ZDICT_isError().+ * Note: Dictionary training will fail if there are not enough samples to construct a+ * dictionary, or if most of the samples are too small (< 8 bytes being the lower limit).+ * If dictionary training fails, you should use zstd without a dictionary, as the dictionary+ * would've been ineffective anyways. If you believe your samples would benefit from a dictionary+ * please open an issue with details, and we can look into it.+ * Note: ZDICT_trainFromBuffer()'s memory usage is about 6 MB.+ * Tips: In general, a reasonable dictionary has a size of ~ 100 KB.+ * It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.+ * In general, it's recommended to provide a few thousands samples, though this can vary a lot.+ * It's recommended that total size of all samples be about ~x100 times the target size of dictionary.+ */ ZDICTLIB_API size_t ZDICT_trainFromBuffer(void* dictBuffer, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples);+ const void* samplesBuffer,+ const size_t* samplesSizes, unsigned nbSamples); /*====== Helper functions ======*/@@ -67,99 +79,166 @@ * ==================================================================================== */ typedef struct {- unsigned selectivityLevel; /* 0 means default; larger => select more => larger dictionary */- int compressionLevel; /* 0 means default; target a specific zstd compression level */- unsigned notificationLevel; /* Write to stderr; 0 = none (default); 1 = errors; 2 = progression; 3 = details; 4 = debug; */- unsigned dictID; /* 0 means auto mode (32-bits random value); other : force dictID value */- unsigned reserved[2]; /* reserved space for future parameters */+ int compressionLevel; /* optimize for a specific zstd compression level; 0 means default */+ unsigned notificationLevel; /* Write log to stderr; 0 = none (default); 1 = errors; 2 = progression; 3 = details; 4 = debug; */+ unsigned dictID; /* force dictID value; 0 means auto mode (32-bits random value) */ } ZDICT_params_t; --/*! ZDICT_trainFromBuffer_advanced() :- Same as ZDICT_trainFromBuffer() with control over more parameters.- `parameters` is optional and can be provided with values set to 0 to mean "default".- @return : size of dictionary stored into `dictBuffer` (<= `dictBufferSize`),- or an error code, which can be tested by ZDICT_isError().- note : ZDICT_trainFromBuffer_advanced() will send notifications into stderr if instructed to, using notificationLevel>0.-*/-ZDICTLIB_API size_t ZDICT_trainFromBuffer_advanced(void* dictBuffer, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,- ZDICT_params_t parameters);--/*! COVER_params_t :- For all values 0 means default.- kMin and d are the only required parameters.-*/+/*! ZDICT_cover_params_t:+ * k and d are the only required parameters.+ * For others, value 0 means default.+ */ typedef struct { unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */ unsigned d; /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */- unsigned steps; /* Number of steps : Only used for optimization : 0 means default (256) : Higher means more parameters checked */-- unsigned notificationLevel; /* Write to stderr; 0 = none (default); 1 = errors; 2 = progression; 3 = details; 4 = debug; */- unsigned dictID; /* 0 means auto mode (32-bits random value); other : force dictID value */- int compressionLevel; /* 0 means default; target a specific zstd compression level */-} COVER_params_t;---/*! COVER_trainFromBuffer() :- Train a dictionary from an array of samples using the COVER algorithm.- Samples must be stored concatenated in a single flat buffer `samplesBuffer`,- supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.- The resulting dictionary will be saved into `dictBuffer`.- @return : size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)- or an error code, which can be tested with ZDICT_isError().- Tips : In general, a reasonable dictionary has a size of ~ 100 KB.- It's obviously possible to target smaller or larger ones, just by specifying different `dictBufferCapacity`.- In general, it's recommended to provide a few thousands samples, but this can vary a lot.- It's recommended that total size of all samples be about ~x100 times the target size of dictionary.-*/-ZDICTLIB_API size_t COVER_trainFromBuffer(void* dictBuffer, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,- COVER_params_t parameters);--/*! COVER_optimizeTrainFromBuffer() :- The same requirements as above hold for all the parameters except `parameters`.- This function tries many parameter combinations and picks the best parameters.- `*parameters` is filled with the best parameters found, and the dictionary- constructed with those parameters is stored in `dictBuffer`.-- All of the parameters d, k, steps are optional.- If d is non-zero then we don't check multiple values of d, otherwise we check d = {6, 8, 10, 12, 14, 16}.- if steps is zero it defaults to its default value.- If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [16, 2048].+ unsigned steps; /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */+ unsigned nbThreads; /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */+ double splitPoint; /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (1.0), 1.0 when all samples are used for both training and testing */+ ZDICT_params_t zParams;+} ZDICT_cover_params_t; - @return : size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)- or an error code, which can be tested with ZDICT_isError().- On success `*parameters` contains the parameters selected.-*/-ZDICTLIB_API size_t COVER_optimizeTrainFromBuffer(void* dictBuffer, size_t dictBufferCapacity,- const void* samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,- COVER_params_t *parameters);+typedef struct {+ unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */+ unsigned d; /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */+ unsigned f; /* log of size of frequency array : constraint: 0 < f <= 31 : 1 means default(20)*/+ unsigned steps; /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */+ unsigned nbThreads; /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */+ double splitPoint; /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (0.75), 1.0 when all samples are used for both training and testing */+ unsigned accel; /* Acceleration level: constraint: 0 < accel <= 10, higher means faster and less accurate, 0 means default(1) */+ ZDICT_params_t zParams;+} ZDICT_fastCover_params_t; -/*! ZDICT_finalizeDictionary() :+/*! ZDICT_trainFromBuffer_cover():+ * Train a dictionary from an array of samples using the COVER algorithm.+ * Samples must be stored concatenated in a single flat buffer `samplesBuffer`,+ * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.+ * The resulting dictionary will be saved into `dictBuffer`.+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)+ * or an error code, which can be tested with ZDICT_isError().+ * See ZDICT_trainFromBuffer() for details on failure modes.+ * Note: ZDICT_trainFromBuffer_cover() requires about 9 bytes of memory for each input byte.+ * Tips: In general, a reasonable dictionary has a size of ~ 100 KB.+ * It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.+ * In general, it's recommended to provide a few thousands samples, though this can vary a lot.+ * It's recommended that total size of all samples be about ~x100 times the target size of dictionary.+ */+ZDICTLIB_API size_t ZDICT_trainFromBuffer_cover(+ void *dictBuffer, size_t dictBufferCapacity,+ const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,+ ZDICT_cover_params_t parameters); - Given a custom content as a basis for dictionary, and a set of samples,- finalize dictionary by adding headers and statistics.+/*! ZDICT_optimizeTrainFromBuffer_cover():+ * The same requirements as above hold for all the parameters except `parameters`.+ * This function tries many parameter combinations and picks the best parameters.+ * `*parameters` is filled with the best parameters found,+ * dictionary constructed with those parameters is stored in `dictBuffer`.+ *+ * All of the parameters d, k, steps are optional.+ * If d is non-zero then we don't check multiple values of d, otherwise we check d = {6, 8}.+ * if steps is zero it defaults to its default value.+ * If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [50, 2000].+ *+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)+ * or an error code, which can be tested with ZDICT_isError().+ * On success `*parameters` contains the parameters selected.+ * See ZDICT_trainFromBuffer() for details on failure modes.+ * Note: ZDICT_optimizeTrainFromBuffer_cover() requires about 8 bytes of memory for each input byte and additionally another 5 bytes of memory for each byte of memory for each thread.+ */+ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(+ void* dictBuffer, size_t dictBufferCapacity,+ const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_cover_params_t* parameters); - Samples must be stored concatenated in a flat buffer `samplesBuffer`,- supplied with an array of sizes `samplesSizes`, providing the size of each sample in order.+/*! ZDICT_trainFromBuffer_fastCover():+ * Train a dictionary from an array of samples using a modified version of COVER algorithm.+ * Samples must be stored concatenated in a single flat buffer `samplesBuffer`,+ * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.+ * d and k are required.+ * All other parameters are optional, will use default values if not provided+ * The resulting dictionary will be saved into `dictBuffer`.+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)+ * or an error code, which can be tested with ZDICT_isError().+ * See ZDICT_trainFromBuffer() for details on failure modes.+ * Note: ZDICT_trainFromBuffer_fastCover() requires 6 * 2^f bytes of memory.+ * Tips: In general, a reasonable dictionary has a size of ~ 100 KB.+ * It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.+ * In general, it's recommended to provide a few thousands samples, though this can vary a lot.+ * It's recommended that total size of all samples be about ~x100 times the target size of dictionary.+ */+ZDICTLIB_API size_t ZDICT_trainFromBuffer_fastCover(void *dictBuffer,+ size_t dictBufferCapacity, const void *samplesBuffer,+ const size_t *samplesSizes, unsigned nbSamples,+ ZDICT_fastCover_params_t parameters); - dictContentSize must be > ZDICT_CONTENTSIZE_MIN bytes.- maxDictSize must be >= dictContentSize, and must be > ZDICT_DICTSIZE_MIN bytes.+/*! ZDICT_optimizeTrainFromBuffer_fastCover():+ * The same requirements as above hold for all the parameters except `parameters`.+ * This function tries many parameter combinations (specifically, k and d combinations)+ * and picks the best parameters. `*parameters` is filled with the best parameters found,+ * dictionary constructed with those parameters is stored in `dictBuffer`.+ * All of the parameters d, k, steps, f, and accel are optional.+ * If d is non-zero then we don't check multiple values of d, otherwise we check d = {6, 8}.+ * if steps is zero it defaults to its default value.+ * If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [50, 2000].+ * If f is zero, default value of 20 is used.+ * If accel is zero, default value of 1 is used.+ *+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)+ * or an error code, which can be tested with ZDICT_isError().+ * On success `*parameters` contains the parameters selected.+ * See ZDICT_trainFromBuffer() for details on failure modes.+ * Note: ZDICT_optimizeTrainFromBuffer_fastCover() requires about 6 * 2^f bytes of memory for each thread.+ */+ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(void* dictBuffer,+ size_t dictBufferCapacity, const void* samplesBuffer,+ const size_t* samplesSizes, unsigned nbSamples,+ ZDICT_fastCover_params_t* parameters); - @return : size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`),- or an error code, which can be tested by ZDICT_isError().- note : ZDICT_finalizeDictionary() will push notifications into stderr if instructed to, using notificationLevel>0.- note 2 : dictBuffer and customDictContent can overlap-*/-#define ZDICT_CONTENTSIZE_MIN 256-#define ZDICT_DICTSIZE_MIN 512+/*! ZDICT_finalizeDictionary():+ * Given a custom content as a basis for dictionary, and a set of samples,+ * finalize dictionary by adding headers and statistics.+ *+ * Samples must be stored concatenated in a flat buffer `samplesBuffer`,+ * supplied with an array of sizes `samplesSizes`, providing the size of each sample in order.+ *+ * dictContentSize must be >= ZDICT_CONTENTSIZE_MIN bytes.+ * maxDictSize must be >= dictContentSize, and must be >= ZDICT_DICTSIZE_MIN bytes.+ *+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`),+ * or an error code, which can be tested by ZDICT_isError().+ * Note: ZDICT_finalizeDictionary() will push notifications into stderr if instructed to, using notificationLevel>0.+ * Note 2: dictBuffer and dictContent can overlap+ */+#define ZDICT_CONTENTSIZE_MIN 128+#define ZDICT_DICTSIZE_MIN 256 ZDICTLIB_API size_t ZDICT_finalizeDictionary(void* dictBuffer, size_t dictBufferCapacity,- const void* customDictContent, size_t dictContentSize,+ const void* dictContent, size_t dictContentSize, const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples, ZDICT_params_t parameters); +typedef struct {+ unsigned selectivityLevel; /* 0 means default; larger => select more => larger dictionary */+ ZDICT_params_t zParams;+} ZDICT_legacy_params_t; +/*! ZDICT_trainFromBuffer_legacy():+ * Train a dictionary from an array of samples.+ * Samples must be stored concatenated in a single flat buffer `samplesBuffer`,+ * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.+ * The resulting dictionary will be saved into `dictBuffer`.+ * `parameters` is optional and can be provided with values set to 0 to mean "default".+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)+ * or an error code, which can be tested with ZDICT_isError().+ * See ZDICT_trainFromBuffer() for details on failure modes.+ * Tips: In general, a reasonable dictionary has a size of ~ 100 KB.+ * It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.+ * In general, it's recommended to provide a few thousands samples, though this can vary a lot.+ * It's recommended that total size of all samples be about ~x100 times the target size of dictionary.+ * Note: ZDICT_trainFromBuffer_legacy() will send notifications into stderr if instructed to, using notificationLevel>0.+ */+ZDICTLIB_API size_t ZDICT_trainFromBuffer_legacy(+ void *dictBuffer, size_t dictBufferCapacity,+ const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,+ ZDICT_legacy_params_t parameters); /* Deprecation warnings */ /* It is generally possible to disable deprecation warnings from compiler,@@ -171,7 +250,7 @@ #else # define ZDICT_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) # if defined (__cplusplus) && (__cplusplus >= 201402) /* C++14 or greater */-# define ZDICT_DEPRECATED(message) ZDICTLIB_API [[deprecated(message)]]+# define ZDICT_DEPRECATED(message) [[deprecated(message)]] ZDICTLIB_API # elif (ZDICT_GCC_VERSION >= 405) || defined(__clang__) # define ZDICT_DEPRECATED(message) ZDICTLIB_API __attribute__((deprecated(message))) # elif (ZDICT_GCC_VERSION >= 301)
zstd/lib/zstd.h view
@@ -2,687 +2,1908 @@ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * All rights reserved. *- * This source code is licensed under the BSD-style license found in the- * LICENSE file in the root directory of this source tree. An additional grant- * of patent rights can be found in the PATENTS file in the same directory.- */--#if defined (__cplusplus)-extern "C" {-#endif--#ifndef ZSTD_H_235446-#define ZSTD_H_235446--/* ====== Dependency ======*/-#include <stddef.h> /* size_t */---/* ===== ZSTDLIB_API : control library symbols visibility ===== */-#if defined(__GNUC__) && (__GNUC__ >= 4)-# define ZSTDLIB_VISIBILITY __attribute__ ((visibility ("default")))-#else-# define ZSTDLIB_VISIBILITY-#endif-#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)-# define ZSTDLIB_API __declspec(dllexport) ZSTDLIB_VISIBILITY-#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)-# define ZSTDLIB_API __declspec(dllimport) ZSTDLIB_VISIBILITY /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/-#else-# define ZSTDLIB_API ZSTDLIB_VISIBILITY-#endif---/*******************************************************************************************************- Introduction-- zstd, short for Zstandard, is a fast lossless compression algorithm, targeting real-time compression scenarios- at zlib-level and better compression ratios. The zstd compression library provides in-memory compression and- decompression functions. The library supports compression levels from 1 up to ZSTD_maxCLevel() which is 22.- Levels >= 20, labelled `--ultra`, should be used with caution, as they require more memory.- Compression can be done in:- - a single step (described as Simple API)- - a single step, reusing a context (described as Explicit memory management)- - unbounded multiple steps (described as Streaming compression)- The compression ratio achievable on small data can be highly improved using compression with a dictionary in:- - a single step (described as Simple dictionary API)- - a single step, reusing a dictionary (described as Fast dictionary API)-- Advanced experimental functions can be accessed using #define ZSTD_STATIC_LINKING_ONLY before including zstd.h.- These APIs shall never be used with a dynamic library.- They are not "stable", their definition may change in the future. Only static linking is allowed.-*********************************************************************************************************/--/*------ Version ------*/-#define ZSTD_VERSION_MAJOR 1-#define ZSTD_VERSION_MINOR 1-#define ZSTD_VERSION_RELEASE 3--#define ZSTD_LIB_VERSION ZSTD_VERSION_MAJOR.ZSTD_VERSION_MINOR.ZSTD_VERSION_RELEASE-#define ZSTD_QUOTE(str) #str-#define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str)-#define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION)--#define ZSTD_VERSION_NUMBER (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)-ZSTDLIB_API unsigned ZSTD_versionNumber(void); /**< library version number; to be used when checking dll version */---/***************************************-* Simple API-***************************************/-/*! ZSTD_compress() :- Compresses `src` content as a single zstd compressed frame into already allocated `dst`.- Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.- @return : compressed size written into `dst` (<= `dstCapacity),- or an error code if it fails (which can be tested using ZSTD_isError()). */-ZSTDLIB_API size_t ZSTD_compress( void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- int compressionLevel);--/*! ZSTD_decompress() :- `compressedSize` : must be the _exact_ size of a single compressed frame.- `dstCapacity` is an upper bound of originalSize.- If user cannot imply a maximum upper bound, it's better to use streaming mode to decompress data.- @return : the number of bytes decompressed into `dst` (<= `dstCapacity`),- or an errorCode if it fails (which can be tested using ZSTD_isError()). */-ZSTDLIB_API size_t ZSTD_decompress( void* dst, size_t dstCapacity,- const void* src, size_t compressedSize);--/*! ZSTD_getDecompressedSize() :-* 'src' is the start of a zstd compressed frame.-* @return : content size to be decompressed, as a 64-bits value _if known_, 0 otherwise.-* note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode.-* When `return==0`, data to decompress could be any size.-* In which case, it's necessary to use streaming mode to decompress data.-* Optionally, application can still use ZSTD_decompress() while relying on implied limits.-* (For example, data may be necessarily cut into blocks <= 16 KB).-* note 2 : decompressed size is always present when compression is done with ZSTD_compress()-* note 3 : decompressed size can be very large (64-bits value),-* potentially larger than what local system can handle as a single memory segment.-* In which case, it's necessary to use streaming mode to decompress data.-* note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified.-* Always ensure result fits within application's authorized limits.-* Each application can set its own limits.-* note 5 : when `return==0`, if precise failure cause is needed, use ZSTD_getFrameParams() to know more. */-ZSTDLIB_API unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);---/*====== Helper functions ======*/-ZSTDLIB_API int ZSTD_maxCLevel(void); /*!< maximum compression level available */-ZSTDLIB_API size_t ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case scenario */-ZSTDLIB_API unsigned ZSTD_isError(size_t code); /*!< tells if a `size_t` function result is an error code */-ZSTDLIB_API const char* ZSTD_getErrorName(size_t code); /*!< provides readable string from an error code */---/***************************************-* Explicit memory management-***************************************/-/*= Compression context-* When compressing many times,-* it is recommended to allocate a context just once, and re-use it for each successive compression operation.-* This will make workload friendlier for system's memory.-* Use one context per thread for parallel execution in multi-threaded environments. */-typedef struct ZSTD_CCtx_s ZSTD_CCtx;-ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void);-ZSTDLIB_API size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx);--/*! ZSTD_compressCCtx() :- Same as ZSTD_compress(), requires an allocated ZSTD_CCtx (see ZSTD_createCCtx()). */-ZSTDLIB_API size_t ZSTD_compressCCtx(ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, int compressionLevel);--/*= Decompression context */-typedef struct ZSTD_DCtx_s ZSTD_DCtx;-ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx(void);-ZSTDLIB_API size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx);--/*! ZSTD_decompressDCtx() :-* Same as ZSTD_decompress(), requires an allocated ZSTD_DCtx (see ZSTD_createDCtx()). */-ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);---/**************************-* Simple dictionary API-***************************/-/*! ZSTD_compress_usingDict() :-* Compression using a predefined Dictionary (see dictBuilder/zdict.h).-* Note : This function loads the dictionary, resulting in significant startup delay.-* Note : When `dict == NULL || dictSize < 8` no dictionary is used. */-ZSTDLIB_API size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const void* dict,size_t dictSize,- int compressionLevel);--/*! ZSTD_decompress_usingDict() :-* Decompression using a predefined Dictionary (see dictBuilder/zdict.h).-* Dictionary must be identical to the one used during compression.-* Note : This function loads the dictionary, resulting in significant startup delay.-* Note : When `dict == NULL || dictSize < 8` no dictionary is used. */-ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const void* dict,size_t dictSize);---/****************************-* Fast dictionary API-****************************/-typedef struct ZSTD_CDict_s ZSTD_CDict;--/*! ZSTD_createCDict() :-* When compressing multiple messages / blocks with the same dictionary, it's recommended to load it just once.-* ZSTD_createCDict() will create a digested dictionary, ready to start future compression operations without startup delay.-* ZSTD_CDict can be created once and used by multiple threads concurrently, as its usage is read-only.-* `dictBuffer` can be released after ZSTD_CDict creation, as its content is copied within CDict */-ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize, int compressionLevel);--/*! ZSTD_freeCDict() :-* Function frees memory allocated by ZSTD_createCDict(). */-ZSTDLIB_API size_t ZSTD_freeCDict(ZSTD_CDict* CDict);--/*! ZSTD_compress_usingCDict() :-* Compression using a digested Dictionary.-* Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.-* Note that compression level is decided during dictionary creation. */-ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const ZSTD_CDict* cdict);---typedef struct ZSTD_DDict_s ZSTD_DDict;--/*! ZSTD_createDDict() :-* Create a digested dictionary, ready to start decompression operation without startup delay.-* dictBuffer can be released after DDict creation, as its content is copied inside DDict */-ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);--/*! ZSTD_freeDDict() :-* Function frees memory allocated with ZSTD_createDDict() */-ZSTDLIB_API size_t ZSTD_freeDDict(ZSTD_DDict* ddict);--/*! ZSTD_decompress_usingDDict() :-* Decompression using a digested Dictionary.-* Faster startup than ZSTD_decompress_usingDict(), recommended when same dictionary is used multiple times. */-ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const ZSTD_DDict* ddict);---/****************************-* Streaming-****************************/--typedef struct ZSTD_inBuffer_s {- const void* src; /**< start of input buffer */- size_t size; /**< size of input buffer */- size_t pos; /**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */-} ZSTD_inBuffer;--typedef struct ZSTD_outBuffer_s {- void* dst; /**< start of output buffer */- size_t size; /**< size of output buffer */- size_t pos; /**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */-} ZSTD_outBuffer;----/*-***********************************************************************-* Streaming compression - HowTo-*-* A ZSTD_CStream object is required to track streaming operation.-* Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.-* ZSTD_CStream objects can be reused multiple times on consecutive compression operations.-* It is recommended to re-use ZSTD_CStream in situations where many streaming operations will be achieved consecutively,-* since it will play nicer with system's memory, by re-using already allocated memory.-* Use one separate ZSTD_CStream per thread for parallel execution.-*-* Start a new compression by initializing ZSTD_CStream.-* Use ZSTD_initCStream() to start a new compression operation.-* Use ZSTD_initCStream_usingDict() or ZSTD_initCStream_usingCDict() for a compression which requires a dictionary (experimental section)-*-* Use ZSTD_compressStream() repetitively to consume input stream.-* The function will automatically update both `pos` fields.-* Note that it may not consume the entire input, in which case `pos < size`,-* and it's up to the caller to present again remaining data.-* @return : a size hint, preferred nb of bytes to use as input for next function call-* or an error code, which can be tested using ZSTD_isError().-* Note 1 : it's just a hint, to help latency a little, any other value will work fine.-* Note 2 : size hint is guaranteed to be <= ZSTD_CStreamInSize()-*-* At any moment, it's possible to flush whatever data remains within internal buffer, using ZSTD_flushStream().-* `output->pos` will be updated.-* Note that some content might still be left within internal buffer if `output->size` is too small.-* @return : nb of bytes still present within internal buffer (0 if it's empty)-* or an error code, which can be tested using ZSTD_isError().-*-* ZSTD_endStream() instructs to finish a frame.-* It will perform a flush and write frame epilogue.-* The epilogue is required for decoders to consider a frame completed.-* Similar to ZSTD_flushStream(), it may not be able to flush the full content if `output->size` is too small.-* In which case, call again ZSTD_endStream() to complete the flush.-* @return : nb of bytes still present within internal buffer (0 if it's empty, hence compression completed)-* or an error code, which can be tested using ZSTD_isError().-*-* *******************************************************************/--typedef struct ZSTD_CStream_s ZSTD_CStream;-ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);-ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);--ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);-ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);-ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);-ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);--ZSTDLIB_API size_t ZSTD_CStreamInSize(void); /**< recommended size for input buffer */-ZSTDLIB_API size_t ZSTD_CStreamOutSize(void); /**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block in all circumstances. */----/*-***************************************************************************-* Streaming decompression - HowTo-*-* A ZSTD_DStream object is required to track streaming operations.-* Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.-* ZSTD_DStream objects can be re-used multiple times.-*-* Use ZSTD_initDStream() to start a new decompression operation,-* or ZSTD_initDStream_usingDict() if decompression requires a dictionary.-* @return : recommended first input size-*-* Use ZSTD_decompressStream() repetitively to consume your input.-* The function will update both `pos` fields.-* If `input.pos < input.size`, some input has not been consumed.-* It's up to the caller to present again remaining data.-* If `output.pos < output.size`, decoder has flushed everything it could.-* @return : 0 when a frame is completely decoded and fully flushed,-* an error code, which can be tested using ZSTD_isError(),-* any other value > 0, which means there is still some decoding to do to complete current frame.-* The return value is a suggested next input size (a hint to improve latency) that will never load more than the current frame.-* *******************************************************************************/--typedef struct ZSTD_DStream_s ZSTD_DStream;-ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);-ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);--ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);-ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);--ZSTDLIB_API size_t ZSTD_DStreamInSize(void); /*!< recommended size for input buffer */-ZSTDLIB_API size_t ZSTD_DStreamOutSize(void); /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */--#endif /* ZSTD_H_235446 */---#if defined(ZSTD_STATIC_LINKING_ONLY) && !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY)-#define ZSTD_H_ZSTD_STATIC_LINKING_ONLY--/****************************************************************************************- * START OF ADVANCED AND EXPERIMENTAL FUNCTIONS- * The definitions in this section are considered experimental.- * They should never be used with a dynamic library, as they may change in the future.- * They are provided for advanced usages.- * Use them only in association with static linking.- * ***************************************************************************************/--/* --- Constants ---*/-#define ZSTD_MAGICNUMBER 0xFD2FB528 /* >= v0.8.0 */-#define ZSTD_MAGIC_SKIPPABLE_START 0x184D2A50U--#define ZSTD_WINDOWLOG_MAX_32 25-#define ZSTD_WINDOWLOG_MAX_64 27-#define ZSTD_WINDOWLOG_MAX ((U32)(MEM_32bits() ? ZSTD_WINDOWLOG_MAX_32 : ZSTD_WINDOWLOG_MAX_64))-#define ZSTD_WINDOWLOG_MIN 10-#define ZSTD_HASHLOG_MAX ZSTD_WINDOWLOG_MAX-#define ZSTD_HASHLOG_MIN 6-#define ZSTD_CHAINLOG_MAX (ZSTD_WINDOWLOG_MAX+1)-#define ZSTD_CHAINLOG_MIN ZSTD_HASHLOG_MIN-#define ZSTD_HASHLOG3_MAX 17-#define ZSTD_SEARCHLOG_MAX (ZSTD_WINDOWLOG_MAX-1)-#define ZSTD_SEARCHLOG_MIN 1-#define ZSTD_SEARCHLENGTH_MAX 7 /* only for ZSTD_fast, other strategies are limited to 6 */-#define ZSTD_SEARCHLENGTH_MIN 3 /* only for ZSTD_btopt, other strategies are limited to 4 */-#define ZSTD_TARGETLENGTH_MIN 4-#define ZSTD_TARGETLENGTH_MAX 999--#define ZSTD_FRAMEHEADERSIZE_MAX 18 /* for static allocation */-#define ZSTD_FRAMEHEADERSIZE_MIN 6-static const size_t ZSTD_frameHeaderSize_prefix = 5;-static const size_t ZSTD_frameHeaderSize_min = ZSTD_FRAMEHEADERSIZE_MIN;-static const size_t ZSTD_frameHeaderSize_max = ZSTD_FRAMEHEADERSIZE_MAX;-static const size_t ZSTD_skippableHeaderSize = 8; /* magic number + skippable frame length */---/*--- Advanced types ---*/-typedef enum { ZSTD_fast, ZSTD_dfast, ZSTD_greedy, ZSTD_lazy, ZSTD_lazy2, ZSTD_btlazy2, ZSTD_btopt, ZSTD_btopt2 } ZSTD_strategy; /* from faster to stronger */--typedef struct {- unsigned windowLog; /**< largest match distance : larger == more compression, more memory needed during decompression */- unsigned chainLog; /**< fully searched segment : larger == more compression, slower, more memory (useless for fast) */- unsigned hashLog; /**< dispatch table : larger == faster, more memory */- unsigned searchLog; /**< nb of searches : larger == more compression, slower */- unsigned searchLength; /**< match length searched : larger == faster decompression, sometimes less compression */- unsigned targetLength; /**< acceptable match size for optimal parser (only) : larger == more compression, slower */- ZSTD_strategy strategy;-} ZSTD_compressionParameters;--typedef struct {- unsigned contentSizeFlag; /**< 1: content size will be in frame header (if known). */- unsigned checksumFlag; /**< 1: will generate a 22-bits checksum at end of frame, to be used for error detection by decompressor */- unsigned noDictIDFlag; /**< 1: no dict ID will be saved into frame header (if dictionary compression) */-} ZSTD_frameParameters;--typedef struct {- ZSTD_compressionParameters cParams;- ZSTD_frameParameters fParams;-} ZSTD_parameters;--/*= Custom memory allocation functions */-typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size);-typedef void (*ZSTD_freeFunction) (void* opaque, void* address);-typedef struct { ZSTD_allocFunction customAlloc; ZSTD_freeFunction customFree; void* opaque; } ZSTD_customMem;---/***************************************-* Advanced compression functions-***************************************/-/*! ZSTD_estimateCCtxSize() :- * Gives the amount of memory allocated for a ZSTD_CCtx given a set of compression parameters.- * `frameContentSize` is an optional parameter, provide `0` if unknown */-ZSTDLIB_API size_t ZSTD_estimateCCtxSize(ZSTD_compressionParameters cParams);--/*! ZSTD_createCCtx_advanced() :- * Create a ZSTD compression context using external alloc and free functions */-ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem);--/*! ZSTD_sizeofCCtx() :- * Gives the amount of memory used by a given ZSTD_CCtx */-ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);--/*! ZSTD_createCDict_byReference() :- * Create a digested dictionary for compression- * Dictionary content is simply referenced, and therefore stays in dictBuffer.- * It is important that dictBuffer outlives CDict, it must remain read accessible throughout the lifetime of CDict */-ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);--/*! ZSTD_createCDict_advanced() :- * Create a ZSTD_CDict using external alloc and free, and customized compression parameters */-ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize, unsigned byReference,- ZSTD_parameters params, ZSTD_customMem customMem);--/*! ZSTD_sizeof_CDict() :- * Gives the amount of memory used by a given ZSTD_sizeof_CDict */-ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);--/*! ZSTD_getCParams() :-* @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.-* `estimatedSrcSize` value is optional, select 0 if not known */-ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);--/*! ZSTD_getParams() :-* same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.-* All fields of `ZSTD_frameParameters` are set to default (0) */-ZSTDLIB_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);--/*! ZSTD_checkCParams() :-* Ensure param values remain within authorized range */-ZSTDLIB_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);--/*! ZSTD_adjustCParams() :-* optimize params for a given `srcSize` and `dictSize`.-* both values are optional, select `0` if unknown. */-ZSTDLIB_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);--/*! ZSTD_compress_advanced() :-* Same as ZSTD_compress_usingDict(), with fine-tune control of each compression parameter */-ZSTDLIB_API size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx,- void* dst, size_t dstCapacity,- const void* src, size_t srcSize,- const void* dict,size_t dictSize,- ZSTD_parameters params);---/*--- Advanced decompression functions ---*/--/*! ZSTD_isFrame() :- * Tells if the content of `buffer` starts with a valid Frame Identifier.- * Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.- * Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.- * Note 3 : Skippable Frame Identifiers are considered valid. */-ZSTDLIB_API unsigned ZSTD_isFrame(const void* buffer, size_t size);--/*! ZSTD_estimateDCtxSize() :- * Gives the potential amount of memory allocated to create a ZSTD_DCtx */-ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);--/*! ZSTD_createDCtx_advanced() :- * Create a ZSTD decompression context using external alloc and free functions */-ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem);--/*! ZSTD_sizeof_DCtx() :- * Gives the amount of memory used by a given ZSTD_DCtx */-ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);--/*! ZSTD_createDDict_byReference() :- * Create a digested dictionary, ready to start decompression operation without startup delay.- * Dictionary content is simply referenced, and therefore stays in dictBuffer.- * It is important that dictBuffer outlives DDict, it must remain read accessible throughout the lifetime of DDict */-ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);--ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,- unsigned byReference, ZSTD_customMem customMem);--/*! ZSTD_sizeof_DDict() :- * Gives the amount of memory used by a given ZSTD_DDict */-ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);--/*! ZSTD_getDictID_fromDict() :- * Provides the dictID stored within dictionary.- * if @return == 0, the dictionary is not conformant with Zstandard specification.- * It can still be loaded, but as a content-only dictionary. */-ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);--/*! ZSTD_getDictID_fromDDict() :- * Provides the dictID of the dictionary loaded into `ddict`.- * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.- * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */-ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);--/*! ZSTD_getDictID_fromFrame() :- * Provides the dictID required to decompressed the frame stored within `src`.- * If @return == 0, the dictID could not be decoded.- * This could for one of the following reasons :- * - The frame does not require a dictionary to be decoded (most common case).- * - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.- * Note : this use case also happens when using a non-conformant dictionary.- * - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).- * - This is not a Zstandard frame.- * When identifying the exact failure cause, it's possible to used ZSTD_getFrameParams(), which will provide a more precise error code. */-ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);---/********************************************************************-* Advanced streaming functions-********************************************************************/--/*===== Advanced Streaming compression functions =====*/-ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem);-ZSTDLIB_API size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize); /**< pledgedSrcSize must be correct */-ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel);-ZSTDLIB_API size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize,- ZSTD_parameters params, unsigned long long pledgedSrcSize); /**< pledgedSrcSize is optional and can be zero == unknown */-ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict); /**< note : cdict will just be referenced, and must outlive compression session */-ZSTDLIB_API size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize); /**< re-use compression parameters from previous init; skip dictionary loading stage; zcs must be init at least once before */-ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);---/*===== Advanced Streaming decompression functions =====*/-typedef enum { ZSTDdsp_maxWindowSize } ZSTD_DStreamParameter_e;-ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem);-ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize);-ZSTDLIB_API size_t ZSTD_setDStreamParameter(ZSTD_DStream* zds, ZSTD_DStreamParameter_e paramType, unsigned paramValue);-ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict); /**< note : ddict will just be referenced, and must outlive decompression session */-ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds); /**< re-use decompression parameters from previous init; saves dictionary loading */-ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);---/*********************************************************************-* Buffer-less and synchronous inner streaming functions-*-* This is an advanced API, giving full control over buffer management, for users which need direct control over memory.-* But it's also a complex one, with many restrictions (documented below).-* Prefer using normal streaming API for an easier experience-********************************************************************* */--/**- Buffer-less streaming compression (synchronous mode)-- A ZSTD_CCtx object is required to track streaming operations.- Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource.- ZSTD_CCtx object can be re-used multiple times within successive compression operations.-- Start by initializing a context.- Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression,- or ZSTD_compressBegin_advanced(), for finer parameter control.- It's also possible to duplicate a reference context which has already been initialized, using ZSTD_copyCCtx()-- Then, consume your input using ZSTD_compressContinue().- There are some important considerations to keep in mind when using this advanced function :- - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffer only.- - Interface is synchronous : input is consumed entirely and produce 1+ (or more) compressed blocks.- - Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario.- Worst case evaluation is provided by ZSTD_compressBound().- ZSTD_compressContinue() doesn't guarantee recover after a failed compression.- - ZSTD_compressContinue() presumes prior input ***is still accessible and unmodified*** (up to maximum distance size, see WindowLog).- It remembers all previous contiguous blocks, plus one separated memory segment (which can itself consists of multiple contiguous blocks)- - ZSTD_compressContinue() detects that prior input has been overwritten when `src` buffer overlaps.- In which case, it will "discard" the relevant memory section from its history.-- Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum.- It's possible to use a NULL,0 src content, in which case, it will write a final empty block to end the frame,- Without last block mark, frames will be considered unfinished (broken) by decoders.-- You can then reuse `ZSTD_CCtx` (ZSTD_compressBegin()) to compress some new frame.-*/--/*===== Buffer-less streaming compression functions =====*/-ZSTDLIB_API size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);-ZSTDLIB_API size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);-ZSTDLIB_API size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize);-ZSTDLIB_API size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize);-ZSTDLIB_API size_t ZSTD_compressContinue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);-ZSTDLIB_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);----/*-- Buffer-less streaming decompression (synchronous mode)-- A ZSTD_DCtx object is required to track streaming operations.- Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.- A ZSTD_DCtx object can be re-used multiple times.-- First typical operation is to retrieve frame parameters, using ZSTD_getFrameParams().- It fills a ZSTD_frameParams structure which provide important information to correctly decode the frame,- such as the minimum rolling buffer size to allocate to decompress data (`windowSize`),- and the dictionary ID used.- (Note : content size is optional, it may not be present. 0 means : content size unknown).- Note that these values could be wrong, either because of data malformation, or because an attacker is spoofing deliberate false information.- As a consequence, check that values remain within valid application range, especially `windowSize`, before allocation.- Each application can set its own limit, depending on local restrictions. For extended interoperability, it is recommended to support at least 8 MB.- Frame parameters are extracted from the beginning of the compressed frame.- Data fragment must be large enough to ensure successful decoding, typically `ZSTD_frameHeaderSize_max` bytes.- @result : 0 : successful decoding, the `ZSTD_frameParams` structure is correctly filled.- >0 : `srcSize` is too small, please provide at least @result bytes on next attempt.- errorCode, which can be tested using ZSTD_isError().-- Start decompression, with ZSTD_decompressBegin() or ZSTD_decompressBegin_usingDict().- Alternatively, you can copy a prepared context, using ZSTD_copyDCtx().-- Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.- ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue().- ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail.-- @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).- It can be zero, which is not an error; it just means ZSTD_decompressContinue() has decoded some metadata item.- It can also be an error code, which can be tested with ZSTD_isError().-- ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize`.- They should preferably be located contiguously, prior to current block.- Alternatively, a round buffer of sufficient size is also possible. Sufficient size is determined by frame parameters.- ZSTD_decompressContinue() is very sensitive to contiguity,- if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,- or that previous contiguous segment is large enough to properly handle maximum back-reference.-- A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.- Context can then be reset to start a new decompression.-- Note : it's possible to know if next input to present is a header or a block, using ZSTD_nextInputType().- This information is not required to properly decode a frame.-- == Special case : skippable frames ==-- Skippable frames allow integration of user-defined data into a flow of concatenated frames.- Skippable frames will be ignored (skipped) by a decompressor. The format of skippable frames is as follows :- a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F- b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits- c) Frame Content - any content (User Data) of length equal to Frame Size- For skippable frames ZSTD_decompressContinue() always returns 0.- For skippable frames ZSTD_getFrameParams() returns fparamsPtr->windowLog==0 what means that a frame is skippable.- It also returns Frame Size as fparamsPtr->frameContentSize.-*/--typedef struct {- unsigned long long frameContentSize;- unsigned windowSize;- unsigned dictID;- unsigned checksumFlag;-} ZSTD_frameParams;--/*===== Buffer-less streaming decompression functions =====*/-ZSTDLIB_API size_t ZSTD_getFrameParams(ZSTD_frameParams* fparamsPtr, const void* src, size_t srcSize); /**< doesn't consume input, see details below */-ZSTDLIB_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx);-ZSTDLIB_API size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);-ZSTDLIB_API void ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx);-ZSTDLIB_API size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx);-ZSTDLIB_API size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);-typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;-ZSTDLIB_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);--/**- Block functions-- Block functions produce and decode raw zstd blocks, without frame metadata.- Frame metadata cost is typically ~18 bytes, which can be non-negligible for very small blocks (< 100 bytes).- User will have to take in charge required information to regenerate data, such as compressed and content sizes.-- A few rules to respect :- - Compressing and decompressing require a context structure- + Use ZSTD_createCCtx() and ZSTD_createDCtx()- - It is necessary to init context before starting- + compression : ZSTD_compressBegin()- + decompression : ZSTD_decompressBegin()- + variants _usingDict() are also allowed- + copyCCtx() and copyDCtx() work too- - Block size is limited, it must be <= ZSTD_getBlockSizeMax()- + If you need to compress more, cut data into multiple blocks- + Consider using the regular ZSTD_compress() instead, as frame metadata costs become negligible when source size is large.- - When a block is considered not compressible enough, ZSTD_compressBlock() result will be zero.- In which case, nothing is produced into `dst`.- + User must test for such outcome and deal directly with uncompressed data- + ZSTD_decompressBlock() doesn't accept uncompressed data as input !!!- + In case of multiple successive blocks, decoder must be informed of uncompressed block existence to follow proper history.- Use ZSTD_insertBlock() in such a case.-*/--#define ZSTD_BLOCKSIZE_ABSOLUTEMAX (128 * 1024) /* define, for static allocation */-/*===== Raw zstd block functions =====*/-ZSTDLIB_API size_t ZSTD_getBlockSizeMax(ZSTD_CCtx* cctx);-ZSTDLIB_API size_t ZSTD_compressBlock (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);-ZSTDLIB_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);-ZSTDLIB_API size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize); /**< insert block into `dctx` history. Useful for uncompressed blocks */+ * This source code is licensed under both the BSD-style license (found in the+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found+ * in the COPYING file in the root directory of this source tree).+ * You may select, at your option, one of the above-listed licenses.+ */+#if defined (__cplusplus)+extern "C" {+#endif++#ifndef ZSTD_H_235446+#define ZSTD_H_235446++/* ====== Dependency ======*/+#include <stddef.h> /* size_t */+++/* ===== ZSTDLIB_API : control library symbols visibility ===== */+#ifndef ZSTDLIB_VISIBILITY+# if defined(__GNUC__) && (__GNUC__ >= 4)+# define ZSTDLIB_VISIBILITY __attribute__ ((visibility ("default")))+# else+# define ZSTDLIB_VISIBILITY+# endif+#endif+#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)+# define ZSTDLIB_API __declspec(dllexport) ZSTDLIB_VISIBILITY+#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)+# define ZSTDLIB_API __declspec(dllimport) ZSTDLIB_VISIBILITY /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/+#else+# define ZSTDLIB_API ZSTDLIB_VISIBILITY+#endif+++/*******************************************************************************+ Introduction++ zstd, short for Zstandard, is a fast lossless compression algorithm, targeting+ real-time compression scenarios at zlib-level and better compression ratios.+ The zstd compression library provides in-memory compression and decompression+ functions.++ The library supports regular compression levels from 1 up to ZSTD_maxCLevel(),+ which is currently 22. Levels >= 20, labeled `--ultra`, should be used with+ caution, as they require more memory. The library also offers negative+ compression levels, which extend the range of speed vs. ratio preferences.+ The lower the level, the faster the speed (at the cost of compression).++ Compression can be done in:+ - a single step (described as Simple API)+ - a single step, reusing a context (described as Explicit context)+ - unbounded multiple steps (described as Streaming compression)++ The compression ratio achievable on small data can be highly improved using+ a dictionary. Dictionary compression can be performed in:+ - a single step (described as Simple dictionary API)+ - a single step, reusing a dictionary (described as Bulk-processing+ dictionary API)++ Advanced experimental functions can be accessed using+ `#define ZSTD_STATIC_LINKING_ONLY` before including zstd.h.++ Advanced experimental APIs should never be used with a dynamically-linked+ library. They are not "stable"; their definitions or signatures may change in+ the future. Only static linking is allowed.+*******************************************************************************/++/*------ Version ------*/+#define ZSTD_VERSION_MAJOR 1+#define ZSTD_VERSION_MINOR 4+#define ZSTD_VERSION_RELEASE 0++#define ZSTD_VERSION_NUMBER (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)+ZSTDLIB_API unsigned ZSTD_versionNumber(void); /**< to check runtime library version */++#define ZSTD_LIB_VERSION ZSTD_VERSION_MAJOR.ZSTD_VERSION_MINOR.ZSTD_VERSION_RELEASE+#define ZSTD_QUOTE(str) #str+#define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str)+#define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION)+ZSTDLIB_API const char* ZSTD_versionString(void); /* requires v1.3.0+ */++/***************************************+* Default constant+***************************************/+#ifndef ZSTD_CLEVEL_DEFAULT+# define ZSTD_CLEVEL_DEFAULT 3+#endif++/***************************************+* Constants+***************************************/++/* All magic numbers are supposed read/written to/from files/memory using little-endian convention */+#define ZSTD_MAGICNUMBER 0xFD2FB528 /* valid since v0.8.0 */+#define ZSTD_MAGIC_DICTIONARY 0xEC30A437 /* valid since v0.7.0 */+#define ZSTD_MAGIC_SKIPPABLE_START 0x184D2A50 /* all 16 values, from 0x184D2A50 to 0x184D2A5F, signal the beginning of a skippable frame */+#define ZSTD_MAGIC_SKIPPABLE_MASK 0xFFFFFFF0++#define ZSTD_BLOCKSIZELOG_MAX 17+#define ZSTD_BLOCKSIZE_MAX (1<<ZSTD_BLOCKSIZELOG_MAX)++++/***************************************+* Simple API+***************************************/+/*! ZSTD_compress() :+ * Compresses `src` content as a single zstd compressed frame into already allocated `dst`.+ * Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.+ * @return : compressed size written into `dst` (<= `dstCapacity),+ * or an error code if it fails (which can be tested using ZSTD_isError()). */+ZSTDLIB_API size_t ZSTD_compress( void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ int compressionLevel);++/*! ZSTD_decompress() :+ * `compressedSize` : must be the _exact_ size of some number of compressed and/or skippable frames.+ * `dstCapacity` is an upper bound of originalSize to regenerate.+ * If user cannot imply a maximum upper bound, it's better to use streaming mode to decompress data.+ * @return : the number of bytes decompressed into `dst` (<= `dstCapacity`),+ * or an errorCode if it fails (which can be tested using ZSTD_isError()). */+ZSTDLIB_API size_t ZSTD_decompress( void* dst, size_t dstCapacity,+ const void* src, size_t compressedSize);++/*! ZSTD_getFrameContentSize() : requires v1.3.0++ * `src` should point to the start of a ZSTD encoded frame.+ * `srcSize` must be at least as large as the frame header.+ * hint : any size >= `ZSTD_frameHeaderSize_max` is large enough.+ * @return : - decompressed size of `src` frame content, if known+ * - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined+ * - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small)+ * note 1 : a 0 return value means the frame is valid but "empty".+ * note 2 : decompressed size is an optional field, it may not be present, typically in streaming mode.+ * When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.+ * In which case, it's necessary to use streaming mode to decompress data.+ * Optionally, application can rely on some implicit limit,+ * as ZSTD_decompress() only needs an upper bound of decompressed size.+ * (For example, data could be necessarily cut into blocks <= 16 KB).+ * note 3 : decompressed size is always present when compression is completed using single-pass functions,+ * such as ZSTD_compress(), ZSTD_compressCCtx() ZSTD_compress_usingDict() or ZSTD_compress_usingCDict().+ * note 4 : decompressed size can be very large (64-bits value),+ * potentially larger than what local system can handle as a single memory segment.+ * In which case, it's necessary to use streaming mode to decompress data.+ * note 5 : If source is untrusted, decompressed size could be wrong or intentionally modified.+ * Always ensure return value fits within application's authorized limits.+ * Each application can set its own limits.+ * note 6 : This function replaces ZSTD_getDecompressedSize() */+#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)+#define ZSTD_CONTENTSIZE_ERROR (0ULL - 2)+ZSTDLIB_API unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);++/*! ZSTD_getDecompressedSize() :+ * NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize().+ * Both functions work the same way, but ZSTD_getDecompressedSize() blends+ * "empty", "unknown" and "error" results to the same return value (0),+ * while ZSTD_getFrameContentSize() gives them separate return values.+ * @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise. */+ZSTDLIB_API unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);++/*! ZSTD_findFrameCompressedSize() :+ * `src` should point to the start of a ZSTD frame or skippable frame.+ * `srcSize` must be >= first frame size+ * @return : the compressed size of the first frame starting at `src`,+ * suitable to pass as `srcSize` to `ZSTD_decompress` or similar,+ * or an error code if input is invalid */+ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);+++/*====== Helper functions ======*/+#define ZSTD_COMPRESSBOUND(srcSize) ((srcSize) + ((srcSize)>>8) + (((srcSize) < (128<<10)) ? (((128<<10) - (srcSize)) >> 11) /* margin, from 64 to 0 */ : 0)) /* this formula ensures that bound(A) + bound(B) <= bound(A+B) as long as A and B >= 128 KB */+ZSTDLIB_API size_t ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case single-pass scenario */+ZSTDLIB_API unsigned ZSTD_isError(size_t code); /*!< tells if a `size_t` function result is an error code */+ZSTDLIB_API const char* ZSTD_getErrorName(size_t code); /*!< provides readable string from an error code */+ZSTDLIB_API int ZSTD_minCLevel(void); /*!< minimum negative compression level allowed */+ZSTDLIB_API int ZSTD_maxCLevel(void); /*!< maximum compression level available */+++/***************************************+* Explicit context+***************************************/+/*= Compression context+ * When compressing many times,+ * it is recommended to allocate a context just once, and re-use it for each successive compression operation.+ * This will make workload friendlier for system's memory.+ * Use one context per thread for parallel execution in multi-threaded environments. */+typedef struct ZSTD_CCtx_s ZSTD_CCtx;+ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void);+ZSTDLIB_API size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx);++/*! ZSTD_compressCCtx() :+ * Same as ZSTD_compress(), using an explicit ZSTD_CCtx+ * The function will compress at requested compression level,+ * ignoring any other parameter */+ZSTDLIB_API size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ int compressionLevel);++/*= Decompression context+ * When decompressing many times,+ * it is recommended to allocate a context only once,+ * and re-use it for each successive compression operation.+ * This will make workload friendlier for system's memory.+ * Use one context per thread for parallel execution. */+typedef struct ZSTD_DCtx_s ZSTD_DCtx;+ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx(void);+ZSTDLIB_API size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx);++/*! ZSTD_decompressDCtx() :+ * Same as ZSTD_decompress(),+ * requires an allocated ZSTD_DCtx.+ * Compatible with sticky parameters.+ */+ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize);+++/***************************************+* Advanced compression API+***************************************/++/* API design :+ * Parameters are pushed one by one into an existing context,+ * using ZSTD_CCtx_set*() functions.+ * Pushed parameters are sticky : they are valid for next compressed frame, and any subsequent frame.+ * "sticky" parameters are applicable to `ZSTD_compress2()` and `ZSTD_compressStream*()` !+ * They do not apply to "simple" one-shot variants such as ZSTD_compressCCtx()+ *+ * It's possible to reset all parameters to "default" using ZSTD_CCtx_reset().+ *+ * This API supercedes all other "advanced" API entry points in the experimental section.+ * In the future, we expect to remove from experimental API entry points which are redundant with this API.+ */+++/* Compression strategies, listed from fastest to strongest */+typedef enum { ZSTD_fast=1,+ ZSTD_dfast=2,+ ZSTD_greedy=3,+ ZSTD_lazy=4,+ ZSTD_lazy2=5,+ ZSTD_btlazy2=6,+ ZSTD_btopt=7,+ ZSTD_btultra=8,+ ZSTD_btultra2=9+ /* note : new strategies _might_ be added in the future.+ Only the order (from fast to strong) is guaranteed */+} ZSTD_strategy;+++typedef enum {++ /* compression parameters+ * Note: When compressing with a ZSTD_CDict these parameters are superseded+ * by the parameters used to construct the ZSTD_CDict. See ZSTD_CCtx_refCDict()+ * for more info (superseded-by-cdict). */+ ZSTD_c_compressionLevel=100, /* Update all compression parameters according to pre-defined cLevel table+ * Default level is ZSTD_CLEVEL_DEFAULT==3.+ * Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT.+ * Note 1 : it's possible to pass a negative compression level.+ * Note 2 : setting a level sets all default values of other compression parameters */+ ZSTD_c_windowLog=101, /* Maximum allowed back-reference distance, expressed as power of 2.+ * Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX.+ * Special: value 0 means "use default windowLog".+ * Note: Using a windowLog greater than ZSTD_WINDOWLOG_LIMIT_DEFAULT+ * requires explicitly allowing such window size at decompression stage if using streaming. */+ ZSTD_c_hashLog=102, /* Size of the initial probe table, as a power of 2.+ * Resulting memory usage is (1 << (hashLog+2)).+ * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX.+ * Larger tables improve compression ratio of strategies <= dFast,+ * and improve speed of strategies > dFast.+ * Special: value 0 means "use default hashLog". */+ ZSTD_c_chainLog=103, /* Size of the multi-probe search table, as a power of 2.+ * Resulting memory usage is (1 << (chainLog+2)).+ * Must be clamped between ZSTD_CHAINLOG_MIN and ZSTD_CHAINLOG_MAX.+ * Larger tables result in better and slower compression.+ * This parameter is useless when using "fast" strategy.+ * It's still useful when using "dfast" strategy,+ * in which case it defines a secondary probe table.+ * Special: value 0 means "use default chainLog". */+ ZSTD_c_searchLog=104, /* Number of search attempts, as a power of 2.+ * More attempts result in better and slower compression.+ * This parameter is useless when using "fast" and "dFast" strategies.+ * Special: value 0 means "use default searchLog". */+ ZSTD_c_minMatch=105, /* Minimum size of searched matches.+ * Note that Zstandard can still find matches of smaller size,+ * it just tweaks its search algorithm to look for this size and larger.+ * Larger values increase compression and decompression speed, but decrease ratio.+ * Must be clamped between ZSTD_MINMATCH_MIN and ZSTD_MINMATCH_MAX.+ * Note that currently, for all strategies < btopt, effective minimum is 4.+ * , for all strategies > fast, effective maximum is 6.+ * Special: value 0 means "use default minMatchLength". */+ ZSTD_c_targetLength=106, /* Impact of this field depends on strategy.+ * For strategies btopt, btultra & btultra2:+ * Length of Match considered "good enough" to stop search.+ * Larger values make compression stronger, and slower.+ * For strategy fast:+ * Distance between match sampling.+ * Larger values make compression faster, and weaker.+ * Special: value 0 means "use default targetLength". */+ ZSTD_c_strategy=107, /* See ZSTD_strategy enum definition.+ * The higher the value of selected strategy, the more complex it is,+ * resulting in stronger and slower compression.+ * Special: value 0 means "use default strategy". */++ /* LDM mode parameters */+ ZSTD_c_enableLongDistanceMatching=160, /* Enable long distance matching.+ * This parameter is designed to improve compression ratio+ * for large inputs, by finding large matches at long distance.+ * It increases memory usage and window size.+ * Note: enabling this parameter increases default ZSTD_c_windowLog to 128 MB+ * except when expressly set to a different value. */+ ZSTD_c_ldmHashLog=161, /* Size of the table for long distance matching, as a power of 2.+ * Larger values increase memory usage and compression ratio,+ * but decrease compression speed.+ * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX+ * default: windowlog - 7.+ * Special: value 0 means "automatically determine hashlog". */+ ZSTD_c_ldmMinMatch=162, /* Minimum match size for long distance matcher.+ * Larger/too small values usually decrease compression ratio.+ * Must be clamped between ZSTD_LDM_MINMATCH_MIN and ZSTD_LDM_MINMATCH_MAX.+ * Special: value 0 means "use default value" (default: 64). */+ ZSTD_c_ldmBucketSizeLog=163, /* Log size of each bucket in the LDM hash table for collision resolution.+ * Larger values improve collision resolution but decrease compression speed.+ * The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX.+ * Special: value 0 means "use default value" (default: 3). */+ ZSTD_c_ldmHashRateLog=164, /* Frequency of inserting/looking up entries into the LDM hash table.+ * Must be clamped between 0 and (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN).+ * Default is MAX(0, (windowLog - ldmHashLog)), optimizing hash table usage.+ * Larger values improve compression speed.+ * Deviating far from default value will likely result in a compression ratio decrease.+ * Special: value 0 means "automatically determine hashRateLog". */++ /* frame parameters */+ ZSTD_c_contentSizeFlag=200, /* Content size will be written into frame header _whenever known_ (default:1)+ * Content size must be known at the beginning of compression.+ * This is automatically the case when using ZSTD_compress2(),+ * For streaming variants, content size must be provided with ZSTD_CCtx_setPledgedSrcSize() */+ ZSTD_c_checksumFlag=201, /* A 32-bits checksum of content is written at end of frame (default:0) */+ ZSTD_c_dictIDFlag=202, /* When applicable, dictionary's ID is written into frame header (default:1) */++ /* multi-threading parameters */+ /* These parameters are only useful if multi-threading is enabled (compiled with build macro ZSTD_MULTITHREAD).+ * They return an error otherwise. */+ ZSTD_c_nbWorkers=400, /* Select how many threads will be spawned to compress in parallel.+ * When nbWorkers >= 1, triggers asynchronous mode when used with ZSTD_compressStream*() :+ * ZSTD_compressStream*() consumes input and flush output if possible, but immediately gives back control to caller,+ * while compression work is performed in parallel, within worker threads.+ * (note : a strong exception to this rule is when first invocation of ZSTD_compressStream2() sets ZSTD_e_end :+ * in which case, ZSTD_compressStream2() delegates to ZSTD_compress2(), which is always a blocking call).+ * More workers improve speed, but also increase memory usage.+ * Default value is `0`, aka "single-threaded mode" : no worker is spawned, compression is performed inside Caller's thread, all invocations are blocking */+ ZSTD_c_jobSize=401, /* Size of a compression job. This value is enforced only when nbWorkers >= 1.+ * Each compression job is completed in parallel, so this value can indirectly impact the nb of active threads.+ * 0 means default, which is dynamically determined based on compression parameters.+ * Job size must be a minimum of overlap size, or 1 MB, whichever is largest.+ * The minimum size is automatically and transparently enforced */+ ZSTD_c_overlapLog=402, /* Control the overlap size, as a fraction of window size.+ * The overlap size is an amount of data reloaded from previous job at the beginning of a new job.+ * It helps preserve compression ratio, while each job is compressed in parallel.+ * This value is enforced only when nbWorkers >= 1.+ * Larger values increase compression ratio, but decrease speed.+ * Possible values range from 0 to 9 :+ * - 0 means "default" : value will be determined by the library, depending on strategy+ * - 1 means "no overlap"+ * - 9 means "full overlap", using a full window size.+ * Each intermediate rank increases/decreases load size by a factor 2 :+ * 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+ * default value varies between 6 and 9, depending on strategy */++ /* note : additional experimental parameters are also available+ * within the experimental section of the API.+ * At the time of this writing, they include :+ * ZSTD_c_rsyncable+ * ZSTD_c_format+ * ZSTD_c_forceMaxWindow+ * ZSTD_c_forceAttachDict+ * ZSTD_c_literalCompressionMode+ * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.+ * note : never ever use experimentalParam? names directly;+ * also, the enums values themselves are unstable and can still change.+ */+ ZSTD_c_experimentalParam1=500,+ ZSTD_c_experimentalParam2=10,+ ZSTD_c_experimentalParam3=1000,+ ZSTD_c_experimentalParam4=1001,+ ZSTD_c_experimentalParam5=1002,+} ZSTD_cParameter;++typedef struct {+ size_t error;+ int lowerBound;+ int upperBound;+} ZSTD_bounds;++/*! ZSTD_cParam_getBounds() :+ * All parameters must belong to an interval with lower and upper bounds,+ * otherwise they will either trigger an error or be automatically clamped.+ * @return : a structure, ZSTD_bounds, which contains+ * - an error status field, which must be tested using ZSTD_isError()+ * - lower and upper bounds, both inclusive+ */+ZSTDLIB_API ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter cParam);++/*! ZSTD_CCtx_setParameter() :+ * Set one compression parameter, selected by enum ZSTD_cParameter.+ * All parameters have valid bounds. Bounds can be queried using ZSTD_cParam_getBounds().+ * Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).+ * Setting a parameter is generally only possible during frame initialization (before starting compression).+ * Exception : when using multi-threading mode (nbWorkers >= 1),+ * the following parameters can be updated _during_ compression (within same frame):+ * => compressionLevel, hashLog, chainLog, searchLog, minMatch, targetLength and strategy.+ * new parameters will be active for next job only (after a flush()).+ * @return : an error code (which can be tested using ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value);++/*! ZSTD_CCtx_setPledgedSrcSize() :+ * Total input data size to be compressed as a single frame.+ * Value will be written in frame header, unless if explicitly forbidden using ZSTD_c_contentSizeFlag.+ * This value will also be controlled at end of frame, and trigger an error if not respected.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Note 1 : pledgedSrcSize==0 actually means zero, aka an empty frame.+ * In order to mean "unknown content size", pass constant ZSTD_CONTENTSIZE_UNKNOWN.+ * ZSTD_CONTENTSIZE_UNKNOWN is default value for any new frame.+ * Note 2 : pledgedSrcSize is only valid once, for the next frame.+ * It's discarded at the end of the frame, and replaced by ZSTD_CONTENTSIZE_UNKNOWN.+ * Note 3 : Whenever all input data is provided and consumed in a single round,+ * for example with ZSTD_compress2(),+ * or invoking immediately ZSTD_compressStream2(,,,ZSTD_e_end),+ * this value is automatically overridden by srcSize instead.+ */+ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);++typedef enum {+ ZSTD_reset_session_only = 1,+ ZSTD_reset_parameters = 2,+ ZSTD_reset_session_and_parameters = 3+} ZSTD_ResetDirective;++/*! ZSTD_CCtx_reset() :+ * There are 2 different things that can be reset, independently or jointly :+ * - The session : will stop compressing current frame, and make CCtx ready to start a new one.+ * Useful after an error, or to interrupt any ongoing compression.+ * Any internal data not yet flushed is cancelled.+ * Compression parameters and dictionary remain unchanged.+ * They will be used to compress next frame.+ * Resetting session never fails.+ * - The parameters : changes all parameters back to "default".+ * This removes any reference to any dictionary too.+ * Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing)+ * otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError())+ * - Both : similar to resetting the session, followed by resetting parameters.+ */+ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);++/*! ZSTD_compress2() :+ * Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.+ * ZSTD_compress2() always starts a new frame.+ * Should cctx hold data from a previously unfinished frame, everything about it is forgotten.+ * - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()+ * - The function is always blocking, returns when compression is completed.+ * Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.+ * @return : compressed size written into `dst` (<= `dstCapacity),+ * or an error code if it fails (which can be tested using ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_compress2( ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize);+++/***************************************+* Advanced decompression API+***************************************/++/* The advanced API pushes parameters one by one into an existing DCtx context.+ * Parameters are sticky, and remain valid for all following frames+ * using the same DCtx context.+ * It's possible to reset parameters to default values using ZSTD_DCtx_reset().+ * Note : This API is compatible with existing ZSTD_decompressDCtx() and ZSTD_decompressStream().+ * Therefore, no new decompression function is necessary.+ */++typedef enum {++ ZSTD_d_windowLogMax=100, /* Select a size limit (in power of 2) beyond which+ * the streaming API will refuse to allocate memory buffer+ * in order to protect the host from unreasonable memory requirements.+ * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.+ * By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT).+ * Special: value 0 means "use default maximum windowLog". */++ /* note : additional experimental parameters are also available+ * within the experimental section of the API.+ * At the time of this writing, they include :+ * ZSTD_c_format+ * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.+ * note : never ever use experimentalParam? names directly+ */+ ZSTD_d_experimentalParam1=1000++} ZSTD_dParameter;++/*! ZSTD_dParam_getBounds() :+ * All parameters must belong to an interval with lower and upper bounds,+ * otherwise they will either trigger an error or be automatically clamped.+ * @return : a structure, ZSTD_bounds, which contains+ * - an error status field, which must be tested using ZSTD_isError()+ * - both lower and upper bounds, inclusive+ */+ZSTDLIB_API ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam);++/*! ZSTD_DCtx_setParameter() :+ * Set one compression parameter, selected by enum ZSTD_dParameter.+ * All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds().+ * Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).+ * Setting a parameter is only possible during frame initialization (before starting decompression).+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value);++/*! ZSTD_DCtx_reset() :+ * Return a DCtx to clean state.+ * Session and parameters can be reset jointly or separately.+ * Parameters can only be reset when no active frame is being decompressed.+ * @return : 0, or an error code, which can be tested with ZSTD_isError()+ */+ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset);+++/****************************+* Streaming+****************************/++typedef struct ZSTD_inBuffer_s {+ const void* src; /**< start of input buffer */+ size_t size; /**< size of input buffer */+ size_t pos; /**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */+} ZSTD_inBuffer;++typedef struct ZSTD_outBuffer_s {+ void* dst; /**< start of output buffer */+ size_t size; /**< size of output buffer */+ size_t pos; /**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */+} ZSTD_outBuffer;++++/*-***********************************************************************+* Streaming compression - HowTo+*+* A ZSTD_CStream object is required to track streaming operation.+* Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.+* ZSTD_CStream objects can be reused multiple times on consecutive compression operations.+* It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.+*+* For parallel execution, use one separate ZSTD_CStream per thread.+*+* note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.+*+* Parameters are sticky : when starting a new compression on the same context,+* it will re-use the same sticky parameters as previous compression session.+* When in doubt, it's recommended to fully initialize the context before usage.+* Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),+* ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to+* set more specific parameters, the pledged source size, or load a dictionary.+*+* Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to+* consume input stream. The function will automatically update both `pos`+* fields within `input` and `output`.+* Note that the function may not consume the entire input, for example, because+* the output buffer is already full, in which case `input.pos < input.size`.+* The caller must check if input has been entirely consumed.+* If not, the caller must make some room to receive more compressed data,+* and then present again remaining input data.+* note: ZSTD_e_continue is guaranteed to make some forward progress when called,+* but doesn't guarantee maximal forward progress. This is especially relevant+* when compressing with multiple threads. The call won't block if it can+* consume some input, but if it can't it will wait for some, but not all,+* output to be flushed.+* @return : provides a minimum amount of data remaining to be flushed from internal buffers+* or an error code, which can be tested using ZSTD_isError().+*+* At any moment, it's possible to flush whatever data might remain stuck within internal buffer,+* using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated.+* Note that, if `output->size` is too small, a single invocation with ZSTD_e_flush might not be enough (return code > 0).+* In which case, make some room to receive more compressed data, and call again ZSTD_compressStream2() with ZSTD_e_flush.+* You must continue calling ZSTD_compressStream2() with ZSTD_e_flush until it returns 0, at which point you can change the+* operation.+* note: ZSTD_e_flush will flush as much output as possible, meaning when compressing with multiple threads, it will+* block until the flush is complete or the output buffer is full.+* @return : 0 if internal buffers are entirely flushed,+* >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),+* or an error code, which can be tested using ZSTD_isError().+*+* Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame.+* It will perform a flush and write frame epilogue.+* The epilogue is required for decoders to consider a frame completed.+* flush operation is the same, and follows same rules as calling ZSTD_compressStream2() with ZSTD_e_flush.+* You must continue calling ZSTD_compressStream2() with ZSTD_e_end until it returns 0, at which point you are free to+* start a new frame.+* note: ZSTD_e_end will flush as much output as possible, meaning when compressing with multiple threads, it will+* block until the flush is complete or the output buffer is full.+* @return : 0 if frame fully completed and fully flushed,+* >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),+* or an error code, which can be tested using ZSTD_isError().+*+* *******************************************************************/++typedef ZSTD_CCtx ZSTD_CStream; /**< CCtx and CStream are now effectively same object (>= v1.3.0) */+ /* Continue to distinguish them for compatibility with older versions <= v1.2.0 */+/*===== ZSTD_CStream management functions =====*/+ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);+ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);++/*===== Streaming compression functions =====*/+typedef enum {+ ZSTD_e_continue=0, /* collect more data, encoder decides when to output compressed result, for optimal compression ratio */+ ZSTD_e_flush=1, /* flush any data provided so far,+ * it creates (at least) one new block, that can be decoded immediately on reception;+ * frame will continue: any future data can still reference previously compressed data, improving compression.+ * note : multithreaded compression will block to flush as much output as possible. */+ ZSTD_e_end=2 /* flush any remaining data _and_ close current frame.+ * note that frame is only closed after compressed data is fully flushed (return value == 0).+ * After that point, any additional data starts a new frame.+ * note : each frame is independent (does not reference any content from previous frame).+ : note : multithreaded compression will block to flush as much output as possible. */+} ZSTD_EndDirective;++/*! ZSTD_compressStream2() :+ * Behaves about the same as ZSTD_compressStream, with additional control on end directive.+ * - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()+ * - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode)+ * - output->pos must be <= dstCapacity, input->pos must be <= srcSize+ * - output->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.+ * - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller.+ * - When nbWorkers>=1, function is non-blocking : it just acquires a copy of input, and distributes jobs to internal worker threads, flush whatever is available,+ * and then immediately returns, just indicating that there is some data remaining to be flushed.+ * The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte.+ * - 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.+ * - @return provides a minimum amount of data remaining to be flushed from internal buffers+ * or an error code, which can be tested using ZSTD_isError().+ * if @return != 0, flush is not fully completed, there is still some data left within internal buffers.+ * This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers.+ * For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed.+ * - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0),+ * only ZSTD_e_end or ZSTD_e_flush operations are allowed.+ * Before starting a new compression job, or changing compression parameters,+ * it is required to fully flush internal buffers.+ */+ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,+ ZSTD_outBuffer* output,+ ZSTD_inBuffer* input,+ ZSTD_EndDirective endOp);++ZSTDLIB_API size_t ZSTD_CStreamInSize(void); /**< recommended size for input buffer */+ZSTDLIB_API size_t ZSTD_CStreamOutSize(void); /**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block in all circumstances. */++/*******************************************************************************+ * This is a legacy streaming API, and can be replaced by ZSTD_CCtx_reset() and+ * ZSTD_compressStream2(). It is redundant, but is still fully supported.+ * Advanced parameters and dictionary compression can only be used through the+ * new API.+ ******************************************************************************/++/**+ * Equivalent to:+ *+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)+ * ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);+ */+ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);+/**+ * Alternative for ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue).+ * NOTE: The return value is different. ZSTD_compressStream() returns a hint for+ * the next read size (if non-zero and not an error). ZSTD_compressStream2()+ * returns the number of bytes left to flush (if non-zero and not an error).+ */+ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);+/** Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_flush). */+ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);+/** Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_end). */+ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);+++/*-***************************************************************************+* Streaming decompression - HowTo+*+* A ZSTD_DStream object is required to track streaming operations.+* Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.+* ZSTD_DStream objects can be re-used multiple times.+*+* Use ZSTD_initDStream() to start a new decompression operation.+* @return : recommended first input size+* Alternatively, use advanced API to set specific properties.+*+* Use ZSTD_decompressStream() repetitively to consume your input.+* The function will update both `pos` fields.+* If `input.pos < input.size`, some input has not been consumed.+* It's up to the caller to present again remaining data.+* The function tries to flush all data decoded immediately, respecting output buffer size.+* If `output.pos < output.size`, decoder has flushed everything it could.+* But if `output.pos == output.size`, there might be some data left within internal buffers.,+* In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer.+* Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX.+* @return : 0 when a frame is completely decoded and fully flushed,+* or an error code, which can be tested using ZSTD_isError(),+* or any other value > 0, which means there is still some decoding or flushing to do to complete current frame :+* the return value is a suggested next input size (just a hint for better latency)+* that will never request more than the remaining frame size.+* *******************************************************************************/++typedef ZSTD_DCtx ZSTD_DStream; /**< DCtx and DStream are now effectively same object (>= v1.3.0) */+ /* For compatibility with versions <= v1.2.0, prefer differentiating them. */+/*===== ZSTD_DStream management functions =====*/+ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);+ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);++/*===== Streaming decompression functions =====*/++/* This function is redundant with the advanced API and equivalent to:+ *+ * ZSTD_DCtx_reset(zds);+ * ZSTD_DCtx_refDDict(zds, NULL);+ */+ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);++ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);++ZSTDLIB_API size_t ZSTD_DStreamInSize(void); /*!< recommended size for input buffer */+ZSTDLIB_API size_t ZSTD_DStreamOutSize(void); /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */+++/**************************+* Simple dictionary API+***************************/+/*! ZSTD_compress_usingDict() :+ * Compression at an explicit compression level using a Dictionary.+ * A dictionary can be any arbitrary data segment (also called a prefix),+ * or a buffer with specified information (see dictBuilder/zdict.h).+ * Note : This function loads the dictionary, resulting in significant startup delay.+ * It's intended for a dictionary used only once.+ * Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used. */+ZSTDLIB_API size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize,+ int compressionLevel);++/*! ZSTD_decompress_usingDict() :+ * Decompression using a known Dictionary.+ * Dictionary must be identical to the one used during compression.+ * Note : This function loads the dictionary, resulting in significant startup delay.+ * It's intended for a dictionary used only once.+ * Note : When `dict == NULL || dictSize < 8` no dictionary is used. */+ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize);+++/***********************************+ * Bulk processing dictionary API+ **********************************/+typedef struct ZSTD_CDict_s ZSTD_CDict;++/*! ZSTD_createCDict() :+ * When compressing multiple messages / blocks using the same dictionary, it's recommended to load it only once.+ * ZSTD_createCDict() will create a digested dictionary, ready to start future compression operations without startup cost.+ * ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.+ * `dictBuffer` can be released after ZSTD_CDict creation, because its content is copied within CDict.+ * Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate `dictBuffer` content.+ * Note : A ZSTD_CDict can be created from an empty dictBuffer, but it is inefficient when used to compress small data. */+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,+ int compressionLevel);++/*! ZSTD_freeCDict() :+ * Function frees memory allocated by ZSTD_createCDict(). */+ZSTDLIB_API size_t ZSTD_freeCDict(ZSTD_CDict* CDict);++/*! ZSTD_compress_usingCDict() :+ * Compression using a digested Dictionary.+ * Recommended when same dictionary is used multiple times.+ * Note : compression level is _decided at dictionary creation time_,+ * and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */+ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict);+++typedef struct ZSTD_DDict_s ZSTD_DDict;++/*! ZSTD_createDDict() :+ * Create a digested dictionary, ready to start decompression operation without startup delay.+ * dictBuffer can be released after DDict creation, as its content is copied inside DDict. */+ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);++/*! ZSTD_freeDDict() :+ * Function frees memory allocated with ZSTD_createDDict() */+ZSTDLIB_API size_t ZSTD_freeDDict(ZSTD_DDict* ddict);++/*! ZSTD_decompress_usingDDict() :+ * Decompression using a digested Dictionary.+ * Recommended when same dictionary is used multiple times. */+ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_DDict* ddict);+++/********************************+ * Dictionary helper functions+ *******************************/++/*! ZSTD_getDictID_fromDict() :+ * Provides the dictID stored within dictionary.+ * if @return == 0, the dictionary is not conformant with Zstandard specification.+ * It can still be loaded, but as a content-only dictionary. */+ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);++/*! ZSTD_getDictID_fromDDict() :+ * Provides the dictID of the dictionary loaded into `ddict`.+ * If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.+ * Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */+ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);++/*! ZSTD_getDictID_fromFrame() :+ * Provides the dictID required to decompressed the frame stored within `src`.+ * If @return == 0, the dictID could not be decoded.+ * This could for one of the following reasons :+ * - The frame does not require a dictionary to be decoded (most common case).+ * - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.+ * Note : this use case also happens when using a non-conformant dictionary.+ * - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).+ * - This is not a Zstandard frame.+ * When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code. */+ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);+++/*******************************************************************************+ * Advanced dictionary and prefix API+ *+ * This API allows dictionaries to be used with ZSTD_compress2(),+ * ZSTD_compressStream2(), and ZSTD_decompress(). Dictionaries are sticky, and+ * only reset with the context is reset with ZSTD_reset_parameters or+ * ZSTD_reset_session_and_parameters. Prefixes are single-use.+ ******************************************************************************/+++/*! ZSTD_CCtx_loadDictionary() :+ * Create an internal CDict from `dict` buffer.+ * Decompression will have to use same dictionary.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary,+ * meaning "return to no-dictionary mode".+ * Note 1 : Dictionary is sticky, it will be used for all future compressed frames.+ * To return to "no-dictionary" situation, load a NULL dictionary (or reset parameters).+ * Note 2 : Loading a dictionary involves building tables.+ * It's also a CPU consuming operation, with non-negligible impact on latency.+ * Tables are dependent on compression parameters, and for this reason,+ * compression parameters can no longer be changed after loading a dictionary.+ * Note 3 :`dict` content will be copied internally.+ * Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead.+ * In such a case, dictionary buffer must outlive its users.+ * Note 4 : Use ZSTD_CCtx_loadDictionary_advanced()+ * to precisely select how dictionary content must be interpreted. */+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);++/*! ZSTD_CCtx_refCDict() :+ * Reference a prepared dictionary, to be used for all next compressed frames.+ * Note that compression parameters are enforced from within CDict,+ * and supersede any compression parameter previously set within CCtx.+ * The parameters ignored are labled as "superseded-by-cdict" in the ZSTD_cParameter enum docs.+ * The ignored parameters will be used again if the CCtx is returned to no-dictionary mode.+ * The dictionary will remain valid for future compressed frames using same CCtx.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Special : Referencing a NULL CDict means "return to no-dictionary mode".+ * Note 1 : Currently, only one dictionary can be managed.+ * Referencing a new dictionary effectively "discards" any previous one.+ * Note 2 : CDict is just referenced, its lifetime must outlive its usage within CCtx. */+ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);++/*! ZSTD_CCtx_refPrefix() :+ * Reference a prefix (single-usage dictionary) for next compressed frame.+ * A prefix is **only used once**. Tables are discarded at end of frame (ZSTD_e_end).+ * Decompression will need same prefix to properly regenerate data.+ * Compressing with a prefix is similar in outcome as performing a diff and compressing it,+ * but performs much faster, especially during decompression (compression speed is tunable with compression level).+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Special: Adding any prefix (including NULL) invalidates any previous prefix or dictionary+ * Note 1 : Prefix buffer is referenced. It **must** outlive compression.+ * Its content must remain unmodified during compression.+ * Note 2 : If the intention is to diff some large src data blob with some prior version of itself,+ * ensure that the window size is large enough to contain the entire source.+ * See ZSTD_c_windowLog.+ * Note 3 : Referencing a prefix involves building tables, which are dependent on compression parameters.+ * It's a CPU consuming operation, with non-negligible impact on latency.+ * If there is a need to use the same prefix multiple times, consider loadDictionary instead.+ * Note 4 : By default, the prefix is interpreted as raw content (ZSTD_dm_rawContent).+ * Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation. */+ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx,+ const void* prefix, size_t prefixSize);++/*! ZSTD_DCtx_loadDictionary() :+ * Create an internal DDict from dict buffer,+ * to be used to decompress next frames.+ * The dictionary remains valid for all future frames, until explicitly invalidated.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,+ * meaning "return to no-dictionary mode".+ * Note 1 : Loading a dictionary involves building tables,+ * which has a non-negligible impact on CPU usage and latency.+ * It's recommended to "load once, use many times", to amortize the cost+ * Note 2 :`dict` content will be copied internally, so `dict` can be released after loading.+ * Use ZSTD_DCtx_loadDictionary_byReference() to reference dictionary content instead.+ * Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to take control of+ * how dictionary content is loaded and interpreted.+ */+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);++/*! ZSTD_DCtx_refDDict() :+ * Reference a prepared dictionary, to be used to decompress next frames.+ * The dictionary remains active for decompression of future frames using same DCtx.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Note 1 : Currently, only one dictionary can be managed.+ * Referencing a new dictionary effectively "discards" any previous one.+ * Special: referencing a NULL DDict means "return to no-dictionary mode".+ * Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx.+ */+ZSTDLIB_API size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);++/*! ZSTD_DCtx_refPrefix() :+ * Reference a prefix (single-usage dictionary) to decompress next frame.+ * This is the reverse operation of ZSTD_CCtx_refPrefix(),+ * and must use the same prefix as the one used during compression.+ * Prefix is **only used once**. Reference is discarded at end of frame.+ * End of frame is reached when ZSTD_decompressStream() returns 0.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ * Note 1 : Adding any prefix (including NULL) invalidates any previously set prefix or dictionary+ * Note 2 : Prefix buffer is referenced. It **must** outlive decompression.+ * Prefix buffer must remain unmodified up to the end of frame,+ * reached when ZSTD_decompressStream() returns 0.+ * Note 3 : By default, the prefix is treated as raw content (ZSTD_dm_rawContent).+ * Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section)+ * Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost.+ * A full dictionary is more costly, as it requires building tables.+ */+ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx,+ const void* prefix, size_t prefixSize);++/* === Memory management === */++/*! ZSTD_sizeof_*() :+ * These functions give the _current_ memory usage of selected object.+ * Note that object memory usage can evolve (increase or decrease) over time. */+ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);+ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);+ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);+ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);+ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);+ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);++#endif /* ZSTD_H_235446 */+++/****************************************************************************************+ * ADVANCED AND EXPERIMENTAL FUNCTIONS+ ****************************************************************************************+ * The definitions in the following section are considered experimental.+ * They are provided for advanced scenarios.+ * They should never be used with a dynamic library, as prototypes may change in the future.+ * Use them only in association with static linking.+ * ***************************************************************************************/++#if defined(ZSTD_STATIC_LINKING_ONLY) && !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY)+#define ZSTD_H_ZSTD_STATIC_LINKING_ONLY++/****************************************************************************************+ * experimental API (static linking only)+ ****************************************************************************************+ * The following symbols and constants+ * are not planned to join "stable API" status in the near future.+ * They can still change in future versions.+ * Some of them are planned to remain in the static_only section indefinitely.+ * Some of them might be removed in the future (especially when redundant with existing stable functions)+ * ***************************************************************************************/++#define ZSTD_FRAMEHEADERSIZE_PREFIX 5 /* minimum input size required to query frame header size */+#define ZSTD_FRAMEHEADERSIZE_MIN 6+#define ZSTD_FRAMEHEADERSIZE_MAX 18 /* can be useful for static allocation */+#define ZSTD_SKIPPABLEHEADERSIZE 8++/* compression parameter bounds */+#define ZSTD_WINDOWLOG_MAX_32 30+#define ZSTD_WINDOWLOG_MAX_64 31+#define ZSTD_WINDOWLOG_MAX ((int)(sizeof(size_t) == 4 ? ZSTD_WINDOWLOG_MAX_32 : ZSTD_WINDOWLOG_MAX_64))+#define ZSTD_WINDOWLOG_MIN 10+#define ZSTD_HASHLOG_MAX ((ZSTD_WINDOWLOG_MAX < 30) ? ZSTD_WINDOWLOG_MAX : 30)+#define ZSTD_HASHLOG_MIN 6+#define ZSTD_CHAINLOG_MAX_32 29+#define ZSTD_CHAINLOG_MAX_64 30+#define ZSTD_CHAINLOG_MAX ((int)(sizeof(size_t) == 4 ? ZSTD_CHAINLOG_MAX_32 : ZSTD_CHAINLOG_MAX_64))+#define ZSTD_CHAINLOG_MIN ZSTD_HASHLOG_MIN+#define ZSTD_SEARCHLOG_MAX (ZSTD_WINDOWLOG_MAX-1)+#define ZSTD_SEARCHLOG_MIN 1+#define ZSTD_MINMATCH_MAX 7 /* only for ZSTD_fast, other strategies are limited to 6 */+#define ZSTD_MINMATCH_MIN 3 /* only for ZSTD_btopt+, faster strategies are limited to 4 */+#define ZSTD_TARGETLENGTH_MAX ZSTD_BLOCKSIZE_MAX+#define ZSTD_TARGETLENGTH_MIN 0 /* note : comparing this constant to an unsigned results in a tautological test */+#define ZSTD_STRATEGY_MIN ZSTD_fast+#define ZSTD_STRATEGY_MAX ZSTD_btultra2+++#define ZSTD_OVERLAPLOG_MIN 0+#define ZSTD_OVERLAPLOG_MAX 9++#define ZSTD_WINDOWLOG_LIMIT_DEFAULT 27 /* by default, the streaming decoder will refuse any frame+ * requiring larger than (1<<ZSTD_WINDOWLOG_LIMIT_DEFAULT) window size,+ * to preserve host's memory from unreasonable requirements.+ * This limit can be overridden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,).+ * The limit does not apply for one-pass decoders (such as ZSTD_decompress()), since no additional memory is allocated */+++/* LDM parameter bounds */+#define ZSTD_LDM_HASHLOG_MIN ZSTD_HASHLOG_MIN+#define ZSTD_LDM_HASHLOG_MAX ZSTD_HASHLOG_MAX+#define ZSTD_LDM_MINMATCH_MIN 4+#define ZSTD_LDM_MINMATCH_MAX 4096+#define ZSTD_LDM_BUCKETSIZELOG_MIN 1+#define ZSTD_LDM_BUCKETSIZELOG_MAX 8+#define ZSTD_LDM_HASHRATELOG_MIN 0+#define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)++/* internal */+#define ZSTD_HASHLOG3_MAX 17+++/* --- Advanced types --- */++typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params;++typedef struct {+ unsigned windowLog; /**< largest match distance : larger == more compression, more memory needed during decompression */+ unsigned chainLog; /**< fully searched segment : larger == more compression, slower, more memory (useless for fast) */+ unsigned hashLog; /**< dispatch table : larger == faster, more memory */+ unsigned searchLog; /**< nb of searches : larger == more compression, slower */+ unsigned minMatch; /**< match length searched : larger == faster decompression, sometimes less compression */+ unsigned targetLength; /**< acceptable match size for optimal parser (only) : larger == more compression, slower */+ ZSTD_strategy strategy; /**< see ZSTD_strategy definition above */+} ZSTD_compressionParameters;++typedef struct {+ int contentSizeFlag; /**< 1: content size will be in frame header (when known) */+ int checksumFlag; /**< 1: generate a 32-bits checksum using XXH64 algorithm at end of frame, for error detection */+ int noDictIDFlag; /**< 1: no dictID will be saved into frame header (dictID is only useful for dictionary compression) */+} ZSTD_frameParameters;++typedef struct {+ ZSTD_compressionParameters cParams;+ ZSTD_frameParameters fParams;+} ZSTD_parameters;++typedef enum {+ ZSTD_dct_auto = 0, /* dictionary is "full" when starting with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */+ ZSTD_dct_rawContent = 1, /* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */+ ZSTD_dct_fullDict = 2 /* refuses to load a dictionary if it does not respect Zstandard's specification, starting with ZSTD_MAGIC_DICTIONARY */+} ZSTD_dictContentType_e;++typedef enum {+ ZSTD_dlm_byCopy = 0, /**< Copy dictionary content internally */+ ZSTD_dlm_byRef = 1, /**< Reference dictionary content -- the dictionary buffer must outlive its users. */+} ZSTD_dictLoadMethod_e;++typedef enum {+ /* Opened question : should we have a format ZSTD_f_auto ?+ * Today, it would mean exactly the same as ZSTD_f_zstd1.+ * But, in the future, should several formats become supported,+ * on the compression side, it would mean "default format".+ * On the decompression side, it would mean "automatic format detection",+ * so that ZSTD_f_zstd1 would mean "accept *only* zstd frames".+ * Since meaning is a little different, another option could be to define different enums for compression and decompression.+ * This question could be kept for later, when there are actually multiple formats to support,+ * but there is also the question of pinning enum values, and pinning value `0` is especially important */+ ZSTD_f_zstd1 = 0, /* zstd frame format, specified in zstd_compression_format.md (default) */+ ZSTD_f_zstd1_magicless = 1, /* Variant of zstd frame format, without initial 4-bytes magic number.+ * Useful to save 4 bytes per generated frame.+ * Decoder cannot recognise automatically this format, requiring this instruction. */+} ZSTD_format_e;++typedef enum {+ /* Note: this enum and the behavior it controls are effectively internal+ * implementation details of the compressor. They are expected to continue+ * to evolve and should be considered only in the context of extremely+ * advanced performance tuning.+ *+ * Zstd currently supports the use of a CDict in two ways:+ *+ * - The contents of the CDict can be copied into the working context. This+ * means that the compression can search both the dictionary and input+ * while operating on a single set of internal tables. This makes+ * the compression faster per-byte of input. However, the initial copy of+ * the CDict's tables incurs a fixed cost at the beginning of the+ * compression. For small compressions (< 8 KB), that copy can dominate+ * the cost of the compression.+ *+ * - The CDict's tables can be used in-place. In this model, compression is+ * slower per input byte, because the compressor has to search two sets of+ * tables. However, this model incurs no start-up cost (as long as the+ * working context's tables can be reused). For small inputs, this can be+ * faster than copying the CDict's tables.+ *+ * Zstd has a simple internal heuristic that selects which strategy to use+ * at the beginning of a compression. However, if experimentation shows that+ * Zstd is making poor choices, it is possible to override that choice with+ * this enum.+ */+ ZSTD_dictDefaultAttach = 0, /* Use the default heuristic. */+ ZSTD_dictForceAttach = 1, /* Never copy the dictionary. */+ ZSTD_dictForceCopy = 2, /* Always copy the dictionary. */+} ZSTD_dictAttachPref_e;++typedef enum {+ ZSTD_lcm_auto = 0, /**< Automatically determine the compression mode based on the compression level.+ * Negative compression levels will be uncompressed, and positive compression+ * levels will be compressed. */+ ZSTD_lcm_huffman = 1, /**< Always attempt Huffman compression. Uncompressed literals will still be+ * emitted if Huffman compression is not profitable. */+ ZSTD_lcm_uncompressed = 2, /**< Always emit uncompressed literals. */+} ZSTD_literalCompressionMode_e;+++/***************************************+* Frame size functions+***************************************/++/*! ZSTD_findDecompressedSize() :+ * `src` should point to the start of a series of ZSTD encoded and/or skippable frames+ * `srcSize` must be the _exact_ size of this series+ * (i.e. there should be a frame boundary at `src + srcSize`)+ * @return : - decompressed size of all data in all successive frames+ * - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN+ * - if an error occurred: ZSTD_CONTENTSIZE_ERROR+ *+ * note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode.+ * When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.+ * In which case, it's necessary to use streaming mode to decompress data.+ * note 2 : decompressed size is always present when compression is done with ZSTD_compress()+ * note 3 : decompressed size can be very large (64-bits value),+ * potentially larger than what local system can handle as a single memory segment.+ * In which case, it's necessary to use streaming mode to decompress data.+ * note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified.+ * Always ensure result fits within application's authorized limits.+ * Each application can set its own limits.+ * note 5 : ZSTD_findDecompressedSize handles multiple frames, and so it must traverse the input to+ * read each contained frame header. This is fast as most of the data is skipped,+ * however it does mean that all frame data must be present and valid. */+ZSTDLIB_API unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize);++/** ZSTD_decompressBound() :+ * `src` should point to the start of a series of ZSTD encoded and/or skippable frames+ * `srcSize` must be the _exact_ size of this series+ * (i.e. there should be a frame boundary at `src + srcSize`)+ * @return : - upper-bound for the decompressed size of all data in all successive frames+ * - if an error occured: ZSTD_CONTENTSIZE_ERROR+ *+ * note 1 : an error can occur if `src` contains an invalid or incorrectly formatted frame.+ * note 2 : the upper-bound is exact when the decompressed size field is available in every ZSTD encoded frame of `src`.+ * in this case, `ZSTD_findDecompressedSize` and `ZSTD_decompressBound` return the same value.+ * note 3 : when the decompressed size field isn't available, the upper-bound for that frame is calculated by:+ * upper-bound = # blocks * min(128 KB, Window_Size)+ */+ZSTDLIB_API unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize);++/*! ZSTD_frameHeaderSize() :+ * srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.+ * @return : size of the Frame Header,+ * or an error code (if srcSize is too small) */+ZSTDLIB_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);+++/***************************************+* Memory management+***************************************/++/*! ZSTD_estimate*() :+ * These functions make it possible to estimate memory usage+ * of a future {D,C}Ctx, before its creation.+ * ZSTD_estimateCCtxSize() will provide a budget large enough for any compression level up to selected one.+ * It will also consider src size to be arbitrarily "large", which is worst case.+ * If srcSize is known to always be small, ZSTD_estimateCCtxSize_usingCParams() can provide a tighter estimation.+ * ZSTD_estimateCCtxSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.+ * ZSTD_estimateCCtxSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.+ * Note : CCtx size estimation is only correct for single-threaded compression. */+ZSTDLIB_API size_t ZSTD_estimateCCtxSize(int compressionLevel);+ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);+ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);+ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);++/*! ZSTD_estimateCStreamSize() :+ * ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one.+ * It will also consider src size to be arbitrarily "large", which is worst case.+ * If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.+ * ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.+ * ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.+ * Note : CStream size estimation is only correct for single-threaded compression.+ * ZSTD_DStream memory budget depends on window Size.+ * This information can be passed manually, using ZSTD_estimateDStreamSize,+ * or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame();+ * Note : if streaming is init with function ZSTD_init?Stream_usingDict(),+ * an internal ?Dict will be created, which additional size is not estimated here.+ * In this case, get total size by adding ZSTD_estimate?DictSize */+ZSTDLIB_API size_t ZSTD_estimateCStreamSize(int compressionLevel);+ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);+ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);+ZSTDLIB_API size_t ZSTD_estimateDStreamSize(size_t windowSize);+ZSTDLIB_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);++/*! ZSTD_estimate?DictSize() :+ * ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict().+ * ZSTD_estimateCDictSize_advanced() makes it possible to control compression parameters precisely, like ZSTD_createCDict_advanced().+ * Note : dictionaries created by reference (`ZSTD_dlm_byRef`) are logically smaller.+ */+ZSTDLIB_API size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel);+ZSTDLIB_API size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod);+ZSTDLIB_API size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod);++/*! ZSTD_initStatic*() :+ * Initialize an object using a pre-allocated fixed-size buffer.+ * workspace: The memory area to emplace the object into.+ * Provided pointer *must be 8-bytes aligned*.+ * Buffer must outlive object.+ * workspaceSize: Use ZSTD_estimate*Size() to determine+ * how large workspace must be to support target scenario.+ * @return : pointer to object (same address as workspace, just different type),+ * or NULL if error (size too small, incorrect alignment, etc.)+ * Note : zstd will never resize nor malloc() when using a static buffer.+ * If the object requires more memory than available,+ * zstd will just error out (typically ZSTD_error_memory_allocation).+ * Note 2 : there is no corresponding "free" function.+ * Since workspace is allocated externally, it must be freed externally too.+ * Note 3 : cParams : use ZSTD_getCParams() to convert a compression level+ * into its associated cParams.+ * Limitation 1 : currently not compatible with internal dictionary creation, triggered by+ * ZSTD_CCtx_loadDictionary(), ZSTD_initCStream_usingDict() or ZSTD_initDStream_usingDict().+ * Limitation 2 : static cctx currently not compatible with multi-threading.+ * Limitation 3 : static dctx is incompatible with legacy support.+ */+ZSTDLIB_API ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize);+ZSTDLIB_API ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize); /**< same as ZSTD_initStaticCCtx() */++ZSTDLIB_API ZSTD_DCtx* ZSTD_initStaticDCtx(void* workspace, size_t workspaceSize);+ZSTDLIB_API ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize); /**< same as ZSTD_initStaticDCtx() */++ZSTDLIB_API const ZSTD_CDict* ZSTD_initStaticCDict(+ void* workspace, size_t workspaceSize,+ const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_compressionParameters cParams);++ZSTDLIB_API const ZSTD_DDict* ZSTD_initStaticDDict(+ void* workspace, size_t workspaceSize,+ const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType);+++/*! Custom memory allocation :+ * These prototypes make it possible to pass your own allocation/free functions.+ * ZSTD_customMem is provided at creation time, using ZSTD_create*_advanced() variants listed below.+ * All allocation/free operations will be completed using these custom variants instead of regular <stdlib.h> ones.+ */+typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size);+typedef void (*ZSTD_freeFunction) (void* opaque, void* address);+typedef struct { ZSTD_allocFunction customAlloc; ZSTD_freeFunction customFree; void* opaque; } ZSTD_customMem;+static ZSTD_customMem const ZSTD_defaultCMem = { NULL, NULL, NULL }; /**< this constant defers to stdlib's functions */++ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem);+ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem);+ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem);+ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem);++ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_compressionParameters cParams,+ ZSTD_customMem customMem);++ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,+ ZSTD_dictLoadMethod_e dictLoadMethod,+ ZSTD_dictContentType_e dictContentType,+ ZSTD_customMem customMem);++++/***************************************+* Advanced compression functions+***************************************/++/*! ZSTD_createCDict_byReference() :+ * Create a digested dictionary for compression+ * Dictionary content is just referenced, not duplicated.+ * As a consequence, `dictBuffer` **must** outlive CDict,+ * and its content must remain unmodified throughout the lifetime of CDict. */+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);++/*! ZSTD_getCParams() :+ * @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.+ * `estimatedSrcSize` value is optional, select 0 if not known */+ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);++/*! ZSTD_getParams() :+ * same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.+ * All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0 */+ZSTDLIB_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);++/*! ZSTD_checkCParams() :+ * Ensure param values remain within authorized range.+ * @return 0 on success, or an error code (can be checked with ZSTD_isError()) */+ZSTDLIB_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);++/*! ZSTD_adjustCParams() :+ * optimize params for a given `srcSize` and `dictSize`.+ * `srcSize` can be unknown, in which case use ZSTD_CONTENTSIZE_UNKNOWN.+ * `dictSize` must be `0` when there is no dictionary.+ * cPar can be invalid : all parameters will be clamped within valid range in the @return struct.+ * This function never fails (wide contract) */+ZSTDLIB_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);++/*! ZSTD_compress_advanced() :+ * Same as ZSTD_compress_usingDict(), with fine-tune control over compression parameters (by structure) */+ZSTDLIB_API size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const void* dict,size_t dictSize,+ ZSTD_parameters params);++/*! ZSTD_compress_usingCDict_advanced() :+ * Same as ZSTD_compress_usingCDict(), with fine-tune control over frame parameters */+ZSTDLIB_API size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity,+ const void* src, size_t srcSize,+ const ZSTD_CDict* cdict,+ ZSTD_frameParameters fParams);+++/*! ZSTD_CCtx_loadDictionary_byReference() :+ * Same as ZSTD_CCtx_loadDictionary(), but dictionary content is referenced, instead of being copied into CCtx.+ * It saves some memory, but also requires that `dict` outlives its usage within `cctx` */+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);++/*! ZSTD_CCtx_loadDictionary_advanced() :+ * Same as ZSTD_CCtx_loadDictionary(), but gives finer control over+ * how to load the dictionary (by copy ? by reference ?)+ * and how to interpret it (automatic ? force raw mode ? full mode only ?) */+ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);++/*! ZSTD_CCtx_refPrefix_advanced() :+ * Same as ZSTD_CCtx_refPrefix(), but gives finer control over+ * how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */+ZSTDLIB_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);++/* === experimental parameters === */+/* these parameters can be used with ZSTD_setParameter()+ * they are not guaranteed to remain supported in the future */++ /* Enables rsyncable mode,+ * which makes compressed files more rsync friendly+ * by adding periodic synchronization points to the compressed data.+ * The target average block size is ZSTD_c_jobSize / 2.+ * It's possible to modify the job size to increase or decrease+ * the granularity of the synchronization point.+ * Once the jobSize is smaller than the window size,+ * it will result in compression ratio degradation.+ * NOTE 1: rsyncable mode only works when multithreading is enabled.+ * NOTE 2: rsyncable performs poorly in combination with long range mode,+ * since it will decrease the effectiveness of synchronization points,+ * though mileage may vary.+ * NOTE 3: Rsyncable mode limits maximum compression speed to ~400 MB/s.+ * If the selected compression level is already running significantly slower,+ * the overall speed won't be significantly impacted.+ */+ #define ZSTD_c_rsyncable ZSTD_c_experimentalParam1++/* Select a compression format.+ * The value must be of type ZSTD_format_e.+ * See ZSTD_format_e enum definition for details */+#define ZSTD_c_format ZSTD_c_experimentalParam2++/* Force back-reference distances to remain < windowSize,+ * even when referencing into Dictionary content (default:0) */+#define ZSTD_c_forceMaxWindow ZSTD_c_experimentalParam3++/* Controls whether the contents of a CDict+ * are used in place, or copied into the working context.+ * Accepts values from the ZSTD_dictAttachPref_e enum.+ * See the comments on that enum for an explanation of the feature. */+#define ZSTD_c_forceAttachDict ZSTD_c_experimentalParam4++/* Controls how the literals are compressed (default is auto).+ * The value must be of type ZSTD_literalCompressionMode_e.+ * See ZSTD_literalCompressionMode_t enum definition for details.+ */+#define ZSTD_c_literalCompressionMode ZSTD_c_experimentalParam5++/*! ZSTD_CCtx_getParameter() :+ * Get the requested compression parameter value, selected by enum ZSTD_cParameter,+ * and store it into int* value.+ * @return : 0, or an error code (which can be tested with ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value);+++/*! ZSTD_CCtx_params :+ * Quick howto :+ * - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure+ * - ZSTD_CCtxParams_setParameter() : Push parameters one by one into+ * an existing ZSTD_CCtx_params structure.+ * This is similar to+ * ZSTD_CCtx_setParameter().+ * - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to+ * an existing CCtx.+ * These parameters will be applied to+ * all subsequent frames.+ * - ZSTD_compressStream2() : Do compression using the CCtx.+ * - ZSTD_freeCCtxParams() : Free the memory.+ *+ * This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()+ * for static allocation of CCtx for single-threaded compression.+ */+ZSTDLIB_API ZSTD_CCtx_params* ZSTD_createCCtxParams(void);+ZSTDLIB_API size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params);++/*! ZSTD_CCtxParams_reset() :+ * Reset params to default values.+ */+ZSTDLIB_API size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params);++/*! ZSTD_CCtxParams_init() :+ * Initializes the compression parameters of cctxParams according to+ * compression level. All other parameters are reset to their default values.+ */+ZSTDLIB_API size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel);++/*! ZSTD_CCtxParams_init_advanced() :+ * Initializes the compression and frame parameters of cctxParams according to+ * params. All other parameters are reset to their default values.+ */+ZSTDLIB_API size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);++/*! ZSTD_CCtxParams_setParameter() :+ * Similar to ZSTD_CCtx_setParameter.+ * Set one compression parameter, selected by enum ZSTD_cParameter.+ * Parameters must be applied to a ZSTD_CCtx using ZSTD_CCtx_setParametersUsingCCtxParams().+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);++/*! ZSTD_CCtxParams_getParameter() :+ * Similar to ZSTD_CCtx_getParameter.+ * Get the requested value of one compression parameter, selected by enum ZSTD_cParameter.+ * @result : 0, or an error code (which can be tested with ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_CCtxParams_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);++/*! ZSTD_CCtx_setParametersUsingCCtxParams() :+ * Apply a set of ZSTD_CCtx_params to the compression context.+ * This can be done even after compression is started,+ * if nbWorkers==0, this will have no impact until a new compression is started.+ * if nbWorkers>=1, new parameters will be picked up at next job,+ * with a few restrictions (windowLog, pledgedSrcSize, nbWorkers, jobSize, and overlapLog are not updated).+ */+ZSTDLIB_API size_t ZSTD_CCtx_setParametersUsingCCtxParams(+ ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params);++/*! ZSTD_compressStream2_simpleArgs() :+ * Same as ZSTD_compressStream2(),+ * but using only integral types as arguments.+ * This variant might be helpful for binders from dynamic languages+ * which have troubles handling structures containing memory pointers.+ */+ZSTDLIB_API size_t ZSTD_compressStream2_simpleArgs (+ ZSTD_CCtx* cctx,+ void* dst, size_t dstCapacity, size_t* dstPos,+ const void* src, size_t srcSize, size_t* srcPos,+ ZSTD_EndDirective endOp);+++/***************************************+* Advanced decompression functions+***************************************/++/*! ZSTD_isFrame() :+ * Tells if the content of `buffer` starts with a valid Frame Identifier.+ * Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.+ * Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.+ * Note 3 : Skippable Frame Identifiers are considered valid. */+ZSTDLIB_API unsigned ZSTD_isFrame(const void* buffer, size_t size);++/*! ZSTD_createDDict_byReference() :+ * Create a digested dictionary, ready to start decompression operation without startup delay.+ * Dictionary content is referenced, and therefore stays in dictBuffer.+ * It is important that dictBuffer outlives DDict,+ * it must remain read accessible throughout the lifetime of DDict */+ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);++/*! ZSTD_DCtx_loadDictionary_byReference() :+ * Same as ZSTD_DCtx_loadDictionary(),+ * but references `dict` content instead of copying it into `dctx`.+ * This saves memory if `dict` remains around.,+ * However, it's imperative that `dict` remains accessible (and unmodified) while being used, so it must outlive decompression. */+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);++/*! ZSTD_DCtx_loadDictionary_advanced() :+ * Same as ZSTD_DCtx_loadDictionary(),+ * but gives direct control over+ * how to load the dictionary (by copy ? by reference ?)+ * and how to interpret it (automatic ? force raw mode ? full mode only ?). */+ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);++/*! ZSTD_DCtx_refPrefix_advanced() :+ * Same as ZSTD_DCtx_refPrefix(), but gives finer control over+ * how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */+ZSTDLIB_API size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);++/*! ZSTD_DCtx_setMaxWindowSize() :+ * Refuses allocating internal buffers for frames requiring a window size larger than provided limit.+ * This protects a decoder context from reserving too much memory for itself (potential attack scenario).+ * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.+ * By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT)+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).+ */+ZSTDLIB_API size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize);++/* ZSTD_d_format+ * experimental parameter,+ * allowing selection between ZSTD_format_e input compression formats+ */+#define ZSTD_d_format ZSTD_d_experimentalParam1++/*! ZSTD_DCtx_setFormat() :+ * Instruct the decoder context about what kind of data to decode next.+ * This instruction is mandatory to decode data without a fully-formed header,+ * such ZSTD_f_zstd1_magicless for example.+ * @return : 0, or an error code (which can be tested using ZSTD_isError()). */+ZSTDLIB_API size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format);++/*! ZSTD_decompressStream_simpleArgs() :+ * Same as ZSTD_decompressStream(),+ * but using only integral types as arguments.+ * This can be helpful for binders from dynamic languages+ * which have troubles handling structures containing memory pointers.+ */+ZSTDLIB_API size_t ZSTD_decompressStream_simpleArgs (+ ZSTD_DCtx* dctx,+ void* dst, size_t dstCapacity, size_t* dstPos,+ const void* src, size_t srcSize, size_t* srcPos);+++/********************************************************************+* Advanced streaming functions+* Warning : most of these functions are now redundant with the Advanced API.+* Once Advanced API reaches "stable" status,+* redundant functions will be deprecated, and then at some point removed.+********************************************************************/++/*===== Advanced Streaming compression functions =====*/+/**! ZSTD_initCStream_srcSize() :+ * This function is deprecated, and equivalent to:+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)+ * ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);+ * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);+ *+ * pledgedSrcSize must be correct. If it is not known at init time, use+ * ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs,+ * "0" also disables frame content size field. It may be enabled in the future.+ */+ZSTDLIB_API size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize);+/**! ZSTD_initCStream_usingDict() :+ * This function is deprecated, and is equivalent to:+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);+ * ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);+ *+ * Creates of an internal CDict (incompatible with static CCtx), except if+ * dict == NULL or dictSize < 8, in which case no dict is used.+ * Note: dict is loaded with ZSTD_dm_auto (treated as a full zstd dictionary if+ * it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.+ */+ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel);+/**! ZSTD_initCStream_advanced() :+ * This function is deprecated, and is approximately equivalent to:+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_setZstdParams(zcs, params); // Set the zstd params and leave the rest as-is+ * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);+ * ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);+ *+ * pledgedSrcSize must be correct. If srcSize is not known at init time, use+ * value ZSTD_CONTENTSIZE_UNKNOWN. dict is loaded with ZSTD_dm_auto and ZSTD_dlm_byCopy.+ */+ZSTDLIB_API size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize,+ ZSTD_parameters params, unsigned long long pledgedSrcSize);+/**! ZSTD_initCStream_usingCDict() :+ * This function is deprecated, and equivalent to:+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_refCDict(zcs, cdict);+ *+ * note : cdict will just be referenced, and must outlive compression session+ */+ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);+/**! ZSTD_initCStream_usingCDict_advanced() :+ * This function is deprecated, and is approximately equivalent to:+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_setZstdFrameParams(zcs, fParams); // Set the zstd frame params and leave the rest as-is+ * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);+ * ZSTD_CCtx_refCDict(zcs, cdict);+ *+ * same as ZSTD_initCStream_usingCDict(), with control over frame parameters.+ * pledgedSrcSize must be correct. If srcSize is not known at init time, use+ * value ZSTD_CONTENTSIZE_UNKNOWN.+ */+ZSTDLIB_API size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize);++/*! ZSTD_resetCStream() :+ * This function is deprecated, and is equivalent to:+ * ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);+ * ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);+ *+ * start a new frame, using same parameters from previous frame.+ * This is typically useful to skip dictionary loading stage, since it will re-use it in-place.+ * Note that zcs must be init at least once before using ZSTD_resetCStream().+ * If pledgedSrcSize is not known at reset time, use macro ZSTD_CONTENTSIZE_UNKNOWN.+ * If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end.+ * For the time being, pledgedSrcSize==0 is interpreted as "srcSize unknown" for compatibility with older programs,+ * but it will change to mean "empty" in future version, so use macro ZSTD_CONTENTSIZE_UNKNOWN instead.+ * @return : 0, or an error code (which can be tested using ZSTD_isError())+ */+ZSTDLIB_API size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize);+++typedef struct {+ unsigned long long ingested; /* nb input bytes read and buffered */+ unsigned long long consumed; /* nb input bytes actually compressed */+ unsigned long long produced; /* nb of compressed bytes generated and buffered */+ unsigned long long flushed; /* nb of compressed bytes flushed : not provided; can be tracked from caller side */+ unsigned currentJobID; /* MT only : latest started job nb */+ unsigned nbActiveWorkers; /* MT only : nb of workers actively compressing at probe time */+} ZSTD_frameProgression;++/* ZSTD_getFrameProgression() :+ * tells how much data has been ingested (read from input)+ * consumed (input actually compressed) and produced (output) for current frame.+ * Note : (ingested - consumed) is amount of input data buffered internally, not yet compressed.+ * Aggregates progression inside active worker threads.+ */+ZSTDLIB_API ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx);++/*! ZSTD_toFlushNow() :+ * Tell how many bytes are ready to be flushed immediately.+ * Useful for multithreading scenarios (nbWorkers >= 1).+ * Probe the oldest active job, defined as oldest job not yet entirely flushed,+ * and check its output buffer.+ * @return : amount of data stored in oldest job and ready to be flushed immediately.+ * if @return == 0, it means either :+ * + there is no active job (could be checked with ZSTD_frameProgression()), or+ * + oldest job is still actively compressing data,+ * but everything it has produced has also been flushed so far,+ * therefore flush speed is limited by production speed of oldest job+ * irrespective of the speed of concurrent (and newer) jobs.+ */+ZSTDLIB_API size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx);+++/*===== Advanced Streaming decompression functions =====*/+/**+ * This function is deprecated, and is equivalent to:+ *+ * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);+ * ZSTD_DCtx_loadDictionary(zds, dict, dictSize);+ *+ * note: no dictionary will be used if dict == NULL or dictSize < 8+ */+ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize);+/**+ * This function is deprecated, and is equivalent to:+ *+ * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);+ * ZSTD_DCtx_refDDict(zds, ddict);+ *+ * note : ddict is referenced, it must outlive decompression session+ */+ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);+/**+ * This function is deprecated, and is equivalent to:+ *+ * ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);+ *+ * re-use decompression parameters from previous init; saves dictionary loading+ */+ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);+++/*********************************************************************+* Buffer-less and synchronous inner streaming functions+*+* This is an advanced API, giving full control over buffer management, for users which need direct control over memory.+* But it's also a complex one, with several restrictions, documented below.+* Prefer normal streaming API for an easier experience.+********************************************************************* */++/**+ Buffer-less streaming compression (synchronous mode)++ A ZSTD_CCtx object is required to track streaming operations.+ Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource.+ ZSTD_CCtx object can be re-used multiple times within successive compression operations.++ Start by initializing a context.+ Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression,+ or ZSTD_compressBegin_advanced(), for finer parameter control.+ It's also possible to duplicate a reference context which has already been initialized, using ZSTD_copyCCtx()++ Then, consume your input using ZSTD_compressContinue().+ There are some important considerations to keep in mind when using this advanced function :+ - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffers only.+ - Interface is synchronous : input is consumed entirely and produces 1+ compressed blocks.+ - Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario.+ Worst case evaluation is provided by ZSTD_compressBound().+ ZSTD_compressContinue() doesn't guarantee recover after a failed compression.+ - ZSTD_compressContinue() presumes prior input ***is still accessible and unmodified*** (up to maximum distance size, see WindowLog).+ It remembers all previous contiguous blocks, plus one separated memory segment (which can itself consists of multiple contiguous blocks)+ - ZSTD_compressContinue() detects that prior input has been overwritten when `src` buffer overlaps.+ In which case, it will "discard" the relevant memory section from its history.++ Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum.+ It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame.+ Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders.++ `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress again.+*/++/*===== Buffer-less streaming compression functions =====*/+ZSTDLIB_API size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);+ZSTDLIB_API size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);+ZSTDLIB_API 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 */+ZSTDLIB_API size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); /**< note: fails if cdict==NULL */+ZSTDLIB_API 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 */+ZSTDLIB_API size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); /**< note: if pledgedSrcSize is not known, use ZSTD_CONTENTSIZE_UNKNOWN */++ZSTDLIB_API size_t ZSTD_compressContinue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);+ZSTDLIB_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);+++/*-+ Buffer-less streaming decompression (synchronous mode)++ A ZSTD_DCtx object is required to track streaming operations.+ Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.+ A ZSTD_DCtx object can be re-used multiple times.++ First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader().+ Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough.+ Data fragment must be large enough to ensure successful decoding.+ `ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough.+ @result : 0 : successful decoding, the `ZSTD_frameHeader` structure is correctly filled.+ >0 : `srcSize` is too small, please provide at least @result bytes on next attempt.+ errorCode, which can be tested using ZSTD_isError().++ It fills a ZSTD_frameHeader structure with important information to correctly decode the frame,+ such as the dictionary ID, content size, or maximum back-reference distance (`windowSize`).+ Note that these values could be wrong, either because of data corruption, or because a 3rd party deliberately spoofs false information.+ As a consequence, check that values remain within valid application range.+ For example, do not allocate memory blindly, check that `windowSize` is within expectation.+ Each application can set its own limits, depending on local restrictions.+ For extended interoperability, it is recommended to support `windowSize` of at least 8 MB.++ ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize` bytes.+ ZSTD_decompressContinue() is very sensitive to contiguity,+ if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,+ or that previous contiguous segment is large enough to properly handle maximum back-reference distance.+ There are multiple ways to guarantee this condition.++ The most memory efficient way is to use a round buffer of sufficient size.+ Sufficient size is determined by invoking ZSTD_decodingBufferSize_min(),+ which can @return an error code if required value is too large for current system (in 32-bits mode).+ In a round buffer methodology, ZSTD_decompressContinue() decompresses each block next to previous one,+ up to the moment there is not enough room left in the buffer to guarantee decoding another full block,+ which maximum size is provided in `ZSTD_frameHeader` structure, field `blockSizeMax`.+ At which point, decoding can resume from the beginning of the buffer.+ Note that already decoded data stored in the buffer should be flushed before being overwritten.++ There are alternatives possible, for example using two or more buffers of size `windowSize` each, though they consume more memory.++ Finally, if you control the compression process, you can also ignore all buffer size rules,+ as long as the encoder and decoder progress in "lock-step",+ aka use exactly the same buffer sizes, break contiguity at the same place, etc.++ Once buffers are setup, start decompression, with ZSTD_decompressBegin().+ If decompression requires a dictionary, use ZSTD_decompressBegin_usingDict() or ZSTD_decompressBegin_usingDDict().++ Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.+ ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue().+ ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail.++ @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).+ It can be zero : it just means ZSTD_decompressContinue() has decoded some metadata item.+ It can also be an error code, which can be tested with ZSTD_isError().++ A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.+ Context can then be reset to start a new decompression.++ Note : it's possible to know if next input to present is a header or a block, using ZSTD_nextInputType().+ This information is not required to properly decode a frame.++ == Special case : skippable frames ==++ Skippable frames allow integration of user-defined data into a flow of concatenated frames.+ Skippable frames will be ignored (skipped) by decompressor.+ The format of skippable frames is as follows :+ a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F+ b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits+ c) Frame Content - any content (User Data) of length equal to Frame Size+ For skippable frames ZSTD_getFrameHeader() returns zfhPtr->frameType==ZSTD_skippableFrame.+ For skippable frames ZSTD_decompressContinue() always returns 0 : it only skips the content.+*/++/*===== Buffer-less streaming decompression functions =====*/+typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;+typedef struct {+ unsigned long long frameContentSize; /* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */+ unsigned long long windowSize; /* can be very large, up to <= frameContentSize */+ unsigned blockSizeMax;+ ZSTD_frameType_e frameType; /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */+ unsigned headerSize;+ unsigned dictID;+ unsigned checksumFlag;+} ZSTD_frameHeader;++/** ZSTD_getFrameHeader() :+ * decode Frame Header, or requires larger `srcSize`.+ * @return : 0, `zfhPtr` is correctly filled,+ * >0, `srcSize` is too small, value is wanted `srcSize` amount,+ * or an error code, which can be tested using ZSTD_isError() */+ZSTDLIB_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize); /**< doesn't consume input */+/*! ZSTD_getFrameHeader_advanced() :+ * same as ZSTD_getFrameHeader(),+ * with added capability to select a format (like ZSTD_f_zstd1_magicless) */+ZSTDLIB_API size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);+ZSTDLIB_API size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize); /**< when frame content size is not known, pass in frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN */++ZSTDLIB_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx);+ZSTDLIB_API size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);+ZSTDLIB_API size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);++ZSTDLIB_API size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx);+ZSTDLIB_API size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);++/* misc */+ZSTDLIB_API void ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx);+typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;+ZSTDLIB_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);+++++/* ============================ */+/** Block level API */+/* ============================ */++/*!+ Block functions produce and decode raw zstd blocks, without frame metadata.+ Frame metadata cost is typically ~18 bytes, which can be non-negligible for very small blocks (< 100 bytes).+ User will have to take in charge required information to regenerate data, such as compressed and content sizes.++ A few rules to respect :+ - Compressing and decompressing require a context structure+ + Use ZSTD_createCCtx() and ZSTD_createDCtx()+ - It is necessary to init context before starting+ + compression : any ZSTD_compressBegin*() variant, including with dictionary+ + decompression : any ZSTD_decompressBegin*() variant, including with dictionary+ + copyCCtx() and copyDCtx() can be used too+ - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB+ + If input is larger than a block size, it's necessary to split input data into multiple blocks+ + For inputs larger than a single block, really consider using regular ZSTD_compress() instead.+ Frame metadata is not that costly, and quickly becomes negligible as source size grows larger.+ - When a block is considered not compressible enough, ZSTD_compressBlock() result will be zero.+ In which case, nothing is produced into `dst` !+ + User must test for such outcome and deal directly with uncompressed data+ + ZSTD_decompressBlock() doesn't accept uncompressed data as input !!!+ + In case of multiple successive blocks, should some of them be uncompressed,+ decoder must be informed of their existence in order to follow proper history.+ Use ZSTD_insertBlock() for such a case.+*/++/*===== Raw zstd block functions =====*/+ZSTDLIB_API size_t ZSTD_getBlockSize (const ZSTD_CCtx* cctx);+ZSTDLIB_API size_t ZSTD_compressBlock (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);+ZSTDLIB_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);+ZSTDLIB_API size_t ZSTD_insertBlock (ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize); /**< insert uncompressed block into `dctx` history. Useful for multi-blocks decompression. */ #endif /* ZSTD_H_ZSTD_STATIC_LINKING_ONLY */