Port lib/legacy/zstd_v01.c (the frozen zstd v0.1 decoder) to
rust/src/legacy/zstd_v01.rs as the first legacy-format port on the new
scaffolding, and reduce the C file to a declaration-only shim that
keeps its header includes for configuration and platform preprocessor
behavior.
Frozen-decoder policy: zstd_v01.c embeds its own v0.1-era FSE and
Huff0 snapshot, distinct from every other release. The Rust port is a
line-by-line translation with the same table layouts (FSE_DTable as a
u32 header word plus packed newState/symbol/nbBits entries, the Huff0
u16 DTable with byte/nbBits pairs), the same arithmetic including
wrap-around and pointer-comparison quirks (e.g. the offset-vs-base
address check in ZSTD_execSequence), the same internal FSE error space
(size_t)-1..-7, and the same public ZSTD error codes. It reuses no
modern Rust entropy module; its only crate dependency is `errors`,
matching the C file's error_private.h include. The 32-bit-only reload
points are kept as compile-time conditions on usize::BITS.
Symbol takeover boundary: all nine ZSTDv01_* entry points from
zstd_v01.h now come from Rust as context-free #[no_mangle] extern "C"
functions (isError, decompress, decompressDCtx,
findFrameSizeInfoLegacy, createDCtx, freeDCtx, resetDCtx,
nextSrcSizeToDecompress, decompressContinue). zstd_legacy.h only uses
the first four for v0.1; streaming for v0.1-v0.3 intentionally returns
version_unsupported there, unchanged. The ZSTDv01_Dctx struct
definition moves entirely into Rust: C code only ever holds an opaque
pointer (zstd_v01.h forward-declares the type), and the context is
malloc/free-allocated exactly like the C version so create/free may
pair across the language boundary.
Byte-identity verification against the pristine pre-migration C build
(f8745da6, pure C, ZSTD_LEGACY_SUPPORT=1):
- Real v0.1 frames were generated by building the v0.1.0 git tag and
compressing text, random, and 426 KB multi-block inputs. A one-shot
ZSTD_decompress harness linked once against the pristine C libzstd.a
and once against the Rust-backed libzstd.a produced bit-identical
outputs for all frames.
- A direct ZSTDv01_* probe (one-shot decode, dst-too-small, truncated
input, bad magic, findFrameSizeInfoLegacy, and the streaming
continue loop) printed identical results, including exact error
codes (-70 dstSize_tooSmall, -72 srcSize_wrong, -10 prefix_unknown)
and identical dBound values.
- zstd -l -v on v0.1 files matches the pristine binary; CLI streaming
decode of v0.1 fails with the same "Version not supported" in both,
by design of zstd_legacy.h.
Unit tests embed three v0.1.0-generated fixtures (entropy-coded,
raw-block, and four-block frames) plus the truncation, bad-magic,
small-destination, and streaming-API cases, all asserting the exact C
error codes above. Note that `make -C tests test-legacy` only covers
v0.4+ frames, so the embedded fixtures and the harness comparison are
the actual v0.1 coverage.
Test plan:
- cd rust && cargo fmt --check && cargo clippy --all-targets
--features legacy-v01 -- -D warnings && cargo test --all-targets
--features legacy-v01 (127 tests, 9 for v0.1)
- cargo clippy/test --no-default-features --features
decompression,legacy-v01 (module builds standalone)
- make -C tests fuzzer && ./tests/fuzzer -i1 --no-big-tests, also with
ZSTD_LEGACY_SUPPORT=1 (mixed Rust v0.1 + C v0.2-0.7 link)
- make -C tests test-rust-lib-smoke && make -C tests test-legacy
- make -C programs zstd (default and ZSTD_LEGACY_SUPPORT=1); nm shows
the nine ZSTDv01_* symbols provided by Rust at level 1
- make -C lib libzstd.a ZSTD_LEGACY_SUPPORT=0 (no legacy symbols) and
meson -Dlegacy_level=1 shared library exporting all nine
* Remove all pointer-overflow suppressions from our UBSAN builds/tests.
* Add `ZSTD_ALLOW_POINTER_OVERFLOW_ATTR` macro to suppress
pointer-overflow at a per-function level. This is a superior approach
because it also applies to users who build zstd with UBSAN.
* Add `ZSTD_wrappedPtr{Diff,Add,Sub}()` that use these suppressions.
The end goal is to only tag these functions with
`ZSTD_ALLOW_POINTER_OVERFLOW`. But we can start by annoting functions
that rely on pointer overflow, and gradually transition to using
these.
* Add `ZSTD_maybeNullPtrAdd()` to simplify pointer addition when the
pointer may be `NULL`.
* Fix all the fuzzer issues that came up. I'm sure there will be a lot
more, but these are the ones that came up within a few minutes of
running the fuzzers, and while running GitHub CI.
```
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
```
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.
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
* Switch to yearless copyright per FB policy
* Fix up SPDX-License-Identifier lines in `contrib/linux-kernel` sources
* Add zstd copyright/license header to the `contrib/linux-kernel` sources
* Update the `tests/test-license.py` to check for yearless copyright
* Improvements to `tests/test-license.py`
* Check `contrib/linux-kernel` in `tests/test-license.py`
When the output buffer is `NULL` with size 0, but the frame content size
is non-zero, we will write to the NULL pointer because our bounds check
underflowed.
This was exposed by a recent PR that allowed an empty frame into the
single-pass shortcut in streaming mode.
* Fix the bug.
* Fix another NULL dereference in zstd-v1.
* Overflow checks in 32-bit mode.
* Add a dedicated test.
* Expose the bug in the dedicated simple_decompress fuzzer.
* Switch all mallocs in fuzzers to return NULL for size=0.
* Fix a new timeout in a fuzzer.
Neither clang nor gcc show a decompression speed regression on x86-64.
On x86-32 clang is slightly positive and gcc loses 2.5% of speed.
Credit to OSS-Fuzz.
* All copyright lines now have -2020 instead of -present
* All copyright lines include "Facebook, Inc"
* All licenses are now standardized
The copyright in `threading.{h,c}` is not changed because it comes from
zstdmt.
The copyright and license of `divsufsort.{h,c}` is not changed.
* Version <= 0.5 could read beyond the end of `dumps`, which points into
the input buffer.
* Check the validity of `dumps` before using it, if it is out of bounds
return garbage values. There is no return code for this function.
* Introduce `MEM_readLE24()` for simplicity, since I don't want to trust
that there is an extra byte after `dumps`.
- Add ZSTD_findDecompressedSize
- Traverses multiple frames to find total output size
- Add ZSTD_getFrameContentSize
- Gets the decompressed size of a single frame by reading header
- Deprecate ZSTD_getDecompressedSize
execSequence relied on pointer overflow to handle cases where
`sequence.matchLength < 8`. Instead of passing an `size_t` to
wildcopy, pass a `ptrdiff_t`.
Allows an adversary to write up to 3 bytes beyond the end of the buffer.
Occurs if the match overlaps the `extDict` and `currentPrefix`, and the
match length in the `currentPrefix` is less than `MINMATCH`, and
`op-(16-MINMATCH) >= oMatchEnd > op-16`.
If `weightTotal == 0`, then `BIT_highbit32(weightTotal)` is
undefined behavior in the case that it calls `__builtin_clz()`.
If `tableLog == HUF_TABLELOG_ABSOLUTEMAX` then we will access one
byte beyond the end of the buffer.
Caused by two things:
1. Not checking that `ip` is in range except for the first byte.
2. `ZSTDv0{5,6}_decodeLiteralsBlock()` could return a value larger than `srcSize`.
ddidderr
Yann Collet