diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2021-11-13 15:32:30 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2021-11-13 15:32:30 -0800 |
commit | c8c109546a19613d323a319d0c921cb1f317e629 (patch) | |
tree | 5cf721ff53b9caba1162065cd16071e5226d40db /lib/zstd/compress/zstd_compress_sequences.c | |
parent | ccfff0a2bd2a30de130b5623d242ddecd0272bc2 (diff) | |
parent | 0a8ea235837cc39f27c45689930aa97ae91d5953 (diff) |
Merge tag 'zstd-for-linus-v5.16' of git://github.com/terrelln/linux
Pull zstd update from Nick Terrell:
"Update to zstd-1.4.10.
Add myself as the maintainer of zstd and update the zstd version in
the kernel, which is now 4 years out of date, to a much more recent
zstd release. This includes bug fixes, much more extensive fuzzing,
and performance improvements. And generates the kernel zstd
automatically from upstream zstd, so it is easier to keep the zstd
verison up to date, and we don't fall so far out of date again.
This includes 5 commits that update the zstd library version:
- Adds a new kernel-style wrapper around zstd.
This wrapper API is functionally equivalent to the subset of the
current zstd API that is currently used. The wrapper API changes to
be kernel style so that the symbols don't collide with zstd's
symbols. The update to zstd-1.4.10 maintains the same API and
preserves the semantics, so that none of the callers need to be
updated. All callers are updated in the commit, because there are
zero functional changes.
- Adds an indirection for `lib/decompress_unzstd.c` so it doesn't
depend on the layout of `lib/zstd/` to include every source file.
This allows the next patch to be automatically generated.
- Imports the zstd-1.4.10 source code. This commit is automatically
generated from upstream zstd (https://github.com/facebook/zstd).
- Adds me (terrelln@fb.com) as the maintainer of `lib/zstd`.
- Fixes a newly added build warning for clang.
The discussion around this patchset has been pretty long, so I've
included a FAQ-style summary of the history of the patchset, and why
we are taking this approach.
Why do we need to update?
-------------------------
The zstd version in the kernel is based off of zstd-1.3.1, which is
was released August 20, 2017. Since then zstd has seen many bug fixes
and performance improvements. And, importantly, upstream zstd is
continuously fuzzed by OSS-Fuzz, and bug fixes aren't backported to
older versions. So the only way to sanely get these fixes is to keep
up to date with upstream zstd.
There are no known security issues that affect the kernel, but we need
to be able to update in case there are. And while there are no known
security issues, there are relevant bug fixes. For example the problem
with large kernel decompression has been fixed upstream for over 2
years [1]
Additionally the performance improvements for kernel use cases are
significant. Measured for x86_64 on my Intel i9-9900k @ 3.6 GHz:
- BtrFS zstd compression at levels 1 and 3 is 5% faster
- BtrFS zstd decompression+read is 15% faster
- SquashFS zstd decompression+read is 15% faster
- F2FS zstd compression+write at level 3 is 8% faster
- F2FS zstd decompression+read is 20% faster
- ZRAM decompression+read is 30% faster
- Kernel zstd decompression is 35% faster
- Initramfs zstd decompression+build is 5% faster
On top of this, there are significant performance improvements coming
down the line in the next zstd release, and the new automated update
patch generation will allow us to pull them easily.
How is the update patch generated?
----------------------------------
The first two patches are preparation for updating the zstd version.
Then the 3rd patch in the series imports upstream zstd into the
kernel. This patch is automatically generated from upstream. A script
makes the necessary changes and imports it into the kernel. The
changes are:
- Replace all libc dependencies with kernel replacements and rewrite
includes.
- Remove unncessary portability macros like: #if defined(_MSC_VER).
- Use the kernel xxhash instead of bundling it.
This automation gets tested every commit by upstream's continuous
integration. When we cut a new zstd release, we will submit a patch to
the kernel to update the zstd version in the kernel.
The automated process makes it easy to keep the kernel version of zstd
up to date. The current zstd in the kernel shares the guts of the
code, but has a lot of API and minor changes to work in the kernel.
This is because at the time upstream zstd was not ready to be used in
the kernel envrionment as-is. But, since then upstream zstd has
evolved to support being used in the kernel as-is.
Why are we updating in one big patch?
-------------------------------------
The 3rd patch in the series is very large. This is because it is
restructuring the code, so it both deletes the existing zstd, and
re-adds the new structure. Future updates will be directly
proportional to the changes in upstream zstd since the last import.
They will admittidly be large, as zstd is an actively developed
project, and has hundreds of commits between every release. However,
there is no other great alternative.
One option ruled out is to replay every upstream zstd commit. This is
not feasible for several reasons:
- There are over 3500 upstream commits since the zstd version in the
kernel.
- The automation to automatically generate the kernel update was only
added recently, so older commits cannot easily be imported.
- Not every upstream zstd commit builds.
- Only zstd releases are "supported", and individual commits may have
bugs that were fixed before a release.
Another option to reduce the patch size would be to first reorganize
to the new file structure, and then apply the patch. However, the
current kernel zstd is formatted with clang-format to be more
"kernel-like". But, the new method imports zstd as-is, without
additional formatting, to allow for closer correlation with upstream,
and easier debugging. So the patch wouldn't be any smaller.
It also doesn't make sense to import upstream zstd commit by commit
going forward. Upstream zstd doesn't support production use cases
running of the development branch. We have a lot of post-commit
fuzzing that catches many bugs, so indiviudal commits may be buggy,
but fixed before a release. So going forward, I intend to import every
(important) zstd release into the Kernel.
So, while it isn't ideal, updating in one big patch is the only patch
I see forward.
Who is responsible for this code?
---------------------------------
I am. This patchset adds me as the maintainer for zstd. Previously,
there was no tree for zstd patches. Because of that, there were
several patches that either got ignored, or took a long time to merge,
since it wasn't clear which tree should pick them up. I'm officially
stepping up as maintainer, and setting up my tree as the path through
which zstd patches get merged. I'll make sure that patches to the
kernel zstd get ported upstream, so they aren't erased when the next
version update happens.
How is this code tested?
------------------------
I tested every caller of zstd on x86_64 (BtrFS, ZRAM, SquashFS, F2FS,
Kernel, InitRAMFS). I also tested Kernel & InitRAMFS on i386 and
aarch64. I checked both performance and correctness.
Also, thanks to many people in the community who have tested these
patches locally.
Lastly, this code will bake in linux-next before being merged into
v5.16.
Why update to zstd-1.4.10 when zstd-1.5.0 has been released?
------------------------------------------------------------
This patchset has been outstanding since 2020, and zstd-1.4.10 was the
latest release when it was created. Since the update patch is
automatically generated from upstream, I could generate it from
zstd-1.5.0.
However, there were some large stack usage regressions in zstd-1.5.0,
and are only fixed in the latest development branch. And the latest
development branch contains some new code that needs to bake in the
fuzzer before I would feel comfortable releasing to the kernel.
Once this patchset has been merged, and we've released zstd-1.5.1, we
can update the kernel to zstd-1.5.1, and exercise the update process.
You may notice that zstd-1.4.10 doesn't exist upstream. This release
is an artifical release based off of zstd-1.4.9, with some fixes for
the kernel backported from the development branch. I will tag the
zstd-1.4.10 release after this patchset is merged, so the Linux Kernel
is running a known version of zstd that can be debugged upstream.
Why was a wrapper API added?
----------------------------
The first versions of this patchset migrated the kernel to the
upstream zstd API. It first added a shim API that supported the new
upstream API with the old code, then updated callers to use the new
shim API, then transitioned to the new code and deleted the shim API.
However, Cristoph Hellwig suggested that we transition to a kernel
style API, and hide zstd's upstream API behind that. This is because
zstd's upstream API is supports many other use cases, and does not
follow the kernel style guide, while the kernel API is focused on the
kernel's use cases, and follows the kernel style guide.
Where is the previous discussion?
---------------------------------
Links for the discussions of the previous versions of the patch set
below. The largest changes in the design of the patchset are driven by
the discussions in v11, v5, and v1. Sorry for the mix of links, I
couldn't find most of the the threads on lkml.org"
Link: https://lkml.org/lkml/2020/9/29/27 [1]
Link: https://www.spinics.net/lists/linux-crypto/msg58189.html [v12]
Link: https://lore.kernel.org/linux-btrfs/20210430013157.747152-1-nickrterrell@gmail.com/ [v11]
Link: https://lore.kernel.org/lkml/20210426234621.870684-2-nickrterrell@gmail.com/ [v10]
Link: https://lore.kernel.org/linux-btrfs/20210330225112.496213-1-nickrterrell@gmail.com/ [v9]
Link: https://lore.kernel.org/linux-f2fs-devel/20210326191859.1542272-1-nickrterrell@gmail.com/ [v8]
Link: https://lkml.org/lkml/2020/12/3/1195 [v7]
Link: https://lkml.org/lkml/2020/12/2/1245 [v6]
Link: https://lore.kernel.org/linux-btrfs/20200916034307.2092020-1-nickrterrell@gmail.com/ [v5]
Link: https://www.spinics.net/lists/linux-btrfs/msg105783.html [v4]
Link: https://lkml.org/lkml/2020/9/23/1074 [v3]
Link: https://www.spinics.net/lists/linux-btrfs/msg105505.html [v2]
Link: https://lore.kernel.org/linux-btrfs/20200916034307.2092020-1-nickrterrell@gmail.com/ [v1]
Signed-off-by: Nick Terrell <terrelln@fb.com>
Tested By: Paul Jones <paul@pauljones.id.au>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com> # LLVM/Clang v13.0.0 on x86-64
Tested-by: Jean-Denis Girard <jd.girard@sysnux.pf>
* tag 'zstd-for-linus-v5.16' of git://github.com/terrelln/linux:
lib: zstd: Add cast to silence clang's -Wbitwise-instead-of-logical
MAINTAINERS: Add maintainer entry for zstd
lib: zstd: Upgrade to latest upstream zstd version 1.4.10
lib: zstd: Add decompress_sources.h for decompress_unzstd
lib: zstd: Add kernel-specific API
Diffstat (limited to 'lib/zstd/compress/zstd_compress_sequences.c')
-rw-r--r-- | lib/zstd/compress/zstd_compress_sequences.c | 439 |
1 files changed, 439 insertions, 0 deletions
diff --git a/lib/zstd/compress/zstd_compress_sequences.c b/lib/zstd/compress/zstd_compress_sequences.c new file mode 100644 index 000000000000..dcfcdc9cc5e8 --- /dev/null +++ b/lib/zstd/compress/zstd_compress_sequences.c @@ -0,0 +1,439 @@ +/* + * Copyright (c) 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 true if we should use ncount=-1 else we should + * use ncount=1 for low probability symbols instead. + */ +static unsigned ZSTD_useLowProbCount(size_t const nbSeq) +{ + /* Heuristic: This should cover most blocks <= 16K and + * start to fade out after 16K to about 32K depending on + * comprssibility. + */ + return nbSeq >= 2048; +} + +/* + * 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, ZSTD_useLowProbCount(nbSeq)), ""); + 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. + */ +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 += (size_t)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. + */ +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) ? (unsigned)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; +} + +typedef struct { + S16 norm[MaxSeq + 1]; + U32 wksp[FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(MaxSeq, MaxFSELog)]; +} ZSTD_BuildCTableWksp; + +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* entropyWorkspace, size_t entropyWorkspaceSize) +{ + 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, "not enough space"); + *op = codeTable[0]; + return 1; + case set_repeat: + ZSTD_memcpy(nextCTable, prevCTable, prevCTableSize); + return 0; + case set_basic: + FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, entropyWorkspace, entropyWorkspaceSize), ""); /* note : could be pre-calculated */ + return 0; + case set_compressed: { + ZSTD_BuildCTableWksp* wksp = (ZSTD_BuildCTableWksp*)entropyWorkspace; + 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); + assert(entropyWorkspaceSize >= sizeof(ZSTD_BuildCTableWksp)); + (void)entropyWorkspaceSize; + FORWARD_IF_ERROR(FSE_normalizeCount(wksp->norm, tableLog, count, nbSeq_1, max, ZSTD_useLowProbCount(nbSeq_1)), ""); + { size_t const NCountSize = FSE_writeNCount(op, oend - op, wksp->norm, max, tableLog); /* overflow protected */ + FORWARD_IF_ERROR(NCountSize, "FSE_writeNCount failed"); + FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, wksp->norm, max, tableLog, wksp->wksp, sizeof(wksp->wksp)), ""); + return NCountSize; + } + } + default: assert(0); RETURN_ERROR(GENERIC, "impossible to reach"); + } +} + +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]; + unsigned 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) { + unsigned 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); +} |