detected by @terrelln,
these issue could be triggered in specific scenarios
namely decompression of certain invalid magic-less frames,
or requested properties from certain invalid skippable frames.
* add check for valid dest buffer and fuzz on random dest ptr when malloc 0
* add uptrval to linux-kernel
* remove bin files
* get rid of uptrval
* restrict max pointer value check to platforms where sizeof(size_t) == sizeof(void*)
decompression features automatic support of sparse files,
aka a form of "compression" where entire blocks consists only of zeroes.
This only works for some compatible file systems (like ext4),
others simply ignore it (like afs).
Triggering this feature relies of `fseek()`.
But `fseek()` is not compatible with non-seekable devices, such as pipes.
Therefore it's disabled for pipes.
However, there are other objects which are not compatible with `fseek()`, such as block devices.
Changed the logic, so that `fseek()` (and therefore sparse write) is only automatically enabled on regular files.
Note that this automatic behavior can always be overridden by explicit commands `--sparse` and `--no-sparse`.
fix#3583
This does the following:
1. Compress test data into multiple frames
2. Perform a series of small decompressions and seeks forward, checking
that compressed data wasn't reread unnecessarily.
3. Perform some seeks forward and backward to ensure correctness.
When decompressing a seekable file, if seeking forward within
a frame (by issuing multiple ZSTD_seekable_decompress calls
with a small gap between them), the frame will be unnecessarily
reread from the beginning. This patch makes it continue using
the current frame data and simply skip over the unneeded bytes.
Looking at the __builtin_expect in ZSTD_decodeSequence:
{ size_t offset;
#if defined(__clang__)
if (LIKELY(ofBits > 1)) {
#else
if (ofBits > 1) {
#endif
ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
From profile-annotated assembly, the probability of ofBits > 1 is about 75%
(101k counts out of 135k counts). This is much smaller than the recommended
likelihood to use __builtin_expect which is 99%. As a result, clang moved the
else block further away which hurts cache locality. Removing this
__built_expect along with two others in ZSTD_decodeSequence gave better
performance when PGO is enabled. I suggest to remove these branch hints and
rely on PGO which leverages runtime profiles from actual workload to calculate
branch probability instead.
Inlining `BIT_reloadDStream` provided >3% decompression speed improvement for
clang PGO-optimized zstd binary, measured using the Silesia corpus with
compression level 1. The win comes from improved register allocation which leads
to fewer spills and reloads. Take a look at this comparison of
profile-annotated hot assembly before and after this change:
https://www.diffchecker.com/UjDGIyLz/. The diff is a bit messy, but notice three
fewer moves after inlining.
In general LLVM's register allocator works better when it can see more code. For
example, when the register allocator sees a call instruction, it partitions the
registers into caller registers and callee registers, and it is not free to do
whatever it wants with all the registers for the current function. Inlining the
callee lets the register allocation access all registers and use them more
flexsibly.
Rather than remove the flag entirely, as proposed in #3499, this commit uses
the newest C++ standard the compiler supports. This retains the selection of
using only standardized features (excluding GNU extensions) and keeps the
recency requirements of the codebase explicit.
Tested with various versions of `g++` and `clang++`.
This fixes compilation with clang-cl on Windows. There
is a bug in cmake so that check_linker_flag() doesn't give
the correct result when using link.exe/lld-link.exe.
Details in CMake's gitlab: https://gitlab.kitware.com/cmake/cmake/-/issues/22023Fixes#3522
Every 256 bytes the lazy match finders process without finding a match,
they will increase their step size by 1. So for bytes [0, 256) they search
every position, for bytes [256, 512) they search every other position,
and so on. However, they currently still insert every position into
their hash tables. This is different from fast & dfast, which only
insert the positions they search.
This PR changes that, so now after we've searched 2KB without finding
any matches, at which point we'll only be searching one in 9 positions,
we'll stop inserting every position, and only insert the positions we
search. The exact cutoff of 2KB isn't terribly important, I've just
selected a cutoff that is reasonably large, to minimize the impact on
"normal" data.
This PR only adds skipping to greedy, lazy, and lazy2, but does not
touch btlazy2.
| Dataset | Level | Compiler | CSize ∆ | Speed ∆ |
|---------|-------|--------------|---------|---------|
| Random | 5 | clang-14.0.6 | 0.0% | +704% |
| Random | 5 | gcc-12.2.0 | 0.0% | +670% |
| Random | 7 | clang-14.0.6 | 0.0% | +679% |
| Random | 7 | gcc-12.2.0 | 0.0% | +657% |
| Random | 12 | clang-14.0.6 | 0.0% | +1355% |
| Random | 12 | gcc-12.2.0 | 0.0% | +1331% |
| Silesia | 5 | clang-14.0.6 | +0.002% | +0.35% |
| Silesia | 5 | gcc-12.2.0 | +0.002% | +2.45% |
| Silesia | 7 | clang-14.0.6 | +0.001% | -1.40% |
| Silesia | 7 | gcc-12.2.0 | +0.007% | +0.13% |
| Silesia | 12 | clang-14.0.6 | +0.011% | +22.70% |
| Silesia | 12 | gcc-12.2.0 | +0.011% | -6.68% |
| Enwik8 | 5 | clang-14.0.6 | 0.0% | -1.02% |
| Enwik8 | 5 | gcc-12.2.0 | 0.0% | +0.34% |
| Enwik8 | 7 | clang-14.0.6 | 0.0% | -1.22% |
| Enwik8 | 7 | gcc-12.2.0 | 0.0% | -0.72% |
| Enwik8 | 12 | clang-14.0.6 | 0.0% | +26.19% |
| Enwik8 | 12 | gcc-12.2.0 | 0.0% | -5.70% |
The speed difference for clang at level 12 is real, but is probably
caused by some sort of alignment or codegen issues. clang is
significantly slower than gcc before this PR, but gets up to parity with
it.
I also measured the ratio difference for the HC match finder, and it
looks basically the same as the row-based match finder. The speedup on
random data looks similar. And performance is about neutral, without the
big difference at level 12 for either clang or gcc.
In Python 3.x, a single element of a bytes array is returned as
an integer number. Thus, NEWLINE is an int variable, and attempting
to add it to the line array will fail with a type mismatch error
that may be demonstrated as follows:
[roam@straylight ~]$ python3 -c 'b"hello" + b"\n"[0]'
Traceback (most recent call last):
File "<string>", line 1, in <module>
TypeError: can't concat int to bytes
[roam@straylight ~]$
* Mark all bufferless and block level functions as deprecated
* Update documentation to suggest not using these functions
* Add `_deprecated()` wrappers for functions that we use internally and
call those instead
* patch-from speed optimization: only load portion of dictionary into normal matchfinders
* test regression for x8 multiplier
* fix off-by-one error for bit shift bound
* restrict patchfrom speed optimization to strategy < ZSTD_btultra
* update results.csv
* update regression test
Part 2 of #3528
Adds hash salt that helps to avoid regressions where consecutive compressions use the same tag space with similar data (running zstd -b5e7 enwik8 -B128K reproduces this regression).
- Adds memory type that is guaranteed to have been initialized at least once in the workspace's lifetime.
- Changes tag space in row hash to be based on init once memory.
#3543 decreases the size of the tagTable by a factor of 2, which requires using the first tag position in each row for head position instead of a tag.
Although position 0 stopped being a valid match, it still persisted in mask calculation resulting in the matches loops possibly terminating before it should have. The fix skips position 0 to solve this problem.
Allocate half the memory for tag space, which means that we get one less slot for an actual tag (needs to be used for next position index).
The results is a slight loss in compression ratio (up to 0.2%) and some regressions/improvements to speed depending on level and sample. In turn, we get to save 16% of the hash table's space (5 bytes per entry instead of 6 bytes per entry).
As reported by @P-E-Meunier in https://github.com/facebook/zstd/issues/2662#issuecomment-1443836186,
seekable format ingestion speed can be particularly slow
when selected `FRAME_SIZE` is very small,
especially in combination with the recent row_hash compression mode.
The specific scenario mentioned was `pijul`,
using frame sizes of 256 bytes and level 10.
This is improved in this PR,
by providing approximate parameter adaptation to the compression process.
Tested locally on a M1 laptop,
ingestion of `enwik8` using `pijul` parameters
went from 35sec. (before this PR) to 2.5sec (with this PR).
For the specific corner case of a file full of zeroes,
this is even more pronounced, going from 45sec. to 0.5sec.
These benefits are unrelated to (and come on top of) other improvement efforts currently being made by @yoniko for the row_hash compression method specifically.
The `seekable_compress` test program has been updated to allows setting compression level,
in order to produce these performance results.
Current timeout is too small for some slower machines, e.g. most modern riscv64 boards,
where tests fail with the following diagnostics:
Traceback (most recent call last):
File "/usr/src/RPM/BUILD/zstd-1.5.4-alt2/tests/./cli-tests/run.py", line 734, in <module>
success = run_tests(tests, opts)
File "/usr/src/RPM/BUILD/zstd-1.5.4-alt2/tests/./cli-tests/run.py", line 601, in run_tests
tests[test_case.name] = test_case.run()
File "/usr/src/RPM/BUILD/zstd-1.5.4-alt2/tests/./cli-tests/run.py", line 285, in run
return self.analyze()
File "/usr/src/RPM/BUILD/zstd-1.5.4-alt2/tests/./cli-tests/run.py", line 275, in analyze
self._join_test()
File "/usr/src/RPM/BUILD/zstd-1.5.4-alt2/tests/./cli-tests/run.py", line 330, in _join_test
(stdout, stderr) = self._test_process.communicate(timeout=self._opts.timeout)
File "/usr/lib64/python3.10/subprocess.py", line 1154, in communicate
stdout, stderr = self._communicate(input, endtime, timeout)
File "/usr/lib64/python3.10/subprocess.py", line 2006, in _communicate
self._check_timeout(endtime, orig_timeout, stdout, stderr)
File "/usr/lib64/python3.10/subprocess.py", line 1198, in _check_timeout
raise TimeoutExpired(
subprocess.TimeoutExpired: Command '['/usr/src/RPM/BUILD/zstd-1.5.4-alt2/tests/cli-tests/compression/window-resize.sh']' timed out after 60 seconds
* Add ZSTD_setFParams() and ZSTD_setParams()
* Modify ZSTD_setCParams() to use ZSTD_setParameter() to avoid a second path setting parameters
* Add unit tests
* Update documentation to suggest using them to replace deprecated functions
Fixes#3396.
- Initializes clevel in `ZSTD_CCtxParams_init`
- Adds CI workflow for msan fuzzers runs without optimization (`-O0`)
- Fixes Makefile to correctly pass on user defined `MOREFLAGS` and `FUZZER_FLAGS` in cases they have been overwritten
The block splitter confuses sequences with literal length == 65536 that use a
repeat offset code. It interprets this as literal length == 0 when deciding the
meaning of the repeat offset, and corrupts the repeat offset history. This is
benign, merely causing suboptimal compression performance, if the confused
history is flushed before the end of the block, e.g. if there are 3 consecutive
non-repeat code sequences after the mistake. It also is only triggered if the
block splitter decided to split the block.
All that to say: This is a rare bug, and requires quite a few conditions to
trigger. However, the good news is that if you have a way to validate that the
decompressed data is correct, e.g. you've enabled zstd's checksum or have a
checksum elsewhere, the original data is very likely recoverable. So if you were
affected by this bug please reach out.
The fix is to remind the block splitter that the literal length is actually 64K.
The test case is a bit tricky to set up, but I've managed to reproduce the issue.
Thanks to @danlark1 for alerting us to the issue and providing us a reproducer!
fix#3500
CMake 3.18 or later was required by #3392. Because it uses
`CheckLinkerFlag`. But requiring CMake 3.18 or later is a bit
aggressive. Because Ubuntu 20.04 LTS still uses CMake 3.16.3:
https://packages.ubuntu.com/search?keywords=cmake
This change disables `-z noexecstack` check with old CMake. This will
not break any existing users. Because users who need `-z noexecstack`
must already use CMake 3.18 or later.
Implemented CI workflow for testing compilation with external compressors and without them. This serves as a sanity check to avoid any code dependencies on libraries that may not always be present. (Reference: #3497 for a bug fix related to this issue.)
Bytef and uInt are zlib types, not available when zlib is disabled
Fixes: 1598e6c634 ("Async write for decompression")
Fixes: cc0657f27d ("AsyncIO compression part 2 - added async read and asyncio to compression code (#3022)")
* Fixes zstd-dll build (https://github.com/facebook/zstd/issues/3492):
- Adds pool.o and threading.o dependency to the zstd-dll target
- Moves custom allocation functions into header to avoid needing to add dependency on common.o
- Adds test target for zstd-dll
- Adds github workflow that buildis zstd-dll
* fix and test MSVC AVX2 build
* treat msbuild warnings as errors
* fix incorrect MSVC 2019 compiler warning
* fix MSVC error D9035: option 'Gm' has been deprecated and will be removed in a future release
We need to run it for the tests, even if programs are disabled. So if
they are disabled, create a build rule for the program, but don't
install it. Just make it available for the test itself.
It depends on the zstd program being built, and passes it as an env
variable. Just like datagen. But for datagen, we explicitly depend on
it, while for zstd, we assume it's built as part of "all".
This can be wrong in two cases:
- when running individual tests, meson can (re)build just what is needed
for that one test
- a later patch will handle building zstd but not by default
This frame is invalid because the `Window_Size = 0`, and the
`Block_Maximum_Size = min(128 KB, Window_Size) = 0`. But the empty
compressed block has a `Block_Content` size of 2, which is invalid.
The fix is to switch to using a `Window_Descriptor` instead of the
`Single_Segment_Flag`. This sets the `Window_Size = 1024`.
Hexdump before this PR: `28b5 2ffd 2000 1500 0000 00`
Hexdump after this PR: `28b5 2ffd 0000 1500 0000 00`
For issue #3482.
such scenario can happen, for example,
when trying a decompression-only benchmark on invalid data.
Other possibilities include an allocation error in an intermediate step.
So far, the benchmark would return immediately, but still return 0.
On command line, this would be confusing, as the program appears successful (though it does not display any successful message).
Now it returns !0, which can be interpreted as an error by command line.
Note that the `fd` is only valid while the file is still open. So we need to
move the setting calls to before we close the file. However! We cannot do so
with the `utime()` call (even though `futimens()` exists) because the follow-
ing `close()` call to the `fd` will reset the atime of the file. So it seems
the `utime()` call has to happen after the file is closed.
Somewhat surprisingly, calling `fchmod()` is non-trivially faster than calling
`chmod()`, and so on.
This commit introduces alternate variants to some common file util functions
that take an optional fd. If present, they call the `f`-variant of the
underlying function. Otherwise, they fall back to the regular filename-taking
version of the function.
In 32-bit mode, ZSTD_getOffsetInfo() can be called when nbSeq == 0, and
in this case the offset table is uninitialized. The function should just
return 0 for both values, because there are no sequences.
Credit to OSS-Fuzz
The 32-bit decoder could corrupt the regenerated data by using regular
offset mode when there were actually long offsets. This is because we
were only considering the window size in the calculation, not the
dictionary size. So a large dictionary could allow longer offsets.
Fix this in two ways:
1. Instead of looking at the window size, look at the total referencable
bytes in the history buffer. Use this in the comparison instead of
the window size. Additionally, we were comparing against the wrong
value, it was too low. Fix that by computing exactly the maximum
offset for regular sequence decoding.
2. If it is possible that we have long offsets due to (1), then check
the offset code decoding table, and if the decoding table's maximum
number of additional bits is no more than STREAM_ACCUMULATOR_MIN,
then we can't have long offsets.
This gates us to be using the long offsets decoder only when we are very
likely to actually have long offsets.
Note that this bug only affects the decoding of the data, and the
original compressed data, if re-read with a patched decoder, will
correctly regenerate the orginal data. Except that the encoder also had
the same issue previously.
This fixes both the open OSS-Fuzz issues.
Credit to OSS-Fuzz
The previous code had an issue when `bitsConsumed == 32` it would read 0
bits for the `ofBits` read, which violates the precondition of
`BIT_readBitsFast()`. This can happen when the stream is corrupted.
Fix thie issue by always reading the maximum possible number of extra
bits. I've measured neutral decoding performance, likely because this
branch is unlikely, but this should be faster anyways. And if not, it is
only 32-bit decoding, so performance isn't as critical.
Credit to OSS-Fuzz
Delete all unused FSE functions, now that we are no longer syncing
to/from upstream.
This avoids confusion about Zstd's stack usage like in Issue #3453.
It also removes dead code, which is always a plus.
The input bounds checks were buggy because they were only breaking from
the inner loop, not the outer loop. The fuzzers found this immediately.
The fix is to use `goto _out` instead of `break`.
This condition can happen on corrupted inputs.
I've benchmarked before and after on x86-64 and there were small changes
in performance, some positive, and some negative, and they end up about
balacing out.
Credit to OSS-Fuzz
Previously, cli-test would, by default, check that a stderr output is strictly identical to a saved outcome.
When there was no instructions on how to interpret stderr, it would default to requiring it to be empty.
There are many tests cases though where stderr content doesn't matter, and we are mainly interested in the return code of the cli.
For these cases, it was possible to set a .ignore document, which would instruct to ignore stderr content.
This PR update the logic, to make .ignore the default.
When willing to check that stderr content is empty, one must now add an empty .strict file.
This will allow status message to evolve without triggering many cli-tests errors.
This is especially important when some of these status include compression results, which may change as a result of compression optimizations.
It also makes it easier to add new tests which only care about the CLI's return code.
[Bugfix] CLI row hash flags set the wrong values
`--[no-]row-match-finder` do the opposite of what they are supposed to.
In effect the no option would activate row hash while the other option will disable it.
This commit fixes the issue and changes the code to use the more readable enum values.
Before calling a dictionary good, make sure that it can compress an
input. If v0.7.3 rejects v0.7.3's dictionary, fall back to the v1.0
dictionary. This is not the job of the verison test to test it, because
we cannot fix this code.
Add generic C versions of the fast decoding loops to serve architectures
that don't have an assembly implementation. Also allow selecting the C
decoding loop over the assembly decoding loop through a zstd
decompression parameter `ZSTD_d_disableHuffmanAssembly`.
I benchmarked on my Intel i9-9900K and my Macbook Air with an M1 processor.
The benchmark command forces zstd to compress without any matches, using
only literals compression, and measures only Huffman decompression speed:
```
zstd -b1e1 --compress-literals --zstd=tlen=131072 silesia.tar
```
The new fast decoding loops outperform the previous implementation uniformly,
but don't beat the x86-64 assembly. Additionally, the fast C decoding loops suffer
from the same stability problems that we've seen in the past, where the assembly
version doesn't. So even though clang gets close to assembly on x86-64, it still
has stability issues.
| Arch | Function | Compiler | Default (MB/s) | Assembly (MB/s) | Fast (MB/s) |
|---------|----------------|--------------|----------------|-----------------|-------------|
| x86-64 | decompress 4X1 | gcc-12.2.0 | 1029.6 | 1308.1 | 1208.1 |
| x86-64 | decompress 4X1 | clang-14.0.6 | 1019.3 | 1305.6 | 1276.3 |
| x86-64 | decompress 4X2 | gcc-12.2.0 | 1348.5 | 1657.0 | 1374.1 |
| x86-64 | decompress 4X2 | clang-14.0.6 | 1027.6 | 1659.9 | 1468.1 |
| aarch64 | decompress 4X1 | clang-12.0.5 | 1081.0 | N/A | 1234.9 |
| aarch64 | decompress 4X2 | clang-12.0.5 | 1270.0 | N/A | 1516.6 |
Fix the following documentation bugs:
* Note that `ZSTD_estimate*` functions are not compatible with the external matchfinder API
* Note that `ZSTD_estimateCStreamSize_usingCCtxParams()` is not compatible with `nbWorkers >= 1`
* Remove incorrect warning that the legacy streaming API is incompatible with advanced parameters and/or dictionary compression
* Note that `ZSTD_initCStream()` is incompatible with dictionary compression
* Warn that
`zstd` CLI has progressively moved to the policy of
ignoring `--rm` command when the output is `stdout`.
The primary drive is to feature a behavior more consistent with `gzip`,
when `--rm` is the default, but is also ignored when output is `stdout`.
Other policies are certainly possible, but would break from this `gzip` convention.
The new policy was inconsistenly enforced, depending on the exact list of commands.
For example, it was possible to circumvent it by using `-c --rm` in this order,
which would re-establish source removal.
- Update the CLI so that it necessarily catch these situations and ensure that `--rm` is always disabled when output is `stdout`.
- Added a warning message in this case (for verbosity 3 `-v`).
- Added an `assert()`, which controls that `--rm` is no longer active with `stdout`
- Added tests, which control the behavior, even when `--rm` is added after `-c`
- Removed some legacy code which where trying to apply a specific policy for the `stdout` + `--rm` case, which is no longer possible
In GCC, we can add a couple more flags to give us confidence that the profile
data is actually being found and used.
Also, my system for example doesn't have a binary installed under the name
`llvm-profdata`, but it does have, e.g., `llvm-profdata-13`, etc. So this
commit adds a variable that can be overridden.
Older versions of zstandard have a bug in the dictionary builder, that
can cause dictionary building to fail. The process still exits 0, but
the dictionary is not created.
For reference, the bug is that it creates a dictionary that starts with
the zstd dictionary magic, in the process of writing the dictionary header,
but the header isn't fully written yet, and zstd fails compressions in
this case, because the dictionary is malformated. We fixed this later on
by trying to load the dictionary as a zstd dictionary, but if that fails
we fallback to content only (by default).
The fix is to:
1. Make the dictionary determinsitic by sorting the input files.
Previously the bug would only sometimes occur, when the input files
were in a particular order.
2. If dictionary creation fails, fallback to the `head` dictionary.
* Cap shortCache chainLog to 24
* Cap row match finder hashLog so that rowLog <= 24
* Add unit tests to expose all cases. The row match finder unit tests
are only run in 64-bit mode, because they allocate ~1GB.
Fixes#3336
The dictionary source files were taken from the `dev` branch before this
commit, which could introduce non-determinism on PR jobs. Instead take
the sources from the PR checkout.
This PR also adds stderr logging, and verbose output for the jobs that
are failing, to help catch the failure if it occurs again.
* deprecate advanced streaming functions
* remove internal usage of the deprecated functions
* nit
* suppress warnings in tests/zstreamtest.c
* purge ZSTD_initDStream_usingDict
* nits
* c90 compat
* zstreamtest.c already disables deprecation warnings!
* fix initDStream() return value
* fix typo
* wasn't able to import private symbol properly, this commit works around that
* new strategy for zbuff
* undo zbuff deprecation warning changes
* move ZSTD_DISABLE_DEPRECATE_WARNINGS from .h to .c
The timer storage type is no longer dependent on OS.
This will make it possible to re-enable posix precise timers
since the timer storage type will no longer be sensible to #include order.
See #3168 for details of pbs of previous interface.
Suggestion by @terrelln
* Add a function and macro ZSTD_decompressionMargin() that computes the
decompression margin for in-place decompression. The function computes
a tight margin that works in all cases, and the macro computes an upper
bound that will only work if flush isn't used.
* When doing in-place decompression, make sure that our output buffer
doesn't overlap with the input buffer. This ensures that we don't
decide to use the portion of the output buffer that overlaps the input
buffer for temporary memory, like for literals.
* Add a simple unit test.
* Add in-place decompression to the simple_round_trip and
stream_round_trip fuzzers. This should help verify that our margin stays
correct.
A minor change in 5434de0 changed a `<=` into a `<`,
and as an indirect consequence allowed compression attempt of literals when there are only 6 literals to compress
(previous limit was effectively 7 literals).
This is not in itself a problem, as the threshold is merely an heuristic,
but it emerged a bug that has always been there, and was just never triggered so far due to the previous limit.
This bug would make the literal compressor believes that all literals are the same symbol,
but for the exact case where nbLiterals==6, plus a pretty wild combination of other limit conditions,
this outcome could be false, resulting in data corruption.
Replaced the blind heuristic by an actual test for all limit cases,
so that even if the threshold is changed again in the future,
the detection of RLE mode will remain reliable.
At Google we fuzz zstd without ZSTD_MULTITHREAD but we want inputs to be as much as reproducible. It allows us to test new fuzzing methods for our fuzz team internally and have more horsepower to find bugs
comparing level 19 to level 22 and expecting a stricter better result from level 22
is not that guaranteed,
because level 19 and 22 are very close to each other,
especially for small files,
so any noise in the final compression result
result in failing this test.
Level 22 could be compared to something much lower, like level 15,
But level 19 is required anyway, because there is a clamping test which depends on it.
Removed level 22, kept level 19
fix#3328
In situations where the alphabet size is very small,
the evaluation of literal costs from the Optimal Parser is initially incorrect.
It takes some time to converge, during which compression is less efficient.
This is especially important for small files,
because there will not be enough data to converge,
so most of the parsing is selected based on incorrect metrics.
After this patch, the scenario ##3328 gets fixed,
delivering the expected 29 bytes compressed size (smallest known compressed size).
Inspired by #3395,
offer a new capability to set all parameters defined in a ZSTD_compressionParameters structure
with a single symbol invocation
to improve user code brevity.
Reported by @shulib :
the specification for 4-streams mode
doesn't work when the amount of literals to compress is 5 bytes.
Extending it, it also doesn't work for sizes 1 or 2.
This patch updates the specification and the implementation
to require a minimum of 6 literals to trigger or accept the 4-streams mode.
The impact is expected to be a no-op :
the 4-streams mode is never triggered for such small quantity of literals anyway,
since it would be wasteful (it costs ~7.3 bytes more than single-stream mode).
An informal lower limit is set at ~256 bytes,
so the technical minimum is very far from this limit.
This is just meant for completeness of the specification.
Basic tests for (de)compressing in the following modes:
* file to file
* file to stdout
* stdin to file
* stdin to stdout
These are basic tests, and aren't testing more advanced scenarios, but
it adds the groundwork for more complex tests as needed.
Fixes#3010.
Thanks to @eli-schwartz for pointing it out!
We should maybe consider adding a helper function for applying
`ZSTD_parameters` and `ZSTD_compressionParameters` to a context.
That would aid the transition to the new API in situations like this.
`./run.py --set-exact-output` will update `stdout.expect` and
`stderr.expect` to match the expected output. This doesn't apply to
outputs which use `.glob` or `.ignore`.
The one that isn't pinned is the OSS-Fuzz builder and runner. They don't
offer tagged releases. I could pin to the current master commit, but I'm not
sure how desirable that is.
Tell CMake to explicitly compile our assembly as C code, because we
require it is compiled by a C compiler, and it is only enabled for
clang/gcc.
Fixes#3193.
```
for f in $(find . \( -path ./.git -o -path ./tests/fuzz/corpora -o -path ./tests/regression/data-cache -o -path ./tests/regression/cache \) -prune -o -type f);
do
sed -i 's/\-2021/-present/' $f;
done
g co HEAD -- .github/workflows/dev-short-tests.yml # fix bad match
```
```
for f in $(find . \( -path ./.git -o -path ./tests/fuzz/corpora \) -prune -o -type f);
do
sed -i 's/Facebook, Inc\./Meta Platforms, Inc. and affiliates./' $f;
done
```
* Add `Portability.h` to fix min/max issues.
* Fix conversion warnings
* Assert that windowLog <= 23, which is currently always the case.
This could be loosened, but we aren't looking to add new functionality.
Fixes on top of PR #3375 by @eli-schwartz, which added Windows CI for contrib & programs.
When spawning a Windows thread we have small worker wrapper function that translates
between the interfaces of Windows and POSIX threads.
This wrapper is given a pointer that might get stale before the worker starts running,
resulting in UB and crashes.
This commit adds synchronization so that we know the wrapper has finished reading the data
it needs before we allow the main thread to resume execution.
1. If threads are resized the threads' `ZSTD_pthread_t` might move
while the worker still holds a pointer into it (see more details in #3120).
2. The join operation was waiting for a thread and then return its `thread.arg`
as a return value, but since the `ZSTD_pthread_t thread` was passed by value it
would have a stale `arg` that wouldn't match the thread's actual return value.
This fix changes the `ZSTD_pthread_join` API and removes support for returning
a value. This means that we are diverging from the `pthread_join` API and this
is no longer just an alias.
In the future, if needed, we could return a Windows thread's return value using
`GetExitCodeThread`, but as this path wouldn't be excised in any case, it's
preferable to not add it right now.
Fix `zdict.h` static linking only section so if you include it twice it
still exposes the static linking only symbols. E.g. this pattern:
```
```
This can easily happen when a header you include includes `zdict.h`.
`ZSTD_LIB_MINIFY` broke in 8bf699aa59.
This commit fixes the macro and the static library shrinks from ~600K to 324K
with ZSTD_LIB_MINIFY set.
Fixes#3066.
1. Follow the scheme introduced in PR #2501 for both `zdict.h` and `zstd_errors.h`.
2. If the `*_VISIBLE` macro isn't set, but the `*_VISIBILITY` macro is, use that.
Also make this change for `zstd.h`, since we probably shouldn't have changed
that macro name without backward compatibility in the first place.
3. Change all references to `*_VISIBILITY` to `*_VISIBLE`.
Fixes#3359.
It seems like with the deletion of Travis CI we didn't successfully transfer the
version compatibility test. Attempt to enable the version compatibility test.
There are a couple of oddities here. We don't attempt to build e.g.
contrib, because that doesn't seem to work at the moment. Also notice
that each command is its own step. This happens because github actions
runs in powershell, which doesn't seem to let you abort on the first
failure.
playTests.sh has an option to run really slow tests. This is enabled by
default in Meson, but what we really want is to do like the Makefile,
and run the fast ones by default, but with an option to run the slow
ones instead.
It's entirely possible some people don't have valgrind installed, but
still want to run the tests. If they don't have it installed, then they
probably don't intend to run those precise test targets anyway.
Also, this solves an error when running the tests in an automated
environment. The valgrind tests have a hard dependency on behavior such
as `./zstd` erroring out with the message "stdin is a console, aborting"
which does not work if the automated environment doesn't have a console.
As a rough heuristic, automated environments lacking a console will
*probably* also not have valgrind, so avoiding that test definition
neatly sidesteps the issue.
Also, valgrind is not easily installable on macOS, at least homebrew
says it isn't available there. This makes it needlessly hard to
enable the testsuite on macOS.
In commit 031de3c69c a feature of Meson
0.50.0 was added, but the minimum specified version of Meson is 0.48.0.
Meson therefore emitted a warning:
WARNING: Project targets '>=0.48.0' but uses feature introduced in '0.50.0': required arg in compiler.has_header.
And if anyone actually used Meson 0.48.0 to build with, it would error
out with mysterious claims that the build file itself is invalid, rather
than telling the user to install a newer version of Meson.
Solve this by bumping the minimum version to align with reality. This
e.g. drops support for Debian oldstable (buster)'s packaged version of
Meson, but still works if backports are enabled, or if the user can
`pip install` a newer version.
In commit 031de3c69c some code was added
that returned a boolean, but was treated as if it returned a dependency
object. This wasn't tested and could not work. Moreover, zstd no longer
built at all unless the entire programs directory was disabled and not
even evaluated.
Fix the return type checking.
It uses non-portable compiler options unconditionally. Elsewhere, we
check the compiler ID and only add the right ones, globally. Do the same
here.
NDEBUG can actually be handled by a core option, so while we are moving
things around, do so.
Unfortunately, this doesn't fix things entirely. The remaining issue is
not Meson's issue though -- MSVC simply does not like this source code
and somehow chokes on innocent code with the inscrutable "syntax error"
and "illegal token".
fixed#3323, reported by @nigeltao
Completed documentation around this risk
(which is largely theoretical,
I can't see that happening in any "real world" scenario,
but an erroneous @srcSize value could indeed trigger it).
Fix an off-by-one error in the compressor that emits corrupt blocks if:
* Zstd is compiled in 32-bit mode
* The windowLog == 25 exactly
* An offset of 2^25-3, 2^25-2, 2^25-1, or 2^25 is emitted
* The bitstream had 7 bits leftover before writing the offset
This bug has been present since before v1.0, but wasn't able to easily
be triggered, since until somewhat recently zstd wasn't able to find
matches that were within 128KB of the window size.
Add a test case, and fix 2 bugs in `ZSTD_compressSequences()`:
* The `ZSTD_isRLE()` check was incorrect. It wouldn't produce
corruption, but it could waste CPU and not emit RLE even if the block
was RLE
* One windowSize was `1 << windowLog`, not `1u << windowLog`
Thanks to @tansy for finding the issue, and giving us a reproducer!
Fixes Issue #3350.
Delete unaligned memory access code from the legacy codebase by removing all the
non-memcpy functions. We don't care about speed at all for this codebase, only
simplicity.
Fix an instance of `NULL + 0` in `ZSTD_decompressStream()`. Also, improve our
`stream_decompress` fuzzer to pass `NULL` in/out buffers to
`ZSTD_decompressStream()`, and fix 2 issues that were immediately surfaced.
Fixes#3351
Split the logic for parameter adaption from the logic to update the display rate.
This decouples the two updates, so changes to display updates don't affect
parameter adaption.
Also add a test case that checks that parameter adaption actually happens.
This fixes Issue #3353, where --adapt is broken when --no-progress is passed.
Instead of using packed attribute hack, just use aligned attribute. It
improves code generation on armv6 and armv7, and slightly improves code
generation on aarch64. GCC generates identical code to regular aligned
access on ARMv6 for all versions between 4.5 and trunk, except GCC 5
which is buggy and generates the same (bad) code as packed access:
https://gcc.godbolt.org/z/hq37rz7sb
* Centralize the logic about whether to print the progress bar or not in
the `*_PROGRESS()` macros.
* Centralize the logc about whether to print the summary line or not in
`FIO_shouldDisplayFileSummary()` and
`FIO_shouldDisplayMultipleFileSummary()`.
* Make `--progress` work for non-zstd (de)compressors.
* Clean up several edge cases in compression and decompression progress
printing along the way. E.g. wrong log level, or missing summary line.
One thing I don't like about stdout mode, which sets the display level
to 1, is that warnings aren't displayed. After this PR, we could change
stdout mode from lowering the display level, to defaulting to implied
`--no-progress`. But, I think that deserves a separate PR.
We've been unable to effectively test cases where stdin/stdout/stderr
are consoles, because in our test cases they generally aren't. Allow the
command line flags `--fake-std{in,out,err}-is-console` to tell the CLI
to pretend that std{in,out,err} is a console.
Newer gcc versions were getting smart and omitting the `memset()`.
Get around this issue by outlining the `memset()` into a different
function. This test is still hacky, but it works...
Run the scraper command to establish the project version immediately,
rather than wait for the build to be configured. This simplifies the
code and ensures that project introspection works correctly.
Disable ASM in the kernel for now. It requires a few changes & setup to
get working. Instead of doing it in a zstd version update, I'd prefer to
package that change as a single patch, and propose it separately from
the version update. This makes the version update easier, and reduces
some risk.
Use a switch statement to select the search function instead of an
indirect function call. This results in a sizable performance win.
This PR is a modification of the approach taken in PR #2828.
When I measured performance for that commit, it was neutral.
However, I now see a performance regression on gcc, but still
neutral on clang. I'm measuring on the same platform, but with
newer compilers. The new approach beats both the current dev
branch and the baseline before PR #2828 was merged.
This PR is necessary for Issue #3275, to update zstd in the kernel.
Without this PR there is a large regression in greedy - btlazy2
compression speed. With this PR it is about neutral.
gcc version: 12.2.0
clang version: 14.0.6
dataset: silesia.tar
| Compiler | Level | Dev Speed (MB/s) | PR Speed (MB/s) | Delta |
|----------|-------|------------------|-----------------|--------|
| gcc | 5 | 102.6 | 113.7 | +10.8% |
| gcc | 7 | 66.6 | 74.8 | +12.3% |
| gcc | 9 | 51.5 | 58.9 | +14.3% |
| gcc | 13 | 14.3 | 14.3 | +0.0% |
| clang | 5 | 108.1 | 114.8 | +6.2% |
| clang | 7 | 68.5 | 72.3 | +5.5% |
| clang | 9 | 53.2 | 56.2 | +5.6% |
| clang | 13 | 14.3 | 14.7 | +2.8% |
The binary size stays just about the same for clang and gcc, measured
using the `size` command:
| Compiler | Branch | Text | Data | BSS | Total |
|----------|--------|---------|------|-----|---------|
| gcc | dev | 1127950 | 3312 | 280 | 1131542 |
| gcc | PR | 1123422 | 2512 | 280 | 1126214 |
| clang | dev | 1046254 | 3256 | 216 | 1049726 |
| clang | PR | 1048198 | 2296 | 216 | 1050710 |
Add a `--spdx` option to the freestanding script to prefix
files with a line like (for `.c` files):
// SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause
or (for `.h` and `.S` files):
/* SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause */
Given the style of the line to be used depends on the extension,
a simple `sed` insert command would not work.
It also skips the file if an existing SPDX line is there,
as well as raising an error if an unexpected SPDX line appears
anywhere else in the file, as well as for unexpected
file extensions.
I double-checked that all currently generated files appear
to be license as expected with:
grep -LRF 'This source code is licensed under both the BSD-style license (found in the' linux/lib/zstd
grep -LRF 'LICENSE file in the root directory of this source tree) and the GPLv2 (found' linux/lib/zstd
but somebody knowledgable on the licensing of the project should
double-check this is the intended case.
Fixes: https://github.com/facebook/zstd/issues/3293
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Currently this function actually reads the dict ID from the dictionary's
header, via `ZSTD_getDictID_fromDict()`. But during decompression the decomp-
ressor actually compares the dict ID in the frame header with the dict ID in
the DDict. Now of course the dict ID in the dictionary contents and the dict
ID in the DDict struct *should* be the same. But in cases of memory corrupt-
ion, where they can drift out of sync, it's misleading for this function to
read it again from the dict buffer rather then return the dict ID that will
actually be used.
Also doing it this way avoids rechecking the magic and so on and so it is a
tiny bit more efficient.
Since your Makefile uses obj/$(HASH_DIR) for object files, this code does not work correctly for GCC. Because profiles are saved in one directory, and are expected in another when reading.
`$(RM) zstd *.o` - this line doesn't delete object files.
Clang stores profiles in the current directory, so the problem doesn't appear when compiling with Clang.
Also this code will work if BUILD_DIR is set.
Clang doesn't allow [[deprecated(...)]] attribute after __attribute__.
Move [[deprecated(...)]] before __attribute__ to fix C++14/C++17 uses
with Clang.
Fix#3250
* add checks to mal-formed numeric values for memory and memlimit parameters
Signed-off-by: Ly Cao <lycao@fb.com>
* changed errorMsg to a literal string instead of static string in main
* moved bogus numeric error to NEXT_UINT32 + add macro NEXT_TSIZE
Signed-off-by: Ly Cao <lycao@fb.com>
Signed-off-by: Ly Cao <lycao@fb.com>
Co-authored-by: Ly Cao <lycao@fb.com>
As mentioned in
https://github.com/facebook/zstd/pull/3252#issuecomment-1251733791 ,
this patch adds a CI job that builds and installs libzstd on the job
runner, and then compiles a sample binary linking against the installed
library; the needed build flags are passed by invoking pkg-config.
Comparing 4B instead of comparing 1B in ZSTD_noDict
mode, thus it can avoid cases like match in match[ml]
but mismatch in match[ml-3]..match[ml-1]. So the call
count of ZSTD_count can be reduced.
Signed-off-by: Jun He <jun.he@arm.com>
Change-Id: I3449ea423d5c8e8344f75341f19a2d1643c703f6
When creating a new `Makefile` target to build,
it's also necessary to update the `clean` target,
which purpose is to remove built targets when they are present.
This process is simple, but it's also easy to forget :
since there is a large distance between the position in the `Makefile` where the new built target is added,
and the place where the list of files to `clean` is defined.
Moreover, the list of files becomes pretty long over time,
hence it's difficult to visually ensure that all built targets are present there,
or that no old target (no longer produced) is no longer in the list
This PR tries to improve this process by adding a CLEAN variable.
Now, when a new built target is added to the `Makefile`,
it should preceded by :
```
CLEAN += newTarget
newTarget:
<TAB> ...recipe...
```
This new requirement is somewhat similar to `.PHONY: newTarget` for non-built targets.
This new method offers the advantage of locality :
there is no separate place in the file to maintain a list of files to clean.
This makes maintenance of `make clean` easier.
With this patch the pkg-config generation when using the CMake build
system is improved in the following ways:
- Libs.private is now filled when needed
- The JoinPaths module is now used to join paths, leading to simpler
code
- The .pc file is always generated, regardless of the platform, as it
can also be consumed on Windows
Here's how the .pc file is affected by these changes, in comparison to
the one generated with the official Makefiles:
$ diff -s lib/libzstd.pc build/cmake/build-old/lib/libzstd.pc
15c15
< Libs.private: -pthread
---
> Libs.private:
$ diff -s lib/libzstd.pc build/cmake/build-new/lib/libzstd.pc
Files lib/libzstd.pc and build/cmake/build-new/lib/libzstd.pc are
identical
Adds documentation to help and man pages for legacy pass-through behavior
when force is set and destination is stdout. Documents --pass-through in
man pages
Fixes#3211.
Adds the `--[no-]pass-through` flag which enables/disables pass-through mode.
* `zstdcat`, `zcat`, and `gzcat` default to `--pass-through`.
Pass-through mode can be disabled by passing `--no-pass-through`.
* All other binaries default to not setting pass-through mode.
However, we preserve the legacy behavior of enabling pass-through
mode when writing to stdout with `-f` set, unless pass-through
mode is explicitly disabled with `--no-pass-through`.
Adds a new test for this behavior that should codify the behavior we want.
fileio_types.h cannot be parsed by itself
because it relies on basic types defined in `lib/common/mem.h`.
As for #3231, it likely wasn't detected because `mem.h` was probably included before within target files.
But this is not proper.
A "easy" solution would be to add the missing include,
but each dependency should be considered "bad" by default,
and only allowed if it brings some tangible value.
In this case, since these types are only used to declare internal structure variables
which are effectively only flags,
I believe it's really not valuable to add a dependency on `mem.h` for this purpose
while the standard `int` type can do the same job.
I was expecting some compiler warnings following this change,
but it turns out we don't use `-Wconversion` by default on `zstd` source code,
so there is none.
Nevertheless, I enabled `-Wconversion` locally and proceeded to fix a few conversion warnings in the process.
Adding `-Wconversion` to the list of flags used for `zstd` is something I would be favorable over the long term,
but it cannot be done overnight,
because the nb of places where this warning is triggered is daunting.
Better progressively reduce the nb of triggered `-Wconversion` warnings before enabling this flag by default.
Fixes#3212.
Long literal and match lengths had an off-by-one error in ZSTD_getSequenceLength.
Fix the off-by-one error, and add a golden compression test that catches the bug.
Also run all the golden tests in the cli-tests framework.
The discussion for this task is here: facebook/zstd#3128.
This task can probably be scoped to the first part: marking these functions deprecated.
We'll later look at removal when we roll out v1.6.0.
When user pass in argument for both decompression and multi-thread, print a warning message
to indicate that multi-threaded decompression is not supported.
* Add warning when multi-thread decompression is requested
* add test case for multi-threaded decoding warning
Expectation is for -d -T0 we will not throw any warning,
and see warning for any other -d -T(>1) inputs
In zlib 1.2.12 the OF macro was changed to _Z_OF breaking any
project that used zlibWrapper. To fix this the OF has been
changed to _Z_OF everywhere and _Z_OF is defined as OF in the
case it is not yet defined for zlib 1.2.11 and older.
Fixes: https://github.com/facebook/zstd/issues/3216
* Fixing compiler warnings
* Replace the old -s flag with the -Wl,-s flag
* Fixing compiler warnings
* Fixing the linker strip flag and tests/code not working as expected on AIX
With statistic data of test data files of silesia
the chance of position beyond highThreshold is very
low (~1.3%@L8 in most cases, all <2.5%), and is in
"lowprob area". Add the branch hint so compiler can
get better pipiline codegen.
With this change it is observed ~1% of mozilla and
xml, and slight (0.3%~0.8%) but consistent uplift on
other files on Arm N1.
Signed-off-by: Jun He <jun.he@arm.com>
Change-Id: Id9ba1d5c767e975290b5c1bf0ecce906544f4ade
match is used for following sequence copy. It is
only updated when extDict is needed, which is a
low probability case. So it can be prefetched to
reduce cache miss.
The benchmarks on various Arm platforms showed
uplift from 1% ~ 14% with gcc-11/clang-14.
Signed-off-by: Jun He <jun.he@arm.com>
Change-Id: If201af4799d2455d74c79f8387404439d7f684ae
Summary:
Freeing an uninitialized pointer is undefined behavior. This caused a segfault
when compiling the benchmark with Clang -O3 and benching decompression.
V2: always create compressInstructions but check if cctxParams is NULL before
setting CCtx params to avoid segfault.
Test Plan:
make and run
Summary:
Added an option -p# where -p0 (default) sets the aggregation method to fastest
speed while -p1 sets the aggregation method to median. Also added a new column
in the csv file to report this option's value.
Test Plan:
``
$ ./largeNbDicts -1 --nbDicts=1 -D ~/benchmarks/html/html_8_16K.32K.dict
~/benchmarks/html/html_8_16K/*
loading 7450 files...
created src buffer of size 83.4 MB
split input into 7450 blocks
loading dictionary /home/zhuhan/benchmarks/html/html_8_16K.32K.dict
compressing at level 1 without dictionary : Ratio=3.03 (28827863 bytes)
compressed using a 32768 bytes dictionary : Ratio=4.28 (20410262 bytes)
generating 1 dictionaries, using 0.1 MB of memory
Compression Speed : 306.0 MB/s
Fastest Speed : 310.6 MB/s
$ ./largeNbDicts -1 --nbDicts=1 -p1 -D ~/benchmarks/html/html_8_16K.32K.dict
~/benchmarks/html/html_8_16K/*
loading 7450 files...
created src buffer of size 83.4 MB
split input into 7450 blocks
loading dictionary /home/zhuhan/benchmarks/html/html_8_16K.32K.dict
compressing at level 1 without dictionary : Ratio=3.03 (28827863 bytes)
compressed using a 32768 bytes dictionary : Ratio=4.28 (20410262 bytes)
generating 1 dictionaries, using 0.1 MB of memory
Compression Speed : 306.9 MB/s
Median Speed : 298.4 MB/s
```
Summary:
Add column headers and data for whether it's a compression or a decompression
run, compression level, nbDicts and dictAttachPref in additional to
compr/decompr speed.
Test Plan:
Example output:
```
./largeNbDicts
Compression/Decompression,Level,nbDicts,dictAttachPref,Speed
Compression,1,1,0,300.9
Compression,1,1,1,296.4
Compression,1,1,2,307.8
Compression,1,10,0,292.3
Compression,1,100,0,293.3
Compression,3,110,0,106.0
Decompression,-1,110,-1,155.6
Decompression,-1,110,-1,709.4
Decompression,-1,120,-1,709.1
Decompression,-1,120,-1,734.6
```
Benchmarking decompression results in a segfault in `createCompressInstructions`
because `cctxParams` is NULL. Skip running that function if we are not benching
compression.
* Intial commit to address 3090. Added support to decompress empty block
* Update zstd_decompress_block.c
Addressed review comments for the case of 'set_basic'
* Update lib/decompress/zstd_decompress_block.c
Co-authored-by: Nick Terrell <nickrterrell@gmail.com>
* Update lib/decompress/zstd_decompress_block.c
Co-authored-by: Nick Terrell <nickrterrell@gmail.com>
Co-authored-by: Nick Terrell <nickrterrell@gmail.com>
The discussion for this task is here: facebook/zstd#3128.
This task can probably be scoped to the first part: marking these functions deprecated.
We'll later look at removal when we roll out v1.6.0.
Streaming decompression used to wait for a minimum of 5 bytes before attempting decoding.
This meant that, in the case that only a few bytes (<5) were provided,
and assuming these bytes are incorrect,
there would be no error reported.
The streaming API would simply request more data, waiting for at least 5 bytes.
This PR makes it possible to detect incorrect Frame IDs as soon as the first byte is provided.
Fix#3169
* first attempt at fast DMS short cache
* significant wins for some scenarios
* fix all clang regressions
* nits
* fix 1.5% gcc11 regression on hot 110Kdict scenario
* fix CI
* nit
* Add tags to doublefast hash table
* use tags in doublefast DMS
* Fix CI
* Clean up some hardcoded logic / constants
* Switch forCCtx to an enum
* nit
* add short cache to ip+1 long search
* Move tag size into hashLog
* Minor nits
* Truncate dictionaries greater than 16MB in short cache mode
* Helper function for tag comparison
* Cap short cache hashLog at 24 to prevent overflow
* size_t dictTagsMatch -> int dictTagsMatch
* nit
* Clean up and comment dictionary truncation
* Move ZSTD_tableFillPurpose_e next to ZSTD_dictTableLoadMethod_e
* Comment and expand helper functions
* Asserts and documentation
* nit
This assert slows the loop down by 10x. We can get similar
coverage by asserting at the beginning & end of the loop.
We need this fix because Debian compiles zstd with asserts
enabled. Separately, we should ask them why, and if they would
consider disabling asserts in their builds. Since we don't
optimize for assert enabled builds.
Fixes Issue #3150.
ZSTD_seqSymbol is a structure with total of 64 bits
wide. So it can be loaded in one operation and
extract its fields by simply shifting or extracting
on aarch64.
GCC doesn't recognize this and generates more
unnecessary ldr/ldrb/ldrh operations that cause
performance drop.
With this change it is observed 2~4% uplift of
silesia and 2.5~6% of cantrbry @L8 on Arm N1.
Signed-off-by: Jun He <jun.he@arm.com>
Change-Id: I7748909204cf78a17eb9d4f2333692d53239daa8
On aarch64 ZSTD_wildcopy uses a simple loop to do
16B based memory copy. There is existing optimized
two stage copy that can achieve better performance.
By applying this to aarch64 it is also observed ~1%
uplift in silesia corpus.
Signed-off-by: Jun He <jun.he@arm.com>
Change-Id: Ic1253308e7a8a7df2d08963ba544e086c81ce8be
We found that movemask is not used properly or consumes too much CPU.
This effort helps to optimize the movemask emulation on ARM.
For level 8-9 we saw 3-5% improvements. For level 10 we say 1.5%
improvement.
The key idea is not to use pure movemasks but to have groups of bits.
For rowEntries == 16, 32 we are going to have groups of size 4 and 2
respectively. It means that each bit will be duplicated within the group
Then we do AND to have only one bit set in the group so that iteration
with lowering bit `a &= (a - 1)` works as well.
Also, aarch64 does not have rotate instructions for 16 bit, only for 32
and 64, that's why we see more improvements for level 8-9.
vshrn_n_u16 instruction is used to achieve that: vshrn_n_u16 shifts by
4 every u16 and narrows to 8 lower bits. See the picture below. It's
also used in
[Folly](https://github.com/facebook/folly/blob/c5702590080aa5d0e8d666d91861d64634065132/folly/container/detail/F14Table.h#L446).
It also uses 2 cycles according to Neoverse-N{1,2} guidelines.
64 bit movemask is already well optimized. We have ongoing experiments
but were not able to validate other implementations work reliably faster.
This commit avoids checking whether a hashtable write is safe in two of the
three match-found paths in `ZSTD_compressBlock_fast_noDict_generic`. This pro-
duces a ~0.5% speed-up in compression.
A comment in the code describes why we can skip this check in the other two
paths (the repcode check and the first match check in the unrolled loop).
A downside is that in the new position where we make this check, we have not
yet computed `mLength`. We therefore have to avoid writing *possibly* dangerous
positions, rather than the old check which only avoids writing *actually*
dangerous positions. This leads to a miniscule loss in ratio (remember that
this scenario can only been triggered in very negative levels or under incomp-
ressibility acceleration).
Partial, Meson-only implementation of #2976 for non-MSVC builds.
Due to the prevalence of private symbol reuse, linking to a shared
library is simply utterly unreliable, but we still want to defer to the
shared library for installable applications. By linking to both, we can
share symbols where possible, and statically link where needed.
This means we no longer need to manually track every file that needs to
be extracted and reused.
The flip side is that MSVC completely does not support this, so for MSVC
builds we just link to a full static copy even where
-Ddefault_library=shared.
As a side benefit, by using library inclusion rather than including
extra explicit object files, the zstd program shrinks in size slightly
(~4kb).
These need to be explicitly included as we use their private symbols,
but we don't need to recompile them when we can reuse the existing
objects.
Minus 7 compile steps.
The poolTests program already linked to libzstd, and later to
libtestcommon with included libzstd objects. So this was redundant.
Minus 4 compile steps.
Adopt the more standard Usage: formatting style
List short and long options alongside where available
Print lists as a table
Use command style description
Newer less versions appear to have changed how stderr
and stdout are showing error messages. hardcode the
expected behavior to make the tests pass with any less version.
Also set locale to C so that the strings are matching.
- As referenced by Nick Terrelln ~ the ZSTD maintainer in the linux kernel, making zstd_reset_cstream() functionally identical to ZSTD_resetCStream() would be the perfect way to fix the warning without touching any core functions or breaking other parts of the code.
Suggested-by: Nick Terrell <terrelln@fb.com>
Signed-off-by: Cyber Knight <cyberknight755@gmail.com>
Set removeSrcFile back to false when -c or --stdout is used to improve
compatibility with gzip(1) behavior.
gzip(1) is removing the original file on compression unless --stdout or
/-c is used. zstd is defaulting to keep the file unless --rm is used or
when it is called via a gzip symlink, in which it is removing by
default. Specifying -c/--stdout turns this behavior off.
- The previous patch throws the following warning:
../linux/lib/zstd/zstd_compress_module.c: In function ‘zstd_reset_cstream’:
../linux/lib/zstd/zstd_compress_module.c:136:34: error: enum conversion when passing argument 2 of ‘ZSTD_CCtx_reset’ is invalid in C++ [-Werror=c++-compat]
136 | return ZSTD_CCtx_reset(cstream, pledged_src_size);
| ^~~~~~~~~~~~~~~~
In file included from ../linux/include/linux/zstd.h:26,
from ../linux/lib/zstd/zstd_compress_module.c:15:
../linux/include/linux/zstd_lib.h:501:20: note: expected ‘ZSTD_ResetDirective’ {aka ‘enum <anonymous>’} but argument is of type ‘long long unsigned int’
501 | ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
| ^~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
Since we have a choice to either use ZSTD_CCtx_reset or ZSTD_CCtx_setPledgedSrcSize instead of ZSTD_resetCStream, let's switch to ZSTD_CCtx_setPledgedSrcSize to not have any unnecessary warns alongside the kernel build and CI test build.
Signed-off-by: Cyber Knight <cyberknight755@gmail.com>
- This fixes the below warning:
../lib/zstd/zstd_compress_module.c: In function 'zstd_reset_cstream':
../lib/zstd/zstd_compress_module.c:136:9: warning: 'ZSTD_resetCStream' is deprecated [-Wdeprecated-declarations]
136 | return ZSTD_resetCStream(cstream, pledged_src_size);
| ^~~~~~
In file included from ../include/linux/zstd.h:26,
from ../lib/zstd/zstd_compress_module.c:15:
../include/linux/zstd_lib.h:2277:8: note: declared here
2277 | size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize);
| ^~~~~~~~~~~~~~~~~
ZSTD_resetCstream is deprecated and zstd_CCtx_reset is suggested to use hence let's switch to it.
Signed-off-by: Cyber Knight <cyberknight755@gmail.com>
* fix the assertion in readLinesFromFile
When the file is not terminated by endline, readLineFromFile will append
a '\0' for the last line. In this case pos + lineLength == dstCapacity.
* test: don't print very long text garbage
We already have BIT_getLowerBits, so use it. Benefits are 2fold:
1) Somewhat cleaner code
2) We are now using bzhi instructions, when available. Performance
delta is too small for microbenchmarks, but avoiding load still helps
larger applications, by reducing data cache pressure.
The zstd symlinks, notably `zstdcat`, weren't working as expected
because only the `tests/cli-tests/bin/zstd` wrapper was symlinked. We
still invoked `zstd` with the name `zstd`. The fix is to create a
directory of zstd symlinks in `tests/cli-tests/bin/symlinks` for each
name that zstd recognizes. And when `tets/cli-tests/bin/zstd` is
invoked, it selects the correct symlink to call.
See the test `zstd-cli/zstdcat.sh` for an example of how it would work.
* playtests.sh: fix for a bug in macos' /bin/sh that persists temporary env vars when introduced before function calls
* cli-tests/run.py: Do not use existing ZSTD* envvars
The use of --long alters the window size internally in the underlying
library (lib/compress/zstd_compress.c:ZSTD_getCParamsFromCCtxParams),
which changes the memory required for decompression. This means that the
reported requirement from the zstd binary when -vv is specified is
incorrect.
A full fix for this would be to add an API call to be able to retrieve
the required decompression memory from the library, but as a
lighterweight fix we can just take account of the fact we've enabled
long mode and update our verbose output appropriately.
Fixes#2968
credit to oss-fuzz
This issue could happen when using the new Sequence Compression API in Explicit Delimiter Mode
with a too small dstCapacity.
In which case, there was one place where the buffer size wasn't checked.
Currently the build errors out with:
```
ERROR: This script does not work on Python 3.6 The minimum supported Python version is 3.7. Please use https://bootstrap.pypa.io/pip/3.6/get-pip.py instead.
```
While in theory this advice could be followed to get a better pip on
xenial, Meson has now deprecated python 3.6 support too, and the next
(unreleased) version requires python 3.7
There are a couple solutions to this:
- hold the version of pip, allow pip to only install 3.6-compatible
versions of meson (effectively freezing meson going forward)
- install python 3.7 on xenial
- update to a 2-year-old image instead of a 4-year-old one
Option 3 is the simplest.
While trying to raise an exception on failures, it instead raised an
exception for misusing the exception. CalledProcessError is only
supposed to be used when given a return code and a command, and it
prints:
Command '{cmd}' returned non-zero exit status {ret}
Passing an error message string instead, just errored out with:
TypeError: __init__() missing 1 required positional argument
Instead use the subprocess module's base error which does accept string
messages. Everything that used to error out, still errors out, but now
they do so with a slightly prettier console message.
libm is not guaranteed to exist. POSIX requires the math functions to
exist, but doesn't require to have it be a standalone library.
On platforms where libm exists as a standalone library, it will always
be found by meson -- it is shipped with libc.
If it is not found, then we can safely assume the linker will make the
math functions available by default.
See https://mesonbuild.com/howtox.html#add-math-library-lm-portably
Fixes building with bin_tests=true on Windows.
* It couldn't detect that the `fastCoverParams` can't be non-null, since it was just an assertion.
* It thought we were accesing `wksp->dtable` beyond the bounds because we were using it to set the `workSpace` value. Instead, compute the workspace size used in a different way.
discovered by oss-fuzz
It's a bug in the test itself :
ZSTD_compressBound() as an upper bound of the compress size
only works for data compressed "normally".
But in situations where many flushes are forcefully introduced,
this creates many more blocks,
each of which has a potential to increase the size by 3 bytes.
In extreme cases (lots of small incompressible blocks), the expansion can go beyond ZSTD_compressBound().
This situation is similar when using the CompressSequences() API
with Explicit Block Delimiters.
In which case, each explicit block acts like a deliberate flush.
When employed by a fuzzer, it's possible to generate scenarios like the one described above,
with tons of incompressible blocks of small sizes,
thus going beyond ZSTD_compressBound().
fix : when using Explicit Block Delimiters, use a larger bound, to account for this scenario.
It's a bug in the test itself, in exceptional circumstances (no more space for additional sequence).
There should be enough room for all cases to work fine from now on,
and if not, we have an additional `assert()` to catch that situation.
Add cli-tests to `make test`. This adds a `python3` dependency to `make
test`, but not `make check`. We could make this dependency optional by
skipping the tests if `python3` is not present.
`datagen` was printing a `\n` even when it had no other output. Raise
the output level for the final `\n` to the minimum output level used.
This minor bug was caught by the new testing framework.
The option `--auto-threads` should still be accepted and parsed, even if
`ZSTD_MULTITHREAD` is not defined. It doesn't mean anything, but we
should still accept the option. Since we want scripts to be able to work
generically.
This bug was caught by tests I added to the new testing framework.
Passing 0 inputs to `DiB_shuffle()` caused an assertion failure where
it should just return.
A test is added in a later commit, with the initial introduction of the
new testing framework.
Fixes#3007.
credit to oss-fuzz
In rare circumstances, the block-splitter might cut a block at the exact beginning of a repcode.
In which case, since litlength=0, if the repcode expected 1+ literals in front, its signification changes.
This scenario is controlled in ZSTD_seqStore_resolveOffCodes(),
and the repcode is transformed into a raw offset when its new meaning is incorrect.
In more complex scenarios, the previous block might be emitted as uncompressed after all,
thus modifying the expected repcode history.
In the case discovered by oss-fuzz, the first block is emitted as uncompressed,
so the repcode history remains at default values: 1,4,8.
But since the starting repcode is repcode3, and the literal length is == 0,
its meaning is : = repcode1 - 1.
Since repcode1==1, it results in an offset value of 0, which is invalid.
So that's what the `assert()` was verifying : the result of the repcode translation should be a valid offset.
But actually, it doesn't matter, because this result will then be compared to reality,
and since it's an invalid offset, it will necessarily be discarded if incorrect,
then the repcode will be replaced by a raw offset.
So the `assert()` is not useful.
Furthermore, it's incorrect, because it assumes this situation cannot happen, but it does, as described in above scenario.
Knowing the version of zlib/lz4/lzma we're linking against is very
useful for debugging issues with those libraries, so print it out in the
verbosity 4 version output.
Also print this information at the top of `playTests.sh`.
which properly represents the maximum bit size of compressed literals (11) as defined in the specification.
To be preferred from HUF_TABLELOG_DEFAULT which represents the same value but by accident.
Name selected to keep the same convention as existing width definitions,
MLFSELog, LLFSELog and OffFSELog.
In rare cases, the default huffman depth selector is a bit too harsh,
requiring brutal adaptations to the tree,
resulting is some loss of compression ratio.
This new heuristic avoids the worse cases, favoring compression ratio.
As an example, compression of a specific distribution of 771 literals
is now improved to 441 bytes, from 601 bytes before.
-r on empty directory resulted in zstd waiting input from stdin. now zstd exits without error and prints a warning message explaining why no processing happened (no files or directories to process).
* Refactored fileio.c:
- Extracted asyncio code to fileio_asyncio.c/.h
- Moved type definitions to fileio_types.h
- Moved common macro definitions needed by both fileio.c and fileio_asyncio.c to fileio_common.h
* Bugfix - rename fileio_asycio to fileio_asyncio
* Added copyrights & license to new files
* CR fixes
* Async IO decompression:
- Added --[no-]asyncio flag for CLI decompression.
- Replaced dstBuffer in decompression with a pool of write jobs.
- Added an ability to execute write jobs in a separate thread.
- Added an ability to wait (join) on all jobs in a thread pool (queued and running).
We previously triggered release artifact generation on release creation. We
sometimes observed that the action failed to run. I hypothesized that we were
hitting rate limiting or something. I just stumbled across [this documentat-
ion](https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows#release), which says:
> Note: Workflows are not triggered for the `created`, `edited`, or `deleted`
> activity types for draft releases. When you create your release through the
> GitHub browser UI, your release may automatically be saved as a draft.
This must have been what was happening. This commit therefore changes the
trigger to the `published` activity. This should be more reliable.
This does have the unfortunate side effect that artifacts won't be generated
or attached until *after* the release has been published, which is what I was
trying to avoid by using the `created` activity. Oh well.
Append -z cet-report=error to LDFLAGS if -fcf-protection is enabled by
default in compiler to catch the missing Intel CET marker:
compiling multi-threaded dynamic library 1.5.1
/usr/local/bin/ld: obj/conf_f408b4c825de923ffc88f7f21b6884b1/dynamic/huf_decompress_amd64.o: error: missing IBT and SHSTK properties
collect2: error: ld returned 1 exit status
...
LINK obj/conf_dbc0b41e36c44111bb0bb918e093d7c1/zstd
/usr/local/bin/ld: obj/conf_dbc0b41e36c44111bb0bb918e093d7c1/huf_decompress_amd64.o: error: missing IBT and SHSTK properties
collect2: error: ld returned 1 exit status
When testing amalgamating other projects it was found: invalid Unicode errors were tripping Python's text IO, and the header search order appears differs from the shell version.
Intel Control-flow Enforcement Technology (CET):
https://en.wikipedia.org/wiki/Control-flow_integrity#Intel_Control-flow_Enforcement_Technology
requires that on Linux, all linker input files are marked as CET enabled
in .note.gnu.property section. For high-level language source codes,
.note.gnu.property section is added by compiler with the -fcf-protection
option. For assembly sources, include <cet.h> to add .note.gnu.property
section.
directly at ZSTD_storeSeq() interface.
In the process, remove ZSTD_REP_MOVE.
This makes it possible, in future commits,
to update and effectively simplify the naming scheme
to properly label the updated processing pipeline :
offset | repcode => offBase => offCode + offBits
2021-12-29 12:03:36 -08:00
492 changed files with 17654 additions and 7803 deletions
@@ -7,7 +7,7 @@ New versions are being developed in the "dev" branch,
or in their own feature branch.
When they are deemed ready for a release, they are merged into "release".
As a consequences, all contributions must stage first through "dev"
As a consequence, all contributions must stage first through "dev"
or their own feature branch.
## Pull Requests
@@ -134,11 +134,11 @@ It can be useful to look at additional static analyzers once in a while (and we
- Static analyzers are full of false positive. The signal to noise ratio is actually pretty low.
- A good CI policy is "zero-warning tolerance". That means that all issues must be solved, including false positives. This quickly becomes a tedious workload.
- Multiple static analyzers will feature multiple kind of false positives, sometimes applying to the same code but in different ways leading to :
+ torteous code, trying to please multiple constraints, hurting readability and therefore maintenance. Sometimes, such complexity introduce other more subtle bugs, that are just out of scope of the analyzers.
+ tortuous code, trying to please multiple constraints, hurting readability and therefore maintenance. Sometimes, such complexity introduce other more subtle bugs, that are just out of scope of the analyzers.
+ sometimes, these constraints are mutually exclusive : if one try to solve one, the other static analyzer will complain, they can't be both happy at the same time.
- As if that was not enough, the list of false positives change with each version. It's hard enough to follow one static analyzer, but multiple ones with their own update agenda, this quickly becomes a massive velocity reducer.
This is different from running a static analyzer once in a while, looking at the output, and __cherry picking__ a few warnings that seem helpful, either because they detected a genuine risk of bug, or because it helps expressing the code in a way which is more readable or more difficult to misuse. These kind of reports can be useful, and are accepted.
This is different from running a static analyzer once in a while, looking at the output, and __cherry picking__ a few warnings that seem helpful, either because they detected a genuine risk of bug, or because it helps expressing the code in a way which is more readable or more difficult to misuse. These kinds of reports can be useful, and are accepted.
## Continuous Integration
CI tests run every time a pull request (PR) is created or updated. The exact tests
@@ -197,7 +197,7 @@ something subtle merged is extensive benchmarking. You will be doing us a great
take the time to run extensive, long-duration, and potentially cross-(os, platform, process, etc)
benchmarks on your end before submitting a PR. Of course, you will not be able to benchmark
your changes on every single processor and os out there (and neither will we) but do that best
you can:) We've adding some things to think about when benchmarking below in the Benchmarking
you can:) We've added some things to think about when benchmarking below in the Benchmarking
Performance section which might be helpful for you.
3. Optimizing performance for a certain OS, processor vendor, compiler, or network system is a perfectly
legitimate thing to do as long as it does not harm the overall performance health of Zstd.
@@ -273,7 +273,7 @@ for that options you have just provided. If you want to look at the internals of
benchmarking script works, you can check out programs/benchzstd.c
For example: say you have made a change that you believe improves the speed of zstd level 1. The
very first thing you should use to asses whether you actually achieved any sort of improvement
very first thing you should use to assess whether you actually achieved any sort of improvement
is `zstd -b`. You might try to do something like this. Note: you can use the `-i` option to
specify a running time for your benchmark in seconds (default is 3 seconds).
Usually, the longer the running time, the more stable your results will be.
@@ -299,7 +299,7 @@ this method of evaluation will not be sufficient.
### Profiling
There are a number of great profilers out there. We're going to briefly mention how you can
profile your code using `instruments` on mac, `perf` on linux and `visual studio profiler`
on windows.
on Windows.
Say you have an idea for a change that you think will provide some good performance gains
for level 1 compression on Zstd. Typically this means, you have identified a section of
@@ -315,8 +315,8 @@ might be).
Most profilers (including the profilers discussed below) will generate a call graph of
functions for you. Your goal will be to find your function of interest in this call graph
and then inspect the time spent inside of it. You might also want to to look at the
annotated assembly which most profilers will provide you with.
and then inspect the time spent inside of it. You might also want to look at the annotated
assembly which most profilers will provide you with.
#### Instruments
We will once again consider the scenario where you think you've identified a piece of code
@@ -330,7 +330,7 @@ Instruments.
* You will want a benchmark that runs for at least a few seconds (5 seconds will
usually be long enough). This way the profiler will have something to work with
and you will have ample time to attach your profiler to this process:)
* I will just use benchzstd as my bencharmking script for this example:
* I will just use benchzstd as my benchmarmking script for this example:
```
$ zstd -b1 -i5 <my-data> # this will run for 5 seconds
```
@@ -455,7 +455,7 @@ This design requirement is fundamental to preserve the portability of the code b
Any variable that can be `const` (aka. read-only) **must** be `const`.
Any pointer which content will not be modified must be `const`.
This property is then controlled at compiler level.
`const` variables are an important signal to readers that this variable isn’t modified.
`const` variables are an important signal to readers that this variable isn't modified.
Conversely, non-const variables are a signal to readers to watch out for modifications later on in the function.
* If a function must be inlined, mention it explicitly,
using project's own portable macros, such as `FORCE_INLINE_ATTR`,
$(MAKE)testCC=clang MOREFLAGS="-g -fsanitize=memory -fno-omit-frame-pointer -Werror"HAVE_LZMA=0# datagen.c fails this test for no obvious reason
$(MAKE)testCC=clang MOREFLAGS="-g -fsanitize=memory -fno-omit-frame-pointer -Werror$(MOREFLAGS)"HAVE_LZMA=0# datagen.c fails this test for no obvious reason
It's possible to create a compressor-only library but since the decompressor is so small in comparison this doesn't bring much of a gain (but for the curious, simply remove the files in the _decompress_ section at the end of `zstd-in.c`).
parser.add_argument("--xxh64-prefix",default=None,help="Alternate XXH64 function prefix (excluding _) e.g. --xxh64-prefix=xxh64")
parser.add_argument("--rewrite-include",default=[],dest="rewritten_includes",action="append",help="Rewrite an include REGEX=NEW (e.g. '<stddef\\.h>=<linux/types.h>')")
parser.add_argument("--sed",default=[],dest="seds",action="append",help="Apply a sed replacement. Format: `s/REGEX/FORMAT/[g]`. REGEX is a Python regex. FORMAT is a Python format string formatted by the regex dict.")
parser.add_argument("-D","--define",default=[],dest="defs",action="append",help="Pre-define this macro (can be passed multiple times)")
parser.add_argument("-U","--undefine",default=[],dest="undefs",action="append",help="Pre-undefine this macro (can be passed mutliple times)")
parser.add_argument("-U","--undefine",default=[],dest="undefs",action="append",help="Pre-undefine this macro (can be passed multiple times)")
parser.add_argument("-R","--replace",default=[],dest="replaces",action="append",help="Pre-define this macro and replace the first ifndef block with its definition")
parser.add_argument("-E","--exclude",default=[],dest="excludes",action="append",help="Exclude all lines between 'BEGIN <EXCLUDE>' and 'END <EXCLUDE>'")
@@ -31,7 +31,7 @@ If this number is not suitable, during compilation you can define `PZSTD_NUM_THR
## Benchmarks
As a reference, PZstandard and Pigz were compared on an Intel Core i7 @ 3.1 GHz, each using 4 threads, with the [Silesia compression corpus](http://sun.aei.polsl.pl/~sdeor/index.php?page=silesia).
As a reference, PZstandard and Pigz were compared on an Intel Core i7 @ 3.1 GHz, each using 4 threads, with the [Silesia compression corpus](https://sun.aei.polsl.pl//~sdeor/index.php?page=silesia).
Compression Speed vs Ratio with 4 Threads | Decompression Speed with 4 Threads
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Yann Collet