This is an archive of the discontinued Mercurial Phabricator instance.

Differential D6858

zstandard: vendor python-zstandard 0.12
ClosedPublic

Authored by indygreg on Sep 15 2019, 11:01 PM.

Details

Reviewers
pulkit
Group Reviewers
hg-reviewers
Commits
rHG69de49c4e39c: zstandard: vendor python-zstandard 0.12
Summary

The upstream source distribution from PyPI was extracted. Unwanted
files were removed.

The clang-format ignore list was updated to reflect the new source
of files.

test-repo-compengines.t was updated to reflect a change in behavior
of the zstd library.

The project contains a vendored copy of zstandard 1.4.3. The old
version was 1.3.8. This should result in some minor performance wins.

  1. no-check-commit because 3rd party code has different style guidelines

Diff Detail

Repository
rHG Mercurial
Lint
Lint Skipped
Unit
Unit Tests Skipped

Event Timeline

indygreg created this revision.Sep 15 2019, 11:01 PM
indygreg retitled this revision from zstandard: vendor python-zstandard 0.11 to zstandard: vendor python-zstandard 0.12.Sep 15 2019, 11:04 PM
pulkit accepted this revision.Sep 16 2019, 1:09 PM
This revision is now accepted and ready to land.Sep 16 2019, 1:09 PM
indygreg added a commit: rHG69de49c4e39c: zstandard: vendor python-zstandard 0.12.Sep 16 2019, 1:56 PM
Closed by commit rHG69de49c4e39c: zstandard: vendor python-zstandard 0.12 (authored by indygreg). · Explain Why
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

PathPackages
Mcontrib/clang-format-ignorelist (4 lines)
Mcontrib/python-zstandard/NEWS.rst (32 lines)
Mcontrib/python-zstandard/README.rst (10 lines)
Mcontrib/python-zstandard/c-ext/compressionparams.c (10 lines)
Mcontrib/python-zstandard/c-ext/decompressionreader.c (2 lines)
Mcontrib/python-zstandard/c-ext/python-zstandard.h (2 lines)
Mcontrib/python-zstandard/make_cffi.py (8 lines)
Mcontrib/python-zstandard/setup.py (1 line)
Mcontrib/python-zstandard/setup_zstd.py (2 lines)
Mcontrib/python-zstandard/tests/test_compressor.py (6 lines)
Mcontrib/python-zstandard/tests/test_data_structures.py (2 lines)
Mcontrib/python-zstandard/tests/test_module_attributes.py (4 lines)
Mcontrib/python-zstandard/tests/test_train_dictionary.py (3 lines)
Mcontrib/python-zstandard/zstandard/__init__.py (2 lines)
Mcontrib/python-zstandard/zstandard/cffi.py (4 lines)
Mcontrib/python-zstandard/zstd.c (2 lines)
Mcontrib/python-zstandard/zstd/common/bitstream.h (4 lines)
Mcontrib/python-zstandard/zstd/common/compiler.h (15 lines)
Mcontrib/python-zstandard/zstd/common/fse.h (2 lines)
Mcontrib/python-zstandard/zstd/common/mem.h (2 lines)
Mcontrib/python-zstandard/zstd/common/threading.c (4 lines)
Mcontrib/python-zstandard/zstd/common/xxhash.c (14 lines)
Mcontrib/python-zstandard/zstd/common/zstd_internal.h (123 lines)
Mcontrib/python-zstandard/zstd/compress/fse_compress.c (4 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_compress.c (1558 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_compress_internal.h (181 lines)
AMcontrib/python-zstandard/zstd/compress/zstd_compress_literals.h (29 lines)
AMcontrib/python-zstandard/zstd/compress/zstd_compress_literals.c (149 lines)
AMcontrib/python-zstandard/zstd/compress/zstd_compress_sequences.h (47 lines)
AMcontrib/python-zstandard/zstd/compress/zstd_compress_sequences.c (415 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_double_fast.c (91 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_fast.c (382 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_lazy.h (2 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_lazy.c (61 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_ldm.c (4 lines)
Mcontrib/python-zstandard/zstd/compress/zstd_opt.c (119 lines)
Mcontrib/python-zstandard/zstd/compress/zstdmt_compress.h (70 lines)
Mcontrib/python-zstandard/zstd/compress/zstdmt_compress.c (115 lines)
Mcontrib/python-zstandard/zstd/decompress/zstd_ddict.c (8 lines)
Mcontrib/python-zstandard/zstd/decompress/zstd_decompress.c (495 lines)
Mcontrib/python-zstandard/zstd/decompress/zstd_decompress_block.c (132 lines)
Mcontrib/python-zstandard/zstd/decompress/zstd_decompress_internal.h (7 lines)
Mcontrib/python-zstandard/zstd/dictBuilder/cover.h (70 lines)
Mcontrib/python-zstandard/zstd/dictBuilder/cover.c (275 lines)
Mcontrib/python-zstandard/zstd/dictBuilder/fastcover.c (123 lines)
Mcontrib/python-zstandard/zstd/dictBuilder/zdict.h (31 lines)
Mcontrib/python-zstandard/zstd/dictBuilder/zdict.c (18 lines)
Mcontrib/python-zstandard/zstd/zstd.h (1145 lines)
Mtests/test-repo-compengines.t (2 lines)

Diff 16552

contrib/clang-format-ignorelist

contrib/python-zstandard/zstd/common/zstd_errors.h contrib/python-zstandard/zstd/common/zstd_errors.h
contrib/python-zstandard/zstd/common/zstd_internal.h contrib/python-zstandard/zstd/common/zstd_internal.h
contrib/python-zstandard/zstd/compress/fse_compress.c contrib/python-zstandard/zstd/compress/fse_compress.c
contrib/python-zstandard/zstd/compress/hist.c contrib/python-zstandard/zstd/compress/hist.c
contrib/python-zstandard/zstd/compress/hist.h contrib/python-zstandard/zstd/compress/hist.h
contrib/python-zstandard/zstd/compress/huf_compress.c contrib/python-zstandard/zstd/compress/huf_compress.c
contrib/python-zstandard/zstd/compress/zstd_compress.c contrib/python-zstandard/zstd/compress/zstd_compress.c
contrib/python-zstandard/zstd/compress/zstd_compress_internal.h contrib/python-zstandard/zstd/compress/zstd_compress_internal.h
contrib/python-zstandard/zstd/compress/zstd_compress_literals.c
contrib/python-zstandard/zstd/compress/zstd_compress_literals.h
contrib/python-zstandard/zstd/compress/zstd_compress_sequences.c
contrib/python-zstandard/zstd/compress/zstd_compress_sequences.h
contrib/python-zstandard/zstd/compress/zstd_double_fast.c contrib/python-zstandard/zstd/compress/zstd_double_fast.c
contrib/python-zstandard/zstd/compress/zstd_double_fast.h contrib/python-zstandard/zstd/compress/zstd_double_fast.h
contrib/python-zstandard/zstd/compress/zstd_fast.c contrib/python-zstandard/zstd/compress/zstd_fast.c
contrib/python-zstandard/zstd/compress/zstd_fast.h contrib/python-zstandard/zstd/compress/zstd_fast.h
contrib/python-zstandard/zstd/compress/zstd_lazy.c contrib/python-zstandard/zstd/compress/zstd_lazy.c
contrib/python-zstandard/zstd/compress/zstd_lazy.h contrib/python-zstandard/zstd/compress/zstd_lazy.h
contrib/python-zstandard/zstd/compress/zstd_ldm.c contrib/python-zstandard/zstd/compress/zstd_ldm.c
contrib/python-zstandard/zstd/compress/zstd_ldm.h contrib/python-zstandard/zstd/compress/zstd_ldm.h

contrib/python-zstandard/NEWS.rst

* Detect memory leaks via bench.py. * Detect memory leaks via bench.py.
* Remove low-level compression parameters from ``ZstdCompressor.__init__`` and * Remove low-level compression parameters from ``ZstdCompressor.__init__`` and
require use of ``CompressionParameters``. require use of ``CompressionParameters``.
* Expose ``ZSTD_getFrameProgression()`` from more compressor types. * Expose ``ZSTD_getFrameProgression()`` from more compressor types.
* Support modifying compression parameters mid operation when supported by * Support modifying compression parameters mid operation when supported by
zstd API. zstd API.
* Expose ``ZSTD_CLEVEL_DEFAULT`` constant. * Expose ``ZSTD_CLEVEL_DEFAULT`` constant.
* Support ``ZSTD_p_forceAttachDict`` compression parameter. * Support ``ZSTD_p_forceAttachDict`` compression parameter.
* Support ``ZSTD_c_literalCompressionMode `` compression parameter.
* Use ``ZSTD_CCtx_getParameter()``/``ZSTD_CCtxParam_getParameter()`` for retrieving * Use ``ZSTD_CCtx_getParameter()``/``ZSTD_CCtxParam_getParameter()`` for retrieving
compression parameters. compression parameters.
* Consider exposing ``ZSTDMT_toFlushNow()``. * Consider exposing ``ZSTDMT_toFlushNow()``.
* Expose ``ZDICT_trainFromBuffer_fastCover()``, * Expose ``ZDICT_trainFromBuffer_fastCover()``,
``ZDICT_optimizeTrainFromBuffer_fastCover``. ``ZDICT_optimizeTrainFromBuffer_fastCover``.
* Expose and enforce ``ZSTD_minCLevel()`` for minimum compression level. * Expose and enforce ``ZSTD_minCLevel()`` for minimum compression level.
* Consider a ``chunker()`` API for decompression. * Consider a ``chunker()`` API for decompression.
* Consider stats for ``chunker()`` API, including finding the last consumed * Consider stats for ``chunker()`` API, including finding the last consumed
offset of input data. offset of input data.
* Consider exposing ``ZSTD_cParam_getBounds()`` and * Consider exposing ``ZSTD_cParam_getBounds()`` and
``ZSTD_dParam_getBounds()`` APIs. ``ZSTD_dParam_getBounds()`` APIs.
* Consider controls over resetting compression contexts (session only, parameters, * Consider controls over resetting compression contexts (session only, parameters,
or session and parameters). or session and parameters).
* Actually use the CFFI backend in fuzzing tests. * Actually use the CFFI backend in fuzzing tests.
Other Actions Not Blocking Release Other Actions Not Blocking Release
--------------------------------------- ---------------------------------------
* Support for block compression APIs. * Support for block compression APIs.
* API for ensuring max memory ceiling isn't exceeded. * API for ensuring max memory ceiling isn't exceeded.
* Move off nose for testing. * Move off nose for testing.
0.12.0 (released 2019-09-15)
============================
Backwards Compatibility Notes
-----------------------------
* Support for Python 3.4 has been dropped since Python 3.4 is no longer
a supported Python version upstream. (But it will likely continue to
work until Python 2.7 support is dropped and we port to Python 3.5+
APIs.)
Bug Fixes
---------
* Fix ``ZstdDecompressor.__init__`` on 64-bit big-endian systems (#91).
* Fix memory leak in ``ZstdDecompressionReader.seek()`` (#82).
Changes
-------
* CI transitioned to Azure Pipelines (from AppVeyor and Travis CI).
* Switched to ``pytest`` for running tests (from ``nose``).
* Bundled zstandard library upgraded from 1.3.8 to 1.4.3.
0.11.1 (released 2019-05-14)
============================
* Fix memory leak in ``ZstdDecompressionReader.seek()`` (#82).
0.11.0 (released 2019-02-24) 0.11.0 (released 2019-02-24)
============================ ============================
Backwards Compatibility Nodes Backwards Compatibility Notes
----------------------------- -----------------------------
* ``ZstdDecompressor.read()`` now allows reading sizes of ``-1`` or ``0`` * ``ZstdDecompressor.read()`` now allows reading sizes of ``-1`` or ``0``
and defaults to ``-1``, per the documented behavior of and defaults to ``-1``, per the documented behavior of
``io.RawIOBase.read()``. Previously, we required an argument that was ``io.RawIOBase.read()``. Previously, we required an argument that was
a positive value. a positive value.
* The ``readline()``, ``readlines()``, ``__iter__``, and ``__next__`` methods * The ``readline()``, ``readlines()``, ``__iter__``, and ``__next__`` methods
of ``ZstdDecompressionReader()`` now raise ``io.UnsupportedOperation`` of ``ZstdDecompressionReader()`` now raise ``io.UnsupportedOperation``

contrib/python-zstandard/README.rst

underlying C API through a Pythonic interface while not sacrificing underlying C API through a Pythonic interface while not sacrificing
performance. This means exposing most of the features and flexibility performance. This means exposing most of the features and flexibility
of the C API while not sacrificing usability or safety that Python provides. of the C API while not sacrificing usability or safety that Python provides.
The canonical home for this project lives in a Mercurial repository run by The canonical home for this project lives in a Mercurial repository run by
the author. For convenience, that repository is frequently synchronized to the author. For convenience, that repository is frequently synchronized to
https://github.com/indygreg/python-zstandard. https://github.com/indygreg/python-zstandard.
| |ci-status| |win-ci-status| | |ci-status|
Requirements Requirements
============ ============
This extension is designed to run with Python 2.7, 3.4, 3.5, 3.6, and 3.7 This extension is designed to run with Python 2.7, 3.4, 3.5, 3.6, and 3.7
on common platforms (Linux, Windows, and OS X). On PyPy (both PyPy2 and PyPy3) we support version 6.0.0 and above. on common platforms (Linux, Windows, and OS X). On PyPy (both PyPy2 and PyPy3) we support version 6.0.0 and above.
x86 and x86_64 are well-tested on Windows. Only x86_64 is well-tested on Linux and macOS. x86 and x86_64 are well-tested on Windows. Only x86_64 is well-tested on Linux and macOS.
If you find this project useful and would like to thank the author for If you find this project useful and would like to thank the author for
their work, consider donating some money. Any amount is appreciated. their work, consider donating some money. Any amount is appreciated.
.. image:: https://www.paypalobjects.com/en_US/i/btn/btn_donate_LG.gif .. image:: https://www.paypalobjects.com/en_US/i/btn/btn_donate_LG.gif
:target: https://www.paypal.com/cgi-bin/webscr?cmd=_donations&business=gregory%2eszorc%40gmail%2ecom&lc=US&item_name=python%2dzstandard&currency_code=USD&bn=PP%2dDonationsBF%3abtn_donate_LG%2egif%3aNonHosted :target: https://www.paypal.com/cgi-bin/webscr?cmd=_donations&business=gregory%2eszorc%40gmail%2ecom&lc=US&item_name=python%2dzstandard&currency_code=USD&bn=PP%2dDonationsBF%3abtn_donate_LG%2egif%3aNonHosted
:alt: Donate via PayPal :alt: Donate via PayPal
.. |ci-status| image:: https://travis-ci.org/indygreg/python-zstandard.svg?branch=master .. |ci-status| image:: https://dev.azure.com/gregoryszorc/python-zstandard/_apis/build/status/indygreg.python-zstandard?branchName=master
:target: https://travis-ci.org/indygreg/python-zstandard :target: https://dev.azure.com/gregoryszorc/python-zstandard/_apis/build/status/indygreg.python-zstandard?branchName=master
.. |win-ci-status| image:: https://ci.appveyor.com/api/projects/status/github/indygreg/python-zstandard?svg=true
:target: https://ci.appveyor.com/project/indygreg/python-zstandard
:alt: Windows build status

contrib/python-zstandard/c-ext/compressionparams.c

/** /**
* Copyright (c) 2016-present, Gregory Szorc * Copyright (c) 2016-present, Gregory Szorc
* All rights reserved. * All rights reserved.
* *
* This software may be modified and distributed under the terms * This software may be modified and distributed under the terms
* of the BSD license. See the LICENSE file for details. * of the BSD license. See the LICENSE file for details.
*/ */
#include "python-zstandard.h" #include "python-zstandard.h"
extern PyObject* ZstdError; extern PyObject* ZstdError;
int set_parameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value) { int set_parameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value) {
size_t zresult = ZSTD_CCtxParam_setParameter(params, param, value); size_t zresult = ZSTD_CCtxParams_setParameter(params, param, value);
if (ZSTD_isError(zresult)) { if (ZSTD_isError(zresult)) {
PyErr_Format(ZstdError, "unable to set compression context parameter: %s", PyErr_Format(ZstdError, "unable to set compression context parameter: %s",
ZSTD_getErrorName(zresult)); ZSTD_getErrorName(zresult));
return 1; return 1;
} }
return 0; return 0;
} }
#define TRY_SET_PARAMETER(params, param, value) if (set_parameter(params, param, value)) return -1; #define TRY_SET_PARAMETER(params, param, value) if (set_parameter(params, param, value)) return -1;
#define TRY_COPY_PARAMETER(source, dest, param) { \ #define TRY_COPY_PARAMETER(source, dest, param) { \
int result; \ int result; \
size_t zresult = ZSTD_CCtxParam_getParameter(source, param, &result); \ size_t zresult = ZSTD_CCtxParams_getParameter(source, param, &result); \
if (ZSTD_isError(zresult)) { \ if (ZSTD_isError(zresult)) { \
return 1; \ return 1; \
} \ } \
zresult = ZSTD_CCtxParam_setParameter(dest, param, result); \ zresult = ZSTD_CCtxParams_setParameter(dest, param, result); \
if (ZSTD_isError(zresult)) { \ if (ZSTD_isError(zresult)) { \
return 1; \ return 1; \
} \ } \
} }
int set_parameters(ZSTD_CCtx_params* params, ZstdCompressionParametersObject* obj) { int set_parameters(ZSTD_CCtx_params* params, ZstdCompressionParametersObject* obj) {
TRY_COPY_PARAMETER(obj->params, params, ZSTD_c_nbWorkers); TRY_COPY_PARAMETER(obj->params, params, ZSTD_c_nbWorkers);
return 1; return 1;
} }
} }
return set_parameters(params->params, params); return set_parameters(params->params, params);
} }
#define TRY_GET_PARAMETER(params, param, value) { \ #define TRY_GET_PARAMETER(params, param, value) { \
size_t zresult = ZSTD_CCtxParam_getParameter(params, param, value); \ size_t zresult = ZSTD_CCtxParams_getParameter(params, param, value); \
if (ZSTD_isError(zresult)) { \ if (ZSTD_isError(zresult)) { \
PyErr_Format(ZstdError, "unable to retrieve parameter: %s", ZSTD_getErrorName(zresult)); \ PyErr_Format(ZstdError, "unable to retrieve parameter: %s", ZSTD_getErrorName(zresult)); \
return 1; \ return 1; \
} \ } \
} }
int to_cparams(ZstdCompressionParametersObject* params, ZSTD_compressionParameters* cparams) { int to_cparams(ZstdCompressionParametersObject* params, ZSTD_compressionParameters* cparams) {
int value; int value;
PyObject_Del(self); PyObject_Del(self);
} }
#define PARAM_GETTER(name, param) PyObject* ZstdCompressionParameters_get_##name(PyObject* self, void* unused) { \ #define PARAM_GETTER(name, param) PyObject* ZstdCompressionParameters_get_##name(PyObject* self, void* unused) { \
int result; \ int result; \
size_t zresult; \ size_t zresult; \
ZstdCompressionParametersObject* p = (ZstdCompressionParametersObject*)(self); \ ZstdCompressionParametersObject* p = (ZstdCompressionParametersObject*)(self); \
zresult = ZSTD_CCtxParam_getParameter(p->params, param, &result); \ zresult = ZSTD_CCtxParams_getParameter(p->params, param, &result); \
if (ZSTD_isError(zresult)) { \ if (ZSTD_isError(zresult)) { \
PyErr_Format(ZstdError, "unable to get compression parameter: %s", \ PyErr_Format(ZstdError, "unable to get compression parameter: %s", \
ZSTD_getErrorName(zresult)); \ ZSTD_getErrorName(zresult)); \
return NULL; \ return NULL; \
} \ } \
return PyLong_FromLong(result); \ return PyLong_FromLong(result); \
} }

contrib/python-zstandard/c-ext/decompressionreader.c

readAmount < defaultOutSize ? readAmount : defaultOutSize); readAmount < defaultOutSize ? readAmount : defaultOutSize);
if (!readResult) { if (!readResult) {
return NULL; return NULL;
} }
readSize = PyBytes_GET_SIZE(readResult); readSize = PyBytes_GET_SIZE(readResult);
Py_CLEAR(readResult);
/* Empty read means EOF. */ /* Empty read means EOF. */
if (!readSize) { if (!readSize) {
break; break;
} }
readAmount -= readSize; readAmount -= readSize;
} }

contrib/python-zstandard/c-ext/python-zstandard.h

#include "structmember.h" #include "structmember.h"
#define ZSTD_STATIC_LINKING_ONLY #define ZSTD_STATIC_LINKING_ONLY
#define ZDICT_STATIC_LINKING_ONLY #define ZDICT_STATIC_LINKING_ONLY
#include <zstd.h> #include <zstd.h>
#include <zdict.h> #include <zdict.h>
/* Remember to change the string in zstandard/__init__ as well */ /* Remember to change the string in zstandard/__init__ as well */
#define PYTHON_ZSTANDARD_VERSION "0.11.0" #define PYTHON_ZSTANDARD_VERSION "0.12.0"
typedef enum { typedef enum {
compressorobj_flush_finish, compressorobj_flush_finish,
compressorobj_flush_block, compressorobj_flush_block,
} CompressorObj_Flush; } CompressorObj_Flush;
/* /*
Represents a ZstdCompressionParameters type. Represents a ZstdCompressionParameters type.

contrib/python-zstandard/make_cffi.py

'common/pool.c', 'common/pool.c',
'common/threading.c', 'common/threading.c',
'common/xxhash.c', 'common/xxhash.c',
'common/zstd_common.c', 'common/zstd_common.c',
'compress/fse_compress.c', 'compress/fse_compress.c',
'compress/hist.c', 'compress/hist.c',
'compress/huf_compress.c', 'compress/huf_compress.c',
'compress/zstd_compress.c', 'compress/zstd_compress.c',
'compress/zstd_compress_literals.c',
'compress/zstd_compress_sequences.c',
'compress/zstd_double_fast.c', 'compress/zstd_double_fast.c',
'compress/zstd_fast.c', 'compress/zstd_fast.c',
'compress/zstd_lazy.c', 'compress/zstd_lazy.c',
'compress/zstd_ldm.c', 'compress/zstd_ldm.c',
'compress/zstd_opt.c', 'compress/zstd_opt.c',
'compress/zstdmt_compress.c', 'compress/zstdmt_compress.c',
'decompress/huf_decompress.c', 'decompress/huf_decompress.c',
'decompress/zstd_ddict.c', 'decompress/zstd_ddict.c',
lines.append(l) lines.append(l)
fd, input_file = tempfile.mkstemp(suffix='.h') fd, input_file = tempfile.mkstemp(suffix='.h')
os.write(fd, b''.join(lines)) os.write(fd, b''.join(lines))
os.close(fd) os.close(fd)
try: try:
process = subprocess.Popen(args + [input_file], stdout=subprocess.PIPE) env = dict(os.environ)
if getattr(compiler, '_paths', None):
env['PATH'] = compiler._paths
process = subprocess.Popen(args + [input_file], stdout=subprocess.PIPE,
env=env)
output = process.communicate()[0] output = process.communicate()[0]
ret = process.poll() ret = process.poll()
if ret: if ret:
raise Exception('preprocessor exited with error') raise Exception('preprocessor exited with error')
return output return output
finally: finally:
os.unlink(input_file) os.unlink(input_file)

contrib/python-zstandard/setup.py

author_email='gregory.szorc@gmail.com', author_email='gregory.szorc@gmail.com',
license='BSD', license='BSD',
classifiers=[ classifiers=[
'Development Status :: 4 - Beta', 'Development Status :: 4 - Beta',
'Intended Audience :: Developers', 'Intended Audience :: Developers',
'License :: OSI Approved :: BSD License', 'License :: OSI Approved :: BSD License',
'Programming Language :: C', 'Programming Language :: C',
'Programming Language :: Python :: 2.7', 'Programming Language :: Python :: 2.7',
'Programming Language :: Python :: 3.4',
'Programming Language :: Python :: 3.5', 'Programming Language :: Python :: 3.5',
'Programming Language :: Python :: 3.6', 'Programming Language :: Python :: 3.6',
'Programming Language :: Python :: 3.7', 'Programming Language :: Python :: 3.7',
], ],
keywords='zstandard zstd compression', keywords='zstandard zstd compression',
packages=['zstandard'], packages=['zstandard'],
ext_modules=extensions, ext_modules=extensions,
test_suite='tests', test_suite='tests',
install_requires=install_requires, install_requires=install_requires,
) )

contrib/python-zstandard/setup_zstd.py

'common/fse_decompress.c', 'common/fse_decompress.c',
'common/pool.c', 'common/pool.c',
'common/threading.c', 'common/threading.c',
'common/xxhash.c', 'common/xxhash.c',
'common/zstd_common.c', 'common/zstd_common.c',
'compress/fse_compress.c', 'compress/fse_compress.c',
'compress/hist.c', 'compress/hist.c',
'compress/huf_compress.c', 'compress/huf_compress.c',
'compress/zstd_compress_literals.c',
'compress/zstd_compress_sequences.c',
'compress/zstd_compress.c', 'compress/zstd_compress.c',
'compress/zstd_double_fast.c', 'compress/zstd_double_fast.c',
'compress/zstd_fast.c', 'compress/zstd_fast.c',
'compress/zstd_lazy.c', 'compress/zstd_lazy.c',
'compress/zstd_ldm.c', 'compress/zstd_ldm.c',
'compress/zstd_opt.c', 'compress/zstd_opt.c',
'compress/zstdmt_compress.c', 'compress/zstdmt_compress.c',
'decompress/huf_decompress.c', 'decompress/huf_decompress.c',

contrib/python-zstandard/tests/test_compressor.py

for i in range(128): for i in range(128):
samples.append(b'foo' * 64) samples.append(b'foo' * 64)
samples.append(b'bar' * 64) samples.append(b'bar' * 64)
samples.append(b'foobar' * 64) samples.append(b'foobar' * 64)
d = zstd.train_dictionary(8192, samples) d = zstd.train_dictionary(8192, samples)
h = hashlib.sha1(d.as_bytes()).hexdigest() h = hashlib.sha1(d.as_bytes()).hexdigest()
self.assertEqual(h, '88ca0d38332aff379d4ced166a51c280a7679aad') self.assertEqual(h, '7a2e59a876db958f74257141045af8f912e00d4e')
buffer = NonClosingBytesIO() buffer = NonClosingBytesIO()
cctx = zstd.ZstdCompressor(level=9, dict_data=d) cctx = zstd.ZstdCompressor(level=9, dict_data=d)
with cctx.stream_writer(buffer) as compressor: with cctx.stream_writer(buffer) as compressor:
self.assertEqual(compressor.write(b'foo'), 0) self.assertEqual(compressor.write(b'foo'), 0)
self.assertEqual(compressor.write(b'bar'), 0) self.assertEqual(compressor.write(b'bar'), 0)
self.assertEqual(compressor.write(b'foo' * 16384), 0) self.assertEqual(compressor.write(b'foo' * 16384), 0)
compressed = buffer.getvalue() compressed = buffer.getvalue()
params = zstd.get_frame_parameters(compressed) params = zstd.get_frame_parameters(compressed)
self.assertEqual(params.content_size, zstd.CONTENTSIZE_UNKNOWN) self.assertEqual(params.content_size, zstd.CONTENTSIZE_UNKNOWN)
self.assertEqual(params.window_size, 2097152) self.assertEqual(params.window_size, 2097152)
self.assertEqual(params.dict_id, d.dict_id()) self.assertEqual(params.dict_id, d.dict_id())
self.assertFalse(params.has_checksum) self.assertFalse(params.has_checksum)
h = hashlib.sha1(compressed).hexdigest() h = hashlib.sha1(compressed).hexdigest()
self.assertEqual(h, '8703b4316f274d26697ea5dd480f29c08e85d940') self.assertEqual(h, '0a7c05635061f58039727cdbe76388c6f4cfef06')
source = b'foo' + b'bar' + (b'foo' * 16384) source = b'foo' + b'bar' + (b'foo' * 16384)
dctx = zstd.ZstdDecompressor(dict_data=d) dctx = zstd.ZstdDecompressor(dict_data=d)
self.assertEqual(dctx.decompress(compressed, max_output_size=len(source)), self.assertEqual(dctx.decompress(compressed, max_output_size=len(source)),
source) source)
params = zstd.get_frame_parameters(compressed) params = zstd.get_frame_parameters(compressed)
self.assertEqual(params.content_size, zstd.CONTENTSIZE_UNKNOWN) self.assertEqual(params.content_size, zstd.CONTENTSIZE_UNKNOWN)
self.assertEqual(params.window_size, 1048576) self.assertEqual(params.window_size, 1048576)
self.assertEqual(params.dict_id, 0) self.assertEqual(params.dict_id, 0)
self.assertFalse(params.has_checksum) self.assertFalse(params.has_checksum)
h = hashlib.sha1(compressed).hexdigest() h = hashlib.sha1(compressed).hexdigest()
self.assertEqual(h, '2a8111d72eb5004cdcecbdac37da9f26720d30ef') self.assertEqual(h, 'dd4bb7d37c1a0235b38a2f6b462814376843ef0b')
def test_write_checksum(self): def test_write_checksum(self):
no_checksum = NonClosingBytesIO() no_checksum = NonClosingBytesIO()
cctx = zstd.ZstdCompressor(level=1) cctx = zstd.ZstdCompressor(level=1)
with cctx.stream_writer(no_checksum) as compressor: with cctx.stream_writer(no_checksum) as compressor:
self.assertEqual(compressor.write(b'foobar'), 0) self.assertEqual(compressor.write(b'foobar'), 0)
with_checksum = NonClosingBytesIO() with_checksum = NonClosingBytesIO()

contrib/python-zstandard/tests/test_data_structures.py

chain_log=16, chain_log=16,
hash_log=17, hash_log=17,
search_log=1, search_log=1,
min_match=5, min_match=5,
target_length=16, target_length=16,
strategy=zstd.STRATEGY_DFAST) strategy=zstd.STRATEGY_DFAST)
# 32-bit has slightly different values from 64-bit. # 32-bit has slightly different values from 64-bit.
self.assertAlmostEqual(p.estimated_compression_context_size(), 1294072, self.assertAlmostEqual(p.estimated_compression_context_size(), 1294144,
delta=250) delta=250)
def test_strategy(self): def test_strategy(self):
with self.assertRaisesRegexp(ValueError, 'cannot specify both compression_strategy'): with self.assertRaisesRegexp(ValueError, 'cannot specify both compression_strategy'):
zstd.ZstdCompressionParameters(strategy=0, compression_strategy=0) zstd.ZstdCompressionParameters(strategy=0, compression_strategy=0)
p = zstd.ZstdCompressionParameters(strategy=2) p = zstd.ZstdCompressionParameters(strategy=2)
self.assertEqual(p.compression_strategy, 2) self.assertEqual(p.compression_strategy, 2)

contrib/python-zstandard/tests/test_module_attributes.py

from __future__ import unicode_literals from __future__ import unicode_literals
import unittest import unittest
import zstandard as zstd import zstandard as zstd
from . common import ( from . common import (
make_cffi, make_cffi,
) )
@make_cffi @make_cffi
class TestModuleAttributes(unittest.TestCase): class TestModuleAttributes(unittest.TestCase):
def test_version(self): def test_version(self):
self.assertEqual(zstd.ZSTD_VERSION, (1, 3, 8)) self.assertEqual(zstd.ZSTD_VERSION, (1, 4, 3))
self.assertEqual(zstd.__version__, '0.11.0') self.assertEqual(zstd.__version__, '0.12.0')
def test_constants(self): def test_constants(self):
self.assertEqual(zstd.MAX_COMPRESSION_LEVEL, 22) self.assertEqual(zstd.MAX_COMPRESSION_LEVEL, 22)
self.assertEqual(zstd.FRAME_HEADER, b'\x28\xb5\x2f\xfd') self.assertEqual(zstd.FRAME_HEADER, b'\x28\xb5\x2f\xfd')
def test_hasattr(self): def test_hasattr(self):
attrs = ( attrs = (
'CONTENTSIZE_UNKNOWN', 'CONTENTSIZE_UNKNOWN',

contrib/python-zstandard/tests/test_train_dictionary.py

import struct import struct
import sys import sys
import unittest import unittest
import zstandard as zstd import zstandard as zstd
from . common import ( from . common import (
generate_samples, generate_samples,
make_cffi, make_cffi,
random_input_data,
) )
if sys.version_info[0] >= 3: if sys.version_info[0] >= 3:
int_type = int int_type = int
else: else:
int_type = long int_type = long
@make_cffi @make_cffi
class TestTrainDictionary(unittest.TestCase): class TestTrainDictionary(unittest.TestCase):
def test_no_args(self): def test_no_args(self):
with self.assertRaises(TypeError): with self.assertRaises(TypeError):
zstd.train_dictionary() zstd.train_dictionary()
def test_bad_args(self): def test_bad_args(self):
with self.assertRaises(TypeError): with self.assertRaises(TypeError):
zstd.train_dictionary(8192, u'foo') zstd.train_dictionary(8192, u'foo')
with self.assertRaises(ValueError): with self.assertRaises(ValueError):
zstd.train_dictionary(8192, [u'foo']) zstd.train_dictionary(8192, [u'foo'])
def test_no_params(self): def test_no_params(self):
d = zstd.train_dictionary(8192, generate_samples()) d = zstd.train_dictionary(8192, random_input_data())
self.assertIsInstance(d.dict_id(), int_type) self.assertIsInstance(d.dict_id(), int_type)
# The dictionary ID may be different across platforms. # The dictionary ID may be different across platforms.
expected = b'\x37\xa4\x30\xec' + struct.pack('<I', d.dict_id()) expected = b'\x37\xa4\x30\xec' + struct.pack('<I', d.dict_id())
data = d.as_bytes() data = d.as_bytes()
self.assertEqual(data[0:8], expected) self.assertEqual(data[0:8], expected)

contrib/python-zstandard/zstandard/__init__.py

elif _module_policy == 'cffi': elif _module_policy == 'cffi':
from .cffi import * from .cffi import *
backend = 'cffi' backend = 'cffi'
else: else:
raise ImportError('unknown module import policy: %s; use default, cffi_fallback, ' raise ImportError('unknown module import policy: %s; use default, cffi_fallback, '
'cext, or cffi' % _module_policy) 'cext, or cffi' % _module_policy)
# Keep this in sync with python-zstandard.h. # Keep this in sync with python-zstandard.h.
__version__ = '0.11.0' __version__ = '0.12.0'

contrib/python-zstandard/zstandard/cffi.py

CompressionParameters = ZstdCompressionParameters CompressionParameters = ZstdCompressionParameters
def estimate_decompression_context_size(): def estimate_decompression_context_size():
return lib.ZSTD_estimateDCtxSize() return lib.ZSTD_estimateDCtxSize()
def _set_compression_parameter(params, param, value): def _set_compression_parameter(params, param, value):
zresult = lib.ZSTD_CCtxParam_setParameter(params, param, value) zresult = lib.ZSTD_CCtxParams_setParameter(params, param, value)
if lib.ZSTD_isError(zresult): if lib.ZSTD_isError(zresult):
raise ZstdError('unable to set compression context parameter: %s' % raise ZstdError('unable to set compression context parameter: %s' %
_zstd_error(zresult)) _zstd_error(zresult))
def _get_compression_parameter(params, param): def _get_compression_parameter(params, param):
result = ffi.new('int *') result = ffi.new('int *')
zresult = lib.ZSTD_CCtxParam_getParameter(params, param, result) zresult = lib.ZSTD_CCtxParams_getParameter(params, param, result)
if lib.ZSTD_isError(zresult): if lib.ZSTD_isError(zresult):
raise ZstdError('unable to get compression context parameter: %s' % raise ZstdError('unable to get compression context parameter: %s' %
_zstd_error(zresult)) _zstd_error(zresult))
return result[0] return result[0]
class ZstdCompressionWriter(object): class ZstdCompressionWriter(object):

contrib/python-zstandard/zstd.c

However, downstream redistributors or packagers may unbundle zstd However, downstream redistributors or packagers may unbundle zstd
from python-zstandard. This can result in a mismatch between zstd from python-zstandard. This can result in a mismatch between zstd
versions and API semantics. This essentially "voids the warranty" versions and API semantics. This essentially "voids the warranty"
of python-zstandard and may cause undefined behavior. of python-zstandard and may cause undefined behavior.
We detect this mismatch here and refuse to load the module if this We detect this mismatch here and refuse to load the module if this
scenario is detected. scenario is detected.
*/ */
if (ZSTD_VERSION_NUMBER != 10308 || ZSTD_versionNumber() != 10308) { if (ZSTD_VERSION_NUMBER != 10403 || ZSTD_versionNumber() != 10403) {
PyErr_SetString(PyExc_ImportError, "zstd C API mismatch; Python bindings not compiled against expected zstd version"); PyErr_SetString(PyExc_ImportError, "zstd C API mismatch; Python bindings not compiled against expected zstd version");
return; return;
} }
bufferutil_module_init(m); bufferutil_module_init(m);
compressionparams_module_init(m); compressionparams_module_init(m);
compressiondict_module_init(m); compressiondict_module_init(m);
compressobj_module_init(m); compressobj_module_init(m);
memcpy(PyBytes_AS_STRING(tmp), PyBytes_AS_STRING(*obj), memcpy(PyBytes_AS_STRING(tmp), PyBytes_AS_STRING(*obj),
PyBytes_GET_SIZE(*obj)); PyBytes_GET_SIZE(*obj));
Py_DECREF(*obj); Py_DECREF(*obj);
*obj = tmp; *obj = tmp;
return 0; return 0;
} }
No newline at end of file No newline at end of file

contrib/python-zstandard/zstd/common/bitstream.h

#include "error_private.h" /* error codes and messages */ #include "error_private.h" /* error codes and messages */
/*========================================= /*=========================================
* Target specific * Target specific
=========================================*/ =========================================*/
#if defined(__BMI__) && defined(__GNUC__) #if defined(__BMI__) && defined(__GNUC__)
# include <immintrin.h> /* support for bextr (experimental) */ # include <immintrin.h> /* support for bextr (experimental) */
#elif defined(__ICCARM__)
# include <intrinsics.h>
#endif #endif
#define STREAM_ACCUMULATOR_MIN_32 25 #define STREAM_ACCUMULATOR_MIN_32 25
#define STREAM_ACCUMULATOR_MIN_64 57 #define STREAM_ACCUMULATOR_MIN_64 57
#define STREAM_ACCUMULATOR_MIN ((U32)(MEM_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64)) #define STREAM_ACCUMULATOR_MIN ((U32)(MEM_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64))
/*-****************************************** /*-******************************************
assert(val != 0); assert(val != 0);
{ {
# if defined(_MSC_VER) /* Visual */ # if defined(_MSC_VER) /* Visual */
unsigned long r=0; unsigned long r=0;
_BitScanReverse ( &r, val ); _BitScanReverse ( &r, val );
return (unsigned) r; return (unsigned) r;
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */ # elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */
return 31 - __builtin_clz (val); return 31 - __builtin_clz (val);
# elif defined(__ICCARM__) /* IAR Intrinsic */
return 31 - __CLZ(val);
# else /* Software version */ # else /* Software version */
static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29,
11, 14, 16, 18, 22, 25, 3, 30, 11, 14, 16, 18, 22, 25, 3, 30,
8, 12, 20, 28, 15, 17, 24, 7, 8, 12, 20, 28, 15, 17, 24, 7,
19, 27, 23, 6, 26, 5, 4, 31 }; 19, 27, 23, 6, 26, 5, 4, 31 };
U32 v = val; U32 v = val;
v |= v >> 1; v |= v >> 1;
v |= v >> 2; v |= v >> 2;

contrib/python-zstandard/zstd/common/compiler.h

#if !defined(ZSTD_NO_INLINE) #if !defined(ZSTD_NO_INLINE)
#if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ #if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */
# define INLINE_KEYWORD inline # define INLINE_KEYWORD inline
#else #else
# define INLINE_KEYWORD # define INLINE_KEYWORD
#endif #endif
#if defined(__GNUC__) #if defined(__GNUC__) || defined(__ICCARM__)
# define FORCE_INLINE_ATTR __attribute__((always_inline)) # define FORCE_INLINE_ATTR __attribute__((always_inline))
#elif defined(_MSC_VER) #elif defined(_MSC_VER)
# define FORCE_INLINE_ATTR __forceinline # define FORCE_INLINE_ATTR __forceinline
#else #else
# define FORCE_INLINE_ATTR # define FORCE_INLINE_ATTR
#endif #endif
#else #else
#define INLINE_KEYWORD #define INLINE_KEYWORD
#define FORCE_INLINE_ATTR #define FORCE_INLINE_ATTR
#endif #endif
/** /**
* FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
* parameters. They must be inlined for the compiler to elimininate the constant * parameters. They must be inlined for the compiler to eliminate the constant
* branches. * branches.
*/ */
#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR #define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
/** /**
* HINT_INLINE is used to help the compiler generate better code. It is *not* * 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 * used for "templates", so it can be tweaked based on the compilers
* performance. * performance.
* *
* gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the * gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the
* always_inline attribute. * always_inline attribute.
* *
* clang up to 5.0.0 (trunk) benefit tremendously from the always_inline * clang up to 5.0.0 (trunk) benefit tremendously from the always_inline
* attribute. * attribute.
*/ */
#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5 #if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5
# define HINT_INLINE static INLINE_KEYWORD # define HINT_INLINE static INLINE_KEYWORD
#else #else
# define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR # define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR
#endif #endif
/* force no inlining */ /* force no inlining */
#ifdef _MSC_VER #ifdef _MSC_VER
# define FORCE_NOINLINE static __declspec(noinline) # define FORCE_NOINLINE static __declspec(noinline)
#else #else
# ifdef __GNUC__ # if defined(__GNUC__) || defined(__ICCARM__)
# define FORCE_NOINLINE static __attribute__((__noinline__)) # define FORCE_NOINLINE static __attribute__((__noinline__))
# else # else
# define FORCE_NOINLINE static # define FORCE_NOINLINE static
# endif # endif
#endif #endif
/* target attribute */ /* target attribute */
#ifndef __has_attribute #ifndef __has_attribute
#define __has_attribute(x) 0 /* Compatibility with non-clang compilers. */ #define __has_attribute(x) 0 /* Compatibility with non-clang compilers. */
#endif #endif
#if defined(__GNUC__) #if defined(__GNUC__) || defined(__ICCARM__)
# define TARGET_ATTRIBUTE(target) __attribute__((__target__(target))) # define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
#else #else
# define TARGET_ATTRIBUTE(target) # define TARGET_ATTRIBUTE(target)
#endif #endif
/* Enable runtime BMI2 dispatch based on the CPU. /* Enable runtime BMI2 dispatch based on the CPU.
* Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default. * Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default.
*/ */
const char* const _ptr = (const char*)(p); \ const char* const _ptr = (const char*)(p); \
size_t const _size = (size_t)(s); \ size_t const _size = (size_t)(s); \
size_t _pos; \ size_t _pos; \
for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \ for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \
PREFETCH_L2(_ptr + _pos); \ PREFETCH_L2(_ptr + _pos); \
} \ } \
} }
/* vectorization */
#if !defined(__clang__) && defined(__GNUC__)
# define DONT_VECTORIZE __attribute__((optimize("no-tree-vectorize")))
#else
# define DONT_VECTORIZE
#endif
/* disable warnings */ /* disable warnings */
#ifdef _MSC_VER /* Visual Studio */ #ifdef _MSC_VER /* Visual Studio */
# include <intrin.h> /* For Visual 2005 */ # include <intrin.h> /* For Visual 2005 */
# pragma warning(disable : 4100) /* disable: C4100: unreferenced formal parameter */ # pragma warning(disable : 4100) /* disable: C4100: unreferenced formal parameter */
# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
# pragma warning(disable : 4204) /* disable: C4204: non-constant aggregate initializer */ # pragma warning(disable : 4204) /* disable: C4204: non-constant aggregate initializer */
# pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */ # pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */
# pragma warning(disable : 4324) /* disable: C4324: padded structure */ # pragma warning(disable : 4324) /* disable: C4324: padded structure */
#endif #endif
#endif /* ZSTD_COMPILER_H */ #endif /* ZSTD_COMPILER_H */

contrib/python-zstandard/zstd/common/fse.h

/**< build a fake FSE_DTable, designed to always generate the same symbolValue */ /**< build a fake FSE_DTable, designed to always generate the same symbolValue */
size_t FSE_decompress_wksp(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, FSE_DTable* workSpace, unsigned maxLog); 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)` */ /**< same as FSE_decompress(), using an externally allocated `workSpace` produced with `FSE_DTABLE_SIZE_U32(maxLog)` */
typedef enum { typedef enum {
FSE_repeat_none, /**< Cannot use the previous table */ FSE_repeat_none, /**< Cannot use the previous table */
FSE_repeat_check, /**< Can use the previous table but it must be checked */ FSE_repeat_check, /**< Can use the previous table but it must be checked */
FSE_repeat_valid /**< Can use the previous table and it is asumed to be valid */ FSE_repeat_valid /**< Can use the previous table and it is assumed to be valid */
} FSE_repeat; } FSE_repeat;
/* ***************************************** /* *****************************************
* FSE symbol compression API * FSE symbol compression API
*******************************************/ *******************************************/
/*! /*!
This API consists of small unitary functions, which highly benefit from being inlined. This API consists of small unitary functions, which highly benefit from being inlined.
Hence their body are included in next section. Hence their body are included in next section.

contrib/python-zstandard/zstd/common/mem.h

* It can generate buggy code on targets depending on alignment. * It can generate buggy code on targets depending on alignment.
* In some circumstances, it's the only known way to get the most performance (i.e. 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. * See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2) * Prefer these methods in priority order (0 > 1 > 2)
*/ */
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ #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__) ) # 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 # define MEM_FORCE_MEMORY_ACCESS 2
# elif defined(__INTEL_COMPILER) || defined(__GNUC__) # elif defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
# define MEM_FORCE_MEMORY_ACCESS 1 # define MEM_FORCE_MEMORY_ACCESS 1
# endif # endif
#endif #endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; } MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; } MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; }
MEM_STATIC unsigned MEM_isLittleEndian(void) MEM_STATIC unsigned MEM_isLittleEndian(void)

contrib/python-zstandard/zstd/common/threading.c

/** /**
* Copyright (c) 2016 Tino Reichardt * Copyright (c) 2016 Tino Reichardt
* All rights reserved. * All rights reserved.
* *
* This source code is licensed under both the BSD-style license (found in the * 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 * 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). * in the COPYING file in the root directory of this source tree).
* *
* You can contact the author at: * You can contact the author at:
* - zstdmt source repository: https://github.com/mcmilk/zstdmt * - zstdmt source repository: https://github.com/mcmilk/zstdmt
*/ */
/** /**
* This file will hold wrapper for systems, which do not support pthreads * This file will hold wrapper for systems, which do not support pthreads
*/ */
/* create fake symbol to avoid empty trnaslation unit warning */ /* create fake symbol to avoid empty translation unit warning */
int g_ZSTD_threading_useles_symbol; int g_ZSTD_threading_useless_symbol;
#if defined(ZSTD_MULTITHREAD) && defined(_WIN32) #if defined(ZSTD_MULTITHREAD) && defined(_WIN32)
/** /**
* Windows minimalist Pthread Wrapper, based on : * Windows minimalist Pthread Wrapper, based on :
* http://www.cse.wustl.edu/~schmidt/win32-cv-1.html * http://www.cse.wustl.edu/~schmidt/win32-cv-1.html
*/ */

contrib/python-zstandard/zstd/common/xxhash.c

* But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6) * But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
* See http://stackoverflow.com/a/32095106/646947 for details. * See http://stackoverflow.com/a/32095106/646947 for details.
* Prefer these methods in priority order (0 > 1 > 2) * Prefer these methods in priority order (0 > 1 > 2)
*/ */
#ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ #ifndef XXH_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__) ) # 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 XXH_FORCE_MEMORY_ACCESS 2 # define XXH_FORCE_MEMORY_ACCESS 2
# elif (defined(__INTEL_COMPILER) && !defined(WIN32)) || \ # elif (defined(__INTEL_COMPILER) && !defined(WIN32)) || \
(defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) || \
defined(__ICCARM__)
# define XXH_FORCE_MEMORY_ACCESS 1 # define XXH_FORCE_MEMORY_ACCESS 1
# endif # endif
#endif #endif
/*!XXH_ACCEPT_NULL_INPUT_POINTER : /*!XXH_ACCEPT_NULL_INPUT_POINTER :
* If the input pointer is a null pointer, xxHash default behavior is to trigger a memory access error, since it is a bad pointer. * If the input pointer is a null pointer, xxHash default behavior is to trigger a memory access error, since it is a bad pointer.
* When this option is enabled, xxHash output for null input pointers will be the same as a null-length input. * When this option is enabled, xxHash output for null input pointers will be the same as a null-length input.
* By default, this option is disabled. To enable it, uncomment below define : * By default, this option is disabled. To enable it, uncomment below define :
*/ */
/* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */ /* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */
/*!XXH_FORCE_NATIVE_FORMAT : /*!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. * 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. * 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. * to improve speed for Big-endian CPU.
* This option has no impact on Little_Endian CPU. * This option has no impact on Little_Endian CPU.
*/ */
#ifndef XXH_FORCE_NATIVE_FORMAT /* can be defined externally */ #ifndef XXH_FORCE_NATIVE_FORMAT /* can be defined externally */
# define XXH_FORCE_NATIVE_FORMAT 0 # define XXH_FORCE_NATIVE_FORMAT 0
#endif #endif
/*!XXH_FORCE_ALIGN_CHECK : /*!XXH_FORCE_ALIGN_CHECK :
* Compiler Specific Options * Compiler Specific Options
***************************************/ ***************************************/
#if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ #if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */
# define INLINE_KEYWORD inline # define INLINE_KEYWORD inline
#else #else
# define INLINE_KEYWORD # define INLINE_KEYWORD
#endif #endif
#if defined(__GNUC__) #if defined(__GNUC__) || defined(__ICCARM__)
# define FORCE_INLINE_ATTR __attribute__((always_inline)) # define FORCE_INLINE_ATTR __attribute__((always_inline))
#elif defined(_MSC_VER) #elif defined(_MSC_VER)
# define FORCE_INLINE_ATTR __forceinline # define FORCE_INLINE_ATTR __forceinline
#else #else
# define FORCE_INLINE_ATTR # define FORCE_INLINE_ATTR
#endif #endif
#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR #define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
******************************************/ ******************************************/
#define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) #define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
/* Note : although _rotl exists for minGW (GCC under windows), performance seems poor */ /* Note : although _rotl exists for minGW (GCC under windows), performance seems poor */
#if defined(_MSC_VER) #if defined(_MSC_VER)
# define XXH_rotl32(x,r) _rotl(x,r) # define XXH_rotl32(x,r) _rotl(x,r)
# define XXH_rotl64(x,r) _rotl64(x,r) # define XXH_rotl64(x,r) _rotl64(x,r)
#else #else
#if defined(__ICCARM__)
# include <intrinsics.h>
# define XXH_rotl32(x,r) __ROR(x,(32 - r))
#else
# define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r))) # define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r)))
#endif
# define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r))) # define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r)))
#endif #endif
#if defined(_MSC_VER) /* Visual Studio */ #if defined(_MSC_VER) /* Visual Studio */
# define XXH_swap32 _byteswap_ulong # define XXH_swap32 _byteswap_ulong
# define XXH_swap64 _byteswap_uint64 # define XXH_swap64 _byteswap_uint64
#elif GCC_VERSION >= 403 #elif GCC_VERSION >= 403
# define XXH_swap32 __builtin_bswap32 # define XXH_swap32 __builtin_bswap32

contrib/python-zstandard/zstd/common/zstd_internal.h

#include "fse.h" #include "fse.h"
#define HUF_STATIC_LINKING_ONLY #define HUF_STATIC_LINKING_ONLY
#include "huf.h" #include "huf.h"
#ifndef XXH_STATIC_LINKING_ONLY #ifndef XXH_STATIC_LINKING_ONLY
# define XXH_STATIC_LINKING_ONLY /* XXH64_state_t */ # define XXH_STATIC_LINKING_ONLY /* XXH64_state_t */
#endif #endif
#include "xxhash.h" /* XXH_reset, update, digest */ #include "xxhash.h" /* XXH_reset, update, digest */
#if defined (__cplusplus) #if defined (__cplusplus)
extern "C" { extern "C" {
#endif #endif
/* ---- static assert (debug) --- */ /* ---- static assert (debug) --- */
#define ZSTD_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c) #define ZSTD_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)
#define ZSTD_isError ERR_isError /* for inlining */ #define ZSTD_isError ERR_isError /* for inlining */
#define FSE_isError ERR_isError #define FSE_isError ERR_isError
#define HUF_isError ERR_isError #define HUF_isError ERR_isError
/*-************************************* /*-*************************************
* shared macros * shared macros
***************************************/ ***************************************/
#undef MIN #undef MIN
#undef MAX #undef MAX
#define MIN(a,b) ((a)<(b) ? (a) : (b)) #define MIN(a,b) ((a)<(b) ? (a) : (b))
#define MAX(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 * Common constants
***************************************/ ***************************************/
#define ZSTD_OPT_NUM (1<<12) #define ZSTD_OPT_NUM (1<<12)
#define ZSTD_REP_NUM 3 /* number of repcodes */ #define ZSTD_REP_NUM 3 /* number of repcodes */
#define OF_DEFAULTNORMLOG 5 /* for static allocation */ #define OF_DEFAULTNORMLOG 5 /* for static allocation */
static const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG; static const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG;
/*-******************************************* /*-*******************************************
* Shared functions to include for inlining * Shared functions to include for inlining
*********************************************/ *********************************************/
static void ZSTD_copy8(void* dst, const void* src) { memcpy(dst, src, 8); } static void ZSTD_copy8(void* dst, const void* src) { memcpy(dst, src, 8); }
#define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; } #define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
static void ZSTD_copy16(void* dst, const void* src) { memcpy(dst, src, 16); }
#define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; }
#define WILDCOPY_OVERLENGTH 8
#define VECLEN 16
typedef enum {
ZSTD_no_overlap,
ZSTD_overlap_src_before_dst,
/* ZSTD_overlap_dst_before_src, */
} ZSTD_overlap_e;
/*! ZSTD_wildcopy() : /*! ZSTD_wildcopy() :
* custom version of memcpy(), can overwrite up to WILDCOPY_OVERLENGTH bytes (if length==0) */ * custom version of memcpy(), can overwrite up to WILDCOPY_OVERLENGTH bytes (if length==0) */
#define WILDCOPY_OVERLENGTH 8 MEM_STATIC FORCE_INLINE_ATTR DONT_VECTORIZE
MEM_STATIC void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length) void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e ovtype)
{ {
ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
const BYTE* ip = (const BYTE*)src; const BYTE* ip = (const BYTE*)src;
BYTE* op = (BYTE*)dst; BYTE* op = (BYTE*)dst;
BYTE* const oend = op + length; BYTE* const oend = op + length;
assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff < -8));
if (length < VECLEN || (ovtype == ZSTD_overlap_src_before_dst && diff < VECLEN)) {
do do
COPY8(op, ip) COPY8(op, ip)
while (op < oend); while (op < oend);
} }
else {
if ((length & 8) == 0)
COPY8(op, ip);
do {
COPY16(op, ip);
}
while (op < oend);
}
}
/*! ZSTD_wildcopy_16min() :
* same semantics as ZSTD_wilcopy() except guaranteed to be able to copy 16 bytes at the start */
MEM_STATIC FORCE_INLINE_ATTR DONT_VECTORIZE
void ZSTD_wildcopy_16min(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e ovtype)
{
ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
const BYTE* ip = (const BYTE*)src;
BYTE* op = (BYTE*)dst;
BYTE* const oend = op + length;
assert(length >= 8);
assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff < -8));
if (ovtype == ZSTD_overlap_src_before_dst && diff < VECLEN) {
do
COPY8(op, ip)
while (op < oend);
}
else {
if ((length & 8) == 0)
COPY8(op, ip);
do {
COPY16(op, ip);
}
while (op < oend);
}
}
MEM_STATIC void ZSTD_wildcopy_e(void* dst, const void* src, void* dstEnd) /* should be faster for decoding, but strangely, not verified on all platform */ MEM_STATIC void ZSTD_wildcopy_e(void* dst, const void* src, void* dstEnd) /* should be faster for decoding, but strangely, not verified on all platform */
{ {
const BYTE* ip = (const BYTE*)src; const BYTE* ip = (const BYTE*)src;
BYTE* op = (BYTE*)dst; BYTE* op = (BYTE*)dst;
BYTE* const oend = (BYTE*)dstEnd; BYTE* const oend = (BYTE*)dstEnd;
do do
COPY8(op, ip) COPY8(op, ip)
BYTE* mlCode; BYTE* mlCode;
BYTE* ofCode; BYTE* ofCode;
size_t maxNbSeq; size_t maxNbSeq;
size_t maxNbLit; size_t maxNbLit;
U32 longLengthID; /* 0 == no longLength; 1 == Lit.longLength; 2 == Match.longLength; */ U32 longLengthID; /* 0 == no longLength; 1 == Lit.longLength; 2 == Match.longLength; */
U32 longLengthPos; U32 longLengthPos;
} seqStore_t; } seqStore_t;
/**
* 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 */ 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) */ void ZSTD_seqToCodes(const seqStore_t* seqStorePtr); /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
/* custom memory allocation functions */ /* custom memory allocation functions */
void* ZSTD_malloc(size_t size, ZSTD_customMem customMem); void* ZSTD_malloc(size_t size, ZSTD_customMem customMem);
void* ZSTD_calloc(size_t size, ZSTD_customMem customMem); void* ZSTD_calloc(size_t size, ZSTD_customMem customMem);
void ZSTD_free(void* ptr, ZSTD_customMem customMem); void ZSTD_free(void* ptr, ZSTD_customMem customMem);
MEM_STATIC U32 ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus */ MEM_STATIC U32 ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus */
{ {
assert(val != 0); assert(val != 0);
{ {
# if defined(_MSC_VER) /* Visual */ # if defined(_MSC_VER) /* Visual */
unsigned long r=0; unsigned long r=0;
_BitScanReverse(&r, val); _BitScanReverse(&r, val);
return (unsigned)r; return (unsigned)r;
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */ # elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */
return 31 - __builtin_clz(val); return 31 - __builtin_clz(val);
# elif defined(__ICCARM__) /* IAR Intrinsic */
return 31 - __CLZ(val);
# else /* Software version */ # else /* Software version */
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 }; 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; U32 v = val;
v |= v >> 1; v |= v >> 1;
v |= v >> 2; v |= v >> 2;
v |= v >> 4; v |= v >> 4;
v |= v >> 8; v |= v >> 8;
v |= v >> 16; v |= v >> 16;

contrib/python-zstandard/zstd/compress/fse_compress.c

} } } }
cumul[maxSymbolValue+1] = tableSize+1; cumul[maxSymbolValue+1] = tableSize+1;
} }
/* Spread symbols */ /* Spread symbols */
{ U32 position = 0; { U32 position = 0;
U32 symbol; U32 symbol;
for (symbol=0; symbol<=maxSymbolValue; symbol++) { for (symbol=0; symbol<=maxSymbolValue; symbol++) {
int nbOccurences; int nbOccurrences;
int const freq = normalizedCounter[symbol]; int const freq = normalizedCounter[symbol];
for (nbOccurences=0; nbOccurences<freq; nbOccurences++) { for (nbOccurrences=0; nbOccurrences<freq; nbOccurrences++) {
tableSymbol[position] = (FSE_FUNCTION_TYPE)symbol; tableSymbol[position] = (FSE_FUNCTION_TYPE)symbol;
position = (position + step) & tableMask; position = (position + step) & tableMask;
while (position > highThreshold) while (position > highThreshold)
position = (position + step) & tableMask; /* Low proba area */ position = (position + step) & tableMask; /* Low proba area */
} } } }
assert(position==0); /* Must have initialized all positions */ assert(position==0); /* Must have initialized all positions */
} }

contrib/python-zstandard/zstd/compress/zstd_compress.c

#include "cpu.h" #include "cpu.h"
#include "mem.h" #include "mem.h"
#include "hist.h" /* HIST_countFast_wksp */ #include "hist.h" /* HIST_countFast_wksp */
#define FSE_STATIC_LINKING_ONLY /* FSE_encodeSymbol */ #define FSE_STATIC_LINKING_ONLY /* FSE_encodeSymbol */
#include "fse.h" #include "fse.h"
#define HUF_STATIC_LINKING_ONLY #define HUF_STATIC_LINKING_ONLY
#include "huf.h" #include "huf.h"
#include "zstd_compress_internal.h" #include "zstd_compress_internal.h"
#include "zstd_compress_sequences.h"
#include "zstd_compress_literals.h"
#include "zstd_fast.h" #include "zstd_fast.h"
#include "zstd_double_fast.h" #include "zstd_double_fast.h"
#include "zstd_lazy.h" #include "zstd_lazy.h"
#include "zstd_opt.h" #include "zstd_opt.h"
#include "zstd_ldm.h" #include "zstd_ldm.h"
/*-************************************* /*-*************************************
{ {
void* const ptr = cctx->blockState.nextCBlock + 1; void* const ptr = cctx->blockState.nextCBlock + 1;
cctx->entropyWorkspace = (U32*)ptr; cctx->entropyWorkspace = (U32*)ptr;
} }
cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid()); cctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());
return cctx; 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) static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx)
{ {
assert(cctx != NULL); assert(cctx != NULL);
assert(cctx->staticSize == 0); assert(cctx->staticSize == 0);
ZSTD_free(cctx->workSpace, cctx->customMem); cctx->workSpace = NULL; ZSTD_free(cctx->workSpace, cctx->customMem); cctx->workSpace = NULL;
ZSTD_freeCDict(cctx->cdictLocal); cctx->cdictLocal = NULL; ZSTD_clearAllDicts(cctx);
#ifdef ZSTD_MULTITHREAD #ifdef ZSTD_MULTITHREAD
ZSTDMT_freeCCtx(cctx->mtctx); cctx->mtctx = NULL; ZSTDMT_freeCCtx(cctx->mtctx); cctx->mtctx = NULL;
#endif #endif
} }
size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx) size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)
{ {
if (cctx==NULL) return 0; /* support free on NULL */ if (cctx==NULL) return 0; /* support free on NULL */
if (cctx->staticSize) return ERROR(memory_allocation); /* not compatible with static CCtx */ RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
"not compatible with static CCtx");
ZSTD_freeCCtxContent(cctx); ZSTD_freeCCtxContent(cctx);
ZSTD_free(cctx, cctx->customMem); ZSTD_free(cctx, cctx->customMem);
return 0; return 0;
} }
static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx) static size_t ZSTD_sizeof_mtctx(const ZSTD_CCtx* cctx)
{ {
#ifdef ZSTD_MULTITHREAD #ifdef ZSTD_MULTITHREAD
return ZSTDMT_sizeof_CCtx(cctx->mtctx); return ZSTDMT_sizeof_CCtx(cctx->mtctx);
#else #else
(void)cctx; (void)cctx;
return 0; return 0;
#endif #endif
} }
size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx) size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx)
{ {
if (cctx==NULL) return 0; /* support sizeof on NULL */ if (cctx==NULL) return 0; /* support sizeof on NULL */
return sizeof(*cctx) + cctx->workSpaceSize return sizeof(*cctx) + cctx->workSpaceSize
+ ZSTD_sizeof_CDict(cctx->cdictLocal) + ZSTD_sizeof_localDict(cctx->localDict)
+ ZSTD_sizeof_mtctx(cctx); + ZSTD_sizeof_mtctx(cctx);
} }
size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs) size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)
{ {
return ZSTD_sizeof_CCtx(zcs); /* same object */ return ZSTD_sizeof_CCtx(zcs); /* same object */
} }
} }
size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params) size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params)
{ {
return ZSTD_CCtxParams_init(params, ZSTD_CLEVEL_DEFAULT); return ZSTD_CCtxParams_init(params, ZSTD_CLEVEL_DEFAULT);
} }
size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel) { size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel) {
if (!cctxParams) { return ERROR(GENERIC); } RETURN_ERROR_IF(!cctxParams, GENERIC);
memset(cctxParams, 0, sizeof(*cctxParams)); memset(cctxParams, 0, sizeof(*cctxParams));
cctxParams->compressionLevel = compressionLevel; cctxParams->compressionLevel = compressionLevel;
cctxParams->fParams.contentSizeFlag = 1; cctxParams->fParams.contentSizeFlag = 1;
return 0; return 0;
} }
size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params) size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params)
{ {
if (!cctxParams) { return ERROR(GENERIC); } RETURN_ERROR_IF(!cctxParams, GENERIC);
CHECK_F( ZSTD_checkCParams(params.cParams) ); FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );
memset(cctxParams, 0, sizeof(*cctxParams)); memset(cctxParams, 0, sizeof(*cctxParams));
cctxParams->cParams = params.cParams; cctxParams->cParams = params.cParams;
cctxParams->fParams = params.fParams; cctxParams->fParams = params.fParams;
cctxParams->compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */ cctxParams->compressionLevel = ZSTD_CLEVEL_DEFAULT; /* should not matter, as all cParams are presumed properly defined */
assert(!ZSTD_checkCParams(params.cParams)); assert(!ZSTD_checkCParams(params.cParams));
return 0; return 0;
} }
return bounds; return bounds;
case ZSTD_c_forceAttachDict: case ZSTD_c_forceAttachDict:
ZSTD_STATIC_ASSERT(ZSTD_dictDefaultAttach < ZSTD_dictForceCopy); ZSTD_STATIC_ASSERT(ZSTD_dictDefaultAttach < ZSTD_dictForceCopy);
bounds.lowerBound = ZSTD_dictDefaultAttach; bounds.lowerBound = ZSTD_dictDefaultAttach;
bounds.upperBound = ZSTD_dictForceCopy; /* note : how to ensure at compile time that this is the highest value enum ? */ bounds.upperBound = ZSTD_dictForceCopy; /* note : how to ensure at compile time that this is the highest value enum ? */
return bounds; 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;
case ZSTD_c_targetCBlockSize:
bounds.lowerBound = ZSTD_TARGETCBLOCKSIZE_MIN;
bounds.upperBound = ZSTD_TARGETCBLOCKSIZE_MAX;
return bounds;
default: default:
{ ZSTD_bounds const boundError = { ERROR(parameter_unsupported), 0, 0 }; { ZSTD_bounds const boundError = { ERROR(parameter_unsupported), 0, 0 };
return boundError; return boundError;
} }
} }
} }
/* ZSTD_cParam_withinBounds: /* ZSTD_cParam_clampBounds:
* @return 1 if value is within cParam bounds, * Clamps the value into the bounded range.
* 0 otherwise */ */
static int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value) static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value)
{ {
ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);
if (ZSTD_isError(bounds.error)) return 0; if (ZSTD_isError(bounds.error)) return bounds.error;
if (value < bounds.lowerBound) return 0; if (*value < bounds.lowerBound) *value = bounds.lowerBound;
if (value > bounds.upperBound) return 0; if (*value > bounds.upperBound) *value = bounds.upperBound;
return 1; return 0;
} }
#define BOUNDCHECK(cParam, val) { \ #define BOUNDCHECK(cParam, val) { \
if (!ZSTD_cParam_withinBounds(cParam,val)) { \ RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \
return ERROR(parameter_outOfBound); \ parameter_outOfBound); \
} } }
static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param) static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param)
{ {
switch(param) switch(param)
{ {
case ZSTD_c_compressionLevel: case ZSTD_c_compressionLevel:
case ZSTD_c_hashLog: case ZSTD_c_hashLog:
case ZSTD_c_overlapLog: case ZSTD_c_overlapLog:
case ZSTD_c_rsyncable: case ZSTD_c_rsyncable:
case ZSTD_c_enableLongDistanceMatching: case ZSTD_c_enableLongDistanceMatching:
case ZSTD_c_ldmHashLog: case ZSTD_c_ldmHashLog:
case ZSTD_c_ldmMinMatch: case ZSTD_c_ldmMinMatch:
case ZSTD_c_ldmBucketSizeLog: case ZSTD_c_ldmBucketSizeLog:
case ZSTD_c_ldmHashRateLog: case ZSTD_c_ldmHashRateLog:
case ZSTD_c_forceAttachDict: case ZSTD_c_forceAttachDict:
case ZSTD_c_literalCompressionMode:
case ZSTD_c_targetCBlockSize:
default: default:
return 0; return 0;
} }
} }
size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value) size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value)
{ {
DEBUGLOG(4, "ZSTD_CCtx_setParameter (%i, %i)", (int)param, value); DEBUGLOG(4, "ZSTD_CCtx_setParameter (%i, %i)", (int)param, value);
if (cctx->streamStage != zcss_init) { if (cctx->streamStage != zcss_init) {
if (ZSTD_isUpdateAuthorized(param)) { if (ZSTD_isUpdateAuthorized(param)) {
cctx->cParamsChanged = 1; cctx->cParamsChanged = 1;
} else { } else {
return ERROR(stage_wrong); RETURN_ERROR(stage_wrong);
} } } }
switch(param) switch(param)
{ {
case ZSTD_c_format : case ZSTD_c_nbWorkers:
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); RETURN_ERROR_IF((value!=0) && cctx->staticSize, parameter_unsupported,
"MT not compatible with static alloc");
break;
case ZSTD_c_compressionLevel: case ZSTD_c_compressionLevel:
if (cctx->cdict) return ERROR(stage_wrong);
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
case ZSTD_c_windowLog: case ZSTD_c_windowLog:
case ZSTD_c_hashLog: case ZSTD_c_hashLog:
case ZSTD_c_chainLog: case ZSTD_c_chainLog:
case ZSTD_c_searchLog: case ZSTD_c_searchLog:
case ZSTD_c_minMatch: case ZSTD_c_minMatch:
case ZSTD_c_targetLength: case ZSTD_c_targetLength:
case ZSTD_c_strategy: case ZSTD_c_strategy:
if (cctx->cdict) return ERROR(stage_wrong); case ZSTD_c_ldmHashRateLog:
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); case ZSTD_c_format:
case ZSTD_c_contentSizeFlag: case ZSTD_c_contentSizeFlag:
case ZSTD_c_checksumFlag: case ZSTD_c_checksumFlag:
case ZSTD_c_dictIDFlag: case ZSTD_c_dictIDFlag:
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); case ZSTD_c_forceMaxWindow:
case ZSTD_c_forceMaxWindow : /* Force back-references to remain < windowSize,
* even when referencing into Dictionary content.
* default : 0 when using a CDict, 1 when using a Prefix */
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
case ZSTD_c_forceAttachDict: case ZSTD_c_forceAttachDict:
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value); case ZSTD_c_literalCompressionMode:
case ZSTD_c_nbWorkers:
if ((value!=0) && cctx->staticSize) {
return ERROR(parameter_unsupported); /* MT not compatible with static alloc */
}
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
case ZSTD_c_jobSize: case ZSTD_c_jobSize:
case ZSTD_c_overlapLog: case ZSTD_c_overlapLog:
case ZSTD_c_rsyncable: case ZSTD_c_rsyncable:
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
case ZSTD_c_enableLongDistanceMatching: case ZSTD_c_enableLongDistanceMatching:
case ZSTD_c_ldmHashLog: case ZSTD_c_ldmHashLog:
case ZSTD_c_ldmMinMatch: case ZSTD_c_ldmMinMatch:
case ZSTD_c_ldmBucketSizeLog: case ZSTD_c_ldmBucketSizeLog:
case ZSTD_c_ldmHashRateLog: case ZSTD_c_targetCBlockSize:
if (cctx->cdict) return ERROR(stage_wrong); break;
return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
default: return ERROR(parameter_unsupported); default: RETURN_ERROR(parameter_unsupported);
} }
return ZSTD_CCtxParams_setParameter(&cctx->requestedParams, param, value);
} }
size_t ZSTD_CCtxParam_setParameter(ZSTD_CCtx_params* CCtxParams, size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
ZSTD_cParameter param, int value) ZSTD_cParameter param, int value)
{ {
DEBUGLOG(4, "ZSTD_CCtxParam_setParameter (%i, %i)", (int)param, value); DEBUGLOG(4, "ZSTD_CCtxParams_setParameter (%i, %i)", (int)param, value);
switch(param) switch(param)
{ {
case ZSTD_c_format : case ZSTD_c_format :
BOUNDCHECK(ZSTD_c_format, value); BOUNDCHECK(ZSTD_c_format, value);
CCtxParams->format = (ZSTD_format_e)value; CCtxParams->format = (ZSTD_format_e)value;
return (size_t)CCtxParams->format; return (size_t)CCtxParams->format;
case ZSTD_c_compressionLevel : { case ZSTD_c_compressionLevel : {
int cLevel = value; FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));
if (cLevel > ZSTD_maxCLevel()) cLevel = ZSTD_maxCLevel(); if (value) { /* 0 : does not change current level */
if (cLevel < ZSTD_minCLevel()) cLevel = ZSTD_minCLevel(); CCtxParams->compressionLevel = value;
if (cLevel) { /* 0 : does not change current level */
CCtxParams->compressionLevel = cLevel;
} }
if (CCtxParams->compressionLevel >= 0) return CCtxParams->compressionLevel; if (CCtxParams->compressionLevel >= 0) return CCtxParams->compressionLevel;
return 0; /* return type (size_t) cannot represent negative values */ return 0; /* return type (size_t) cannot represent negative values */
} }
case ZSTD_c_windowLog : case ZSTD_c_windowLog :
if (value!=0) /* 0 => use default */ if (value!=0) /* 0 => use default */
BOUNDCHECK(ZSTD_c_windowLog, value); BOUNDCHECK(ZSTD_c_windowLog, value);
case ZSTD_c_forceAttachDict : { case ZSTD_c_forceAttachDict : {
const ZSTD_dictAttachPref_e pref = (ZSTD_dictAttachPref_e)value; const ZSTD_dictAttachPref_e pref = (ZSTD_dictAttachPref_e)value;
BOUNDCHECK(ZSTD_c_forceAttachDict, pref); BOUNDCHECK(ZSTD_c_forceAttachDict, pref);
CCtxParams->attachDictPref = pref; CCtxParams->attachDictPref = pref;
return CCtxParams->attachDictPref; 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 : case ZSTD_c_nbWorkers :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
if (value!=0) return ERROR(parameter_unsupported); RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
return 0; return 0;
#else #else
return ZSTDMT_CCtxParam_setNbWorkers(CCtxParams, value); FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));
CCtxParams->nbWorkers = value;
return CCtxParams->nbWorkers;
#endif #endif
case ZSTD_c_jobSize : case ZSTD_c_jobSize :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
return ERROR(parameter_unsupported); RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
return 0;
#else #else
return ZSTDMT_CCtxParam_setMTCtxParameter(CCtxParams, ZSTDMT_p_jobSize, value); /* 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 #endif
case ZSTD_c_overlapLog : case ZSTD_c_overlapLog :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
return ERROR(parameter_unsupported); RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
return 0;
#else #else
return ZSTDMT_CCtxParam_setMTCtxParameter(CCtxParams, ZSTDMT_p_overlapLog, value); FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value));
CCtxParams->overlapLog = value;
return CCtxParams->overlapLog;
#endif #endif
case ZSTD_c_rsyncable : case ZSTD_c_rsyncable :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
return ERROR(parameter_unsupported); RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
return 0;
#else #else
return ZSTDMT_CCtxParam_setMTCtxParameter(CCtxParams, ZSTDMT_p_rsyncable, value); FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value));
CCtxParams->rsyncable = value;
return CCtxParams->rsyncable;
#endif #endif
case ZSTD_c_enableLongDistanceMatching : case ZSTD_c_enableLongDistanceMatching :
CCtxParams->ldmParams.enableLdm = (value!=0); CCtxParams->ldmParams.enableLdm = (value!=0);
return CCtxParams->ldmParams.enableLdm; return CCtxParams->ldmParams.enableLdm;
case ZSTD_c_ldmHashLog : case ZSTD_c_ldmHashLog :
if (value!=0) /* 0 ==> auto */ if (value!=0) /* 0 ==> auto */
case ZSTD_c_ldmBucketSizeLog : case ZSTD_c_ldmBucketSizeLog :
if (value!=0) /* 0 ==> default */ if (value!=0) /* 0 ==> default */
BOUNDCHECK(ZSTD_c_ldmBucketSizeLog, value); BOUNDCHECK(ZSTD_c_ldmBucketSizeLog, value);
CCtxParams->ldmParams.bucketSizeLog = value; CCtxParams->ldmParams.bucketSizeLog = value;
return CCtxParams->ldmParams.bucketSizeLog; return CCtxParams->ldmParams.bucketSizeLog;
case ZSTD_c_ldmHashRateLog : case ZSTD_c_ldmHashRateLog :
if (value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN) RETURN_ERROR_IF(value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN,
return ERROR(parameter_outOfBound); parameter_outOfBound);
CCtxParams->ldmParams.hashRateLog = value; CCtxParams->ldmParams.hashRateLog = value;
return CCtxParams->ldmParams.hashRateLog; return CCtxParams->ldmParams.hashRateLog;
default: return ERROR(parameter_unsupported); case ZSTD_c_targetCBlockSize :
if (value!=0) /* 0 ==> default */
BOUNDCHECK(ZSTD_c_targetCBlockSize, value);
CCtxParams->targetCBlockSize = value;
return CCtxParams->targetCBlockSize;
default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
} }
} }
size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value) size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value)
{ {
return ZSTD_CCtxParam_getParameter(&cctx->requestedParams, param, value); return ZSTD_CCtxParams_getParameter(&cctx->requestedParams, param, value);
} }
size_t ZSTD_CCtxParam_getParameter( size_t ZSTD_CCtxParams_getParameter(
ZSTD_CCtx_params* CCtxParams, ZSTD_cParameter param, int* value) ZSTD_CCtx_params* CCtxParams, ZSTD_cParameter param, int* value)
{ {
switch(param) switch(param)
{ {
case ZSTD_c_format : case ZSTD_c_format :
*value = CCtxParams->format; *value = CCtxParams->format;
break; break;
case ZSTD_c_compressionLevel : case ZSTD_c_compressionLevel :
*value = CCtxParams->compressionLevel; *value = CCtxParams->compressionLevel;
break; break;
case ZSTD_c_windowLog : case ZSTD_c_windowLog :
*value = CCtxParams->cParams.windowLog; *value = (int)CCtxParams->cParams.windowLog;
break; break;
case ZSTD_c_hashLog : case ZSTD_c_hashLog :
*value = CCtxParams->cParams.hashLog; *value = (int)CCtxParams->cParams.hashLog;
break; break;
case ZSTD_c_chainLog : case ZSTD_c_chainLog :
*value = CCtxParams->cParams.chainLog; *value = (int)CCtxParams->cParams.chainLog;
break; break;
case ZSTD_c_searchLog : case ZSTD_c_searchLog :
*value = CCtxParams->cParams.searchLog; *value = CCtxParams->cParams.searchLog;
break; break;
case ZSTD_c_minMatch : case ZSTD_c_minMatch :
*value = CCtxParams->cParams.minMatch; *value = CCtxParams->cParams.minMatch;
break; break;
case ZSTD_c_targetLength : case ZSTD_c_targetLength :
*value = !CCtxParams->fParams.noDictIDFlag; *value = !CCtxParams->fParams.noDictIDFlag;
break; break;
case ZSTD_c_forceMaxWindow : case ZSTD_c_forceMaxWindow :
*value = CCtxParams->forceWindow; *value = CCtxParams->forceWindow;
break; break;
case ZSTD_c_forceAttachDict : case ZSTD_c_forceAttachDict :
*value = CCtxParams->attachDictPref; *value = CCtxParams->attachDictPref;
break; break;
case ZSTD_c_literalCompressionMode :
*value = CCtxParams->literalCompressionMode;
break;
case ZSTD_c_nbWorkers : case ZSTD_c_nbWorkers :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
assert(CCtxParams->nbWorkers == 0); assert(CCtxParams->nbWorkers == 0);
#endif #endif
*value = CCtxParams->nbWorkers; *value = CCtxParams->nbWorkers;
break; break;
case ZSTD_c_jobSize : case ZSTD_c_jobSize :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
return ERROR(parameter_unsupported); RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
#else #else
assert(CCtxParams->jobSize <= INT_MAX); assert(CCtxParams->jobSize <= INT_MAX);
*value = (int)CCtxParams->jobSize; *value = (int)CCtxParams->jobSize;
break; break;
#endif #endif
case ZSTD_c_overlapLog : case ZSTD_c_overlapLog :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
return ERROR(parameter_unsupported); RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
#else #else
*value = CCtxParams->overlapLog; *value = CCtxParams->overlapLog;
break; break;
#endif #endif
case ZSTD_c_rsyncable : case ZSTD_c_rsyncable :
#ifndef ZSTD_MULTITHREAD #ifndef ZSTD_MULTITHREAD
return ERROR(parameter_unsupported); RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
#else #else
*value = CCtxParams->rsyncable; *value = CCtxParams->rsyncable;
break; break;
#endif #endif
case ZSTD_c_enableLongDistanceMatching : case ZSTD_c_enableLongDistanceMatching :
*value = CCtxParams->ldmParams.enableLdm; *value = CCtxParams->ldmParams.enableLdm;
break; break;
case ZSTD_c_ldmHashLog : case ZSTD_c_ldmHashLog :
*value = CCtxParams->ldmParams.hashLog; *value = CCtxParams->ldmParams.hashLog;
break; break;
case ZSTD_c_ldmMinMatch : case ZSTD_c_ldmMinMatch :
*value = CCtxParams->ldmParams.minMatchLength; *value = CCtxParams->ldmParams.minMatchLength;
break; break;
case ZSTD_c_ldmBucketSizeLog : case ZSTD_c_ldmBucketSizeLog :
*value = CCtxParams->ldmParams.bucketSizeLog; *value = CCtxParams->ldmParams.bucketSizeLog;
break; break;
case ZSTD_c_ldmHashRateLog : case ZSTD_c_ldmHashRateLog :
*value = CCtxParams->ldmParams.hashRateLog; *value = CCtxParams->ldmParams.hashRateLog;
break; break;
default: return ERROR(parameter_unsupported); case ZSTD_c_targetCBlockSize :
*value = (int)CCtxParams->targetCBlockSize;
break;
default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
} }
return 0; return 0;
} }
/** ZSTD_CCtx_setParametersUsingCCtxParams() : /** ZSTD_CCtx_setParametersUsingCCtxParams() :
* just applies `params` into `cctx` * just applies `params` into `cctx`
* no action is performed, parameters are merely stored. * no action is performed, parameters are merely stored.
* If ZSTDMT is enabled, parameters are pushed to cctx->mtctx. * If ZSTDMT is enabled, parameters are pushed to cctx->mtctx.
* This is possible even if a compression is ongoing. * 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. * In which case, new parameters will be applied on the fly, starting with next compression job.
*/ */
size_t ZSTD_CCtx_setParametersUsingCCtxParams( size_t ZSTD_CCtx_setParametersUsingCCtxParams(
ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params) ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params)
{ {
DEBUGLOG(4, "ZSTD_CCtx_setParametersUsingCCtxParams"); DEBUGLOG(4, "ZSTD_CCtx_setParametersUsingCCtxParams");
if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
if (cctx->cdict) return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->cdict, stage_wrong);
cctx->requestedParams = *params; cctx->requestedParams = *params;
return 0; return 0;
} }
ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize) ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize)
{ {
DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %u bytes", (U32)pledgedSrcSize); DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %u bytes", (U32)pledgedSrcSize);
if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1; cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;
return 0; 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( size_t ZSTD_CCtx_loadDictionary_advanced(
ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_CCtx* cctx, const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType) ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType)
{ {
if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
if (cctx->staticSize) return ERROR(memory_allocation); /* no malloc for static CCtx */ RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
"no malloc for static CCtx");
DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize); DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize);
ZSTD_freeCDict(cctx->cdictLocal); /* in case one already exists */ ZSTD_clearAllDicts(cctx); /* in case one already exists */
if (dict==NULL || dictSize==0) { /* no dictionary mode */ if (dict == NULL || dictSize == 0) /* no dictionary mode */
cctx->cdictLocal = NULL; return 0;
cctx->cdict = NULL; if (dictLoadMethod == ZSTD_dlm_byRef) {
cctx->localDict.dict = dict;
} else { } else {
ZSTD_compressionParameters const cParams = void* dictBuffer = ZSTD_malloc(dictSize, cctx->customMem);
ZSTD_getCParamsFromCCtxParams(&cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, dictSize); RETURN_ERROR_IF(!dictBuffer, memory_allocation);
cctx->cdictLocal = ZSTD_createCDict_advanced( memcpy(dictBuffer, dict, dictSize);
dict, dictSize, cctx->localDict.dictBuffer = dictBuffer;
dictLoadMethod, dictContentType, cctx->localDict.dict = dictBuffer;
cParams, cctx->customMem);
cctx->cdict = cctx->cdictLocal;
if (cctx->cdictLocal == NULL)
return ERROR(memory_allocation);
} }
cctx->localDict.dictSize = dictSize;
cctx->localDict.dictContentType = dictContentType;
return 0; return 0;
} }
ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference( ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(
ZSTD_CCtx* cctx, const void* dict, size_t dictSize) ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
{ {
return ZSTD_CCtx_loadDictionary_advanced( return ZSTD_CCtx_loadDictionary_advanced(
cctx, dict, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto); 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) ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize)
{ {
return ZSTD_CCtx_loadDictionary_advanced( return ZSTD_CCtx_loadDictionary_advanced(
cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto); cctx, dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto);
} }
size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict) size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
{ {
if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); 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; cctx->cdict = cdict;
memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); /* exclusive */
return 0; return 0;
} }
size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize) 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); return ZSTD_CCtx_refPrefix_advanced(cctx, prefix, prefixSize, ZSTD_dct_rawContent);
} }
size_t ZSTD_CCtx_refPrefix_advanced( size_t ZSTD_CCtx_refPrefix_advanced(
ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType) ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)
{ {
if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
cctx->cdict = NULL; /* prefix discards any prior cdict */ ZSTD_clearAllDicts(cctx);
cctx->prefixDict.dict = prefix; cctx->prefixDict.dict = prefix;
cctx->prefixDict.dictSize = prefixSize; cctx->prefixDict.dictSize = prefixSize;
cctx->prefixDict.dictContentType = dictContentType; cctx->prefixDict.dictContentType = dictContentType;
return 0; return 0;
} }
/*! ZSTD_CCtx_reset() : /*! ZSTD_CCtx_reset() :
* Also dumps dictionary */ * Also dumps dictionary */
size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset) size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset)
{ {
if ( (reset == ZSTD_reset_session_only) if ( (reset == ZSTD_reset_session_only)
|| (reset == ZSTD_reset_session_and_parameters) ) { || (reset == ZSTD_reset_session_and_parameters) ) {
cctx->streamStage = zcss_init; cctx->streamStage = zcss_init;
cctx->pledgedSrcSizePlusOne = 0; cctx->pledgedSrcSizePlusOne = 0;
} }
if ( (reset == ZSTD_reset_parameters) if ( (reset == ZSTD_reset_parameters)
|| (reset == ZSTD_reset_session_and_parameters) ) { || (reset == ZSTD_reset_session_and_parameters) ) {
if (cctx->streamStage != zcss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
cctx->cdict = NULL; ZSTD_clearAllDicts(cctx);
return ZSTD_CCtxParams_reset(&cctx->requestedParams); return ZSTD_CCtxParams_reset(&cctx->requestedParams);
} }
return 0; return 0;
} }
/** ZSTD_checkCParams() : /** ZSTD_checkCParams() :
control CParam values remain within authorized range. control CParam values remain within authorized range.
@return : 0, or an error code if one value is beyond authorized range */ @return : 0, or an error code if one value is beyond authorized range */
size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams) size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
{ {
BOUNDCHECK(ZSTD_c_windowLog, cParams.windowLog); BOUNDCHECK(ZSTD_c_windowLog, (int)cParams.windowLog);
BOUNDCHECK(ZSTD_c_chainLog, cParams.chainLog); BOUNDCHECK(ZSTD_c_chainLog, (int)cParams.chainLog);
BOUNDCHECK(ZSTD_c_hashLog, cParams.hashLog); BOUNDCHECK(ZSTD_c_hashLog, (int)cParams.hashLog);
BOUNDCHECK(ZSTD_c_searchLog, cParams.searchLog); BOUNDCHECK(ZSTD_c_searchLog, (int)cParams.searchLog);
BOUNDCHECK(ZSTD_c_minMatch, cParams.minMatch); BOUNDCHECK(ZSTD_c_minMatch, (int)cParams.minMatch);
BOUNDCHECK(ZSTD_c_targetLength,cParams.targetLength); BOUNDCHECK(ZSTD_c_targetLength,(int)cParams.targetLength);
BOUNDCHECK(ZSTD_c_strategy, cParams.strategy); BOUNDCHECK(ZSTD_c_strategy, cParams.strategy);
return 0; return 0;
} }
/** ZSTD_clampCParams() : /** ZSTD_clampCParams() :
* make CParam values within valid range. * make CParam values within valid range.
* @return : valid CParams */ * @return : valid CParams */
static ZSTD_compressionParameters static ZSTD_compressionParameters
ZSTD_clampCParams(ZSTD_compressionParameters cParams) ZSTD_clampCParams(ZSTD_compressionParameters cParams)
{ {
# define CLAMP_TYPE(cParam, val, type) { \ # define CLAMP_TYPE(cParam, val, type) { \
ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); \ ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); \
if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound; \ if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound; \
else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \ else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \
} }
# define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, int) # define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, unsigned)
CLAMP(ZSTD_c_windowLog, cParams.windowLog); CLAMP(ZSTD_c_windowLog, cParams.windowLog);
CLAMP(ZSTD_c_chainLog, cParams.chainLog); CLAMP(ZSTD_c_chainLog, cParams.chainLog);
CLAMP(ZSTD_c_hashLog, cParams.hashLog); CLAMP(ZSTD_c_hashLog, cParams.hashLog);
CLAMP(ZSTD_c_searchLog, cParams.searchLog); CLAMP(ZSTD_c_searchLog, cParams.searchLog);
CLAMP(ZSTD_c_minMatch, cParams.minMatch); CLAMP(ZSTD_c_minMatch, cParams.minMatch);
CLAMP(ZSTD_c_targetLength,cParams.targetLength); CLAMP(ZSTD_c_targetLength,cParams.targetLength);
CLAMP_TYPE(ZSTD_c_strategy,cParams.strategy, ZSTD_strategy); CLAMP_TYPE(ZSTD_c_strategy,cParams.strategy, ZSTD_strategy);
return cParams; return cParams;
} }
/** ZSTD_cycleLog() : /** ZSTD_cycleLog() :
* condition for correct operation : hashLog > 1 */ * condition for correct operation : hashLog > 1 */
static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat) static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)
{ {
U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2); U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);
return hashLog - btScale; return hashLog - btScale;
} }
/** ZSTD_adjustCParams_internal() : /** ZSTD_adjustCParams_internal() :
optimize `cPar` for a given input (`srcSize` and `dictSize`). * optimize `cPar` for a specified input (`srcSize` and `dictSize`).
mostly downsizing to reduce memory consumption and initialization latency. * mostly downsize to reduce memory consumption and initialization latency.
Both `srcSize` and `dictSize` are optional (use 0 if unknown). * `srcSize` can be ZSTD_CONTENTSIZE_UNKNOWN when not known.
Note : cPar is assumed validated. Use ZSTD_checkCParams() to ensure this condition. */ * 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 static ZSTD_compressionParameters
ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar, ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar,
unsigned long long srcSize, unsigned long long srcSize,
size_t dictSize) size_t dictSize)
{ {
static const U64 minSrcSize = 513; /* (1<<9) + 1 */ static const U64 minSrcSize = 513; /* (1<<9) + 1 */
static const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1); static const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1);
assert(ZSTD_checkCParams(cPar)==0); assert(ZSTD_checkCParams(cPar)==0);
if (dictSize && (srcSize+1<2) /* srcSize unknown */ ) if (dictSize && (srcSize+1<2) /* ZSTD_CONTENTSIZE_UNKNOWN and 0 mean "unknown" */ )
srcSize = minSrcSize; /* presumed small when there is a dictionary */ srcSize = minSrcSize; /* presumed small when there is a dictionary */
else if (srcSize == 0) else if (srcSize == 0)
srcSize = ZSTD_CONTENTSIZE_UNKNOWN; /* 0 == unknown : presumed large */ srcSize = ZSTD_CONTENTSIZE_UNKNOWN; /* 0 == unknown : presumed large */
/* resize windowLog if input is small enough, to use less memory */ /* resize windowLog if input is small enough, to use less memory */
if ( (srcSize < maxWindowResize) if ( (srcSize < maxWindowResize)
&& (dictSize < maxWindowResize) ) { && (dictSize < maxWindowResize) ) {
U32 const tSize = (U32)(srcSize + dictSize); U32 const tSize = (U32)(srcSize + dictSize);
static U32 const hashSizeMin = 1 << ZSTD_HASHLOG_MIN; static U32 const hashSizeMin = 1 << ZSTD_HASHLOG_MIN;
U32 const srcLog = (tSize < hashSizeMin) ? ZSTD_HASHLOG_MIN : U32 const srcLog = (tSize < hashSizeMin) ? ZSTD_HASHLOG_MIN :
ZSTD_highbit32(tSize-1) + 1; ZSTD_highbit32(tSize-1) + 1;
if (cPar.windowLog > srcLog) cPar.windowLog = srcLog; if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;
} }
if (cPar.hashLog > cPar.windowLog+1) cPar.hashLog = cPar.windowLog+1; if (cPar.hashLog > cPar.windowLog+1) cPar.hashLog = cPar.windowLog+1;
{ U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy); { U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);
if (cycleLog > cPar.windowLog) if (cycleLog > cPar.windowLog)
cPar.chainLog -= (cycleLog - cPar.windowLog); cPar.chainLog -= (cycleLog - cPar.windowLog);
} }
if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN) if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)
cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* required for frame header */ cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* minimum wlog required for valid frame header */
return cPar; return cPar;
} }
ZSTD_compressionParameters ZSTD_compressionParameters
ZSTD_adjustCParams(ZSTD_compressionParameters cPar, ZSTD_adjustCParams(ZSTD_compressionParameters cPar,
unsigned long long srcSize, unsigned long long srcSize,
size_t dictSize) size_t dictSize)
{ {
cPar = ZSTD_clampCParams(cPar); cPar = ZSTD_clampCParams(cPar); /* resulting cPar is necessarily valid (all parameters within range) */
return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize); return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize);
} }
ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams( ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(
const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize) const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize)
{ {
ZSTD_compressionParameters cParams = ZSTD_getCParams(CCtxParams->compressionLevel, srcSizeHint, dictSize); ZSTD_compressionParameters cParams = ZSTD_getCParams(CCtxParams->compressionLevel, srcSizeHint, dictSize);
if (CCtxParams->ldmParams.enableLdm) cParams.windowLog = ZSTD_LDM_DEFAULT_WINDOW_LOG; if (CCtxParams->ldmParams.enableLdm) cParams.windowLog = ZSTD_LDM_DEFAULT_WINDOW_LOG;
: 0; : 0;
DEBUGLOG(4, "chainSize: %u - hSize: %u - h3Size: %u", DEBUGLOG(4, "chainSize: %u - hSize: %u - h3Size: %u",
(U32)chainSize, (U32)hSize, (U32)h3Size); (U32)chainSize, (U32)hSize, (U32)h3Size);
return tableSpace + optSpace; return tableSpace + optSpace;
} }
size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params) size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params)
{ {
/* Estimate CCtx size is supported for single-threaded compression only. */ RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");
if (params->nbWorkers > 0) { return ERROR(GENERIC); }
{ ZSTD_compressionParameters const cParams = { ZSTD_compressionParameters const cParams =
ZSTD_getCParamsFromCCtxParams(params, 0, 0); ZSTD_getCParamsFromCCtxParams(params, 0, 0);
size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog); size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
U32 const divider = (cParams.minMatch==3) ? 3 : 4; U32 const divider = (cParams.minMatch==3) ? 3 : 4;
size_t const maxNbSeq = blockSize / divider; size_t const maxNbSeq = blockSize / divider;
size_t const tokenSpace = WILDCOPY_OVERLENGTH + blockSize + 11*maxNbSeq; size_t const tokenSpace = WILDCOPY_OVERLENGTH + blockSize + 11*maxNbSeq;
size_t const entropySpace = HUF_WORKSPACE_SIZE; size_t const entropySpace = HUF_WORKSPACE_SIZE;
size_t const blockStateSpace = 2 * sizeof(ZSTD_compressedBlockState_t); size_t const blockStateSpace = 2 * sizeof(ZSTD_compressedBlockState_t);
size_t const newMB = ZSTD_estimateCCtxSize_internal(level); size_t const newMB = ZSTD_estimateCCtxSize_internal(level);
if (newMB > memBudget) memBudget = newMB; if (newMB > memBudget) memBudget = newMB;
} }
return memBudget; return memBudget;
} }
size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params) size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)
{ {
if (params->nbWorkers > 0) { return ERROR(GENERIC); } RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");
{ size_t const CCtxSize = ZSTD_estimateCCtxSize_usingCCtxParams(params); { ZSTD_compressionParameters const cParams =
size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << params->cParams.windowLog); ZSTD_getCParamsFromCCtxParams(params, 0, 0);
size_t const inBuffSize = ((size_t)1 << params->cParams.windowLog) + blockSize; 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 outBuffSize = ZSTD_compressBound(blockSize) + 1;
size_t const streamingSize = inBuffSize + outBuffSize; size_t const streamingSize = inBuffSize + outBuffSize;
return CCtxSize + streamingSize; return CCtxSize + streamingSize;
} }
} }
size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams) size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams)
bs->rep[i] = repStartValue[i]; bs->rep[i] = repStartValue[i];
bs->entropy.huf.repeatMode = HUF_repeat_none; bs->entropy.huf.repeatMode = HUF_repeat_none;
bs->entropy.fse.offcode_repeatMode = FSE_repeat_none; bs->entropy.fse.offcode_repeatMode = FSE_repeat_none;
bs->entropy.fse.matchlength_repeatMode = FSE_repeat_none; bs->entropy.fse.matchlength_repeatMode = FSE_repeat_none;
bs->entropy.fse.litlength_repeatMode = FSE_repeat_none; bs->entropy.fse.litlength_repeatMode = FSE_repeat_none;
} }
/*! ZSTD_invalidateMatchState() /*! ZSTD_invalidateMatchState()
* Invalidate all the matches in the match finder tables. * Invalidate all the matches in the match finder tables.
* Requires nextSrc and base to be set (can be NULL). * Requires nextSrc and base to be set (can be NULL).
*/ */
static void ZSTD_invalidateMatchState(ZSTD_matchState_t* ms) static void ZSTD_invalidateMatchState(ZSTD_matchState_t* ms)
{ {
ZSTD_window_clear(&ms->window); ZSTD_window_clear(&ms->window);
ms->nextToUpdate = ms->window.dictLimit; ms->nextToUpdate = ms->window.dictLimit;
ms->nextToUpdate3 = ms->window.dictLimit;
ms->loadedDictEnd = 0; ms->loadedDictEnd = 0;
ms->opt.litLengthSum = 0; /* force reset of btopt stats */ ms->opt.litLengthSum = 0; /* force reset of btopt stats */
ms->dictMatchState = NULL; ms->dictMatchState = NULL;
} }
/*! ZSTD_continueCCtx() : /*! ZSTD_continueCCtx() :
* reuse CCtx without reset (note : requires no dictionary) */ * reuse CCtx without reset (note : requires no dictionary) */
static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_CCtx_params params, U64 pledgedSrcSize) static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_CCtx_params params, U64 pledgedSrcSize)
ZSTD_invalidateMatchState(&cctx->blockState.matchState); ZSTD_invalidateMatchState(&cctx->blockState.matchState);
ZSTD_reset_compressedBlockState(cctx->blockState.prevCBlock); ZSTD_reset_compressedBlockState(cctx->blockState.prevCBlock);
XXH64_reset(&cctx->xxhState, 0); XXH64_reset(&cctx->xxhState, 0);
return 0; return 0;
} }
typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset } ZSTD_compResetPolicy_e; typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset } ZSTD_compResetPolicy_e;
typedef enum { ZSTD_resetTarget_CDict, ZSTD_resetTarget_CCtx } ZSTD_resetTarget_e;
static void* static void*
ZSTD_reset_matchState(ZSTD_matchState_t* ms, ZSTD_reset_matchState(ZSTD_matchState_t* ms,
void* ptr, void* ptr,
const ZSTD_compressionParameters* cParams, const ZSTD_compressionParameters* cParams,
ZSTD_compResetPolicy_e const crp, U32 const forCCtx) ZSTD_compResetPolicy_e const crp, ZSTD_resetTarget_e const forWho)
{ {
size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog); size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog);
size_t const hSize = ((size_t)1) << cParams->hashLog; size_t const hSize = ((size_t)1) << cParams->hashLog;
U32 const hashLog3 = (forCCtx && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0; U32 const hashLog3 = ((forWho == ZSTD_resetTarget_CCtx) && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;
size_t const h3Size = ((size_t)1) << hashLog3; size_t const h3Size = ((size_t)1) << hashLog3;
size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32); size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
assert(((size_t)ptr & 3) == 0); assert(((size_t)ptr & 3) == 0);
ms->hashLog3 = hashLog3; ms->hashLog3 = hashLog3;
memset(&ms->window, 0, sizeof(ms->window)); memset(&ms->window, 0, sizeof(ms->window));
ms->window.dictLimit = 1; /* start from 1, so that 1st position is valid */ 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.lowLimit = 1; /* it ensures first and later CCtx usages compress the same */
ms->window.nextSrc = ms->window.base + 1; /* see issue #1241 */ ms->window.nextSrc = ms->window.base + 1; /* see issue #1241 */
ZSTD_invalidateMatchState(ms); ZSTD_invalidateMatchState(ms);
/* opt parser space */ /* opt parser space */
if (forCCtx && (cParams->strategy >= ZSTD_btopt)) { if ((forWho == ZSTD_resetTarget_CCtx) && (cParams->strategy >= ZSTD_btopt)) {
DEBUGLOG(4, "reserving optimal parser space"); DEBUGLOG(4, "reserving optimal parser space");
ms->opt.litFreq = (unsigned*)ptr; ms->opt.litFreq = (unsigned*)ptr;
ms->opt.litLengthFreq = ms->opt.litFreq + (1<<Litbits); ms->opt.litLengthFreq = ms->opt.litFreq + (1<<Litbits);
ms->opt.matchLengthFreq = ms->opt.litLengthFreq + (MaxLL+1); ms->opt.matchLengthFreq = ms->opt.litLengthFreq + (MaxLL+1);
ms->opt.offCodeFreq = ms->opt.matchLengthFreq + (MaxML+1); ms->opt.offCodeFreq = ms->opt.matchLengthFreq + (MaxML+1);
ptr = ms->opt.offCodeFreq + (MaxOff+1); ptr = ms->opt.offCodeFreq + (MaxOff+1);
ms->opt.matchTable = (ZSTD_match_t*)ptr; ms->opt.matchTable = (ZSTD_match_t*)ptr;
ptr = ms->opt.matchTable + ZSTD_OPT_NUM+1; ptr = ms->opt.matchTable + ZSTD_OPT_NUM+1;
ptr = ms->hashTable3 + h3Size; ptr = ms->hashTable3 + h3Size;
ms->cParams = *cParams; ms->cParams = *cParams;
assert(((size_t)ptr & 3) == 0); assert(((size_t)ptr & 3) == 0);
return ptr; return ptr;
} }
/* ZSTD_indexTooCloseToMax() :
* minor optimization : prefer memset() rather than reduceIndex()
* which is measurably slow in some circumstances (reported for Visual Studio).
* Works when re-using a context for a lot of smallish inputs :
* if all inputs are smaller than ZSTD_INDEXOVERFLOW_MARGIN,
* memset() will be triggered before reduceIndex().
*/
#define ZSTD_INDEXOVERFLOW_MARGIN (16 MB)
static int ZSTD_indexTooCloseToMax(ZSTD_window_t w)
{
return (size_t)(w.nextSrc - w.base) > (ZSTD_CURRENT_MAX - ZSTD_INDEXOVERFLOW_MARGIN);
}
#define ZSTD_WORKSPACETOOLARGE_FACTOR 3 /* define "workspace is too large" as this number of times larger than needed */ #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 #define ZSTD_WORKSPACETOOLARGE_MAXDURATION 128 /* when workspace is continuously too large
* during at least this number of times, * during at least this number of times,
* context's memory usage is considered wasteful, * context's memory usage is considered wasteful,
* because it's sized to handle a worst case scenario which rarely happens. * because it's sized to handle a worst case scenario which rarely happens.
* In which case, resize it down to free some memory */ * In which case, resize it down to free some memory */
/*! ZSTD_resetCCtx_internal() : /*! ZSTD_resetCCtx_internal() :
note : `params` are assumed fully validated at this stage */ note : `params` are assumed fully validated at this stage */
static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc, static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
ZSTD_CCtx_params params, ZSTD_CCtx_params params,
U64 pledgedSrcSize, U64 const pledgedSrcSize,
ZSTD_compResetPolicy_e const crp, ZSTD_compResetPolicy_e const crp,
ZSTD_buffered_policy_e const zbuff) ZSTD_buffered_policy_e const zbuff)
{ {
DEBUGLOG(4, "ZSTD_resetCCtx_internal: pledgedSrcSize=%u, wlog=%u", DEBUGLOG(4, "ZSTD_resetCCtx_internal: pledgedSrcSize=%u, wlog=%u",
(U32)pledgedSrcSize, params.cParams.windowLog); (U32)pledgedSrcSize, params.cParams.windowLog);
assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
if (crp == ZSTDcrp_continue) { if (crp == ZSTDcrp_continue) {
if (ZSTD_equivalentParams(zc->appliedParams, params, if (ZSTD_equivalentParams(zc->appliedParams, params,
zc->inBuffSize, zc->inBuffSize,
zc->seqStore.maxNbSeq, zc->seqStore.maxNbLit, zc->seqStore.maxNbSeq, zc->seqStore.maxNbLit,
zbuff, pledgedSrcSize)) { zbuff, pledgedSrcSize) ) {
DEBUGLOG(4, "ZSTD_equivalentParams()==1 -> continue mode (wLog1=%u, blockSize1=%zu)", DEBUGLOG(4, "ZSTD_equivalentParams()==1 -> consider continue mode");
zc->appliedParams.cParams.windowLog, zc->blockSize);
zc->workSpaceOversizedDuration += (zc->workSpaceOversizedDuration > 0); /* if it was too large, it still is */ zc->workSpaceOversizedDuration += (zc->workSpaceOversizedDuration > 0); /* if it was too large, it still is */
if (zc->workSpaceOversizedDuration <= ZSTD_WORKSPACETOOLARGE_MAXDURATION) if (zc->workSpaceOversizedDuration <= ZSTD_WORKSPACETOOLARGE_MAXDURATION) {
DEBUGLOG(4, "continue mode confirmed (wLog1=%u, blockSize1=%zu)",
zc->appliedParams.cParams.windowLog, zc->blockSize);
if (ZSTD_indexTooCloseToMax(zc->blockState.matchState.window)) {
/* prefer a reset, faster than a rescale */
ZSTD_reset_matchState(&zc->blockState.matchState,
zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32,
&params.cParams,
crp, ZSTD_resetTarget_CCtx);
}
return ZSTD_continueCCtx(zc, params, pledgedSrcSize); return ZSTD_continueCCtx(zc, params, pledgedSrcSize);
} } } } }
DEBUGLOG(4, "ZSTD_equivalentParams()==0 -> reset CCtx"); DEBUGLOG(4, "ZSTD_equivalentParams()==0 -> reset CCtx");
if (params.ldmParams.enableLdm) { if (params.ldmParams.enableLdm) {
/* Adjust long distance matching parameters */ /* Adjust long distance matching parameters */
ZSTD_ldm_adjustParameters(&params.ldmParams, &params.cParams); ZSTD_ldm_adjustParameters(&params.ldmParams, &params.cParams);
assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog); assert(params.ldmParams.hashLog >= params.ldmParams.bucketSizeLog);
assert(params.ldmParams.hashRateLog < 32); assert(params.ldmParams.hashRateLog < 32);
zc->ldmState.hashPower = ZSTD_rollingHash_primePower(params.ldmParams.minMatchLength); zc->ldmState.hashPower = ZSTD_rollingHash_primePower(params.ldmParams.minMatchLength);
int const workSpaceWasteful = workSpaceTooLarge && (zc->workSpaceOversizedDuration > ZSTD_WORKSPACETOOLARGE_MAXDURATION); int const workSpaceWasteful = workSpaceTooLarge && (zc->workSpaceOversizedDuration > ZSTD_WORKSPACETOOLARGE_MAXDURATION);
zc->workSpaceOversizedDuration = workSpaceTooLarge ? zc->workSpaceOversizedDuration+1 : 0; zc->workSpaceOversizedDuration = workSpaceTooLarge ? zc->workSpaceOversizedDuration+1 : 0;
DEBUGLOG(4, "Need %zuKB workspace, including %zuKB for match state, and %zuKB for buffers", DEBUGLOG(4, "Need %zuKB workspace, including %zuKB for match state, and %zuKB for buffers",
neededSpace>>10, matchStateSize>>10, bufferSpace>>10); neededSpace>>10, matchStateSize>>10, bufferSpace>>10);
DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize); DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize);
if (workSpaceTooSmall || workSpaceWasteful) { if (workSpaceTooSmall || workSpaceWasteful) {
DEBUGLOG(4, "Need to resize workSpaceSize from %zuKB to %zuKB", DEBUGLOG(4, "Resize workSpaceSize from %zuKB to %zuKB",
zc->workSpaceSize >> 10, zc->workSpaceSize >> 10,
neededSpace >> 10); neededSpace >> 10);
/* static cctx : no resize, error out */
if (zc->staticSize) return ERROR(memory_allocation); RETURN_ERROR_IF(zc->staticSize, memory_allocation, "static cctx : no resize");
zc->workSpaceSize = 0; zc->workSpaceSize = 0;
ZSTD_free(zc->workSpace, zc->customMem); ZSTD_free(zc->workSpace, zc->customMem);
zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem); zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem);
if (zc->workSpace == NULL) return ERROR(memory_allocation); RETURN_ERROR_IF(zc->workSpace == NULL, memory_allocation);
zc->workSpaceSize = neededSpace; zc->workSpaceSize = neededSpace;
zc->workSpaceOversizedDuration = 0; zc->workSpaceOversizedDuration = 0;
/* Statically sized space. /* Statically sized space.
* entropyWorkspace never moves, * entropyWorkspace never moves,
* though prev/next block swap places */ * though prev/next block swap places */
assert(((size_t)zc->workSpace & 3) == 0); /* ensure correct alignment */ assert(((size_t)zc->workSpace & 3) == 0); /* ensure correct alignment */
assert(zc->workSpaceSize >= 2 * sizeof(ZSTD_compressedBlockState_t)); assert(zc->workSpaceSize >= 2 * sizeof(ZSTD_compressedBlockState_t));
zc->blockSize = blockSize; zc->blockSize = blockSize;
XXH64_reset(&zc->xxhState, 0); XXH64_reset(&zc->xxhState, 0);
zc->stage = ZSTDcs_init; zc->stage = ZSTDcs_init;
zc->dictID = 0; zc->dictID = 0;
ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock); ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock);
ptr = zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32; ptr = ZSTD_reset_matchState(&zc->blockState.matchState,
zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32,
&params.cParams,
crp, ZSTD_resetTarget_CCtx);
/* ldm hash table */ /* ldm hash table */
/* initialize bucketOffsets table later for pointer alignment */ /* initialize bucketOffsets table later for pointer alignment */
if (params.ldmParams.enableLdm) { if (params.ldmParams.enableLdm) {
size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog; size_t const ldmHSize = ((size_t)1) << params.ldmParams.hashLog;
memset(ptr, 0, ldmHSize * sizeof(ldmEntry_t)); memset(ptr, 0, ldmHSize * sizeof(ldmEntry_t));
assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */ assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */
zc->ldmState.hashTable = (ldmEntry_t*)ptr; zc->ldmState.hashTable = (ldmEntry_t*)ptr;
ptr = zc->ldmState.hashTable + ldmHSize; ptr = zc->ldmState.hashTable + ldmHSize;
zc->ldmSequences = (rawSeq*)ptr; zc->ldmSequences = (rawSeq*)ptr;
ptr = zc->ldmSequences + maxNbLdmSeq; ptr = zc->ldmSequences + maxNbLdmSeq;
zc->maxNbLdmSequences = maxNbLdmSeq; zc->maxNbLdmSequences = maxNbLdmSeq;
memset(&zc->ldmState.window, 0, sizeof(zc->ldmState.window)); memset(&zc->ldmState.window, 0, sizeof(zc->ldmState.window));
} }
assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */ assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */
ptr = ZSTD_reset_matchState(&zc->blockState.matchState, ptr, &params.cParams, crp, /* forCCtx */ 1);
/* sequences storage */ /* sequences storage */
zc->seqStore.maxNbSeq = maxNbSeq; zc->seqStore.maxNbSeq = maxNbSeq;
zc->seqStore.sequencesStart = (seqDef*)ptr; zc->seqStore.sequencesStart = (seqDef*)ptr;
ptr = zc->seqStore.sequencesStart + maxNbSeq; ptr = zc->seqStore.sequencesStart + maxNbSeq;
zc->seqStore.llCode = (BYTE*) ptr; zc->seqStore.llCode = (BYTE*) ptr;
zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq; zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq;
zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq; zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq;
zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq; zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq;
return ( pledgedSrcSize <= cutoff return ( pledgedSrcSize <= cutoff
|| pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN || pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN
|| params.attachDictPref == ZSTD_dictForceAttach ) || params.attachDictPref == ZSTD_dictForceAttach )
&& params.attachDictPref != ZSTD_dictForceCopy && params.attachDictPref != ZSTD_dictForceCopy
&& !params.forceWindow; /* dictMatchState isn't correctly && !params.forceWindow; /* dictMatchState isn't correctly
* handled in _enforceMaxDist */ * handled in _enforceMaxDist */
} }
static size_t ZSTD_resetCCtx_byAttachingCDict( static size_t
ZSTD_CCtx* cctx, ZSTD_resetCCtx_byAttachingCDict(ZSTD_CCtx* cctx,
const ZSTD_CDict* cdict, const ZSTD_CDict* cdict,
ZSTD_CCtx_params params, ZSTD_CCtx_params params,
U64 pledgedSrcSize, U64 pledgedSrcSize,
ZSTD_buffered_policy_e zbuff) ZSTD_buffered_policy_e zbuff)
{ {
{ { const ZSTD_compressionParameters* const cdict_cParams = &cdict->matchState.cParams;
const ZSTD_compressionParameters *cdict_cParams = &cdict->matchState.cParams;
unsigned const windowLog = params.cParams.windowLog; unsigned const windowLog = params.cParams.windowLog;
assert(windowLog != 0); assert(windowLog != 0);
/* Resize working context table params for input only, since the dict /* Resize working context table params for input only, since the dict
* has its own tables. */ * has its own tables. */
params.cParams = ZSTD_adjustCParams_internal(*cdict_cParams, pledgedSrcSize, 0); params.cParams = ZSTD_adjustCParams_internal(*cdict_cParams, pledgedSrcSize, 0);
params.cParams.windowLog = windowLog; params.cParams.windowLog = windowLog;
ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize, ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,
ZSTDcrp_continue, zbuff); ZSTDcrp_continue, zbuff);
assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy); assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy);
} }
{ { const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc
const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc
- cdict->matchState.window.base); - cdict->matchState.window.base);
const U32 cdictLen = cdictEnd - cdict->matchState.window.dictLimit; const U32 cdictLen = cdictEnd - cdict->matchState.window.dictLimit;
if (cdictLen == 0) { if (cdictLen == 0) {
/* don't even attach dictionaries with no contents */ /* don't even attach dictionaries with no contents */
DEBUGLOG(4, "skipping attaching empty dictionary"); DEBUGLOG(4, "skipping attaching empty dictionary");
} else { } else {
DEBUGLOG(4, "attaching dictionary into context"); DEBUGLOG(4, "attaching dictionary into context");
cctx->blockState.matchState.dictMatchState = &cdict->matchState; cctx->blockState.matchState.dictMatchState = &cdict->matchState;
/* prep working match state so dict matches never have negative indices /* prep working match state so dict matches never have negative indices
* when they are translated to the working context's index space. */ * when they are translated to the working context's index space. */
if (cctx->blockState.matchState.window.dictLimit < cdictEnd) { if (cctx->blockState.matchState.window.dictLimit < cdictEnd) {
cctx->blockState.matchState.window.nextSrc = cctx->blockState.matchState.window.nextSrc =
cctx->blockState.matchState.window.base + cdictEnd; cctx->blockState.matchState.window.base + cdictEnd;
ZSTD_window_clear(&cctx->blockState.matchState.window); ZSTD_window_clear(&cctx->blockState.matchState.window);
} }
/* loadedDictEnd is expressed within the referential of the active context */
cctx->blockState.matchState.loadedDictEnd = cctx->blockState.matchState.window.dictLimit; cctx->blockState.matchState.loadedDictEnd = cctx->blockState.matchState.window.dictLimit;
} } }
}
cctx->dictID = cdict->dictID; cctx->dictID = cdict->dictID;
/* copy block state */ /* copy block state */
memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState)); memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState));
return 0; return 0;
} }
memset(cctx->blockState.matchState.hashTable3, 0, h3Size * sizeof(U32)); memset(cctx->blockState.matchState.hashTable3, 0, h3Size * sizeof(U32));
} }
/* copy dictionary offsets */ /* copy dictionary offsets */
{ ZSTD_matchState_t const* srcMatchState = &cdict->matchState; { ZSTD_matchState_t const* srcMatchState = &cdict->matchState;
ZSTD_matchState_t* dstMatchState = &cctx->blockState.matchState; ZSTD_matchState_t* dstMatchState = &cctx->blockState.matchState;
dstMatchState->window = srcMatchState->window; dstMatchState->window = srcMatchState->window;
dstMatchState->nextToUpdate = srcMatchState->nextToUpdate; dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;
dstMatchState->nextToUpdate3= srcMatchState->nextToUpdate3;
dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd; dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;
} }
cctx->dictID = cdict->dictID; cctx->dictID = cdict->dictID;
/* copy block state */ /* copy block state */
memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState)); memcpy(cctx->blockState.prevCBlock, &cdict->cBlockState, sizeof(cdict->cBlockState));
* @return : 0, or an error code */ * @return : 0, or an error code */
static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx, static size_t ZSTD_copyCCtx_internal(ZSTD_CCtx* dstCCtx,
const ZSTD_CCtx* srcCCtx, const ZSTD_CCtx* srcCCtx,
ZSTD_frameParameters fParams, ZSTD_frameParameters fParams,
U64 pledgedSrcSize, U64 pledgedSrcSize,
ZSTD_buffered_policy_e zbuff) ZSTD_buffered_policy_e zbuff)
{ {
DEBUGLOG(5, "ZSTD_copyCCtx_internal"); DEBUGLOG(5, "ZSTD_copyCCtx_internal");
if (srcCCtx->stage!=ZSTDcs_init) return ERROR(stage_wrong); RETURN_ERROR_IF(srcCCtx->stage!=ZSTDcs_init, stage_wrong);
memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem)); memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
{ ZSTD_CCtx_params params = dstCCtx->requestedParams; { ZSTD_CCtx_params params = dstCCtx->requestedParams;
/* Copy only compression parameters related to tables. */ /* Copy only compression parameters related to tables. */
params.cParams = srcCCtx->appliedParams.cParams; params.cParams = srcCCtx->appliedParams.cParams;
params.fParams = fParams; params.fParams = fParams;
ZSTD_resetCCtx_internal(dstCCtx, params, pledgedSrcSize, ZSTD_resetCCtx_internal(dstCCtx, params, pledgedSrcSize,
ZSTDcrp_noMemset, zbuff); ZSTDcrp_noMemset, zbuff);
} }
/* copy dictionary offsets */ /* copy dictionary offsets */
{ {
const ZSTD_matchState_t* srcMatchState = &srcCCtx->blockState.matchState; const ZSTD_matchState_t* srcMatchState = &srcCCtx->blockState.matchState;
ZSTD_matchState_t* dstMatchState = &dstCCtx->blockState.matchState; ZSTD_matchState_t* dstMatchState = &dstCCtx->blockState.matchState;
dstMatchState->window = srcMatchState->window; dstMatchState->window = srcMatchState->window;
dstMatchState->nextToUpdate = srcMatchState->nextToUpdate; dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;
dstMatchState->nextToUpdate3= srcMatchState->nextToUpdate3;
dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd; dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;
} }
dstCCtx->dictID = srcCCtx->dictID; dstCCtx->dictID = srcCCtx->dictID;
/* copy block state */ /* copy block state */
memcpy(dstCCtx->blockState.prevCBlock, srcCCtx->blockState.prevCBlock, sizeof(*srcCCtx->blockState.prevCBlock)); memcpy(dstCCtx->blockState.prevCBlock, srcCCtx->blockState.prevCBlock, sizeof(*srcCCtx->blockState.prevCBlock));
return 0; return 0;
static void ZSTD_reduceTable_btlazy2(U32* const table, U32 const size, U32 const reducerValue) static void ZSTD_reduceTable_btlazy2(U32* const table, U32 const size, U32 const reducerValue)
{ {
ZSTD_reduceTable_internal(table, size, reducerValue, 1); ZSTD_reduceTable_internal(table, size, reducerValue, 1);
} }
/*! ZSTD_reduceIndex() : /*! ZSTD_reduceIndex() :
* rescale all indexes to avoid future overflow (indexes are U32) */ * rescale all indexes to avoid future overflow (indexes are U32) */
static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue) static void ZSTD_reduceIndex (ZSTD_matchState_t* ms, ZSTD_CCtx_params const* params, const U32 reducerValue)
{ {
ZSTD_matchState_t* const ms = &zc->blockState.matchState; { U32 const hSize = (U32)1 << params->cParams.hashLog;
{ U32 const hSize = (U32)1 << zc->appliedParams.cParams.hashLog;
ZSTD_reduceTable(ms->hashTable, hSize, reducerValue); ZSTD_reduceTable(ms->hashTable, hSize, reducerValue);
} }
if (zc->appliedParams.cParams.strategy != ZSTD_fast) { if (params->cParams.strategy != ZSTD_fast) {
U32 const chainSize = (U32)1 << zc->appliedParams.cParams.chainLog; U32 const chainSize = (U32)1 << params->cParams.chainLog;
if (zc->appliedParams.cParams.strategy == ZSTD_btlazy2) if (params->cParams.strategy == ZSTD_btlazy2)
ZSTD_reduceTable_btlazy2(ms->chainTable, chainSize, reducerValue); ZSTD_reduceTable_btlazy2(ms->chainTable, chainSize, reducerValue);
else else
ZSTD_reduceTable(ms->chainTable, chainSize, reducerValue); ZSTD_reduceTable(ms->chainTable, chainSize, reducerValue);
} }
if (ms->hashLog3) { if (ms->hashLog3) {
U32 const h3Size = (U32)1 << ms->hashLog3; U32 const h3Size = (U32)1 << ms->hashLog3;
ZSTD_reduceTable(ms->hashTable3, h3Size, reducerValue); ZSTD_reduceTable(ms->hashTable3, h3Size, reducerValue);
} }
} }
/*-******************************************************* /*-*******************************************************
* Block entropic compression * Block entropic compression
*********************************************************/ *********************************************************/
/* See doc/zstd_compression_format.md for detailed format description */ /* 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) 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); U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(srcSize << 3);
if (srcSize + ZSTD_blockHeaderSize > dstCapacity) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(srcSize + ZSTD_blockHeaderSize > dstCapacity,
dstSize_tooSmall);
MEM_writeLE24(dst, cBlockHeader24); MEM_writeLE24(dst, cBlockHeader24);
memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize); memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize);
return ZSTD_blockHeaderSize + 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);
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;
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);
}
if (dstCapacity < lhSize+1) return ERROR(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) void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)
{ {
const seqDef* const sequences = seqStorePtr->sequencesStart; const seqDef* const sequences = seqStorePtr->sequencesStart;
BYTE* const llCodeTable = seqStorePtr->llCode; BYTE* const llCodeTable = seqStorePtr->llCode;
BYTE* const ofCodeTable = seqStorePtr->ofCode; BYTE* const ofCodeTable = seqStorePtr->ofCode;
BYTE* const mlCodeTable = seqStorePtr->mlCode; BYTE* const mlCodeTable = seqStorePtr->mlCode;
U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
U32 u; U32 u;
assert(nbSeq <= seqStorePtr->maxNbSeq); assert(nbSeq <= seqStorePtr->maxNbSeq);
for (u=0; u<nbSeq; u++) { for (u=0; u<nbSeq; u++) {
U32 const llv = sequences[u].litLength; U32 const llv = sequences[u].litLength;
U32 const mlv = sequences[u].matchLength; U32 const mlv = sequences[u].matchLength;
llCodeTable[u] = (BYTE)ZSTD_LLcode(llv); llCodeTable[u] = (BYTE)ZSTD_LLcode(llv);
ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset); ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset);
mlCodeTable[u] = (BYTE)ZSTD_MLcode(mlv); mlCodeTable[u] = (BYTE)ZSTD_MLcode(mlv);
} }
if (seqStorePtr->longLengthID==1) if (seqStorePtr->longLengthID==1)
llCodeTable[seqStorePtr->longLengthPos] = MaxLL; llCodeTable[seqStorePtr->longLengthPos] = MaxLL;
if (seqStorePtr->longLengthID==2) if (seqStorePtr->longLengthID==2)
mlCodeTable[seqStorePtr->longLengthPos] = MaxML; mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
} }
static int ZSTD_disableLiteralsCompression(const ZSTD_CCtx_params* cctxParams)
/**
* -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 kInverseProbabiltyLog256[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] * kInverseProbabiltyLog256[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] * kInverseProbabiltyLog256[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);
if (ZSTD_getFSEMaxSymbolValue(ctable) < max) {
DEBUGLOG(5, "Repeat FSE_CTable has maxSymbolValue %u < %u",
ZSTD_getFSEMaxSymbolValue(ctable), max);
return ERROR(GENERIC);
}
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;
if (bitCost >= badCost) {
DEBUGLOG(5, "Repeat FSE_CTable has Prob[%u] == 0", s);
return ERROR(GENERIC);
}
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]; switch (cctxParams->literalCompressionMode) {
S16 norm[MaxSeq + 1]; case ZSTD_lcm_huffman:
const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
CHECK_F(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:
CHECK_F(FSE_buildCTable_rle(nextCTable, (BYTE)max));
if (dstCapacity==0) return ERROR(dstSize_tooSmall);
*op = codeTable[0];
return 1;
case set_repeat:
memcpy(nextCTable, prevCTable, prevCTableSize);
return 0;
case set_basic:
CHECK_F(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, workspace, workspaceSize)); /* note : could be pre-calculated */
return 0; return 0;
case set_compressed: { case ZSTD_lcm_uncompressed:
S16 norm[MaxSeq + 1]; return 1;
size_t nbSeq_1 = nbSeq; default:
const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max); assert(0 /* impossible: pre-validated */);
if (count[codeTable[nbSeq-1]] > 1) { /* fall-through */
count[codeTable[nbSeq-1]]--; case ZSTD_lcm_auto:
nbSeq_1--; return (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0);
}
assert(nbSeq_1 > 1);
CHECK_F(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 NCountSize;
CHECK_F(FSE_buildCTable_wksp(nextCTable, norm, max, tableLog, workspace, workspaceSize));
return NCountSize;
}
}
default: return assert(0), 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;
CHECK_E(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);
if (streamSize==0) return ERROR(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);
} }
/* ZSTD_compressSequences_internal(): /* ZSTD_compressSequences_internal():
* actually compresses both literals and sequences */ * actually compresses both literals and sequences */
MEM_STATIC size_t MEM_STATIC size_t
ZSTD_compressSequences_internal(seqStore_t* seqStorePtr, ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy, const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy, ZSTD_entropyCTables_t* nextEntropy,
U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */ U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */
const seqDef* const sequences = seqStorePtr->sequencesStart; const seqDef* const sequences = seqStorePtr->sequencesStart;
const BYTE* const ofCodeTable = seqStorePtr->ofCode; const BYTE* const ofCodeTable = seqStorePtr->ofCode;
const BYTE* const llCodeTable = seqStorePtr->llCode; const BYTE* const llCodeTable = seqStorePtr->llCode;
const BYTE* const mlCodeTable = seqStorePtr->mlCode; const BYTE* const mlCodeTable = seqStorePtr->mlCode;
BYTE* const ostart = (BYTE*)dst; BYTE* const ostart = (BYTE*)dst;
BYTE* const oend = ostart + dstCapacity; BYTE* const oend = ostart + dstCapacity;
BYTE* op = ostart; BYTE* op = ostart;
size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart; size_t const nbSeq = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
BYTE* seqHead; BYTE* seqHead;
BYTE* lastNCount = NULL; BYTE* lastNCount = NULL;
DEBUGLOG(5, "ZSTD_compressSequences_internal (nbSeq=%zu)", nbSeq);
ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog))); ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog)));
DEBUGLOG(5, "ZSTD_compressSequences_internal");
/* Compress literals */ /* Compress literals */
{ const BYTE* const literals = seqStorePtr->litStart; { const BYTE* const literals = seqStorePtr->litStart;
size_t const litSize = seqStorePtr->lit - literals; size_t const litSize = (size_t)(seqStorePtr->lit - literals);
int const disableLiteralCompression = (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0);
size_t const cSize = ZSTD_compressLiterals( size_t const cSize = ZSTD_compressLiterals(
&prevEntropy->huf, &nextEntropy->huf, &prevEntropy->huf, &nextEntropy->huf,
cctxParams->cParams.strategy, disableLiteralCompression, cctxParams->cParams.strategy,
ZSTD_disableLiteralsCompression(cctxParams),
op, dstCapacity, op, dstCapacity,
literals, litSize, literals, litSize,
workspace, wkspSize, workspace, wkspSize,
bmi2); bmi2);
if (ZSTD_isError(cSize)) FORWARD_IF_ERROR(cSize);
return cSize;
assert(cSize <= dstCapacity); assert(cSize <= dstCapacity);
op += cSize; op += cSize;
} }
/* Sequences Header */ /* Sequences Header */
if ((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,
dstSize_tooSmall);
if (nbSeq < 0x7F) if (nbSeq < 0x7F)
*op++ = (BYTE)nbSeq; *op++ = (BYTE)nbSeq;
else if (nbSeq < LONGNBSEQ) else if (nbSeq < LONGNBSEQ)
op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2; op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;
else else
op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3; op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;
assert(op <= oend);
if (nbSeq==0) { if (nbSeq==0) {
/* Copy the old tables over as if we repeated them */ /* Copy the old tables over as if we repeated them */
memcpy(&nextEntropy->fse, &prevEntropy->fse, sizeof(prevEntropy->fse)); memcpy(&nextEntropy->fse, &prevEntropy->fse, sizeof(prevEntropy->fse));
return op - ostart; return (size_t)(op - ostart);
} }
/* seqHead : flags for FSE encoding type */ /* seqHead : flags for FSE encoding type */
seqHead = op++; seqHead = op++;
assert(op <= oend);
/* convert length/distances into codes */ /* convert length/distances into codes */
ZSTD_seqToCodes(seqStorePtr); ZSTD_seqToCodes(seqStorePtr);
/* build CTable for Literal Lengths */ /* build CTable for Literal Lengths */
{ unsigned max = MaxLL; { unsigned max = MaxLL;
size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace, wkspSize); /* can't fail */ size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace, wkspSize); /* can't fail */
DEBUGLOG(5, "Building LL table"); DEBUGLOG(5, "Building LL table");
nextEntropy->fse.litlength_repeatMode = prevEntropy->fse.litlength_repeatMode; nextEntropy->fse.litlength_repeatMode = prevEntropy->fse.litlength_repeatMode;
LLtype = ZSTD_selectEncodingType(&nextEntropy->fse.litlength_repeatMode, LLtype = ZSTD_selectEncodingType(&nextEntropy->fse.litlength_repeatMode,
count, max, mostFrequent, nbSeq, count, max, mostFrequent, nbSeq,
LLFSELog, prevEntropy->fse.litlengthCTable, LLFSELog, prevEntropy->fse.litlengthCTable,
LL_defaultNorm, LL_defaultNormLog, LL_defaultNorm, LL_defaultNormLog,
ZSTD_defaultAllowed, strategy); ZSTD_defaultAllowed, strategy);
assert(set_basic < set_compressed && set_rle < set_compressed); 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 */ 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, { size_t const countSize = ZSTD_buildCTable(op, (size_t)(oend - op), CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL, count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL,
prevEntropy->fse.litlengthCTable, sizeof(prevEntropy->fse.litlengthCTable), prevEntropy->fse.litlengthCTable, sizeof(prevEntropy->fse.litlengthCTable),
workspace, wkspSize); workspace, wkspSize);
if (ZSTD_isError(countSize)) return countSize; FORWARD_IF_ERROR(countSize);
if (LLtype == set_compressed) if (LLtype == set_compressed)
lastNCount = op; lastNCount = op;
op += countSize; op += countSize;
assert(op <= oend);
} } } }
/* build CTable for Offsets */ /* build CTable for Offsets */
{ unsigned max = MaxOff; { unsigned max = MaxOff;
size_t const mostFrequent = HIST_countFast_wksp(count, &max, ofCodeTable, nbSeq, workspace, wkspSize); /* can't fail */ 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 */ /* 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; ZSTD_defaultPolicy_e const defaultPolicy = (max <= DefaultMaxOff) ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed;
DEBUGLOG(5, "Building OF table"); DEBUGLOG(5, "Building OF table");
nextEntropy->fse.offcode_repeatMode = prevEntropy->fse.offcode_repeatMode; nextEntropy->fse.offcode_repeatMode = prevEntropy->fse.offcode_repeatMode;
Offtype = ZSTD_selectEncodingType(&nextEntropy->fse.offcode_repeatMode, Offtype = ZSTD_selectEncodingType(&nextEntropy->fse.offcode_repeatMode,
count, max, mostFrequent, nbSeq, count, max, mostFrequent, nbSeq,
OffFSELog, prevEntropy->fse.offcodeCTable, OffFSELog, prevEntropy->fse.offcodeCTable,
OF_defaultNorm, OF_defaultNormLog, OF_defaultNorm, OF_defaultNormLog,
defaultPolicy, strategy); defaultPolicy, strategy);
assert(!(Offtype < set_compressed && nextEntropy->fse.offcode_repeatMode != FSE_repeat_none)); /* We don't copy tables */ 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, { size_t const countSize = ZSTD_buildCTable(op, (size_t)(oend - op), CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff, count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
prevEntropy->fse.offcodeCTable, sizeof(prevEntropy->fse.offcodeCTable), prevEntropy->fse.offcodeCTable, sizeof(prevEntropy->fse.offcodeCTable),
workspace, wkspSize); workspace, wkspSize);
if (ZSTD_isError(countSize)) return countSize; FORWARD_IF_ERROR(countSize);
if (Offtype == set_compressed) if (Offtype == set_compressed)
lastNCount = op; lastNCount = op;
op += countSize; op += countSize;
assert(op <= oend);
} } } }
/* build CTable for MatchLengths */ /* build CTable for MatchLengths */
{ unsigned max = MaxML; { unsigned max = MaxML;
size_t const mostFrequent = HIST_countFast_wksp(count, &max, mlCodeTable, nbSeq, workspace, wkspSize); /* can't fail */ 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)); DEBUGLOG(5, "Building ML table (remaining space : %i)", (int)(oend-op));
nextEntropy->fse.matchlength_repeatMode = prevEntropy->fse.matchlength_repeatMode; nextEntropy->fse.matchlength_repeatMode = prevEntropy->fse.matchlength_repeatMode;
MLtype = ZSTD_selectEncodingType(&nextEntropy->fse.matchlength_repeatMode, MLtype = ZSTD_selectEncodingType(&nextEntropy->fse.matchlength_repeatMode,
count, max, mostFrequent, nbSeq, count, max, mostFrequent, nbSeq,
MLFSELog, prevEntropy->fse.matchlengthCTable, MLFSELog, prevEntropy->fse.matchlengthCTable,
ML_defaultNorm, ML_defaultNormLog, ML_defaultNorm, ML_defaultNormLog,
ZSTD_defaultAllowed, strategy); ZSTD_defaultAllowed, strategy);
assert(!(MLtype < set_compressed && nextEntropy->fse.matchlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */ 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, { size_t const countSize = ZSTD_buildCTable(op, (size_t)(oend - op), CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML, count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML,
prevEntropy->fse.matchlengthCTable, sizeof(prevEntropy->fse.matchlengthCTable), prevEntropy->fse.matchlengthCTable, sizeof(prevEntropy->fse.matchlengthCTable),
workspace, wkspSize); workspace, wkspSize);
if (ZSTD_isError(countSize)) return countSize; FORWARD_IF_ERROR(countSize);
if (MLtype == set_compressed) if (MLtype == set_compressed)
lastNCount = op; lastNCount = op;
op += countSize; op += countSize;
assert(op <= oend);
} } } }
*seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2)); *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2));
{ size_t const bitstreamSize = ZSTD_encodeSequences( { size_t const bitstreamSize = ZSTD_encodeSequences(
op, oend - op, op, (size_t)(oend - op),
CTable_MatchLength, mlCodeTable, CTable_MatchLength, mlCodeTable,
CTable_OffsetBits, ofCodeTable, CTable_OffsetBits, ofCodeTable,
CTable_LitLength, llCodeTable, CTable_LitLength, llCodeTable,
sequences, nbSeq, sequences, nbSeq,
longOffsets, bmi2); longOffsets, bmi2);
if (ZSTD_isError(bitstreamSize)) return bitstreamSize; FORWARD_IF_ERROR(bitstreamSize);
op += bitstreamSize; op += bitstreamSize;
assert(op <= oend);
/* zstd versions <= 1.3.4 mistakenly report corruption when /* zstd versions <= 1.3.4 mistakenly report corruption when
* FSE_readNCount() recieves a buffer < 4 bytes. * FSE_readNCount() receives a buffer < 4 bytes.
* Fixed by https://github.com/facebook/zstd/pull/1146. * Fixed by https://github.com/facebook/zstd/pull/1146.
* This can happen when the last set_compressed table present is 2 * This can happen when the last set_compressed table present is 2
* bytes and the bitstream is only one byte. * bytes and the bitstream is only one byte.
* In this exceedingly rare case, we will simply emit an uncompressed * In this exceedingly rare case, we will simply emit an uncompressed
* block, since it isn't worth optimizing. * block, since it isn't worth optimizing.
*/ */
if (lastNCount && (op - lastNCount) < 4) { if (lastNCount && (op - lastNCount) < 4) {
/* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */ /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */
assert(op - lastNCount == 3); assert(op - lastNCount == 3);
DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by " DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by "
"emitting an uncompressed block."); "emitting an uncompressed block.");
return 0; return 0;
} }
} }
DEBUGLOG(5, "compressed block size : %u", (unsigned)(op - ostart)); DEBUGLOG(5, "compressed block size : %u", (unsigned)(op - ostart));
return op - ostart; return (size_t)(op - ostart);
} }
MEM_STATIC size_t MEM_STATIC size_t
ZSTD_compressSequences(seqStore_t* seqStorePtr, ZSTD_compressSequences(seqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy, const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy, ZSTD_entropyCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams, const ZSTD_CCtx_params* cctxParams,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
size_t srcSize, size_t srcSize,
void* workspace, size_t wkspSize, void* workspace, size_t wkspSize,
int bmi2) int bmi2)
{ {
size_t const cSize = ZSTD_compressSequences_internal( size_t const cSize = ZSTD_compressSequences_internal(
seqStorePtr, prevEntropy, nextEntropy, cctxParams, seqStorePtr, prevEntropy, nextEntropy, cctxParams,
dst, dstCapacity, dst, dstCapacity,
workspace, wkspSize, bmi2); workspace, wkspSize, bmi2);
if (cSize == 0) return 0; if (cSize == 0) return 0;
/* When srcSize <= dstCapacity, there is enough space to write a raw uncompressed block. /* 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. * 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)) if ((cSize == ERROR(dstSize_tooSmall)) & (srcSize <= dstCapacity))
return 0; /* block not compressed */ return 0; /* block not compressed */
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
/* Check compressibility */ /* Check compressibility */
{ size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, cctxParams->cParams.strategy); { size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, cctxParams->cParams.strategy);
if (cSize >= maxCSize) return 0; /* block not compressed */ if (cSize >= maxCSize) return 0; /* block not compressed */
} }
return cSize; return cSize;
} }
void ZSTD_resetSeqStore(seqStore_t* ssPtr) void ZSTD_resetSeqStore(seqStore_t* ssPtr)
{ {
ssPtr->lit = ssPtr->litStart; ssPtr->lit = ssPtr->litStart;
ssPtr->sequences = ssPtr->sequencesStart; ssPtr->sequences = ssPtr->sequencesStart;
ssPtr->longLengthID = 0; ssPtr->longLengthID = 0;
} }
static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc, typedef enum { ZSTDbss_compress, ZSTDbss_noCompress } ZSTD_buildSeqStore_e;
void* dst, size_t dstCapacity,
const void* src, size_t srcSize) static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
{ {
ZSTD_matchState_t* const ms = &zc->blockState.matchState; ZSTD_matchState_t* const ms = &zc->blockState.matchState;
size_t cSize; DEBUGLOG(5, "ZSTD_buildSeqStore (srcSize=%zu)", srcSize);
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(srcSize <= ZSTD_BLOCKSIZE_MAX);
/* Assert that we have correctly flushed the ctx params into the ms's copy */ /* Assert that we have correctly flushed the ctx params into the ms's copy */
ZSTD_assertEqualCParams(zc->appliedParams.cParams, ms->cParams); ZSTD_assertEqualCParams(zc->appliedParams.cParams, ms->cParams);
if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) { if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) {
ZSTD_ldm_skipSequences(&zc->externSeqStore, srcSize, zc->appliedParams.cParams.minMatch); ZSTD_ldm_skipSequences(&zc->externSeqStore, srcSize, zc->appliedParams.cParams.minMatch);
cSize = 0; return ZSTDbss_noCompress; /* don't even attempt compression below a certain srcSize */
goto out; /* don't even attempt compression below a certain srcSize */
} }
ZSTD_resetSeqStore(&(zc->seqStore)); ZSTD_resetSeqStore(&(zc->seqStore));
ms->opt.symbolCosts = &zc->blockState.prevCBlock->entropy; /* required for optimal parser to read stats from dictionary */ /* 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, /* a gap between an attached dict and the current window is not safe,
* they must remain adjacent, * they must remain adjacent,
* and when that stops being the case, the dict must be unset */ * and when that stops being the case, the dict must be unset */
assert(ms->dictMatchState == NULL || ms->loadedDictEnd == ms->window.dictLimit); assert(ms->dictMatchState == NULL || ms->loadedDictEnd == ms->window.dictLimit);
/* limited update after a very long match */ /* limited update after a very long match */
{ const BYTE* const base = ms->window.base; { const BYTE* const base = ms->window.base;
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
src, srcSize); src, srcSize);
assert(zc->externSeqStore.pos <= zc->externSeqStore.size); assert(zc->externSeqStore.pos <= zc->externSeqStore.size);
} else if (zc->appliedParams.ldmParams.enableLdm) { } else if (zc->appliedParams.ldmParams.enableLdm) {
rawSeqStore_t ldmSeqStore = {NULL, 0, 0, 0}; rawSeqStore_t ldmSeqStore = {NULL, 0, 0, 0};
ldmSeqStore.seq = zc->ldmSequences; ldmSeqStore.seq = zc->ldmSequences;
ldmSeqStore.capacity = zc->maxNbLdmSequences; ldmSeqStore.capacity = zc->maxNbLdmSequences;
/* Updates ldmSeqStore.size */ /* Updates ldmSeqStore.size */
CHECK_F(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore, FORWARD_IF_ERROR(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore,
&zc->appliedParams.ldmParams, &zc->appliedParams.ldmParams,
src, srcSize)); src, srcSize));
/* Updates ldmSeqStore.pos */ /* Updates ldmSeqStore.pos */
lastLLSize = lastLLSize =
ZSTD_ldm_blockCompress(&ldmSeqStore, ZSTD_ldm_blockCompress(&ldmSeqStore,
ms, &zc->seqStore, ms, &zc->seqStore,
zc->blockState.nextCBlock->rep, zc->blockState.nextCBlock->rep,
src, srcSize); src, srcSize);
assert(ldmSeqStore.pos == ldmSeqStore.size); assert(ldmSeqStore.pos == ldmSeqStore.size);
} else { /* not long range mode */ } else { /* not long range mode */
ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, dictMode); ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->appliedParams.cParams.strategy, dictMode);
lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize); lastLLSize = blockCompressor(ms, &zc->seqStore, zc->blockState.nextCBlock->rep, src, srcSize);
} }
{ const BYTE* const lastLiterals = (const BYTE*)src + srcSize - lastLLSize; { const BYTE* const lastLiterals = (const BYTE*)src + srcSize - lastLLSize;
ZSTD_storeLastLiterals(&zc->seqStore, lastLiterals, lastLLSize); ZSTD_storeLastLiterals(&zc->seqStore, lastLiterals, lastLLSize);
} } } }
return ZSTDbss_compress;
}
static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize)
{
size_t cSize;
DEBUGLOG(5, "ZSTD_compressBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",
(unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit,
(unsigned)zc->blockState.matchState.nextToUpdate);
{ const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize);
FORWARD_IF_ERROR(bss);
if (bss == ZSTDbss_noCompress) { cSize = 0; goto out; }
}
/* encode sequences and literals */ /* encode sequences and literals */
cSize = ZSTD_compressSequences(&zc->seqStore, cSize = ZSTD_compressSequences(&zc->seqStore,
&zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy, &zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy,
&zc->appliedParams, &zc->appliedParams,
dst, dstCapacity, dst, dstCapacity,
srcSize, srcSize,
zc->entropyWorkspace, HUF_WORKSPACE_SIZE /* statically allocated in resetCCtx */, zc->entropyWorkspace, HUF_WORKSPACE_SIZE /* statically allocated in resetCCtx */,
*/ */
if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid) if (zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode == FSE_repeat_valid)
zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check; zc->blockState.prevCBlock->entropy.fse.offcode_repeatMode = FSE_repeat_check;
return cSize; return cSize;
} }
static void ZSTD_overflowCorrectIfNeeded(ZSTD_matchState_t* ms, ZSTD_CCtx_params const* params, void const* ip, void const* iend)
{
if (ZSTD_window_needOverflowCorrection(ms->window, iend)) {
U32 const maxDist = (U32)1 << params->cParams.windowLog;
U32 const cycleLog = ZSTD_cycleLog(params->cParams.chainLog, params->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(ms, params, correction);
if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;
else ms->nextToUpdate -= correction;
/* invalidate dictionaries on overflow correction */
ms->loadedDictEnd = 0;
ms->dictMatchState = NULL;
}
}
/*! ZSTD_compress_frameChunk() : /*! ZSTD_compress_frameChunk() :
* Compress a chunk of data into one or multiple blocks. * Compress a chunk of data into one or multiple blocks.
* All blocks will be terminated, all input will be consumed. * 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. * Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.
* Frame is supposed already started (header already produced) * Frame is supposed already started (header already produced)
* @return : compressed size, or an error code * @return : compressed size, or an error code
*/ */
static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx, static size_t ZSTD_compress_frameChunk (ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
U32 lastFrameChunk) U32 lastFrameChunk)
{ {
size_t blockSize = cctx->blockSize; size_t blockSize = cctx->blockSize;
size_t remaining = srcSize; size_t remaining = srcSize;
const BYTE* ip = (const BYTE*)src; const BYTE* ip = (const BYTE*)src;
BYTE* const ostart = (BYTE*)dst; BYTE* const ostart = (BYTE*)dst;
BYTE* op = ostart; BYTE* op = ostart;
U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog; U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog;
assert(cctx->appliedParams.cParams.windowLog <= 31); assert(cctx->appliedParams.cParams.windowLog <= ZSTD_WINDOWLOG_MAX);
DEBUGLOG(5, "ZSTD_compress_frameChunk (blockSize=%u)", (unsigned)blockSize); DEBUGLOG(5, "ZSTD_compress_frameChunk (blockSize=%u)", (unsigned)blockSize);
if (cctx->appliedParams.fParams.checksumFlag && srcSize) if (cctx->appliedParams.fParams.checksumFlag && srcSize)
XXH64_update(&cctx->xxhState, src, srcSize); XXH64_update(&cctx->xxhState, src, srcSize);
while (remaining) { while (remaining) {
ZSTD_matchState_t* const ms = &cctx->blockState.matchState; ZSTD_matchState_t* const ms = &cctx->blockState.matchState;
U32 const lastBlock = lastFrameChunk & (blockSize >= remaining); U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);
if (dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE) RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE,
return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */ dstSize_tooSmall,
"not enough space to store compressed block");
if (remaining < blockSize) blockSize = remaining; if (remaining < blockSize) blockSize = remaining;
if (ZSTD_window_needOverflowCorrection(ms->window, ip + blockSize)) { ZSTD_overflowCorrectIfNeeded(ms, &cctx->appliedParams, ip, ip + blockSize);
U32 const cycleLog = ZSTD_cycleLog(cctx->appliedParams.cParams.chainLog, cctx->appliedParams.cParams.strategy); ZSTD_checkDictValidity(&ms->window, ip + blockSize, maxDist, &ms->loadedDictEnd, &ms->dictMatchState);
U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, maxDist, ip);
ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30); /* Ensure hash/chain table insertion resumes no sooner than lowlimit */
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; if (ms->nextToUpdate < ms->window.lowLimit) ms->nextToUpdate = ms->window.lowLimit;
{ size_t cSize = ZSTD_compressBlock_internal(cctx, { size_t cSize = ZSTD_compressBlock_internal(cctx,
op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize, op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize,
ip, blockSize); ip, blockSize);
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
if (cSize == 0) { /* block is not compressible */ if (cSize == 0) { /* block is not compressible */
cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock); cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock);
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
} else { } else {
U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3); U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
MEM_writeLE24(op, cBlockHeader24); MEM_writeLE24(op, cBlockHeader24);
cSize += ZSTD_blockHeaderSize; cSize += ZSTD_blockHeaderSize;
} }
ip += blockSize; ip += blockSize;
assert(remaining >= blockSize); assert(remaining >= blockSize);
remaining -= blockSize; remaining -= blockSize;
op += cSize; op += cSize;
assert(dstCapacity >= cSize); assert(dstCapacity >= cSize);
dstCapacity -= cSize; dstCapacity -= cSize;
DEBUGLOG(5, "ZSTD_compress_frameChunk: adding a block of size %u", DEBUGLOG(5, "ZSTD_compress_frameChunk: adding a block of size %u",
(unsigned)cSize); (unsigned)cSize);
} } } }
if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending; if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending;
return op-ostart; return (size_t)(op-ostart);
} }
static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity, static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity,
ZSTD_CCtx_params params, U64 pledgedSrcSize, U32 dictID) ZSTD_CCtx_params params, U64 pledgedSrcSize, U32 dictID)
{ BYTE* const op = (BYTE*)dst; { BYTE* const op = (BYTE*)dst;
U32 const dictIDSizeCodeLength = (dictID>0) + (dictID>=256) + (dictID>=65536); /* 0-3 */ U32 const dictIDSizeCodeLength = (dictID>0) + (dictID>=256) + (dictID>=65536); /* 0-3 */
U32 const dictIDSizeCode = params.fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength; /* 0-3 */ U32 const dictIDSizeCode = params.fParams.noDictIDFlag ? 0 : dictIDSizeCodeLength; /* 0-3 */
U32 const checksumFlag = params.fParams.checksumFlag>0; U32 const checksumFlag = params.fParams.checksumFlag>0;
U32 const windowSize = (U32)1 << params.cParams.windowLog; U32 const windowSize = (U32)1 << params.cParams.windowLog;
U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize); U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);
BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3); BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);
U32 const fcsCode = params.fParams.contentSizeFlag ? U32 const fcsCode = params.fParams.contentSizeFlag ?
(pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : 0; /* 0-3 */ (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) : 0; /* 0-3 */
BYTE const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) ); BYTE const frameHeaderDescriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );
size_t pos=0; size_t pos=0;
assert(!(params.fParams.contentSizeFlag && pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN)); assert(!(params.fParams.contentSizeFlag && pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN));
if (dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX, dstSize_tooSmall);
DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u", DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u",
!params.fParams.noDictIDFlag, (unsigned)dictID, (unsigned)dictIDSizeCode); !params.fParams.noDictIDFlag, (unsigned)dictID, (unsigned)dictIDSizeCode);
if (params.format == ZSTD_f_zstd1) { if (params.format == ZSTD_f_zstd1) {
MEM_writeLE32(dst, ZSTD_MAGICNUMBER); MEM_writeLE32(dst, ZSTD_MAGICNUMBER);
pos = 4; pos = 4;
} }
op[pos++] = frameHeaderDecriptionByte; op[pos++] = frameHeaderDescriptionByte;
if (!singleSegment) op[pos++] = windowLogByte; if (!singleSegment) op[pos++] = windowLogByte;
switch(dictIDSizeCode) switch(dictIDSizeCode)
{ {
default: assert(0); /* impossible */ default: assert(0); /* impossible */
case 0 : break; case 0 : break;
case 1 : op[pos] = (BYTE)(dictID); pos++; break; case 1 : op[pos] = (BYTE)(dictID); pos++; break;
case 2 : MEM_writeLE16(op+pos, (U16)dictID); pos+=2; break; case 2 : MEM_writeLE16(op+pos, (U16)dictID); pos+=2; break;
case 3 : MEM_writeLE32(op+pos, dictID); pos+=4; break; case 3 : MEM_writeLE32(op+pos, dictID); pos+=4; break;
} }
switch(fcsCode) switch(fcsCode)
{ {
default: assert(0); /* impossible */ default: assert(0); /* impossible */
case 0 : if (singleSegment) op[pos++] = (BYTE)(pledgedSrcSize); break; case 0 : if (singleSegment) op[pos++] = (BYTE)(pledgedSrcSize); break;
case 1 : MEM_writeLE16(op+pos, (U16)(pledgedSrcSize-256)); pos+=2; 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 2 : MEM_writeLE32(op+pos, (U32)(pledgedSrcSize)); pos+=4; break;
case 3 : MEM_writeLE64(op+pos, (U64)(pledgedSrcSize)); pos+=8; break; case 3 : MEM_writeLE64(op+pos, (U64)(pledgedSrcSize)); pos+=8; break;
} }
return pos; return pos;
} }
/* ZSTD_writeLastEmptyBlock() : /* ZSTD_writeLastEmptyBlock() :
* output an empty Block with end-of-frame mark to complete a frame * 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)) * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))
* or an error code if `dstCapcity` is too small (<ZSTD_blockHeaderSize) * or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)
*/ */
size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity) size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity)
{ {
if (dstCapacity < ZSTD_blockHeaderSize) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize, dstSize_tooSmall);
{ U32 const cBlockHeader24 = 1 /*lastBlock*/ + (((U32)bt_raw)<<1); /* 0 size */ { U32 const cBlockHeader24 = 1 /*lastBlock*/ + (((U32)bt_raw)<<1); /* 0 size */
MEM_writeLE24(dst, cBlockHeader24); MEM_writeLE24(dst, cBlockHeader24);
return ZSTD_blockHeaderSize; return ZSTD_blockHeaderSize;
} }
} }
size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq) size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq)
{ {
if (cctx->stage != ZSTDcs_init) RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong);
return ERROR(stage_wrong); RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm,
if (cctx->appliedParams.ldmParams.enableLdm) parameter_unsupported);
return ERROR(parameter_unsupported);
cctx->externSeqStore.seq = seq; cctx->externSeqStore.seq = seq;
cctx->externSeqStore.size = nbSeq; cctx->externSeqStore.size = nbSeq;
cctx->externSeqStore.capacity = nbSeq; cctx->externSeqStore.capacity = nbSeq;
cctx->externSeqStore.pos = 0; cctx->externSeqStore.pos = 0;
return 0; return 0;
} }
static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx, static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
U32 frame, U32 lastFrameChunk) U32 frame, U32 lastFrameChunk)
{ {
ZSTD_matchState_t* const ms = &cctx->blockState.matchState; ZSTD_matchState_t* const ms = &cctx->blockState.matchState;
size_t fhSize = 0; size_t fhSize = 0;
DEBUGLOG(5, "ZSTD_compressContinue_internal, stage: %u, srcSize: %u", DEBUGLOG(5, "ZSTD_compressContinue_internal, stage: %u, srcSize: %u",
cctx->stage, (unsigned)srcSize); cctx->stage, (unsigned)srcSize);
if (cctx->stage==ZSTDcs_created) return ERROR(stage_wrong); /* missing init (ZSTD_compressBegin) */ RETURN_ERROR_IF(cctx->stage==ZSTDcs_created, stage_wrong,
"missing init (ZSTD_compressBegin)");
if (frame && (cctx->stage==ZSTDcs_init)) { if (frame && (cctx->stage==ZSTDcs_init)) {
fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams,
cctx->pledgedSrcSizePlusOne-1, cctx->dictID); cctx->pledgedSrcSizePlusOne-1, cctx->dictID);
if (ZSTD_isError(fhSize)) return fhSize; FORWARD_IF_ERROR(fhSize);
assert(fhSize <= dstCapacity);
dstCapacity -= fhSize; dstCapacity -= fhSize;
dst = (char*)dst + fhSize; dst = (char*)dst + fhSize;
cctx->stage = ZSTDcs_ongoing; cctx->stage = ZSTDcs_ongoing;
} }
if (!srcSize) return fhSize; /* do not generate an empty block if no input */ if (!srcSize) return fhSize; /* do not generate an empty block if no input */
if (!ZSTD_window_update(&ms->window, src, srcSize)) { if (!ZSTD_window_update(&ms->window, src, srcSize)) {
ms->nextToUpdate = ms->window.dictLimit; ms->nextToUpdate = ms->window.dictLimit;
} }
if (cctx->appliedParams.ldmParams.enableLdm) { if (cctx->appliedParams.ldmParams.enableLdm) {
ZSTD_window_update(&cctx->ldmState.window, src, srcSize); ZSTD_window_update(&cctx->ldmState.window, src, srcSize);
} }
if (!frame) { if (!frame) {
/* overflow check and correction for block mode */ /* overflow check and correction for block mode */
if (ZSTD_window_needOverflowCorrection(ms->window, (const char*)src + srcSize)) { ZSTD_overflowCorrectIfNeeded(ms, &cctx->appliedParams, src, (BYTE const*)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); DEBUGLOG(5, "ZSTD_compressContinue_internal (blockSize=%u)", (unsigned)cctx->blockSize);
{ size_t const cSize = frame ? { size_t const cSize = frame ?
ZSTD_compress_frameChunk (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) : ZSTD_compress_frameChunk (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) :
ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize); ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize);
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
cctx->consumedSrcSize += srcSize; cctx->consumedSrcSize += srcSize;
cctx->producedCSize += (cSize + fhSize); cctx->producedCSize += (cSize + fhSize);
assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0)); assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));
if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */ if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */
ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1); ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);
if (cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne) { RETURN_ERROR_IF(
DEBUGLOG(4, "error : pledgedSrcSize = %u, while realSrcSize >= %u", cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne,
(unsigned)cctx->pledgedSrcSizePlusOne-1, (unsigned)cctx->consumedSrcSize); srcSize_wrong,
return ERROR(srcSize_wrong); "error : pledgedSrcSize = %u, while realSrcSize >= %u",
} (unsigned)cctx->pledgedSrcSizePlusOne-1,
(unsigned)cctx->consumedSrcSize);
} }
return cSize + fhSize; return cSize + fhSize;
} }
} }
size_t ZSTD_compressContinue (ZSTD_CCtx* cctx, size_t ZSTD_compressContinue (ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
DEBUGLOG(5, "ZSTD_compressContinue (srcSize=%u)", (unsigned)srcSize); DEBUGLOG(5, "ZSTD_compressContinue (srcSize=%u)", (unsigned)srcSize);
return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 0 /* last chunk */); return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1 /* frame mode */, 0 /* last chunk */);
} }
size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx) size_t ZSTD_getBlockSize(const ZSTD_CCtx* cctx)
{ {
ZSTD_compressionParameters const cParams = cctx->appliedParams.cParams; ZSTD_compressionParameters const cParams = cctx->appliedParams.cParams;
assert(!ZSTD_checkCParams(cParams)); assert(!ZSTD_checkCParams(cParams));
return MIN (ZSTD_BLOCKSIZE_MAX, (U32)1 << cParams.windowLog); 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 ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
{ {
size_t const blockSizeMax = ZSTD_getBlockSize(cctx); DEBUGLOG(5, "ZSTD_compressBlock: srcSize = %u", (unsigned)srcSize);
if (srcSize > blockSizeMax) return ERROR(srcSize_wrong); { 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 */); return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0 /* frame mode */, 0 /* last chunk */);
} }
/*! ZSTD_loadDictionaryContent() : /*! ZSTD_loadDictionaryContent() :
* @return : 0, or an error code * @return : 0, or an error code
*/ */
static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms, static size_t ZSTD_loadDictionaryContent(ZSTD_matchState_t* ms,
ZSTD_CCtx_params const* params, ZSTD_CCtx_params const* params,
const void* src, size_t srcSize, const void* src, size_t srcSize,
ZSTD_dictTableLoadMethod_e dtlm) ZSTD_dictTableLoadMethod_e dtlm)
{ {
const BYTE* const ip = (const BYTE*) src; const BYTE* ip = (const BYTE*) src;
const BYTE* const iend = ip + srcSize; const BYTE* const iend = ip + srcSize;
ZSTD_window_update(&ms->window, src, srcSize); ZSTD_window_update(&ms->window, src, srcSize);
ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base); ms->loadedDictEnd = params->forceWindow ? 0 : (U32)(iend - ms->window.base);
/* Assert that we the ms params match the params we're being given */ /* Assert that we the ms params match the params we're being given */
ZSTD_assertEqualCParams(params->cParams, ms->cParams); ZSTD_assertEqualCParams(params->cParams, ms->cParams);
if (srcSize <= HASH_READ_SIZE) return 0; if (srcSize <= HASH_READ_SIZE) return 0;
while (iend - ip > HASH_READ_SIZE) {
size_t const remaining = (size_t)(iend - ip);
size_t const chunk = MIN(remaining, ZSTD_CHUNKSIZE_MAX);
const BYTE* const ichunk = ip + chunk;
ZSTD_overflowCorrectIfNeeded(ms, params, ip, ichunk);
switch(params->cParams.strategy) switch(params->cParams.strategy)
{ {
case ZSTD_fast: case ZSTD_fast:
ZSTD_fillHashTable(ms, iend, dtlm); ZSTD_fillHashTable(ms, ichunk, dtlm);
break; break;
case ZSTD_dfast: case ZSTD_dfast:
ZSTD_fillDoubleHashTable(ms, iend, dtlm); ZSTD_fillDoubleHashTable(ms, ichunk, dtlm);
break; break;
case ZSTD_greedy: case ZSTD_greedy:
case ZSTD_lazy: case ZSTD_lazy:
case ZSTD_lazy2: case ZSTD_lazy2:
if (srcSize >= HASH_READ_SIZE) if (chunk >= HASH_READ_SIZE)
ZSTD_insertAndFindFirstIndex(ms, iend-HASH_READ_SIZE); ZSTD_insertAndFindFirstIndex(ms, ichunk-HASH_READ_SIZE);
break; break;
case ZSTD_btlazy2: /* we want the dictionary table fully sorted */ case ZSTD_btlazy2: /* we want the dictionary table fully sorted */
case ZSTD_btopt: case ZSTD_btopt:
case ZSTD_btultra: case ZSTD_btultra:
case ZSTD_btultra2: case ZSTD_btultra2:
if (srcSize >= HASH_READ_SIZE) if (chunk >= HASH_READ_SIZE)
ZSTD_updateTree(ms, iend-HASH_READ_SIZE, iend); ZSTD_updateTree(ms, ichunk-HASH_READ_SIZE, ichunk);
break; break;
default: default:
assert(0); /* not possible : not a valid strategy id */ assert(0); /* not possible : not a valid strategy id */
} }
ip = ichunk;
}
ms->nextToUpdate = (U32)(iend - ms->window.base); ms->nextToUpdate = (U32)(iend - ms->window.base);
return 0; return 0;
} }
/* Dictionaries that assign zero probability to symbols that show up causes problems /* Dictionaries that assign zero probability to symbols that show up causes problems
when FSE encoding. Refuse dictionaries that assign zero probability to symbols when FSE encoding. Refuse dictionaries that assign zero probability to symbols
that we may encounter during compression. that we may encounter during compression.
NOTE: This behavior is not standard and could be improved in the future. */ NOTE: This behavior is not standard and could be improved in the future. */
static size_t ZSTD_checkDictNCount(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue) { static size_t ZSTD_checkDictNCount(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue) {
U32 s; U32 s;
if (dictMaxSymbolValue < maxSymbolValue) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(dictMaxSymbolValue < maxSymbolValue, dictionary_corrupted);
for (s = 0; s <= maxSymbolValue; ++s) { for (s = 0; s <= maxSymbolValue; ++s) {
if (normalizedCounter[s] == 0) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(normalizedCounter[s] == 0, dictionary_corrupted);
} }
return 0; return 0;
} }
/* Dictionary format : /* Dictionary format :
* See : * See :
* https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format
assert(MEM_readLE32(dictPtr) == ZSTD_MAGIC_DICTIONARY); assert(MEM_readLE32(dictPtr) == ZSTD_MAGIC_DICTIONARY);
dictPtr += 4; /* skip magic number */ dictPtr += 4; /* skip magic number */
dictID = params->fParams.noDictIDFlag ? 0 : MEM_readLE32(dictPtr); dictID = params->fParams.noDictIDFlag ? 0 : MEM_readLE32(dictPtr);
dictPtr += 4; dictPtr += 4;
{ unsigned maxSymbolValue = 255; { unsigned maxSymbolValue = 255;
size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)bs->entropy.huf.CTable, &maxSymbolValue, dictPtr, dictEnd-dictPtr); size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)bs->entropy.huf.CTable, &maxSymbolValue, dictPtr, dictEnd-dictPtr);
if (HUF_isError(hufHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted);
if (maxSymbolValue < 255) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted);
dictPtr += hufHeaderSize; dictPtr += hufHeaderSize;
} }
{ unsigned offcodeLog; { unsigned offcodeLog;
size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr); size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);
if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted);
if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted);
/* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */ /* 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 */ /* fill all offset symbols to avoid garbage at end of table */
CHECK_E( FSE_buildCTable_wksp(bs->entropy.fse.offcodeCTable, RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
bs->entropy.fse.offcodeCTable,
offcodeNCount, MaxOff, offcodeLog, offcodeNCount, MaxOff, offcodeLog,
workspace, HUF_WORKSPACE_SIZE), workspace, HUF_WORKSPACE_SIZE)),
dictionary_corrupted); dictionary_corrupted);
dictPtr += offcodeHeaderSize; dictPtr += offcodeHeaderSize;
} }
{ short matchlengthNCount[MaxML+1]; { short matchlengthNCount[MaxML+1];
unsigned matchlengthMaxValue = MaxML, matchlengthLog; unsigned matchlengthMaxValue = MaxML, matchlengthLog;
size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr); size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);
if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted);
if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted);
/* Every match length code must have non-zero probability */ /* Every match length code must have non-zero probability */
CHECK_F( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML)); FORWARD_IF_ERROR( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));
CHECK_E( FSE_buildCTable_wksp(bs->entropy.fse.matchlengthCTable, RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
bs->entropy.fse.matchlengthCTable,
matchlengthNCount, matchlengthMaxValue, matchlengthLog, matchlengthNCount, matchlengthMaxValue, matchlengthLog,
workspace, HUF_WORKSPACE_SIZE), workspace, HUF_WORKSPACE_SIZE)),
dictionary_corrupted); dictionary_corrupted);
dictPtr += matchlengthHeaderSize; dictPtr += matchlengthHeaderSize;
} }
{ short litlengthNCount[MaxLL+1]; { short litlengthNCount[MaxLL+1];
unsigned litlengthMaxValue = MaxLL, litlengthLog; unsigned litlengthMaxValue = MaxLL, litlengthLog;
size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr); size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);
if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted);
if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted);
/* Every literal length code must have non-zero probability */ /* Every literal length code must have non-zero probability */
CHECK_F( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL)); FORWARD_IF_ERROR( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));
CHECK_E( FSE_buildCTable_wksp(bs->entropy.fse.litlengthCTable, RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
bs->entropy.fse.litlengthCTable,
litlengthNCount, litlengthMaxValue, litlengthLog, litlengthNCount, litlengthMaxValue, litlengthLog,
workspace, HUF_WORKSPACE_SIZE), workspace, HUF_WORKSPACE_SIZE)),
dictionary_corrupted); dictionary_corrupted);
dictPtr += litlengthHeaderSize; dictPtr += litlengthHeaderSize;
} }
if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted);
bs->rep[0] = MEM_readLE32(dictPtr+0); bs->rep[0] = MEM_readLE32(dictPtr+0);
bs->rep[1] = MEM_readLE32(dictPtr+4); bs->rep[1] = MEM_readLE32(dictPtr+4);
bs->rep[2] = MEM_readLE32(dictPtr+8); bs->rep[2] = MEM_readLE32(dictPtr+8);
dictPtr += 12; dictPtr += 12;
{ size_t const dictContentSize = (size_t)(dictEnd - dictPtr); { size_t const dictContentSize = (size_t)(dictEnd - dictPtr);
U32 offcodeMax = MaxOff; U32 offcodeMax = MaxOff;
if (dictContentSize <= ((U32)-1) - 128 KB) { if (dictContentSize <= ((U32)-1) - 128 KB) {
U32 const maxOffset = (U32)dictContentSize + 128 KB; /* The maximum offset that must be supported */ 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 */ offcodeMax = ZSTD_highbit32(maxOffset); /* Calculate minimum offset code required to represent maxOffset */
} }
/* All offset values <= dictContentSize + 128 KB must be representable */ /* All offset values <= dictContentSize + 128 KB must be representable */
CHECK_F (ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff))); FORWARD_IF_ERROR(ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));
/* All repCodes must be <= dictContentSize and != 0*/ /* All repCodes must be <= dictContentSize and != 0*/
{ U32 u; { U32 u;
for (u=0; u<3; u++) { for (u=0; u<3; u++) {
if (bs->rep[u] == 0) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(bs->rep[u] == 0, dictionary_corrupted);
if (bs->rep[u] > dictContentSize) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(bs->rep[u] > dictContentSize, dictionary_corrupted);
} } } }
bs->entropy.huf.repeatMode = HUF_repeat_valid; bs->entropy.huf.repeatMode = HUF_repeat_valid;
bs->entropy.fse.offcode_repeatMode = FSE_repeat_valid; bs->entropy.fse.offcode_repeatMode = FSE_repeat_valid;
bs->entropy.fse.matchlength_repeatMode = FSE_repeat_valid; bs->entropy.fse.matchlength_repeatMode = FSE_repeat_valid;
bs->entropy.fse.litlength_repeatMode = FSE_repeat_valid; bs->entropy.fse.litlength_repeatMode = FSE_repeat_valid;
CHECK_F(ZSTD_loadDictionaryContent(ms, params, dictPtr, dictContentSize, dtlm)); FORWARD_IF_ERROR(ZSTD_loadDictionaryContent(ms, params, dictPtr, dictContentSize, dtlm));
return dictID; return dictID;
} }
} }
/** ZSTD_compress_insertDictionary() : /** ZSTD_compress_insertDictionary() :
* @return : dictID, or an error code */ * @return : dictID, or an error code */
static size_t static size_t
ZSTD_compress_insertDictionary(ZSTD_compressedBlockState_t* bs, ZSTD_compress_insertDictionary(ZSTD_compressedBlockState_t* bs,
if (dictContentType == ZSTD_dct_rawContent) if (dictContentType == ZSTD_dct_rawContent)
return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm); return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm);
if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) { if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) {
if (dictContentType == ZSTD_dct_auto) { if (dictContentType == ZSTD_dct_auto) {
DEBUGLOG(4, "raw content dictionary detected"); DEBUGLOG(4, "raw content dictionary detected");
return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm); return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm);
} }
if (dictContentType == ZSTD_dct_fullDict) RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong);
return ERROR(dictionary_wrong);
assert(0); /* impossible */ assert(0); /* impossible */
} }
/* dict as full zstd dictionary */ /* dict as full zstd dictionary */
return ZSTD_loadZstdDictionary(bs, ms, params, dict, dictSize, dtlm, workspace); return ZSTD_loadZstdDictionary(bs, ms, params, dict, dictSize, dtlm, workspace);
} }
/*! ZSTD_compressBegin_internal() : /*! ZSTD_compressBegin_internal() :
/* params are supposed to be fully validated at this point */ /* params are supposed to be fully validated at this point */
assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
assert(!((dict) && (cdict))); /* either dict or cdict, not both */ assert(!((dict) && (cdict))); /* either dict or cdict, not both */
if (cdict && cdict->dictContentSize>0) { if (cdict && cdict->dictContentSize>0) {
return ZSTD_resetCCtx_usingCDict(cctx, cdict, params, pledgedSrcSize, zbuff); return ZSTD_resetCCtx_usingCDict(cctx, cdict, params, pledgedSrcSize, zbuff);
} }
CHECK_F( ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize, FORWARD_IF_ERROR( ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,
ZSTDcrp_continue, zbuff) ); ZSTDcrp_continue, zbuff) );
{ { size_t const dictID = ZSTD_compress_insertDictionary(
size_t const dictID = ZSTD_compress_insertDictionary(
cctx->blockState.prevCBlock, &cctx->blockState.matchState, cctx->blockState.prevCBlock, &cctx->blockState.matchState,
&params, dict, dictSize, dictContentType, dtlm, cctx->entropyWorkspace); &params, dict, dictSize, dictContentType, dtlm, cctx->entropyWorkspace);
if (ZSTD_isError(dictID)) return dictID; FORWARD_IF_ERROR(dictID);
assert(dictID <= (size_t)(U32)-1); assert(dictID <= UINT_MAX);
cctx->dictID = (U32)dictID; cctx->dictID = (U32)dictID;
} }
return 0; return 0;
} }
size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx, size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,
const void* dict, size_t dictSize, const void* dict, size_t dictSize,
ZSTD_dictContentType_e dictContentType, ZSTD_dictContentType_e dictContentType,
ZSTD_dictTableLoadMethod_e dtlm, ZSTD_dictTableLoadMethod_e dtlm,
const ZSTD_CDict* cdict, const ZSTD_CDict* cdict,
ZSTD_CCtx_params params, ZSTD_CCtx_params params,
unsigned long long pledgedSrcSize) unsigned long long pledgedSrcSize)
{ {
DEBUGLOG(4, "ZSTD_compressBegin_advanced_internal: wlog=%u", params.cParams.windowLog); DEBUGLOG(4, "ZSTD_compressBegin_advanced_internal: wlog=%u", params.cParams.windowLog);
/* compression parameters verification and optimization */ /* compression parameters verification and optimization */
CHECK_F( ZSTD_checkCParams(params.cParams) ); FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );
return ZSTD_compressBegin_internal(cctx, return ZSTD_compressBegin_internal(cctx,
dict, dictSize, dictContentType, dtlm, dict, dictSize, dictContentType, dtlm,
cdict, cdict,
params, pledgedSrcSize, params, pledgedSrcSize,
ZSTDb_not_buffered); ZSTDb_not_buffered);
} }
/*! ZSTD_compressBegin_advanced() : /*! ZSTD_compressBegin_advanced() :
* @return : nb of bytes written into dst (or an error code) */ * @return : nb of bytes written into dst (or an error code) */
static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity) static size_t ZSTD_writeEpilogue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity)
{ {
BYTE* const ostart = (BYTE*)dst; BYTE* const ostart = (BYTE*)dst;
BYTE* op = ostart; BYTE* op = ostart;
size_t fhSize = 0; size_t fhSize = 0;
DEBUGLOG(4, "ZSTD_writeEpilogue"); DEBUGLOG(4, "ZSTD_writeEpilogue");
if (cctx->stage == ZSTDcs_created) return ERROR(stage_wrong); /* init missing */ RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing");
/* special case : empty frame */ /* special case : empty frame */
if (cctx->stage == ZSTDcs_init) { if (cctx->stage == ZSTDcs_init) {
fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, 0, 0); fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, 0, 0);
if (ZSTD_isError(fhSize)) return fhSize; FORWARD_IF_ERROR(fhSize);
dstCapacity -= fhSize; dstCapacity -= fhSize;
op += fhSize; op += fhSize;
cctx->stage = ZSTDcs_ongoing; cctx->stage = ZSTDcs_ongoing;
} }
if (cctx->stage != ZSTDcs_ending) { if (cctx->stage != ZSTDcs_ending) {
/* write one last empty block, make it the "last" block */ /* write one last empty block, make it the "last" block */
U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0; U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0;
if (dstCapacity<4) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall);
MEM_writeLE32(op, cBlockHeader24); MEM_writeLE32(op, cBlockHeader24);
op += ZSTD_blockHeaderSize; op += ZSTD_blockHeaderSize;
dstCapacity -= ZSTD_blockHeaderSize; dstCapacity -= ZSTD_blockHeaderSize;
} }
if (cctx->appliedParams.fParams.checksumFlag) { if (cctx->appliedParams.fParams.checksumFlag) {
U32 const checksum = (U32) XXH64_digest(&cctx->xxhState); U32 const checksum = (U32) XXH64_digest(&cctx->xxhState);
if (dstCapacity<4) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall);
DEBUGLOG(4, "ZSTD_writeEpilogue: write checksum : %08X", (unsigned)checksum); DEBUGLOG(4, "ZSTD_writeEpilogue: write checksum : %08X", (unsigned)checksum);
MEM_writeLE32(op, checksum); MEM_writeLE32(op, checksum);
op += 4; op += 4;
} }
cctx->stage = ZSTDcs_created; /* return to "created but no init" status */ cctx->stage = ZSTDcs_created; /* return to "created but no init" status */
return op-ostart; return op-ostart;
} }
size_t ZSTD_compressEnd (ZSTD_CCtx* cctx, size_t ZSTD_compressEnd (ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
size_t endResult; size_t endResult;
size_t const cSize = ZSTD_compressContinue_internal(cctx, size_t const cSize = ZSTD_compressContinue_internal(cctx,
dst, dstCapacity, src, srcSize, dst, dstCapacity, src, srcSize,
1 /* frame mode */, 1 /* last chunk */); 1 /* frame mode */, 1 /* last chunk */);
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize); endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize);
if (ZSTD_isError(endResult)) return endResult; FORWARD_IF_ERROR(endResult);
assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0)); assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));
if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */ if (cctx->pledgedSrcSizePlusOne != 0) { /* control src size */
ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1); ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);
DEBUGLOG(4, "end of frame : controlling src size"); DEBUGLOG(4, "end of frame : controlling src size");
if (cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1) { RETURN_ERROR_IF(
DEBUGLOG(4, "error : pledgedSrcSize = %u, while realSrcSize = %u", cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1,
(unsigned)cctx->pledgedSrcSizePlusOne-1, (unsigned)cctx->consumedSrcSize); srcSize_wrong,
return ERROR(srcSize_wrong); "error : pledgedSrcSize = %u, while realSrcSize = %u",
} } (unsigned)cctx->pledgedSrcSizePlusOne-1,
(unsigned)cctx->consumedSrcSize);
}
return cSize + endResult; return cSize + endResult;
} }
static size_t ZSTD_compress_internal (ZSTD_CCtx* cctx, static size_t ZSTD_compress_internal (ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const void* dict,size_t dictSize, const void* dict,size_t dictSize,
size_t ZSTD_compress_advanced (ZSTD_CCtx* cctx, size_t ZSTD_compress_advanced (ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const void* dict,size_t dictSize, const void* dict,size_t dictSize,
ZSTD_parameters params) ZSTD_parameters params)
{ {
DEBUGLOG(4, "ZSTD_compress_advanced"); DEBUGLOG(4, "ZSTD_compress_advanced");
CHECK_F(ZSTD_checkCParams(params.cParams)); FORWARD_IF_ERROR(ZSTD_checkCParams(params.cParams));
return ZSTD_compress_internal(cctx, return ZSTD_compress_internal(cctx,
dst, dstCapacity, dst, dstCapacity,
src, srcSize, src, srcSize,
dict, dictSize, dict, dictSize,
params); params);
} }
/* Internal */ /* Internal */
size_t ZSTD_compress_advanced_internal( size_t ZSTD_compress_advanced_internal(
ZSTD_CCtx* cctx, ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const void* dict,size_t dictSize, const void* dict,size_t dictSize,
ZSTD_CCtx_params params) ZSTD_CCtx_params params)
{ {
DEBUGLOG(4, "ZSTD_compress_advanced_internal (srcSize:%u)", (unsigned)srcSize); DEBUGLOG(4, "ZSTD_compress_advanced_internal (srcSize:%u)", (unsigned)srcSize);
CHECK_F( ZSTD_compressBegin_internal(cctx, FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL, dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,
params, srcSize, ZSTDb_not_buffered) ); params, srcSize, ZSTDb_not_buffered) );
return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize); return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
} }
size_t ZSTD_compress_usingDict(ZSTD_CCtx* cctx, size_t ZSTD_compress_usingDict(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
cdict->matchState.cParams = cParams; cdict->matchState.cParams = cParams;
if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dictBuffer) || (!dictSize)) { if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dictBuffer) || (!dictSize)) {
cdict->dictBuffer = NULL; cdict->dictBuffer = NULL;
cdict->dictContent = dictBuffer; cdict->dictContent = dictBuffer;
} else { } else {
void* const internalBuffer = ZSTD_malloc(dictSize, cdict->customMem); void* const internalBuffer = ZSTD_malloc(dictSize, cdict->customMem);
cdict->dictBuffer = internalBuffer; cdict->dictBuffer = internalBuffer;
cdict->dictContent = internalBuffer; cdict->dictContent = internalBuffer;
if (!internalBuffer) return ERROR(memory_allocation); RETURN_ERROR_IF(!internalBuffer, memory_allocation);
memcpy(internalBuffer, dictBuffer, dictSize); memcpy(internalBuffer, dictBuffer, dictSize);
} }
cdict->dictContentSize = dictSize; cdict->dictContentSize = dictSize;
/* Reset the state to no dictionary */ /* Reset the state to no dictionary */
ZSTD_reset_compressedBlockState(&cdict->cBlockState); ZSTD_reset_compressedBlockState(&cdict->cBlockState);
{ void* const end = ZSTD_reset_matchState( { void* const end = ZSTD_reset_matchState(&cdict->matchState,
&cdict->matchState,
(U32*)cdict->workspace + HUF_WORKSPACE_SIZE_U32, (U32*)cdict->workspace + HUF_WORKSPACE_SIZE_U32,
&cParams, ZSTDcrp_continue, /* forCCtx */ 0); &cParams,
ZSTDcrp_continue, ZSTD_resetTarget_CDict);
assert(end == (char*)cdict->workspace + cdict->workspaceSize); assert(end == (char*)cdict->workspace + cdict->workspaceSize);
(void)end; (void)end;
} }
/* (Maybe) load the dictionary /* (Maybe) load the dictionary
* Skips loading the dictionary if it is <= 8 bytes. * Skips loading the dictionary if it is <= 8 bytes.
*/ */
{ ZSTD_CCtx_params params; { ZSTD_CCtx_params params;
memset(&params, 0, sizeof(params)); memset(&params, 0, sizeof(params));
params.compressionLevel = ZSTD_CLEVEL_DEFAULT; params.compressionLevel = ZSTD_CLEVEL_DEFAULT;
params.fParams.contentSizeFlag = 1; params.fParams.contentSizeFlag = 1;
params.cParams = cParams; params.cParams = cParams;
{ size_t const dictID = ZSTD_compress_insertDictionary( { size_t const dictID = ZSTD_compress_insertDictionary(
&cdict->cBlockState, &cdict->matchState, &params, &cdict->cBlockState, &cdict->matchState, &params,
cdict->dictContent, cdict->dictContentSize, cdict->dictContent, cdict->dictContentSize,
dictContentType, ZSTD_dtlm_full, cdict->workspace); dictContentType, ZSTD_dtlm_full, cdict->workspace);
if (ZSTD_isError(dictID)) return dictID; FORWARD_IF_ERROR(dictID);
assert(dictID <= (size_t)(U32)-1); assert(dictID <= (size_t)(U32)-1);
cdict->dictID = (U32)dictID; cdict->dictID = (U32)dictID;
} }
} }
return 0; return 0;
} }
/* ZSTD_compressBegin_usingCDict_advanced() : /* ZSTD_compressBegin_usingCDict_advanced() :
* cdict must be != NULL */ * cdict must be != NULL */
size_t ZSTD_compressBegin_usingCDict_advanced( size_t ZSTD_compressBegin_usingCDict_advanced(
ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict, ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict,
ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize) ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize)
{ {
DEBUGLOG(4, "ZSTD_compressBegin_usingCDict_advanced"); DEBUGLOG(4, "ZSTD_compressBegin_usingCDict_advanced");
if (cdict==NULL) return ERROR(dictionary_wrong); RETURN_ERROR_IF(cdict==NULL, dictionary_wrong);
{ ZSTD_CCtx_params params = cctx->requestedParams; { ZSTD_CCtx_params params = cctx->requestedParams;
params.cParams = ZSTD_getCParamsFromCDict(cdict); params.cParams = ZSTD_getCParamsFromCDict(cdict);
/* Increase window log to fit the entire dictionary and source if the /* Increase window log to fit the entire dictionary and source if the
* source size is known. Limit the increase to 19, which is the * source size is known. Limit the increase to 19, which is the
* window log for compression level 1 with the largest source size. * window log for compression level 1 with the largest source size.
*/ */
if (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN) { if (pledgedSrcSize != ZSTD_CONTENTSIZE_UNKNOWN) {
U32 const limitedSrcSize = (U32)MIN(pledgedSrcSize, 1U << 19); U32 const limitedSrcSize = (U32)MIN(pledgedSrcSize, 1U << 19);
return ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN); return ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN);
} }
size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx, size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const ZSTD_CDict* cdict, ZSTD_frameParameters fParams) const ZSTD_CDict* cdict, ZSTD_frameParameters fParams)
{ {
CHECK_F (ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize)); /* will check if cdict != NULL */ FORWARD_IF_ERROR(ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize)); /* will check if cdict != NULL */
return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize); return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
} }
/*! ZSTD_compress_usingCDict() : /*! ZSTD_compress_usingCDict() :
* Compression using a digested Dictionary. * Compression using a digested Dictionary.
* Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times. * Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.
* Note that compression parameters are decided at CDict creation time * Note that compression parameters are decided at CDict creation time
* while frame parameters are hardcoded */ * while frame parameters are hardcoded */
{ {
DEBUGLOG(4, "ZSTD_resetCStream_internal"); DEBUGLOG(4, "ZSTD_resetCStream_internal");
/* Finalize the compression parameters */ /* Finalize the compression parameters */
params.cParams = ZSTD_getCParamsFromCCtxParams(&params, pledgedSrcSize, dictSize); params.cParams = ZSTD_getCParamsFromCCtxParams(&params, pledgedSrcSize, dictSize);
/* params are supposed to be fully validated at this point */ /* params are supposed to be fully validated at this point */
assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
assert(!((dict) && (cdict))); /* either dict or cdict, not both */ assert(!((dict) && (cdict))); /* either dict or cdict, not both */
CHECK_F( ZSTD_compressBegin_internal(cctx, FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
dict, dictSize, dictContentType, ZSTD_dtlm_fast, dict, dictSize, dictContentType, ZSTD_dtlm_fast,
cdict, cdict,
params, pledgedSrcSize, params, pledgedSrcSize,
ZSTDb_buffered) ); ZSTDb_buffered) );
cctx->inToCompress = 0; cctx->inToCompress = 0;
cctx->inBuffPos = 0; cctx->inBuffPos = 0;
cctx->inBuffTarget = cctx->blockSize 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->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->outBuffContentSize = cctx->outBuffFlushedSize = 0;
cctx->streamStage = zcss_load; cctx->streamStage = zcss_load;
cctx->frameEnded = 0; cctx->frameEnded = 0;
return 0; /* ready to go */ return 0; /* ready to go */
} }
/* ZSTD_resetCStream(): /* ZSTD_resetCStream():
* pledgedSrcSize == 0 means "unknown" */ * pledgedSrcSize == 0 means "unknown" */
size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize) size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pss)
{ {
ZSTD_CCtx_params params = zcs->requestedParams; /* 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); DEBUGLOG(4, "ZSTD_resetCStream: pledgedSrcSize = %u", (unsigned)pledgedSrcSize);
if (pledgedSrcSize==0) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN; FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
params.fParams.contentSizeFlag = 1; FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dct_auto, zcs->cdict, params, pledgedSrcSize); return 0;
} }
/*! ZSTD_initCStream_internal() : /*! ZSTD_initCStream_internal() :
* Note : for lib/compress only. Used by zstdmt_compress.c. * Note : for lib/compress only. Used by zstdmt_compress.c.
* Assumption 1 : params are valid * Assumption 1 : params are valid
* Assumption 2 : either dict, or cdict, is defined, not both */ * Assumption 2 : either dict, or cdict, is defined, not both */
size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs, size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
const void* dict, size_t dictSize, const ZSTD_CDict* cdict, const void* dict, size_t dictSize, const ZSTD_CDict* cdict,
ZSTD_CCtx_params params, unsigned long long pledgedSrcSize) ZSTD_CCtx_params params, unsigned long long pledgedSrcSize)
{ {
DEBUGLOG(4, "ZSTD_initCStream_internal"); DEBUGLOG(4, "ZSTD_initCStream_internal");
params.cParams = ZSTD_getCParamsFromCCtxParams(&params, pledgedSrcSize, dictSize); 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))); assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
zcs->requestedParams = params;
assert(!((dict) && (cdict))); /* either dict or cdict, not both */ assert(!((dict) && (cdict))); /* either dict or cdict, not both */
if (dict) {
if (dict && dictSize >= 8) { FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );
DEBUGLOG(4, "loading dictionary of size %u", (unsigned)dictSize);
if (zcs->staticSize) { /* static CCtx : never uses malloc */
/* incompatible with internal cdict creation */
return ERROR(memory_allocation);
}
ZSTD_freeCDict(zcs->cdictLocal);
zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,
ZSTD_dlm_byCopy, ZSTD_dct_auto,
params.cParams, zcs->customMem);
zcs->cdict = zcs->cdictLocal;
if (zcs->cdictLocal == NULL) return ERROR(memory_allocation);
} else { } else {
if (cdict) { /* Dictionary is cleared if !cdict */
params.cParams = ZSTD_getCParamsFromCDict(cdict); /* cParams are enforced from cdict; it includes windowLog */ FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );
}
ZSTD_freeCDict(zcs->cdictLocal);
zcs->cdictLocal = NULL;
zcs->cdict = cdict;
} }
return 0;
return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dct_auto, zcs->cdict, params, pledgedSrcSize);
} }
/* ZSTD_initCStream_usingCDict_advanced() : /* ZSTD_initCStream_usingCDict_advanced() :
* same as ZSTD_initCStream_usingCDict(), with control over frame parameters */ * same as ZSTD_initCStream_usingCDict(), with control over frame parameters */
size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
const ZSTD_CDict* cdict, const ZSTD_CDict* cdict,
ZSTD_frameParameters fParams, ZSTD_frameParameters fParams,
unsigned long long pledgedSrcSize) unsigned long long pledgedSrcSize)
{ {
DEBUGLOG(4, "ZSTD_initCStream_usingCDict_advanced"); DEBUGLOG(4, "ZSTD_initCStream_usingCDict_advanced");
if (!cdict) return ERROR(dictionary_wrong); /* cannot handle NULL cdict (does not know what to do) */ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
{ ZSTD_CCtx_params params = zcs->requestedParams; FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
params.cParams = ZSTD_getCParamsFromCDict(cdict); zcs->requestedParams.fParams = fParams;
params.fParams = fParams; FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );
return ZSTD_initCStream_internal(zcs, return 0;
NULL, 0, cdict,
params, pledgedSrcSize);
}
} }
/* note : cdict must outlive compression session */ /* note : cdict must outlive compression session */
size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict) size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict)
{ {
ZSTD_frameParameters const fParams = { 0 /* contentSizeFlag */, 0 /* checksum */, 0 /* hideDictID */ };
DEBUGLOG(4, "ZSTD_initCStream_usingCDict"); DEBUGLOG(4, "ZSTD_initCStream_usingCDict");
return ZSTD_initCStream_usingCDict_advanced(zcs, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN); /* note : will check that cdict != NULL */ FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );
return 0;
} }
/* ZSTD_initCStream_advanced() : /* ZSTD_initCStream_advanced() :
* pledgedSrcSize must be exact. * pledgedSrcSize must be exact.
* if srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN. * 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. */ * dict is loaded with default parameters ZSTD_dm_auto and ZSTD_dlm_byCopy. */
size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
const void* dict, size_t dictSize, const void* dict, size_t dictSize,
ZSTD_parameters params, unsigned long long pledgedSrcSize) ZSTD_parameters params, unsigned long long pss)
{ {
DEBUGLOG(4, "ZSTD_initCStream_advanced: pledgedSrcSize=%u, flag=%u", /* for compatibility with older programs relying on this behavior.
(unsigned)pledgedSrcSize, params.fParams.contentSizeFlag); * Users should now specify ZSTD_CONTENTSIZE_UNKNOWN.
CHECK_F( ZSTD_checkCParams(params.cParams) ); * This line will be removed in the future.
if ((pledgedSrcSize==0) && (params.fParams.contentSizeFlag==0)) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN; /* 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); zcs->requestedParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL /*cdict*/, zcs->requestedParams, pledgedSrcSize); 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) size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)
{ {
ZSTD_CCtxParams_init(&zcs->requestedParams, compressionLevel); DEBUGLOG(4, "ZSTD_initCStream_usingDict");
return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, zcs->requestedParams, ZSTD_CONTENTSIZE_UNKNOWN); 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) size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pss)
{ {
U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : 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 */ /* temporary : 0 interpreted as "unknown" during transition period.
ZSTD_CCtxParams_init(&zcs->requestedParams, compressionLevel); * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.
return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, zcs->requestedParams, pledgedSrcSize); * 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) size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
{ {
DEBUGLOG(4, "ZSTD_initCStream"); DEBUGLOG(4, "ZSTD_initCStream");
return ZSTD_initCStream_srcSize(zcs, compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN); 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 ======*/ /*====== Compression ======*/
static size_t ZSTD_nextInputSizeHint(const ZSTD_CCtx* cctx) static size_t ZSTD_nextInputSizeHint(const ZSTD_CCtx* cctx)
{ {
size_t hintInSize = cctx->inBuffTarget - cctx->inBuffPos; size_t hintInSize = cctx->inBuffTarget - cctx->inBuffPos;
if (hintInSize==0) hintInSize = cctx->blockSize; if (hintInSize==0) hintInSize = cctx->blockSize;
return hintInSize; return hintInSize;
} }
static size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, static size_t ZSTD_limitCopy(void* dst, size_t dstCapacity,
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
size_t const length = MIN(dstCapacity, srcSize); size_t const length = MIN(dstCapacity, srcSize);
if (length) memcpy(dst, src, length); if (length) memcpy(dst, src, length);
return length; return length;
} }
/** ZSTD_compressStream_generic(): /** ZSTD_compressStream_generic():
* internal function for all *compressStream*() variants * internal function for all *compressStream*() variants
* non-static, because can be called from zstdmt_compress.c * non-static, because can be called from zstdmt_compress.c
* @return : hint size for next input */ * @return : hint size for next input */
size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs, static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
ZSTD_outBuffer* output, ZSTD_outBuffer* output,
ZSTD_inBuffer* input, ZSTD_inBuffer* input,
ZSTD_EndDirective const flushMode) ZSTD_EndDirective const flushMode)
{ {
const char* const istart = (const char*)input->src; const char* const istart = (const char*)input->src;
const char* const iend = istart + input->size; const char* const iend = istart + input->size;
const char* ip = istart + input->pos; const char* ip = istart + input->pos;
char* const ostart = (char*)output->dst; char* const ostart = (char*)output->dst;
char* const oend = ostart + output->size; char* const oend = ostart + output->size;
char* op = ostart + output->pos; char* op = ostart + output->pos;
U32 someMoreWork = 1; U32 someMoreWork = 1;
/* check expectations */ /* check expectations */
DEBUGLOG(5, "ZSTD_compressStream_generic, flush=%u", (unsigned)flushMode); DEBUGLOG(5, "ZSTD_compressStream_generic, flush=%u", (unsigned)flushMode);
assert(zcs->inBuff != NULL); assert(zcs->inBuff != NULL);
assert(zcs->inBuffSize > 0); assert(zcs->inBuffSize > 0);
assert(zcs->outBuff != NULL); assert(zcs->outBuff != NULL);
assert(zcs->outBuffSize > 0); assert(zcs->outBuffSize > 0);
assert(output->pos <= output->size); assert(output->pos <= output->size);
assert(input->pos <= input->size); assert(input->pos <= input->size);
while (someMoreWork) { while (someMoreWork) {
switch(zcs->streamStage) switch(zcs->streamStage)
{ {
case zcss_init: case zcss_init:
/* call ZSTD_initCStream() first ! */ RETURN_ERROR(init_missing, "call ZSTD_initCStream() first!");
return ERROR(init_missing);
case zcss_load: case zcss_load:
if ( (flushMode == ZSTD_e_end) if ( (flushMode == ZSTD_e_end)
&& ((size_t)(oend-op) >= ZSTD_compressBound(iend-ip)) /* enough dstCapacity */ && ((size_t)(oend-op) >= ZSTD_compressBound(iend-ip)) /* enough dstCapacity */
&& (zcs->inBuffPos == 0) ) { && (zcs->inBuffPos == 0) ) {
/* shortcut to compression pass directly into output buffer */ /* shortcut to compression pass directly into output buffer */
size_t const cSize = ZSTD_compressEnd(zcs, size_t const cSize = ZSTD_compressEnd(zcs,
op, oend-op, ip, iend-ip); op, oend-op, ip, iend-ip);
DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize); DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize);
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
ip = iend; ip = iend;
op += cSize; op += cSize;
zcs->frameEnded = 1; zcs->frameEnded = 1;
ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only); ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
someMoreWork = 0; break; someMoreWork = 0; break;
} }
/* complete loading into inBuffer */ /* complete loading into inBuffer */
{ size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos; { size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos;
cDst = op; /* compress into output buffer, to skip flush stage */ cDst = op; /* compress into output buffer, to skip flush stage */
else else
cDst = zcs->outBuff, oSize = zcs->outBuffSize; cDst = zcs->outBuff, oSize = zcs->outBuffSize;
cSize = lastBlock ? cSize = lastBlock ?
ZSTD_compressEnd(zcs, cDst, oSize, ZSTD_compressEnd(zcs, cDst, oSize,
zcs->inBuff + zcs->inToCompress, iSize) : zcs->inBuff + zcs->inToCompress, iSize) :
ZSTD_compressContinue(zcs, cDst, oSize, ZSTD_compressContinue(zcs, cDst, oSize,
zcs->inBuff + zcs->inToCompress, iSize); zcs->inBuff + zcs->inToCompress, iSize);
if (ZSTD_isError(cSize)) return cSize; FORWARD_IF_ERROR(cSize);
zcs->frameEnded = lastBlock; zcs->frameEnded = lastBlock;
/* prepare next block */ /* prepare next block */
zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize; zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;
if (zcs->inBuffTarget > zcs->inBuffSize) if (zcs->inBuffTarget > zcs->inBuffSize)
zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize; zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize;
DEBUGLOG(5, "inBuffTarget:%u / inBuffSize:%u", DEBUGLOG(5, "inBuffTarget:%u / inBuffSize:%u",
(unsigned)zcs->inBuffTarget, (unsigned)zcs->inBuffSize); (unsigned)zcs->inBuffTarget, (unsigned)zcs->inBuffSize);
if (!lastBlock) if (!lastBlock)
zcs->outBuffContentSize = cSize; zcs->outBuffContentSize = cSize;
zcs->outBuffFlushedSize = 0; zcs->outBuffFlushedSize = 0;
zcs->streamStage = zcss_flush; /* pass-through to flush stage */ zcs->streamStage = zcss_flush; /* pass-through to flush stage */
} }
/* fall-through */ /* fall-through */
case zcss_flush: case zcss_flush:
DEBUGLOG(5, "flush stage"); DEBUGLOG(5, "flush stage");
{ size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize; { size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
size_t const flushed = ZSTD_limitCopy(op, oend-op, size_t const flushed = ZSTD_limitCopy(op, (size_t)(oend-op),
zcs->outBuff + zcs->outBuffFlushedSize, toFlush); zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
DEBUGLOG(5, "toFlush: %u into %u ==> flushed: %u", DEBUGLOG(5, "toFlush: %u into %u ==> flushed: %u",
(unsigned)toFlush, (unsigned)(oend-op), (unsigned)flushed); (unsigned)toFlush, (unsigned)(oend-op), (unsigned)flushed);
op += flushed; op += flushed;
zcs->outBuffFlushedSize += flushed; zcs->outBuffFlushedSize += flushed;
if (toFlush!=flushed) { if (toFlush!=flushed) {
/* flush not fully completed, presumably because dst is too small */ /* flush not fully completed, presumably because dst is too small */
assert(op==oend); assert(op==oend);
} }
#endif #endif
return ZSTD_nextInputSizeHint(cctx); return ZSTD_nextInputSizeHint(cctx);
} }
size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input) size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
{ {
CHECK_F( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) ); FORWARD_IF_ERROR( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) );
return ZSTD_nextInputSizeHint_MTorST(zcs); return ZSTD_nextInputSizeHint_MTorST(zcs);
} }
size_t ZSTD_compressStream2( ZSTD_CCtx* cctx, size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
ZSTD_outBuffer* output, ZSTD_outBuffer* output,
ZSTD_inBuffer* input, ZSTD_inBuffer* input,
ZSTD_EndDirective endOp) ZSTD_EndDirective endOp)
{ {
DEBUGLOG(5, "ZSTD_compressStream2, endOp=%u ", (unsigned)endOp); DEBUGLOG(5, "ZSTD_compressStream2, endOp=%u ", (unsigned)endOp);
/* check conditions */ /* check conditions */
if (output->pos > output->size) return ERROR(GENERIC); RETURN_ERROR_IF(output->pos > output->size, GENERIC);
if (input->pos > input->size) return ERROR(GENERIC); RETURN_ERROR_IF(input->pos > input->size, GENERIC);
assert(cctx!=NULL); assert(cctx!=NULL);
/* transparent initialization stage */ /* transparent initialization stage */
if (cctx->streamStage == zcss_init) { if (cctx->streamStage == zcss_init) {
ZSTD_CCtx_params params = cctx->requestedParams; ZSTD_CCtx_params params = cctx->requestedParams;
ZSTD_prefixDict const prefixDict = cctx->prefixDict; 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 */ memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict)); /* single usage */
assert(prefixDict.dict==NULL || cctx->cdict==NULL); /* only one can be set */ assert(prefixDict.dict==NULL || cctx->cdict==NULL); /* only one can be set */
DEBUGLOG(4, "ZSTD_compressStream2 : transparent init stage"); DEBUGLOG(4, "ZSTD_compressStream2 : transparent init stage");
if (endOp == ZSTD_e_end) cctx->pledgedSrcSizePlusOne = input->size + 1; /* auto-fix pledgedSrcSize */ if (endOp == ZSTD_e_end) cctx->pledgedSrcSizePlusOne = input->size + 1; /* auto-fix pledgedSrcSize */
params.cParams = ZSTD_getCParamsFromCCtxParams( params.cParams = ZSTD_getCParamsFromCCtxParams(
&cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/); &cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, 0 /*dictSize*/);
#ifdef ZSTD_MULTITHREAD #ifdef ZSTD_MULTITHREAD
if ((cctx->pledgedSrcSizePlusOne-1) <= ZSTDMT_JOBSIZE_MIN) { if ((cctx->pledgedSrcSizePlusOne-1) <= ZSTDMT_JOBSIZE_MIN) {
params.nbWorkers = 0; /* do not invoke multi-threading when src size is too small */ params.nbWorkers = 0; /* do not invoke multi-threading when src size is too small */
} }
if (params.nbWorkers > 0) { if (params.nbWorkers > 0) {
/* mt context creation */ /* mt context creation */
if (cctx->mtctx == NULL) { if (cctx->mtctx == NULL) {
DEBUGLOG(4, "ZSTD_compressStream2: creating new mtctx for nbWorkers=%u", DEBUGLOG(4, "ZSTD_compressStream2: creating new mtctx for nbWorkers=%u",
params.nbWorkers); params.nbWorkers);
cctx->mtctx = ZSTDMT_createCCtx_advanced(params.nbWorkers, cctx->customMem); cctx->mtctx = ZSTDMT_createCCtx_advanced(params.nbWorkers, cctx->customMem);
if (cctx->mtctx == NULL) return ERROR(memory_allocation); RETURN_ERROR_IF(cctx->mtctx == NULL, memory_allocation);
} }
/* mt compression */ /* mt compression */
DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbWorkers=%u", params.nbWorkers); DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbWorkers=%u", params.nbWorkers);
CHECK_F( ZSTDMT_initCStream_internal( FORWARD_IF_ERROR( ZSTDMT_initCStream_internal(
cctx->mtctx, cctx->mtctx,
prefixDict.dict, prefixDict.dictSize, ZSTD_dct_rawContent, prefixDict.dict, prefixDict.dictSize, ZSTD_dct_rawContent,
cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) ); cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) );
cctx->streamStage = zcss_load; cctx->streamStage = zcss_load;
cctx->appliedParams.nbWorkers = params.nbWorkers; cctx->appliedParams.nbWorkers = params.nbWorkers;
} else } else
#endif #endif
{ CHECK_F( ZSTD_resetCStream_internal(cctx, { FORWARD_IF_ERROR( ZSTD_resetCStream_internal(cctx,
prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType, prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType,
cctx->cdict, cctx->cdict,
params, cctx->pledgedSrcSizePlusOne-1) ); params, cctx->pledgedSrcSizePlusOne-1) );
assert(cctx->streamStage == zcss_load); assert(cctx->streamStage == zcss_load);
assert(cctx->appliedParams.nbWorkers == 0); assert(cctx->appliedParams.nbWorkers == 0);
} } } }
/* end of transparent initialization stage */ /* end of transparent initialization stage */
/* compression stage */ /* compression stage */
#ifdef ZSTD_MULTITHREAD #ifdef ZSTD_MULTITHREAD
if (cctx->appliedParams.nbWorkers > 0) { 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) { if (cctx->cParamsChanged) {
ZSTDMT_updateCParams_whileCompressing(cctx->mtctx, &cctx->requestedParams); ZSTDMT_updateCParams_whileCompressing(cctx->mtctx, &cctx->requestedParams);
cctx->cParamsChanged = 0; cctx->cParamsChanged = 0;
} }
{ size_t const flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp); do {
flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp);
if ( ZSTD_isError(flushMin) if ( ZSTD_isError(flushMin)
|| (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */ || (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */
ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only); 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"); 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; return flushMin;
} } }
#endif #endif
CHECK_F( ZSTD_compressStream_generic(cctx, output, input, endOp) ); FORWARD_IF_ERROR( ZSTD_compressStream_generic(cctx, output, input, endOp) );
DEBUGLOG(5, "completed ZSTD_compressStream2"); DEBUGLOG(5, "completed ZSTD_compressStream2");
return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */ return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */
} }
size_t ZSTD_compressStream2_simpleArgs ( size_t ZSTD_compressStream2_simpleArgs (
ZSTD_CCtx* cctx, ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, size_t* dstPos, void* dst, size_t dstCapacity, size_t* dstPos,
const void* src, size_t srcSize, size_t* srcPos, const void* src, size_t srcSize, size_t* srcPos,
{ {
ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only); ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only);
{ size_t oPos = 0; { size_t oPos = 0;
size_t iPos = 0; size_t iPos = 0;
size_t const result = ZSTD_compressStream2_simpleArgs(cctx, size_t const result = ZSTD_compressStream2_simpleArgs(cctx,
dst, dstCapacity, &oPos, dst, dstCapacity, &oPos,
src, srcSize, &iPos, src, srcSize, &iPos,
ZSTD_e_end); ZSTD_e_end);
if (ZSTD_isError(result)) return result; FORWARD_IF_ERROR(result);
if (result != 0) { /* compression not completed, due to lack of output space */ if (result != 0) { /* compression not completed, due to lack of output space */
assert(oPos == dstCapacity); assert(oPos == dstCapacity);
return ERROR(dstSize_tooSmall); RETURN_ERROR(dstSize_tooSmall);
} }
assert(iPos == srcSize); /* all input is expected consumed */ assert(iPos == srcSize); /* all input is expected consumed */
return oPos; return oPos;
} }
} }
/*====== Finalize ======*/ /*====== Finalize ======*/
/*! ZSTD_flushStream() : /*! ZSTD_flushStream() :
* @return : amount of data remaining to flush */ * @return : amount of data remaining to flush */
size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output) size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)
{ {
ZSTD_inBuffer input = { NULL, 0, 0 }; ZSTD_inBuffer input = { NULL, 0, 0 };
return ZSTD_compressStream2(zcs, output, &input, ZSTD_e_flush); return ZSTD_compressStream2(zcs, output, &input, ZSTD_e_flush);
} }
size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output) size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)
{ {
ZSTD_inBuffer input = { NULL, 0, 0 }; ZSTD_inBuffer input = { NULL, 0, 0 };
size_t const remainingToFlush = ZSTD_compressStream2(zcs, output, &input, ZSTD_e_end); size_t const remainingToFlush = ZSTD_compressStream2(zcs, output, &input, ZSTD_e_end);
CHECK_F( remainingToFlush ); FORWARD_IF_ERROR( remainingToFlush );
if (zcs->appliedParams.nbWorkers > 0) return remainingToFlush; /* minimal estimation */ if (zcs->appliedParams.nbWorkers > 0) return remainingToFlush; /* minimal estimation */
/* single thread mode : attempt to calculate remaining to flush more precisely */ /* single thread mode : attempt to calculate remaining to flush more precisely */
{ size_t const lastBlockSize = zcs->frameEnded ? 0 : ZSTD_BLOCKHEADERSIZE; { size_t const lastBlockSize = zcs->frameEnded ? 0 : ZSTD_BLOCKHEADERSIZE;
size_t const checksumSize = zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4; size_t const checksumSize = (size_t)(zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4);
size_t const toFlush = remainingToFlush + lastBlockSize + checksumSize; size_t const toFlush = remainingToFlush + lastBlockSize + checksumSize;
DEBUGLOG(4, "ZSTD_endStream : remaining to flush : %u", (unsigned)toFlush); DEBUGLOG(4, "ZSTD_endStream : remaining to flush : %u", (unsigned)toFlush);
return toFlush; return toFlush;
} }
} }
/*-===== Pre-defined compression levels =====-*/ /*-===== Pre-defined compression levels =====-*/
#define ZSTD_MAX_CLEVEL 22 #define ZSTD_MAX_CLEVEL 22
int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; } int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; }
int ZSTD_minCLevel(void) { return (int)-ZSTD_TARGETLENGTH_MAX; } int ZSTD_minCLevel(void) { return (int)-ZSTD_TARGETLENGTH_MAX; }
static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = { static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = {
{ /* "default" - guarantees a monotonically increasing memory budget */ { /* "default" - for any srcSize > 256 KB */
/* W, C, H, S, L, TL, strat */ /* W, C, H, S, L, TL, strat */
{ 19, 12, 13, 1, 6, 1, ZSTD_fast }, /* base for negative levels */ { 19, 12, 13, 1, 6, 1, ZSTD_fast }, /* base for negative levels */
{ 19, 13, 14, 1, 7, 0, ZSTD_fast }, /* level 1 */ { 19, 13, 14, 1, 7, 0, ZSTD_fast }, /* level 1 */
{ 20, 15, 16, 1, 6, 0, ZSTD_fast }, /* level 2 */ { 20, 15, 16, 1, 6, 0, ZSTD_fast }, /* level 2 */
{ 21, 16, 17, 1, 5, 1, ZSTD_dfast }, /* level 3 */ { 21, 16, 17, 1, 5, 1, ZSTD_dfast }, /* level 3 */
{ 21, 18, 18, 1, 5, 1, ZSTD_dfast }, /* level 4 */ { 21, 18, 18, 1, 5, 1, ZSTD_dfast }, /* level 4 */
{ 21, 18, 19, 2, 5, 2, ZSTD_greedy }, /* level 5 */ { 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, 4, ZSTD_greedy }, /* level 6 */
{ 14, 15, 15, 8, 3,256, ZSTD_btultra2}, /* level 19.*/ { 14, 15, 15, 8, 3,256, ZSTD_btultra2}, /* level 19.*/
{ 14, 15, 15, 8, 3,512, ZSTD_btultra2}, /* level 20.*/ { 14, 15, 15, 8, 3,512, ZSTD_btultra2}, /* level 20.*/
{ 14, 15, 15, 9, 3,512, ZSTD_btultra2}, /* level 21.*/ { 14, 15, 15, 9, 3,512, ZSTD_btultra2}, /* level 21.*/
{ 14, 15, 15, 10, 3,999, ZSTD_btultra2}, /* level 22.*/ { 14, 15, 15, 10, 3,999, ZSTD_btultra2}, /* level 22.*/
}, },
}; };
/*! ZSTD_getCParams() : /*! ZSTD_getCParams() :
* @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize. * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize.
* Size values are optional, provide 0 if not known or unused */ * Size values are optional, provide 0 if not known or unused */
ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize)
{ {
size_t const addedSize = srcSizeHint ? 0 : 500; size_t const addedSize = srcSizeHint ? 0 : 500;
U64 const rSize = srcSizeHint+dictSize ? srcSizeHint+dictSize+addedSize : (U64)-1; 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); /* intentional underflow for srcSizeHint == 0 */ U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB);
int row = compressionLevel; int row = compressionLevel;
DEBUGLOG(5, "ZSTD_getCParams (cLevel=%i)", compressionLevel); DEBUGLOG(5, "ZSTD_getCParams (cLevel=%i)", compressionLevel);
if (compressionLevel == 0) row = ZSTD_CLEVEL_DEFAULT; /* 0 == default */ if (compressionLevel == 0) row = ZSTD_CLEVEL_DEFAULT; /* 0 == default */
if (compressionLevel < 0) row = 0; /* entry 0 is baseline for fast mode */ if (compressionLevel < 0) row = 0; /* entry 0 is baseline for fast mode */
if (compressionLevel > ZSTD_MAX_CLEVEL) row = ZSTD_MAX_CLEVEL; if (compressionLevel > ZSTD_MAX_CLEVEL) row = ZSTD_MAX_CLEVEL;
{ ZSTD_compressionParameters cp = ZSTD_defaultCParameters[tableID][row]; { ZSTD_compressionParameters cp = ZSTD_defaultCParameters[tableID][row];
if (compressionLevel < 0) cp.targetLength = (unsigned)(-compressionLevel); /* acceleration factor */ if (compressionLevel < 0) cp.targetLength = (unsigned)(-compressionLevel); /* acceleration factor */
return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize); return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize); /* refine parameters based on srcSize & dictSize */
} }
} }
/*! ZSTD_getParams() : /*! ZSTD_getParams() :
* same as ZSTD_getCParams(), but @return a `ZSTD_parameters` object (instead of `ZSTD_compressionParameters`). * same idea as ZSTD_getCParams()
* All fields of `ZSTD_frameParameters` are set to default (0) */ * @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 ZSTD_getParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) {
ZSTD_parameters params; ZSTD_parameters params;
ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSizeHint, dictSize); ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSizeHint, dictSize);
DEBUGLOG(5, "ZSTD_getParams (cLevel=%i)", compressionLevel); DEBUGLOG(5, "ZSTD_getParams (cLevel=%i)", compressionLevel);
memset(&params, 0, sizeof(params)); memset(&params, 0, sizeof(params));
params.cParams = cParams; params.cParams = cParams;
params.fParams.contentSizeFlag = 1; params.fParams.contentSizeFlag = 1;
return params; return params;
} }

contrib/python-zstandard/zstd/compress/zstd_compress_internal.h

#endif #endif
/*-************************************* /*-*************************************
* Constants * Constants
***************************************/ ***************************************/
#define kSearchStrength 8 #define kSearchStrength 8
#define HASH_READ_SIZE 8 #define HASH_READ_SIZE 8
#define ZSTD_DUBT_UNSORTED_MARK 1 /* For btlazy2 strategy, index 1 now means "unsorted". #define ZSTD_DUBT_UNSORTED_MARK 1 /* For btlazy2 strategy, index ZSTD_DUBT_UNSORTED_MARK==1 means "unsorted".
It could be confused for a real successor at index "1", if sorted as larger than its predecessor. 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. It's not a big deal though : candidate will just be sorted again.
Additionnally, candidate position 1 will be lost. Additionally, candidate position 1 will be lost.
But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss. 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 misdhandled after table re-use with a different strategy 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() */ This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */
/*-************************************* /*-*************************************
* Context memory management * Context memory management
***************************************/ ***************************************/
typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e; 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 enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage;
typedef struct ZSTD_prefixDict_s { typedef struct ZSTD_prefixDict_s {
const void* dict; const void* dict;
size_t dictSize; size_t dictSize;
ZSTD_dictContentType_e dictContentType; ZSTD_dictContentType_e dictContentType;
} ZSTD_prefixDict; } ZSTD_prefixDict;
typedef struct { 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)]; U32 CTable[HUF_CTABLE_SIZE_U32(255)];
HUF_repeat repeatMode; HUF_repeat repeatMode;
} ZSTD_hufCTables_t; } ZSTD_hufCTables_t;
typedef struct { typedef struct {
FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)]; FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];
FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)]; FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];
FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)]; FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];
U32 matchLengthSum; /* nb of matchLength codes */ U32 matchLengthSum; /* nb of matchLength codes */
U32 offCodeSum; /* nb of offset codes */ U32 offCodeSum; /* nb of offset codes */
U32 litSumBasePrice; /* to compare to log2(litfreq) */ U32 litSumBasePrice; /* to compare to log2(litfreq) */
U32 litLengthSumBasePrice; /* to compare to log2(llfreq) */ U32 litLengthSumBasePrice; /* to compare to log2(llfreq) */
U32 matchLengthSumBasePrice;/* to compare to log2(mlfreq) */ U32 matchLengthSumBasePrice;/* to compare to log2(mlfreq) */
U32 offCodeSumBasePrice; /* to compare to log2(offreq) */ U32 offCodeSumBasePrice; /* to compare to log2(offreq) */
ZSTD_OptPrice_e priceType; /* prices can be determined dynamically, or follow a pre-defined cost structure */ 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 */ const ZSTD_entropyCTables_t* symbolCosts; /* pre-calculated dictionary statistics */
ZSTD_literalCompressionMode_e literalCompressionMode;
} optState_t; } optState_t;
typedef struct { typedef struct {
ZSTD_entropyCTables_t entropy; ZSTD_entropyCTables_t entropy;
U32 rep[ZSTD_REP_NUM]; U32 rep[ZSTD_REP_NUM];
} ZSTD_compressedBlockState_t; } ZSTD_compressedBlockState_t;
typedef struct { typedef struct {
BYTE const* nextSrc; /* next block here to continue on current prefix */ BYTE const* nextSrc; /* next block here to continue on current prefix */
BYTE const* base; /* All regular indexes relative to this position */ BYTE const* base; /* All regular indexes relative to this position */
BYTE const* dictBase; /* extDict indexes relative to this position */ BYTE const* dictBase; /* extDict indexes relative to this position */
U32 dictLimit; /* below that point, need extDict */ U32 dictLimit; /* below that point, need extDict */
U32 lowLimit; /* below that point, no more data */ U32 lowLimit; /* below that point, no more valid data */
} ZSTD_window_t; } ZSTD_window_t;
typedef struct ZSTD_matchState_t ZSTD_matchState_t; typedef struct ZSTD_matchState_t ZSTD_matchState_t;
struct ZSTD_matchState_t { struct ZSTD_matchState_t {
ZSTD_window_t window; /* State for window round buffer management */ ZSTD_window_t window; /* State for window round buffer management */
U32 loadedDictEnd; /* index of end of dictionary */ U32 loadedDictEnd; /* index of end of dictionary, within context's referential.
* When loadedDictEnd != 0, a dictionary is in use, and still valid.
* This relies on a mechanism to set loadedDictEnd=0 when dictionary is no longer within distance.
* Such mechanism is provided within ZSTD_window_enforceMaxDist() and ZSTD_checkDictValidity().
* When dict referential is copied into active context (i.e. not attached),
* loadedDictEnd == dictSize, since referential starts from zero.
*/
U32 nextToUpdate; /* index from which to continue table update */ U32 nextToUpdate; /* index from which to continue table update */
U32 nextToUpdate3; /* index from which to continue table update */ U32 hashLog3; /* dispatch table for matches of len==3 : larger == faster, more memory */
U32 hashLog3; /* dispatch table : larger == faster, more memory */
U32* hashTable; U32* hashTable;
U32* hashTable3; U32* hashTable3;
U32* chainTable; U32* chainTable;
optState_t opt; /* optimal parser state */ optState_t opt; /* optimal parser state */
const ZSTD_matchState_t * dictMatchState; const ZSTD_matchState_t* dictMatchState;
ZSTD_compressionParameters cParams; ZSTD_compressionParameters cParams;
}; };
typedef struct { typedef struct {
ZSTD_compressedBlockState_t* prevCBlock; ZSTD_compressedBlockState_t* prevCBlock;
ZSTD_compressedBlockState_t* nextCBlock; ZSTD_compressedBlockState_t* nextCBlock;
ZSTD_matchState_t matchState; ZSTD_matchState_t matchState;
} ZSTD_blockState_t; } ZSTD_blockState_t;
struct ZSTD_CCtx_params_s { struct ZSTD_CCtx_params_s {
ZSTD_format_e format; ZSTD_format_e format;
ZSTD_compressionParameters cParams; ZSTD_compressionParameters cParams;
ZSTD_frameParameters fParams; ZSTD_frameParameters fParams;
int compressionLevel; int compressionLevel;
int forceWindow; /* force back-references to respect limit of int forceWindow; /* force back-references to respect limit of
* 1<<wLog, even for dictionary */ * 1<<wLog, even for dictionary */
size_t targetCBlockSize; /* Tries to fit compressed block size to be around targetCBlockSize.
* No target when targetCBlockSize == 0.
* There is no guarantee on compressed block size */
ZSTD_dictAttachPref_e attachDictPref; ZSTD_dictAttachPref_e attachDictPref;
ZSTD_literalCompressionMode_e literalCompressionMode;
/* Multithreading: used to pass parameters to mtctx */ /* Multithreading: used to pass parameters to mtctx */
int nbWorkers; int nbWorkers;
size_t jobSize; size_t jobSize;
int overlapLog; int overlapLog;
int rsyncable; int rsyncable;
/* Long distance matching parameters */ /* Long distance matching parameters */
char* outBuff; char* outBuff;
size_t outBuffSize; size_t outBuffSize;
size_t outBuffContentSize; size_t outBuffContentSize;
size_t outBuffFlushedSize; size_t outBuffFlushedSize;
ZSTD_cStreamStage streamStage; ZSTD_cStreamStage streamStage;
U32 frameEnded; U32 frameEnded;
/* Dictionary */ /* Dictionary */
ZSTD_CDict* cdictLocal; ZSTD_localDict localDict;
const ZSTD_CDict* cdict; const ZSTD_CDict* cdict;
ZSTD_prefixDict prefixDict; /* single-usage dictionary */ ZSTD_prefixDict prefixDict; /* single-usage dictionary */
/* Multi-threading */ /* Multi-threading */
#ifdef ZSTD_MULTITHREAD #ifdef ZSTD_MULTITHREAD
ZSTDMT_CCtx* mtctx; ZSTDMT_CCtx* mtctx;
#endif #endif
}; };
40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 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, 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,
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; static const U32 ML_deltaCode = 36;
return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase]; return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase];
} }
/* ZSTD_cParam_withinBounds:
* @return 1 if value is within cParam bounds,
* 0 otherwise */
MEM_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_minGain() :
* minimum compression required
* to generate a compress block or a compressed literals section.
* note : use same formula for both situations */
MEM_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;
}
/*! ZSTD_storeSeq() : /*! ZSTD_storeSeq() :
* Store a sequence (literal length, literals, offset code and match length code) into seqStore_t. * 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). * `offsetCode` : distance to match + 3 (values 1-3 are repCodes).
* `mlBase` : matchLength - MINMATCH * `mlBase` : matchLength - MINMATCH
*/ */
MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t mlBase) MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t mlBase)
{ {
#if defined(DEBUGLEVEL) && (DEBUGLEVEL >= 6) #if defined(DEBUGLEVEL) && (DEBUGLEVEL >= 6)
static const BYTE* g_start = NULL; 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 */ 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); { U32 const pos = (U32)((const BYTE*)literals - g_start);
DEBUGLOG(6, "Cpos%7u :%3u literals, match%4u bytes at offCode%7u", DEBUGLOG(6, "Cpos%7u :%3u literals, match%4u bytes at offCode%7u",
pos, (U32)litLength, (U32)mlBase+MINMATCH, (U32)offsetCode); pos, (U32)litLength, (U32)mlBase+MINMATCH, (U32)offsetCode);
} }
#endif #endif
assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq); assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq);
/* copy Literals */ /* copy Literals */
assert(seqStorePtr->maxNbLit <= 128 KB); assert(seqStorePtr->maxNbLit <= 128 KB);
assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit); assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit);
ZSTD_wildcopy(seqStorePtr->lit, literals, litLength); ZSTD_wildcopy(seqStorePtr->lit, literals, (ptrdiff_t)litLength, ZSTD_no_overlap);
seqStorePtr->lit += litLength; seqStorePtr->lit += litLength;
/* literal Length */ /* literal Length */
if (litLength>0xFFFF) { if (litLength>0xFFFF) {
assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */ assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */
seqStorePtr->longLengthID = 1; seqStorePtr->longLengthID = 1;
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
} }
hash *= prime8bytes; hash *= prime8bytes;
hash += toAdd + ZSTD_ROLL_HASH_CHAR_OFFSET; hash += toAdd + ZSTD_ROLL_HASH_CHAR_OFFSET;
return hash; return hash;
} }
/*-************************************* /*-*************************************
* Round buffer management * Round buffer management
***************************************/ ***************************************/
#if (ZSTD_WINDOWLOG_MAX_64 > 31)
# error "ZSTD_WINDOWLOG_MAX is too large : would overflow ZSTD_CURRENT_MAX"
#endif
/* Max current allowed */ /* Max current allowed */
#define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX)) #define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX))
/* Maximum chunk size before overflow correction needs to be called again */ /* Maximum chunk size before overflow correction needs to be called again */
#define ZSTD_CHUNKSIZE_MAX \ #define ZSTD_CHUNKSIZE_MAX \
( ((U32)-1) /* Maximum ending current index */ \ ( ((U32)-1) /* Maximum ending current index */ \
- ZSTD_CURRENT_MAX) /* Maximum beginning lowLimit */ - ZSTD_CURRENT_MAX) /* Maximum beginning lowLimit */
/** /**
return correction; return correction;
} }
/** /**
* ZSTD_window_enforceMaxDist(): * ZSTD_window_enforceMaxDist():
* Updates lowLimit so that: * Updates lowLimit so that:
* (srcEnd - base) - lowLimit == maxDist + loadedDictEnd * (srcEnd - base) - lowLimit == maxDist + loadedDictEnd
* *
* This allows a simple check that index >= lowLimit to see if index is valid. * It ensures index is valid as long as index >= lowLimit.
* This must be called before a block compression call, with srcEnd as the block * This must be called before a block compression call.
* source end. *
* loadedDictEnd is only defined if a dictionary is in use for current compression.
* As the name implies, loadedDictEnd represents the index at end of dictionary.
* The value lies within context's referential, it can be directly compared to blockEndIdx.
* *
* If loadedDictEndPtr is not NULL, we set it to zero once we update lowLimit. * If loadedDictEndPtr is NULL, no dictionary is in use, and we use loadedDictEnd == 0.
* This is because dictionaries are allowed to be referenced as long as the last * If loadedDictEndPtr is not NULL, we set it to zero after updating lowLimit.
* byte of the dictionary is in the window, but once they are out of range, * This is because dictionaries are allowed to be referenced fully
* they cannot be referenced. If loadedDictEndPtr is NULL, we use * as long as the last byte of the dictionary is in the window.
* loadedDictEnd == 0. * Once input has progressed beyond window size, dictionary cannot be referenced anymore.
* *
* In normal dict mode, the dict is between lowLimit and dictLimit. In * In normal dict mode, the dictionary lies between lowLimit and dictLimit.
* dictMatchState mode, lowLimit and dictLimit are the same, and the dictionary * In dictMatchState mode, lowLimit and dictLimit are the same,
* is below them. forceWindow and dictMatchState are therefore incompatible. * and the dictionary is below them.
* forceWindow and dictMatchState are therefore incompatible.
*/ */
MEM_STATIC void MEM_STATIC void
ZSTD_window_enforceMaxDist(ZSTD_window_t* window, ZSTD_window_enforceMaxDist(ZSTD_window_t* window,
void const* srcEnd, const void* blockEnd,
U32 maxDist, U32 maxDist,
U32* loadedDictEndPtr, U32* loadedDictEndPtr,
const ZSTD_matchState_t** dictMatchStatePtr) const ZSTD_matchState_t** dictMatchStatePtr)
{ {
U32 const blockEndIdx = (U32)((BYTE const*)srcEnd - window->base); U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
U32 loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0; U32 const loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0;
DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u", DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u",
(unsigned)blockEndIdx, (unsigned)maxDist); (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd);
/* - When there is no dictionary : loadedDictEnd == 0.
In which case, the test (blockEndIdx > maxDist) is merely to avoid
overflowing next operation `newLowLimit = blockEndIdx - maxDist`.
- When there is a standard dictionary :
Index referential is copied from the dictionary,
which means it starts from 0.
In which case, loadedDictEnd == dictSize,
and it makes sense to compare `blockEndIdx > maxDist + dictSize`
since `blockEndIdx` also starts from zero.
- When there is an attached dictionary :
loadedDictEnd is expressed within the referential of the context,
so it can be directly compared against blockEndIdx.
*/
if (blockEndIdx > maxDist + loadedDictEnd) { if (blockEndIdx > maxDist + loadedDictEnd) {
U32 const newLowLimit = blockEndIdx - maxDist; U32 const newLowLimit = blockEndIdx - maxDist;
if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit; if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit;
if (window->dictLimit < window->lowLimit) { if (window->dictLimit < window->lowLimit) {
DEBUGLOG(5, "Update dictLimit to match lowLimit, from %u to %u", DEBUGLOG(5, "Update dictLimit to match lowLimit, from %u to %u",
(unsigned)window->dictLimit, (unsigned)window->lowLimit); (unsigned)window->dictLimit, (unsigned)window->lowLimit);
window->dictLimit = window->lowLimit; window->dictLimit = window->lowLimit;
} }
if (loadedDictEndPtr) /* On reaching window size, dictionaries are invalidated */
if (loadedDictEndPtr) *loadedDictEndPtr = 0;
if (dictMatchStatePtr) *dictMatchStatePtr = NULL;
}
}
/* Similar to ZSTD_window_enforceMaxDist(),
* but only invalidates dictionary
* when input progresses beyond window size.
* assumption : loadedDictEndPtr and dictMatchStatePtr are valid (non NULL)
* loadedDictEnd uses same referential as window->base
* maxDist is the window size */
MEM_STATIC void
ZSTD_checkDictValidity(const ZSTD_window_t* window,
const void* blockEnd,
U32 maxDist,
U32* loadedDictEndPtr,
const ZSTD_matchState_t** dictMatchStatePtr)
{
assert(loadedDictEndPtr != NULL);
assert(dictMatchStatePtr != NULL);
{ U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
U32 const loadedDictEnd = *loadedDictEndPtr;
DEBUGLOG(5, "ZSTD_checkDictValidity: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u",
(unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd);
assert(blockEndIdx >= loadedDictEnd);
if (blockEndIdx > loadedDictEnd + maxDist) {
/* On reaching window size, dictionaries are invalidated.
* For simplification, if window size is reached anywhere within next block,
* the dictionary is invalidated for the full block.
*/
DEBUGLOG(6, "invalidating dictionary for current block (distance > windowSize)");
*loadedDictEndPtr = 0; *loadedDictEndPtr = 0;
if (dictMatchStatePtr)
*dictMatchStatePtr = NULL; *dictMatchStatePtr = NULL;
} } else {
if (*loadedDictEndPtr != 0) {
DEBUGLOG(6, "dictionary considered valid for current block");
} } }
} }
/** /**
* ZSTD_window_update(): * ZSTD_window_update():
* Updates the window by appending [src, src + srcSize) to the window. * Updates the window by appending [src, src + srcSize) to the window.
* If it is not contiguous, the current prefix becomes the extDict, and we * If it is not contiguous, the current prefix becomes the extDict, and we
* forget about the extDict. Handles overlap of the prefix and extDict. * forget about the extDict. Handles overlap of the prefix and extDict.
* Returns non-zero if the segment is contiguous. * Returns non-zero if the segment is contiguous.
ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase; ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase;
U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx; U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx;
window->lowLimit = lowLimitMax; window->lowLimit = lowLimitMax;
DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit); DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit);
} }
return contiguous; return contiguous;
} }
MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 current, unsigned windowLog)
{
U32 const maxDistance = 1U << windowLog;
U32 const lowestValid = ms->window.lowLimit;
U32 const withinWindow = (current - lowestValid > maxDistance) ? current - maxDistance : lowestValid;
U32 const isDictionary = (ms->loadedDictEnd != 0);
U32 const matchLowest = isDictionary ? lowestValid : withinWindow;
return matchLowest;
}
/* debug functions */ /* debug functions */
#if (DEBUGLEVEL>=2) #if (DEBUGLEVEL>=2)
MEM_STATIC double ZSTD_fWeight(U32 rawStat) MEM_STATIC double ZSTD_fWeight(U32 rawStat)
{ {
U32 const fp_accuracy = 8; U32 const fp_accuracy = 8;
U32 const fp_multiplier = (1 << fp_accuracy); U32 const fp_multiplier = (1 << fp_accuracy);
* @return : 0, or an error code */ * @return : 0, or an error code */
size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs, size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
const void* dict, size_t dictSize, const void* dict, size_t dictSize,
const ZSTD_CDict* cdict, const ZSTD_CDict* cdict,
ZSTD_CCtx_params params, unsigned long long pledgedSrcSize); ZSTD_CCtx_params params, unsigned long long pledgedSrcSize);
void ZSTD_resetSeqStore(seqStore_t* ssPtr); void ZSTD_resetSeqStore(seqStore_t* ssPtr);
/*! ZSTD_compressStream_generic() :
* Private use only. To be called from zstdmt_compress.c in single-thread mode. */
size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
ZSTD_outBuffer* output,
ZSTD_inBuffer* input,
ZSTD_EndDirective const flushMode);
/*! ZSTD_getCParamsFromCDict() : /*! ZSTD_getCParamsFromCDict() :
* as the name implies */ * as the name implies */
ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict); ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict);
/* ZSTD_compressBegin_advanced_internal() : /* ZSTD_compressBegin_advanced_internal() :
* Private use only. To be called from zstdmt_compress.c. */ * Private use only. To be called from zstdmt_compress.c. */
size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx, size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx,
const void* dict, size_t dictSize, const void* dict, size_t dictSize,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const void* dict,size_t dictSize, const void* dict,size_t dictSize,
ZSTD_CCtx_params params); ZSTD_CCtx_params params);
/* ZSTD_writeLastEmptyBlock() : /* ZSTD_writeLastEmptyBlock() :
* output an empty Block with end-of-frame mark to complete a frame * 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)) * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))
* or an error code if `dstCapcity` is too small (<ZSTD_blockHeaderSize) * or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)
*/ */
size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity); size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity);
/* ZSTD_referenceExternalSequences() : /* ZSTD_referenceExternalSequences() :
* Must be called before starting a compression operation. * Must be called before starting a compression operation.
* seqs must parse a prefix of the source. * seqs must parse a prefix of the source.
* This cannot be used when long range matching is enabled. * This cannot be used when long range matching is enabled.

contrib/python-zstandard/zstd/compress/zstd_compress_literals.h

  • This file was added.
/*
* 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_COMPRESS_LITERALS_H
#define ZSTD_COMPRESS_LITERALS_H
#include "zstd_compress_internal.h" /* ZSTD_hufCTables_t, ZSTD_minGain() */
size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
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);
#endif /* ZSTD_COMPRESS_LITERALS_H */

contrib/python-zstandard/zstd/compress/zstd_compress_literals.c

  • This file was added.
/*
* 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 "zstd_compress_literals.h"
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;
}
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;
}
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;
}

contrib/python-zstandard/zstd/compress/zstd_compress_sequences.h

  • This file was added.
/*
* 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_COMPRESS_SEQUENCES_H
#define ZSTD_COMPRESS_SEQUENCES_H
#include "fse.h" /* FSE_repeat, FSE_CTable */
#include "zstd_internal.h" /* symbolEncodingType_e, ZSTD_strategy */
typedef enum {
ZSTD_defaultDisallowed = 0,
ZSTD_defaultAllowed = 1
} ZSTD_defaultPolicy_e;
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);
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);
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);
#endif /* ZSTD_COMPRESS_SEQUENCES_H */

contrib/python-zstandard/zstd/compress/zstd_compress_sequences.c

  • This file was added.
/*
* 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 "zstd_compress_sequences.h"
/**
* -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,
};
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 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);
}
/**
* 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 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 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;
}
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;
}
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
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);
}

contrib/python-zstandard/zstd/compress/zstd_double_fast.c

size_t const lgHash = ZSTD_hashPtr(ip + i, hBitsL, 8); size_t const lgHash = ZSTD_hashPtr(ip + i, hBitsL, 8);
if (i == 0) if (i == 0)
hashSmall[smHash] = current + i; hashSmall[smHash] = current + i;
if (i == 0 || hashLarge[lgHash] == 0) if (i == 0 || hashLarge[lgHash] == 0)
hashLarge[lgHash] = current + i; hashLarge[lgHash] = current + i;
/* Only load extra positions for ZSTD_dtlm_full */ /* Only load extra positions for ZSTD_dtlm_full */
if (dtlm == ZSTD_dtlm_fast) if (dtlm == ZSTD_dtlm_fast)
break; break;
} } }
}
} }
FORCE_INLINE_TEMPLATE FORCE_INLINE_TEMPLATE
size_t ZSTD_compressBlock_doubleFast_generic( size_t ZSTD_compressBlock_doubleFast_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize, void const* src, size_t srcSize,
U32 const mls /* template */, ZSTD_dictMode_e const dictMode) U32 const mls /* template */, ZSTD_dictMode_e const dictMode)
{ {
ZSTD_compressionParameters const* cParams = &ms->cParams; ZSTD_compressionParameters const* cParams = &ms->cParams;
U32* const hashLong = ms->hashTable; U32* const hashLong = ms->hashTable;
const U32 hBitsL = cParams->hashLog; const U32 hBitsL = cParams->hashLog;
U32* const hashSmall = ms->chainTable; U32* const hashSmall = ms->chainTable;
const U32 hBitsS = cParams->chainLog; const U32 hBitsS = cParams->chainLog;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart; const BYTE* ip = istart;
const BYTE* anchor = istart; const BYTE* anchor = istart;
const U32 prefixLowestIndex = ms->window.dictLimit; const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
const U32 lowestValid = ms->window.dictLimit;
const U32 maxDistance = 1U << cParams->windowLog;
/* presumes that, if there is a dictionary, it must be using Attach mode */
const U32 prefixLowestIndex = (endIndex - lowestValid > maxDistance) ? endIndex - maxDistance : lowestValid;
const BYTE* const prefixLowest = base + prefixLowestIndex; const BYTE* const prefixLowest = base + prefixLowestIndex;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - HASH_READ_SIZE; const BYTE* const ilimit = iend - HASH_READ_SIZE;
U32 offset_1=rep[0], offset_2=rep[1]; U32 offset_1=rep[0], offset_2=rep[1];
U32 offsetSaved = 0; U32 offsetSaved = 0;
const ZSTD_matchState_t* const dms = ms->dictMatchState; const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_compressionParameters* const dictCParams = const ZSTD_compressionParameters* const dictCParams =
prefixLowestIndex - (U32)(dictEnd - dictBase) : prefixLowestIndex - (U32)(dictEnd - dictBase) :
0; 0;
const U32 dictHBitsL = dictMode == ZSTD_dictMatchState ? const U32 dictHBitsL = dictMode == ZSTD_dictMatchState ?
dictCParams->hashLog : hBitsL; dictCParams->hashLog : hBitsL;
const U32 dictHBitsS = dictMode == ZSTD_dictMatchState ? const U32 dictHBitsS = dictMode == ZSTD_dictMatchState ?
dictCParams->chainLog : hBitsS; dictCParams->chainLog : hBitsS;
const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictStart); const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictStart);
DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_generic");
assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState); assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);
/* if a dictionary is attached, it must be within window range */
if (dictMode == ZSTD_dictMatchState) {
assert(lowestValid + maxDistance >= endIndex);
}
/* init */ /* init */
ip += (dictAndPrefixLength == 0); ip += (dictAndPrefixLength == 0);
if (dictMode == ZSTD_noDict) { if (dictMode == ZSTD_noDict) {
U32 const maxRep = (U32)(ip - prefixLowest); U32 const maxRep = (U32)(ip - prefixLowest);
if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0; if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0; if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
} }
if (dictMode == ZSTD_dictMatchState) { if (dictMode == ZSTD_dictMatchState) {
/* check dictMatchState repcode */ /* check dictMatchState repcode */
if (dictMode == ZSTD_dictMatchState if (dictMode == ZSTD_dictMatchState
&& ((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */) && ((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) { && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend; const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4; mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
ip++; ip++;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
goto _match_stored; goto _match_stored;
} }
/* check noDict repcode */ /* check noDict repcode */
if ( dictMode == ZSTD_noDict if ( dictMode == ZSTD_noDict
&& ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) { && ((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; mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
ip++; ip++;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
goto _match_stored; goto _match_stored;
} }
if (matchIndexL > prefixLowestIndex) { if (matchIndexL > prefixLowestIndex) {
/* check prefix long match */ /* check prefix long match */
if (MEM_read64(matchLong) == MEM_read64(ip)) { if (MEM_read64(matchLong) == MEM_read64(ip)) {
mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8; mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
offset = (U32)(ip-matchLong); offset = (U32)(ip-matchLong);
while (((ip>anchor) & (matchLong>prefixLowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */ while (((ip>anchor) & (matchLong>prefixLowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
goto _match_found; goto _match_found;
} }
} else if (dictMode == ZSTD_dictMatchState) { } else if (dictMode == ZSTD_dictMatchState) {
/* check dictMatchState long match */ /* check dictMatchState long match */
U32 const dictMatchIndexL = dictHashLong[dictHL]; U32 const dictMatchIndexL = dictHashLong[dictHL];
const BYTE* dictMatchL = dictBase + dictMatchIndexL; const BYTE* dictMatchL = dictBase + dictMatchIndexL;
assert(dictMatchL < dictEnd); assert(dictMatchL < dictEnd);
if (dictMatchL > dictStart && MEM_read64(dictMatchL) == MEM_read64(ip)) { if (dictMatchL > dictStart && MEM_read64(dictMatchL) == MEM_read64(ip)) {
mLength = ZSTD_count_2segments(ip+8, dictMatchL+8, iend, dictEnd, prefixLowest) + 8; mLength = ZSTD_count_2segments(ip+8, dictMatchL+8, iend, dictEnd, prefixLowest) + 8;
offset = (U32)(current - dictMatchIndexL - dictIndexDelta); offset = (U32)(current - dictMatchIndexL - dictIndexDelta);
while (((ip>anchor) & (dictMatchL>dictStart)) && (ip[-1] == dictMatchL[-1])) { ip--; dictMatchL--; mLength++; } /* catch up */ while (((ip>anchor) & (dictMatchL>dictStart)) && (ip[-1] == dictMatchL[-1])) { ip--; dictMatchL--; mLength++; } /* catch up */
goto _match_found; goto _match_found;
} } }
}
if (matchIndexS > prefixLowestIndex) { if (matchIndexS > prefixLowestIndex) {
/* check prefix short match */ /* check prefix short match */
if (MEM_read32(match) == MEM_read32(ip)) { if (MEM_read32(match) == MEM_read32(ip)) {
goto _search_next_long; goto _search_next_long;
} }
} else if (dictMode == ZSTD_dictMatchState) { } else if (dictMode == ZSTD_dictMatchState) {
/* check dictMatchState short match */ /* check dictMatchState short match */
U32 const dictMatchIndexS = dictHashSmall[dictHS]; U32 const dictMatchIndexS = dictHashSmall[dictHS];
match = dictBase + dictMatchIndexS; match = dictBase + dictMatchIndexS;
matchIndexS = dictMatchIndexS + dictIndexDelta; matchIndexS = dictMatchIndexS + dictIndexDelta;
if (match > dictStart && MEM_read32(match) == MEM_read32(ip)) { if (match > dictStart && MEM_read32(match) == MEM_read32(ip)) {
goto _search_next_long; goto _search_next_long;
} } }
}
ip += ((ip-anchor) >> kSearchStrength) + 1; ip += ((ip-anchor) >> kSearchStrength) + 1;
continue; continue;
_search_next_long: _search_next_long:
{ { size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
size_t const dictHLNext = ZSTD_hashPtr(ip+1, dictHBitsL, 8); size_t const dictHLNext = ZSTD_hashPtr(ip+1, dictHBitsL, 8);
U32 const matchIndexL3 = hashLong[hl3]; U32 const matchIndexL3 = hashLong[hl3];
const BYTE* matchL3 = base + matchIndexL3; const BYTE* matchL3 = base + matchIndexL3;
hashLong[hl3] = current + 1; hashLong[hl3] = current + 1;
/* check prefix long +1 match */ /* check prefix long +1 match */
if (matchIndexL3 > prefixLowestIndex) { if (matchIndexL3 > prefixLowestIndex) {
if (MEM_read64(matchL3) == MEM_read64(ip+1)) { if (MEM_read64(matchL3) == MEM_read64(ip+1)) {
const BYTE* dictMatchL3 = dictBase + dictMatchIndexL3; const BYTE* dictMatchL3 = dictBase + dictMatchIndexL3;
assert(dictMatchL3 < dictEnd); assert(dictMatchL3 < dictEnd);
if (dictMatchL3 > dictStart && MEM_read64(dictMatchL3) == MEM_read64(ip+1)) { if (dictMatchL3 > dictStart && MEM_read64(dictMatchL3) == MEM_read64(ip+1)) {
mLength = ZSTD_count_2segments(ip+1+8, dictMatchL3+8, iend, dictEnd, prefixLowest) + 8; mLength = ZSTD_count_2segments(ip+1+8, dictMatchL3+8, iend, dictEnd, prefixLowest) + 8;
ip++; ip++;
offset = (U32)(current + 1 - dictMatchIndexL3 - dictIndexDelta); offset = (U32)(current + 1 - dictMatchIndexL3 - dictIndexDelta);
while (((ip>anchor) & (dictMatchL3>dictStart)) && (ip[-1] == dictMatchL3[-1])) { ip--; dictMatchL3--; mLength++; } /* catch up */ while (((ip>anchor) & (dictMatchL3>dictStart)) && (ip[-1] == dictMatchL3[-1])) { ip--; dictMatchL3--; mLength++; } /* catch up */
goto _match_found; goto _match_found;
} } } }
}
}
/* if no long +1 match, explore the short match we found */ /* if no long +1 match, explore the short match we found */
if (dictMode == ZSTD_dictMatchState && matchIndexS < prefixLowestIndex) { if (dictMode == ZSTD_dictMatchState && matchIndexS < prefixLowestIndex) {
mLength = ZSTD_count_2segments(ip+4, match+4, iend, dictEnd, prefixLowest) + 4; mLength = ZSTD_count_2segments(ip+4, match+4, iend, dictEnd, prefixLowest) + 4;
offset = (U32)(current - matchIndexS); offset = (U32)(current - matchIndexS);
while (((ip>anchor) & (match>dictStart)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */ while (((ip>anchor) & (match>dictStart)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
} else { } else {
mLength = ZSTD_count(ip+4, match+4, iend) + 4; mLength = ZSTD_count(ip+4, match+4, iend) + 4;
offset = (U32)(ip - match); offset = (U32)(ip - match);
while (((ip>anchor) & (match>prefixLowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */ while (((ip>anchor) & (match>prefixLowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
} }
/* fall-through */ /* fall-through */
_match_found: _match_found:
offset_2 = offset_1; offset_2 = offset_1;
offset_1 = offset; offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
_match_stored: _match_stored:
/* match found */ /* match found */
ip += mLength; ip += mLength;
anchor = ip; anchor = ip;
if (ip <= ilimit) { if (ip <= ilimit) {
/* Fill Table */ /* Complementary insertion */
hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = /* done after iLimit test, as candidates could be > iend-8 */
hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2; /* here because current+2 could be > iend-8 */ { U32 const indexToInsert = current+2;
hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base); hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
}
/* check immediate repcode */ /* check immediate repcode */
if (dictMode == ZSTD_dictMatchState) { if (dictMode == ZSTD_dictMatchState) {
while (ip <= ilimit) { while (ip <= ilimit) {
U32 const current2 = (U32)(ip-base); U32 const current2 = (U32)(ip-base);
U32 const repIndex2 = current2 - offset_2; U32 const repIndex2 = current2 - offset_2;
const BYTE* repMatch2 = dictMode == ZSTD_dictMatchState const BYTE* repMatch2 = dictMode == ZSTD_dictMatchState
&& repIndex2 < prefixLowestIndex ? && repIndex2 < prefixLowestIndex ?
dictBase - dictIndexDelta + repIndex2 : dictBase - dictIndexDelta + repIndex2 :
base + repIndex2; base + repIndex2;
if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */) if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) { && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex2 < prefixLowestIndex ? dictEnd : iend; const BYTE* const repEnd2 = repIndex2 < prefixLowestIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixLowest) + 4; 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 */ U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH); ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2; hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2; hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
ip += repLength2; ip += repLength2;
anchor = ip; anchor = ip;
continue; continue;
} }
break; break;
} } }
}
if (dictMode == ZSTD_noDict) { if (dictMode == ZSTD_noDict) {
while ( (ip <= ilimit) while ( (ip <= ilimit)
&& ( (offset_2>0) && ( (offset_2>0)
& (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) { & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
/* store sequence */ /* store sequence */
size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4; 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 */ 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); hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);
hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base); hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);
ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH); ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
ip += rLength; ip += rLength;
anchor = ip; anchor = ip;
continue; /* faster when present ... (?) */ continue; /* faster when present ... (?) */
} } } } } } }
} /* while (ip < ilimit) */
/* save reps for next block */ /* save reps for next block */
rep[0] = offset_1 ? offset_1 : offsetSaved; rep[0] = offset_1 ? offset_1 : offsetSaved;
rep[1] = offset_2 ? offset_2 : offsetSaved; rep[1] = offset_2 ? offset_2 : offsetSaved;
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(iend - anchor);
} }
size_t ZSTD_compressBlock_doubleFast( size_t ZSTD_compressBlock_doubleFast(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
const U32 mls = ms->cParams.minMatch; const U32 mls = ms->cParams.minMatch;
U32 const hBitsL = cParams->hashLog; U32 const hBitsL = cParams->hashLog;
U32* const hashSmall = ms->chainTable; U32* const hashSmall = ms->chainTable;
U32 const hBitsS = cParams->chainLog; U32 const hBitsS = cParams->chainLog;
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart; const BYTE* ip = istart;
const BYTE* anchor = istart; const BYTE* anchor = istart;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - 8; const BYTE* const ilimit = iend - 8;
const U32 prefixStartIndex = ms->window.dictLimit;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
const U32 lowLimit = ZSTD_getLowestMatchIndex(ms, endIndex, cParams->windowLog);
const U32 dictStartIndex = lowLimit;
const U32 dictLimit = ms->window.dictLimit;
const U32 prefixStartIndex = (dictLimit > lowLimit) ? dictLimit : lowLimit;
const BYTE* const prefixStart = base + prefixStartIndex; const BYTE* const prefixStart = base + prefixStartIndex;
const U32 dictStartIndex = ms->window.lowLimit;
const BYTE* const dictBase = ms->window.dictBase; const BYTE* const dictBase = ms->window.dictBase;
const BYTE* const dictStart = dictBase + dictStartIndex; const BYTE* const dictStart = dictBase + dictStartIndex;
const BYTE* const dictEnd = dictBase + prefixStartIndex; const BYTE* const dictEnd = dictBase + prefixStartIndex;
U32 offset_1=rep[0], offset_2=rep[1]; U32 offset_1=rep[0], offset_2=rep[1];
DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_extDict_generic (srcSize=%zu)", srcSize); DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_extDict_generic (srcSize=%zu)", srcSize);
/* if extDict is invalidated due to maxDistance, switch to "regular" variant */
if (prefixStartIndex == dictStartIndex)
return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, mls, ZSTD_noDict);
/* Search Loop */ /* Search Loop */
while (ip < ilimit) { /* < instead of <=, because (ip+1) */ while (ip < ilimit) { /* < instead of <=, because (ip+1) */
const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls); const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
const U32 matchIndex = hashSmall[hSmall]; const U32 matchIndex = hashSmall[hSmall];
const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base; const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;
const BYTE* match = matchBase + matchIndex; const BYTE* match = matchBase + matchIndex;
const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8); const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
hashSmall[hSmall] = hashLong[hLong] = current; /* update hash table */ hashSmall[hSmall] = hashLong[hLong] = current; /* update hash table */
if ((((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex doesn't overlap dict + prefix */ if ((((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex doesn't overlap dict + prefix */
& (repIndex > dictStartIndex)) & (repIndex > dictStartIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) { && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend; const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4; mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
ip++; ip++;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
} else { } else {
if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) { if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend; const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;
const BYTE* const lowMatchPtr = matchLongIndex < prefixStartIndex ? dictStart : prefixStart; const BYTE* const lowMatchPtr = matchLongIndex < prefixStartIndex ? dictStart : prefixStart;
U32 offset; U32 offset;
mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, prefixStart) + 8; mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, prefixStart) + 8;
offset = current - matchLongIndex; offset = current - matchLongIndex;
while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */ while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
offset_2 = offset_1; offset_2 = offset_1;
offset_1 = offset; offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) { } else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {
size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8); size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
U32 const matchIndex3 = hashLong[h3]; U32 const matchIndex3 = hashLong[h3];
const BYTE* const match3Base = matchIndex3 < prefixStartIndex ? dictBase : base; const BYTE* const match3Base = matchIndex3 < prefixStartIndex ? dictBase : base;
const BYTE* match3 = match3Base + matchIndex3; const BYTE* match3 = match3Base + matchIndex3;
U32 offset; U32 offset;
hashLong[h3] = current + 1; hashLong[h3] = current + 1;
if ( (matchIndex3 > dictStartIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) { if ( (matchIndex3 > dictStartIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {
const BYTE* const matchEnd = matchIndex3 < prefixStartIndex ? dictEnd : iend; const BYTE* const matchEnd = matchIndex3 < prefixStartIndex ? dictEnd : iend;
const BYTE* const lowMatchPtr = matchIndex3 < prefixStartIndex ? dictStart : prefixStart; const BYTE* const lowMatchPtr = matchIndex3 < prefixStartIndex ? dictStart : prefixStart;
mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, prefixStart) + 8; mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, prefixStart) + 8;
ip++; ip++;
offset = current+1 - matchIndex3; offset = current+1 - matchIndex3;
while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */ while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */
} else { } else {
const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend; const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart; const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4; mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
offset = current - matchIndex; offset = current - matchIndex;
while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */ while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
} }
offset_2 = offset_1; offset_2 = offset_1;
offset_1 = offset; offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} else { } else {
ip += ((ip-anchor) >> kSearchStrength) + 1; ip += ((ip-anchor) >> kSearchStrength) + 1;
continue; continue;
} } } }
/* found a match : store it */ /* move to next sequence start */
ip += mLength; ip += mLength;
anchor = ip; anchor = ip;
if (ip <= ilimit) { if (ip <= ilimit) {
/* Fill Table */ /* Complementary insertion */
hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2; /* done after iLimit test, as candidates could be > iend-8 */
hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2; { U32 const indexToInsert = current+2;
hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base); hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base); hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
}
/* check immediate repcode */ /* check immediate repcode */
while (ip <= ilimit) { while (ip <= ilimit) {
U32 const current2 = (U32)(ip-base); U32 const current2 = (U32)(ip-base);
U32 const repIndex2 = current2 - offset_2; U32 const repIndex2 = current2 - offset_2;
const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2; const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) /* intentional overflow : ensure repIndex2 doesn't overlap dict + prefix */ if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) /* intentional overflow : ensure repIndex2 doesn't overlap dict + prefix */
& (repIndex2 > dictStartIndex)) & (repIndex2 > dictStartIndex))
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) { && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
break; break;
} } } } } }
/* save reps for next block */ /* save reps for next block */
rep[0] = offset_1; rep[0] = offset_1;
rep[1] = offset_2; rep[1] = offset_2;
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(iend - anchor);
} }
size_t ZSTD_compressBlock_doubleFast_extDict( size_t ZSTD_compressBlock_doubleFast_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
U32 const mls = ms->cParams.minMatch; U32 const mls = ms->cParams.minMatch;

contrib/python-zstandard/zstd/compress/zstd_fast.c

/* /*
* Copyright (c) 2016-present, Yann Collet, Facebook, Inc. * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
* All rights reserved. * All rights reserved.
* *
* This source code is licensed under both the BSD-style license (found in the * 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 * 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). * in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses. * You may select, at your option, one of the above-listed licenses.
*/ */
#include "zstd_compress_internal.h" #include "zstd_compress_internal.h"
#include "zstd_fast.h" #include "zstd_fast.h"
void ZSTD_fillHashTable(ZSTD_matchState_t* ms, void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
void const* end, ZSTD_dictTableLoadMethod_e dtlm) const void* const end,
ZSTD_dictTableLoadMethod_e dtlm)
{ {
const ZSTD_compressionParameters* const cParams = &ms->cParams; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable; U32* const hashTable = ms->hashTable;
U32 const hBits = cParams->hashLog; U32 const hBits = cParams->hashLog;
U32 const mls = cParams->minMatch; U32 const mls = cParams->minMatch;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const BYTE* ip = base + ms->nextToUpdate; const BYTE* ip = base + ms->nextToUpdate;
const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE; const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
{ U32 p; { U32 p;
for (p = 1; p < fastHashFillStep; ++p) { for (p = 1; p < fastHashFillStep; ++p) {
size_t const hash = ZSTD_hashPtr(ip + p, hBits, mls); size_t const hash = ZSTD_hashPtr(ip + p, hBits, mls);
if (hashTable[hash] == 0) { /* not yet filled */ if (hashTable[hash] == 0) { /* not yet filled */
hashTable[hash] = current + p; hashTable[hash] = current + p;
} } } } } } } }
} }
FORCE_INLINE_TEMPLATE FORCE_INLINE_TEMPLATE
size_t ZSTD_compressBlock_fast_generic( size_t ZSTD_compressBlock_fast_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize, void const* src, size_t srcSize,
U32 const mls, ZSTD_dictMode_e const dictMode) 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 endIndex = (U32)((size_t)(istart - base) + srcSize);
const U32 maxDistance = 1U << cParams->windowLog;
const U32 validStartIndex = ms->window.dictLimit;
const U32 prefixStartIndex = (endIndex - validStartIndex > maxDistance) ? endIndex - maxDistance : validStartIndex;
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 */
DEBUGLOG(5, "ZSTD_compressBlock_fast_generic");
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 (size_t)(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; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable; U32* const hashTable = ms->hashTable;
U32 const hlog = cParams->hashLog; U32 const hlog = cParams->hashLog;
/* support stepSize of 0 */ /* support stepSize of 0 */
U32 const stepSize = cParams->targetLength + !(cParams->targetLength); U32 const stepSize = cParams->targetLength + !(cParams->targetLength);
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart; const BYTE* ip = istart;
const BYTE* anchor = istart; const BYTE* anchor = istart;
const U32 prefixStartIndex = ms->window.dictLimit; const U32 prefixStartIndex = ms->window.dictLimit;
const BYTE* const prefixStart = base + prefixStartIndex; const BYTE* const prefixStart = base + prefixStartIndex;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - HASH_READ_SIZE; const BYTE* const ilimit = iend - HASH_READ_SIZE;
U32 offset_1=rep[0], offset_2=rep[1]; U32 offset_1=rep[0], offset_2=rep[1];
U32 offsetSaved = 0; U32 offsetSaved = 0;
const ZSTD_matchState_t* const dms = ms->dictMatchState; const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_compressionParameters* const dictCParams = const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;
dictMode == ZSTD_dictMatchState ? const U32* const dictHashTable = dms->hashTable;
&dms->cParams : NULL; const U32 dictStartIndex = dms->window.dictLimit;
const U32* const dictHashTable = dictMode == ZSTD_dictMatchState ? const BYTE* const dictBase = dms->window.base;
dms->hashTable : NULL; const BYTE* const dictStart = dictBase + dictStartIndex;
const U32 dictStartIndex = dictMode == ZSTD_dictMatchState ? const BYTE* const dictEnd = dms->window.nextSrc;
dms->window.dictLimit : 0; const U32 dictIndexDelta = prefixStartIndex - (U32)(dictEnd - dictBase);
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 ?
prefixStartIndex - (U32)(dictEnd - dictBase) :
0;
const U32 dictAndPrefixLength = (U32)(ip - prefixStart + dictEnd - dictStart); const U32 dictAndPrefixLength = (U32)(ip - prefixStart + dictEnd - dictStart);
const U32 dictHLog = dictMode == ZSTD_dictMatchState ? const U32 dictHLog = dictCParams->hashLog;
dictCParams->hashLog : hlog;
assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState); /* if a dictionary is still attached, it necessarily means that
* it is within window size. So we just check it. */
/* otherwise, we would get index underflow when translating a dict index const U32 maxDistance = 1U << cParams->windowLog;
* into a local index */ const U32 endIndex = (U32)((size_t)(ip - base) + srcSize);
assert(dictMode != ZSTD_dictMatchState assert(endIndex - prefixStartIndex <= maxDistance);
|| prefixStartIndex >= (U32)(dictEnd - dictBase)); (void)maxDistance; (void)endIndex; /* these variables are not used when assert() is disabled */
/* ensure there will be no no underflow
* when translating a dict index into a local index */
assert(prefixStartIndex >= (U32)(dictEnd - dictBase));
/* init */ /* init */
DEBUGLOG(5, "ZSTD_compressBlock_fast_dictMatchState_generic");
ip += (dictAndPrefixLength == 0); ip += (dictAndPrefixLength == 0);
if (dictMode == ZSTD_noDict) {
U32 const maxRep = (U32)(ip - prefixStart);
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 /* dictMatchState repCode checks don't currently handle repCode == 0
* disabling. */ * disabling. */
assert(offset_1 <= dictAndPrefixLength); assert(offset_1 <= dictAndPrefixLength);
assert(offset_2 <= dictAndPrefixLength); assert(offset_2 <= dictAndPrefixLength);
}
/* Main Search Loop */ /* Main Search Loop */
while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */ while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
size_t mLength; size_t mLength;
size_t const h = ZSTD_hashPtr(ip, hlog, mls); size_t const h = ZSTD_hashPtr(ip, hlog, mls);
U32 const current = (U32)(ip-base); U32 const current = (U32)(ip-base);
U32 const matchIndex = hashTable[h]; U32 const matchIndex = hashTable[h];
const BYTE* match = base + matchIndex; const BYTE* match = base + matchIndex;
const U32 repIndex = current + 1 - offset_1; const U32 repIndex = current + 1 - offset_1;
const BYTE* repMatch = (dictMode == ZSTD_dictMatchState const BYTE* repMatch = (repIndex < prefixStartIndex) ?
&& repIndex < prefixStartIndex) ?
dictBase + (repIndex - dictIndexDelta) : dictBase + (repIndex - dictIndexDelta) :
base + repIndex; base + repIndex;
hashTable[h] = current; /* update hash table */ hashTable[h] = current; /* update hash table */
if ( (dictMode == ZSTD_dictMatchState) if ( ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
&& ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) { && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend; const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4; mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
ip++; ip++;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
} else 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);
} else if ( (matchIndex <= prefixStartIndex) ) { } else if ( (matchIndex <= prefixStartIndex) ) {
if (dictMode == ZSTD_dictMatchState) {
size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls); size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
U32 const dictMatchIndex = dictHashTable[dictHash]; U32 const dictMatchIndex = dictHashTable[dictHash];
const BYTE* dictMatch = dictBase + dictMatchIndex; const BYTE* dictMatch = dictBase + dictMatchIndex;
if (dictMatchIndex <= dictStartIndex || if (dictMatchIndex <= dictStartIndex ||
MEM_read32(dictMatch) != MEM_read32(ip)) { MEM_read32(dictMatch) != MEM_read32(ip)) {
assert(stepSize >= 1); assert(stepSize >= 1);
ip += ((ip-anchor) >> kSearchStrength) + stepSize; ip += ((ip-anchor) >> kSearchStrength) + stepSize;
continue; continue;
} else { } else {
/* found a dict match */ /* found a dict match */
U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta); U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);
mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4; mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
while (((ip>anchor) & (dictMatch>dictStart)) while (((ip>anchor) & (dictMatch>dictStart))
&& (ip[-1] == dictMatch[-1])) { && (ip[-1] == dictMatch[-1])) {
ip--; dictMatch--; mLength++; ip--; dictMatch--; mLength++;
} /* catch up */ } /* catch up */
offset_2 = offset_1; offset_2 = offset_1;
offset_1 = offset; offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
}
} else {
assert(stepSize >= 1);
ip += ((ip-anchor) >> kSearchStrength) + stepSize;
continue;
} }
} else if (MEM_read32(match) != MEM_read32(ip)) { } else if (MEM_read32(match) != MEM_read32(ip)) {
/* it's not a match, and we're not going to check the dictionary */ /* it's not a match, and we're not going to check the dictionary */
assert(stepSize >= 1); assert(stepSize >= 1);
ip += ((ip-anchor) >> kSearchStrength) + stepSize; ip += ((ip-anchor) >> kSearchStrength) + stepSize;
continue; continue;
} else { } else {
/* found a regular match */ /* found a regular match */
U32 const offset = (U32)(ip-match); U32 const offset = (U32)(ip-match);
mLength = ZSTD_count(ip+4, match+4, iend) + 4; mLength = ZSTD_count(ip+4, match+4, iend) + 4;
while (((ip>anchor) & (match>prefixStart)) while (((ip>anchor) & (match>prefixStart))
&& (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */ && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
offset_2 = offset_1; offset_2 = offset_1;
offset_1 = offset; offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} }
/* match found */ /* match found */
ip += mLength; ip += mLength;
anchor = ip; anchor = ip;
if (ip <= ilimit) { if (ip <= ilimit) {
/* Fill Table */ /* Fill Table */
assert(base+current+2 > istart); /* check base overflow */ 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(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); hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
/* check immediate repcode */ /* check immediate repcode */
if (dictMode == ZSTD_dictMatchState) {
while (ip <= ilimit) { while (ip <= ilimit) {
U32 const current2 = (U32)(ip-base); U32 const current2 = (U32)(ip-base);
U32 const repIndex2 = current2 - offset_2; U32 const repIndex2 = current2 - offset_2;
const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
dictBase - dictIndexDelta + repIndex2 : dictBase - dictIndexDelta + repIndex2 :
base + repIndex2; base + repIndex2;
if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */) if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) { && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend; const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4; 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 */ U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH); ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2; hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
ip += repLength2; ip += repLength2;
anchor = ip; anchor = ip;
continue; continue;
} }
break; 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 */
hashTable[ZSTD_hashPtr(ip, hlog, mls)] = (U32)(ip-base);
ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
ip += rLength;
anchor = ip;
continue; /* faster when present ... (?) */
} } } }
/* save reps for next block */ /* save reps for next block */
rep[0] = offset_1 ? offset_1 : offsetSaved; rep[0] = offset_1 ? offset_1 : offsetSaved;
rep[1] = offset_2 ? offset_2 : offsetSaved; rep[1] = offset_2 ? offset_2 : offsetSaved;
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(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, ZSTD_noDict);
case 5 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_noDict);
case 6 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_noDict);
case 7 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_noDict);
}
} }
size_t ZSTD_compressBlock_fast_dictMatchState( size_t ZSTD_compressBlock_fast_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
ZSTD_compressionParameters const* cParams = &ms->cParams; ZSTD_compressionParameters const* cParams = &ms->cParams;
U32 const mls = cParams->minMatch; U32 const mls = cParams->minMatch;
assert(ms->dictMatchState != NULL); assert(ms->dictMatchState != NULL);
switch(mls) switch(mls)
{ {
default: /* includes case 3 */ default: /* includes case 3 */
case 4 : case 4 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_dictMatchState); return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 4);
case 5 : case 5 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_dictMatchState); return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 5);
case 6 : case 6 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_dictMatchState); return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 6);
case 7 : case 7 :
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_dictMatchState); return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 7);
} }
} }
static size_t ZSTD_compressBlock_fast_extDict_generic( static size_t ZSTD_compressBlock_fast_extDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize, U32 const mls) void const* src, size_t srcSize, U32 const mls)
{ {
const ZSTD_compressionParameters* const cParams = &ms->cParams; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable; U32* const hashTable = ms->hashTable;
U32 const hlog = cParams->hashLog; U32 const hlog = cParams->hashLog;
/* support stepSize of 0 */ /* support stepSize of 0 */
U32 const stepSize = cParams->targetLength + !(cParams->targetLength); U32 const stepSize = cParams->targetLength + !(cParams->targetLength);
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const BYTE* const dictBase = ms->window.dictBase; const BYTE* const dictBase = ms->window.dictBase;
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart; const BYTE* ip = istart;
const BYTE* anchor = istart; const BYTE* anchor = istart;
const U32 dictStartIndex = ms->window.lowLimit; const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
const U32 lowLimit = ZSTD_getLowestMatchIndex(ms, endIndex, cParams->windowLog);
const U32 dictStartIndex = lowLimit;
const BYTE* const dictStart = dictBase + dictStartIndex; const BYTE* const dictStart = dictBase + dictStartIndex;
const U32 prefixStartIndex = ms->window.dictLimit; const U32 dictLimit = ms->window.dictLimit;
const U32 prefixStartIndex = dictLimit < lowLimit ? lowLimit : dictLimit;
const BYTE* const prefixStart = base + prefixStartIndex; const BYTE* const prefixStart = base + prefixStartIndex;
const BYTE* const dictEnd = dictBase + prefixStartIndex; const BYTE* const dictEnd = dictBase + prefixStartIndex;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - 8; const BYTE* const ilimit = iend - 8;
U32 offset_1=rep[0], offset_2=rep[1]; U32 offset_1=rep[0], offset_2=rep[1];
DEBUGLOG(5, "ZSTD_compressBlock_fast_extDict_generic");
/* switch to "regular" variant if extDict is invalidated due to maxDistance */
if (prefixStartIndex == dictStartIndex)
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, mls);
/* Search Loop */ /* Search Loop */
while (ip < ilimit) { /* < instead of <=, because (ip+1) */ while (ip < ilimit) { /* < instead of <=, because (ip+1) */
const size_t h = ZSTD_hashPtr(ip, hlog, mls); const size_t h = ZSTD_hashPtr(ip, hlog, mls);
const U32 matchIndex = hashTable[h]; const U32 matchIndex = hashTable[h];
const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base; const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;
const BYTE* match = matchBase + matchIndex; const BYTE* match = matchBase + matchIndex;
const U32 current = (U32)(ip-base); const U32 current = (U32)(ip-base);
const U32 repIndex = current + 1 - offset_1; const U32 repIndex = current + 1 - offset_1;
const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base; const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;
const BYTE* const repMatch = repBase + repIndex; const BYTE* const repMatch = repBase + repIndex;
size_t mLength; size_t mLength;
hashTable[h] = current; /* update hash table */ hashTable[h] = current; /* update hash table */
assert(offset_1 <= current +1); /* check repIndex */ assert(offset_1 <= current +1); /* check repIndex */
if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex)) if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) { && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend; const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4; mLength = ZSTD_count_2segments(ip+1 +4, repMatch +4, iend, repMatchEnd, prefixStart) + 4;
ip++; ip++;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
} else { } else {
if ( (matchIndex < dictStartIndex) || if ( (matchIndex < dictStartIndex) ||
(MEM_read32(match) != MEM_read32(ip)) ) { (MEM_read32(match) != MEM_read32(ip)) ) {
assert(stepSize >= 1); assert(stepSize >= 1);
ip += ((ip-anchor) >> kSearchStrength) + stepSize; ip += ((ip-anchor) >> kSearchStrength) + stepSize;
continue; continue;
} }
{ const BYTE* matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend; { const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
const BYTE* lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart; const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
U32 offset; U32 offset;
mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4; 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 */ while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
offset = current - matchIndex; offset = current - matchIndex;
offset_2 = offset_1; offset_2 = offset_1;
offset_1 = offset; offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH); ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} } } }
/* found a match : store it */ /* found a match : store it */
ip += mLength; ip += mLength;
anchor = ip; anchor = ip;
if (ip <= ilimit) { if (ip <= ilimit) {
/* Fill Table */ /* Fill Table */
hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2; hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2;
hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base); hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
/* check immediate repcode */ /* check immediate repcode */
while (ip <= ilimit) { while (ip <= ilimit) {
U32 const current2 = (U32)(ip-base); U32 const current2 = (U32)(ip-base);
U32 const repIndex2 = current2 - offset_2; U32 const repIndex2 = current2 - offset_2;
const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2; const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (repIndex2 > dictStartIndex)) /* intentional overflow */ if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (repIndex2 > dictStartIndex)) /* intentional overflow */
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) { && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend; const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4; 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 */ 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); ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2; hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
ip += repLength2; ip += repLength2;
anchor = ip; anchor = ip;
continue; continue;
} }
break; break;
} } } } } }
/* save reps for next block */ /* save reps for next block */
rep[0] = offset_1; rep[0] = offset_1;
rep[1] = offset_2; rep[1] = offset_2;
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(iend - anchor);
} }
size_t ZSTD_compressBlock_fast_extDict( size_t ZSTD_compressBlock_fast_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
ZSTD_compressionParameters const* cParams = &ms->cParams; ZSTD_compressionParameters const* cParams = &ms->cParams;

contrib/python-zstandard/zstd/compress/zstd_lazy.h

#if defined (__cplusplus) #if defined (__cplusplus)
extern "C" { extern "C" {
#endif #endif
#include "zstd_compress_internal.h" #include "zstd_compress_internal.h"
U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip); 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(). pre-emptively increase value of ZSTD_DUBT_UNSORTED_MARK */ 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( size_t ZSTD_compressBlock_btlazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize); void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2( size_t ZSTD_compressBlock_lazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize); void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy( size_t ZSTD_compressBlock_lazy(

contrib/python-zstandard/zstd/compress/zstd_lazy.c

const BYTE* const iend = (current>=dictLimit) ? inputEnd : dictBase + dictLimit; const BYTE* const iend = (current>=dictLimit) ? inputEnd : dictBase + dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit; const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const prefixStart = base + dictLimit; const BYTE* const prefixStart = base + dictLimit;
const BYTE* match; const BYTE* match;
U32* smallerPtr = bt + 2*(current&btMask); U32* smallerPtr = bt + 2*(current&btMask);
U32* largerPtr = smallerPtr + 1; 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 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 dummy32; /* to be nullified at the end */
U32 const windowLow = ms->window.lowLimit; U32 const windowValid = ms->window.lowLimit;
U32 const maxDistance = 1U << cParams->windowLog;
U32 const windowLow = (current - windowValid > maxDistance) ? current - maxDistance : windowValid;
DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)", DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)",
current, dictLimit, windowLow); current, dictLimit, windowLow);
assert(current >= btLow); assert(current >= btLow);
assert(ip < iend); /* condition for ZSTD_count */ assert(ip < iend); /* condition for ZSTD_count */
while (nbCompares-- && (matchIndex > windowLow)) { while (nbCompares-- && (matchIndex > windowLow)) {
U32* const nextPtr = bt + 2*(matchIndex & btMask); U32* const nextPtr = bt + 2*(matchIndex & btMask);
const ZSTD_compressionParameters* const cParams = &ms->cParams; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable; U32* const hashTable = ms->hashTable;
U32 const hashLog = cParams->hashLog; U32 const hashLog = cParams->hashLog;
size_t const h = ZSTD_hashPtr(ip, hashLog, mls); size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
U32 matchIndex = hashTable[h]; U32 matchIndex = hashTable[h];
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
U32 const current = (U32)(ip-base); U32 const current = (U32)(ip-base);
U32 const windowLow = ms->window.lowLimit; U32 const windowLow = ZSTD_getLowestMatchIndex(ms, current, cParams->windowLog);
U32* const bt = ms->chainTable; U32* const bt = ms->chainTable;
U32 const btLog = cParams->chainLog - 1; U32 const btLog = cParams->chainLog - 1;
U32 const btMask = (1 << btLog) - 1; U32 const btMask = (1 << btLog) - 1;
U32 const btLow = (btMask >= current) ? 0 : current - btMask; U32 const btLow = (btMask >= current) ? 0 : current - btMask;
U32 const unsortLimit = MAX(btLow, windowLow); U32 const unsortLimit = MAX(btLow, windowLow);
U32* nextCandidate = bt + 2*(matchIndex&btMask); U32* nextCandidate = bt + 2*(matchIndex&btMask);
U32* const chainTable = ms->chainTable; U32* const chainTable = ms->chainTable;
const U32 chainSize = (1 << cParams->chainLog); const U32 chainSize = (1 << cParams->chainLog);
const U32 chainMask = chainSize-1; const U32 chainMask = chainSize-1;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const BYTE* const dictBase = ms->window.dictBase; const BYTE* const dictBase = ms->window.dictBase;
const U32 dictLimit = ms->window.dictLimit; const U32 dictLimit = ms->window.dictLimit;
const BYTE* const prefixStart = base + dictLimit; const BYTE* const prefixStart = base + dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit; const BYTE* const dictEnd = dictBase + dictLimit;
const U32 lowLimit = ms->window.lowLimit;
const U32 current = (U32)(ip-base); const U32 current = (U32)(ip-base);
const U32 maxDistance = 1U << cParams->windowLog;
const U32 lowestValid = ms->window.lowLimit;
const U32 withinMaxDistance = (current - lowestValid > maxDistance) ? current - maxDistance : lowestValid;
const U32 isDictionary = (ms->loadedDictEnd != 0);
const U32 lowLimit = isDictionary ? lowestValid : withinMaxDistance;
const U32 minChain = current > chainSize ? current - chainSize : 0; const U32 minChain = current > chainSize ? current - chainSize : 0;
U32 nbAttempts = 1U << cParams->searchLog; U32 nbAttempts = 1U << cParams->searchLog;
size_t ml=4-1; size_t ml=4-1;
/* HC4 match finder */ /* HC4 match finder */
U32 matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls); U32 matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls);
for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) { for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) {
case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_extDict); case 6 : return ZSTD_HcFindBestMatch_generic(ms, ip, iLimit, offsetPtr, 6, ZSTD_extDict);
} }
} }
/* ******************************* /* *******************************
* Common parser - lazy strategy * Common parser - lazy strategy
*********************************/ *********************************/
FORCE_INLINE_TEMPLATE typedef enum { search_hashChain, search_binaryTree } searchMethod_e;
size_t ZSTD_compressBlock_lazy_generic(
FORCE_INLINE_TEMPLATE size_t
ZSTD_compressBlock_lazy_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, ZSTD_matchState_t* ms, seqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM], U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize, const void* src, size_t srcSize,
const U32 searchMethod, const U32 depth, const searchMethod_e searchMethod, const U32 depth,
ZSTD_dictMode_e const dictMode) ZSTD_dictMode_e const dictMode)
{ {
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart; const BYTE* ip = istart;
const BYTE* anchor = istart; const BYTE* anchor = istart;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - 8; const BYTE* const ilimit = iend - 8;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const U32 prefixLowestIndex = ms->window.dictLimit; const U32 prefixLowestIndex = ms->window.dictLimit;
const BYTE* const prefixLowest = base + prefixLowestIndex; const BYTE* const prefixLowest = base + prefixLowestIndex;
typedef size_t (*searchMax_f)( typedef size_t (*searchMax_f)(
ZSTD_matchState_t* ms, ZSTD_matchState_t* ms,
const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr); const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);
searchMax_f const searchMax = dictMode == ZSTD_dictMatchState ? searchMax_f const searchMax = dictMode == ZSTD_dictMatchState ?
(searchMethod ? ZSTD_BtFindBestMatch_dictMatchState_selectMLS : ZSTD_HcFindBestMatch_dictMatchState_selectMLS) : (searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_dictMatchState_selectMLS
(searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS); : ZSTD_HcFindBestMatch_dictMatchState_selectMLS) :
(searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_selectMLS
: ZSTD_HcFindBestMatch_selectMLS);
U32 offset_1 = rep[0], offset_2 = rep[1], savedOffset=0; U32 offset_1 = rep[0], offset_2 = rep[1], savedOffset=0;
const ZSTD_matchState_t* const dms = ms->dictMatchState; const ZSTD_matchState_t* const dms = ms->dictMatchState;
const U32 dictLowestIndex = dictMode == ZSTD_dictMatchState ? const U32 dictLowestIndex = dictMode == ZSTD_dictMatchState ?
dms->window.dictLimit : 0; dms->window.dictLimit : 0;
const BYTE* const dictBase = dictMode == ZSTD_dictMatchState ? const BYTE* const dictBase = dictMode == ZSTD_dictMatchState ?
dms->window.base : NULL; dms->window.base : NULL;
const BYTE* const dictLowest = dictMode == ZSTD_dictMatchState ? const BYTE* const dictLowest = dictMode == ZSTD_dictMatchState ?
dictBase + dictLowestIndex : NULL; dictBase + dictLowestIndex : NULL;
const BYTE* const dictEnd = dictMode == ZSTD_dictMatchState ? const BYTE* const dictEnd = dictMode == ZSTD_dictMatchState ?
dms->window.nextSrc : NULL; dms->window.nextSrc : NULL;
const U32 dictIndexDelta = dictMode == ZSTD_dictMatchState ? const U32 dictIndexDelta = dictMode == ZSTD_dictMatchState ?
prefixLowestIndex - (U32)(dictEnd - dictBase) : prefixLowestIndex - (U32)(dictEnd - dictBase) :
0; 0;
const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictLowest); const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictLowest);
/* init */ /* init */
ip += (dictAndPrefixLength == 0); ip += (dictAndPrefixLength == 0);
ms->nextToUpdate3 = ms->nextToUpdate;
if (dictMode == ZSTD_noDict) { if (dictMode == ZSTD_noDict) {
U32 const maxRep = (U32)(ip - prefixLowest); U32 const maxRep = (U32)(ip - prefixLowest);
if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0; if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;
if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0; if (offset_1 > maxRep) savedOffset = offset_1, offset_1 = 0;
} }
if (dictMode == ZSTD_dictMatchState) { if (dictMode == ZSTD_dictMatchState) {
/* dictMatchState repCode checks don't currently handle repCode == 0 /* dictMatchState repCode checks don't currently handle repCode == 0
* disabling. */ * disabling. */
continue; /* faster when present ... (?) */ continue; /* faster when present ... (?) */
} } } } } }
/* Save reps for next block */ /* Save reps for next block */
rep[0] = offset_1 ? offset_1 : savedOffset; rep[0] = offset_1 ? offset_1 : savedOffset;
rep[1] = offset_2 ? offset_2 : savedOffset; rep[1] = offset_2 ? offset_2 : savedOffset;
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(iend - anchor);
} }
size_t ZSTD_compressBlock_btlazy2( size_t ZSTD_compressBlock_btlazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 1, 2, ZSTD_noDict); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
} }
size_t ZSTD_compressBlock_lazy2( size_t ZSTD_compressBlock_lazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 2, ZSTD_noDict); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
} }
size_t ZSTD_compressBlock_lazy( size_t ZSTD_compressBlock_lazy(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 1, ZSTD_noDict); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
} }
size_t ZSTD_compressBlock_greedy( size_t ZSTD_compressBlock_greedy(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 0, ZSTD_noDict); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict);
} }
size_t ZSTD_compressBlock_btlazy2_dictMatchState( size_t ZSTD_compressBlock_btlazy2_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 1, 2, ZSTD_dictMatchState); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
} }
size_t ZSTD_compressBlock_lazy2_dictMatchState( size_t ZSTD_compressBlock_lazy2_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 2, ZSTD_dictMatchState); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
} }
size_t ZSTD_compressBlock_lazy_dictMatchState( size_t ZSTD_compressBlock_lazy_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 1, ZSTD_dictMatchState); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
} }
size_t ZSTD_compressBlock_greedy_dictMatchState( size_t ZSTD_compressBlock_greedy_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 0, ZSTD_dictMatchState); return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState);
} }
FORCE_INLINE_TEMPLATE FORCE_INLINE_TEMPLATE
size_t ZSTD_compressBlock_lazy_extDict_generic( size_t ZSTD_compressBlock_lazy_extDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, ZSTD_matchState_t* ms, seqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM], U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize, const void* src, size_t srcSize,
const U32 searchMethod, const U32 depth) const searchMethod_e searchMethod, const U32 depth)
{ {
const BYTE* const istart = (const BYTE*)src; const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart; const BYTE* ip = istart;
const BYTE* anchor = istart; const BYTE* anchor = istart;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - 8; const BYTE* const ilimit = iend - 8;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const U32 dictLimit = ms->window.dictLimit; const U32 dictLimit = ms->window.dictLimit;
const U32 lowestIndex = ms->window.lowLimit; const U32 lowestIndex = ms->window.lowLimit;
const BYTE* const prefixStart = base + dictLimit; const BYTE* const prefixStart = base + dictLimit;
const BYTE* const dictBase = ms->window.dictBase; const BYTE* const dictBase = ms->window.dictBase;
const BYTE* const dictEnd = dictBase + dictLimit; const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const dictStart = dictBase + lowestIndex; const BYTE* const dictStart = dictBase + lowestIndex;
typedef size_t (*searchMax_f)( typedef size_t (*searchMax_f)(
ZSTD_matchState_t* ms, ZSTD_matchState_t* ms,
const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr); const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);
searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_extDict_selectMLS : ZSTD_HcFindBestMatch_extDict_selectMLS; searchMax_f searchMax = searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_extDict_selectMLS : ZSTD_HcFindBestMatch_extDict_selectMLS;
U32 offset_1 = rep[0], offset_2 = rep[1]; U32 offset_1 = rep[0], offset_2 = rep[1];
/* init */ /* init */
ms->nextToUpdate3 = ms->nextToUpdate;
ip += (ip == prefixStart); ip += (ip == prefixStart);
/* Match Loop */ /* Match Loop */
while (ip < ilimit) { while (ip < ilimit) {
size_t matchLength=0; size_t matchLength=0;
size_t offset=0; size_t offset=0;
const BYTE* start=ip+1; const BYTE* start=ip+1;
U32 current = (U32)(ip-base); U32 current = (U32)(ip-base);
break; break;
} } } }
/* Save reps for next block */ /* Save reps for next block */
rep[0] = offset_1; rep[0] = offset_1;
rep[1] = offset_2; rep[1] = offset_2;
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(iend - anchor);
} }
size_t ZSTD_compressBlock_greedy_extDict( size_t ZSTD_compressBlock_greedy_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 0); return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0);
} }
size_t ZSTD_compressBlock_lazy_extDict( size_t ZSTD_compressBlock_lazy_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 1); return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1);
} }
size_t ZSTD_compressBlock_lazy2_extDict( size_t ZSTD_compressBlock_lazy2_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 2); return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2);
} }
size_t ZSTD_compressBlock_btlazy2_extDict( size_t ZSTD_compressBlock_btlazy2_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
{ {
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 1, 2); return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2);
} }

contrib/python-zstandard/zstd/compress/zstd_ldm.c

size_t leftoverSize = 0; size_t leftoverSize = 0;
assert(ZSTD_CHUNKSIZE_MAX >= kMaxChunkSize); assert(ZSTD_CHUNKSIZE_MAX >= kMaxChunkSize);
/* Check that ZSTD_window_update() has been called for this chunk prior /* Check that ZSTD_window_update() has been called for this chunk prior
* to passing it to this function. * to passing it to this function.
*/ */
assert(ldmState->window.nextSrc >= (BYTE const*)src + srcSize); assert(ldmState->window.nextSrc >= (BYTE const*)src + srcSize);
/* The input could be very large (in zstdmt), so it must be broken up into /* The input could be very large (in zstdmt), so it must be broken up into
* chunks to enforce the maximmum distance and handle overflow correction. * chunks to enforce the maximum distance and handle overflow correction.
*/ */
assert(sequences->pos <= sequences->size); assert(sequences->pos <= sequences->size);
assert(sequences->size <= sequences->capacity); assert(sequences->size <= sequences->capacity);
for (chunk = 0; chunk < nbChunks && sequences->size < sequences->capacity; ++chunk) { for (chunk = 0; chunk < nbChunks && sequences->size < sequences->capacity; ++chunk) {
BYTE const* const chunkStart = istart + chunk * kMaxChunkSize; BYTE const* const chunkStart = istart + chunk * kMaxChunkSize;
size_t const remaining = (size_t)(iend - chunkStart); size_t const remaining = (size_t)(iend - chunkStart);
BYTE const *const chunkEnd = BYTE const *const chunkEnd =
(remaining < kMaxChunkSize) ? iend : chunkStart + kMaxChunkSize; (remaining < kMaxChunkSize) ? iend : chunkStart + kMaxChunkSize;
size_t const chunkSize = chunkEnd - chunkStart; size_t const chunkSize = chunkEnd - chunkStart;
size_t newLeftoverSize; size_t newLeftoverSize;
size_t const prevSize = sequences->size; size_t const prevSize = sequences->size;
assert(chunkStart < iend); assert(chunkStart < iend);
/* 1. Perform overflow correction if necessary. */ /* 1. Perform overflow correction if necessary. */
if (ZSTD_window_needOverflowCorrection(ldmState->window, chunkEnd)) { if (ZSTD_window_needOverflowCorrection(ldmState->window, chunkEnd)) {
U32 const ldmHSize = 1U << params->hashLog; U32 const ldmHSize = 1U << params->hashLog;
U32 const correction = ZSTD_window_correctOverflow( U32 const correction = ZSTD_window_correctOverflow(
&ldmState->window, /* cycleLog */ 0, maxDist, src); &ldmState->window, /* cycleLog */ 0, maxDist, chunkStart);
ZSTD_ldm_reduceTable(ldmState->hashTable, ldmHSize, correction); ZSTD_ldm_reduceTable(ldmState->hashTable, ldmHSize, correction);
} }
/* 2. We enforce the maximum offset allowed. /* 2. We enforce the maximum offset allowed.
* *
* kMaxChunkSize should be small enough that we don't lose too much of * kMaxChunkSize should be small enough that we don't lose too much of
* the window through early invalidation. * the window through early invalidation.
* TODO: * Test the chunk size. * TODO: * Test the chunk size.
* * Try invalidation after the sequence generation and test the * * Try invalidation after the sequence generation and test the

contrib/python-zstandard/zstd/compress/zstd_opt.c

* @return price in bytes as fractional value * @return price in bytes as fractional value
* for debug messages only */ * for debug messages only */
MEM_STATIC double ZSTD_fCost(U32 price) MEM_STATIC double ZSTD_fCost(U32 price)
{ {
return (double)price / (BITCOST_MULTIPLIER*8); return (double)price / (BITCOST_MULTIPLIER*8);
} }
#endif #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) static void ZSTD_setBasePrices(optState_t* optPtr, int optLevel)
{ {
if (ZSTD_compressedLiterals(optPtr))
optPtr->litSumBasePrice = WEIGHT(optPtr->litSum, optLevel); optPtr->litSumBasePrice = WEIGHT(optPtr->litSum, optLevel);
optPtr->litLengthSumBasePrice = WEIGHT(optPtr->litLengthSum, optLevel); optPtr->litLengthSumBasePrice = WEIGHT(optPtr->litLengthSum, optLevel);
optPtr->matchLengthSumBasePrice = WEIGHT(optPtr->matchLengthSum, optLevel); optPtr->matchLengthSumBasePrice = WEIGHT(optPtr->matchLengthSum, optLevel);
optPtr->offCodeSumBasePrice = WEIGHT(optPtr->offCodeSum, optLevel); optPtr->offCodeSumBasePrice = WEIGHT(optPtr->offCodeSum, optLevel);
} }
/* ZSTD_downscaleStat() : /* ZSTD_downscaleStat() :
* reduce all elements in table by a factor 2^(ZSTD_FREQ_DIV+malus) * reduce all elements in table by a factor 2^(ZSTD_FREQ_DIV+malus)
* or init from zero, using src for literals stats, or flat 1 for match symbols * 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. * otherwise downscale existing stats, to be used as seed for next block.
*/ */
static void static void
ZSTD_rescaleFreqs(optState_t* const optPtr, ZSTD_rescaleFreqs(optState_t* const optPtr,
const BYTE* const src, size_t const srcSize, const BYTE* const src, size_t const srcSize,
int const optLevel) int const optLevel)
{ {
int const compressedLiterals = ZSTD_compressedLiterals(optPtr);
DEBUGLOG(5, "ZSTD_rescaleFreqs (srcSize=%u)", (unsigned)srcSize); DEBUGLOG(5, "ZSTD_rescaleFreqs (srcSize=%u)", (unsigned)srcSize);
optPtr->priceType = zop_dynamic; optPtr->priceType = zop_dynamic;
if (optPtr->litLengthSum == 0) { /* first block : init */ if (optPtr->litLengthSum == 0) { /* first block : init */
if (srcSize <= ZSTD_PREDEF_THRESHOLD) { /* heuristic */ if (srcSize <= ZSTD_PREDEF_THRESHOLD) { /* heuristic */
DEBUGLOG(5, "(srcSize <= ZSTD_PREDEF_THRESHOLD) => zop_predef"); DEBUGLOG(5, "(srcSize <= ZSTD_PREDEF_THRESHOLD) => zop_predef");
optPtr->priceType = zop_predef; optPtr->priceType = zop_predef;
} }
assert(optPtr->symbolCosts != NULL); assert(optPtr->symbolCosts != NULL);
if (optPtr->symbolCosts->huf.repeatMode == HUF_repeat_valid) { if (optPtr->symbolCosts->huf.repeatMode == HUF_repeat_valid) {
/* huffman table presumed generated by dictionary */ /* huffman table presumed generated by dictionary */
optPtr->priceType = zop_dynamic; optPtr->priceType = zop_dynamic;
if (compressedLiterals) {
unsigned lit;
assert(optPtr->litFreq != NULL); assert(optPtr->litFreq != NULL);
optPtr->litSum = 0; optPtr->litSum = 0;
{ unsigned lit;
for (lit=0; lit<=MaxLit; lit++) { for (lit=0; lit<=MaxLit; lit++) {
U32 const scaleLog = 11; /* scale to 2K */ U32 const scaleLog = 11; /* scale to 2K */
U32 const bitCost = HUF_getNbBits(optPtr->symbolCosts->huf.CTable, lit); U32 const bitCost = HUF_getNbBits(optPtr->symbolCosts->huf.CTable, lit);
assert(bitCost <= scaleLog); assert(bitCost <= scaleLog);
optPtr->litFreq[lit] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/; optPtr->litFreq[lit] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;
optPtr->litSum += optPtr->litFreq[lit]; optPtr->litSum += optPtr->litFreq[lit];
} } } }
assert(bitCost < scaleLog); assert(bitCost < scaleLog);
optPtr->offCodeFreq[of] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/; optPtr->offCodeFreq[of] = bitCost ? 1 << (scaleLog-bitCost) : 1 /*minimum to calculate cost*/;
optPtr->offCodeSum += optPtr->offCodeFreq[of]; optPtr->offCodeSum += optPtr->offCodeFreq[of];
} } } }
} else { /* not a dictionary */ } else { /* not a dictionary */
assert(optPtr->litFreq != NULL); assert(optPtr->litFreq != NULL);
{ unsigned lit = MaxLit; if (compressedLiterals) {
unsigned lit = MaxLit;
HIST_count_simple(optPtr->litFreq, &lit, src, srcSize); /* use raw first block to init statistics */ HIST_count_simple(optPtr->litFreq, &lit, src, srcSize); /* use raw first block to init statistics */
}
optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1); optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
}
{ unsigned ll; { unsigned ll;
for (ll=0; ll<=MaxLL; ll++) for (ll=0; ll<=MaxLL; ll++)
optPtr->litLengthFreq[ll] = 1; optPtr->litLengthFreq[ll] = 1;
} }
optPtr->litLengthSum = MaxLL+1; optPtr->litLengthSum = MaxLL+1;
{ unsigned ml; { unsigned ml;
for (ml=0; ml<=MaxML; ml++) for (ml=0; ml<=MaxML; ml++)
optPtr->matchLengthFreq[ml] = 1; optPtr->matchLengthFreq[ml] = 1;
} }
optPtr->matchLengthSum = MaxML+1; optPtr->matchLengthSum = MaxML+1;
{ unsigned of; { unsigned of;
for (of=0; of<=MaxOff; of++) for (of=0; of<=MaxOff; of++)
optPtr->offCodeFreq[of] = 1; optPtr->offCodeFreq[of] = 1;
} }
optPtr->offCodeSum = MaxOff+1; optPtr->offCodeSum = MaxOff+1;
} }
} else { /* new block : re-use previous statistics, scaled down */ } else { /* new block : re-use previous statistics, scaled down */
if (compressedLiterals)
optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1); optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
optPtr->litLengthSum = ZSTD_downscaleStat(optPtr->litLengthFreq, MaxLL, 0); optPtr->litLengthSum = ZSTD_downscaleStat(optPtr->litLengthFreq, MaxLL, 0);
optPtr->matchLengthSum = ZSTD_downscaleStat(optPtr->matchLengthFreq, MaxML, 0); optPtr->matchLengthSum = ZSTD_downscaleStat(optPtr->matchLengthFreq, MaxML, 0);
optPtr->offCodeSum = ZSTD_downscaleStat(optPtr->offCodeFreq, MaxOff, 0); optPtr->offCodeSum = ZSTD_downscaleStat(optPtr->offCodeFreq, MaxOff, 0);
} }
ZSTD_setBasePrices(optPtr, optLevel); ZSTD_setBasePrices(optPtr, optLevel);
} }
/* ZSTD_rawLiteralsCost() : /* ZSTD_rawLiteralsCost() :
* price of literals (only) in specified segment (which length can be 0). * price of literals (only) in specified segment (which length can be 0).
* does not include price of literalLength symbol */ * does not include price of literalLength symbol */
static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength, static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength,
const optState_t* const optPtr, const optState_t* const optPtr,
int optLevel) int optLevel)
{ {
if (litLength == 0) return 0; if (litLength == 0) return 0;
if (!ZSTD_compressedLiterals(optPtr))
return (litLength << 3) * BITCOST_MULTIPLIER; /* Uncompressed - 8 bytes per literal. */
if (optPtr->priceType == zop_predef) if (optPtr->priceType == zop_predef)
return (litLength*6) * BITCOST_MULTIPLIER; /* 6 bit per literal - no statistic used */ return (litLength*6) * BITCOST_MULTIPLIER; /* 6 bit per literal - no statistic used */
/* dynamic statistics */ /* dynamic statistics */
{ U32 price = litLength * optPtr->litSumBasePrice; { U32 price = litLength * optPtr->litSumBasePrice;
U32 u; U32 u;
for (u=0; u < litLength; u++) { for (u=0; u < litLength; u++) {
assert(WEIGHT(optPtr->litFreq[literals[u]], optLevel) <= optPtr->litSumBasePrice); /* literal cost should never be negative */ assert(WEIGHT(optPtr->litFreq[literals[u]], optLevel) <= optPtr->litSumBasePrice); /* literal cost should never be negative */
} }
/* ZSTD_litLengthContribution() : /* ZSTD_litLengthContribution() :
* @return ( cost(litlength) - cost(0) ) * @return ( cost(litlength) - cost(0) )
* this value can then be added to rawLiteralsCost() * this value can then be added to rawLiteralsCost()
* to provide a cost which is directly comparable to a match ending at same position */ * 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) static int ZSTD_litLengthContribution(U32 const litLength, const optState_t* const optPtr, int optLevel)
{ {
if (optPtr->priceType >= zop_predef) return WEIGHT(litLength, optLevel); if (optPtr->priceType >= zop_predef) return (int)WEIGHT(litLength, optLevel);
/* dynamic statistics */ /* dynamic statistics */
{ U32 const llCode = ZSTD_LLcode(litLength); { U32 const llCode = ZSTD_LLcode(litLength);
int const contribution = (LL_bits[llCode] * BITCOST_MULTIPLIER) int const contribution = (int)(LL_bits[llCode] * BITCOST_MULTIPLIER)
+ WEIGHT(optPtr->litLengthFreq[0], optLevel) /* note: log2litLengthSum cancel out */ + (int)WEIGHT(optPtr->litLengthFreq[0], optLevel) /* note: log2litLengthSum cancel out */
- WEIGHT(optPtr->litLengthFreq[llCode], optLevel); - (int)WEIGHT(optPtr->litLengthFreq[llCode], optLevel);
#if 1 #if 1
return contribution; return contribution;
#else #else
return MAX(0, contribution); /* sometimes better, sometimes not ... */ return MAX(0, contribution); /* sometimes better, sometimes not ... */
#endif #endif
} }
} }
/* ZSTD_literalsContribution() : /* ZSTD_literalsContribution() :
* creates a fake cost for the literals part of a sequence * creates a fake cost for the literals part of a sequence
* which can be compared to the ending cost of a match * which can be compared to the ending cost of a match
* should a new match start at this position */ * should a new match start at this position */
static int ZSTD_literalsContribution(const BYTE* const literals, U32 const litLength, static int ZSTD_literalsContribution(const BYTE* const literals, U32 const litLength,
const optState_t* const optPtr, const optState_t* const optPtr,
int optLevel) int optLevel)
{ {
int const contribution = ZSTD_rawLiteralsCost(literals, litLength, optPtr, optLevel) int const contribution = (int)ZSTD_rawLiteralsCost(literals, litLength, optPtr, optLevel)
+ ZSTD_litLengthContribution(litLength, optPtr, optLevel); + ZSTD_litLengthContribution(litLength, optPtr, optLevel);
return contribution; return contribution;
} }
/* ZSTD_getMatchPrice() : /* ZSTD_getMatchPrice() :
* Provides the cost of the match part (offset + matchLength) of a sequence * 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. * 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) */ * optLevel: when <2, favors small offset for decompression speed (improved cache efficiency) */
/* ZSTD_updateStats() : /* ZSTD_updateStats() :
* assumption : literals + litLengtn <= iend */ * assumption : literals + litLengtn <= iend */
static void ZSTD_updateStats(optState_t* const optPtr, static void ZSTD_updateStats(optState_t* const optPtr,
U32 litLength, const BYTE* literals, U32 litLength, const BYTE* literals,
U32 offsetCode, U32 matchLength) U32 offsetCode, U32 matchLength)
{ {
/* literals */ /* literals */
{ U32 u; if (ZSTD_compressedLiterals(optPtr)) {
U32 u;
for (u=0; u < litLength; u++) for (u=0; u < litLength; u++)
optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD; optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;
optPtr->litSum += litLength*ZSTD_LITFREQ_ADD; optPtr->litSum += litLength*ZSTD_LITFREQ_ADD;
} }
/* literal Length */ /* literal Length */
{ U32 const llCode = ZSTD_LLcode(litLength); { U32 const llCode = ZSTD_LLcode(litLength);
optPtr->litLengthFreq[llCode]++; optPtr->litLengthFreq[llCode]++;
else else
return MEM_read32(memPtr)>>8; return MEM_read32(memPtr)>>8;
} }
} }
/* Update hashTable3 up to ip (excluded) /* Update hashTable3 up to ip (excluded)
Assumption : always within prefix (i.e. not within extDict) */ Assumption : always within prefix (i.e. not within extDict) */
static U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_matchState_t* ms, const BYTE* const ip) static U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_matchState_t* ms,
U32* nextToUpdate3,
const BYTE* const ip)
{ {
U32* const hashTable3 = ms->hashTable3; U32* const hashTable3 = ms->hashTable3;
U32 const hashLog3 = ms->hashLog3; U32 const hashLog3 = ms->hashLog3;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
U32 idx = ms->nextToUpdate3; U32 idx = *nextToUpdate3;
U32 const target = ms->nextToUpdate3 = (U32)(ip - base); U32 const target = (U32)(ip - base);
size_t const hash3 = ZSTD_hash3Ptr(ip, hashLog3); size_t const hash3 = ZSTD_hash3Ptr(ip, hashLog3);
assert(hashLog3 > 0); assert(hashLog3 > 0);
while(idx < target) { while(idx < target) {
hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx; hashTable3[ZSTD_hash3Ptr(base+idx, hashLog3)] = idx;
idx++; idx++;
} }
*nextToUpdate3 = target;
return hashTable3[hash3]; return hashTable3[hash3];
} }
/*-************************************* /*-*************************************
* Binary Tree search * Binary Tree search
***************************************/ ***************************************/
/** ZSTD_insertBt1() : add one or multiple positions to tree. /** ZSTD_insertBt1() : add one or multiple positions to tree.
*largerPtr = matchIndex; *largerPtr = matchIndex;
commonLengthLarger = matchLength; commonLengthLarger = matchLength;
if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop searching */ if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop searching */
largerPtr = nextPtr; largerPtr = nextPtr;
matchIndex = nextPtr[0]; matchIndex = nextPtr[0];
} } } }
*smallerPtr = *largerPtr = 0; *smallerPtr = *largerPtr = 0;
if (bestLength > 384) return MIN(192, (U32)(bestLength - 384)); /* speed optimization */ { U32 positions = 0;
if (bestLength > 384) positions = MIN(192, (U32)(bestLength - 384)); /* speed optimization */
assert(matchEndIdx > current + 8); assert(matchEndIdx > current + 8);
return matchEndIdx - (current + 8); return MAX(positions, matchEndIdx - (current + 8));
}
} }
FORCE_INLINE_TEMPLATE FORCE_INLINE_TEMPLATE
void ZSTD_updateTree_internal( void ZSTD_updateTree_internal(
ZSTD_matchState_t* ms, ZSTD_matchState_t* ms,
const BYTE* const ip, const BYTE* const iend, const BYTE* const ip, const BYTE* const iend,
const U32 mls, const ZSTD_dictMode_e dictMode) const U32 mls, const ZSTD_dictMode_e dictMode)
{ {
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
U32 const target = (U32)(ip - base); U32 const target = (U32)(ip - base);
U32 idx = ms->nextToUpdate; U32 idx = ms->nextToUpdate;
DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)", DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)",
idx, target, dictMode); idx, target, dictMode);
while(idx < target) while(idx < target) {
idx += ZSTD_insertBt1(ms, base+idx, iend, mls, dictMode == ZSTD_extDict); U32 const forward = ZSTD_insertBt1(ms, base+idx, iend, mls, dictMode == ZSTD_extDict);
assert(idx < (U32)(idx + forward));
idx += forward;
}
assert((size_t)(ip - base) <= (size_t)(U32)(-1));
assert((size_t)(iend - base) <= (size_t)(U32)(-1));
ms->nextToUpdate = target; ms->nextToUpdate = target;
} }
void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) { void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) {
ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict); ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict);
} }
FORCE_INLINE_TEMPLATE FORCE_INLINE_TEMPLATE
U32 ZSTD_insertBtAndGetAllMatches ( U32 ZSTD_insertBtAndGetAllMatches (
ZSTD_match_t* matches, /* store result (found matches) in this table (presumed large enough) */
ZSTD_matchState_t* ms, ZSTD_matchState_t* ms,
U32* nextToUpdate3,
const BYTE* const ip, const BYTE* const iLimit, const ZSTD_dictMode_e dictMode, const BYTE* const ip, const BYTE* const iLimit, const ZSTD_dictMode_e dictMode,
U32 rep[ZSTD_REP_NUM], const U32 rep[ZSTD_REP_NUM],
U32 const ll0, /* tells if associated literal length is 0 or not. This value must be 0 or 1 */ 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, const U32 lengthToBeat,
U32 const mls /* template */) U32 const mls /* template */)
{ {
const ZSTD_compressionParameters* const cParams = &ms->cParams; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1); U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
U32 const current = (U32)(ip-base); U32 const current = (U32)(ip-base);
U32 const hashLog = cParams->hashLog; U32 const hashLog = cParams->hashLog;
U32 const minMatch = (mls==3) ? 3 : 4; U32 const minMatch = (mls==3) ? 3 : 4;
U32* const hashTable = ms->hashTable; U32* const hashTable = ms->hashTable;
size_t const h = ZSTD_hashPtr(ip, hashLog, mls); size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
U32 matchIndex = hashTable[h]; U32 matchIndex = hashTable[h];
U32* const bt = ms->chainTable; U32* const bt = ms->chainTable;
U32 const btLog = cParams->chainLog - 1; U32 const btLog = cParams->chainLog - 1;
U32 const btMask= (1U << btLog) - 1; U32 const btMask= (1U << btLog) - 1;
size_t commonLengthSmaller=0, commonLengthLarger=0; size_t commonLengthSmaller=0, commonLengthLarger=0;
const BYTE* const dictBase = ms->window.dictBase; const BYTE* const dictBase = ms->window.dictBase;
U32 const dictLimit = ms->window.dictLimit; U32 const dictLimit = ms->window.dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit; const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const prefixStart = base + dictLimit; const BYTE* const prefixStart = base + dictLimit;
U32 const btLow = btMask >= current ? 0 : current - btMask; U32 const btLow = (btMask >= current) ? 0 : current - btMask;
U32 const windowLow = ms->window.lowLimit; U32 const windowLow = ZSTD_getLowestMatchIndex(ms, current, cParams->windowLog);
U32 const matchLow = windowLow ? windowLow : 1; U32 const matchLow = windowLow ? windowLow : 1;
U32* smallerPtr = bt + 2*(current&btMask); U32* smallerPtr = bt + 2*(current&btMask);
U32* largerPtr = bt + 2*(current&btMask) + 1; U32* largerPtr = bt + 2*(current&btMask) + 1;
U32 matchEndIdx = current+8+1; /* farthest referenced position of any match => detects repetitive patterns */ U32 matchEndIdx = current+8+1; /* farthest referenced position of any match => detects repetitive patterns */
U32 dummy32; /* to be nullified at the end */ U32 dummy32; /* to be nullified at the end */
U32 mnum = 0; U32 mnum = 0;
U32 nbCompares = 1U << cParams->searchLog; U32 nbCompares = 1U << cParams->searchLog;
mnum++; mnum++;
if ( (repLen > sufficient_len) if ( (repLen > sufficient_len)
| (ip+repLen == iLimit) ) { /* best possible */ | (ip+repLen == iLimit) ) { /* best possible */
return mnum; return mnum;
} } } } } } } }
/* HC3 match finder */ /* HC3 match finder */
if ((mls == 3) /*static*/ && (bestLength < mls)) { if ((mls == 3) /*static*/ && (bestLength < mls)) {
U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(ms, ip); U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(ms, nextToUpdate3, ip);
if ((matchIndex3 >= matchLow) if ((matchIndex3 >= matchLow)
& (current - matchIndex3 < (1<<18)) /*heuristic : longer distance likely too expensive*/ ) { & (current - matchIndex3 < (1<<18)) /*heuristic : longer distance likely too expensive*/ ) {
size_t mlen; size_t mlen;
if ((dictMode == ZSTD_noDict) /*static*/ || (dictMode == ZSTD_dictMatchState) /*static*/ || (matchIndex3 >= dictLimit)) { if ((dictMode == ZSTD_noDict) /*static*/ || (dictMode == ZSTD_dictMatchState) /*static*/ || (matchIndex3 >= dictLimit)) {
const BYTE* const match = base + matchIndex3; const BYTE* const match = base + matchIndex3;
mlen = ZSTD_count(ip, match, iLimit); mlen = ZSTD_count(ip, match, iLimit);
} else { } else {
const BYTE* const match = dictBase + matchIndex3; const BYTE* const match = dictBase + matchIndex3;
assert(mnum==0); /* no prior solution */ assert(mnum==0); /* no prior solution */
matches[0].off = (current - matchIndex3) + ZSTD_REP_MOVE; matches[0].off = (current - matchIndex3) + ZSTD_REP_MOVE;
matches[0].len = (U32)mlen; matches[0].len = (U32)mlen;
mnum = 1; mnum = 1;
if ( (mlen > sufficient_len) | if ( (mlen > sufficient_len) |
(ip+mlen == iLimit) ) { /* best possible length */ (ip+mlen == iLimit) ) { /* best possible length */
ms->nextToUpdate = current+1; /* skip insertion */ ms->nextToUpdate = current+1; /* skip insertion */
return 1; return 1;
} } } }
}
}
/* no dictMatchState lookup: dicts don't have a populated HC3 table */ /* no dictMatchState lookup: dicts don't have a populated HC3 table */
} }
hashTable[h] = current; /* Update Hash Table */ hashTable[h] = current; /* Update Hash Table */
while (nbCompares-- && (matchIndex >= matchLow)) { while (nbCompares-- && (matchIndex >= matchLow)) {
U32* const nextPtr = bt + 2*(matchIndex & btMask); U32* const nextPtr = bt + 2*(matchIndex & btMask);
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
const BYTE* match; const BYTE* match;
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
assert(current > matchIndex); assert(current > matchIndex);
if ((dictMode == ZSTD_noDict) || (dictMode == ZSTD_dictMatchState) || (matchIndex+matchLength >= dictLimit)) { if ((dictMode == ZSTD_noDict) || (dictMode == ZSTD_dictMatchState) || (matchIndex+matchLength >= dictLimit)) {
assert(matchIndex+matchLength >= dictLimit); /* ensure the condition is correct when !extDict */ assert(matchIndex+matchLength >= dictLimit); /* ensure the condition is correct when !extDict */
match = base + matchIndex; match = base + matchIndex;
if (matchIndex >= dictLimit) assert(memcmp(match, ip, matchLength) == 0); /* ensure early section of match is equal as expected */
matchLength += ZSTD_count(ip+matchLength, match+matchLength, iLimit); matchLength += ZSTD_count(ip+matchLength, match+matchLength, iLimit);
} else { } else {
match = dictBase + matchIndex; match = dictBase + matchIndex;
assert(memcmp(match, ip, matchLength) == 0); /* ensure early section of match is equal as expected */
matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart); matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);
if (matchIndex+matchLength >= dictLimit) if (matchIndex+matchLength >= dictLimit)
match = base + matchIndex; /* prepare for match[matchLength] */ match = base + matchIndex; /* prepare for match[matchLength] read */
} }
if (matchLength > bestLength) { if (matchLength > bestLength) {
DEBUGLOG(8, "found match of length %u at distance %u (offCode=%u)", DEBUGLOG(8, "found match of length %u at distance %u (offCode=%u)",
(U32)matchLength, current - matchIndex, current - matchIndex + ZSTD_REP_MOVE); (U32)matchLength, current - matchIndex, current - matchIndex + ZSTD_REP_MOVE);
assert(matchEndIdx > matchIndex); assert(matchEndIdx > matchIndex);
if (matchLength > matchEndIdx - matchIndex) if (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength; matchEndIdx = matchIndex + (U32)matchLength;
assert(matchEndIdx > current+8); assert(matchEndIdx > current+8);
ms->nextToUpdate = matchEndIdx - 8; /* skip repetitive patterns */ ms->nextToUpdate = matchEndIdx - 8; /* skip repetitive patterns */
return mnum; return mnum;
} }
FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllMatches ( FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllMatches (
ZSTD_match_t* matches, /* store result (match found, increasing size) in this table */
ZSTD_matchState_t* ms, ZSTD_matchState_t* ms,
U32* nextToUpdate3,
const BYTE* ip, const BYTE* const iHighLimit, const ZSTD_dictMode_e dictMode, const BYTE* ip, const BYTE* const iHighLimit, const ZSTD_dictMode_e dictMode,
U32 rep[ZSTD_REP_NUM], U32 const ll0, const U32 rep[ZSTD_REP_NUM],
ZSTD_match_t* matches, U32 const lengthToBeat) U32 const ll0,
U32 const lengthToBeat)
{ {
const ZSTD_compressionParameters* const cParams = &ms->cParams; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32 const matchLengthSearch = cParams->minMatch; U32 const matchLengthSearch = cParams->minMatch;
DEBUGLOG(8, "ZSTD_BtGetAllMatches"); DEBUGLOG(8, "ZSTD_BtGetAllMatches");
if (ip < ms->window.base + ms->nextToUpdate) return 0; /* skipped area */ if (ip < ms->window.base + ms->nextToUpdate) return 0; /* skipped area */
ZSTD_updateTree_internal(ms, ip, iHighLimit, matchLengthSearch, dictMode); ZSTD_updateTree_internal(ms, ip, iHighLimit, matchLengthSearch, dictMode);
switch(matchLengthSearch) switch(matchLengthSearch)
{ {
case 3 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 3); case 3 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 3);
default : default :
case 4 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 4); case 4 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 4);
case 5 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 5); case 5 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 5);
case 7 : case 7 :
case 6 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 6); case 6 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 6);
} }
} }
/*-******************************* /*-*******************************
* Optimal parser * Optimal parser
*********************************/ *********************************/
typedef struct repcodes_s { typedef struct repcodes_s {
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - 8; const BYTE* const ilimit = iend - 8;
const BYTE* const base = ms->window.base; const BYTE* const base = ms->window.base;
const BYTE* const prefixStart = base + ms->window.dictLimit; const BYTE* const prefixStart = base + ms->window.dictLimit;
const ZSTD_compressionParameters* const cParams = &ms->cParams; const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1); U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);
U32 const minMatch = (cParams->minMatch == 3) ? 3 : 4; U32 const minMatch = (cParams->minMatch == 3) ? 3 : 4;
U32 nextToUpdate3 = ms->nextToUpdate;
ZSTD_optimal_t* const opt = optStatePtr->priceTable; ZSTD_optimal_t* const opt = optStatePtr->priceTable;
ZSTD_match_t* const matches = optStatePtr->matchTable; ZSTD_match_t* const matches = optStatePtr->matchTable;
ZSTD_optimal_t lastSequence; ZSTD_optimal_t lastSequence;
/* init */ /* init */
DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u", DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u",
(U32)(ip - base), ms->window.dictLimit, ms->nextToUpdate); (U32)(ip - base), ms->window.dictLimit, ms->nextToUpdate);
assert(optLevel <= 2); assert(optLevel <= 2);
ms->nextToUpdate3 = ms->nextToUpdate;
ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize, optLevel); ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize, optLevel);
ip += (ip==prefixStart); ip += (ip==prefixStart);
/* Match Loop */ /* Match Loop */
while (ip < ilimit) { while (ip < ilimit) {
U32 cur, last_pos = 0; U32 cur, last_pos = 0;
/* find first match */ /* find first match */
{ U32 const litlen = (U32)(ip - anchor); { U32 const litlen = (U32)(ip - anchor);
U32 const ll0 = !litlen; U32 const ll0 = !litlen;
U32 const nbMatches = ZSTD_BtGetAllMatches(ms, ip, iend, dictMode, rep, ll0, matches, minMatch); U32 const nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, ip, iend, dictMode, rep, ll0, minMatch);
if (!nbMatches) { ip++; continue; } if (!nbMatches) { ip++; continue; }
/* initialize opt[0] */ /* initialize opt[0] */
{ U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; } { 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].mlen = 0; /* means is_a_literal */
opt[0].litlen = litlen; opt[0].litlen = litlen;
opt[0].price = ZSTD_literalsContribution(anchor, litlen, optStatePtr, optLevel); opt[0].price = ZSTD_literalsContribution(anchor, litlen, optStatePtr, optLevel);
/* large match -> immediate encoding */ /* large match -> immediate encoding */
{ U32 const maxML = matches[nbMatches-1].len; { U32 const maxML = matches[nbMatches-1].len;
U32 const maxOffset = matches[nbMatches-1].off; U32 const maxOffset = matches[nbMatches-1].off;
DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new serie", DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new series",
nbMatches, maxML, maxOffset, (U32)(ip-prefixStart)); nbMatches, maxML, maxOffset, (U32)(ip-prefixStart));
if (maxML > sufficient_len) { if (maxML > sufficient_len) {
lastSequence.litlen = litlen; lastSequence.litlen = litlen;
lastSequence.mlen = maxML; lastSequence.mlen = maxML;
lastSequence.off = maxOffset; lastSequence.off = maxOffset;
DEBUGLOG(6, "large match (%u>%u), immediate encoding", DEBUGLOG(6, "large match (%u>%u), immediate encoding",
maxML, sufficient_len); maxML, sufficient_len);
DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1); DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1);
continue; /* skip unpromising positions; about ~+6% speed, -0.01 ratio */ continue; /* skip unpromising positions; about ~+6% speed, -0.01 ratio */
} }
{ U32 const ll0 = (opt[cur].mlen != 0); { U32 const ll0 = (opt[cur].mlen != 0);
U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0; U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0;
U32 const previousPrice = opt[cur].price; U32 const previousPrice = opt[cur].price;
U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel); 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 const nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, inr, iend, dictMode, opt[cur].rep, ll0, minMatch);
U32 matchNb; U32 matchNb;
if (!nbMatches) { if (!nbMatches) {
DEBUGLOG(7, "rPos:%u : no match found", cur); DEBUGLOG(7, "rPos:%u : no match found", cur);
continue; continue;
} }
{ U32 const maxML = matches[nbMatches-1].len; { U32 const maxML = matches[nbMatches-1].len;
DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u", DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u",
ip = anchor; ip = anchor;
} } } }
ZSTD_setBasePrices(optStatePtr, optLevel); ZSTD_setBasePrices(optStatePtr, optLevel);
} }
} /* while (ip < ilimit) */ } /* while (ip < ilimit) */
/* Return the last literals size */ /* Return the last literals size */
return iend - anchor; return (size_t)(iend - anchor);
} }
size_t ZSTD_compressBlock_btopt( size_t ZSTD_compressBlock_btopt(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
DEBUGLOG(5, "ZSTD_compressBlock_btopt"); DEBUGLOG(5, "ZSTD_compressBlock_btopt");
sum += table[s]; sum += table[s];
} }
return sum; return sum;
} }
/* used in 2-pass strategy */ /* used in 2-pass strategy */
MEM_STATIC void ZSTD_upscaleStats(optState_t* optPtr) MEM_STATIC void ZSTD_upscaleStats(optState_t* optPtr)
{ {
if (ZSTD_compressedLiterals(optPtr))
optPtr->litSum = ZSTD_upscaleStat(optPtr->litFreq, MaxLit, 0); optPtr->litSum = ZSTD_upscaleStat(optPtr->litFreq, MaxLit, 0);
optPtr->litLengthSum = ZSTD_upscaleStat(optPtr->litLengthFreq, MaxLL, 0); optPtr->litLengthSum = ZSTD_upscaleStat(optPtr->litLengthFreq, MaxLL, 0);
optPtr->matchLengthSum = ZSTD_upscaleStat(optPtr->matchLengthFreq, MaxML, 0); optPtr->matchLengthSum = ZSTD_upscaleStat(optPtr->matchLengthFreq, MaxML, 0);
optPtr->offCodeSum = ZSTD_upscaleStat(optPtr->offCodeFreq, MaxOff, 0); optPtr->offCodeSum = ZSTD_upscaleStat(optPtr->offCodeFreq, MaxOff, 0);
} }
/* ZSTD_initStats_ultra(): /* ZSTD_initStats_ultra():
* make a first compression pass, just to seed stats with more accurate starting values. * 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. * only works on first block, with no dictionary and no ldm.
* this function cannot error, hence its constract must be respected. * this function cannot error, hence its contract must be respected.
*/ */
static void static void
ZSTD_initStats_ultra(ZSTD_matchState_t* ms, ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
seqStore_t* seqStore, seqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM], U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
U32 tmpRep[ZSTD_REP_NUM]; /* updated rep codes will sink here */ U32 tmpRep[ZSTD_REP_NUM]; /* updated rep codes will sink here */
memcpy(tmpRep, rep, sizeof(tmpRep)); memcpy(tmpRep, rep, sizeof(tmpRep));
DEBUGLOG(4, "ZSTD_initStats_ultra (srcSize=%zu)", srcSize); DEBUGLOG(4, "ZSTD_initStats_ultra (srcSize=%zu)", srcSize);
assert(ms->opt.litLengthSum == 0); /* first block */ assert(ms->opt.litLengthSum == 0); /* first block */
assert(seqStore->sequences == seqStore->sequencesStart); /* no ldm */ assert(seqStore->sequences == seqStore->sequencesStart); /* no ldm */
assert(ms->window.dictLimit == ms->window.lowLimit); /* no dictionary */ 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) */ 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*/ ZSTD_compressBlock_opt_generic(ms, seqStore, tmpRep, src, srcSize, 2 /*optLevel*/, ZSTD_noDict); /* generate stats into ms->opt*/
/* invalidate first scan from history */ /* invalidate first scan from history */
ZSTD_resetSeqStore(seqStore); ZSTD_resetSeqStore(seqStore);
ms->window.base -= srcSize; ms->window.base -= srcSize;
ms->window.dictLimit += (U32)srcSize; ms->window.dictLimit += (U32)srcSize;
ms->window.lowLimit = ms->window.dictLimit; ms->window.lowLimit = ms->window.dictLimit;
ms->nextToUpdate = ms->window.dictLimit; ms->nextToUpdate = ms->window.dictLimit;
ms->nextToUpdate3 = ms->window.dictLimit;
/* re-inforce weight of collected statistics */ /* re-inforce weight of collected statistics */
ZSTD_upscaleStats(&ms->opt); ZSTD_upscaleStats(&ms->opt);
} }
size_t ZSTD_compressBlock_btultra( size_t ZSTD_compressBlock_btultra(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize) const void* src, size_t srcSize)

contrib/python-zstandard/zstd/compress/zstdmt_compress.h

#define ZSTDMT_COMPRESS_H #define ZSTDMT_COMPRESS_H
#if defined (__cplusplus) #if defined (__cplusplus)
extern "C" { extern "C" {
#endif #endif
/* Note : This is an internal API. /* Note : This is an internal API.
* Some methods are still exposed (ZSTDLIB_API), * These APIs used to be exposed with ZSTDLIB_API,
* because it used to be the only way to invoke MT compression. * because it used to be the only way to invoke MT compression.
* Now, it's recommended to use ZSTD_compress_generic() instead. * Now, it's recommended to use ZSTD_compress2 and ZSTD_compressStream2()
* These methods will stop being exposed in a future version */ * 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 === */ /* === Dependencies === */
#include <stddef.h> /* size_t */ #include <stddef.h> /* size_t */
#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_parameters */ #define ZSTD_STATIC_LINKING_ONLY /* ZSTD_parameters */
#include "zstd.h" /* ZSTD_inBuffer, ZSTD_outBuffer, ZSTDLIB_API */ #include "zstd.h" /* ZSTD_inBuffer, ZSTD_outBuffer, ZSTDLIB_API */
/* === Constants === */ /* === Constants === */
#ifndef ZSTDMT_NBWORKERS_MAX #ifndef ZSTDMT_NBWORKERS_MAX
# define ZSTDMT_NBWORKERS_MAX 200 # define ZSTDMT_NBWORKERS_MAX 200
#endif #endif
#ifndef ZSTDMT_JOBSIZE_MIN #ifndef ZSTDMT_JOBSIZE_MIN
# define ZSTDMT_JOBSIZE_MIN (1 MB) # define ZSTDMT_JOBSIZE_MIN (1 MB)
#endif #endif
#define ZSTDMT_JOBLOG_MAX (MEM_32bits() ? 29 : 30)
#define ZSTDMT_JOBSIZE_MAX (MEM_32bits() ? (512 MB) : (1024 MB)) #define ZSTDMT_JOBSIZE_MAX (MEM_32bits() ? (512 MB) : (1024 MB))
/* === Memory management === */ /* === Memory management === */
typedef struct ZSTDMT_CCtx_s ZSTDMT_CCtx; typedef struct ZSTDMT_CCtx_s ZSTDMT_CCtx;
ZSTDLIB_API ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers); /* Requires ZSTD_MULTITHREAD to be defined during compilation, otherwise it will return NULL. */
ZSTDLIB_API ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers, 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); ZSTD_customMem cMem);
ZSTDLIB_API size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx); ZSTDMT_API size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx);
ZSTDLIB_API size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx); ZSTDMT_API size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx);
/* === Simple one-pass compression function === */ /* === Simple one-pass compression function === */
ZSTDLIB_API size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx, ZSTDMT_API size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
int compressionLevel); int compressionLevel);
/* === Streaming functions === */ /* === Streaming functions === */
ZSTDLIB_API size_t ZSTDMT_initCStream(ZSTDMT_CCtx* mtctx, int compressionLevel); ZSTDMT_API size_t ZSTDMT_initCStream(ZSTDMT_CCtx* mtctx, int compressionLevel);
ZSTDLIB_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_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" */
ZSTDLIB_API size_t ZSTDMT_nextInputSizeHint(const ZSTDMT_CCtx* mtctx); ZSTDMT_API size_t ZSTDMT_nextInputSizeHint(const ZSTDMT_CCtx* mtctx);
ZSTDLIB_API size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input); ZSTDMT_API size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
ZSTDLIB_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_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output); /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */
ZSTDLIB_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()) */ 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 === */ /* === Advanced functions and parameters === */
ZSTDLIB_API size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx, ZSTDMT_API size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const ZSTD_CDict* cdict, const ZSTD_CDict* cdict,
ZSTD_parameters params, ZSTD_parameters params,
int overlapLog); int overlapLog);
ZSTDLIB_API size_t ZSTDMT_initCStream_advanced(ZSTDMT_CCtx* mtctx, 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 */ const void* dict, size_t dictSize, /* dict can be released after init, a local copy is preserved within zcs */
ZSTD_parameters params, ZSTD_parameters params,
unsigned long long pledgedSrcSize); /* pledgedSrcSize is optional and can be zero == unknown */ unsigned long long pledgedSrcSize); /* pledgedSrcSize is optional and can be zero == unknown */
ZSTDLIB_API size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx, ZSTDMT_API size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
const ZSTD_CDict* cdict, const ZSTD_CDict* cdict,
ZSTD_frameParameters fparams, ZSTD_frameParameters fparams,
unsigned long long pledgedSrcSize); /* note : zero means empty */ unsigned long long pledgedSrcSize); /* note : zero means empty */
/* ZSTDMT_parameter : /* ZSTDMT_parameter :
* List of parameters that can be set using ZSTDMT_setMTCtxParameter() */ * List of parameters that can be set using ZSTDMT_setMTCtxParameter() */
typedef enum { 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_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 compressionr 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_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_p_rsyncable /* Enables rsyncable mode. */
} ZSTDMT_parameter; } ZSTDMT_parameter;
/* ZSTDMT_setMTCtxParameter() : /* ZSTDMT_setMTCtxParameter() :
* allow setting individual parameters, one at a time, among a list of enums defined in ZSTDMT_parameter. * 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*() !__ * 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. * 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()) */ * @return : 0, or an error code (which can be tested using ZSTD_isError()) */
ZSTDLIB_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value); ZSTDMT_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value);
/* ZSTDMT_getMTCtxParameter() : /* ZSTDMT_getMTCtxParameter() :
* Query the ZSTDMT_CCtx for a parameter value. * Query the ZSTDMT_CCtx for a parameter value.
* @return : 0, or an error code (which can be tested using ZSTD_isError()) */ * @return : 0, or an error code (which can be tested using ZSTD_isError()) */
ZSTDLIB_API size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value); ZSTDMT_API size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value);
/*! ZSTDMT_compressStream_generic() : /*! ZSTDMT_compressStream_generic() :
* Combines ZSTDMT_compressStream() with optional ZSTDMT_flushStream() or ZSTDMT_endStream() * Combines ZSTDMT_compressStream() with optional ZSTDMT_flushStream() or ZSTDMT_endStream()
* depending on flush directive. * depending on flush directive.
* @return : minimum amount of data still to be flushed * @return : minimum amount of data still to be flushed
* 0 if fully flushed * 0 if fully flushed
* or an error code * or an error code
* note : needs to be init using any ZSTD_initCStream*() variant */ * note : needs to be init using any ZSTD_initCStream*() variant */
ZSTDLIB_API size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx, ZSTDMT_API size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
ZSTD_outBuffer* output, ZSTD_outBuffer* output,
ZSTD_inBuffer* input, ZSTD_inBuffer* input,
ZSTD_EndDirective endOp); ZSTD_EndDirective endOp);
/* ======================================================== /* ========================================================
* === Private interface, for use by ZSTD_compress.c === * === Private interface, for use by ZSTD_compress.c ===
* === Not exposed in libzstd. Never invoke directly === * === Not exposed in libzstd. Never invoke directly ===

contrib/python-zstandard/zstd/compress/zstdmt_compress.c

/* ====== Constants ====== */ /* ====== Constants ====== */
#define ZSTDMT_OVERLAPLOG_DEFAULT 0 #define ZSTDMT_OVERLAPLOG_DEFAULT 0
/* ====== Dependencies ====== */ /* ====== Dependencies ====== */
#include <string.h> /* memcpy, memset */ #include <string.h> /* memcpy, memset */
#include <limits.h> /* INT_MAX, UINT_MAX */ #include <limits.h> /* INT_MAX, UINT_MAX */
#include "mem.h" /* MEM_STATIC */
#include "pool.h" /* threadpool */ #include "pool.h" /* threadpool */
#include "threading.h" /* mutex */ #include "threading.h" /* mutex */
#include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */ #include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */
#include "zstd_ldm.h" #include "zstd_ldm.h"
#include "zstdmt_compress.h" #include "zstdmt_compress.h"
/* Guards code to support resizing the SeqPool. /* Guards code to support resizing the SeqPool.
* We will want to resize the SeqPool to save memory in the future. * We will want to resize the SeqPool to save memory in the future.
ZSTD_CCtx_params params; ZSTD_CCtx_params params;
ldmState_t ldmState; ldmState_t ldmState;
XXH64_state_t xxhState; XXH64_state_t xxhState;
unsigned nextJobID; unsigned nextJobID;
/* Protects ldmWindow. /* Protects ldmWindow.
* Must be acquired after the main mutex when acquiring both. * Must be acquired after the main mutex when acquiring both.
*/ */
ZSTD_pthread_mutex_t ldmWindowMutex; ZSTD_pthread_mutex_t ldmWindowMutex;
ZSTD_pthread_cond_t ldmWindowCond; /* Signaled when ldmWindow is udpated */ ZSTD_pthread_cond_t ldmWindowCond; /* Signaled when ldmWindow is updated */
ZSTD_window_t ldmWindow; /* A thread-safe copy of ldmState.window */ ZSTD_window_t ldmWindow; /* A thread-safe copy of ldmState.window */
} serialState_t; } serialState_t;
static int ZSTDMT_serialState_reset(serialState_t* serialState, ZSTDMT_seqPool* seqPool, ZSTD_CCtx_params params, size_t jobSize) static int ZSTDMT_serialState_reset(serialState_t* serialState, ZSTDMT_seqPool* seqPool, ZSTD_CCtx_params params, size_t jobSize)
{ {
/* Adjust parameters */ /* Adjust parameters */
if (params.ldmParams.enableLdm) { if (params.ldmParams.enableLdm) {
DEBUGLOG(4, "LDM window size = %u KB", (1U << params.cParams.windowLog) >> 10); DEBUGLOG(4, "LDM window size = %u KB", (1U << params.cParams.windowLog) >> 10);
{ {
ZSTDMT_jobDescription* const job = (ZSTDMT_jobDescription*)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_params jobParams = job->params; /* do not modify job->params ! copy it, modify the copy */
ZSTD_CCtx* const cctx = ZSTDMT_getCCtx(job->cctxPool); ZSTD_CCtx* const cctx = ZSTDMT_getCCtx(job->cctxPool);
rawSeqStore_t rawSeqStore = ZSTDMT_getSeq(job->seqPool); rawSeqStore_t rawSeqStore = ZSTDMT_getSeq(job->seqPool);
buffer_t dstBuff = job->dstBuff; buffer_t dstBuff = job->dstBuff;
size_t lastCBlockSize = 0; size_t lastCBlockSize = 0;
/* ressources */ /* resources */
if (cctx==NULL) JOB_ERROR(ERROR(memory_allocation)); if (cctx==NULL) JOB_ERROR(ERROR(memory_allocation));
if (dstBuff.start == NULL) { /* streaming job : doesn't provide a dstBuffer */ if (dstBuff.start == NULL) { /* streaming job : doesn't provide a dstBuffer */
dstBuff = ZSTDMT_getBuffer(job->bufPool); dstBuff = ZSTDMT_getBuffer(job->bufPool);
if (dstBuff.start==NULL) JOB_ERROR(ERROR(memory_allocation)); 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 */ job->dstBuff = dstBuff; /* this value can be read in ZSTDMT_flush, when it copies the whole job */
} }
if (jobParams.ldmParams.enableLdm && rawSeqStore.seq == NULL) if (jobParams.ldmParams.enableLdm && rawSeqStore.seq == NULL)
JOB_ERROR(ERROR(memory_allocation)); JOB_ERROR(ERROR(memory_allocation));
/* Don't compute the checksum for chunks, since we compute it externally, /* Don't compute the checksum for chunks, since we compute it externally,
* but write it in the header. * but write it in the header.
*/ */
if (job->jobID != 0) jobParams.fParams.checksumFlag = 0; if (job->jobID != 0) jobParams.fParams.checksumFlag = 0;
/* Don't run LDM for the chunks, since we handle it externally */ /* Don't run LDM for the chunks, since we handle it externally */
jobParams.ldmParams.enableLdm = 0; jobParams.ldmParams.enableLdm = 0;
/* init */ /* init */
if (job->cdict) { 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); 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 */ assert(job->firstJob); /* only allowed for first job */
if (ZSTD_isError(initError)) JOB_ERROR(initError); if (ZSTD_isError(initError)) JOB_ERROR(initError);
} else { /* srcStart points at reloaded section */ } else { /* srcStart points at reloaded section */
U64 const pledgedSrcSize = job->firstJob ? job->fullFrameSize : job->src.size; U64 const pledgedSrcSize = job->firstJob ? job->fullFrameSize : job->src.size;
{ size_t const forceWindowError = ZSTD_CCtxParam_setParameter(&jobParams, ZSTD_c_forceMaxWindow, !job->firstJob); { size_t const forceWindowError = ZSTD_CCtxParams_setParameter(&jobParams, ZSTD_c_forceMaxWindow, !job->firstJob);
if (ZSTD_isError(forceWindowError)) JOB_ERROR(forceWindowError); if (ZSTD_isError(forceWindowError)) JOB_ERROR(forceWindowError);
} }
{ size_t const initError = ZSTD_compressBegin_advanced_internal(cctx, { 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) */ job->prefix.start, job->prefix.size, ZSTD_dct_rawContent, /* load dictionary in "content-only" mode (no header analysis) */
ZSTD_dtlm_fast, ZSTD_dtlm_fast,
NULL, /*cdict*/ NULL, /*cdict*/
jobParams, pledgedSrcSize); jobParams, pledgedSrcSize);
if (ZSTD_isError(initError)) JOB_ERROR(initError); if (ZSTD_isError(initError)) JOB_ERROR(initError);
return 0; return 0;
} }
/* ZSTDMT_CCtxParam_setNbWorkers(): /* ZSTDMT_CCtxParam_setNbWorkers():
* Internal use only */ * Internal use only */
size_t ZSTDMT_CCtxParam_setNbWorkers(ZSTD_CCtx_params* params, unsigned nbWorkers) size_t ZSTDMT_CCtxParam_setNbWorkers(ZSTD_CCtx_params* params, unsigned nbWorkers)
{ {
if (nbWorkers > ZSTDMT_NBWORKERS_MAX) nbWorkers = ZSTDMT_NBWORKERS_MAX; return ZSTD_CCtxParams_setParameter(params, ZSTD_c_nbWorkers, (int)nbWorkers);
params->nbWorkers = nbWorkers;
params->overlapLog = ZSTDMT_OVERLAPLOG_DEFAULT;
params->jobSize = 0;
return nbWorkers;
} }
ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers, ZSTD_customMem cMem) MEM_STATIC ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced_internal(unsigned nbWorkers, ZSTD_customMem cMem)
{ {
ZSTDMT_CCtx* mtctx; ZSTDMT_CCtx* mtctx;
U32 nbJobs = nbWorkers + 2; U32 nbJobs = nbWorkers + 2;
int initError; int initError;
DEBUGLOG(3, "ZSTDMT_createCCtx_advanced (nbWorkers = %u)", nbWorkers); DEBUGLOG(3, "ZSTDMT_createCCtx_advanced (nbWorkers = %u)", nbWorkers);
if (nbWorkers < 1) return NULL; if (nbWorkers < 1) return NULL;
nbWorkers = MIN(nbWorkers , ZSTDMT_NBWORKERS_MAX); nbWorkers = MIN(nbWorkers , ZSTDMT_NBWORKERS_MAX);
if (!mtctx->factory | !mtctx->jobs | !mtctx->bufPool | !mtctx->cctxPool | !mtctx->seqPool | initError) { if (!mtctx->factory | !mtctx->jobs | !mtctx->bufPool | !mtctx->cctxPool | !mtctx->seqPool | initError) {
ZSTDMT_freeCCtx(mtctx); ZSTDMT_freeCCtx(mtctx);
return NULL; return NULL;
} }
DEBUGLOG(3, "mt_cctx created, for %u threads", nbWorkers); DEBUGLOG(3, "mt_cctx created, for %u threads", nbWorkers);
return mtctx; 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) ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers)
{ {
return ZSTDMT_createCCtx_advanced(nbWorkers, ZSTD_defaultCMem); return ZSTDMT_createCCtx_advanced(nbWorkers, ZSTD_defaultCMem);
} }
/* ZSTDMT_releaseAllJobResources() : /* ZSTDMT_releaseAllJobResources() :
* note : ensure all workers are killed first ! */ * note : ensure all workers are killed first ! */
ZSTDMT_parameter parameter, ZSTDMT_parameter parameter,
int value) int value)
{ {
DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter"); DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter");
switch(parameter) switch(parameter)
{ {
case ZSTDMT_p_jobSize : case ZSTDMT_p_jobSize :
DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter : set jobSize to %i", value); DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter : set jobSize to %i", value);
if ( value != 0 /* default */ return ZSTD_CCtxParams_setParameter(params, ZSTD_c_jobSize, value);
&& value < ZSTDMT_JOBSIZE_MIN)
value = ZSTDMT_JOBSIZE_MIN;
assert(value >= 0);
if (value > ZSTDMT_JOBSIZE_MAX) value = ZSTDMT_JOBSIZE_MAX;
params->jobSize = value;
return value;
case ZSTDMT_p_overlapLog : case ZSTDMT_p_overlapLog :
DEBUGLOG(4, "ZSTDMT_p_overlapLog : %i", value); DEBUGLOG(4, "ZSTDMT_p_overlapLog : %i", value);
if (value < ZSTD_OVERLAPLOG_MIN) value = ZSTD_OVERLAPLOG_MIN; return ZSTD_CCtxParams_setParameter(params, ZSTD_c_overlapLog, value);
if (value > ZSTD_OVERLAPLOG_MAX) value = ZSTD_OVERLAPLOG_MAX;
params->overlapLog = value;
return value;
case ZSTDMT_p_rsyncable : case ZSTDMT_p_rsyncable :
value = (value != 0); DEBUGLOG(4, "ZSTD_p_rsyncable : %i", value);
params->rsyncable = value; return ZSTD_CCtxParams_setParameter(params, ZSTD_c_rsyncable, value);
return value;
default : default :
return ERROR(parameter_unsupported); return ERROR(parameter_unsupported);
} }
} }
size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value) size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value)
{ {
DEBUGLOG(4, "ZSTDMT_setMTCtxParameter"); DEBUGLOG(4, "ZSTDMT_setMTCtxParameter");
return ZSTDMT_CCtxParam_setMTCtxParameter(&mtctx->params, parameter, value); return ZSTDMT_CCtxParam_setMTCtxParameter(&mtctx->params, parameter, value);
} }
size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value) size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value)
{ {
switch (parameter) { switch (parameter) {
case ZSTDMT_p_jobSize: case ZSTDMT_p_jobSize:
assert(mtctx->params.jobSize <= INT_MAX); return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_jobSize, value);
*value = (int)(mtctx->params.jobSize);
break;
case ZSTDMT_p_overlapLog: case ZSTDMT_p_overlapLog:
*value = mtctx->params.overlapLog; return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_overlapLog, value);
break;
case ZSTDMT_p_rsyncable: case ZSTDMT_p_rsyncable:
*value = mtctx->params.rsyncable; return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_rsyncable, value);
break;
default: default:
return ERROR(parameter_unsupported); return ERROR(parameter_unsupported);
} }
return 0;
} }
/* Sets parameters relevant to the compression job, /* Sets parameters relevant to the compression job,
* initializing others to default values. */ * initializing others to default values. */
static ZSTD_CCtx_params ZSTDMT_initJobCCtxParams(ZSTD_CCtx_params const params) static ZSTD_CCtx_params ZSTDMT_initJobCCtxParams(ZSTD_CCtx_params const params)
{ {
ZSTD_CCtx_params jobParams; ZSTD_CCtx_params jobParams = params;
memset(&jobParams, 0, sizeof(jobParams)); /* Clear parameters related to multithreading */
jobParams.forceWindow = 0;
jobParams.cParams = params.cParams; jobParams.nbWorkers = 0;
jobParams.fParams = params.fParams; jobParams.jobSize = 0;
jobParams.compressionLevel = params.compressionLevel; jobParams.overlapLog = 0;
jobParams.rsyncable = 0;
memset(&jobParams.ldmParams, 0, sizeof(ldmParams_t));
memset(&jobParams.customMem, 0, sizeof(ZSTD_customMem));
return jobParams; return jobParams;
} }
/* ZSTDMT_resize() : /* ZSTDMT_resize() :
* @return : error code if fails, 0 on success */ * @return : error code if fails, 0 on success */
static size_t ZSTDMT_resize(ZSTDMT_CCtx* mtctx, unsigned nbWorkers) static size_t ZSTDMT_resize(ZSTDMT_CCtx* mtctx, unsigned nbWorkers)
{ {
if (POOL_resize(mtctx->factory, nbWorkers)) return ERROR(memory_allocation); if (POOL_resize(mtctx->factory, nbWorkers)) return ERROR(memory_allocation);
CHECK_F( ZSTDMT_expandJobsTable(mtctx, nbWorkers) ); FORWARD_IF_ERROR( ZSTDMT_expandJobsTable(mtctx, nbWorkers) );
mtctx->bufPool = ZSTDMT_expandBufferPool(mtctx->bufPool, nbWorkers); mtctx->bufPool = ZSTDMT_expandBufferPool(mtctx->bufPool, nbWorkers);
if (mtctx->bufPool == NULL) return ERROR(memory_allocation); if (mtctx->bufPool == NULL) return ERROR(memory_allocation);
mtctx->cctxPool = ZSTDMT_expandCCtxPool(mtctx->cctxPool, nbWorkers); mtctx->cctxPool = ZSTDMT_expandCCtxPool(mtctx->cctxPool, nbWorkers);
if (mtctx->cctxPool == NULL) return ERROR(memory_allocation); if (mtctx->cctxPool == NULL) return ERROR(memory_allocation);
mtctx->seqPool = ZSTDMT_expandSeqPool(mtctx->seqPool, nbWorkers); mtctx->seqPool = ZSTDMT_expandSeqPool(mtctx->seqPool, nbWorkers);
if (mtctx->seqPool == NULL) return ERROR(memory_allocation); if (mtctx->seqPool == NULL) return ERROR(memory_allocation);
ZSTDMT_CCtxParam_setNbWorkers(&mtctx->params, nbWorkers); ZSTDMT_CCtxParam_setNbWorkers(&mtctx->params, nbWorkers);
return 0; return 0;
/* look into oldest non-fully-flushed job */ /* look into oldest non-fully-flushed job */
{ unsigned const wJobID = jobID & mtctx->jobIDMask; { unsigned const wJobID = jobID & mtctx->jobIDMask;
ZSTDMT_jobDescription* const jobPtr = &mtctx->jobs[wJobID]; ZSTDMT_jobDescription* const jobPtr = &mtctx->jobs[wJobID];
ZSTD_pthread_mutex_lock(&jobPtr->job_mutex); ZSTD_pthread_mutex_lock(&jobPtr->job_mutex);
{ size_t const cResult = jobPtr->cSize; { size_t const cResult = jobPtr->cSize;
size_t const produced = ZSTD_isError(cResult) ? 0 : cResult; size_t const produced = ZSTD_isError(cResult) ? 0 : cResult;
size_t const flushed = ZSTD_isError(cResult) ? 0 : jobPtr->dstFlushed; size_t const flushed = ZSTD_isError(cResult) ? 0 : jobPtr->dstFlushed;
assert(flushed <= produced); assert(flushed <= produced);
assert(jobPtr->consumed <= jobPtr->src.size);
toFlush = produced - flushed; toFlush = produced - flushed;
if (toFlush==0 && (jobPtr->consumed >= jobPtr->src.size)) { /* if toFlush==0, nothing is available to flush.
/* doneJobID is not-fully-flushed, but toFlush==0 : doneJobID should be compressing some more data */ * However, jobID is expected to still be active:
* if jobID was already completed and fully flushed,
* ZSTDMT_flushProduced() should have already moved onto next job.
* Therefore, some input has not yet been consumed. */
if (toFlush==0) {
assert(jobPtr->consumed < jobPtr->src.size); assert(jobPtr->consumed < jobPtr->src.size);
} }
} }
ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex); ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex);
} }
return toFlush; return toFlush;
} }
/* ------------------------------------------ */ /* ------------------------------------------ */
/* ===== Multi-threaded compression ===== */ /* ===== Multi-threaded compression ===== */
/* ------------------------------------------ */ /* ------------------------------------------ */
static unsigned ZSTDMT_computeTargetJobLog(ZSTD_CCtx_params const params) static unsigned ZSTDMT_computeTargetJobLog(ZSTD_CCtx_params const params)
{ {
if (params.ldmParams.enableLdm) unsigned jobLog;
if (params.ldmParams.enableLdm) {
/* In Long Range Mode, the windowLog is typically oversized. /* In Long Range Mode, the windowLog is typically oversized.
* In which case, it's preferable to determine the jobSize * In which case, it's preferable to determine the jobSize
* based on chainLog instead. */ * based on chainLog instead. */
return MAX(21, params.cParams.chainLog + 4); jobLog = MAX(21, params.cParams.chainLog + 4);
return MAX(20, params.cParams.windowLog + 2); } else {
jobLog = MAX(20, params.cParams.windowLog + 2);
}
return MIN(jobLog, (unsigned)ZSTDMT_JOBLOG_MAX);
} }
static int ZSTDMT_overlapLog_default(ZSTD_strategy strat) static int ZSTDMT_overlapLog_default(ZSTD_strategy strat)
{ {
switch(strat) switch(strat)
{ {
case ZSTD_btultra2: case ZSTD_btultra2:
return 9; return 9;
if (params.ldmParams.enableLdm) { if (params.ldmParams.enableLdm) {
/* In Long Range Mode, the windowLog is typically oversized. /* In Long Range Mode, the windowLog is typically oversized.
* In which case, it's preferable to determine the jobSize * In which case, it's preferable to determine the jobSize
* based on chainLog instead. * based on chainLog instead.
* Then, ovLog becomes a fraction of the jobSize, rather than windowSize */ * Then, ovLog becomes a fraction of the jobSize, rather than windowSize */
ovLog = MIN(params.cParams.windowLog, ZSTDMT_computeTargetJobLog(params) - 2) ovLog = MIN(params.cParams.windowLog, ZSTDMT_computeTargetJobLog(params) - 2)
- overlapRLog; - overlapRLog;
} }
assert(0 <= ovLog && ovLog <= 30); assert(0 <= ovLog && ovLog <= ZSTD_WINDOWLOG_MAX);
DEBUGLOG(4, "overlapLog : %i", params.overlapLog); DEBUGLOG(4, "overlapLog : %i", params.overlapLog);
DEBUGLOG(4, "overlap size : %i", 1 << ovLog); DEBUGLOG(4, "overlap size : %i", 1 << ovLog);
return (ovLog==0) ? 0 : (size_t)1 << ovLog; return (ovLog==0) ? 0 : (size_t)1 << ovLog;
} }
static unsigned static unsigned
ZSTDMT_computeNbJobs(ZSTD_CCtx_params params, size_t srcSize, unsigned nbWorkers) ZSTDMT_computeNbJobs(ZSTD_CCtx_params params, size_t srcSize, unsigned nbWorkers)
{ {
return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, NULL, 0, jobParams); 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) */ 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) ); ZSTDMT_setBufferSize(mtctx->bufPool, ZSTD_compressBound(avgJobSize) );
if (ZSTDMT_serialState_reset(&mtctx->serial, mtctx->seqPool, params, avgJobSize)) if (ZSTDMT_serialState_reset(&mtctx->serial, mtctx->seqPool, params, avgJobSize))
return ERROR(memory_allocation); return ERROR(memory_allocation);
CHECK_F( ZSTDMT_expandJobsTable(mtctx, nbJobs) ); /* only expands if necessary */ FORWARD_IF_ERROR( ZSTDMT_expandJobsTable(mtctx, nbJobs) ); /* only expands if necessary */
{ unsigned u; { unsigned u;
for (u=0; u<nbJobs; u++) { for (u=0; u<nbJobs; u++) {
size_t const jobSize = MIN(remainingSrcSize, avgJobSize); size_t const jobSize = MIN(remainingSrcSize, avgJobSize);
size_t const dstBufferCapacity = ZSTD_compressBound(jobSize); size_t const dstBufferCapacity = ZSTD_compressBound(jobSize);
buffer_t const dstAsBuffer = { (char*)dst + dstBufferPos, dstBufferCapacity }; buffer_t const dstAsBuffer = { (char*)dst + dstBufferPos, dstBufferCapacity };
buffer_t const dstBuffer = u < compressWithinDst ? dstAsBuffer : g_nullBuffer; buffer_t const dstBuffer = u < compressWithinDst ? dstAsBuffer : g_nullBuffer;
size_t dictSize = u ? overlapSize : 0; size_t dictSize = u ? overlapSize : 0;
(U32)pledgedSrcSize, params.nbWorkers, mtctx->cctxPool->totalCCtx); (U32)pledgedSrcSize, params.nbWorkers, mtctx->cctxPool->totalCCtx);
/* params supposed partially fully validated at this point */ /* params supposed partially fully validated at this point */
assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams))); assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
assert(!((dict) && (cdict))); /* either dict or cdict, not both */ assert(!((dict) && (cdict))); /* either dict or cdict, not both */
/* init */ /* init */
if (params.nbWorkers != mtctx->params.nbWorkers) if (params.nbWorkers != mtctx->params.nbWorkers)
CHECK_F( ZSTDMT_resize(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 != 0 && params.jobSize < ZSTDMT_JOBSIZE_MIN) params.jobSize = ZSTDMT_JOBSIZE_MIN;
if (params.jobSize > (size_t)ZSTDMT_JOBSIZE_MAX) params.jobSize = ZSTDMT_JOBSIZE_MAX; if (params.jobSize > (size_t)ZSTDMT_JOBSIZE_MAX) params.jobSize = (size_t)ZSTDMT_JOBSIZE_MAX;
mtctx->singleBlockingThread = (pledgedSrcSize <= ZSTDMT_JOBSIZE_MIN); /* do not trigger multi-threading when srcSize is too small */ mtctx->singleBlockingThread = (pledgedSrcSize <= ZSTDMT_JOBSIZE_MIN); /* do not trigger multi-threading when srcSize is too small */
if (mtctx->singleBlockingThread) { if (mtctx->singleBlockingThread) {
ZSTD_CCtx_params const singleThreadParams = ZSTDMT_initJobCCtxParams(params); ZSTD_CCtx_params const singleThreadParams = ZSTDMT_initJobCCtxParams(params);
DEBUGLOG(5, "ZSTDMT_initCStream_internal: switch to single blocking thread mode"); DEBUGLOG(5, "ZSTDMT_initCStream_internal: switch to single blocking thread mode");
assert(singleThreadParams.nbWorkers == 0); assert(singleThreadParams.nbWorkers == 0);
return ZSTD_initCStream_internal(mtctx->cctxPool->cctx[0], return ZSTD_initCStream_internal(mtctx->cctxPool->cctx[0],
dict, dictSize, cdict, dict, dictSize, cdict,
} }
mtctx->targetPrefixSize = ZSTDMT_computeOverlapSize(params); mtctx->targetPrefixSize = ZSTDMT_computeOverlapSize(params);
DEBUGLOG(4, "overlapLog=%i => %u KB", params.overlapLog, (U32)(mtctx->targetPrefixSize>>10)); DEBUGLOG(4, "overlapLog=%i => %u KB", params.overlapLog, (U32)(mtctx->targetPrefixSize>>10));
mtctx->targetSectionSize = params.jobSize; mtctx->targetSectionSize = params.jobSize;
if (mtctx->targetSectionSize == 0) { if (mtctx->targetSectionSize == 0) {
mtctx->targetSectionSize = 1ULL << ZSTDMT_computeTargetJobLog(params); mtctx->targetSectionSize = 1ULL << ZSTDMT_computeTargetJobLog(params);
} }
assert(mtctx->targetSectionSize <= (size_t)ZSTDMT_JOBSIZE_MAX);
if (params.rsyncable) { if (params.rsyncable) {
/* Aim for the targetsectionSize as the average job size. */ /* Aim for the targetsectionSize as the average job size. */
U32 const jobSizeMB = (U32)(mtctx->targetSectionSize >> 20); U32 const jobSizeMB = (U32)(mtctx->targetSectionSize >> 20);
U32 const rsyncBits = ZSTD_highbit32(jobSizeMB) + 20; U32 const rsyncBits = ZSTD_highbit32(jobSizeMB) + 20;
assert(jobSizeMB >= 1); assert(jobSizeMB >= 1);
DEBUGLOG(4, "rsyncLog = %u", rsyncBits); DEBUGLOG(4, "rsyncLog = %u", rsyncBits);
mtctx->rsync.hash = 0; mtctx->rsync.hash = 0;
mtctx->rsync.hitMask = (1ULL << rsyncBits) - 1; mtctx->rsync.hitMask = (1ULL << rsyncBits) - 1;
cctxParams.fParams = params.fParams; cctxParams.fParams = params.fParams;
return ZSTDMT_initCStream_internal(mtctx, NULL, 0, ZSTD_dct_auto, NULL, cctxParams, ZSTD_CONTENTSIZE_UNKNOWN); return ZSTDMT_initCStream_internal(mtctx, NULL, 0, ZSTD_dct_auto, NULL, cctxParams, ZSTD_CONTENTSIZE_UNKNOWN);
} }
/* ZSTDMT_writeLastEmptyBlock() /* ZSTDMT_writeLastEmptyBlock()
* Write a single empty block with an end-of-frame to finish a frame. * Write a single empty block with an end-of-frame to finish a frame.
* Job must be created from streaming variant. * Job must be created from streaming variant.
* This function is always successfull if expected conditions are fulfilled. * This function is always successful if expected conditions are fulfilled.
*/ */
static void ZSTDMT_writeLastEmptyBlock(ZSTDMT_jobDescription* job) static void ZSTDMT_writeLastEmptyBlock(ZSTDMT_jobDescription* job)
{ {
assert(job->lastJob == 1); assert(job->lastJob == 1);
assert(job->src.size == 0); /* last job is empty -> will be simplified into a last empty block */ 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->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) */ assert(job->dstBuff.start == NULL); /* invoked from streaming variant only (otherwise, dstBuff might be user's output) */
job->dstBuff = ZSTDMT_getBuffer(job->bufPool); job->dstBuff = ZSTDMT_getBuffer(job->bufPool);
{ {
unsigned forwardInputProgress = 0; unsigned forwardInputProgress = 0;
DEBUGLOG(5, "ZSTDMT_compressStream_generic (endOp=%u, srcSize=%u)", DEBUGLOG(5, "ZSTDMT_compressStream_generic (endOp=%u, srcSize=%u)",
(U32)endOp, (U32)(input->size - input->pos)); (U32)endOp, (U32)(input->size - input->pos));
assert(output->pos <= output->size); assert(output->pos <= output->size);
assert(input->pos <= input->size); assert(input->pos <= input->size);
if (mtctx->singleBlockingThread) { /* delegate to single-thread (synchronous) */ if (mtctx->singleBlockingThread) { /* delegate to single-thread (synchronous) */
return ZSTD_compressStream_generic(mtctx->cctxPool->cctx[0], output, input, endOp); return ZSTD_compressStream2(mtctx->cctxPool->cctx[0], output, input, endOp);
} }
if ((mtctx->frameEnded) && (endOp==ZSTD_e_continue)) { if ((mtctx->frameEnded) && (endOp==ZSTD_e_continue)) {
/* current frame being ended. Only flush/end are allowed */ /* current frame being ended. Only flush/end are allowed */
return ERROR(stage_wrong); return ERROR(stage_wrong);
} }
/* single-pass shortcut (note : synchronous-mode) */ /* single-pass shortcut (note : synchronous-mode) */
} }
if ( (mtctx->jobReady) if ( (mtctx->jobReady)
|| (mtctx->inBuff.filled >= mtctx->targetSectionSize) /* filled enough : let's compress */ || (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_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 */ || ((endOp == ZSTD_e_end) && (!mtctx->frameEnded)) ) { /* must finish the frame with a zero-size block */
size_t const jobSize = mtctx->inBuff.filled; size_t const jobSize = mtctx->inBuff.filled;
assert(mtctx->inBuff.filled <= mtctx->targetSectionSize); assert(mtctx->inBuff.filled <= mtctx->targetSectionSize);
CHECK_F( ZSTDMT_createCompressionJob(mtctx, jobSize, endOp) ); FORWARD_IF_ERROR( ZSTDMT_createCompressionJob(mtctx, jobSize, endOp) );
} }
/* check for potential compressed data ready to be flushed */ /* 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 */ { 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 */ 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); DEBUGLOG(5, "end of ZSTDMT_compressStream_generic: remainingToFlush = %u", (U32)remainingToFlush);
return remainingToFlush; return remainingToFlush;
} }
} }
size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input) size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
{ {
CHECK_F( ZSTDMT_compressStream_generic(mtctx, output, input, ZSTD_e_continue) ); FORWARD_IF_ERROR( ZSTDMT_compressStream_generic(mtctx, output, input, ZSTD_e_continue) );
/* recommended next input size : fill current input buffer */ /* 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 */ 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) static size_t ZSTDMT_flushStream_internal(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_EndDirective endFrame)
{ {
size_t const srcSize = mtctx->inBuff.filled; size_t const srcSize = mtctx->inBuff.filled;
DEBUGLOG(5, "ZSTDMT_flushStream_internal"); DEBUGLOG(5, "ZSTDMT_flushStream_internal");
if ( mtctx->jobReady /* one job ready for a worker to pick up */ if ( mtctx->jobReady /* one job ready for a worker to pick up */
|| (srcSize > 0) /* still some data within input buffer */ || (srcSize > 0) /* still some data within input buffer */
|| ((endFrame==ZSTD_e_end) && !mtctx->frameEnded)) { /* need a last 0-size block to end frame */ || ((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)", DEBUGLOG(5, "ZSTDMT_flushStream_internal : create a new job (%u bytes, end:%u)",
(U32)srcSize, (U32)endFrame); (U32)srcSize, (U32)endFrame);
CHECK_F( ZSTDMT_createCompressionJob(mtctx, srcSize, endFrame) ); FORWARD_IF_ERROR( ZSTDMT_createCompressionJob(mtctx, srcSize, endFrame) );
} }
/* check if there is any data available to flush */ /* check if there is any data available to flush */
return ZSTDMT_flushProduced(mtctx, output, 1 /* blockToFlush */, endFrame); return ZSTDMT_flushProduced(mtctx, output, 1 /* blockToFlush */, endFrame);
} }
size_t ZSTDMT_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output) size_t ZSTDMT_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output)

contrib/python-zstandard/zstd/decompress/zstd_ddict.c

if (dictContentType == ZSTD_dct_fullDict) if (dictContentType == ZSTD_dct_fullDict)
return ERROR(dictionary_corrupted); /* only accept specified dictionaries */ return ERROR(dictionary_corrupted); /* only accept specified dictionaries */
return 0; /* pure content mode */ return 0; /* pure content mode */
} }
} }
ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_FRAMEIDSIZE); ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_FRAMEIDSIZE);
/* load entropy tables */ /* load entropy tables */
CHECK_E( ZSTD_loadDEntropy(&ddict->entropy, RETURN_ERROR_IF(ZSTD_isError(ZSTD_loadDEntropy(
ddict->dictContent, ddict->dictSize), &ddict->entropy, ddict->dictContent, ddict->dictSize)),
dictionary_corrupted ); dictionary_corrupted);
ddict->entropyPresent = 1; ddict->entropyPresent = 1;
return 0; return 0;
} }
static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict, static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict,
const void* dict, size_t dictSize, const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictLoadMethod_e dictLoadMethod,
ddict->dictContent = internalBuffer; ddict->dictContent = internalBuffer;
if (!internalBuffer) return ERROR(memory_allocation); if (!internalBuffer) return ERROR(memory_allocation);
memcpy(internalBuffer, dict, dictSize); memcpy(internalBuffer, dict, dictSize);
} }
ddict->dictSize = dictSize; ddict->dictSize = dictSize;
ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */ ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */
/* parse dictionary content */ /* parse dictionary content */
CHECK_F( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) ); FORWARD_IF_ERROR( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) );
return 0; return 0;
} }
ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize, ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictLoadMethod_e dictLoadMethod,
ZSTD_dictContentType_e dictContentType, ZSTD_dictContentType_e dictContentType,
ZSTD_customMem customMem) ZSTD_customMem customMem)

contrib/python-zstandard/zstd/decompress/zstd_decompress.c

{ {
dctx->format = ZSTD_f_zstd1; /* ZSTD_decompressBegin() invokes ZSTD_startingInputLength() with argument dctx->format */ dctx->format = ZSTD_f_zstd1; /* ZSTD_decompressBegin() invokes ZSTD_startingInputLength() with argument dctx->format */
dctx->staticSize = 0; dctx->staticSize = 0;
dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT; dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
dctx->ddict = NULL; dctx->ddict = NULL;
dctx->ddictLocal = NULL; dctx->ddictLocal = NULL;
dctx->dictEnd = NULL; dctx->dictEnd = NULL;
dctx->ddictIsCold = 0; dctx->ddictIsCold = 0;
dctx->dictUses = ZSTD_dont_use;
dctx->inBuff = NULL; dctx->inBuff = NULL;
dctx->inBuffSize = 0; dctx->inBuffSize = 0;
dctx->outBuffSize = 0; dctx->outBuffSize = 0;
dctx->streamStage = zdss_init; dctx->streamStage = zdss_init;
dctx->legacyContext = NULL; dctx->legacyContext = NULL;
dctx->previousLegacyVersion = 0; dctx->previousLegacyVersion = 0;
dctx->noForwardProgress = 0; dctx->noForwardProgress = 0;
dctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid()); dctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());
} }
ZSTD_DCtx* ZSTD_createDCtx(void) ZSTD_DCtx* ZSTD_createDCtx(void)
{ {
DEBUGLOG(3, "ZSTD_createDCtx"); DEBUGLOG(3, "ZSTD_createDCtx");
return ZSTD_createDCtx_advanced(ZSTD_defaultCMem); 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) size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx)
{ {
if (dctx==NULL) return 0; /* support free on NULL */ if (dctx==NULL) return 0; /* support free on NULL */
if (dctx->staticSize) return ERROR(memory_allocation); /* not compatible with static DCtx */ RETURN_ERROR_IF(dctx->staticSize, memory_allocation, "not compatible with static DCtx");
{ ZSTD_customMem const cMem = dctx->customMem; { ZSTD_customMem const cMem = dctx->customMem;
ZSTD_freeDDict(dctx->ddictLocal); ZSTD_clearDict(dctx);
dctx->ddictLocal = NULL;
ZSTD_free(dctx->inBuff, cMem); ZSTD_free(dctx->inBuff, cMem);
dctx->inBuff = NULL; dctx->inBuff = NULL;
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1) #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
if (dctx->legacyContext) if (dctx->legacyContext)
ZSTD_freeLegacyStreamContext(dctx->legacyContext, dctx->previousLegacyVersion); ZSTD_freeLegacyStreamContext(dctx->legacyContext, dctx->previousLegacyVersion);
#endif #endif
ZSTD_free(dctx, cMem); ZSTD_free(dctx, cMem);
return 0; return 0;
/** ZSTD_frameHeaderSize_internal() : /** ZSTD_frameHeaderSize_internal() :
* srcSize must be large enough to reach header size fields. * srcSize must be large enough to reach header size fields.
* note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless. * note : only works for formats ZSTD_f_zstd1 and ZSTD_f_zstd1_magicless.
* @return : size of the Frame Header * @return : size of the Frame Header
* or an error code, which can be tested with ZSTD_isError() */ * 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) static size_t ZSTD_frameHeaderSize_internal(const void* src, size_t srcSize, ZSTD_format_e format)
{ {
size_t const minInputSize = ZSTD_startingInputLength(format); size_t const minInputSize = ZSTD_startingInputLength(format);
if (srcSize < minInputSize) return ERROR(srcSize_wrong); RETURN_ERROR_IF(srcSize < minInputSize, srcSize_wrong);
{ BYTE const fhd = ((const BYTE*)src)[minInputSize-1]; { BYTE const fhd = ((const BYTE*)src)[minInputSize-1];
U32 const dictID= fhd & 3; U32 const dictID= fhd & 3;
U32 const singleSegment = (fhd >> 5) & 1; U32 const singleSegment = (fhd >> 5) & 1;
U32 const fcsId = fhd >> 6; U32 const fcsId = fhd >> 6;
return minInputSize + !singleSegment return minInputSize + !singleSegment
+ ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId] + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId]
+ (singleSegment && !fcsId); + (singleSegment && !fcsId);
* or an error code, which can be tested using ZSTD_isError() */ * 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) size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format)
{ {
const BYTE* ip = (const BYTE*)src; const BYTE* ip = (const BYTE*)src;
size_t const minInputSize = ZSTD_startingInputLength(format); 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 */ 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; if (srcSize < minInputSize) return minInputSize;
if (src==NULL) return ERROR(GENERIC); /* invalid parameter */ RETURN_ERROR_IF(src==NULL, GENERIC, "invalid parameter");
if ( (format != ZSTD_f_zstd1_magicless) if ( (format != ZSTD_f_zstd1_magicless)
&& (MEM_readLE32(src) != ZSTD_MAGICNUMBER) ) { && (MEM_readLE32(src) != ZSTD_MAGICNUMBER) ) {
if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
/* skippable frame */ /* skippable frame */
if (srcSize < ZSTD_SKIPPABLEHEADERSIZE) if (srcSize < ZSTD_SKIPPABLEHEADERSIZE)
return ZSTD_SKIPPABLEHEADERSIZE; /* magic number + frame length */ return ZSTD_SKIPPABLEHEADERSIZE; /* magic number + frame length */
memset(zfhPtr, 0, sizeof(*zfhPtr)); memset(zfhPtr, 0, sizeof(*zfhPtr));
zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_FRAMEIDSIZE); zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_FRAMEIDSIZE);
zfhPtr->frameType = ZSTD_skippableFrame; zfhPtr->frameType = ZSTD_skippableFrame;
return 0; return 0;
} }
return ERROR(prefix_unknown); RETURN_ERROR(prefix_unknown);
} }
/* ensure there is enough `srcSize` to fully read/decode frame header */ /* ensure there is enough `srcSize` to fully read/decode frame header */
{ size_t const fhsize = ZSTD_frameHeaderSize_internal(src, srcSize, format); { size_t const fhsize = ZSTD_frameHeaderSize_internal(src, srcSize, format);
if (srcSize < fhsize) return fhsize; if (srcSize < fhsize) return fhsize;
zfhPtr->headerSize = (U32)fhsize; zfhPtr->headerSize = (U32)fhsize;
} }
{ BYTE const fhdByte = ip[minInputSize-1]; { BYTE const fhdByte = ip[minInputSize-1];
size_t pos = minInputSize; size_t pos = minInputSize;
U32 const dictIDSizeCode = fhdByte&3; U32 const dictIDSizeCode = fhdByte&3;
U32 const checksumFlag = (fhdByte>>2)&1; U32 const checksumFlag = (fhdByte>>2)&1;
U32 const singleSegment = (fhdByte>>5)&1; U32 const singleSegment = (fhdByte>>5)&1;
U32 const fcsID = fhdByte>>6; U32 const fcsID = fhdByte>>6;
U64 windowSize = 0; U64 windowSize = 0;
U32 dictID = 0; U32 dictID = 0;
U64 frameContentSize = ZSTD_CONTENTSIZE_UNKNOWN; U64 frameContentSize = ZSTD_CONTENTSIZE_UNKNOWN;
if ((fhdByte & 0x08) != 0) RETURN_ERROR_IF((fhdByte & 0x08) != 0, frameParameter_unsupported,
return ERROR(frameParameter_unsupported); /* reserved bits, must be zero */ "reserved bits, must be zero");
if (!singleSegment) { if (!singleSegment) {
BYTE const wlByte = ip[pos++]; BYTE const wlByte = ip[pos++];
U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN; U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;
if (windowLog > ZSTD_WINDOWLOG_MAX) RETURN_ERROR_IF(windowLog > ZSTD_WINDOWLOG_MAX, frameParameter_windowTooLarge);
return ERROR(frameParameter_windowTooLarge);
windowSize = (1ULL << windowLog); windowSize = (1ULL << windowLog);
windowSize += (windowSize >> 3) * (wlByte&7); windowSize += (windowSize >> 3) * (wlByte&7);
} }
switch(dictIDSizeCode) switch(dictIDSizeCode)
{ {
default: assert(0); /* impossible */ default: assert(0); /* impossible */
case 0 : break; case 0 : break;
case 1 : dictID = ip[pos]; pos++; break; case 1 : dictID = ip[pos]; pos++; break;
} } } }
} }
static size_t readSkippableFrameSize(void const* src, size_t srcSize) static size_t readSkippableFrameSize(void const* src, size_t srcSize)
{ {
size_t const skippableHeaderSize = ZSTD_SKIPPABLEHEADERSIZE; size_t const skippableHeaderSize = ZSTD_SKIPPABLEHEADERSIZE;
U32 sizeU32; U32 sizeU32;
if (srcSize < ZSTD_SKIPPABLEHEADERSIZE) RETURN_ERROR_IF(srcSize < ZSTD_SKIPPABLEHEADERSIZE, srcSize_wrong);
return ERROR(srcSize_wrong);
sizeU32 = MEM_readLE32((BYTE const*)src + ZSTD_FRAMEIDSIZE); sizeU32 = MEM_readLE32((BYTE const*)src + ZSTD_FRAMEIDSIZE);
if ((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32) RETURN_ERROR_IF((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32,
return ERROR(frameParameter_unsupported); frameParameter_unsupported);
{
return skippableHeaderSize + sizeU32; size_t const skippableSize = skippableHeaderSize + sizeU32;
RETURN_ERROR_IF(skippableSize > srcSize, srcSize_wrong);
return skippableSize;
}
} }
/** ZSTD_findDecompressedSize() : /** ZSTD_findDecompressedSize() :
* compatible with legacy mode * compatible with legacy mode
* `srcSize` must be the exact length of some number of ZSTD compressed and/or * `srcSize` must be the exact length of some number of ZSTD compressed and/or
* skippable frames * skippable frames
* @return : decompressed size of the frames contained */ * @return : decompressed size of the frames contained */
unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize) unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize)
{ {
unsigned long long totalDstSize = 0; unsigned long long totalDstSize = 0;
while (srcSize >= ZSTD_FRAMEHEADERSIZE_PREFIX) { while (srcSize >= ZSTD_FRAMEHEADERSIZE_PREFIX) {
U32 const magicNumber = MEM_readLE32(src); U32 const magicNumber = MEM_readLE32(src);
if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
size_t const skippableSize = readSkippableFrameSize(src, srcSize); size_t const skippableSize = readSkippableFrameSize(src, srcSize);
if (ZSTD_isError(skippableSize)) if (ZSTD_isError(skippableSize)) {
return skippableSize;
if (srcSize < skippableSize) {
return ZSTD_CONTENTSIZE_ERROR; return ZSTD_CONTENTSIZE_ERROR;
} }
assert(skippableSize <= srcSize);
src = (const BYTE *)src + skippableSize; src = (const BYTE *)src + skippableSize;
srcSize -= skippableSize; srcSize -= skippableSize;
continue; continue;
} }
{ unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize); { unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
if (ret >= ZSTD_CONTENTSIZE_ERROR) return ret; if (ret >= ZSTD_CONTENTSIZE_ERROR) return ret;
/** ZSTD_decodeFrameHeader() : /** ZSTD_decodeFrameHeader() :
* `headerSize` must be the size provided by ZSTD_frameHeaderSize(). * `headerSize` must be the size provided by ZSTD_frameHeaderSize().
* @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */ * @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) 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); size_t const result = ZSTD_getFrameHeader_advanced(&(dctx->fParams), src, headerSize, dctx->format);
if (ZSTD_isError(result)) return result; /* invalid header */ if (ZSTD_isError(result)) return result; /* invalid header */
if (result>0) return ERROR(srcSize_wrong); /* headerSize too small */ RETURN_ERROR_IF(result>0, srcSize_wrong, "headerSize too small");
if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID)) #ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
return ERROR(dictionary_wrong); /* 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); if (dctx->fParams.checksumFlag) XXH64_reset(&dctx->xxhState, 0);
return 0; return 0;
} }
static ZSTD_frameSizeInfo ZSTD_errorFrameSizeInfo(size_t ret)
{
ZSTD_frameSizeInfo frameSizeInfo;
frameSizeInfo.compressedSize = ret;
frameSizeInfo.decompressedBound = ZSTD_CONTENTSIZE_ERROR;
return frameSizeInfo;
}
/** ZSTD_findFrameCompressedSize() : static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize)
* 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 frameSizeInfo;
memset(&frameSizeInfo, 0, sizeof(ZSTD_frameSizeInfo));
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1) #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
if (ZSTD_isLegacy(src, srcSize)) if (ZSTD_isLegacy(src, srcSize))
return ZSTD_findFrameCompressedSizeLegacy(src, srcSize); return ZSTD_findFrameSizeInfoLegacy(src, srcSize);
#endif #endif
if ( (srcSize >= ZSTD_SKIPPABLEHEADERSIZE) if ((srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
&& (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START ) { && (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
return readSkippableFrameSize(src, srcSize); frameSizeInfo.compressedSize = readSkippableFrameSize(src, srcSize);
assert(ZSTD_isError(frameSizeInfo.compressedSize) ||
frameSizeInfo.compressedSize <= srcSize);
return frameSizeInfo;
} else { } else {
const BYTE* ip = (const BYTE*)src; const BYTE* ip = (const BYTE*)src;
const BYTE* const ipstart = ip; const BYTE* const ipstart = ip;
size_t remainingSize = srcSize; size_t remainingSize = srcSize;
size_t nbBlocks = 0;
ZSTD_frameHeader zfh; ZSTD_frameHeader zfh;
/* Extract Frame Header */ /* Extract Frame Header */
{ size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize); { size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);
if (ZSTD_isError(ret)) return ret; if (ZSTD_isError(ret))
if (ret > 0) return ERROR(srcSize_wrong); return ZSTD_errorFrameSizeInfo(ret);
if (ret > 0)
return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
} }
ip += zfh.headerSize; ip += zfh.headerSize;
remainingSize -= zfh.headerSize; remainingSize -= zfh.headerSize;
/* Loop on each block */ /* Iterate over each block */
while (1) { while (1) {
blockProperties_t blockProperties; blockProperties_t blockProperties;
size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties); size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
if (ZSTD_isError(cBlockSize)) return cBlockSize; if (ZSTD_isError(cBlockSize))
return ZSTD_errorFrameSizeInfo(cBlockSize);
if (ZSTD_blockHeaderSize + cBlockSize > remainingSize) if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)
return ERROR(srcSize_wrong); return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
ip += ZSTD_blockHeaderSize + cBlockSize; ip += ZSTD_blockHeaderSize + cBlockSize;
remainingSize -= ZSTD_blockHeaderSize + cBlockSize; remainingSize -= ZSTD_blockHeaderSize + cBlockSize;
nbBlocks++;
if (blockProperties.lastBlock) break; if (blockProperties.lastBlock) break;
} }
if (zfh.checksumFlag) { /* Final frame content checksum */ /* Final frame content checksum */
if (remainingSize < 4) return ERROR(srcSize_wrong); if (zfh.checksumFlag) {
if (remainingSize < 4)
return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
ip += 4; ip += 4;
} }
return ip - ipstart; 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 * Frame decoding
***************************************************************/ ***************************************************************/
void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst) void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst)
{ {
if (dst != dctx->previousDstEnd) { /* not contiguous */ if (dst != dctx->previousDstEnd) { /* not contiguous */
dctx->dictEnd = dctx->previousDstEnd; dctx->dictEnd = dctx->previousDstEnd;
dctx->virtualStart = (const char*)dst - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart)); dctx->virtualStart = (const char*)dst - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart));
dctx->prefixStart = dst; dctx->prefixStart = dst;
dctx->previousDstEnd = dst; dctx->previousDstEnd = dst;
} }
} }
/** ZSTD_insertBlock() : /** ZSTD_insertBlock() :
insert `src` block into `dctx` history. Useful to track uncompressed blocks. */ * insert `src` block into `dctx` history. Useful to track uncompressed blocks. */
size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize) size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize)
{ {
DEBUGLOG(5, "ZSTD_insertBlock: %u bytes", (unsigned)blockSize);
ZSTD_checkContinuity(dctx, blockStart); ZSTD_checkContinuity(dctx, blockStart);
dctx->previousDstEnd = (const char*)blockStart + blockSize; dctx->previousDstEnd = (const char*)blockStart + blockSize;
return blockSize; return blockSize;
} }
static size_t ZSTD_copyRawBlock(void* dst, size_t dstCapacity, static size_t ZSTD_copyRawBlock(void* dst, size_t dstCapacity,
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
DEBUGLOG(5, "ZSTD_copyRawBlock"); DEBUGLOG(5, "ZSTD_copyRawBlock");
if (dst == NULL) { if (dst == NULL) {
if (srcSize == 0) return 0; if (srcSize == 0) return 0;
return ERROR(dstBuffer_null); RETURN_ERROR(dstBuffer_null);
} }
if (srcSize > dstCapacity) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(srcSize > dstCapacity, dstSize_tooSmall);
memcpy(dst, src, srcSize); memcpy(dst, src, srcSize);
return srcSize; return srcSize;
} }
static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity, static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity,
BYTE b, BYTE b,
size_t regenSize) size_t regenSize)
{ {
if (dst == NULL) { if (dst == NULL) {
if (regenSize == 0) return 0; if (regenSize == 0) return 0;
return ERROR(dstBuffer_null); RETURN_ERROR(dstBuffer_null);
} }
if (regenSize > dstCapacity) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(regenSize > dstCapacity, dstSize_tooSmall);
memset(dst, b, regenSize); memset(dst, b, regenSize);
return regenSize; return regenSize;
} }
/*! ZSTD_decompressFrame() : /*! ZSTD_decompressFrame() :
* @dctx must be properly initialized * @dctx must be properly initialized
* will update *srcPtr and *srcSizePtr, * will update *srcPtr and *srcSizePtr,
* to make *srcPtr progress by one frame. */ * to make *srcPtr progress by one frame. */
static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx, static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void** srcPtr, size_t *srcSizePtr) const void** srcPtr, size_t *srcSizePtr)
{ {
const BYTE* ip = (const BYTE*)(*srcPtr); const BYTE* ip = (const BYTE*)(*srcPtr);
BYTE* const ostart = (BYTE* const)dst; BYTE* const ostart = (BYTE* const)dst;
BYTE* const oend = ostart + dstCapacity; BYTE* const oend = ostart + dstCapacity;
BYTE* op = ostart; BYTE* op = ostart;
size_t remainingSrcSize = *srcSizePtr; size_t remainingSrcSize = *srcSizePtr;
DEBUGLOG(4, "ZSTD_decompressFrame (srcSize:%i)", (int)*srcSizePtr); DEBUGLOG(4, "ZSTD_decompressFrame (srcSize:%i)", (int)*srcSizePtr);
/* check */ /* check */
if (remainingSrcSize < ZSTD_FRAMEHEADERSIZE_MIN+ZSTD_blockHeaderSize) RETURN_ERROR_IF(
return ERROR(srcSize_wrong); remainingSrcSize < ZSTD_FRAMEHEADERSIZE_MIN+ZSTD_blockHeaderSize,
srcSize_wrong);
/* Frame Header */ /* Frame Header */
{ size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_FRAMEHEADERSIZE_PREFIX); { size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_FRAMEHEADERSIZE_PREFIX);
if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize; if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize;
if (remainingSrcSize < frameHeaderSize+ZSTD_blockHeaderSize) RETURN_ERROR_IF(remainingSrcSize < frameHeaderSize+ZSTD_blockHeaderSize,
return ERROR(srcSize_wrong); srcSize_wrong);
CHECK_F( ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize) ); FORWARD_IF_ERROR( ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize) );
ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize; ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize;
} }
/* Loop on each block */ /* Loop on each block */
while (1) { while (1) {
size_t decodedSize; size_t decodedSize;
blockProperties_t blockProperties; blockProperties_t blockProperties;
size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSrcSize, &blockProperties); size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSrcSize, &blockProperties);
if (ZSTD_isError(cBlockSize)) return cBlockSize; if (ZSTD_isError(cBlockSize)) return cBlockSize;
ip += ZSTD_blockHeaderSize; ip += ZSTD_blockHeaderSize;
remainingSrcSize -= ZSTD_blockHeaderSize; remainingSrcSize -= ZSTD_blockHeaderSize;
if (cBlockSize > remainingSrcSize) return ERROR(srcSize_wrong); RETURN_ERROR_IF(cBlockSize > remainingSrcSize, srcSize_wrong);
switch(blockProperties.blockType) switch(blockProperties.blockType)
{ {
case bt_compressed: case bt_compressed:
decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend-op, ip, cBlockSize, /* frame */ 1); decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend-op, ip, cBlockSize, /* frame */ 1);
break; break;
case bt_raw : case bt_raw :
decodedSize = ZSTD_copyRawBlock(op, oend-op, ip, cBlockSize); decodedSize = ZSTD_copyRawBlock(op, oend-op, ip, cBlockSize);
break; break;
case bt_rle : case bt_rle :
decodedSize = ZSTD_setRleBlock(op, oend-op, *ip, blockProperties.origSize); decodedSize = ZSTD_setRleBlock(op, oend-op, *ip, blockProperties.origSize);
break; break;
case bt_reserved : case bt_reserved :
default: default:
return ERROR(corruption_detected); RETURN_ERROR(corruption_detected);
} }
if (ZSTD_isError(decodedSize)) return decodedSize; if (ZSTD_isError(decodedSize)) return decodedSize;
if (dctx->fParams.checksumFlag) if (dctx->fParams.checksumFlag)
XXH64_update(&dctx->xxhState, op, decodedSize); XXH64_update(&dctx->xxhState, op, decodedSize);
op += decodedSize; op += decodedSize;
ip += cBlockSize; ip += cBlockSize;
remainingSrcSize -= cBlockSize; remainingSrcSize -= cBlockSize;
if (blockProperties.lastBlock) break; if (blockProperties.lastBlock) break;
} }
if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) { if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {
if ((U64)(op-ostart) != dctx->fParams.frameContentSize) { RETURN_ERROR_IF((U64)(op-ostart) != dctx->fParams.frameContentSize,
return ERROR(corruption_detected); corruption_detected);
} } }
if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */ if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */
U32 const checkCalc = (U32)XXH64_digest(&dctx->xxhState); U32 const checkCalc = (U32)XXH64_digest(&dctx->xxhState);
U32 checkRead; U32 checkRead;
if (remainingSrcSize<4) return ERROR(checksum_wrong); RETURN_ERROR_IF(remainingSrcSize<4, checksum_wrong);
checkRead = MEM_readLE32(ip); checkRead = MEM_readLE32(ip);
if (checkRead != checkCalc) return ERROR(checksum_wrong); RETURN_ERROR_IF(checkRead != checkCalc, checksum_wrong);
ip += 4; ip += 4;
remainingSrcSize -= 4; remainingSrcSize -= 4;
} }
/* Allow caller to get size read */ /* Allow caller to get size read */
*srcPtr = ip; *srcPtr = ip;
*srcSizePtr = remainingSrcSize; *srcSizePtr = remainingSrcSize;
return op-ostart; return op-ostart;
while (srcSize >= ZSTD_FRAMEHEADERSIZE_PREFIX) { while (srcSize >= ZSTD_FRAMEHEADERSIZE_PREFIX) {
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1) #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
if (ZSTD_isLegacy(src, srcSize)) { if (ZSTD_isLegacy(src, srcSize)) {
size_t decodedSize; size_t decodedSize;
size_t const frameSize = ZSTD_findFrameCompressedSizeLegacy(src, srcSize); size_t const frameSize = ZSTD_findFrameCompressedSizeLegacy(src, srcSize);
if (ZSTD_isError(frameSize)) return frameSize; if (ZSTD_isError(frameSize)) return frameSize;
/* legacy support is not compatible with static dctx */ RETURN_ERROR_IF(dctx->staticSize, memory_allocation,
if (dctx->staticSize) return ERROR(memory_allocation); "legacy support is not compatible with static dctx");
decodedSize = ZSTD_decompressLegacy(dst, dstCapacity, src, frameSize, dict, dictSize); decodedSize = ZSTD_decompressLegacy(dst, dstCapacity, src, frameSize, dict, dictSize);
if (ZSTD_isError(decodedSize)) return decodedSize; if (ZSTD_isError(decodedSize)) return decodedSize;
assert(decodedSize <=- dstCapacity); assert(decodedSize <=- dstCapacity);
dst = (BYTE*)dst + decodedSize; dst = (BYTE*)dst + decodedSize;
dstCapacity -= decodedSize; dstCapacity -= decodedSize;
src = (const BYTE*)src + frameSize; src = (const BYTE*)src + frameSize;
srcSize -= frameSize; srcSize -= frameSize;
continue; continue;
} }
#endif #endif
{ U32 const magicNumber = MEM_readLE32(src); { U32 const magicNumber = MEM_readLE32(src);
DEBUGLOG(4, "reading magic number %08X (expecting %08X)", DEBUGLOG(4, "reading magic number %08X (expecting %08X)",
(unsigned)magicNumber, ZSTD_MAGICNUMBER); (unsigned)magicNumber, ZSTD_MAGICNUMBER);
if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
size_t const skippableSize = readSkippableFrameSize(src, srcSize); size_t const skippableSize = readSkippableFrameSize(src, srcSize);
if (ZSTD_isError(skippableSize)) FORWARD_IF_ERROR(skippableSize);
return skippableSize; assert(skippableSize <= srcSize);
if (srcSize < skippableSize) return ERROR(srcSize_wrong);
src = (const BYTE *)src + skippableSize; src = (const BYTE *)src + skippableSize;
srcSize -= skippableSize; srcSize -= skippableSize;
continue; continue;
} } } }
if (ddict) { if (ddict) {
/* we were called from ZSTD_decompress_usingDDict */ /* we were called from ZSTD_decompress_usingDDict */
CHECK_F(ZSTD_decompressBegin_usingDDict(dctx, ddict)); FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(dctx, ddict));
} else { } else {
/* this will initialize correctly with no dict if dict == NULL, so /* this will initialize correctly with no dict if dict == NULL, so
* use this in all cases but ddict */ * use this in all cases but ddict */
CHECK_F(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize)); FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize));
} }
ZSTD_checkContinuity(dctx, dst); ZSTD_checkContinuity(dctx, dst);
{ const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity, { const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity,
&src, &srcSize); &src, &srcSize);
if ( (ZSTD_getErrorCode(res) == ZSTD_error_prefix_unknown) RETURN_ERROR_IF(
&& (moreThan1Frame==1) ) { (ZSTD_getErrorCode(res) == ZSTD_error_prefix_unknown)
/* at least one frame successfully completed, && (moreThan1Frame==1),
* but following bytes are garbage : srcSize_wrong,
* it's more likely to be a srcSize error, "at least one frame successfully completed, but following "
* specifying more bytes than compressed size of frame(s). "bytes are garbage: it's more likely to be a srcSize error, "
* This error message replaces ERROR(prefix_unknown), "specifying more bytes than compressed size of frame(s). This "
* which would be confusing, as the first header is actually correct. "error message replaces ERROR(prefix_unknown), which would be "
* Note that one could be unlucky, it might be a corruption error instead, "confusing, as the first header is actually correct. Note that "
* happening right at the place where we expect zstd magic bytes. "one could be unlucky, it might be a corruption error instead, "
* But this is _much_ less likely than a srcSize field error. */ "happening right at the place where we expect zstd magic "
return ERROR(srcSize_wrong); "bytes. But this is _much_ less likely than a srcSize field "
} "error.");
if (ZSTD_isError(res)) return res; if (ZSTD_isError(res)) return res;
assert(res <= dstCapacity); assert(res <= dstCapacity);
dst = (BYTE*)dst + res; dst = (BYTE*)dst + res;
dstCapacity -= res; dstCapacity -= res;
} }
moreThan1Frame = 1; moreThan1Frame = 1;
} /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */ } /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */
if (srcSize) return ERROR(srcSize_wrong); /* input not entirely consumed */ RETURN_ERROR_IF(srcSize, srcSize_wrong, "input not entirely consumed");
return (BYTE*)dst - (BYTE*)dststart; return (BYTE*)dst - (BYTE*)dststart;
} }
size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx, size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const void* dict, size_t dictSize) const void* dict, size_t dictSize)
{ {
return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, dict, dictSize, NULL); 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) 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); 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) size_t ZSTD_decompress(void* dst, size_t dstCapacity, const void* src, size_t srcSize)
{ {
#if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE>=1) #if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE>=1)
size_t regenSize; size_t regenSize;
ZSTD_DCtx* const dctx = ZSTD_createDCtx(); ZSTD_DCtx* const dctx = ZSTD_createDCtx();
if (dctx==NULL) return ERROR(memory_allocation); RETURN_ERROR_IF(dctx==NULL, memory_allocation);
regenSize = ZSTD_decompressDCtx(dctx, dst, dstCapacity, src, srcSize); regenSize = ZSTD_decompressDCtx(dctx, dst, dstCapacity, src, srcSize);
ZSTD_freeDCtx(dctx); ZSTD_freeDCtx(dctx);
return regenSize; return regenSize;
#else /* stack mode */ #else /* stack mode */
ZSTD_DCtx dctx; ZSTD_DCtx dctx;
ZSTD_initDCtx_internal(&dctx); ZSTD_initDCtx_internal(&dctx);
return ZSTD_decompressDCtx(&dctx, dst, dstCapacity, src, srcSize); return ZSTD_decompressDCtx(&dctx, dst, dstCapacity, src, srcSize);
#endif #endif
/** ZSTD_decompressContinue() : /** ZSTD_decompressContinue() :
* srcSize : must be the exact nb of bytes expected (see ZSTD_nextSrcSizeToDecompress()) * srcSize : must be the exact nb of bytes expected (see ZSTD_nextSrcSizeToDecompress())
* @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity) * @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity)
* or an error code, which can be tested using ZSTD_isError() */ * 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) 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); DEBUGLOG(5, "ZSTD_decompressContinue (srcSize:%u)", (unsigned)srcSize);
/* Sanity check */ /* Sanity check */
if (srcSize != dctx->expected) RETURN_ERROR_IF(srcSize != dctx->expected, srcSize_wrong, "not allowed");
return ERROR(srcSize_wrong); /* not allowed */
if (dstCapacity) ZSTD_checkContinuity(dctx, dst); if (dstCapacity) ZSTD_checkContinuity(dctx, dst);
switch (dctx->stage) switch (dctx->stage)
{ {
case ZSTDds_getFrameHeaderSize : case ZSTDds_getFrameHeaderSize :
assert(src != NULL); assert(src != NULL);
if (dctx->format == ZSTD_f_zstd1) { /* allows header */ if (dctx->format == ZSTD_f_zstd1) { /* allows header */
assert(srcSize >= ZSTD_FRAMEIDSIZE); /* to read skippable magic number */ assert(srcSize >= ZSTD_FRAMEIDSIZE); /* to read skippable magic number */
if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */ if ((MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */
memcpy(dctx->headerBuffer, src, srcSize); memcpy(dctx->headerBuffer, src, srcSize);
dctx->expected = ZSTD_SKIPPABLEHEADERSIZE - srcSize; /* remaining to load to get full skippable frame header */ dctx->expected = ZSTD_SKIPPABLEHEADERSIZE - srcSize; /* remaining to load to get full skippable frame header */
dctx->stage = ZSTDds_decodeSkippableHeader; dctx->stage = ZSTDds_decodeSkippableHeader;
return 0; return 0;
} } } }
dctx->headerSize = ZSTD_frameHeaderSize_internal(src, srcSize, dctx->format); dctx->headerSize = ZSTD_frameHeaderSize_internal(src, srcSize, dctx->format);
if (ZSTD_isError(dctx->headerSize)) return dctx->headerSize; if (ZSTD_isError(dctx->headerSize)) return dctx->headerSize;
memcpy(dctx->headerBuffer, src, srcSize); memcpy(dctx->headerBuffer, src, srcSize);
dctx->expected = dctx->headerSize - srcSize; dctx->expected = dctx->headerSize - srcSize;
dctx->stage = ZSTDds_decodeFrameHeader; dctx->stage = ZSTDds_decodeFrameHeader;
return 0; return 0;
case ZSTDds_decodeFrameHeader: case ZSTDds_decodeFrameHeader:
assert(src != NULL); assert(src != NULL);
memcpy(dctx->headerBuffer + (dctx->headerSize - srcSize), src, srcSize); memcpy(dctx->headerBuffer + (dctx->headerSize - srcSize), src, srcSize);
CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize)); FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));
dctx->expected = ZSTD_blockHeaderSize; dctx->expected = ZSTD_blockHeaderSize;
dctx->stage = ZSTDds_decodeBlockHeader; dctx->stage = ZSTDds_decodeBlockHeader;
return 0; return 0;
case ZSTDds_decodeBlockHeader: case ZSTDds_decodeBlockHeader:
{ blockProperties_t bp; { blockProperties_t bp;
size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp); size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);
if (ZSTD_isError(cBlockSize)) return cBlockSize; if (ZSTD_isError(cBlockSize)) return cBlockSize;
RETURN_ERROR_IF(cBlockSize > dctx->fParams.blockSizeMax, corruption_detected, "Block Size Exceeds Maximum");
dctx->expected = cBlockSize; dctx->expected = cBlockSize;
dctx->bType = bp.blockType; dctx->bType = bp.blockType;
dctx->rleSize = bp.origSize; dctx->rleSize = bp.origSize;
if (cBlockSize) { if (cBlockSize) {
dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock; dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock;
return 0; return 0;
} }
/* empty block */ /* empty block */
case bt_raw : case bt_raw :
rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize); rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize);
break; break;
case bt_rle : case bt_rle :
rSize = ZSTD_setRleBlock(dst, dstCapacity, *(const BYTE*)src, dctx->rleSize); rSize = ZSTD_setRleBlock(dst, dstCapacity, *(const BYTE*)src, dctx->rleSize);
break; break;
case bt_reserved : /* should never happen */ case bt_reserved : /* should never happen */
default: default:
return ERROR(corruption_detected); RETURN_ERROR(corruption_detected);
} }
if (ZSTD_isError(rSize)) return rSize; if (ZSTD_isError(rSize)) return rSize;
RETURN_ERROR_IF(rSize > dctx->fParams.blockSizeMax, corruption_detected, "Decompressed Block Size Exceeds Maximum");
DEBUGLOG(5, "ZSTD_decompressContinue: decoded size from block : %u", (unsigned)rSize); DEBUGLOG(5, "ZSTD_decompressContinue: decoded size from block : %u", (unsigned)rSize);
dctx->decodedSize += rSize; dctx->decodedSize += rSize;
if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, dst, rSize); if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, dst, rSize);
if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */ if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */
DEBUGLOG(4, "ZSTD_decompressContinue: decoded size from frame : %u", (unsigned)dctx->decodedSize); DEBUGLOG(4, "ZSTD_decompressContinue: decoded size from frame : %u", (unsigned)dctx->decodedSize);
if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) { RETURN_ERROR_IF(
if (dctx->decodedSize != dctx->fParams.frameContentSize) { dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
return ERROR(corruption_detected); && dctx->decodedSize != dctx->fParams.frameContentSize,
} } corruption_detected);
if (dctx->fParams.checksumFlag) { /* another round for frame checksum */ if (dctx->fParams.checksumFlag) { /* another round for frame checksum */
dctx->expected = 4; dctx->expected = 4;
dctx->stage = ZSTDds_checkChecksum; dctx->stage = ZSTDds_checkChecksum;
} else { } else {
dctx->expected = 0; /* ends here */ dctx->expected = 0; /* ends here */
dctx->stage = ZSTDds_getFrameHeaderSize; dctx->stage = ZSTDds_getFrameHeaderSize;
} }
} else { } else {
dctx->stage = ZSTDds_decodeBlockHeader; dctx->stage = ZSTDds_decodeBlockHeader;
dctx->expected = ZSTD_blockHeaderSize; dctx->expected = ZSTD_blockHeaderSize;
dctx->previousDstEnd = (char*)dst + rSize; dctx->previousDstEnd = (char*)dst + rSize;
} }
return rSize; return rSize;
} }
case ZSTDds_checkChecksum: case ZSTDds_checkChecksum:
assert(srcSize == 4); /* guaranteed by dctx->expected */ assert(srcSize == 4); /* guaranteed by dctx->expected */
{ U32 const h32 = (U32)XXH64_digest(&dctx->xxhState); { U32 const h32 = (U32)XXH64_digest(&dctx->xxhState);
U32 const check32 = MEM_readLE32(src); U32 const check32 = MEM_readLE32(src);
DEBUGLOG(4, "ZSTD_decompressContinue: checksum : calculated %08X :: %08X read", (unsigned)h32, (unsigned)check32); DEBUGLOG(4, "ZSTD_decompressContinue: checksum : calculated %08X :: %08X read", (unsigned)h32, (unsigned)check32);
if (check32 != h32) return ERROR(checksum_wrong); RETURN_ERROR_IF(check32 != h32, checksum_wrong);
dctx->expected = 0; dctx->expected = 0;
dctx->stage = ZSTDds_getFrameHeaderSize; dctx->stage = ZSTDds_getFrameHeaderSize;
return 0; return 0;
} }
case ZSTDds_decodeSkippableHeader: case ZSTDds_decodeSkippableHeader:
assert(src != NULL); assert(src != NULL);
assert(srcSize <= ZSTD_SKIPPABLEHEADERSIZE); assert(srcSize <= ZSTD_SKIPPABLEHEADERSIZE);
memcpy(dctx->headerBuffer + (ZSTD_SKIPPABLEHEADERSIZE - srcSize), src, srcSize); /* complete skippable header */ 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->expected = MEM_readLE32(dctx->headerBuffer + ZSTD_FRAMEIDSIZE); /* note : dctx->expected can grow seriously large, beyond local buffer size */
dctx->stage = ZSTDds_skipFrame; dctx->stage = ZSTDds_skipFrame;
return 0; return 0;
case ZSTDds_skipFrame: case ZSTDds_skipFrame:
dctx->expected = 0; dctx->expected = 0;
dctx->stage = ZSTDds_getFrameHeaderSize; dctx->stage = ZSTDds_getFrameHeaderSize;
return 0; return 0;
default: default:
assert(0); /* impossible */ assert(0); /* impossible */
return ERROR(GENERIC); /* some compiler require default to do something */ RETURN_ERROR(GENERIC); /* some compiler require default to do something */
} }
} }
static size_t ZSTD_refDictContent(ZSTD_DCtx* dctx, const void* dict, size_t dictSize) static size_t ZSTD_refDictContent(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
{ {
dctx->dictEnd = dctx->previousDstEnd; dctx->dictEnd = dctx->previousDstEnd;
dctx->virtualStart = (const char*)dict - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart)); dctx->virtualStart = (const char*)dict - ((const char*)(dctx->previousDstEnd) - (const char*)(dctx->prefixStart));
dctx->prefixStart = dict; dctx->prefixStart = dict;
dctx->previousDstEnd = (const char*)dict + dictSize; dctx->previousDstEnd = (const char*)dict + dictSize;
return 0; return 0;
} }
/*! ZSTD_loadDEntropy() : /*! ZSTD_loadDEntropy() :
* dict : must point at beginning of a valid zstd dictionary. * dict : must point at beginning of a valid zstd dictionary.
* @return : size of entropy tables read */ * @return : size of entropy tables read */
size_t size_t
ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy, ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy,
const void* const dict, size_t const dictSize) const void* const dict, size_t const dictSize)
{ {
const BYTE* dictPtr = (const BYTE*)dict; const BYTE* dictPtr = (const BYTE*)dict;
const BYTE* const dictEnd = dictPtr + dictSize; const BYTE* const dictEnd = dictPtr + dictSize;
if (dictSize <= 8) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(dictSize <= 8, dictionary_corrupted);
assert(MEM_readLE32(dict) == ZSTD_MAGIC_DICTIONARY); /* dict must be valid */ assert(MEM_readLE32(dict) == ZSTD_MAGIC_DICTIONARY); /* dict must be valid */
dictPtr += 8; /* skip header = magic + dictID */ 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, 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(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); 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 */ { 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); size_t const workspaceSize = sizeof(entropy->LLTable) + sizeof(entropy->OFTable) + sizeof(entropy->MLTable);
#ifdef HUF_FORCE_DECOMPRESS_X1 #ifdef HUF_FORCE_DECOMPRESS_X1
/* in minimal huffman, we always use X1 variants */ /* in minimal huffman, we always use X1 variants */
size_t const hSize = HUF_readDTableX1_wksp(entropy->hufTable, size_t const hSize = HUF_readDTableX1_wksp(entropy->hufTable,
dictPtr, dictEnd - dictPtr, dictPtr, dictEnd - dictPtr,
workspace, workspaceSize); workspace, workspaceSize);
#else #else
size_t const hSize = HUF_readDTableX2_wksp(entropy->hufTable, size_t const hSize = HUF_readDTableX2_wksp(entropy->hufTable,
dictPtr, dictEnd - dictPtr, dictPtr, dictEnd - dictPtr,
workspace, workspaceSize); workspace, workspaceSize);
#endif #endif
if (HUF_isError(hSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(HUF_isError(hSize), dictionary_corrupted);
dictPtr += hSize; dictPtr += hSize;
} }
{ short offcodeNCount[MaxOff+1]; { short offcodeNCount[MaxOff+1];
unsigned offcodeMaxValue = MaxOff, offcodeLog; unsigned offcodeMaxValue = MaxOff, offcodeLog;
size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr); size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);
if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted);
if (offcodeMaxValue > MaxOff) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(offcodeMaxValue > MaxOff, dictionary_corrupted);
if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted);
ZSTD_buildFSETable( entropy->OFTable, ZSTD_buildFSETable( entropy->OFTable,
offcodeNCount, offcodeMaxValue, offcodeNCount, offcodeMaxValue,
OF_base, OF_bits, OF_base, OF_bits,
offcodeLog); offcodeLog);
dictPtr += offcodeHeaderSize; dictPtr += offcodeHeaderSize;
} }
{ short matchlengthNCount[MaxML+1]; { short matchlengthNCount[MaxML+1];
unsigned matchlengthMaxValue = MaxML, matchlengthLog; unsigned matchlengthMaxValue = MaxML, matchlengthLog;
size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr); size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);
if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted);
if (matchlengthMaxValue > MaxML) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(matchlengthMaxValue > MaxML, dictionary_corrupted);
if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted);
ZSTD_buildFSETable( entropy->MLTable, ZSTD_buildFSETable( entropy->MLTable,
matchlengthNCount, matchlengthMaxValue, matchlengthNCount, matchlengthMaxValue,
ML_base, ML_bits, ML_base, ML_bits,
matchlengthLog); matchlengthLog);
dictPtr += matchlengthHeaderSize; dictPtr += matchlengthHeaderSize;
} }
{ short litlengthNCount[MaxLL+1]; { short litlengthNCount[MaxLL+1];
unsigned litlengthMaxValue = MaxLL, litlengthLog; unsigned litlengthMaxValue = MaxLL, litlengthLog;
size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr); size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);
if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted);
if (litlengthMaxValue > MaxLL) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(litlengthMaxValue > MaxLL, dictionary_corrupted);
if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted);
ZSTD_buildFSETable( entropy->LLTable, ZSTD_buildFSETable( entropy->LLTable,
litlengthNCount, litlengthMaxValue, litlengthNCount, litlengthMaxValue,
LL_base, LL_bits, LL_base, LL_bits,
litlengthLog); litlengthLog);
dictPtr += litlengthHeaderSize; dictPtr += litlengthHeaderSize;
} }
if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted);
{ int i; { int i;
size_t const dictContentSize = (size_t)(dictEnd - (dictPtr+12)); size_t const dictContentSize = (size_t)(dictEnd - (dictPtr+12));
for (i=0; i<3; i++) { for (i=0; i<3; i++) {
U32 const rep = MEM_readLE32(dictPtr); dictPtr += 4; U32 const rep = MEM_readLE32(dictPtr); dictPtr += 4;
if (rep==0 || rep >= dictContentSize) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(rep==0 || rep >= dictContentSize,
dictionary_corrupted);
entropy->rep[i] = rep; entropy->rep[i] = rep;
} } } }
return dictPtr - (const BYTE*)dict; return dictPtr - (const BYTE*)dict;
} }
static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize) static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
{ {
if (dictSize < 8) return ZSTD_refDictContent(dctx, dict, dictSize); if (dictSize < 8) return ZSTD_refDictContent(dctx, dict, dictSize);
{ U32 const magic = MEM_readLE32(dict); { U32 const magic = MEM_readLE32(dict);
if (magic != ZSTD_MAGIC_DICTIONARY) { if (magic != ZSTD_MAGIC_DICTIONARY) {
return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */ return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */
} } } }
dctx->dictID = MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE); dctx->dictID = MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE);
/* load entropy tables */ /* load entropy tables */
{ size_t const eSize = ZSTD_loadDEntropy(&dctx->entropy, dict, dictSize); { size_t const eSize = ZSTD_loadDEntropy(&dctx->entropy, dict, dictSize);
if (ZSTD_isError(eSize)) return ERROR(dictionary_corrupted); RETURN_ERROR_IF(ZSTD_isError(eSize), dictionary_corrupted);
dict = (const char*)dict + eSize; dict = (const char*)dict + eSize;
dictSize -= eSize; dictSize -= eSize;
} }
dctx->litEntropy = dctx->fseEntropy = 1; dctx->litEntropy = dctx->fseEntropy = 1;
/* reference dictionary content */ /* reference dictionary content */
return ZSTD_refDictContent(dctx, dict, dictSize); return ZSTD_refDictContent(dctx, dict, dictSize);
} }
dctx->MLTptr = dctx->entropy.MLTable; dctx->MLTptr = dctx->entropy.MLTable;
dctx->OFTptr = dctx->entropy.OFTable; dctx->OFTptr = dctx->entropy.OFTable;
dctx->HUFptr = dctx->entropy.hufTable; dctx->HUFptr = dctx->entropy.hufTable;
return 0; return 0;
} }
size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize) size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
{ {
CHECK_F( ZSTD_decompressBegin(dctx) ); FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) );
if (dict && dictSize) if (dict && dictSize)
CHECK_E(ZSTD_decompress_insertDictionary(dctx, dict, dictSize), dictionary_corrupted); RETURN_ERROR_IF(
ZSTD_isError(ZSTD_decompress_insertDictionary(dctx, dict, dictSize)),
dictionary_corrupted);
return 0; return 0;
} }
/* ====== ZSTD_DDict ====== */ /* ====== ZSTD_DDict ====== */
size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict) size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
{ {
DEBUGLOG(4, "ZSTD_decompressBegin_usingDDict"); DEBUGLOG(4, "ZSTD_decompressBegin_usingDDict");
assert(dctx != NULL); assert(dctx != NULL);
if (ddict) { if (ddict) {
const char* const dictStart = (const char*)ZSTD_DDict_dictContent(ddict); const char* const dictStart = (const char*)ZSTD_DDict_dictContent(ddict);
size_t const dictSize = ZSTD_DDict_dictSize(ddict); size_t const dictSize = ZSTD_DDict_dictSize(ddict);
const void* const dictEnd = dictStart + dictSize; const void* const dictEnd = dictStart + dictSize;
dctx->ddictIsCold = (dctx->dictEnd != dictEnd); dctx->ddictIsCold = (dctx->dictEnd != dictEnd);
DEBUGLOG(4, "DDict is %s", DEBUGLOG(4, "DDict is %s",
dctx->ddictIsCold ? "~cold~" : "hot!"); dctx->ddictIsCold ? "~cold~" : "hot!");
} }
CHECK_F( ZSTD_decompressBegin(dctx) ); FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) );
if (ddict) { /* NULL ddict is equivalent to no dictionary */ if (ddict) { /* NULL ddict is equivalent to no dictionary */
ZSTD_copyDDictParameters(dctx, ddict); ZSTD_copyDDictParameters(dctx, ddict);
} }
return 0; return 0;
} }
/*! ZSTD_getDictID_fromDict() : /*! ZSTD_getDictID_fromDict() :
* Provides the dictID stored within dictionary. * Provides the dictID stored within dictionary.
* if @return == 0, the dictionary is not conformant with Zstandard specification. * if @return == 0, the dictionary is not conformant with Zstandard specification.
* It can still be loaded, but as a content-only dictionary. */ * It can still be loaded, but as a content-only dictionary. */
unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize) unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize)
{ {
if (dictSize < 8) return 0; if (dictSize < 8) return 0;
if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) return 0; if (MEM_readLE32(dict) != ZSTD_MAGIC_DICTIONARY) return 0;
return MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE); return MEM_readLE32((const char*)dict + ZSTD_FRAMEIDSIZE);
} }
/*! ZSTD_getDictID_fromFrame() : /*! ZSTD_getDictID_fromFrame() :
* Provides the dictID required to decompresse frame stored within `src`. * Provides the dictID required to decompress frame stored within `src`.
* If @return == 0, the dictID could not be decoded. * If @return == 0, the dictID could not be decoded.
* This could for one of the following reasons : * This could for one of the following reasons :
* - The frame does not require a dictionary (most common case). * - The frame does not require a dictionary (most common case).
* - The frame was built with dictID intentionally removed. * - The frame was built with dictID intentionally removed.
* Needed dictionary is a hidden information. * Needed dictionary is a hidden information.
* Note : this use case also happens when using a non-conformant dictionary. * 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. * - `srcSize` is too small, and as a result, frame header could not be decoded.
* Note : possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`. * Note : possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`.
size_t ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_MAX + ZSTD_blockHeaderSize; } size_t ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_MAX + ZSTD_blockHeaderSize; }
size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_MAX; } size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_MAX; }
size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx,
const void* dict, size_t dictSize, const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictLoadMethod_e dictLoadMethod,
ZSTD_dictContentType_e dictContentType) ZSTD_dictContentType_e dictContentType)
{ {
if (dctx->streamStage != zdss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
ZSTD_freeDDict(dctx->ddictLocal); ZSTD_clearDict(dctx);
if (dict && dictSize >= 8) { if (dict && dictSize >= 8) {
dctx->ddictLocal = ZSTD_createDDict_advanced(dict, dictSize, dictLoadMethod, dictContentType, dctx->customMem); dctx->ddictLocal = ZSTD_createDDict_advanced(dict, dictSize, dictLoadMethod, dictContentType, dctx->customMem);
if (dctx->ddictLocal == NULL) return ERROR(memory_allocation); RETURN_ERROR_IF(dctx->ddictLocal == NULL, memory_allocation);
} else {
dctx->ddictLocal = NULL;
}
dctx->ddict = dctx->ddictLocal; dctx->ddict = dctx->ddictLocal;
dctx->dictUses = ZSTD_use_indefinitely;
}
return 0; return 0;
} }
size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize) 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); 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) 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); 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) size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)
{ {
return ZSTD_DCtx_loadDictionary_advanced(dctx, prefix, prefixSize, ZSTD_dlm_byRef, 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) 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); return ZSTD_DCtx_refPrefix_advanced(dctx, prefix, prefixSize, ZSTD_dct_rawContent);
} }
/* ZSTD_initDStream_usingDict() : /* ZSTD_initDStream_usingDict() :
* return : expected size, aka ZSTD_FRAMEHEADERSIZE_PREFIX. * return : expected size, aka ZSTD_FRAMEHEADERSIZE_PREFIX.
* this function cannot fail */ * this function cannot fail */
size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize) size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize)
{ {
DEBUGLOG(4, "ZSTD_initDStream_usingDict"); DEBUGLOG(4, "ZSTD_initDStream_usingDict");
zds->streamStage = zdss_init; FORWARD_IF_ERROR( ZSTD_DCtx_reset(zds, ZSTD_reset_session_only) );
zds->noForwardProgress = 0; FORWARD_IF_ERROR( ZSTD_DCtx_loadDictionary(zds, dict, dictSize) );
CHECK_F( ZSTD_DCtx_loadDictionary(zds, dict, dictSize) );
return ZSTD_FRAMEHEADERSIZE_PREFIX; return ZSTD_FRAMEHEADERSIZE_PREFIX;
} }
/* note : this variant can't fail */ /* note : this variant can't fail */
size_t ZSTD_initDStream(ZSTD_DStream* zds) size_t ZSTD_initDStream(ZSTD_DStream* zds)
{ {
DEBUGLOG(4, "ZSTD_initDStream"); DEBUGLOG(4, "ZSTD_initDStream");
return ZSTD_initDStream_usingDict(zds, NULL, 0); return ZSTD_initDStream_usingDDict(zds, NULL);
} }
/* ZSTD_initDStream_usingDDict() : /* ZSTD_initDStream_usingDDict() :
* ddict will just be referenced, and must outlive decompression session * ddict will just be referenced, and must outlive decompression session
* this function cannot fail */ * this function cannot fail */
size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* dctx, const ZSTD_DDict* ddict) size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* dctx, const ZSTD_DDict* ddict)
{ {
size_t const initResult = ZSTD_initDStream(dctx); FORWARD_IF_ERROR( ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only) );
dctx->ddict = ddict; FORWARD_IF_ERROR( ZSTD_DCtx_refDDict(dctx, ddict) );
return initResult; return ZSTD_FRAMEHEADERSIZE_PREFIX;
} }
/* ZSTD_resetDStream() : /* ZSTD_resetDStream() :
* return : expected size, aka ZSTD_FRAMEHEADERSIZE_PREFIX. * return : expected size, aka ZSTD_FRAMEHEADERSIZE_PREFIX.
* this function cannot fail */ * this function cannot fail */
size_t ZSTD_resetDStream(ZSTD_DStream* dctx) size_t ZSTD_resetDStream(ZSTD_DStream* dctx)
{ {
DEBUGLOG(4, "ZSTD_resetDStream"); FORWARD_IF_ERROR(ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only));
dctx->streamStage = zdss_loadHeader;
dctx->lhSize = dctx->inPos = dctx->outStart = dctx->outEnd = 0;
dctx->legacyVersion = 0;
dctx->hostageByte = 0;
return ZSTD_FRAMEHEADERSIZE_PREFIX; return ZSTD_FRAMEHEADERSIZE_PREFIX;
} }
size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict) size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
{ {
if (dctx->streamStage != zdss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
ZSTD_clearDict(dctx);
if (ddict) {
dctx->ddict = ddict; dctx->ddict = ddict;
dctx->dictUses = ZSTD_use_indefinitely;
}
return 0; return 0;
} }
/* ZSTD_DCtx_setMaxWindowSize() : /* ZSTD_DCtx_setMaxWindowSize() :
* note : no direct equivalence in ZSTD_DCtx_setParameter, * note : no direct equivalence in ZSTD_DCtx_setParameter,
* since this version sets windowSize, and the other sets windowLog */ * since this version sets windowSize, and the other sets windowLog */
size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize) size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize)
{ {
ZSTD_bounds const bounds = ZSTD_dParam_getBounds(ZSTD_d_windowLogMax); ZSTD_bounds const bounds = ZSTD_dParam_getBounds(ZSTD_d_windowLogMax);
size_t const min = (size_t)1 << bounds.lowerBound; size_t const min = (size_t)1 << bounds.lowerBound;
size_t const max = (size_t)1 << bounds.upperBound; size_t const max = (size_t)1 << bounds.upperBound;
if (dctx->streamStage != zdss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
if (maxWindowSize < min) return ERROR(parameter_outOfBound); RETURN_ERROR_IF(maxWindowSize < min, parameter_outOfBound);
if (maxWindowSize > max) return ERROR(parameter_outOfBound); RETURN_ERROR_IF(maxWindowSize > max, parameter_outOfBound);
dctx->maxWindowSize = maxWindowSize; dctx->maxWindowSize = maxWindowSize;
return 0; return 0;
} }
size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format) size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format)
{ {
return ZSTD_DCtx_setParameter(dctx, ZSTD_d_format, format); return ZSTD_DCtx_setParameter(dctx, ZSTD_d_format, format);
} }
ZSTD_bounds const bounds = ZSTD_dParam_getBounds(dParam); ZSTD_bounds const bounds = ZSTD_dParam_getBounds(dParam);
if (ZSTD_isError(bounds.error)) return 0; if (ZSTD_isError(bounds.error)) return 0;
if (value < bounds.lowerBound) return 0; if (value < bounds.lowerBound) return 0;
if (value > bounds.upperBound) return 0; if (value > bounds.upperBound) return 0;
return 1; return 1;
} }
#define CHECK_DBOUNDS(p,v) { \ #define CHECK_DBOUNDS(p,v) { \
if (!ZSTD_dParam_withinBounds(p, v)) \ RETURN_ERROR_IF(!ZSTD_dParam_withinBounds(p, v), parameter_outOfBound); \
return ERROR(parameter_outOfBound); \
} }
size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value) size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value)
{ {
if (dctx->streamStage != zdss_init) return ERROR(stage_wrong); RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
switch(dParam) { switch(dParam) {
case ZSTD_d_windowLogMax: case ZSTD_d_windowLogMax:
if (value == 0) value = ZSTD_WINDOWLOG_LIMIT_DEFAULT;
CHECK_DBOUNDS(ZSTD_d_windowLogMax, value); CHECK_DBOUNDS(ZSTD_d_windowLogMax, value);
dctx->maxWindowSize = ((size_t)1) << value; dctx->maxWindowSize = ((size_t)1) << value;
return 0; return 0;
case ZSTD_d_format: case ZSTD_d_format:
CHECK_DBOUNDS(ZSTD_d_format, value); CHECK_DBOUNDS(ZSTD_d_format, value);
dctx->format = (ZSTD_format_e)value; dctx->format = (ZSTD_format_e)value;
return 0; return 0;
default:; default:;
} }
return ERROR(parameter_unsupported); RETURN_ERROR(parameter_unsupported);
} }
size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset) size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset)
{ {
if ( (reset == ZSTD_reset_session_only) if ( (reset == ZSTD_reset_session_only)
|| (reset == ZSTD_reset_session_and_parameters) ) { || (reset == ZSTD_reset_session_and_parameters) ) {
(void)ZSTD_initDStream(dctx); dctx->streamStage = zdss_init;
dctx->noForwardProgress = 0;
} }
if ( (reset == ZSTD_reset_parameters) if ( (reset == ZSTD_reset_parameters)
|| (reset == ZSTD_reset_session_and_parameters) ) { || (reset == ZSTD_reset_session_and_parameters) ) {
if (dctx->streamStage != zdss_init) RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
return ERROR(stage_wrong); ZSTD_clearDict(dctx);
dctx->format = ZSTD_f_zstd1; dctx->format = ZSTD_f_zstd1;
dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT; dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
} }
return 0; return 0;
} }
size_t ZSTD_sizeof_DStream(const ZSTD_DStream* dctx) size_t ZSTD_sizeof_DStream(const ZSTD_DStream* dctx)
{ {
return ZSTD_sizeof_DCtx(dctx); return ZSTD_sizeof_DCtx(dctx);
} }
size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize) size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
{ {
size_t const blockSize = (size_t) MIN(windowSize, ZSTD_BLOCKSIZE_MAX); 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 neededRBSize = windowSize + blockSize + (WILDCOPY_OVERLENGTH * 2);
unsigned long long const neededSize = MIN(frameContentSize, neededRBSize); unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);
size_t const minRBSize = (size_t) neededSize; size_t const minRBSize = (size_t) neededSize;
if ((unsigned long long)minRBSize != neededSize) return ERROR(frameParameter_windowTooLarge); RETURN_ERROR_IF((unsigned long long)minRBSize != neededSize,
frameParameter_windowTooLarge);
return minRBSize; return minRBSize;
} }
size_t ZSTD_estimateDStreamSize(size_t windowSize) size_t ZSTD_estimateDStreamSize(size_t windowSize)
{ {
size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX); size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
size_t const inBuffSize = blockSize; /* no block can be larger */ size_t const inBuffSize = blockSize; /* no block can be larger */
size_t const outBuffSize = ZSTD_decodingBufferSize_min(windowSize, ZSTD_CONTENTSIZE_UNKNOWN); size_t const outBuffSize = ZSTD_decodingBufferSize_min(windowSize, ZSTD_CONTENTSIZE_UNKNOWN);
return ZSTD_estimateDCtxSize() + inBuffSize + outBuffSize; return ZSTD_estimateDCtxSize() + inBuffSize + outBuffSize;
} }
size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize) 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) */ U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX; /* note : should be user-selectable, but requires an additional parameter (or a dctx) */
ZSTD_frameHeader zfh; ZSTD_frameHeader zfh;
size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize); size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize);
if (ZSTD_isError(err)) return err; if (ZSTD_isError(err)) return err;
if (err>0) return ERROR(srcSize_wrong); RETURN_ERROR_IF(err>0, srcSize_wrong);
if (zfh.windowSize > windowSizeMax) RETURN_ERROR_IF(zfh.windowSize > windowSizeMax,
return ERROR(frameParameter_windowTooLarge); frameParameter_windowTooLarge);
return ZSTD_estimateDStreamSize((size_t)zfh.windowSize); return ZSTD_estimateDStreamSize((size_t)zfh.windowSize);
} }
/* ***** Decompression ***** */ /* ***** Decompression ***** */
MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize) MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize)
{ {
const char* const iend = (const char*)(input->src) + input->size; const char* const iend = (const char*)(input->src) + input->size;
const char* ip = istart; const char* ip = istart;
char* const ostart = (char*)(output->dst) + output->pos; char* const ostart = (char*)(output->dst) + output->pos;
char* const oend = (char*)(output->dst) + output->size; char* const oend = (char*)(output->dst) + output->size;
char* op = ostart; char* op = ostart;
U32 someMoreWork = 1; U32 someMoreWork = 1;
DEBUGLOG(5, "ZSTD_decompressStream"); DEBUGLOG(5, "ZSTD_decompressStream");
if (input->pos > input->size) { /* forbidden */ RETURN_ERROR_IF(
DEBUGLOG(5, "in: pos: %u vs size: %u", input->pos > input->size,
srcSize_wrong,
"forbidden. in: pos: %u vs size: %u",
(U32)input->pos, (U32)input->size); (U32)input->pos, (U32)input->size);
return ERROR(srcSize_wrong); RETURN_ERROR_IF(
} output->pos > output->size,
if (output->pos > output->size) { /* forbidden */ dstSize_tooSmall,
DEBUGLOG(5, "out: pos: %u vs size: %u", "forbidden. out: pos: %u vs size: %u",
(U32)output->pos, (U32)output->size); (U32)output->pos, (U32)output->size);
return ERROR(dstSize_tooSmall);
}
DEBUGLOG(5, "input size : %u", (U32)(input->size - input->pos)); DEBUGLOG(5, "input size : %u", (U32)(input->size - input->pos));
while (someMoreWork) { while (someMoreWork) {
switch(zds->streamStage) switch(zds->streamStage)
{ {
case zdss_init : case zdss_init :
DEBUGLOG(5, "stage zdss_init => transparent reset "); DEBUGLOG(5, "stage zdss_init => transparent reset ");
ZSTD_resetDStream(zds); /* transparent reset on starting decoding a new frame */ zds->streamStage = zdss_loadHeader;
zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
zds->legacyVersion = 0;
zds->hostageByte = 0;
/* fall-through */ /* fall-through */
case zdss_loadHeader : case zdss_loadHeader :
DEBUGLOG(5, "stage zdss_loadHeader (srcSize : %u)", (U32)(iend - ip)); DEBUGLOG(5, "stage zdss_loadHeader (srcSize : %u)", (U32)(iend - ip));
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1) #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
if (zds->legacyVersion) { if (zds->legacyVersion) {
/* legacy support is incompatible with static dctx */ RETURN_ERROR_IF(zds->staticSize, memory_allocation,
if (zds->staticSize) return ERROR(memory_allocation); "legacy support is incompatible with static dctx");
{ size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, zds->legacyVersion, output, input); { size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, zds->legacyVersion, output, input);
if (hint==0) zds->streamStage = zdss_init; if (hint==0) zds->streamStage = zdss_init;
return hint; return hint;
} } } }
#endif #endif
{ size_t const hSize = ZSTD_getFrameHeader_advanced(&zds->fParams, zds->headerBuffer, zds->lhSize, zds->format); { size_t const hSize = ZSTD_getFrameHeader_advanced(&zds->fParams, zds->headerBuffer, zds->lhSize, zds->format);
DEBUGLOG(5, "header size : %u", (U32)hSize); DEBUGLOG(5, "header size : %u", (U32)hSize);
if (ZSTD_isError(hSize)) { if (ZSTD_isError(hSize)) {
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1) #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
U32 const legacyVersion = ZSTD_isLegacy(istart, iend-istart); U32 const legacyVersion = ZSTD_isLegacy(istart, iend-istart);
if (legacyVersion) { if (legacyVersion) {
const void* const dict = zds->ddict ? ZSTD_DDict_dictContent(zds->ddict) : NULL; ZSTD_DDict const* const ddict = ZSTD_getDDict(zds);
size_t const dictSize = zds->ddict ? ZSTD_DDict_dictSize(zds->ddict) : 0; 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); DEBUGLOG(5, "ZSTD_decompressStream: detected legacy version v0.%u", legacyVersion);
/* legacy support is incompatible with static dctx */ RETURN_ERROR_IF(zds->staticSize, memory_allocation,
if (zds->staticSize) return ERROR(memory_allocation); "legacy support is incompatible with static dctx");
CHECK_F(ZSTD_initLegacyStream(&zds->legacyContext, FORWARD_IF_ERROR(ZSTD_initLegacyStream(&zds->legacyContext,
zds->previousLegacyVersion, legacyVersion, zds->previousLegacyVersion, legacyVersion,
dict, dictSize)); dict, dictSize));
zds->legacyVersion = zds->previousLegacyVersion = legacyVersion; zds->legacyVersion = zds->previousLegacyVersion = legacyVersion;
{ size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, legacyVersion, output, input); { size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, legacyVersion, output, input);
if (hint==0) zds->streamStage = zdss_init; /* or stay in stage zdss_loadHeader */ if (hint==0) zds->streamStage = zdss_init; /* or stay in stage zdss_loadHeader */
return hint; return hint;
} } } }
#endif #endif
} } } }
/* check for single-pass mode opportunity */ /* check for single-pass mode opportunity */
if (zds->fParams.frameContentSize && zds->fParams.windowSize /* skippable frame if == 0 */ if (zds->fParams.frameContentSize && zds->fParams.windowSize /* skippable frame if == 0 */
&& (U64)(size_t)(oend-op) >= zds->fParams.frameContentSize) { && (U64)(size_t)(oend-op) >= zds->fParams.frameContentSize) {
size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend-istart); size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend-istart);
if (cSize <= (size_t)(iend-istart)) { if (cSize <= (size_t)(iend-istart)) {
/* shortcut : using single-pass mode */ /* shortcut : using single-pass mode */
size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, oend-op, istart, cSize, zds->ddict); size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, oend-op, istart, cSize, ZSTD_getDDict(zds));
if (ZSTD_isError(decompressedSize)) return decompressedSize; if (ZSTD_isError(decompressedSize)) return decompressedSize;
DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()") DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()")
ip = istart + cSize; ip = istart + cSize;
op += decompressedSize; op += decompressedSize;
zds->expected = 0; zds->expected = 0;
zds->streamStage = zdss_init; zds->streamStage = zdss_init;
someMoreWork = 0; someMoreWork = 0;
break; break;
} } } }
/* Consume header (see ZSTDds_decodeFrameHeader) */ /* Consume header (see ZSTDds_decodeFrameHeader) */
DEBUGLOG(4, "Consume header"); DEBUGLOG(4, "Consume header");
CHECK_F(ZSTD_decompressBegin_usingDDict(zds, zds->ddict)); FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(zds, ZSTD_getDDict(zds)));
if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */ if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */
zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE); zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE);
zds->stage = ZSTDds_skipFrame; zds->stage = ZSTDds_skipFrame;
} else { } else {
CHECK_F(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize)); FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize));
zds->expected = ZSTD_blockHeaderSize; zds->expected = ZSTD_blockHeaderSize;
zds->stage = ZSTDds_decodeBlockHeader; zds->stage = ZSTDds_decodeBlockHeader;
} }
/* control buffer memory usage */ /* control buffer memory usage */
DEBUGLOG(4, "Control max memory usage (%u KB <= max %u KB)", DEBUGLOG(4, "Control max memory usage (%u KB <= max %u KB)",
(U32)(zds->fParams.windowSize >>10), (U32)(zds->fParams.windowSize >>10),
(U32)(zds->maxWindowSize >> 10) ); (U32)(zds->maxWindowSize >> 10) );
zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN); zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
if (zds->fParams.windowSize > zds->maxWindowSize) return ERROR(frameParameter_windowTooLarge); RETURN_ERROR_IF(zds->fParams.windowSize > zds->maxWindowSize,
frameParameter_windowTooLarge);
/* Adapt buffer sizes to frame header instructions */ /* Adapt buffer sizes to frame header instructions */
{ size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */); { size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);
size_t const neededOutBuffSize = ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize); size_t const neededOutBuffSize = ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize);
if ((zds->inBuffSize < neededInBuffSize) || (zds->outBuffSize < neededOutBuffSize)) { if ((zds->inBuffSize < neededInBuffSize) || (zds->outBuffSize < neededOutBuffSize)) {
size_t const bufferSize = neededInBuffSize + neededOutBuffSize; size_t const bufferSize = neededInBuffSize + neededOutBuffSize;
DEBUGLOG(4, "inBuff : from %u to %u", DEBUGLOG(4, "inBuff : from %u to %u",
(U32)zds->inBuffSize, (U32)neededInBuffSize); (U32)zds->inBuffSize, (U32)neededInBuffSize);
DEBUGLOG(4, "outBuff : from %u to %u", DEBUGLOG(4, "outBuff : from %u to %u",
(U32)zds->outBuffSize, (U32)neededOutBuffSize); (U32)zds->outBuffSize, (U32)neededOutBuffSize);
if (zds->staticSize) { /* static DCtx */ if (zds->staticSize) { /* static DCtx */
DEBUGLOG(4, "staticSize : %u", (U32)zds->staticSize); DEBUGLOG(4, "staticSize : %u", (U32)zds->staticSize);
assert(zds->staticSize >= sizeof(ZSTD_DCtx)); /* controlled at init */ assert(zds->staticSize >= sizeof(ZSTD_DCtx)); /* controlled at init */
if (bufferSize > zds->staticSize - sizeof(ZSTD_DCtx)) RETURN_ERROR_IF(
return ERROR(memory_allocation); bufferSize > zds->staticSize - sizeof(ZSTD_DCtx),
memory_allocation);
} else { } else {
ZSTD_free(zds->inBuff, zds->customMem); ZSTD_free(zds->inBuff, zds->customMem);
zds->inBuffSize = 0; zds->inBuffSize = 0;
zds->outBuffSize = 0; zds->outBuffSize = 0;
zds->inBuff = (char*)ZSTD_malloc(bufferSize, zds->customMem); zds->inBuff = (char*)ZSTD_malloc(bufferSize, zds->customMem);
if (zds->inBuff == NULL) return ERROR(memory_allocation); RETURN_ERROR_IF(zds->inBuff == NULL, memory_allocation);
} }
zds->inBuffSize = neededInBuffSize; zds->inBuffSize = neededInBuffSize;
zds->outBuff = zds->inBuff + zds->inBuffSize; zds->outBuff = zds->inBuff + zds->inBuffSize;
zds->outBuffSize = neededOutBuffSize; zds->outBuffSize = neededOutBuffSize;
} } } }
zds->streamStage = zdss_read; zds->streamStage = zdss_read;
/* fall-through */ /* fall-through */
case zdss_load: case zdss_load:
{ size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds); { size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds);
size_t const toLoad = neededInSize - zds->inPos; size_t const toLoad = neededInSize - zds->inPos;
int const isSkipFrame = ZSTD_isSkipFrame(zds); int const isSkipFrame = ZSTD_isSkipFrame(zds);
size_t loadedSize; size_t loadedSize;
if (isSkipFrame) { if (isSkipFrame) {
loadedSize = MIN(toLoad, (size_t)(iend-ip)); loadedSize = MIN(toLoad, (size_t)(iend-ip));
} else { } else {
if (toLoad > zds->inBuffSize - zds->inPos) return ERROR(corruption_detected); /* should never happen */ RETURN_ERROR_IF(toLoad > zds->inBuffSize - zds->inPos,
corruption_detected,
"should never happen");
loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend-ip); loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend-ip);
} }
ip += loadedSize; ip += loadedSize;
zds->inPos += loadedSize; zds->inPos += loadedSize;
if (loadedSize < toLoad) { someMoreWork = 0; break; } /* not enough input, wait for more */ if (loadedSize < toLoad) { someMoreWork = 0; break; } /* not enough input, wait for more */
/* decode loaded input */ /* decode loaded input */
{ size_t const decodedSize = ZSTD_decompressContinue(zds, { size_t const decodedSize = ZSTD_decompressContinue(zds,
break; break;
} } } }
/* cannot complete flush */ /* cannot complete flush */
someMoreWork = 0; someMoreWork = 0;
break; break;
default: default:
assert(0); /* impossible */ assert(0); /* impossible */
return ERROR(GENERIC); /* some compiler require default to do something */ RETURN_ERROR(GENERIC); /* some compiler require default to do something */
} } } }
/* result */ /* result */
input->pos = (size_t)(ip - (const char*)(input->src)); input->pos = (size_t)(ip - (const char*)(input->src));
output->pos = (size_t)(op - (char*)(output->dst)); output->pos = (size_t)(op - (char*)(output->dst));
if ((ip==istart) && (op==ostart)) { /* no forward progress */ if ((ip==istart) && (op==ostart)) { /* no forward progress */
zds->noForwardProgress ++; zds->noForwardProgress ++;
if (zds->noForwardProgress >= ZSTD_NO_FORWARD_PROGRESS_MAX) { if (zds->noForwardProgress >= ZSTD_NO_FORWARD_PROGRESS_MAX) {
if (op==oend) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(op==oend, dstSize_tooSmall);
if (ip==iend) return ERROR(srcSize_wrong); RETURN_ERROR_IF(ip==iend, srcSize_wrong);
assert(0); assert(0);
} }
} else { } else {
zds->noForwardProgress = 0; zds->noForwardProgress = 0;
} }
{ size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds); { size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds);
if (!nextSrcSizeHint) { /* frame fully decoded */ if (!nextSrcSizeHint) { /* frame fully decoded */
if (zds->outEnd == zds->outStart) { /* output fully flushed */ if (zds->outEnd == zds->outStart) { /* output fully flushed */

contrib/python-zstandard/zstd/decompress/zstd_decompress_block.c

* Block decoding * Block decoding
***************************************************************/ ***************************************************************/
/*! ZSTD_getcBlockSize() : /*! ZSTD_getcBlockSize() :
* Provides the size of compressed block from block header `src` */ * Provides the size of compressed block from block header `src` */
size_t ZSTD_getcBlockSize(const void* src, size_t srcSize, size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
blockProperties_t* bpPtr) blockProperties_t* bpPtr)
{ {
if (srcSize < ZSTD_blockHeaderSize) return ERROR(srcSize_wrong); RETURN_ERROR_IF(srcSize < ZSTD_blockHeaderSize, srcSize_wrong);
{ U32 const cBlockHeader = MEM_readLE24(src); { U32 const cBlockHeader = MEM_readLE24(src);
U32 const cSize = cBlockHeader >> 3; U32 const cSize = cBlockHeader >> 3;
bpPtr->lastBlock = cBlockHeader & 1; bpPtr->lastBlock = cBlockHeader & 1;
bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3); bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);
bpPtr->origSize = cSize; /* only useful for RLE */ bpPtr->origSize = cSize; /* only useful for RLE */
if (bpPtr->blockType == bt_rle) return 1; if (bpPtr->blockType == bt_rle) return 1;
if (bpPtr->blockType == bt_reserved) return ERROR(corruption_detected); RETURN_ERROR_IF(bpPtr->blockType == bt_reserved, corruption_detected);
return cSize; return cSize;
} }
} }
/* Hidden declaration for fullbench */ /* Hidden declaration for fullbench */
size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx, size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
const void* src, size_t srcSize); const void* src, size_t srcSize);
/*! ZSTD_decodeLiteralsBlock() : /*! ZSTD_decodeLiteralsBlock() :
* @return : nb of bytes read from src (< srcSize ) * @return : nb of bytes read from src (< srcSize )
* note : symbol not declared but exposed for fullbench */ * note : symbol not declared but exposed for fullbench */
size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx, size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
const void* src, size_t srcSize) /* note : srcSize < BLOCKSIZE */ const void* src, size_t srcSize) /* note : srcSize < BLOCKSIZE */
{ {
if (srcSize < MIN_CBLOCK_SIZE) return ERROR(corruption_detected); DEBUGLOG(5, "ZSTD_decodeLiteralsBlock");
RETURN_ERROR_IF(srcSize < MIN_CBLOCK_SIZE, corruption_detected);
{ const BYTE* const istart = (const BYTE*) src; { const BYTE* const istart = (const BYTE*) src;
symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3); symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
switch(litEncType) switch(litEncType)
{ {
case set_repeat: case set_repeat:
if (dctx->litEntropy==0) return ERROR(dictionary_corrupted); DEBUGLOG(5, "set_repeat flag : re-using stats from previous compressed literals block");
RETURN_ERROR_IF(dctx->litEntropy==0, dictionary_corrupted);
/* fall-through */ /* fall-through */
case set_compressed: case set_compressed:
if (srcSize < 5) return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3 */ 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; { size_t lhSize, litSize, litCSize;
U32 singleStream=0; U32 singleStream=0;
U32 const lhlCode = (istart[0] >> 2) & 3; U32 const lhlCode = (istart[0] >> 2) & 3;
U32 const lhc = MEM_readLE32(istart); U32 const lhc = MEM_readLE32(istart);
size_t hufSuccess; size_t hufSuccess;
switch(lhlCode) switch(lhlCode)
{ {
case 0: case 1: default: /* note : default is impossible, since lhlCode into [0..3] */ case 0: case 1: default: /* note : default is impossible, since lhlCode into [0..3] */
/* 2 - 2 - 10 - 10 */ /* 2 - 2 - 10 - 10 */
singleStream = !lhlCode; singleStream = !lhlCode;
lhSize = 3; lhSize = 3;
litSize = (lhc >> 4) & 0x3FF; litSize = (lhc >> 4) & 0x3FF;
litCSize = (lhc >> 14) & 0x3FF; litCSize = (lhc >> 14) & 0x3FF;
break; break;
case 2: case 2:
/* 2 - 2 - 14 - 14 */ /* 2 - 2 - 14 - 14 */
lhSize = 4; lhSize = 4;
litSize = (lhc >> 4) & 0x3FFF; litSize = (lhc >> 4) & 0x3FFF;
litCSize = lhc >> 18; litCSize = lhc >> 18;
break; break;
case 3: case 3:
/* 2 - 2 - 18 - 18 */ /* 2 - 2 - 18 - 18 */
lhSize = 5; lhSize = 5;
litSize = (lhc >> 4) & 0x3FFFF; litSize = (lhc >> 4) & 0x3FFFF;
litCSize = (lhc >> 22) + (istart[4] << 10); litCSize = (lhc >> 22) + ((size_t)istart[4] << 10);
break; break;
} }
if (litSize > ZSTD_BLOCKSIZE_MAX) return ERROR(corruption_detected); RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected);
if (litCSize + lhSize > srcSize) return ERROR(corruption_detected); RETURN_ERROR_IF(litCSize + lhSize > srcSize, corruption_detected);
/* prefetch huffman table if cold */ /* prefetch huffman table if cold */
if (dctx->ddictIsCold && (litSize > 768 /* heuristic */)) { if (dctx->ddictIsCold && (litSize > 768 /* heuristic */)) {
PREFETCH_AREA(dctx->HUFptr, sizeof(dctx->entropy.hufTable)); PREFETCH_AREA(dctx->HUFptr, sizeof(dctx->entropy.hufTable));
} }
if (litEncType==set_repeat) { if (litEncType==set_repeat) {
if (singleStream) { if (singleStream) {
} else { } else {
hufSuccess = HUF_decompress4X_hufOnly_wksp_bmi2( hufSuccess = HUF_decompress4X_hufOnly_wksp_bmi2(
dctx->entropy.hufTable, dctx->litBuffer, litSize, dctx->entropy.hufTable, dctx->litBuffer, litSize,
istart+lhSize, litCSize, dctx->workspace, istart+lhSize, litCSize, dctx->workspace,
sizeof(dctx->workspace), dctx->bmi2); sizeof(dctx->workspace), dctx->bmi2);
} }
} }
if (HUF_isError(hufSuccess)) return ERROR(corruption_detected); RETURN_ERROR_IF(HUF_isError(hufSuccess), corruption_detected);
dctx->litPtr = dctx->litBuffer; dctx->litPtr = dctx->litBuffer;
dctx->litSize = litSize; dctx->litSize = litSize;
dctx->litEntropy = 1; dctx->litEntropy = 1;
if (litEncType==set_compressed) dctx->HUFptr = dctx->entropy.hufTable; if (litEncType==set_compressed) dctx->HUFptr = dctx->entropy.hufTable;
memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH); memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
return litCSize + lhSize; return litCSize + lhSize;
} }
break; break;
case 3: case 3:
lhSize = 3; lhSize = 3;
litSize = MEM_readLE24(istart) >> 4; litSize = MEM_readLE24(istart) >> 4;
break; break;
} }
if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */ if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */
if (litSize+lhSize > srcSize) return ERROR(corruption_detected); RETURN_ERROR_IF(litSize+lhSize > srcSize, corruption_detected);
memcpy(dctx->litBuffer, istart+lhSize, litSize); memcpy(dctx->litBuffer, istart+lhSize, litSize);
dctx->litPtr = dctx->litBuffer; dctx->litPtr = dctx->litBuffer;
dctx->litSize = litSize; dctx->litSize = litSize;
memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH); memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
return lhSize+litSize; return lhSize+litSize;
} }
/* direct reference into compressed stream */ /* direct reference into compressed stream */
dctx->litPtr = istart+lhSize; dctx->litPtr = istart+lhSize;
break; break;
case 1: case 1:
lhSize = 2; lhSize = 2;
litSize = MEM_readLE16(istart) >> 4; litSize = MEM_readLE16(istart) >> 4;
break; break;
case 3: case 3:
lhSize = 3; lhSize = 3;
litSize = MEM_readLE24(istart) >> 4; litSize = MEM_readLE24(istart) >> 4;
if (srcSize<4) return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4 */ RETURN_ERROR_IF(srcSize<4, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4");
break; break;
} }
if (litSize > ZSTD_BLOCKSIZE_MAX) return ERROR(corruption_detected); RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected);
memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH); memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);
dctx->litPtr = dctx->litBuffer; dctx->litPtr = dctx->litBuffer;
dctx->litSize = litSize; dctx->litSize = litSize;
return lhSize+1; return lhSize+1;
} }
default: default:
return ERROR(corruption_detected); /* impossible */ RETURN_ERROR(corruption_detected, "impossible");
} }
} }
} }
/* Default FSE distribution tables. /* Default FSE distribution tables.
* These are pre-calculated FSE decoding tables using default distributions as defined in specification : * 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 * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#default-distributions
* They were generated programmatically with following method : * They were generated programmatically with following method :
{ S16 const largeLimit= (S16)(1 << (tableLog-1)); { S16 const largeLimit= (S16)(1 << (tableLog-1));
U32 s; U32 s;
for (s=0; s<maxSV1; s++) { for (s=0; s<maxSV1; s++) {
if (normalizedCounter[s]==-1) { if (normalizedCounter[s]==-1) {
tableDecode[highThreshold--].baseValue = s; tableDecode[highThreshold--].baseValue = s;
symbolNext[s] = 1; symbolNext[s] = 1;
} else { } else {
if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0; if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;
symbolNext[s] = normalizedCounter[s]; assert(normalizedCounter[s]>=0);
symbolNext[s] = (U16)normalizedCounter[s];
} } } } } }
memcpy(dt, &DTableH, sizeof(DTableH)); memcpy(dt, &DTableH, sizeof(DTableH));
} }
/* Spread symbols */ /* Spread symbols */
{ U32 const tableMask = tableSize-1; { U32 const tableMask = tableSize-1;
U32 const step = FSE_TABLESTEP(tableSize); U32 const step = FSE_TABLESTEP(tableSize);
U32 s, position = 0; U32 s, position = 0;
const void* src, size_t srcSize, const void* src, size_t srcSize,
const U32* baseValue, const U32* nbAdditionalBits, const U32* baseValue, const U32* nbAdditionalBits,
const ZSTD_seqSymbol* defaultTable, U32 flagRepeatTable, const ZSTD_seqSymbol* defaultTable, U32 flagRepeatTable,
int ddictIsCold, int nbSeq) int ddictIsCold, int nbSeq)
{ {
switch(type) switch(type)
{ {
case set_rle : case set_rle :
if (!srcSize) return ERROR(srcSize_wrong); RETURN_ERROR_IF(!srcSize, srcSize_wrong);
if ( (*(const BYTE*)src) > max) return ERROR(corruption_detected); RETURN_ERROR_IF((*(const BYTE*)src) > max, corruption_detected);
{ U32 const symbol = *(const BYTE*)src; { U32 const symbol = *(const BYTE*)src;
U32 const baseline = baseValue[symbol]; U32 const baseline = baseValue[symbol];
U32 const nbBits = nbAdditionalBits[symbol]; U32 const nbBits = nbAdditionalBits[symbol];
ZSTD_buildSeqTable_rle(DTableSpace, baseline, nbBits); ZSTD_buildSeqTable_rle(DTableSpace, baseline, nbBits);
} }
*DTablePtr = DTableSpace; *DTablePtr = DTableSpace;
return 1; return 1;
case set_basic : case set_basic :
*DTablePtr = defaultTable; *DTablePtr = defaultTable;
return 0; return 0;
case set_repeat: case set_repeat:
if (!flagRepeatTable) return ERROR(corruption_detected); RETURN_ERROR_IF(!flagRepeatTable, corruption_detected);
/* prefetch FSE table if used */ /* prefetch FSE table if used */
if (ddictIsCold && (nbSeq > 24 /* heuristic */)) { if (ddictIsCold && (nbSeq > 24 /* heuristic */)) {
const void* const pStart = *DTablePtr; const void* const pStart = *DTablePtr;
size_t const pSize = sizeof(ZSTD_seqSymbol) * (SEQSYMBOL_TABLE_SIZE(maxLog)); size_t const pSize = sizeof(ZSTD_seqSymbol) * (SEQSYMBOL_TABLE_SIZE(maxLog));
PREFETCH_AREA(pStart, pSize); PREFETCH_AREA(pStart, pSize);
} }
return 0; return 0;
case set_compressed : case set_compressed :
{ unsigned tableLog; { unsigned tableLog;
S16 norm[MaxSeq+1]; S16 norm[MaxSeq+1];
size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize); size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);
if (FSE_isError(headerSize)) return ERROR(corruption_detected); RETURN_ERROR_IF(FSE_isError(headerSize), corruption_detected);
if (tableLog > maxLog) return ERROR(corruption_detected); RETURN_ERROR_IF(tableLog > maxLog, corruption_detected);
ZSTD_buildFSETable(DTableSpace, norm, max, baseValue, nbAdditionalBits, tableLog); ZSTD_buildFSETable(DTableSpace, norm, max, baseValue, nbAdditionalBits, tableLog);
*DTablePtr = DTableSpace; *DTablePtr = DTableSpace;
return headerSize; return headerSize;
} }
default : /* impossible */ default :
assert(0); assert(0);
return ERROR(GENERIC); RETURN_ERROR(GENERIC, "impossible");
} }
} }
size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr, size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
const void* src, size_t srcSize) const void* src, size_t srcSize)
{ {
const BYTE* const istart = (const BYTE* const)src; const BYTE* const istart = (const BYTE* const)src;
const BYTE* const iend = istart + srcSize; const BYTE* const iend = istart + srcSize;
const BYTE* ip = istart; const BYTE* ip = istart;
int nbSeq; int nbSeq;
DEBUGLOG(5, "ZSTD_decodeSeqHeaders"); DEBUGLOG(5, "ZSTD_decodeSeqHeaders");
/* check */ /* check */
if (srcSize < MIN_SEQUENCES_SIZE) return ERROR(srcSize_wrong); RETURN_ERROR_IF(srcSize < MIN_SEQUENCES_SIZE, srcSize_wrong);
/* SeqHead */ /* SeqHead */
nbSeq = *ip++; nbSeq = *ip++;
if (!nbSeq) { if (!nbSeq) {
*nbSeqPtr=0; *nbSeqPtr=0;
if (srcSize != 1) return ERROR(srcSize_wrong); RETURN_ERROR_IF(srcSize != 1, srcSize_wrong);
return 1; return 1;
} }
if (nbSeq > 0x7F) { if (nbSeq > 0x7F) {
if (nbSeq == 0xFF) { if (nbSeq == 0xFF) {
if (ip+2 > iend) return ERROR(srcSize_wrong); RETURN_ERROR_IF(ip+2 > iend, srcSize_wrong);
nbSeq = MEM_readLE16(ip) + LONGNBSEQ, ip+=2; nbSeq = MEM_readLE16(ip) + LONGNBSEQ, ip+=2;
} else { } else {
if (ip >= iend) return ERROR(srcSize_wrong); RETURN_ERROR_IF(ip >= iend, srcSize_wrong);
nbSeq = ((nbSeq-0x80)<<8) + *ip++; nbSeq = ((nbSeq-0x80)<<8) + *ip++;
} }
} }
*nbSeqPtr = nbSeq; *nbSeqPtr = nbSeq;
/* FSE table descriptors */ /* FSE table descriptors */
if (ip+4 > iend) return ERROR(srcSize_wrong); /* minimum possible size */ RETURN_ERROR_IF(ip+1 > iend, srcSize_wrong); /* minimum possible size: 1 byte for symbol encoding types */
{ symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6); { symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3); symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3); symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
ip++; ip++;
/* Build DTables */ /* Build DTables */
{ size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr, { size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr,
LLtype, MaxLL, LLFSELog, LLtype, MaxLL, LLFSELog,
ip, iend-ip, ip, iend-ip,
LL_base, LL_bits, LL_base, LL_bits,
LL_defaultDTable, dctx->fseEntropy, LL_defaultDTable, dctx->fseEntropy,
dctx->ddictIsCold, nbSeq); dctx->ddictIsCold, nbSeq);
if (ZSTD_isError(llhSize)) return ERROR(corruption_detected); RETURN_ERROR_IF(ZSTD_isError(llhSize), corruption_detected);
ip += llhSize; ip += llhSize;
} }
{ size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr, { size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr,
OFtype, MaxOff, OffFSELog, OFtype, MaxOff, OffFSELog,
ip, iend-ip, ip, iend-ip,
OF_base, OF_bits, OF_base, OF_bits,
OF_defaultDTable, dctx->fseEntropy, OF_defaultDTable, dctx->fseEntropy,
dctx->ddictIsCold, nbSeq); dctx->ddictIsCold, nbSeq);
if (ZSTD_isError(ofhSize)) return ERROR(corruption_detected); RETURN_ERROR_IF(ZSTD_isError(ofhSize), corruption_detected);
ip += ofhSize; ip += ofhSize;
} }
{ size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr, { size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr,
MLtype, MaxML, MLFSELog, MLtype, MaxML, MLFSELog,
ip, iend-ip, ip, iend-ip,
ML_base, ML_bits, ML_base, ML_bits,
ML_defaultDTable, dctx->fseEntropy, ML_defaultDTable, dctx->fseEntropy,
dctx->ddictIsCold, nbSeq); dctx->ddictIsCold, nbSeq);
if (ZSTD_isError(mlhSize)) return ERROR(corruption_detected); RETURN_ERROR_IF(ZSTD_isError(mlhSize), corruption_detected);
ip += mlhSize; ip += mlhSize;
} }
} }
return ip-istart; return ip-istart;
} }
{ {
BYTE* const oLitEnd = op + sequence.litLength; BYTE* const oLitEnd = op + sequence.litLength;
size_t const sequenceLength = sequence.litLength + sequence.matchLength; size_t const sequenceLength = sequence.litLength + sequence.matchLength;
BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */ BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
const BYTE* const iLitEnd = *litPtr + sequence.litLength; const BYTE* const iLitEnd = *litPtr + sequence.litLength;
const BYTE* match = oLitEnd - sequence.offset; const BYTE* match = oLitEnd - sequence.offset;
/* check */ /* check */
if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must fit within dstBuffer */ RETURN_ERROR_IF(oMatchEnd>oend, dstSize_tooSmall, "last match must fit within dstBuffer");
if (iLitEnd > litLimit) return ERROR(corruption_detected); /* try to read beyond literal buffer */ RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "try to read beyond literal buffer");
/* copy literals */ /* copy literals */
while (op < oLitEnd) *op++ = *(*litPtr)++; while (op < oLitEnd) *op++ = *(*litPtr)++;
/* copy Match */ /* copy Match */
if (sequence.offset > (size_t)(oLitEnd - base)) { if (sequence.offset > (size_t)(oLitEnd - base)) {
/* offset beyond prefix */ /* offset beyond prefix */
if (sequence.offset > (size_t)(oLitEnd - vBase)) return ERROR(corruption_detected); RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - vBase),corruption_detected);
match = dictEnd - (base-match); match = dictEnd - (base-match);
if (match + sequence.matchLength <= dictEnd) { if (match + sequence.matchLength <= dictEnd) {
memmove(oLitEnd, match, sequence.matchLength); memmove(oLitEnd, match, sequence.matchLength);
return sequenceLength; return sequenceLength;
} }
/* span extDict & currentPrefixSegment */ /* span extDict & currentPrefixSegment */
{ size_t const length1 = dictEnd - match; { size_t const length1 = dictEnd - match;
memmove(oLitEnd, match, length1); memmove(oLitEnd, match, length1);
BYTE* const oLitEnd = op + sequence.litLength; BYTE* const oLitEnd = op + sequence.litLength;
size_t const sequenceLength = sequence.litLength + sequence.matchLength; size_t const sequenceLength = sequence.litLength + sequence.matchLength;
BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */ BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH; BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;
const BYTE* const iLitEnd = *litPtr + sequence.litLength; const BYTE* const iLitEnd = *litPtr + sequence.litLength;
const BYTE* match = oLitEnd - sequence.offset; const BYTE* match = oLitEnd - sequence.offset;
/* check */ /* check */
if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */ RETURN_ERROR_IF(oMatchEnd>oend, 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 */ 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); if (oLitEnd>oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, prefixStart, virtualStart, dictEnd);
/* copy Literals */ /* copy Literals */
ZSTD_copy8(op, *litPtr);
if (sequence.litLength > 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 */ ZSTD_wildcopy_16min(op, (*litPtr), sequence.litLength, ZSTD_no_overlap); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
else
ZSTD_copy8(op, *litPtr);
op = oLitEnd; op = oLitEnd;
*litPtr = iLitEnd; /* update for next sequence */ *litPtr = iLitEnd; /* update for next sequence */
/* copy Match */ /* copy Match */
if (sequence.offset > (size_t)(oLitEnd - prefixStart)) { if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {
/* offset beyond prefix -> go into extDict */ /* offset beyond prefix -> go into extDict */
if (sequence.offset > (size_t)(oLitEnd - virtualStart)) RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - virtualStart), corruption_detected);
return ERROR(corruption_detected);
match = dictEnd + (match - prefixStart); match = dictEnd + (match - prefixStart);
if (match + sequence.matchLength <= dictEnd) { if (match + sequence.matchLength <= dictEnd) {
memmove(oLitEnd, match, sequence.matchLength); memmove(oLitEnd, match, sequence.matchLength);
return sequenceLength; return sequenceLength;
} }
/* span extDict & currentPrefixSegment */ /* span extDict & currentPrefixSegment */
{ size_t const length1 = dictEnd - match; { size_t const length1 = dictEnd - match;
memmove(oLitEnd, match, length1); memmove(oLitEnd, match, length1);
match -= sub2; match -= sub2;
} else { } else {
ZSTD_copy8(op, match); ZSTD_copy8(op, match);
} }
op += 8; match += 8; op += 8; match += 8;
if (oMatchEnd > oend-(16-MINMATCH)) { if (oMatchEnd > oend-(16-MINMATCH)) {
if (op < oend_w) { if (op < oend_w) {
ZSTD_wildcopy(op, match, oend_w - op); ZSTD_wildcopy(op, match, oend_w - op, ZSTD_overlap_src_before_dst);
match += oend_w - op; match += oend_w - op;
op = oend_w; op = oend_w;
} }
while (op < oMatchEnd) *op++ = *match++; while (op < oMatchEnd) *op++ = *match++;
} else { } else {
ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */ ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8, ZSTD_overlap_src_before_dst); /* works even if matchLength < 8 */
} }
return sequenceLength; return sequenceLength;
} }
HINT_INLINE HINT_INLINE
size_t ZSTD_execSequenceLong(BYTE* op, size_t ZSTD_execSequenceLong(BYTE* op,
BYTE* const oend, seq_t sequence, BYTE* const oend, seq_t sequence,
const BYTE** litPtr, const BYTE* const litLimit, const BYTE** litPtr, const BYTE* const litLimit,
const BYTE* const prefixStart, const BYTE* const dictStart, const BYTE* const dictEnd) const BYTE* const prefixStart, const BYTE* const dictStart, const BYTE* const dictEnd)
{ {
BYTE* const oLitEnd = op + sequence.litLength; BYTE* const oLitEnd = op + sequence.litLength;
size_t const sequenceLength = sequence.litLength + sequence.matchLength; size_t const sequenceLength = sequence.litLength + sequence.matchLength;
BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */ BYTE* const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH; BYTE* const oend_w = oend - WILDCOPY_OVERLENGTH;
const BYTE* const iLitEnd = *litPtr + sequence.litLength; const BYTE* const iLitEnd = *litPtr + sequence.litLength;
const BYTE* match = sequence.match; const BYTE* match = sequence.match;
/* check */ /* check */
if (oMatchEnd > oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */ RETURN_ERROR_IF(oMatchEnd > oend, 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 */ 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); if (oLitEnd > oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, prefixStart, dictStart, dictEnd);
/* copy Literals */ /* copy Literals */
ZSTD_copy8(op, *litPtr); /* note : op <= oLitEnd <= oend_w == oend - 8 */
if (sequence.litLength > 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 */ ZSTD_wildcopy_16min(op, *litPtr, sequence.litLength, ZSTD_no_overlap); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
else
ZSTD_copy8(op, *litPtr); /* note : op <= oLitEnd <= oend_w == oend - 8 */
op = oLitEnd; op = oLitEnd;
*litPtr = iLitEnd; /* update for next sequence */ *litPtr = iLitEnd; /* update for next sequence */
/* copy Match */ /* copy Match */
if (sequence.offset > (size_t)(oLitEnd - prefixStart)) { if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {
/* offset beyond prefix */ /* offset beyond prefix */
if (sequence.offset > (size_t)(oLitEnd - dictStart)) return ERROR(corruption_detected); RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - dictStart), corruption_detected);
if (match + sequence.matchLength <= dictEnd) { if (match + sequence.matchLength <= dictEnd) {
memmove(oLitEnd, match, sequence.matchLength); memmove(oLitEnd, match, sequence.matchLength);
return sequenceLength; return sequenceLength;
} }
/* span extDict & currentPrefixSegment */ /* span extDict & currentPrefixSegment */
{ size_t const length1 = dictEnd - match; { size_t const length1 = dictEnd - match;
memmove(oLitEnd, match, length1); memmove(oLitEnd, match, length1);
op = oLitEnd + length1; op = oLitEnd + length1;
match -= sub2; match -= sub2;
} else { } else {
ZSTD_copy8(op, match); ZSTD_copy8(op, match);
} }
op += 8; match += 8; op += 8; match += 8;
if (oMatchEnd > oend-(16-MINMATCH)) { if (oMatchEnd > oend-(16-MINMATCH)) {
if (op < oend_w) { if (op < oend_w) {
ZSTD_wildcopy(op, match, oend_w - op); ZSTD_wildcopy(op, match, oend_w - op, ZSTD_overlap_src_before_dst);
match += oend_w - op; match += oend_w - op;
op = oend_w; op = oend_w;
} }
while (op < oMatchEnd) *op++ = *match++; while (op < oMatchEnd) *op++ = *match++;
} else { } else {
ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8); /* works even if matchLength < 8 */ ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8, ZSTD_overlap_src_before_dst); /* works even if matchLength < 8 */
} }
return sequenceLength; return sequenceLength;
} }
static void static void
ZSTD_initFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, const ZSTD_seqSymbol* dt) ZSTD_initFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, const ZSTD_seqSymbol* dt)
{ {
const void* ptr = dt; const void* ptr = dt;
U32 const nbBits = DInfo.nbBits; U32 const nbBits = DInfo.nbBits;
size_t const lowBits = BIT_readBits(bitD, nbBits); size_t const lowBits = BIT_readBits(bitD, nbBits);
DStatePtr->state = DInfo.nextState + lowBits; DStatePtr->state = DInfo.nextState + lowBits;
} }
/* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum /* 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) * 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 * bits before reloading. This value is the maximum number of bytes we read
* after reloading when we are decoding long offets. * after reloading when we are decoding long offsets.
*/ */
#define LONG_OFFSETS_MAX_EXTRA_BITS_32 \ #define LONG_OFFSETS_MAX_EXTRA_BITS_32 \
(ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32 \ (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32 \
? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32 \ ? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32 \
: 0) : 0)
typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e; typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
ZSTD_updateFseState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */ ZSTD_updateFseState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */
if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */ ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */
return seq; return seq;
} }
FORCE_INLINE_TEMPLATE size_t FORCE_INLINE_TEMPLATE size_t
DONT_VECTORIZE
ZSTD_decompressSequences_body( ZSTD_DCtx* dctx, ZSTD_decompressSequences_body( ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize, void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq, const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset) const ZSTD_longOffset_e isLongOffset)
{ {
const BYTE* ip = (const BYTE*)seqStart; const BYTE* ip = (const BYTE*)seqStart;
const BYTE* const iend = ip + seqSize; const BYTE* const iend = ip + seqSize;
BYTE* const ostart = (BYTE* const)dst; BYTE* const ostart = (BYTE* const)dst;
BYTE* const oend = ostart + maxDstSize; BYTE* const oend = ostart + maxDstSize;
BYTE* op = ostart; BYTE* op = ostart;
const BYTE* litPtr = dctx->litPtr; const BYTE* litPtr = dctx->litPtr;
const BYTE* const litEnd = litPtr + dctx->litSize; const BYTE* const litEnd = litPtr + dctx->litSize;
const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart); const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
const BYTE* const vBase = (const BYTE*) (dctx->virtualStart); const BYTE* const vBase = (const BYTE*) (dctx->virtualStart);
const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd); const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
DEBUGLOG(5, "ZSTD_decompressSequences_body"); DEBUGLOG(5, "ZSTD_decompressSequences_body");
/* Regen sequences */ /* Regen sequences */
if (nbSeq) { if (nbSeq) {
seqState_t seqState; seqState_t seqState;
dctx->fseEntropy = 1; dctx->fseEntropy = 1;
{ U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; } { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected); 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.stateLL, &seqState.DStream, dctx->LLTptr);
ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr); ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr); ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
ZSTD_STATIC_ASSERT(
BIT_DStream_unfinished < BIT_DStream_completed &&
BIT_DStream_endOfBuffer < BIT_DStream_completed &&
BIT_DStream_completed < BIT_DStream_overflow);
for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) { for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {
nbSeq--; nbSeq--;
{ seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset); { seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, prefixStart, vBase, dictEnd); size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, prefixStart, vBase, dictEnd);
DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize); DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
if (ZSTD_isError(oneSeqSize)) return oneSeqSize; if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
op += oneSeqSize; op += oneSeqSize;
} } } }
/* check if reached exact end */ /* check if reached exact end */
DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq); DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq);
if (nbSeq) return ERROR(corruption_detected); RETURN_ERROR_IF(nbSeq, corruption_detected);
RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected);
/* save reps for next block */ /* save reps for next block */
{ U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); } { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
} }
/* last literal segment */ /* last literal segment */
{ size_t const lastLLSize = litEnd - litPtr; { size_t const lastLLSize = litEnd - litPtr;
if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall);
memcpy(op, litPtr, lastLLSize); memcpy(op, litPtr, lastLLSize);
op += lastLLSize; op += lastLLSize;
} }
return op-ostart; return op-ostart;
} }
static size_t static size_t
seqState_t seqState; seqState_t seqState;
int seqNb; int seqNb;
dctx->fseEntropy = 1; dctx->fseEntropy = 1;
{ int i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; } { int i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
seqState.prefixStart = prefixStart; seqState.prefixStart = prefixStart;
seqState.pos = (size_t)(op-prefixStart); seqState.pos = (size_t)(op-prefixStart);
seqState.dictEnd = dictEnd; seqState.dictEnd = dictEnd;
assert(iend >= ip); assert(iend >= ip);
CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected); 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.stateLL, &seqState.DStream, dctx->LLTptr);
ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr); ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr); ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
/* prepare in advance */ /* prepare in advance */
for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) { for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {
sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, isLongOffset); 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 */ 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 */
} }
if (seqNb<seqAdvance) return ERROR(corruption_detected); RETURN_ERROR_IF(seqNb<seqAdvance, corruption_detected);
/* decode and decompress */ /* decode and decompress */
for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) { for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) {
seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, isLongOffset); 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); 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; 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 */ 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; sequences[seqNb & STORED_SEQS_MASK] = sequence;
op += oneSeqSize; op += oneSeqSize;
} }
if (seqNb<nbSeq) return ERROR(corruption_detected); RETURN_ERROR_IF(seqNb<nbSeq, corruption_detected);
/* finish queue */ /* finish queue */
seqNb -= seqAdvance; seqNb -= seqAdvance;
for ( ; seqNb<nbSeq ; seqNb++) { for ( ; seqNb<nbSeq ; seqNb++) {
size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb&STORED_SEQS_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd); size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb&STORED_SEQS_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
if (ZSTD_isError(oneSeqSize)) return oneSeqSize; if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
op += oneSeqSize; op += oneSeqSize;
} }
/* save reps for next block */ /* save reps for next block */
{ U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); } { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
} }
/* last literal segment */ /* last literal segment */
{ size_t const lastLLSize = litEnd - litPtr; { size_t const lastLLSize = litEnd - litPtr;
if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall); RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall);
memcpy(op, litPtr, lastLLSize); memcpy(op, litPtr, lastLLSize);
op += lastLLSize; op += lastLLSize;
} }
return op-ostart; return op-ostart;
} }
static size_t static size_t
ZSTD_decompressSequencesLong_default(ZSTD_DCtx* dctx, ZSTD_decompressSequencesLong_default(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize, void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq, const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset) const ZSTD_longOffset_e isLongOffset)
{ {
return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset); return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
} }
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */ #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
#if DYNAMIC_BMI2 #if DYNAMIC_BMI2
#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
static TARGET_ATTRIBUTE("bmi2") size_t static TARGET_ATTRIBUTE("bmi2") size_t
DONT_VECTORIZE
ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx, ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize, void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq, const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset) const ZSTD_longOffset_e isLongOffset)
{ {
return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset); return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
} }
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */ #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
} }
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */ #endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
/* ZSTD_decompressSequencesLong() : /* ZSTD_decompressSequencesLong() :
* decompression function triggered when a minimum share of offsets is considered "long", * decompression function triggered when a minimum share of offsets is considered "long",
* aka out of cache. * aka out of cache.
* note : "long" definition seems overloaded here, sometimes meaning "wider than bitstream register", and sometimes mearning "farther than memory cache distance". * 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 */ * This function will try to mitigate main memory latency through the use of prefetching */
static size_t static size_t
ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx, ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize, void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq, const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset) const ZSTD_longOffset_e isLongOffset)
{ {
DEBUGLOG(5, "ZSTD_decompressSequencesLong"); DEBUGLOG(5, "ZSTD_decompressSequencesLong");
* Offsets are long if they are larger than 2^STREAM_ACCUMULATOR_MIN. * 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. * 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. * In block mode, window size is not known, so we have to be conservative.
* (note: but it could be evaluated from current-lowLimit) * (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)))); 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); DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);
if (srcSize >= ZSTD_BLOCKSIZE_MAX) return ERROR(srcSize_wrong); RETURN_ERROR_IF(srcSize >= ZSTD_BLOCKSIZE_MAX, srcSize_wrong);
/* Decode literals section */ /* Decode literals section */
{ size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize); { size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);
DEBUGLOG(5, "ZSTD_decodeLiteralsBlock : %u", (U32)litCSize); DEBUGLOG(5, "ZSTD_decodeLiteralsBlock : %u", (U32)litCSize);
if (ZSTD_isError(litCSize)) return litCSize; if (ZSTD_isError(litCSize)) return litCSize;
ip += litCSize; ip += litCSize;
srcSize -= litCSize; srcSize -= litCSize;
} }

contrib/python-zstandard/zstd/decompress/zstd_decompress_internal.h

typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader, typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader,
ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock, ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock,
ZSTDds_decompressLastBlock, ZSTDds_checkChecksum, ZSTDds_decompressLastBlock, ZSTDds_checkChecksum,
ZSTDds_decodeSkippableHeader, ZSTDds_skipFrame } ZSTD_dStage; ZSTDds_decodeSkippableHeader, ZSTDds_skipFrame } ZSTD_dStage;
typedef enum { zdss_init=0, zdss_loadHeader, typedef enum { zdss_init=0, zdss_loadHeader,
zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage; 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 struct ZSTD_DCtx_s
{ {
const ZSTD_seqSymbol* LLTptr; const ZSTD_seqSymbol* LLTptr;
const ZSTD_seqSymbol* MLTptr; const ZSTD_seqSymbol* MLTptr;
const ZSTD_seqSymbol* OFTptr; const ZSTD_seqSymbol* OFTptr;
const HUF_DTable* HUFptr; const HUF_DTable* HUFptr;
ZSTD_entropyDTables_t entropy; ZSTD_entropyDTables_t entropy;
U32 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32]; /* space needed when building huffman tables */ U32 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32]; /* space needed when building huffman tables */
size_t staticSize; size_t staticSize;
int bmi2; /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */ int bmi2; /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */
/* dictionary */ /* dictionary */
ZSTD_DDict* ddictLocal; ZSTD_DDict* ddictLocal;
const ZSTD_DDict* ddict; /* set by ZSTD_initDStream_usingDDict(), or ZSTD_DCtx_refDDict() */ const ZSTD_DDict* ddict; /* set by ZSTD_initDStream_usingDDict(), or ZSTD_DCtx_refDDict() */
U32 dictID; U32 dictID;
int ddictIsCold; /* if == 1 : dictionary is "new" for working context, and presumed "cold" (not in cpu cache) */ int ddictIsCold; /* if == 1 : dictionary is "new" for working context, and presumed "cold" (not in cpu cache) */
ZSTD_dictUses_e dictUses;
/* streaming */ /* streaming */
ZSTD_dStreamStage streamStage; ZSTD_dStreamStage streamStage;
char* inBuff; char* inBuff;
size_t inBuffSize; size_t inBuffSize;
size_t inPos; size_t inPos;
size_t maxWindowSize; size_t maxWindowSize;
char* outBuff; char* outBuff;

contrib/python-zstandard/zstd/dictBuilder/cover.h

*/ */
typedef struct { typedef struct {
U32 begin; U32 begin;
U32 end; U32 end;
U32 score; U32 score;
} COVER_segment_t; } COVER_segment_t;
/** /**
*Number of epochs and size of each epoch.
*/
typedef struct {
U32 num;
U32 size;
} COVER_epoch_info_t;
/**
* Struct used for the dictionary selection function.
*/
typedef struct COVER_dictSelection {
BYTE* dictContent;
size_t dictSize;
size_t totalCompressedSize;
} COVER_dictSelection_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 * Checks total compressed size of a dictionary
*/ */
size_t COVER_checkTotalCompressedSize(const ZDICT_cover_params_t parameters, size_t COVER_checkTotalCompressedSize(const ZDICT_cover_params_t parameters,
const size_t *samplesSizes, const BYTE *samples, const size_t *samplesSizes, const BYTE *samples,
size_t *offsets, size_t *offsets,
size_t nbTrainSamples, size_t nbSamples, size_t nbTrainSamples, size_t nbSamples,
BYTE *const dict, size_t dictBufferCapacity); BYTE *const dict, size_t dictBufferCapacity);
*/ */
void COVER_best_start(COVER_best_t *best); void COVER_best_start(COVER_best_t *best);
/** /**
* Called when a thread finishes executing, both on error or success. * Called when a thread finishes executing, both on error or success.
* Decrements liveJobs and signals any waiting threads if liveJobs == 0. * Decrements liveJobs and signals any waiting threads if liveJobs == 0.
* If this dictionary is the best so far save it and its parameters. * If this dictionary is the best so far save it and its parameters.
*/ */
void COVER_best_finish(COVER_best_t *best, size_t compressedSize, void COVER_best_finish(COVER_best_t *best, ZDICT_cover_params_t parameters,
ZDICT_cover_params_t parameters, void *dict, COVER_dictSelection_t selection);
size_t dictSize); /**
* Error function for COVER_selectDict function. Checks if the return
* value is an error.
*/
unsigned COVER_dictSelectionIsError(COVER_dictSelection_t selection);
/**
* Error function for COVER_selectDict function. Returns a struct where
* return.totalCompressedSize is a ZSTD error.
*/
COVER_dictSelection_t COVER_dictSelectionError(size_t error);
/**
* Always call after selectDict is called to free up used memory from
* newly created dictionary.
*/
void COVER_dictSelectionFree(COVER_dictSelection_t selection);
/**
* Called to finalize the dictionary and select one based on whether or not
* the shrink-dict flag was enabled. If enabled the dictionary used is the
* smallest dictionary within a specified regression of the compressed size
* from the largest dictionary.
*/
COVER_dictSelection_t COVER_selectDict(BYTE* customDictContent,
size_t dictContentSize, const BYTE* samplesBuffer, const size_t* samplesSizes, unsigned nbFinalizeSamples,
size_t nbCheckSamples, size_t nbSamples, ZDICT_cover_params_t params, size_t* offsets, size_t totalCompressedSize);

contrib/python-zstandard/zstd/dictBuilder/cover.c

* Selects the best segment in an epoch. * Selects the best segment in an epoch.
* Segments of are scored according to the function: * Segments of are scored according to the function:
* *
* Let F(d) be the frequency of dmer d. * Let F(d) be the frequency of dmer d.
* Let S_i be the dmer at position i of segment S which has length k. * Let S_i be 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}) * 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, static COVER_segment_t COVER_selectSegment(const COVER_ctx_t *ctx, U32 *freqs,
COVER_map_t *activeDmers, U32 begin, COVER_map_t *activeDmers, U32 begin,
U32 end, U32 end,
ZDICT_cover_params_t parameters) { ZDICT_cover_params_t parameters) {
/* Constants */ /* Constants */
const U32 k = parameters.k; const U32 k = parameters.k;
const U32 d = parameters.d; const U32 d = parameters.d;
*newDmerOcc += 1; *newDmerOcc += 1;
/* If the window is now too large, drop the first position */ /* If the window is now too large, drop the first position */
if (activeSegment.end - activeSegment.begin == dmersInK + 1) { if (activeSegment.end - activeSegment.begin == dmersInK + 1) {
U32 delDmer = ctx->dmerAt[activeSegment.begin]; U32 delDmer = ctx->dmerAt[activeSegment.begin];
U32 *delDmerOcc = COVER_map_at(activeDmers, delDmer); U32 *delDmerOcc = COVER_map_at(activeDmers, delDmer);
activeSegment.begin += 1; activeSegment.begin += 1;
*delDmerOcc -= 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) { if (*delDmerOcc == 0) {
COVER_map_remove(activeDmers, delDmer); COVER_map_remove(activeDmers, delDmer);
activeSegment.score -= freqs[delDmer]; activeSegment.score -= freqs[delDmer];
} }
} }
/* If this segment is the best so far save it */ /* If this segment is the best so far save it */
if (activeSegment.score > bestSegment.score) { if (activeSegment.score > bestSegment.score) {
ctx->offsets = NULL; ctx->offsets = NULL;
} }
} }
/** /**
* Prepare a context for dictionary building. * Prepare a context for dictionary building.
* The context is only dependent on the parameter `d` and can used multiple * The context is only dependent on the parameter `d` and can used multiple
* times. * times.
* Returns 1 on success or zero on error. * Returns 0 on success or error code on error.
* The context must be destroyed with `COVER_ctx_destroy()`. * The context must be destroyed with `COVER_ctx_destroy()`.
*/ */
static int COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer, static size_t COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
const size_t *samplesSizes, unsigned nbSamples, const size_t *samplesSizes, unsigned nbSamples,
unsigned d, double splitPoint) { unsigned d, double splitPoint) {
const BYTE *const samples = (const BYTE *)samplesBuffer; const BYTE *const samples = (const BYTE *)samplesBuffer;
const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples); const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);
/* Split samples into testing and training sets */ /* Split samples into testing and training sets */
const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples; const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples;
const unsigned nbTestSamples = splitPoint < 1.0 ? nbSamples - nbTrainSamples : 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 trainingSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes, nbTrainSamples) : totalSamplesSize;
const size_t testSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize; const size_t testSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize;
/* Checks */ /* Checks */
if (totalSamplesSize < MAX(d, sizeof(U64)) || if (totalSamplesSize < MAX(d, sizeof(U64)) ||
totalSamplesSize >= (size_t)COVER_MAX_SAMPLES_SIZE) { totalSamplesSize >= (size_t)COVER_MAX_SAMPLES_SIZE) {
DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n", DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",
(unsigned)(totalSamplesSize>>20), (COVER_MAX_SAMPLES_SIZE >> 20)); (unsigned)(totalSamplesSize>>20), (COVER_MAX_SAMPLES_SIZE >> 20));
return 0; return ERROR(srcSize_wrong);
} }
/* Check if there are at least 5 training samples */ /* Check if there are at least 5 training samples */
if (nbTrainSamples < 5) { if (nbTrainSamples < 5) {
DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid.", nbTrainSamples); DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid.", nbTrainSamples);
return 0; return ERROR(srcSize_wrong);
} }
/* Check if there's testing sample */ /* Check if there's testing sample */
if (nbTestSamples < 1) { if (nbTestSamples < 1) {
DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.", nbTestSamples); DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.", nbTestSamples);
return 0; return ERROR(srcSize_wrong);
} }
/* Zero the context */ /* Zero the context */
memset(ctx, 0, sizeof(*ctx)); memset(ctx, 0, sizeof(*ctx));
DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples, DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples,
(unsigned)trainingSamplesSize); (unsigned)trainingSamplesSize);
DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples, DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples,
(unsigned)testSamplesSize); (unsigned)testSamplesSize);
ctx->samples = samples; ctx->samples = samples;
ctx->samplesSizes = samplesSizes; ctx->samplesSizes = samplesSizes;
ctx->nbSamples = nbSamples; ctx->nbSamples = nbSamples;
ctx->nbTrainSamples = nbTrainSamples; ctx->nbTrainSamples = nbTrainSamples;
ctx->nbTestSamples = nbTestSamples; ctx->nbTestSamples = nbTestSamples;
/* Partial suffix array */ /* Partial suffix array */
ctx->suffixSize = trainingSamplesSize - MAX(d, sizeof(U64)) + 1; ctx->suffixSize = trainingSamplesSize - MAX(d, sizeof(U64)) + 1;
ctx->suffix = (U32 *)malloc(ctx->suffixSize * sizeof(U32)); ctx->suffix = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
/* Maps index to the dmerID */ /* Maps index to the dmerID */
ctx->dmerAt = (U32 *)malloc(ctx->suffixSize * sizeof(U32)); ctx->dmerAt = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
/* The offsets of each file */ /* The offsets of each file */
ctx->offsets = (size_t *)malloc((nbSamples + 1) * sizeof(size_t)); ctx->offsets = (size_t *)malloc((nbSamples + 1) * sizeof(size_t));
if (!ctx->suffix || !ctx->dmerAt || !ctx->offsets) { if (!ctx->suffix || !ctx->dmerAt || !ctx->offsets) {
DISPLAYLEVEL(1, "Failed to allocate scratch buffers\n"); DISPLAYLEVEL(1, "Failed to allocate scratch buffers\n");
COVER_ctx_destroy(ctx); COVER_ctx_destroy(ctx);
return 0; return ERROR(memory_allocation);
} }
ctx->freqs = NULL; ctx->freqs = NULL;
ctx->d = d; ctx->d = d;
/* Fill offsets from the samplesSizes */ /* Fill offsets from the samplesSizes */
{ {
U32 i; U32 i;
ctx->offsets[0] = 0; ctx->offsets[0] = 0;
* dmerID so we can look up values in freq. * dmerID so we can look up values in freq.
* 2. We calculate how many samples the dmer occurs in and save it in * 2. We calculate how many samples the dmer occurs in and save it in
* freqs[dmerId]. * freqs[dmerId].
*/ */
COVER_groupBy(ctx->suffix, ctx->suffixSize, sizeof(U32), ctx, COVER_groupBy(ctx->suffix, ctx->suffixSize, sizeof(U32), ctx,
(ctx->d <= 8 ? &COVER_cmp8 : &COVER_cmp), &COVER_group); (ctx->d <= 8 ? &COVER_cmp8 : &COVER_cmp), &COVER_group);
ctx->freqs = ctx->suffix; ctx->freqs = ctx->suffix;
ctx->suffix = NULL; ctx->suffix = NULL;
return 1; return 0;
}
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. * Given the prepared context build the dictionary.
*/ */
static size_t COVER_buildDictionary(const COVER_ctx_t *ctx, U32 *freqs, static size_t COVER_buildDictionary(const COVER_ctx_t *ctx, U32 *freqs,
COVER_map_t *activeDmers, void *dictBuffer, COVER_map_t *activeDmers, void *dictBuffer,
size_t dictBufferCapacity, size_t dictBufferCapacity,
ZDICT_cover_params_t parameters) { ZDICT_cover_params_t parameters) {
BYTE *const dict = (BYTE *)dictBuffer; BYTE *const dict = (BYTE *)dictBuffer;
size_t tail = dictBufferCapacity; size_t tail = dictBufferCapacity;
/* Divide the data up into epochs of equal size. /* Divide the data into epochs. We will select one segment from each epoch. */
* We will select at least one segment from each epoch. const COVER_epoch_info_t epochs = COVER_computeEpochs(
*/ (U32)dictBufferCapacity, (U32)ctx->suffixSize, parameters.k, 4);
const unsigned epochs = MAX(1, (U32)(dictBufferCapacity / parameters.k / 4)); const size_t maxZeroScoreRun = MAX(10, MIN(100, epochs.num >> 3));
const unsigned epochSize = (U32)(ctx->suffixSize / epochs); size_t zeroScoreRun = 0;
size_t epoch; size_t epoch;
DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n", DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n",
epochs, epochSize); (U32)epochs.num, (U32)epochs.size);
/* Loop through the epochs until there are no more segments or the dictionary /* Loop through the epochs until there are no more segments or the dictionary
* is full. * is full.
*/ */
for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) { for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs.num) {
const U32 epochBegin = (U32)(epoch * epochSize); const U32 epochBegin = (U32)(epoch * epochs.size);
const U32 epochEnd = epochBegin + epochSize; const U32 epochEnd = epochBegin + epochs.size;
size_t segmentSize; size_t segmentSize;
/* Select a segment */ /* Select a segment */
COVER_segment_t segment = COVER_selectSegment( COVER_segment_t segment = COVER_selectSegment(
ctx, freqs, activeDmers, epochBegin, epochEnd, parameters); ctx, freqs, activeDmers, epochBegin, epochEnd, parameters);
/* If the segment covers no dmers, then we are out of content */ /* 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 (segment.score == 0) {
if (++zeroScoreRun >= maxZeroScoreRun) {
break; break;
} }
continue;
}
zeroScoreRun = 0;
/* Trim the segment if necessary and if it is too small then we are done */ /* 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); segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);
if (segmentSize < parameters.d) { if (segmentSize < parameters.d) {
break; break;
} }
/* We fill the dictionary from the back to allow the best segments to be /* We fill the dictionary from the back to allow the best segments to be
* referenced with the smallest offsets. * referenced with the smallest offsets.
*/ */
COVER_ctx_t ctx; COVER_ctx_t ctx;
COVER_map_t activeDmers; COVER_map_t activeDmers;
parameters.splitPoint = 1.0; parameters.splitPoint = 1.0;
/* Initialize global data */ /* Initialize global data */
g_displayLevel = parameters.zParams.notificationLevel; g_displayLevel = parameters.zParams.notificationLevel;
/* Checks */ /* Checks */
if (!COVER_checkParameters(parameters, dictBufferCapacity)) { if (!COVER_checkParameters(parameters, dictBufferCapacity)) {
DISPLAYLEVEL(1, "Cover parameters incorrect\n"); DISPLAYLEVEL(1, "Cover parameters incorrect\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (nbSamples == 0) { if (nbSamples == 0) {
DISPLAYLEVEL(1, "Cover must have at least one input file\n"); DISPLAYLEVEL(1, "Cover must have at least one input file\n");
return ERROR(GENERIC); return ERROR(srcSize_wrong);
} }
if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) { if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n", DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
ZDICT_DICTSIZE_MIN); ZDICT_DICTSIZE_MIN);
return ERROR(dstSize_tooSmall); return ERROR(dstSize_tooSmall);
} }
/* Initialize context and activeDmers */ /* Initialize context and activeDmers */
if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, {
parameters.d, parameters.splitPoint)) { size_t const initVal = COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
return ERROR(GENERIC); parameters.d, parameters.splitPoint);
if (ZSTD_isError(initVal)) {
return initVal;
} }
}
COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.suffixSize, g_displayLevel);
if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) { if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) {
DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n"); DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n");
COVER_ctx_destroy(&ctx); COVER_ctx_destroy(&ctx);
return ERROR(GENERIC); return ERROR(memory_allocation);
} }
DISPLAYLEVEL(2, "Building dictionary\n"); DISPLAYLEVEL(2, "Building dictionary\n");
{ {
const size_t tail = const size_t tail =
COVER_buildDictionary(&ctx, ctx.freqs, &activeDmers, dictBuffer, COVER_buildDictionary(&ctx, ctx.freqs, &activeDmers, dictBuffer,
dictBufferCapacity, parameters); dictBufferCapacity, parameters);
const size_t dictionarySize = ZDICT_finalizeDictionary( const size_t dictionarySize = ZDICT_finalizeDictionary(
/* Compress each sample and sum their sizes (or error) */ /* Compress each sample and sum their sizes (or error) */
totalCompressedSize = dictBufferCapacity; totalCompressedSize = dictBufferCapacity;
i = parameters.splitPoint < 1.0 ? nbTrainSamples : 0; i = parameters.splitPoint < 1.0 ? nbTrainSamples : 0;
for (; i < nbSamples; ++i) { for (; i < nbSamples; ++i) {
const size_t size = ZSTD_compress_usingCDict( const size_t size = ZSTD_compress_usingCDict(
cctx, dst, dstCapacity, samples + offsets[i], cctx, dst, dstCapacity, samples + offsets[i],
samplesSizes[i], cdict); samplesSizes[i], cdict);
if (ZSTD_isError(size)) { if (ZSTD_isError(size)) {
totalCompressedSize = ERROR(GENERIC); totalCompressedSize = size;
goto _compressCleanup; goto _compressCleanup;
} }
totalCompressedSize += size; totalCompressedSize += size;
} }
_compressCleanup: _compressCleanup:
ZSTD_freeCCtx(cctx); ZSTD_freeCCtx(cctx);
ZSTD_freeCDict(cdict); ZSTD_freeCDict(cdict);
if (dst) { if (dst) {
ZSTD_pthread_mutex_unlock(&best->mutex); ZSTD_pthread_mutex_unlock(&best->mutex);
} }
/** /**
* Called when a thread finishes executing, both on error or success. * Called when a thread finishes executing, both on error or success.
* Decrements liveJobs and signals any waiting threads if liveJobs == 0. * Decrements liveJobs and signals any waiting threads if liveJobs == 0.
* If this dictionary is the best so far save it and its parameters. * If this dictionary is the best so far save it and its parameters.
*/ */
void COVER_best_finish(COVER_best_t *best, size_t compressedSize, void COVER_best_finish(COVER_best_t *best, ZDICT_cover_params_t parameters,
ZDICT_cover_params_t parameters, void *dict, COVER_dictSelection_t selection) {
size_t dictSize) { void* dict = selection.dictContent;
size_t compressedSize = selection.totalCompressedSize;
size_t dictSize = selection.dictSize;
if (!best) { if (!best) {
return; return;
} }
{ {
size_t liveJobs; size_t liveJobs;
ZSTD_pthread_mutex_lock(&best->mutex); ZSTD_pthread_mutex_lock(&best->mutex);
--best->liveJobs; --best->liveJobs;
liveJobs = best->liveJobs; liveJobs = best->liveJobs;
best->compressedSize = ERROR(GENERIC); best->compressedSize = ERROR(GENERIC);
best->dictSize = 0; best->dictSize = 0;
ZSTD_pthread_cond_signal(&best->cond); ZSTD_pthread_cond_signal(&best->cond);
ZSTD_pthread_mutex_unlock(&best->mutex); ZSTD_pthread_mutex_unlock(&best->mutex);
return; return;
} }
} }
/* Save the dictionary, parameters, and size */ /* Save the dictionary, parameters, and size */
if (dict) {
memcpy(best->dict, dict, dictSize); memcpy(best->dict, dict, dictSize);
best->dictSize = dictSize; best->dictSize = dictSize;
best->parameters = parameters; best->parameters = parameters;
best->compressedSize = compressedSize; best->compressedSize = compressedSize;
} }
}
if (liveJobs == 0) { if (liveJobs == 0) {
ZSTD_pthread_cond_broadcast(&best->cond); ZSTD_pthread_cond_broadcast(&best->cond);
} }
ZSTD_pthread_mutex_unlock(&best->mutex); ZSTD_pthread_mutex_unlock(&best->mutex);
} }
} }
COVER_dictSelection_t COVER_dictSelectionError(size_t error) {
COVER_dictSelection_t selection = { NULL, 0, error };
return selection;
}
unsigned COVER_dictSelectionIsError(COVER_dictSelection_t selection) {
return (ZSTD_isError(selection.totalCompressedSize) || !selection.dictContent);
}
void COVER_dictSelectionFree(COVER_dictSelection_t selection){
free(selection.dictContent);
}
COVER_dictSelection_t COVER_selectDict(BYTE* customDictContent,
size_t dictContentSize, const BYTE* samplesBuffer, const size_t* samplesSizes, unsigned nbFinalizeSamples,
size_t nbCheckSamples, size_t nbSamples, ZDICT_cover_params_t params, size_t* offsets, size_t totalCompressedSize) {
size_t largestDict = 0;
size_t largestCompressed = 0;
BYTE* customDictContentEnd = customDictContent + dictContentSize;
BYTE * largestDictbuffer = (BYTE *)malloc(dictContentSize);
BYTE * candidateDictBuffer = (BYTE *)malloc(dictContentSize);
double regressionTolerance = ((double)params.shrinkDictMaxRegression / 100.0) + 1.00;
if (!largestDictbuffer || !candidateDictBuffer) {
free(largestDictbuffer);
free(candidateDictBuffer);
return COVER_dictSelectionError(dictContentSize);
}
/* Initial dictionary size and compressed size */
memcpy(largestDictbuffer, customDictContent, dictContentSize);
dictContentSize = ZDICT_finalizeDictionary(
largestDictbuffer, dictContentSize, customDictContent, dictContentSize,
samplesBuffer, samplesSizes, nbFinalizeSamples, params.zParams);
if (ZDICT_isError(dictContentSize)) {
free(largestDictbuffer);
free(candidateDictBuffer);
return COVER_dictSelectionError(dictContentSize);
}
totalCompressedSize = COVER_checkTotalCompressedSize(params, samplesSizes,
samplesBuffer, offsets,
nbCheckSamples, nbSamples,
largestDictbuffer, dictContentSize);
if (ZSTD_isError(totalCompressedSize)) {
free(largestDictbuffer);
free(candidateDictBuffer);
return COVER_dictSelectionError(totalCompressedSize);
}
if (params.shrinkDict == 0) {
COVER_dictSelection_t selection = { largestDictbuffer, dictContentSize, totalCompressedSize };
free(candidateDictBuffer);
return selection;
}
largestDict = dictContentSize;
largestCompressed = totalCompressedSize;
dictContentSize = ZDICT_DICTSIZE_MIN;
/* Largest dict is initially at least ZDICT_DICTSIZE_MIN */
while (dictContentSize < largestDict) {
memcpy(candidateDictBuffer, largestDictbuffer, largestDict);
dictContentSize = ZDICT_finalizeDictionary(
candidateDictBuffer, dictContentSize, customDictContentEnd - dictContentSize, dictContentSize,
samplesBuffer, samplesSizes, nbFinalizeSamples, params.zParams);
if (ZDICT_isError(dictContentSize)) {
free(largestDictbuffer);
free(candidateDictBuffer);
return COVER_dictSelectionError(dictContentSize);
}
totalCompressedSize = COVER_checkTotalCompressedSize(params, samplesSizes,
samplesBuffer, offsets,
nbCheckSamples, nbSamples,
candidateDictBuffer, dictContentSize);
if (ZSTD_isError(totalCompressedSize)) {
free(largestDictbuffer);
free(candidateDictBuffer);
return COVER_dictSelectionError(totalCompressedSize);
}
if (totalCompressedSize <= largestCompressed * regressionTolerance) {
COVER_dictSelection_t selection = { candidateDictBuffer, dictContentSize, totalCompressedSize };
free(largestDictbuffer);
return selection;
}
dictContentSize *= 2;
}
dictContentSize = largestDict;
totalCompressedSize = largestCompressed;
{
COVER_dictSelection_t selection = { largestDictbuffer, dictContentSize, totalCompressedSize };
free(candidateDictBuffer);
return selection;
}
}
/** /**
* Parameters for COVER_tryParameters(). * Parameters for COVER_tryParameters().
*/ */
typedef struct COVER_tryParameters_data_s { typedef struct COVER_tryParameters_data_s {
const COVER_ctx_t *ctx; const COVER_ctx_t *ctx;
COVER_best_t *best; COVER_best_t *best;
size_t dictBufferCapacity; size_t dictBufferCapacity;
ZDICT_cover_params_t parameters; ZDICT_cover_params_t parameters;
COVER_tryParameters_data_t *const data = (COVER_tryParameters_data_t *)opaque; COVER_tryParameters_data_t *const data = (COVER_tryParameters_data_t *)opaque;
const COVER_ctx_t *const ctx = data->ctx; const COVER_ctx_t *const ctx = data->ctx;
const ZDICT_cover_params_t parameters = data->parameters; const ZDICT_cover_params_t parameters = data->parameters;
size_t dictBufferCapacity = data->dictBufferCapacity; size_t dictBufferCapacity = data->dictBufferCapacity;
size_t totalCompressedSize = ERROR(GENERIC); size_t totalCompressedSize = ERROR(GENERIC);
/* Allocate space for hash table, dict, and freqs */ /* Allocate space for hash table, dict, and freqs */
COVER_map_t activeDmers; COVER_map_t activeDmers;
BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity); BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);
COVER_dictSelection_t selection = COVER_dictSelectionError(ERROR(GENERIC));
U32 *freqs = (U32 *)malloc(ctx->suffixSize * sizeof(U32)); U32 *freqs = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) { if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) {
DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n"); DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n");
goto _cleanup; goto _cleanup;
} }
if (!dict || !freqs) { if (!dict || !freqs) {
DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n"); DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n");
goto _cleanup; goto _cleanup;
} }
/* Copy the frequencies because we need to modify them */ /* Copy the frequencies because we need to modify them */
memcpy(freqs, ctx->freqs, ctx->suffixSize * sizeof(U32)); memcpy(freqs, ctx->freqs, ctx->suffixSize * sizeof(U32));
/* Build the dictionary */ /* Build the dictionary */
{ {
const size_t tail = COVER_buildDictionary(ctx, freqs, &activeDmers, dict, const size_t tail = COVER_buildDictionary(ctx, freqs, &activeDmers, dict,
dictBufferCapacity, parameters); dictBufferCapacity, parameters);
dictBufferCapacity = ZDICT_finalizeDictionary( selection = COVER_selectDict(dict + tail, dictBufferCapacity - tail,
dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail, ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbTrainSamples, ctx->nbTrainSamples, ctx->nbSamples, parameters, ctx->offsets,
ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbTrainSamples, totalCompressedSize);
parameters.zParams);
if (ZDICT_isError(dictBufferCapacity)) { if (COVER_dictSelectionIsError(selection)) {
DISPLAYLEVEL(1, "Failed to finalize dictionary\n"); DISPLAYLEVEL(1, "Failed to select dictionary\n");
goto _cleanup; goto _cleanup;
} }
} }
/* Check total compressed size */
totalCompressedSize = COVER_checkTotalCompressedSize(parameters, ctx->samplesSizes,
ctx->samples, ctx->offsets,
ctx->nbTrainSamples, ctx->nbSamples,
dict, dictBufferCapacity);
_cleanup: _cleanup:
COVER_best_finish(data->best, totalCompressedSize, parameters, dict, free(dict);
dictBufferCapacity); COVER_best_finish(data->best, parameters, selection);
free(data); free(data);
COVER_map_destroy(&activeDmers); COVER_map_destroy(&activeDmers);
if (dict) { COVER_dictSelectionFree(selection);
free(dict);
}
if (freqs) { if (freqs) {
free(freqs); free(freqs);
} }
} }
ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover( ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(
void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer, void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer,
const size_t *samplesSizes, unsigned nbSamples, const size_t *samplesSizes, unsigned nbSamples,
ZDICT_cover_params_t *parameters) { ZDICT_cover_params_t *parameters) {
/* constants */ /* constants */
const unsigned nbThreads = parameters->nbThreads; const unsigned nbThreads = parameters->nbThreads;
const double splitPoint = const double splitPoint =
parameters->splitPoint <= 0.0 ? DEFAULT_SPLITPOINT : parameters->splitPoint; parameters->splitPoint <= 0.0 ? DEFAULT_SPLITPOINT : parameters->splitPoint;
const unsigned kMinD = parameters->d == 0 ? 6 : parameters->d; const unsigned kMinD = parameters->d == 0 ? 6 : parameters->d;
const unsigned kMaxD = parameters->d == 0 ? 8 : parameters->d; const unsigned kMaxD = parameters->d == 0 ? 8 : parameters->d;
const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k; const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k;
const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k; const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k;
const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps; const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps;
const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1); const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1);
const unsigned kIterations = const unsigned kIterations =
(1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize); (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);
const unsigned shrinkDict = 0;
/* Local variables */ /* Local variables */
const int displayLevel = parameters->zParams.notificationLevel; const int displayLevel = parameters->zParams.notificationLevel;
unsigned iteration = 1; unsigned iteration = 1;
unsigned d; unsigned d;
unsigned k; unsigned k;
COVER_best_t best; COVER_best_t best;
POOL_ctx *pool = NULL; POOL_ctx *pool = NULL;
int warned = 0;
/* Checks */ /* Checks */
if (splitPoint <= 0 || splitPoint > 1) { if (splitPoint <= 0 || splitPoint > 1) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (kMinK < kMaxD || kMaxK < kMinK) { if (kMinK < kMaxD || kMaxK < kMinK) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (nbSamples == 0) { if (nbSamples == 0) {
DISPLAYLEVEL(1, "Cover must have at least one input file\n"); DISPLAYLEVEL(1, "Cover must have at least one input file\n");
return ERROR(GENERIC); return ERROR(srcSize_wrong);
} }
if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) { if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n", DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
ZDICT_DICTSIZE_MIN); ZDICT_DICTSIZE_MIN);
return ERROR(dstSize_tooSmall); return ERROR(dstSize_tooSmall);
} }
if (nbThreads > 1) { if (nbThreads > 1) {
pool = POOL_create(nbThreads, 1); pool = POOL_create(nbThreads, 1);
if (!pool) { if (!pool) {
return ERROR(memory_allocation); return ERROR(memory_allocation);
} }
} }
/* Initialization */ /* Initialization */
COVER_best_init(&best); COVER_best_init(&best);
/* Turn down global display level to clean up display at level 2 and below */ /* Turn down global display level to clean up display at level 2 and below */
g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1; g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1;
/* Loop through d first because each new value needs a new context */ /* Loop through d first because each new value needs a new context */
LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n", LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n",
kIterations); kIterations);
for (d = kMinD; d <= kMaxD; d += 2) { for (d = kMinD; d <= kMaxD; d += 2) {
/* Initialize the context for this value of d */ /* Initialize the context for this value of d */
COVER_ctx_t ctx; COVER_ctx_t ctx;
LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d); LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);
if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint)) { {
const size_t initVal = COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint);
if (ZSTD_isError(initVal)) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
COVER_best_destroy(&best); COVER_best_destroy(&best);
POOL_free(pool); POOL_free(pool);
return ERROR(GENERIC); return initVal;
}
}
if (!warned) {
COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.suffixSize, displayLevel);
warned = 1;
} }
/* Loop through k reusing the same context */ /* Loop through k reusing the same context */
for (k = kMinK; k <= kMaxK; k += kStepSize) { for (k = kMinK; k <= kMaxK; k += kStepSize) {
/* Prepare the arguments */ /* Prepare the arguments */
COVER_tryParameters_data_t *data = (COVER_tryParameters_data_t *)malloc( COVER_tryParameters_data_t *data = (COVER_tryParameters_data_t *)malloc(
sizeof(COVER_tryParameters_data_t)); sizeof(COVER_tryParameters_data_t));
LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k); LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k);
if (!data) { if (!data) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n");
COVER_best_destroy(&best); COVER_best_destroy(&best);
COVER_ctx_destroy(&ctx); COVER_ctx_destroy(&ctx);
POOL_free(pool); POOL_free(pool);
return ERROR(GENERIC); return ERROR(memory_allocation);
} }
data->ctx = &ctx; data->ctx = &ctx;
data->best = &best; data->best = &best;
data->dictBufferCapacity = dictBufferCapacity; data->dictBufferCapacity = dictBufferCapacity;
data->parameters = *parameters; data->parameters = *parameters;
data->parameters.k = k; data->parameters.k = k;
data->parameters.d = d; data->parameters.d = d;
data->parameters.splitPoint = splitPoint; data->parameters.splitPoint = splitPoint;
data->parameters.steps = kSteps; data->parameters.steps = kSteps;
data->parameters.shrinkDict = shrinkDict;
data->parameters.zParams.notificationLevel = g_displayLevel; data->parameters.zParams.notificationLevel = g_displayLevel;
/* Check the parameters */ /* Check the parameters */
if (!COVER_checkParameters(data->parameters, dictBufferCapacity)) { if (!COVER_checkParameters(data->parameters, dictBufferCapacity)) {
DISPLAYLEVEL(1, "Cover parameters incorrect\n"); DISPLAYLEVEL(1, "Cover parameters incorrect\n");
free(data); free(data);
continue; continue;
} }
/* Call the function and pass ownership of data to it */ /* Call the function and pass ownership of data to it */

contrib/python-zstandard/zstd/dictBuilder/fastcover.c

* Selects the best segment in an epoch. * Selects the best segment in an epoch.
* Segments of are scored according to the function: * Segments of are scored according to the function:
* *
* Let F(d) be the frequency of all dmers with hash value d. * 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. * 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}) * Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1})
* *
* Once the dmer with hash value d is in the dictionay we set F(d) = 0. * 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, static COVER_segment_t FASTCOVER_selectSegment(const FASTCOVER_ctx_t *ctx,
U32 *freqs, U32 begin, U32 end, U32 *freqs, U32 begin, U32 end,
ZDICT_cover_params_t parameters, ZDICT_cover_params_t parameters,
U16* segmentFreqs) { U16* segmentFreqs) {
/* Constants */ /* Constants */
const U32 k = parameters.k; const U32 k = parameters.k;
const U32 d = parameters.d; const U32 d = parameters.d;
/* Slide the activeSegment through the whole epoch. /* Slide the activeSegment through the whole epoch.
* Save the best segment in bestSegment. * Save the best segment in bestSegment.
*/ */
while (activeSegment.end < end) { while (activeSegment.end < end) {
/* Get hash value of current dmer */ /* Get hash value of current dmer */
const size_t idx = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.end, f, d); const size_t idx = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.end, f, d);
/* Add frequency of this index to score if this is the first occurence of index in active segment */ /* Add frequency of this index to score if this is the first occurrence of index in active segment */
if (segmentFreqs[idx] == 0) { if (segmentFreqs[idx] == 0) {
activeSegment.score += freqs[idx]; activeSegment.score += freqs[idx];
} }
/* Increment end of segment and segmentFreqs*/ /* Increment end of segment and segmentFreqs*/
activeSegment.end += 1; activeSegment.end += 1;
segmentFreqs[idx] += 1; segmentFreqs[idx] += 1;
/* If the window is now too large, drop the first position */ /* If the window is now too large, drop the first position */
if (activeSegment.end - activeSegment.begin == dmersInK + 1) { if (activeSegment.end - activeSegment.begin == dmersInK + 1) {
} }
} }
/** /**
* Prepare a context for dictionary building. * Prepare a context for dictionary building.
* The context is only dependent on the parameter `d` and can used multiple * The context is only dependent on the parameter `d` and can used multiple
* times. * times.
* Returns 1 on success or zero on error. * Returns 0 on success or error code on error.
* The context must be destroyed with `FASTCOVER_ctx_destroy()`. * The context must be destroyed with `FASTCOVER_ctx_destroy()`.
*/ */
static int static size_t
FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx, FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx,
const void* samplesBuffer, const void* samplesBuffer,
const size_t* samplesSizes, unsigned nbSamples, const size_t* samplesSizes, unsigned nbSamples,
unsigned d, double splitPoint, unsigned f, unsigned d, double splitPoint, unsigned f,
FASTCOVER_accel_t accelParams) FASTCOVER_accel_t accelParams)
{ {
const BYTE* const samples = (const BYTE*)samplesBuffer; const BYTE* const samples = (const BYTE*)samplesBuffer;
const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples); const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);
/* Split samples into testing and training sets */ /* Split samples into testing and training sets */
const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples; const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples;
const unsigned nbTestSamples = splitPoint < 1.0 ? nbSamples - nbTrainSamples : 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 trainingSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes, nbTrainSamples) : totalSamplesSize;
const size_t testSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize; const size_t testSamplesSize = splitPoint < 1.0 ? COVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize;
/* Checks */ /* Checks */
if (totalSamplesSize < MAX(d, sizeof(U64)) || if (totalSamplesSize < MAX(d, sizeof(U64)) ||
totalSamplesSize >= (size_t)FASTCOVER_MAX_SAMPLES_SIZE) { totalSamplesSize >= (size_t)FASTCOVER_MAX_SAMPLES_SIZE) {
DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n", DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",
(unsigned)(totalSamplesSize >> 20), (FASTCOVER_MAX_SAMPLES_SIZE >> 20)); (unsigned)(totalSamplesSize >> 20), (FASTCOVER_MAX_SAMPLES_SIZE >> 20));
return 0; return ERROR(srcSize_wrong);
} }
/* Check if there are at least 5 training samples */ /* Check if there are at least 5 training samples */
if (nbTrainSamples < 5) { if (nbTrainSamples < 5) {
DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid\n", nbTrainSamples); DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid\n", nbTrainSamples);
return 0; return ERROR(srcSize_wrong);
} }
/* Check if there's testing sample */ /* Check if there's testing sample */
if (nbTestSamples < 1) { if (nbTestSamples < 1) {
DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.\n", nbTestSamples); DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.\n", nbTestSamples);
return 0; return ERROR(srcSize_wrong);
} }
/* Zero the context */ /* Zero the context */
memset(ctx, 0, sizeof(*ctx)); memset(ctx, 0, sizeof(*ctx));
DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples, DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples,
(unsigned)trainingSamplesSize); (unsigned)trainingSamplesSize);
DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples, DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples,
(unsigned)testSamplesSize); (unsigned)testSamplesSize);
ctx->samples = samples; ctx->samples = samples;
ctx->samplesSizes = samplesSizes; ctx->samplesSizes = samplesSizes;
ctx->nbSamples = nbSamples; ctx->nbSamples = nbSamples;
ctx->nbTrainSamples = nbTrainSamples; ctx->nbTrainSamples = nbTrainSamples;
ctx->nbTestSamples = nbTestSamples; ctx->nbTestSamples = nbTestSamples;
ctx->nbDmers = trainingSamplesSize - MAX(d, sizeof(U64)) + 1; ctx->nbDmers = trainingSamplesSize - MAX(d, sizeof(U64)) + 1;
ctx->d = d; ctx->d = d;
ctx->f = f; ctx->f = f;
ctx->accelParams = accelParams; ctx->accelParams = accelParams;
/* The offsets of each file */ /* The offsets of each file */
ctx->offsets = (size_t*)calloc((nbSamples + 1), sizeof(size_t)); ctx->offsets = (size_t*)calloc((nbSamples + 1), sizeof(size_t));
if (ctx->offsets == NULL) { if (ctx->offsets == NULL) {
DISPLAYLEVEL(1, "Failed to allocate scratch buffers \n"); DISPLAYLEVEL(1, "Failed to allocate scratch buffers \n");
FASTCOVER_ctx_destroy(ctx); FASTCOVER_ctx_destroy(ctx);
return 0; return ERROR(memory_allocation);
} }
/* Fill offsets from the samplesSizes */ /* Fill offsets from the samplesSizes */
{ U32 i; { U32 i;
ctx->offsets[0] = 0; ctx->offsets[0] = 0;
assert(nbSamples >= 5); assert(nbSamples >= 5);
for (i = 1; i <= nbSamples; ++i) { for (i = 1; i <= nbSamples; ++i) {
ctx->offsets[i] = ctx->offsets[i - 1] + samplesSizes[i - 1]; ctx->offsets[i] = ctx->offsets[i - 1] + samplesSizes[i - 1];
} }
} }
/* Initialize frequency array of size 2^f */ /* Initialize frequency array of size 2^f */
ctx->freqs = (U32*)calloc(((U64)1 << f), sizeof(U32)); ctx->freqs = (U32*)calloc(((U64)1 << f), sizeof(U32));
if (ctx->freqs == NULL) { if (ctx->freqs == NULL) {
DISPLAYLEVEL(1, "Failed to allocate frequency table \n"); DISPLAYLEVEL(1, "Failed to allocate frequency table \n");
FASTCOVER_ctx_destroy(ctx); FASTCOVER_ctx_destroy(ctx);
return 0; return ERROR(memory_allocation);
} }
DISPLAYLEVEL(2, "Computing frequencies\n"); DISPLAYLEVEL(2, "Computing frequencies\n");
FASTCOVER_computeFrequency(ctx->freqs, ctx); FASTCOVER_computeFrequency(ctx->freqs, ctx);
return 1; return 0;
} }
/** /**
* Given the prepared context build the dictionary. * Given the prepared context build the dictionary.
*/ */
static size_t static size_t
FASTCOVER_buildDictionary(const FASTCOVER_ctx_t* ctx, FASTCOVER_buildDictionary(const FASTCOVER_ctx_t* ctx,
U32* freqs, U32* freqs,
void* dictBuffer, size_t dictBufferCapacity, void* dictBuffer, size_t dictBufferCapacity,
ZDICT_cover_params_t parameters, ZDICT_cover_params_t parameters,
U16* segmentFreqs) U16* segmentFreqs)
{ {
BYTE *const dict = (BYTE *)dictBuffer; BYTE *const dict = (BYTE *)dictBuffer;
size_t tail = dictBufferCapacity; size_t tail = dictBufferCapacity;
/* Divide the data up into epochs of equal size. /* Divide the data into epochs. We will select one segment from each epoch. */
* We will select at least one segment from each epoch. const COVER_epoch_info_t epochs = COVER_computeEpochs(
*/ (U32)dictBufferCapacity, (U32)ctx->nbDmers, parameters.k, 1);
const unsigned epochs = MAX(1, (U32)(dictBufferCapacity / parameters.k)); const size_t maxZeroScoreRun = 10;
const unsigned epochSize = (U32)(ctx->nbDmers / epochs); size_t zeroScoreRun = 0;
size_t epoch; size_t epoch;
DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n", DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n",
epochs, epochSize); (U32)epochs.num, (U32)epochs.size);
/* Loop through the epochs until there are no more segments or the dictionary /* Loop through the epochs until there are no more segments or the dictionary
* is full. * is full.
*/ */
for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) { for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs.num) {
const U32 epochBegin = (U32)(epoch * epochSize); const U32 epochBegin = (U32)(epoch * epochs.size);
const U32 epochEnd = epochBegin + epochSize; const U32 epochEnd = epochBegin + epochs.size;
size_t segmentSize; size_t segmentSize;
/* Select a segment */ /* Select a segment */
COVER_segment_t segment = FASTCOVER_selectSegment( COVER_segment_t segment = FASTCOVER_selectSegment(
ctx, freqs, epochBegin, epochEnd, parameters, segmentFreqs); ctx, freqs, epochBegin, epochEnd, parameters, segmentFreqs);
/* If the segment covers no dmers, then we are out of content */ /* 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 (segment.score == 0) {
if (++zeroScoreRun >= maxZeroScoreRun) {
break; break;
} }
continue;
}
zeroScoreRun = 0;
/* Trim the segment if necessary and if it is too small then we are done */ /* 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); segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);
if (segmentSize < parameters.d) { if (segmentSize < parameters.d) {
break; break;
} }
/* We fill the dictionary from the back to allow the best segments to be /* We fill the dictionary from the back to allow the best segments to be
* referenced with the smallest offsets. * referenced with the smallest offsets.
*/ */
tail -= segmentSize; tail -= segmentSize;
memcpy(dict + tail, ctx->samples + segment.begin, segmentSize); memcpy(dict + tail, ctx->samples + segment.begin, segmentSize);
DISPLAYUPDATE( DISPLAYUPDATE(
2, "\r%u%% ", 2, "\r%u%% ",
(unsigned)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity)); (unsigned)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity));
} }
DISPLAYLEVEL(2, "\r%79s\r", ""); DISPLAYLEVEL(2, "\r%79s\r", "");
return tail; return tail;
} }
/** /**
* Parameters for FASTCOVER_tryParameters(). * Parameters for FASTCOVER_tryParameters().
*/ */
typedef struct FASTCOVER_tryParameters_data_s { typedef struct FASTCOVER_tryParameters_data_s {
const FASTCOVER_ctx_t* ctx; const FASTCOVER_ctx_t* ctx;
COVER_best_t* best; COVER_best_t* best;
size_t dictBufferCapacity; size_t dictBufferCapacity;
ZDICT_cover_params_t parameters; ZDICT_cover_params_t parameters;
const FASTCOVER_ctx_t *const ctx = data->ctx; const FASTCOVER_ctx_t *const ctx = data->ctx;
const ZDICT_cover_params_t parameters = data->parameters; const ZDICT_cover_params_t parameters = data->parameters;
size_t dictBufferCapacity = data->dictBufferCapacity; size_t dictBufferCapacity = data->dictBufferCapacity;
size_t totalCompressedSize = ERROR(GENERIC); size_t totalCompressedSize = ERROR(GENERIC);
/* Initialize array to keep track of frequency of dmer within activeSegment */ /* Initialize array to keep track of frequency of dmer within activeSegment */
U16* segmentFreqs = (U16 *)calloc(((U64)1 << ctx->f), sizeof(U16)); U16* segmentFreqs = (U16 *)calloc(((U64)1 << ctx->f), sizeof(U16));
/* Allocate space for hash table, dict, and freqs */ /* Allocate space for hash table, dict, and freqs */
BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity); BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);
COVER_dictSelection_t selection = COVER_dictSelectionError(ERROR(GENERIC));
U32 *freqs = (U32*) malloc(((U64)1 << ctx->f) * sizeof(U32)); U32 *freqs = (U32*) malloc(((U64)1 << ctx->f) * sizeof(U32));
if (!segmentFreqs || !dict || !freqs) { if (!segmentFreqs || !dict || !freqs) {
DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n"); DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n");
goto _cleanup; goto _cleanup;
} }
/* Copy the frequencies because we need to modify them */ /* Copy the frequencies because we need to modify them */
memcpy(freqs, ctx->freqs, ((U64)1 << ctx->f) * sizeof(U32)); memcpy(freqs, ctx->freqs, ((U64)1 << ctx->f) * sizeof(U32));
/* Build the dictionary */ /* Build the dictionary */
{ const size_t tail = FASTCOVER_buildDictionary(ctx, freqs, dict, dictBufferCapacity, { const size_t tail = FASTCOVER_buildDictionary(ctx, freqs, dict, dictBufferCapacity,
parameters, segmentFreqs); parameters, segmentFreqs);
const unsigned nbFinalizeSamples = (unsigned)(ctx->nbTrainSamples * ctx->accelParams.finalize / 100); const unsigned nbFinalizeSamples = (unsigned)(ctx->nbTrainSamples * ctx->accelParams.finalize / 100);
dictBufferCapacity = ZDICT_finalizeDictionary( selection = COVER_selectDict(dict + tail, dictBufferCapacity - tail,
dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail, ctx->samples, ctx->samplesSizes, nbFinalizeSamples, ctx->nbTrainSamples, ctx->nbSamples, parameters, ctx->offsets,
ctx->samples, ctx->samplesSizes, nbFinalizeSamples, parameters.zParams); totalCompressedSize);
if (ZDICT_isError(dictBufferCapacity)) {
DISPLAYLEVEL(1, "Failed to finalize dictionary\n"); if (COVER_dictSelectionIsError(selection)) {
DISPLAYLEVEL(1, "Failed to select dictionary\n");
goto _cleanup; goto _cleanup;
} }
} }
/* Check total compressed size */
totalCompressedSize = COVER_checkTotalCompressedSize(parameters, ctx->samplesSizes,
ctx->samples, ctx->offsets,
ctx->nbTrainSamples, ctx->nbSamples,
dict, dictBufferCapacity);
_cleanup: _cleanup:
COVER_best_finish(data->best, totalCompressedSize, parameters, dict, free(dict);
dictBufferCapacity); COVER_best_finish(data->best, parameters, selection);
free(data); free(data);
free(segmentFreqs); free(segmentFreqs);
free(dict); COVER_dictSelectionFree(selection);
free(freqs); free(freqs);
} }
static void static void
FASTCOVER_convertToCoverParams(ZDICT_fastCover_params_t fastCoverParams, FASTCOVER_convertToCoverParams(ZDICT_fastCover_params_t fastCoverParams,
ZDICT_cover_params_t* coverParams) ZDICT_cover_params_t* coverParams)
{ {
coverParams->k = fastCoverParams.k; coverParams->k = fastCoverParams.k;
coverParams->d = fastCoverParams.d; coverParams->d = fastCoverParams.d;
coverParams->steps = fastCoverParams.steps; coverParams->steps = fastCoverParams.steps;
coverParams->nbThreads = fastCoverParams.nbThreads; coverParams->nbThreads = fastCoverParams.nbThreads;
coverParams->splitPoint = fastCoverParams.splitPoint; coverParams->splitPoint = fastCoverParams.splitPoint;
coverParams->zParams = fastCoverParams.zParams; coverParams->zParams = fastCoverParams.zParams;
coverParams->shrinkDict = fastCoverParams.shrinkDict;
} }
static void static void
FASTCOVER_convertToFastCoverParams(ZDICT_cover_params_t coverParams, FASTCOVER_convertToFastCoverParams(ZDICT_cover_params_t coverParams,
ZDICT_fastCover_params_t* fastCoverParams, ZDICT_fastCover_params_t* fastCoverParams,
unsigned f, unsigned accel) unsigned f, unsigned accel)
{ {
fastCoverParams->k = coverParams.k; fastCoverParams->k = coverParams.k;
fastCoverParams->d = coverParams.d; fastCoverParams->d = coverParams.d;
fastCoverParams->steps = coverParams.steps; fastCoverParams->steps = coverParams.steps;
fastCoverParams->nbThreads = coverParams.nbThreads; fastCoverParams->nbThreads = coverParams.nbThreads;
fastCoverParams->splitPoint = coverParams.splitPoint; fastCoverParams->splitPoint = coverParams.splitPoint;
fastCoverParams->f = f; fastCoverParams->f = f;
fastCoverParams->accel = accel; fastCoverParams->accel = accel;
fastCoverParams->zParams = coverParams.zParams; fastCoverParams->zParams = coverParams.zParams;
fastCoverParams->shrinkDict = coverParams.shrinkDict;
} }
ZDICTLIB_API size_t ZDICTLIB_API size_t
ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity, ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity,
const void* samplesBuffer, const void* samplesBuffer,
const size_t* samplesSizes, unsigned nbSamples, const size_t* samplesSizes, unsigned nbSamples,
ZDICT_fastCover_params_t parameters) ZDICT_fastCover_params_t parameters)
parameters.accel = parameters.accel == 0 ? DEFAULT_ACCEL : parameters.accel; parameters.accel = parameters.accel == 0 ? DEFAULT_ACCEL : parameters.accel;
/* Convert to cover parameter */ /* Convert to cover parameter */
memset(&coverParams, 0 , sizeof(coverParams)); memset(&coverParams, 0 , sizeof(coverParams));
FASTCOVER_convertToCoverParams(parameters, &coverParams); FASTCOVER_convertToCoverParams(parameters, &coverParams);
/* Checks */ /* Checks */
if (!FASTCOVER_checkParameters(coverParams, dictBufferCapacity, parameters.f, if (!FASTCOVER_checkParameters(coverParams, dictBufferCapacity, parameters.f,
parameters.accel)) { parameters.accel)) {
DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n"); DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (nbSamples == 0) { if (nbSamples == 0) {
DISPLAYLEVEL(1, "FASTCOVER must have at least one input file\n"); DISPLAYLEVEL(1, "FASTCOVER must have at least one input file\n");
return ERROR(GENERIC); return ERROR(srcSize_wrong);
} }
if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) { if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n", DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
ZDICT_DICTSIZE_MIN); ZDICT_DICTSIZE_MIN);
return ERROR(dstSize_tooSmall); return ERROR(dstSize_tooSmall);
} }
/* Assign corresponding FASTCOVER_accel_t to accelParams*/ /* Assign corresponding FASTCOVER_accel_t to accelParams*/
accelParams = FASTCOVER_defaultAccelParameters[parameters.accel]; accelParams = FASTCOVER_defaultAccelParameters[parameters.accel];
/* Initialize context */ /* Initialize context */
if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, {
size_t const initVal = FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
coverParams.d, parameters.splitPoint, parameters.f, coverParams.d, parameters.splitPoint, parameters.f,
accelParams)) { accelParams);
if (ZSTD_isError(initVal)) {
DISPLAYLEVEL(1, "Failed to initialize context\n"); DISPLAYLEVEL(1, "Failed to initialize context\n");
return ERROR(GENERIC); return initVal;
}
} }
COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.nbDmers, g_displayLevel);
/* Build the dictionary */ /* Build the dictionary */
DISPLAYLEVEL(2, "Building dictionary\n"); DISPLAYLEVEL(2, "Building dictionary\n");
{ {
/* Initialize array to keep track of frequency of dmer within activeSegment */ /* Initialize array to keep track of frequency of dmer within activeSegment */
U16* segmentFreqs = (U16 *)calloc(((U64)1 << parameters.f), sizeof(U16)); U16* segmentFreqs = (U16 *)calloc(((U64)1 << parameters.f), sizeof(U16));
const size_t tail = FASTCOVER_buildDictionary(&ctx, ctx.freqs, dictBuffer, const size_t tail = FASTCOVER_buildDictionary(&ctx, ctx.freqs, dictBuffer,
dictBufferCapacity, coverParams, segmentFreqs); dictBufferCapacity, coverParams, segmentFreqs);
const unsigned nbFinalizeSamples = (unsigned)(ctx.nbTrainSamples * ctx.accelParams.finalize / 100); const unsigned nbFinalizeSamples = (unsigned)(ctx.nbTrainSamples * ctx.accelParams.finalize / 100);
const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k; const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k;
const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k; const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k;
const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps; const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps;
const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1); const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1);
const unsigned kIterations = const unsigned kIterations =
(1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize); (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);
const unsigned f = parameters->f == 0 ? DEFAULT_F : parameters->f; const unsigned f = parameters->f == 0 ? DEFAULT_F : parameters->f;
const unsigned accel = parameters->accel == 0 ? DEFAULT_ACCEL : parameters->accel; const unsigned accel = parameters->accel == 0 ? DEFAULT_ACCEL : parameters->accel;
const unsigned shrinkDict = 0;
/* Local variables */ /* Local variables */
const int displayLevel = parameters->zParams.notificationLevel; const int displayLevel = parameters->zParams.notificationLevel;
unsigned iteration = 1; unsigned iteration = 1;
unsigned d; unsigned d;
unsigned k; unsigned k;
COVER_best_t best; COVER_best_t best;
POOL_ctx *pool = NULL; POOL_ctx *pool = NULL;
int warned = 0;
/* Checks */ /* Checks */
if (splitPoint <= 0 || splitPoint > 1) { if (splitPoint <= 0 || splitPoint > 1) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect splitPoint\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect splitPoint\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (accel == 0 || accel > FASTCOVER_MAX_ACCEL) { if (accel == 0 || accel > FASTCOVER_MAX_ACCEL) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect accel\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect accel\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (kMinK < kMaxD || kMaxK < kMinK) { if (kMinK < kMaxD || kMaxK < kMinK) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect k\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect k\n");
return ERROR(GENERIC); return ERROR(parameter_outOfBound);
} }
if (nbSamples == 0) { if (nbSamples == 0) {
LOCALDISPLAYLEVEL(displayLevel, 1, "FASTCOVER must have at least one input file\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "FASTCOVER must have at least one input file\n");
return ERROR(GENERIC); return ERROR(srcSize_wrong);
} }
if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) { if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
LOCALDISPLAYLEVEL(displayLevel, 1, "dictBufferCapacity must be at least %u\n", LOCALDISPLAYLEVEL(displayLevel, 1, "dictBufferCapacity must be at least %u\n",
ZDICT_DICTSIZE_MIN); ZDICT_DICTSIZE_MIN);
return ERROR(dstSize_tooSmall); return ERROR(dstSize_tooSmall);
} }
if (nbThreads > 1) { if (nbThreads > 1) {
pool = POOL_create(nbThreads, 1); pool = POOL_create(nbThreads, 1);
g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1; g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1;
/* Loop through d first because each new value needs a new context */ /* Loop through d first because each new value needs a new context */
LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n", LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n",
kIterations); kIterations);
for (d = kMinD; d <= kMaxD; d += 2) { for (d = kMinD; d <= kMaxD; d += 2) {
/* Initialize the context for this value of d */ /* Initialize the context for this value of d */
FASTCOVER_ctx_t ctx; FASTCOVER_ctx_t ctx;
LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d); LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);
if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint, f, accelParams)) { {
size_t const initVal = FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint, f, accelParams);
if (ZSTD_isError(initVal)) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
COVER_best_destroy(&best); COVER_best_destroy(&best);
POOL_free(pool); POOL_free(pool);
return ERROR(GENERIC); return initVal;
}
}
if (!warned) {
COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.nbDmers, displayLevel);
warned = 1;
} }
/* Loop through k reusing the same context */ /* Loop through k reusing the same context */
for (k = kMinK; k <= kMaxK; k += kStepSize) { for (k = kMinK; k <= kMaxK; k += kStepSize) {
/* Prepare the arguments */ /* Prepare the arguments */
FASTCOVER_tryParameters_data_t *data = (FASTCOVER_tryParameters_data_t *)malloc( FASTCOVER_tryParameters_data_t *data = (FASTCOVER_tryParameters_data_t *)malloc(
sizeof(FASTCOVER_tryParameters_data_t)); sizeof(FASTCOVER_tryParameters_data_t));
LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k); LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k);
if (!data) { if (!data) {
LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n"); LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n");
COVER_best_destroy(&best); COVER_best_destroy(&best);
FASTCOVER_ctx_destroy(&ctx); FASTCOVER_ctx_destroy(&ctx);
POOL_free(pool); POOL_free(pool);
return ERROR(GENERIC); return ERROR(memory_allocation);
} }
data->ctx = &ctx; data->ctx = &ctx;
data->best = &best; data->best = &best;
data->dictBufferCapacity = dictBufferCapacity; data->dictBufferCapacity = dictBufferCapacity;
data->parameters = coverParams; data->parameters = coverParams;
data->parameters.k = k; data->parameters.k = k;
data->parameters.d = d; data->parameters.d = d;
data->parameters.splitPoint = splitPoint; data->parameters.splitPoint = splitPoint;
data->parameters.steps = kSteps; data->parameters.steps = kSteps;
data->parameters.shrinkDict = shrinkDict;
data->parameters.zParams.notificationLevel = g_displayLevel; data->parameters.zParams.notificationLevel = g_displayLevel;
/* Check the parameters */ /* Check the parameters */
if (!FASTCOVER_checkParameters(data->parameters, dictBufferCapacity, if (!FASTCOVER_checkParameters(data->parameters, dictBufferCapacity,
data->ctx->f, accel)) { data->ctx->f, accel)) {
DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n"); DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n");
free(data); free(data);
continue; continue;
} }

contrib/python-zstandard/zstd/dictBuilder/zdict.h

* Train a dictionary from an array of samples. * Train a dictionary from an array of samples.
* Redirect towards ZDICT_optimizeTrainFromBuffer_fastCover() single-threaded, with d=8, steps=4, * Redirect towards ZDICT_optimizeTrainFromBuffer_fastCover() single-threaded, with d=8, steps=4,
* f=20, and accel=1. * f=20, and accel=1.
* Samples must be stored concatenated in a single flat buffer `samplesBuffer`, * 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. * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
* The resulting dictionary will be saved into `dictBuffer`. * The resulting dictionary will be saved into `dictBuffer`.
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`) * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
* or an error code, which can be tested with ZDICT_isError(). * or an error code, which can be tested with ZDICT_isError().
* Note: ZDICT_trainFromBuffer() requires about 9 bytes of memory for each input byte. * 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. * 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`. * 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. * 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. * 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, ZDICTLIB_API size_t ZDICT_trainFromBuffer(void* dictBuffer, size_t dictBufferCapacity,
const void* samplesBuffer, const void* samplesBuffer,
const size_t* samplesSizes, unsigned nbSamples); const size_t* samplesSizes, unsigned nbSamples);
* For others, value 0 means default. * For others, value 0 means default.
*/ */
typedef struct { typedef struct {
unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */ unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
unsigned d; /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */ unsigned d; /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */
unsigned steps; /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */ 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 */ 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 */ 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 */
unsigned shrinkDict; /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking */
unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
ZDICT_params_t zParams; ZDICT_params_t zParams;
} ZDICT_cover_params_t; } ZDICT_cover_params_t;
typedef struct { typedef struct {
unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */ unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
unsigned d; /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */ 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 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 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 */ 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 */ 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) */ unsigned accel; /* Acceleration level: constraint: 0 < accel <= 10, higher means faster and less accurate, 0 means default(1) */
unsigned shrinkDict; /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking */
unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
ZDICT_params_t zParams; ZDICT_params_t zParams;
} ZDICT_fastCover_params_t; } ZDICT_fastCover_params_t;
/*! ZDICT_trainFromBuffer_cover(): /*! ZDICT_trainFromBuffer_cover():
* Train a dictionary from an array of samples using the COVER algorithm. * Train a dictionary from an array of samples using the COVER algorithm.
* Samples must be stored concatenated in a single flat buffer `samplesBuffer`, * 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. * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
* The resulting dictionary will be saved into `dictBuffer`. * The resulting dictionary will be saved into `dictBuffer`.
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`) * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
* or an error code, which can be tested with ZDICT_isError(). * 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. * 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. * 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`. * 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. * 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. * 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( ZDICTLIB_API size_t ZDICT_trainFromBuffer_cover(
void *dictBuffer, size_t dictBufferCapacity, void *dictBuffer, size_t dictBufferCapacity,
const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples, const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
ZDICT_cover_params_t parameters); ZDICT_cover_params_t parameters);
/*! ZDICT_optimizeTrainFromBuffer_cover(): /*! ZDICT_optimizeTrainFromBuffer_cover():
* The same requirements as above hold for all the parameters except `parameters`. * The same requirements as above hold for all the parameters except `parameters`.
* This function tries many parameter combinations and picks the best parameters. * This function tries many parameter combinations and picks the best parameters.
* `*parameters` is filled with the best parameters found, * `*parameters` is filled with the best parameters found,
* dictionary constructed with those parameters is stored in `dictBuffer`. * dictionary constructed with those parameters is stored in `dictBuffer`.
* *
* All of the parameters d, k, steps are optional. * 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 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 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 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`) * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
* or an error code, which can be tested with ZDICT_isError(). * or an error code, which can be tested with ZDICT_isError().
* On success `*parameters` contains the parameters selected. * 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. * 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( ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(
void* dictBuffer, size_t dictBufferCapacity, void* dictBuffer, size_t dictBufferCapacity,
const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples, const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
ZDICT_cover_params_t* parameters); ZDICT_cover_params_t* parameters);
/*! ZDICT_trainFromBuffer_fastCover(): /*! ZDICT_trainFromBuffer_fastCover():
* Train a dictionary from an array of samples using a modified version of COVER algorithm. * 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`, * 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. * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
* d and k are required. * d and k are required.
* All other parameters are optional, will use default values if not provided * All other parameters are optional, will use default values if not provided
* The resulting dictionary will be saved into `dictBuffer`. * The resulting dictionary will be saved into `dictBuffer`.
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`) * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
* or an error code, which can be tested with ZDICT_isError(). * or an error code, which can be tested with ZDICT_isError().
* Note: ZDICT_trainFromBuffer_fastCover() requires about 1 bytes of memory for each input byte and additionally another 6 * 2^f bytes of memory . * 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. * 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`. * 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. * 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. * 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, ZDICTLIB_API size_t ZDICT_trainFromBuffer_fastCover(void *dictBuffer,
size_t dictBufferCapacity, const void *samplesBuffer, size_t dictBufferCapacity, const void *samplesBuffer,
const size_t *samplesSizes, unsigned nbSamples, const size_t *samplesSizes, unsigned nbSamples,
ZDICT_fastCover_params_t parameters); ZDICT_fastCover_params_t parameters);
/*! ZDICT_optimizeTrainFromBuffer_fastCover(): /*! ZDICT_optimizeTrainFromBuffer_fastCover():
* The same requirements as above hold for all the parameters except `parameters`. * The same requirements as above hold for all the parameters except `parameters`.
* This function tries many parameter combinations (specifically, k and d combinations) * This function tries many parameter combinations (specifically, k and d combinations)
* and picks the best parameters. `*parameters` is filled with the best parameters found, * and picks the best parameters. `*parameters` is filled with the best parameters found,
* dictionary constructed with those parameters is stored in `dictBuffer`. * dictionary constructed with those parameters is stored in `dictBuffer`.
* All of the parameters d, k, steps, f, and accel are optional. * 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 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 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 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 f is zero, default value of 20 is used.
* If accel is zero, default value of 1 is used. * If accel is zero, default value of 1 is used.
* *
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`) * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
* or an error code, which can be tested with ZDICT_isError(). * or an error code, which can be tested with ZDICT_isError().
* On success `*parameters` contains the parameters selected. * On success `*parameters` contains the parameters selected.
* Note: ZDICT_optimizeTrainFromBuffer_fastCover() requires about 1 byte of memory for each input byte and additionally another 6 * 2^f bytes of memory for each thread. * 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, ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(void* dictBuffer,
size_t dictBufferCapacity, const void* samplesBuffer, size_t dictBufferCapacity, const void* samplesBuffer,
const size_t* samplesSizes, unsigned nbSamples, const size_t* samplesSizes, unsigned nbSamples,
ZDICT_fastCover_params_t* parameters); ZDICT_fastCover_params_t* parameters);
/*! ZDICT_finalizeDictionary(): /*! ZDICT_finalizeDictionary():
* Given a custom content as a basis for dictionary, and a set of samples, * Given a custom content as a basis for dictionary, and a set of samples,
* finalize dictionary by adding headers and statistics. * finalize dictionary by adding headers and statistics.
* *
* Samples must be stored concatenated in a flat buffer `samplesBuffer`, * 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. * supplied with an array of sizes `samplesSizes`, providing the size of each sample in order.
* *
* dictContentSize must be >= ZDICT_CONTENTSIZE_MIN bytes. * dictContentSize must be >= ZDICT_CONTENTSIZE_MIN bytes.
* maxDictSize must be >= dictContentSize, and must be >= ZDICT_DICTSIZE_MIN bytes. * maxDictSize must be >= dictContentSize, and must be >= ZDICT_DICTSIZE_MIN bytes.
* *
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`), * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`),
* or an error code, which can be tested by ZDICT_isError(). * 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: ZDICT_finalizeDictionary() will push notifications into stderr if instructed to, using notificationLevel>0.
* Note 2: dictBuffer and dictContent can overlap * Note 2: dictBuffer and dictContent can overlap
*/ */
#define ZDICT_CONTENTSIZE_MIN 128 #define ZDICT_CONTENTSIZE_MIN 128
#define ZDICT_DICTSIZE_MIN 256 #define ZDICT_DICTSIZE_MIN 256
ZDICTLIB_API size_t ZDICT_finalizeDictionary(void* dictBuffer, size_t dictBufferCapacity, ZDICTLIB_API size_t ZDICT_finalizeDictionary(void* dictBuffer, size_t dictBufferCapacity,
const void* dictContent, size_t dictContentSize, const void* dictContent, size_t dictContentSize,
const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples, const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
ZDICT_params_t parameters); ZDICT_params_t parameters);
typedef struct { typedef struct {
unsigned selectivityLevel; /* 0 means default; larger => select more => larger dictionary */ unsigned selectivityLevel; /* 0 means default; larger => select more => larger dictionary */
ZDICT_params_t zParams; ZDICT_params_t zParams;
} ZDICT_legacy_params_t; } ZDICT_legacy_params_t;
/*! ZDICT_trainFromBuffer_legacy(): /*! ZDICT_trainFromBuffer_legacy():
* Train a dictionary from an array of samples. * Train a dictionary from an array of samples.
* Samples must be stored concatenated in a single flat buffer `samplesBuffer`, * 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. * supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
* The resulting dictionary will be saved into `dictBuffer`. * The resulting dictionary will be saved into `dictBuffer`.
* `parameters` is optional and can be provided with values set to 0 to mean "default". * `parameters` is optional and can be provided with values set to 0 to mean "default".
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`) * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
* or an error code, which can be tested with ZDICT_isError(). * 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. * 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`. * 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. * 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. * 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. * Note: ZDICT_trainFromBuffer_legacy() will send notifications into stderr if instructed to, using notificationLevel>0.
*/ */
ZDICTLIB_API size_t ZDICT_trainFromBuffer_legacy( ZDICTLIB_API size_t ZDICT_trainFromBuffer_legacy(
void *dictBuffer, size_t dictBufferCapacity, void *dictBuffer, size_t dictBufferCapacity,

contrib/python-zstandard/zstd/dictBuilder/zdict.c

(const char*)srcBuffer + pos, fileSizes[u], (const char*)srcBuffer + pos, fileSizes[u],
notificationLevel); notificationLevel);
pos += fileSizes[u]; pos += fileSizes[u];
} }
/* analyze, build stats, starting with literals */ /* analyze, build stats, starting with literals */
{ size_t maxNbBits = HUF_buildCTable (hufTable, countLit, 255, huffLog); { size_t maxNbBits = HUF_buildCTable (hufTable, countLit, 255, huffLog);
if (HUF_isError(maxNbBits)) { if (HUF_isError(maxNbBits)) {
eSize = ERROR(GENERIC); eSize = maxNbBits;
DISPLAYLEVEL(1, " HUF_buildCTable error \n"); DISPLAYLEVEL(1, " HUF_buildCTable error \n");
goto _cleanup; goto _cleanup;
} }
if (maxNbBits==8) { /* not compressible : will fail on HUF_writeCTable() */ 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"); 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 */ 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); maxNbBits = HUF_buildCTable (hufTable, countLit, 255, huffLog);
assert(maxNbBits==9); assert(maxNbBits==9);
} }
huffLog = (U32)maxNbBits; huffLog = (U32)maxNbBits;
} }
/* looking for most common first offsets */ /* looking for most common first offsets */
{ U32 offset; { U32 offset;
for (offset=1; offset<MAXREPOFFSET; offset++) for (offset=1; offset<MAXREPOFFSET; offset++)
ZDICT_insertSortCount(bestRepOffset, offset, repOffset[offset]); ZDICT_insertSortCount(bestRepOffset, offset, repOffset[offset]);
} }
/* note : the result of this phase should be used to better appreciate the impact on statistics */ /* note : the result of this phase should be used to better appreciate the impact on statistics */
total=0; for (u=0; u<=offcodeMax; u++) total+=offcodeCount[u]; total=0; for (u=0; u<=offcodeMax; u++) total+=offcodeCount[u];
errorCode = FSE_normalizeCount(offcodeNCount, Offlog, offcodeCount, total, offcodeMax); errorCode = FSE_normalizeCount(offcodeNCount, Offlog, offcodeCount, total, offcodeMax);
if (FSE_isError(errorCode)) { if (FSE_isError(errorCode)) {
eSize = ERROR(GENERIC); eSize = errorCode;
DISPLAYLEVEL(1, "FSE_normalizeCount error with offcodeCount \n"); DISPLAYLEVEL(1, "FSE_normalizeCount error with offcodeCount \n");
goto _cleanup; goto _cleanup;
} }
Offlog = (U32)errorCode; Offlog = (U32)errorCode;
total=0; for (u=0; u<=MaxML; u++) total+=matchLengthCount[u]; total=0; for (u=0; u<=MaxML; u++) total+=matchLengthCount[u];
errorCode = FSE_normalizeCount(matchLengthNCount, mlLog, matchLengthCount, total, MaxML); errorCode = FSE_normalizeCount(matchLengthNCount, mlLog, matchLengthCount, total, MaxML);
if (FSE_isError(errorCode)) { if (FSE_isError(errorCode)) {
eSize = ERROR(GENERIC); eSize = errorCode;
DISPLAYLEVEL(1, "FSE_normalizeCount error with matchLengthCount \n"); DISPLAYLEVEL(1, "FSE_normalizeCount error with matchLengthCount \n");
goto _cleanup; goto _cleanup;
} }
mlLog = (U32)errorCode; mlLog = (U32)errorCode;
total=0; for (u=0; u<=MaxLL; u++) total+=litLengthCount[u]; total=0; for (u=0; u<=MaxLL; u++) total+=litLengthCount[u];
errorCode = FSE_normalizeCount(litLengthNCount, llLog, litLengthCount, total, MaxLL); errorCode = FSE_normalizeCount(litLengthNCount, llLog, litLengthCount, total, MaxLL);
if (FSE_isError(errorCode)) { if (FSE_isError(errorCode)) {
eSize = ERROR(GENERIC); eSize = errorCode;
DISPLAYLEVEL(1, "FSE_normalizeCount error with litLengthCount \n"); DISPLAYLEVEL(1, "FSE_normalizeCount error with litLengthCount \n");
goto _cleanup; goto _cleanup;
} }
llLog = (U32)errorCode; llLog = (U32)errorCode;
/* write result to buffer */ /* write result to buffer */
{ size_t const hhSize = HUF_writeCTable(dstPtr, maxDstSize, hufTable, 255, huffLog); { size_t const hhSize = HUF_writeCTable(dstPtr, maxDstSize, hufTable, 255, huffLog);
if (HUF_isError(hhSize)) { if (HUF_isError(hhSize)) {
eSize = ERROR(GENERIC); eSize = hhSize;
DISPLAYLEVEL(1, "HUF_writeCTable error \n"); DISPLAYLEVEL(1, "HUF_writeCTable error \n");
goto _cleanup; goto _cleanup;
} }
dstPtr += hhSize; dstPtr += hhSize;
maxDstSize -= hhSize; maxDstSize -= hhSize;
eSize += hhSize; eSize += hhSize;
} }
{ size_t const ohSize = FSE_writeNCount(dstPtr, maxDstSize, offcodeNCount, OFFCODE_MAX, Offlog); { size_t const ohSize = FSE_writeNCount(dstPtr, maxDstSize, offcodeNCount, OFFCODE_MAX, Offlog);
if (FSE_isError(ohSize)) { if (FSE_isError(ohSize)) {
eSize = ERROR(GENERIC); eSize = ohSize;
DISPLAYLEVEL(1, "FSE_writeNCount error with offcodeNCount \n"); DISPLAYLEVEL(1, "FSE_writeNCount error with offcodeNCount \n");
goto _cleanup; goto _cleanup;
} }
dstPtr += ohSize; dstPtr += ohSize;
maxDstSize -= ohSize; maxDstSize -= ohSize;
eSize += ohSize; eSize += ohSize;
} }
{ size_t const mhSize = FSE_writeNCount(dstPtr, maxDstSize, matchLengthNCount, MaxML, mlLog); { size_t const mhSize = FSE_writeNCount(dstPtr, maxDstSize, matchLengthNCount, MaxML, mlLog);
if (FSE_isError(mhSize)) { if (FSE_isError(mhSize)) {
eSize = ERROR(GENERIC); eSize = mhSize;
DISPLAYLEVEL(1, "FSE_writeNCount error with matchLengthNCount \n"); DISPLAYLEVEL(1, "FSE_writeNCount error with matchLengthNCount \n");
goto _cleanup; goto _cleanup;
} }
dstPtr += mhSize; dstPtr += mhSize;
maxDstSize -= mhSize; maxDstSize -= mhSize;
eSize += mhSize; eSize += mhSize;
} }
{ size_t const lhSize = FSE_writeNCount(dstPtr, maxDstSize, litLengthNCount, MaxLL, llLog); { size_t const lhSize = FSE_writeNCount(dstPtr, maxDstSize, litLengthNCount, MaxLL, llLog);
if (FSE_isError(lhSize)) { if (FSE_isError(lhSize)) {
eSize = ERROR(GENERIC); eSize = lhSize;
DISPLAYLEVEL(1, "FSE_writeNCount error with litlengthNCount \n"); DISPLAYLEVEL(1, "FSE_writeNCount error with litlengthNCount \n");
goto _cleanup; goto _cleanup;
} }
dstPtr += lhSize; dstPtr += lhSize;
maxDstSize -= lhSize; maxDstSize -= lhSize;
eSize += lhSize; eSize += lhSize;
} }
if (maxDstSize<12) { if (maxDstSize<12) {
eSize = ERROR(GENERIC); eSize = ERROR(dstSize_tooSmall);
DISPLAYLEVEL(1, "not enough space to write RepOffsets \n"); DISPLAYLEVEL(1, "not enough space to write RepOffsets \n");
goto _cleanup; goto _cleanup;
} }
# if 0 # if 0
MEM_writeLE32(dstPtr+0, bestRepOffset[0].offset); MEM_writeLE32(dstPtr+0, bestRepOffset[0].offset);
MEM_writeLE32(dstPtr+4, bestRepOffset[1].offset); MEM_writeLE32(dstPtr+4, bestRepOffset[1].offset);
MEM_writeLE32(dstPtr+8, bestRepOffset[2].offset); MEM_writeLE32(dstPtr+8, bestRepOffset[2].offset);
#else #else

contrib/python-zstandard/zstd/zstd.h

Advanced experimental APIs should never be used with a dynamically-linked Advanced experimental APIs should never be used with a dynamically-linked
library. They are not "stable"; their definitions or signatures may change in library. They are not "stable"; their definitions or signatures may change in
the future. Only static linking is allowed. the future. Only static linking is allowed.
*******************************************************************************/ *******************************************************************************/
/*------ Version ------*/ /*------ Version ------*/
#define ZSTD_VERSION_MAJOR 1 #define ZSTD_VERSION_MAJOR 1
#define ZSTD_VERSION_MINOR 3 #define ZSTD_VERSION_MINOR 4
#define ZSTD_VERSION_RELEASE 8 #define ZSTD_VERSION_RELEASE 3
#define ZSTD_VERSION_NUMBER (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE) #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 */ 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_LIB_VERSION ZSTD_VERSION_MAJOR.ZSTD_VERSION_MINOR.ZSTD_VERSION_RELEASE
#define ZSTD_QUOTE(str) #str #define ZSTD_QUOTE(str) #str
#define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str) #define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str)
#define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION) #define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION)
ZSTDLIB_API const char* ZSTD_versionString(void); /* requires v1.3.0+ */ ZSTDLIB_API const char* ZSTD_versionString(void); /* requires v1.3.0+ */
/*************************************** /* *************************************
* Default constant * Default constant
***************************************/ ***************************************/
#ifndef ZSTD_CLEVEL_DEFAULT #ifndef ZSTD_CLEVEL_DEFAULT
# define ZSTD_CLEVEL_DEFAULT 3 # define ZSTD_CLEVEL_DEFAULT 3
#endif #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 * Simple API
***************************************/ ***************************************/
/*! ZSTD_compress() : /*! ZSTD_compress() :
* Compresses `src` content as a single zstd compressed frame into already allocated `dst`. * Compresses `src` content as a single zstd compressed frame into already allocated `dst`.
* Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`. * Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.
* @return : compressed size written into `dst` (<= `dstCapacity), * @return : compressed size written into `dst` (<= `dstCapacity),
* or an error code if it fails (which can be tested using ZSTD_isError()). */ * or an error code if it fails (which can be tested using ZSTD_isError()). */
/*! ZSTD_getDecompressedSize() : /*! ZSTD_getDecompressedSize() :
* NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize(). * NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize().
* Both functions work the same way, but ZSTD_getDecompressedSize() blends * Both functions work the same way, but ZSTD_getDecompressedSize() blends
* "empty", "unknown" and "error" results to the same return value (0), * "empty", "unknown" and "error" results to the same return value (0),
* while ZSTD_getFrameContentSize() gives them separate return values. * while ZSTD_getFrameContentSize() gives them separate return values.
* @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise. */ * @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); 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 ======*/ /*====== 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 */ #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 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 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 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 */ ZSTDLIB_API int ZSTD_maxCLevel(void); /*!< maximum compression level available */
/*************************************** /***************************************
* Explicit context * Explicit context
***************************************/ ***************************************/
/*= Compression context /*= Compression context
* When compressing many times, * When compressing many times,
* it is recommended to allocate a context just once, and re-use it for each successive compression operation. * 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. * This will make workload friendlier for system's memory.
* Use one context per thread for parallel execution in multi-threaded environments. */ * Note : re-using context is just a speed / resource optimization.
* It doesn't change the compression ratio, which remains identical.
* Note 2 : In multi-threaded environments,
* use one different context per thread for parallel execution.
*/
typedef struct ZSTD_CCtx_s ZSTD_CCtx; typedef struct ZSTD_CCtx_s ZSTD_CCtx;
ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void); ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void);
ZSTDLIB_API size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx); ZSTDLIB_API size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx);
/*! ZSTD_compressCCtx() : /*! ZSTD_compressCCtx() :
* Same as ZSTD_compress(), using an explicit ZSTD_CCtx * Same as ZSTD_compress(), using an explicit ZSTD_CCtx
* The function will compress at requested compression level, * The function will compress at requested compression level,
* ignoring any other parameter */ * ignoring any other parameter */
* requires an allocated ZSTD_DCtx. * requires an allocated ZSTD_DCtx.
* Compatible with sticky parameters. * Compatible with sticky parameters.
*/ */
ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx, ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize); const void* src, size_t srcSize);
/************************** /***************************************
* Simple dictionary API * Advanced compression 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() : /* API design :
* Decompression using a known Dictionary. * Parameters are pushed one by one into an existing context,
* Dictionary must be identical to the one used during compression. * using ZSTD_CCtx_set*() functions.
* Note : This function loads the dictionary, resulting in significant startup delay. * Pushed parameters are sticky : they are valid for next compressed frame, and any subsequent frame.
* It's intended for a dictionary used only once. * "sticky" parameters are applicable to `ZSTD_compress2()` and `ZSTD_compressStream*()` !
* Note : When `dict == NULL || dictSize < 8` no dictionary is used. */ * They do not apply to "simple" one-shot variants such as ZSTD_compressCCtx()
ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx, *
void* dst, size_t dstCapacity, * It's possible to reset all parameters to "default" using ZSTD_CCtx_reset().
const void* src, size_t srcSize, *
const void* dict,size_t dictSize); * 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 */
* Bulk processing dictionary API typedef enum { ZSTD_fast=1,
**********************************/ ZSTD_dfast=2,
typedef struct ZSTD_CDict_s ZSTD_CDict; 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;
/*! 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() : typedef enum {
* Function frees memory allocated by ZSTD_createCDict(). */
ZSTDLIB_API size_t ZSTD_freeCDict(ZSTD_CDict* CDict);
/*! ZSTD_compress_usingCDict() : /* compression parameters
* Compression using a digested Dictionary. * Note: When compressing with a ZSTD_CDict these parameters are superseded
* Recommended when same dictionary is used multiple times. * by the parameters used to construct the ZSTD_CDict. See ZSTD_CCtx_refCDict()
* Note : compression level is _decided at dictionary creation time_, * for more info (superseded-by-cdict). */
* and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */ ZSTD_c_compressionLevel=100, /* Update all compression parameters according to pre-defined cLevel table
ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx, * Default level is ZSTD_CLEVEL_DEFAULT==3.
void* dst, size_t dstCapacity, * Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT.
const void* src, size_t srcSize, * Note 1 : it's possible to pass a negative compression level.
const ZSTD_CDict* cdict); * 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". */
typedef struct ZSTD_DDict_s ZSTD_DDict; /* 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) */
/*! ZSTD_createDDict() : /* multi-threading parameters */
* Create a digested dictionary, ready to start decompression operation without startup delay. /* These parameters are only useful if multi-threading is enabled (compiled with build macro ZSTD_MULTITHREAD).
* dictBuffer can be released after DDict creation, as its content is copied inside DDict. */ * They return an error otherwise. */
ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize); 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 */
/*! ZSTD_freeDDict() : /* note : additional experimental parameters are also available
* Function frees memory allocated with ZSTD_createDDict() */ * within the experimental section of the API.
ZSTDLIB_API size_t ZSTD_freeDDict(ZSTD_DDict* ddict); * At the time of this writing, they include :
* ZSTD_c_rsyncable
* ZSTD_c_format
* ZSTD_c_forceMaxWindow
* ZSTD_c_forceAttachDict
* ZSTD_c_literalCompressionMode
* ZSTD_c_targetCBlockSize
* 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_c_experimentalParam6=1003,
} ZSTD_cParameter;
/*! ZSTD_decompress_usingDDict() : typedef struct {
* Decompression using a digested Dictionary. size_t error;
* Recommended when same dictionary is used multiple times. */ int lowerBound;
ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx, int upperBound;
void* dst, size_t dstCapacity, } ZSTD_bounds;
const void* src, size_t srcSize,
const ZSTD_DDict* ddict); /*! 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 * Streaming
****************************/ ****************************/
typedef struct ZSTD_inBuffer_s { typedef struct ZSTD_inBuffer_s {
const void* src; /**< start of input buffer */ const void* src; /**< start of input buffer */
* *
* For parallel execution, use one separate ZSTD_CStream per thread. * For parallel execution, use one separate ZSTD_CStream per thread.
* *
* note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing. * 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, * Parameters are sticky : when starting a new compression on the same context,
* it will re-use the same sticky parameters as previous compression session. * it will re-use the same sticky parameters as previous compression session.
* When in doubt, it's recommended to fully initialize the context before usage. * When in doubt, it's recommended to fully initialize the context before usage.
* Use ZSTD_initCStream() to set the parameter to a selected compression level. * Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),
* Use advanced API (ZSTD_CCtx_setParameter(), etc.) to set more specific parameters. * ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to
* * set more specific parameters, the pledged source size, or load a dictionary.
* Use ZSTD_compressStream() as many times as necessary to consume input stream. *
* The function will automatically update both `pos` fields within `input` and `output`. * Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to
* Note that the function may not consume the entire input, * consume input stream. The function will automatically update both `pos`
* for example, because the output buffer is already full, * fields within `input` and `output`.
* in which case `input.pos < input.size`. * 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. * The caller must check if input has been entirely consumed.
* If not, the caller must make some room to receive more compressed data, * If not, the caller must make some room to receive more compressed data,
* and then present again remaining input data. * and then present again remaining input data.
* @return : a size hint, preferred nb of bytes to use as input for next function call * 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(). * or an error code, which can be tested using ZSTD_isError().
* Note 1 : it's just a hint, to help latency a little, any value will work fine.
* Note 2 : size hint is guaranteed to be <= ZSTD_CStreamInSize()
* *
* At any moment, it's possible to flush whatever data might remain stuck within internal buffer, * At any moment, it's possible to flush whatever data might remain stuck within internal buffer,
* using ZSTD_flushStream(). `output->pos` will be updated. * using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated.
* Note that, if `output->size` is too small, a single invocation of ZSTD_flushStream() might not be enough (return code > 0). * 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_flushStream(). * 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, * @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), * >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(). * or an error code, which can be tested using ZSTD_isError().
* *
* ZSTD_endStream() instructs to finish a frame. * Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame.
* It will perform a flush and write frame epilogue. * It will perform a flush and write frame epilogue.
* The epilogue is required for decoders to consider a frame completed. * The epilogue is required for decoders to consider a frame completed.
* flush() operation is the same, and follows same rules as ZSTD_flushStream(). * 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, * @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), * >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(). * 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) */ 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 */ /* Continue to distinguish them for compatibility with older versions <= v1.2.0 */
/*===== ZSTD_CStream management functions =====*/ /*===== ZSTD_CStream management functions =====*/
ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void); ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);
ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs); ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);
/*===== Streaming compression functions =====*/ /*===== 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);
/* These buffer sizes are softly recommended.
* They are not required : ZSTD_compressStream*() happily accepts any buffer size, for both input and output.
* Respecting the recommended size just makes it a bit easier for ZSTD_compressStream*(),
* reducing the amount of memory shuffling and buffering, resulting in minor performance savings.
*
* However, note that these recommendations are from the perspective of a C caller program.
* If the streaming interface is invoked from some other language,
* especially managed ones such as Java or Go, through a foreign function interface such as jni or cgo,
* a major performance rule is to reduce crossing such interface to an absolute minimum.
* It's not rare that performance ends being spent more into the interface, rather than compression itself.
* In which cases, prefer using large buffers, as large as practical,
* for both input and output, to reduce the nb of roundtrips.
*/
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. */
/* *****************************************************************************
* This following is a legacy streaming API.
* It can be replaced by ZSTD_CCtx_reset() and ZSTD_compressStream2().
* It is redundant, but remains 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); 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 minimum nb 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); 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); 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); 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 * Streaming decompression - HowTo
* *
* A ZSTD_DStream object is required to track streaming operations. * A ZSTD_DStream object is required to track streaming operations.
* Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources. * Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
* ZSTD_DStream objects can be re-used multiple times. * ZSTD_DStream objects can be re-used multiple times.
* *
typedef ZSTD_DCtx ZSTD_DStream; /**< DCtx and DStream are now effectively same object (>= v1.3.0) */ 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. */ /* For compatibility with versions <= v1.2.0, prefer differentiating them. */
/*===== ZSTD_DStream management functions =====*/ /*===== ZSTD_DStream management functions =====*/
ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void); ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);
ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds); ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);
/*===== Streaming decompression functions =====*/ /*===== 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_initDStream(ZSTD_DStream* zds);
ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input); 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_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. */ 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 */
/****************************************************************************************
* 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
/****************************************************************************************
* Candidate API for promotion to stable status
****************************************************************************************
* The following symbols and constants form the "staging area" :
* they are considered to join "stable API" by v1.4.0.
* The proposal is written so that it can be made stable "as is",
* though it's still possible to suggest improvements.
* Staging is in fact last chance for changes,
* the API is locked once reaching "stable" status.
* ***************************************************************************************/
/* === 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)
/* === query limits === */
ZSTDLIB_API int ZSTD_minCLevel(void); /*!< minimum negative compression level allowed */
/* === frame size === */
/*! 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);
/**************************
* 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);
/* === Memory management === */ /*! 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);
/*! 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);
/***********************************
* Bulk processing dictionary API
**********************************/
typedef struct ZSTD_CDict_s ZSTD_CDict;
/*************************************** /*! ZSTD_createCDict() :
* Advanced compression API * 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);
/* API design : /*! ZSTD_freeCDict() :
* Parameters are pushed one by one into an existing context, * Function frees memory allocated by ZSTD_createCDict(). */
* using ZSTD_CCtx_set*() functions. ZSTDLIB_API size_t ZSTD_freeCDict(ZSTD_CDict* CDict);
* 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.
*/
/*! 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);
/* 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 struct ZSTD_DDict_s ZSTD_DDict;
typedef enum { /*! 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);
/* compression parameters */ /*! ZSTD_freeDDict() :
ZSTD_c_compressionLevel=100, /* Update all compression parameters according to pre-defined cLevel table * Function frees memory allocated with ZSTD_createDDict() */
* Default level is ZSTD_CLEVEL_DEFAULT==3. ZSTDLIB_API size_t ZSTD_freeDDict(ZSTD_DDict* ddict);
* 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_decompress_usingDDict() :
ZSTD_c_enableLongDistanceMatching=160, /* Enable long distance matching. * Decompression using a digested Dictionary.
* This parameter is designed to improve compression ratio * Recommended when same dictionary is used multiple times. */
* for large inputs, by finding large matches at long distance. ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
* It increases memory usage and window size. void* dst, size_t dstCapacity,
* Note: enabling this parameter increases default ZSTD_c_windowLog to 128 MB const void* src, size_t srcSize,
* except when expressly set to a different value. */ const ZSTD_DDict* ddict);
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). * Dictionary helper functions
* 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 /*! ZSTD_getDictID_fromDict() :
* within the experimental section of the API. * Provides the dictID stored within dictionary.
* At the time of this writing, they include : * if @return == 0, the dictionary is not conformant with Zstandard specification.
* ZSTD_c_rsyncable * It can still be loaded, but as a content-only dictionary. */
* ZSTD_c_format ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
* ZSTD_c_forceMaxWindow
* ZSTD_c_forceAttachDict
* 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_cParameter;
/*! 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);
typedef struct { /*! ZSTD_getDictID_fromFrame() :
size_t error; * Provides the dictID required to decompressed the frame stored within `src`.
int lowerBound; * If @return == 0, the dictID could not be decoded.
int upperBound; * This could for one of the following reasons :
} ZSTD_bounds; * - 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);
/*! 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. * Advanced dictionary and prefix API
* 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). * This API allows dictionaries to be used with ZSTD_compress2(),
* Setting a parameter is generally only possible during frame initialization (before starting compression). * ZSTD_compressStream2(), and ZSTD_decompress(). Dictionaries are sticky, and
* Exception : when using multi-threading mode (nbWorkers >= 1), * only reset with the context is reset with ZSTD_reset_parameters or
* the following parameters can be updated _during_ compression (within same frame): * ZSTD_reset_session_and_parameters. Prefixes are single-use.
* => 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 overriden by srcSize instead.
*/
ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);
/*! ZSTD_CCtx_loadDictionary() : /*! ZSTD_CCtx_loadDictionary() :
* Create an internal CDict from `dict` buffer. * Create an internal CDict from `dict` buffer.
* Decompression will have to use same dictionary. * Decompression will have to use same dictionary.
* @result : 0, or an error code (which can be tested with ZSTD_isError()). * @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary, * Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary,
* meaning "return to no-dictionary mode". * meaning "return to no-dictionary mode".
* Note 1 : Dictionary is sticky, it will be used for all future compressed frames. * 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). * To return to "no-dictionary" situation, load a NULL dictionary (or reset parameters).
* Note 2 : Loading a dictionary involves building tables. * Note 2 : Loading a dictionary involves building tables.
* It's also a CPU consuming operation, with non-negligible impact on latency. * It's also a CPU consuming operation, with non-negligible impact on latency.
* Tables are dependent on compression parameters, and for this reason, * Tables are dependent on compression parameters, and for this reason,
* compression parameters can no longer be changed after loading a dictionary. * compression parameters can no longer be changed after loading a dictionary.
* Note 3 :`dict` content will be copied internally. * Note 3 :`dict` content will be copied internally.
* Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead. * Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead.
* In such a case, dictionary buffer must outlive its users. * In such a case, dictionary buffer must outlive its users.
* Note 4 : Use ZSTD_CCtx_loadDictionary_advanced() * Note 4 : Use ZSTD_CCtx_loadDictionary_advanced()
* to precisely select how dictionary content must be interpreted. */ * to precisely select how dictionary content must be interpreted. */
ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize); ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
/*! ZSTD_CCtx_refCDict() : /*! ZSTD_CCtx_refCDict() :
* Reference a prepared dictionary, to be used for all next compressed frames. * Reference a prepared dictionary, to be used for all next compressed frames.
* Note that compression parameters are enforced from within CDict, * Note that compression parameters are enforced from within CDict,
* and supercede any compression parameter previously set within CCtx. * 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. * 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()). * @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Special : Referencing a NULL CDict means "return to no-dictionary mode". * Special : Referencing a NULL CDict means "return to no-dictionary mode".
* Note 1 : Currently, only one dictionary can be managed. * Note 1 : Currently, only one dictionary can be managed.
* Referencing a new dictionary effectively "discards" any previous one. * Referencing a new dictionary effectively "discards" any previous one.
* Note 2 : CDict is just referenced, its lifetime must outlive its usage within CCtx. */ * 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); ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
* Note 3 : Referencing a prefix involves building tables, which are dependent on compression parameters. * 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. * 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. * 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). * Note 4 : By default, the prefix is interpreted as raw content (ZSTD_dm_rawContent).
* Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation. */ * Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation. */
ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx,
const void* prefix, size_t prefixSize); const void* prefix, size_t prefixSize);
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);
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. */
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). */
} 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)
* - outpot->pos must be <= dstCapacity, input->pos must be <= srcSize
* - outpot->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);
/* ============================== */
/* 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) */
/* 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_loadDictionary() : /*! ZSTD_DCtx_loadDictionary() :
* Create an internal DDict from dict buffer, * Create an internal DDict from dict buffer,
* to be used to decompress next frames. * to be used to decompress next frames.
* The dictionary remains valid for all future frames, until explicitly invalidated. * The dictionary remains valid for all future frames, until explicitly invalidated.
* @result : 0, or an error code (which can be tested with ZSTD_isError()). * @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, * Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,
* meaning "return to no-dictionary mode". * meaning "return to no-dictionary mode".
* Note 1 : Loading a dictionary involves building tables, * Note 1 : Loading a dictionary involves building tables,
* Note 3 : By default, the prefix is treated as raw content (ZSTD_dm_rawContent). * Note 3 : By default, the prefix is treated as raw content (ZSTD_dm_rawContent).
* Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section) * Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section)
* Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost. * 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. * A full dictionary is more costly, as it requires building tables.
*/ */
ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx,
const void* prefix, size_t prefixSize); const void* prefix, size_t prefixSize);
/*! ZSTD_DCtx_reset() : /* === Memory management === */
* Return a DCtx to clean state.
* Session and parameters can be reset jointly or separately. /*! ZSTD_sizeof_*() :
* Parameters can only be reset when no active frame is being decompressed. * These functions give the _current_ memory usage of selected object.
* @return : 0, or an error code, which can be tested with ZSTD_isError() * 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_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset); 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) * experimental API (static linking only)
**************************************************************************************** ****************************************************************************************
* The following symbols and constants * The following symbols and constants
* are not planned to join "stable API" status in the near future. * are not planned to join "stable API" status in the near future.
* They can still change in future versions. * They can still change in future versions.
* Some of them are planned to remain in the static_only section indefinitely. * Some of them are planned to remain in the static_only section indefinitely.
#define ZSTD_OVERLAPLOG_MIN 0 #define ZSTD_OVERLAPLOG_MIN 0
#define ZSTD_OVERLAPLOG_MAX 9 #define ZSTD_OVERLAPLOG_MAX 9
#define ZSTD_WINDOWLOG_LIMIT_DEFAULT 27 /* by default, the streaming decoder will refuse any frame #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, * requiring larger than (1<<ZSTD_WINDOWLOG_LIMIT_DEFAULT) window size,
* to preserve host's memory from unreasonable requirements. * to preserve host's memory from unreasonable requirements.
* This limit can be overriden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,). * 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 */ * The limit does not apply for one-pass decoders (such as ZSTD_decompress()), since no additional memory is allocated */
/* LDM parameter bounds */ /* LDM parameter bounds */
#define ZSTD_LDM_HASHLOG_MIN ZSTD_HASHLOG_MIN #define ZSTD_LDM_HASHLOG_MIN ZSTD_HASHLOG_MIN
#define ZSTD_LDM_HASHLOG_MAX ZSTD_HASHLOG_MAX #define ZSTD_LDM_HASHLOG_MAX ZSTD_HASHLOG_MAX
#define ZSTD_LDM_MINMATCH_MIN 4 #define ZSTD_LDM_MINMATCH_MIN 4
#define ZSTD_LDM_MINMATCH_MAX 4096 #define ZSTD_LDM_MINMATCH_MAX 4096
#define ZSTD_LDM_BUCKETSIZELOG_MIN 1 #define ZSTD_LDM_BUCKETSIZELOG_MIN 1
#define ZSTD_LDM_BUCKETSIZELOG_MAX 8 #define ZSTD_LDM_BUCKETSIZELOG_MAX 8
#define ZSTD_LDM_HASHRATELOG_MIN 0 #define ZSTD_LDM_HASHRATELOG_MIN 0
#define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN) #define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)
/* Advanced parameter bounds */
#define ZSTD_TARGETCBLOCKSIZE_MIN 64
#define ZSTD_TARGETCBLOCKSIZE_MAX ZSTD_BLOCKSIZE_MAX
/* internal */ /* internal */
#define ZSTD_HASHLOG3_MAX 17 #define ZSTD_HASHLOG3_MAX 17
/* --- Advanced types --- */ /* --- Advanced types --- */
typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params; typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params;
* Zstd is making poor choices, it is possible to override that choice with * Zstd is making poor choices, it is possible to override that choice with
* this enum. * this enum.
*/ */
ZSTD_dictDefaultAttach = 0, /* Use the default heuristic. */ ZSTD_dictDefaultAttach = 0, /* Use the default heuristic. */
ZSTD_dictForceAttach = 1, /* Never copy the dictionary. */ ZSTD_dictForceAttach = 1, /* Never copy the dictionary. */
ZSTD_dictForceCopy = 2, /* Always copy the dictionary. */ ZSTD_dictForceCopy = 2, /* Always copy the dictionary. */
} ZSTD_dictAttachPref_e; } 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 * Frame size functions
***************************************/ ***************************************/
/*! ZSTD_findDecompressedSize() : /*! ZSTD_findDecompressedSize() :
* `src` should point the start of a series of ZSTD encoded and/or skippable frames * `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 * `srcSize` must be the _exact_ size of this series
* (i.e. there should be a frame boundary exactly at `srcSize` bytes after `src`) * (i.e. there should be a frame boundary at `src + srcSize`)
* @return : - decompressed size of all data in all successive frames * @return : - decompressed size of all data in all successive frames
* - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN * - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN
* - if an error occurred: ZSTD_CONTENTSIZE_ERROR * - if an error occurred: ZSTD_CONTENTSIZE_ERROR
* *
* note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode. * 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. * 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. * 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 2 : decompressed size is always present when compression is done with ZSTD_compress()
* note 3 : decompressed size can be very large (64-bits value), * note 3 : decompressed size can be very large (64-bits value),
* potentially larger than what local system can handle as a single memory segment. * 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. * 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. * note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified.
* Always ensure result fits within application's authorized limits. * Always ensure result fits within application's authorized limits.
* Each application can set its own limits. * Each application can set its own limits.
* note 5 : ZSTD_findDecompressedSize handles multiple frames, and so it must traverse the input to * 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, * 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. */ * 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); 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() : /*! ZSTD_frameHeaderSize() :
* srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX. * srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.
* @return : size of the Frame Header, * @return : size of the Frame Header,
* or an error code (if srcSize is too small) */ * or an error code (if srcSize is too small) */
ZSTDLIB_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize); ZSTDLIB_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
/*************************************** /***************************************
* Memory management * Memory management
***************************************/ ***************************************/
/*! ZSTD_estimate*() : /*! ZSTD_estimate*() :
* These functions make it possible to estimate memory usage * These functions make it possible to estimate memory usage
* of a future {D,C}Ctx, before its creation. * 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. * 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. * 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. * 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_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
* ZSTD_estimateCCtxSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1. * 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. */ * 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(int compressionLevel);
ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams); 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_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void); ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);
/*! ZSTD_estimateCStreamSize() : /*! ZSTD_estimateCStreamSize() :
* ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one. * 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. * 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. * 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_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
* ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1. * 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. * Note : CStream size estimation is only correct for single-threaded compression.
* ZSTD_DStream memory budget depends on window Size. * ZSTD_DStream memory budget depends on window Size.
* This information can be passed manually, using ZSTD_estimateDStreamSize, * This information can be passed manually, using ZSTD_estimateDStreamSize,
* or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame(); * or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame();
* Note : if streaming is init with function ZSTD_init?Stream_usingDict(), * Note : if streaming is init with function ZSTD_init?Stream_usingDict(),
* an internal ?Dict will be created, which additional size is not estimated here. * an internal ?Dict will be created, which additional size is not estimated here.
* In this case, get total size by adding ZSTD_estimate?DictSize */ * 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(int compressionLevel);
/*! ZSTD_createCDict_byReference() : /*! ZSTD_createCDict_byReference() :
* Create a digested dictionary for compression * Create a digested dictionary for compression
* Dictionary content is just referenced, not duplicated. * Dictionary content is just referenced, not duplicated.
* As a consequence, `dictBuffer` **must** outlive CDict, * As a consequence, `dictBuffer` **must** outlive CDict,
* and its content must remain unmodified throughout the lifetime of 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); ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
/*! ZSTD_getCParams() : /*! ZSTD_getCParams() :
* @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize. * @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
* `estimatedSrcSize` value is optional, select 0 if not known */ * `estimatedSrcSize` value is optional, select 0 if not known */
ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize); ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
/*! ZSTD_getParams() : /*! ZSTD_getParams() :
* same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`. * 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 */ * 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); ZSTDLIB_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
/*! ZSTD_checkCParams() : /*! ZSTD_checkCParams() :
* Ensure param values remain within authorized range */ * 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); ZSTDLIB_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
/*! ZSTD_adjustCParams() : /*! ZSTD_adjustCParams() :
* optimize params for a given `srcSize` and `dictSize`. * optimize params for a given `srcSize` and `dictSize`.
* both values are optional, select `0` if unknown. */ * `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); ZSTDLIB_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
/*! ZSTD_compress_advanced() : /*! ZSTD_compress_advanced() :
* Same as ZSTD_compress_usingDict(), with fine-tune control over compression parameters (by structure) */ * Same as ZSTD_compress_usingDict(), with fine-tune control over compression parameters (by structure) */
ZSTDLIB_API size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx, ZSTDLIB_API size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const void* src, size_t srcSize,
const void* dict,size_t dictSize, const void* dict,size_t dictSize,
#define ZSTD_c_forceMaxWindow ZSTD_c_experimentalParam3 #define ZSTD_c_forceMaxWindow ZSTD_c_experimentalParam3
/* Controls whether the contents of a CDict /* Controls whether the contents of a CDict
* are used in place, or copied into the working context. * are used in place, or copied into the working context.
* Accepts values from the ZSTD_dictAttachPref_e enum. * Accepts values from the ZSTD_dictAttachPref_e enum.
* See the comments on that enum for an explanation of the feature. */ * See the comments on that enum for an explanation of the feature. */
#define ZSTD_c_forceAttachDict ZSTD_c_experimentalParam4 #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
/* Tries to fit compressed block size to be around targetCBlockSize.
* No target when targetCBlockSize == 0.
* There is no guarantee on compressed block size (default:0) */
#define ZSTD_c_targetCBlockSize ZSTD_c_experimentalParam6
/*! ZSTD_CCtx_getParameter() : /*! ZSTD_CCtx_getParameter() :
* Get the requested compression parameter value, selected by enum ZSTD_cParameter, * Get the requested compression parameter value, selected by enum ZSTD_cParameter,
* and store it into int* value. * and store it into int* value.
* @return : 0, or an error code (which can be tested with ZSTD_isError()). * @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); ZSTDLIB_API size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value);
/*! ZSTD_CCtx_params : /*! ZSTD_CCtx_params :
* Quick howto : * Quick howto :
* - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure * - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
* - ZSTD_CCtxParam_setParameter() : Push parameters one by one into * - ZSTD_CCtxParams_setParameter() : Push parameters one by one into
* an existing ZSTD_CCtx_params structure. * an existing ZSTD_CCtx_params structure.
* This is similar to * This is similar to
* ZSTD_CCtx_setParameter(). * ZSTD_CCtx_setParameter().
* - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to * - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to
* an existing CCtx. * an existing CCtx.
* These parameters will be applied to * These parameters will be applied to
* all subsequent frames. * all subsequent frames.
* - ZSTD_compressStream2() : Do compression using the CCtx. * - ZSTD_compressStream2() : Do compression using the CCtx.
* - ZSTD_freeCCtxParams() : Free the memory. * - ZSTD_freeCCtxParams() : Free the memory.
* *
* This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams() * This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()
ZSTDLIB_API size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel); ZSTDLIB_API size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel);
/*! ZSTD_CCtxParams_init_advanced() : /*! ZSTD_CCtxParams_init_advanced() :
* Initializes the compression and frame parameters of cctxParams according to * Initializes the compression and frame parameters of cctxParams according to
* params. All other parameters are reset to their default values. * 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); ZSTDLIB_API size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
/*! ZSTD_CCtxParam_setParameter() : /*! ZSTD_CCtxParams_setParameter() :
* Similar to ZSTD_CCtx_setParameter. * Similar to ZSTD_CCtx_setParameter.
* Set one compression parameter, selected by enum ZSTD_cParameter. * Set one compression parameter, selected by enum ZSTD_cParameter.
* Parameters must be applied to a ZSTD_CCtx using ZSTD_CCtx_setParametersUsingCCtxParams(). * 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()). * @result : 0, or an error code (which can be tested with ZSTD_isError()).
*/ */
ZSTDLIB_API size_t ZSTD_CCtxParam_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value); ZSTDLIB_API size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);
/*! ZSTD_CCtxParam_getParameter() : /*! ZSTD_CCtxParams_getParameter() :
* Similar to ZSTD_CCtx_getParameter. * Similar to ZSTD_CCtx_getParameter.
* Get the requested value of one compression parameter, selected by enum ZSTD_cParameter. * 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()). * @result : 0, or an error code (which can be tested with ZSTD_isError()).
*/ */
ZSTDLIB_API size_t ZSTD_CCtxParam_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value); ZSTDLIB_API size_t ZSTD_CCtxParams_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);
/*! ZSTD_CCtx_setParametersUsingCCtxParams() : /*! ZSTD_CCtx_setParametersUsingCCtxParams() :
* Apply a set of ZSTD_CCtx_params to the compression context. * Apply a set of ZSTD_CCtx_params to the compression context.
* This can be done even after compression is started, * 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==0, this will have no impact until a new compression is started.
* if nbWorkers>=1, new parameters will be picked up at next job, * 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). * with a few restrictions (windowLog, pledgedSrcSize, nbWorkers, jobSize, and overlapLog are not updated).
*/ */
/*! ZSTD_createDDict_byReference() : /*! ZSTD_createDDict_byReference() :
* Create a digested dictionary, ready to start decompression operation without startup delay. * Create a digested dictionary, ready to start decompression operation without startup delay.
* Dictionary content is referenced, and therefore stays in dictBuffer. * Dictionary content is referenced, and therefore stays in dictBuffer.
* It is important that dictBuffer outlives DDict, * It is important that dictBuffer outlives DDict,
* it must remain read accessible throughout the lifetime of 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_byReference(const void* dictBuffer, size_t dictSize);
/*! 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);
/*! ZSTD_DCtx_loadDictionary_byReference() : /*! ZSTD_DCtx_loadDictionary_byReference() :
* Same as ZSTD_DCtx_loadDictionary(), * Same as ZSTD_DCtx_loadDictionary(),
* but references `dict` content instead of copying it into `dctx`. * but references `dict` content instead of copying it into `dctx`.
* This saves memory if `dict` remains around., * 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. */ * 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); ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
/*! ZSTD_DCtx_loadDictionary_advanced() : /*! ZSTD_DCtx_loadDictionary_advanced() :
/******************************************************************** /********************************************************************
* Advanced streaming functions * Advanced streaming functions
* Warning : most of these functions are now redundant with the Advanced API. * Warning : most of these functions are now redundant with the Advanced API.
* Once Advanced API reaches "stable" status, * Once Advanced API reaches "stable" status,
* redundant functions will be deprecated, and then at some point removed. * redundant functions will be deprecated, and then at some point removed.
********************************************************************/ ********************************************************************/
/*===== Advanced Streaming compression functions =====*/ /*===== Advanced Streaming compression functions =====*/
ZSTDLIB_API size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long 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. */ /**! ZSTD_initCStream_srcSize() :
ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel); /**< 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.*/ * 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, ZSTDLIB_API size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize,
ZSTD_parameters params, unsigned long long pledgedSrcSize); /**< 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. */ ZSTD_parameters params, unsigned long long pledgedSrcSize);
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 */ /**! ZSTD_initCStream_usingCDict() :
ZSTDLIB_API size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize); /**< 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. */ * 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() : /*! 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. * start a new frame, using same parameters from previous frame.
* This is typically useful to skip dictionary loading stage, since it will re-use it in-place. * This is typically useful to skip dictionary loading stage, since it will re-use it in-place.
* Note that zcs must be init at least once before using ZSTD_resetCStream(). * 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 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. * 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, * 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. * 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()) * @return : 0, or an error code (which can be tested using ZSTD_isError())
* but everything it has produced has also been flushed so far, * but everything it has produced has also been flushed so far,
* therefore flush speed is limited by production speed of oldest job * therefore flush speed is limited by production speed of oldest job
* irrespective of the speed of concurrent (and newer) jobs. * irrespective of the speed of concurrent (and newer) jobs.
*/ */
ZSTDLIB_API size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx); ZSTDLIB_API size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx);
/*===== Advanced Streaming decompression functions =====*/ /*===== Advanced Streaming decompression functions =====*/
ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize); /**< note: no dictionary will be used if dict == NULL or dictSize < 8 */ /**
ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict); /**< note : ddict is referenced, it must outlive decompression session */ * This function is deprecated, and is equivalent to:
ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds); /**< re-use decompression parameters from previous init; saves dictionary loading */ *
* 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 * 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. * 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. * But it's also a complex one, with several restrictions, documented below.
* Prefer normal streaming API for an easier experience. * Prefer normal streaming API for an easier experience.
unsigned long long windowSize; /* can be very large, up to <= frameContentSize */ unsigned long long windowSize; /* can be very large, up to <= frameContentSize */
unsigned blockSizeMax; unsigned blockSizeMax;
ZSTD_frameType_e frameType; /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */ ZSTD_frameType_e frameType; /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */
unsigned headerSize; unsigned headerSize;
unsigned dictID; unsigned dictID;
unsigned checksumFlag; unsigned checksumFlag;
} ZSTD_frameHeader; } ZSTD_frameHeader;
/** ZSTD_getFrameHeader() : /*! ZSTD_getFrameHeader() :
* decode Frame Header, or requires larger `srcSize`. * decode Frame Header, or requires larger `srcSize`.
* @return : 0, `zfhPtr` is correctly filled, * @return : 0, `zfhPtr` is correctly filled,
* >0, `srcSize` is too small, value is wanted `srcSize` amount, * >0, `srcSize` is too small, value is wanted `srcSize` amount,
* or an error code, which can be tested using ZSTD_isError() */ * 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 */ ZSTDLIB_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize); /**< doesn't consume input */
/*! ZSTD_getFrameHeader_advanced() : /*! ZSTD_getFrameHeader_advanced() :
* same as ZSTD_getFrameHeader(), * same as ZSTD_getFrameHeader(),
* with added capability to select a format (like ZSTD_f_zstd1_magicless) */ * with added capability to select a format (like ZSTD_f_zstd1_magicless) */
/* ============================ */ /* ============================ */
/** Block level API */ /** Block level API */
/* ============================ */ /* ============================ */
/*! /*!
Block functions produce and decode raw zstd blocks, without frame metadata. 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). 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. But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes.
A few rules to respect : A few rules to respect :
- Compressing and decompressing require a context structure - Compressing and decompressing require a context structure
+ Use ZSTD_createCCtx() and ZSTD_createDCtx() + Use ZSTD_createCCtx() and ZSTD_createDCtx()
- It is necessary to init context before starting - It is necessary to init context before starting
+ compression : any ZSTD_compressBegin*() variant, including with dictionary + compression : any ZSTD_compressBegin*() variant, including with dictionary
+ decompression : any ZSTD_decompressBegin*() variant, including with dictionary + decompression : any ZSTD_decompressBegin*() variant, including with dictionary
+ copyCCtx() and copyDCtx() can be used too + copyCCtx() and copyDCtx() can be used too
- Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB - 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 + 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. + For inputs larger than a single block, consider using regular ZSTD_compress() instead.
Frame metadata is not that costly, and quickly becomes negligible as source size grows larger. Frame metadata is not that costly, and quickly becomes negligible as source size grows larger than a block.
- When a block is considered not compressible enough, ZSTD_compressBlock() result will be zero. - When a block is considered not compressible enough, ZSTD_compressBlock() result will be 0 (zero) !
In which case, nothing is produced into `dst` ! ===> In which case, nothing is produced into `dst` !
+ User must test for such outcome and deal directly with uncompressed data + User __must__ test for such outcome and deal directly with uncompressed data
+ ZSTD_decompressBlock() doesn't accept uncompressed data as input !!! + A block cannot be declared incompressible if ZSTD_compressBlock() return value was != 0.
Doing so would mess up with statistics history, leading to potential data corruption.
+ ZSTD_decompressBlock() _doesn't accept uncompressed data as input_ !!
+ In case of multiple successive blocks, should some of them be uncompressed, + 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. decoder must be informed of their existence in order to follow proper history.
Use ZSTD_insertBlock() for such a case. Use ZSTD_insertBlock() for such a case.
*/ */
/*===== Raw zstd block functions =====*/ /*===== Raw zstd block functions =====*/
ZSTDLIB_API size_t ZSTD_getBlockSize (const ZSTD_CCtx* cctx); 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_compressBlock (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);

tests/test-repo-compengines.t

> hg -R $repo commit -m some-commit > hg -R $repo commit -m some-commit
> } > }
$ for repo in zstd-level-default zstd-level-1 zstd-level-22; do $ for repo in zstd-level-default zstd-level-1 zstd-level-22; do
> commitone $repo > commitone $repo
> done > done
$ $RUNTESTDIR/f -s zstd-*/.hg/store/data/* $ $RUNTESTDIR/f -s zstd-*/.hg/store/data/*
zstd-level-1/.hg/store/data/a.i: size=4097 zstd-level-1/.hg/store/data/a.i: size=4114
zstd-level-22/.hg/store/data/a.i: size=4091 zstd-level-22/.hg/store/data/a.i: size=4091
zstd-level-default/\.hg/store/data/a\.i: size=(4094|4102) (re) zstd-level-default/\.hg/store/data/a\.i: size=(4094|4102) (re)
Test error cases Test error cases
$ hg init zstd-level-invalid --config format.revlog-compression=zstd $ hg init zstd-level-invalid --config format.revlog-compression=zstd
$ cat << EOF >> zstd-level-invalid/.hg/hgrc $ cat << EOF >> zstd-level-invalid/.hg/hgrc
> [storage] > [storage]