* Flatten ZSTD_row_getMatchMask
* Remove the SIMD abstraction layer.
* Add big endian support.
* Align `hashTags` within `tagRow` to a 16-byte boundary.
* Switch SSE2 to use aligned reads.
* Optimize scalar path using SWAR.
* Optimize neon path for `n == 32`
* Work around minor clang issue for NEON (https://bugs.llvm.org/show_bug.cgi?id=49577)
* replace memcpy with MEM_readST
* silence alignment warnings
* fix neon casts
* Update zstd_lazy.c
* unify simd preprocessor detection (#3)
* remove duplicate asserts
* tweak rotates
* improve endian detection
* add cast
there is a fun little catch-22 with gcc: result from pmovmskb has to be cast to uint32_t to avoid a zero-extension
but must be uint16_t to get gcc to generate a rotate instruction..
* more casts
* fix casts
better work-around for the (bogus) warning: unary minus on unsigned
Call `ZSTD_enforceMaxDist()` before each block with the beginning of the
block. This ensures that `lowLimit` is updated to `dictLimit` whenever
the ext-dict is out of range, so we can use prefix mode for speed.
This can cause non-determinism because prefix mode and ext-dict mode
match finders can return different results. It can also hurt speed
because ext-dict match finders are slower.
The scenario is:
1. Compress large data with a dictionary.
2. The dictionary goes out of bounds, so we invalidate it.
3. However, we still have `lowLimit < dictLimit`, since it is
never updated.
4. We will call the ext-dict match finder instead of the prefix one.
The repcode checks disallowed repcodes that are equal to `windowLow`.
This is slightly inefficient, but isn't a problem on its own. Together
with the next commit, it cause non-determinism.
`ZSTD_insertBt1()` has a speed optimization that skips the prefix of
very long matches.
https://github.com/facebook/zstd/blob/40def70387f99b239f3f566ba399275a8fd44cde/lib/compress/zstd_opt.c#L476
This optimization is based off the length longest match found. However,
when indices are reset, we only ensure that we can reference the whole
window starting from `ip`. If the previous block ended with a long match
then `nextToUpdate` could be much less than `ip`. It might be far enough
back that `nextToUpdate < maxDist`, so it doesn't have a full window of
data to reference. This can cause non-determinism bugs, because we may
find a match that is beyond `ip - maxDist`, and may sometimes be
un-referencable, and that match triggers the speed optimization.
The fix is to base the `windowLow` off of the `target` of
`ZSTD_updateTree_internal()`, because anything below that value will be
obsolete by the time `ZSTD_updateTree_internal()` completes.
and restored limit to 256 when in 64-bit mode
(it was reduced to 200 to give more room for 32-bit).
This should fix test instability issues
using lot of threads in 32-bit environments.
Linearly back off the frequency of overflow correction based on the
number of times the `ZSTD_window_t` has been overflow corrected. This
will still allow the fuzzer to quickly find overflow correction bugs,
while also keeping good speed for larger inputs.
Additionally, the `nbOverflowCorrections` variable can be useful for
debugging coredumps, since we can inspect the `ZSTD_CCtx` to see if
overflow correction has happened yet.
I've verified this fixes the timeouts in OSS-Fuzz (176 seconds -> 6
seconds). I've also verified that fuzzers and `fuzzer` and `zstreamtest`
still catch the row-hash overflow correction bug.
This flag forces zstd to always load the prefix in ext-dict mode, even
if it happens to be contiguous, to force determinism. It also applies to
dictionaries that are re-processed.
A determinism test case is also added, which fails without
`ZSTD_c_deterministicRefPrefix` and passes with it set.
Question: Should this be the default behavior? It isn't in this PR.
* Take `params` by const reference in `ZSTD_resetCCtx_internal()`.
* Add `simpleApiParams` to the CCtx and use them in the simple API
functions, instead of creating those parameters on the stack.
I think this is a good direction to move in, because we shouldn't need
to worry about adding parameters to `ZSTD_CCtx_params`, since it should
always be on the heap (unless they become absoultely gigantic).
Some `ZSTD_CCtx_params` are still on the stack in the CDict functions,
but I've left them for now, because it was a little more complex, and we
don't use those functions in stack-constrained currently.
Dictionaries larger than `ZSTD_CHUNKSIZE_MAX` used to have to be loaded
in multiple segments. Instead, when we detect large dictionaries, ensure
that we reset the context's indicies. Then, for dictionaries larger than
`ZSTD_CURRENT_MAX - 1`, only load the suffix of the dictionary. Finally,
enable DDS for large dictionaries, since we no longer load in multiple
segments.
This simplifes the dictionary loading code, and reduces opportunities
for non-determinism to slip in.
previous lower limit was 1 MB.
Note : by default, the lowest job size is 2 MB, achieved at level 1.
Even lower job sizes can be achieved by manipulating this value directly,
or manually modifying window sizes to lower amounts.
Updated unit test to ensure that this new limit works fine
(test would fail with previous 1 MB limit).
LDM does especially poorly on repetitive data when that data's hash happens
to have `(hash & stopMask) == 0`. Either because the `stopMask == 0` or
random chance. Optimize this case by skipping over repetitive patterns.
The detection is very simplistic, but should catch most of the offending
cases.
```
head -c 1G /dev/zero | perf stat -- ./zstd -1 -o /dev/null -v --zstd=ldmHashRateLog=1 --long
21.187881087 seconds time elapsed
head -c 1G /dev/zero | perf stat -- ./zstd -1 -o /dev/null -v --zstd=ldmHashRateLog=1 --long
1.149707921 seconds time elapsed
```
* Fix overflow correction when `windowLog < cycleLog`. Previously, we
got the correction wrong in this case, and our chain tables and binary
trees would be corrupted. Now, we work as long as `maxDist` is a power
of two, by adding `MAX(maxDist, cycleSize)` to our indices.
* When `ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY` is defined to non-zero
run overflow correction as frequently as allowed without impacting
compression ratio.
* Enable `ZSTD_WINDOW_OVERFLOW_CORRECT_FREQUENTLY` in `fuzzer` and
`zstreamtest` as well as all the OSS-Fuzz fuzzers. This has a 5-10%
speed penalty at most, which seems reasonable.
Instead of providing a default no-op implementation, check the symbols
for `NULL` before accessing them. Providing a default implementation
doesn't reliably work with dynamic linking. Depending on link order the
default implementations may not be overridden. By skipping the default
implementation, all link order issues are resolved. If the symbols
aren't provided the weak function will be `NULL`.
