Re: [GIT PULL][PATCH v9 0/3] Update to zstd-1.4.10

From: Oleksandr Natalenko
Date: Wed Mar 31 2021 - 09:31:53 EST


Hello.

On Tue, Mar 30, 2021 at 03:51:09PM -0700, Nick Terrell wrote:
> From: Nick Terrell <terrelln@xxxxxx>
>
> Please pull from
>
> git@xxxxxxxxxx:terrelln/linux.git tags/v9-zstd-1.4.10
>
> to get these changes. Alternatively the patchset is included.
>
> This patchset upgrades the zstd library to the latest upstream release. The
> current zstd version in the kernel is a modified version of upstream zstd-1.3.1.
> At the time it was integrated, zstd wasn't ready to be used in the kernel as-is.
> But, it is now possible to use upstream zstd directly in the kernel.
>
> I have not yet released zstd-1.4.10 upstream. I want the zstd version in the
> kernel to match up with a known upstream release, so we know exactly what code
> is running. Whenever this patchset is ready for merge, I will cut a release at
> the upstream commit that gets merged. This should not be necessary for future
> releases.
>
> The kernel zstd library is automatically generated from upstream zstd. A script
> makes the necessary changes and imports it into the kernel. The changes are:
>
> 1. Replace all libc dependencies with kernel replacements and rewrite includes.
> 2. Remove unncessary portability macros like: #if defined(_MSC_VER).
> 3. Use the kernel xxhash instead of bundling it.
>
> This automation gets tested every commit by upstream's continuous integration.
> When we cut a new zstd release, we will submit a patch to the kernel to update
> the zstd version in the kernel.
>
> I've updated zstd to upstream with one big patch because every commit must build,
> so that precludes partial updates. Since the commit is 100% generated, I hope the
> review burden is lightened. I considered replaying upstream commits, but that is
> not possible because there have been ~3500 upstream commits since the last zstd
> import, and the commits don't all build individually. The bulk update preserves
> bisectablity because bugs can be bisected to the zstd version update. At that
> point the update can be reverted, and we can work with upstream to find and fix
> the bug. After this big switch in how the kernel consumes zstd, future patches
> will be smaller, because they will only have one upstream release worth of
> changes each.
>
> This patchset adds a new kernel-style wrapper around zstd. This wrapper API is
> functionally equivalent to the subset of the current zstd API that is currently
> used. The wrapper API changes to be kernel style so that the symbols don't
> collide with zstd's symbols. The update to zstd-1.4.6 maintains the same API
> and preserves the semantics, so that none of the callers need to be updated.
>
> This patchset comes in 2 parts:
> 1. The first 2 patches prepare for the zstd upgrade. The first patch adds the
> new kernel style API so zstd can be upgraded without modifying any callers.
> The second patch adds an indirection for the lib/decompress_unzstd.c
> including of all decompression source files.
> 2. Import zstd-1.4.10. This patch is completely generated from upstream using
> automated tooling.
>
> I tested every caller of zstd on x86_64. I tested both after the 1.4.10 upgrade
> using the compatibility wrapper, and after the final patch in this series.
>
> I tested kernel and initramfs decompression in i386 and arm.
>
> I ran benchmarks to compare the current zstd in the kernel with zstd-1.4.6.
> I benchmarked on x86_64 using QEMU with KVM enabled on an Intel i9-9900k.
> I found:
> * BtrFS zstd compression at levels 1 and 3 is 5% faster
> * BtrFS zstd decompression+read is 15% faster
> * SquashFS zstd decompression+read is 15% faster
> * F2FS zstd compression+write at level 3 is 8% faster
> * F2FS zstd decompression+read is 20% faster
> * ZRAM decompression+read is 30% faster
> * Kernel zstd decompression is 35% faster
> * Initramfs zstd decompression+build is 5% faster
>
> The latest zstd also offers bug fixes. For example the problem with large kernel
> decompression has been fixed upstream for over 2 years
> https://lkml.org/lkml/2020/9/29/27.
>
> Please let me know if there is anything that I can do to ease the way for these
> patches. I think it is important because it gets large performance improvements,
> contains bug fixes, and is switching to a more maintainable model of consuming
> upstream zstd directly, making it easy to keep up to date.
>
> Best,
> Nick Terrell
>
> v1 -> v2:
> * Successfully tested F2FS with help from Chao Yu to fix my test.
> * (1/9) Fix ZSTD_initCStream() wrapper to handle pledged_src_size=0 means unknown.
> This fixes F2FS with the zstd-1.4.6 compatibility wrapper, exposed by the test.
>
> v2 -> v3:
> * (3/9) Silence warnings by Kernel Test Robot:
> https://github.com/facebook/zstd/pull/2324
> Stack size warnings remain, but these aren't new, and the functions it warns on
> are either unused or not in the maximum stack path. This patchset reduces zstd
> compression stack usage by 1 KB overall. I've gotten the low hanging fruit, and
> more stack reduction would require significant changes that have the potential
> to introduce new bugs. However, I do hope to continue to reduce zstd stack
> usage in future versions.
>
> v3 -> v4:
> * (3/9) Fix errors and warnings reported by Kernel Test Robot:
> https://github.com/facebook/zstd/pull/2326
> - Replace mem.h with a custom kernel implementation that matches the current
> lib/zstd/mem.h in the kernel. This avoids calls to __builtin_bswap*() which
> don't work on certain architectures, as exposed by the Kernel Test Robot.
> - Remove ASAN/MSAN (un)poisoning code which doesn't work in the kernel, as
> exposed by the Kernel Test Robot.
> - I've fixed all of the valid cppcheck warnings reported, but there were many
> false positives, where cppcheck was incorrectly analyzing the situation,
> which I did not fix. I don't believe it is reasonable to expect that upstream
> zstd silences all the static analyzer false positives. Upstream zstd uses
> clang scan-build for its static analysis. We find that supporting multiple
> static analysis tools multiplies the burden of silencing false positives,
> without providing enough marginal value over running a single static analysis
> tool.
>
> v4 -> v5:
> * Rebase onto v5.10-rc2
> * (6/9) Merge with other F2FS changes (no functional change in patch).
>
> v5 -> v6:
> * Rebase onto v5.10-rc6.
> * Switch to using a kernel style wrapper API as suggested by Cristoph.
>
> v6 -> v7:
> * Expose the upstream library header as `include/linux/zstd_lib.h`.
> Instead of creating new structs mirroring the upstream zstd structs
> use upstream's structs directly with a typedef to get a kernel style name.
> This removes the memcpy cruft.
> * (1/3) Undo ZSTD_WINDOWLOG_MAX and handle_zstd_error changes.
> * (3/3) Expose zstd_errors.h as `include/linux/zstd_errors.h` because it
> is needed by the kernel wrapper API.
>
> v7 -> v8:
> * (1/3) Fix typo in EXPORT_SYMBOL().
> * (1/3) Fix typo in zstd.h comments.
> * (3/3) Update to latest zstd release: 1.4.6 -> 1.4.10
> This includes ~1KB of stack space reductions.
>
> v8 -> v9:
> * (1/3) Rebase onto v5.12-rc5
> * (1/3) Add zstd_min_clevel() & zstd_max_clevel() and use in f2fs.
> Thanks to Oleksandr Natalenko for spotting it!
> * (1/3) Move lib/decompress_unzstd.c usage of ZSTD_getErrorCode()
> to zstd_get_error_code().
> * (1/3) Update modified zstd headers to yearless copyright.
> * (2/3) Add copyright/license header to decompress_sources.h for consistency.
> * (3/3) Update to yearless copyright for all zstd files. Thanks to
> Mike Dolan for spotting it!
>
> Nick Terrell (3):
> lib: zstd: Add kernel-specific API
> lib: zstd: Add decompress_sources.h for decompress_unzstd
> lib: zstd: Upgrade to latest upstream zstd version 1.4.10
>
> crypto/zstd.c | 28 +-
> fs/btrfs/zstd.c | 68 +-
> fs/f2fs/compress.c | 56 +-
> fs/f2fs/super.c | 2 +-
> fs/pstore/platform.c | 2 +-
> fs/squashfs/zstd_wrapper.c | 16 +-
> include/linux/zstd.h | 1252 +---
> include/linux/zstd_errors.h | 77 +
> include/linux/zstd_lib.h | 2432 ++++++++
> lib/decompress_unzstd.c | 48 +-
> lib/zstd/Makefile | 44 +-
> lib/zstd/bitstream.h | 380 --
> lib/zstd/common/bitstream.h | 437 ++
> lib/zstd/common/compiler.h | 151 +
> lib/zstd/common/cpu.h | 194 +
> lib/zstd/common/debug.c | 24 +
> lib/zstd/common/debug.h | 101 +
> lib/zstd/common/entropy_common.c | 357 ++
> lib/zstd/common/error_private.c | 56 +
> lib/zstd/common/error_private.h | 66 +
> lib/zstd/common/fse.h | 710 +++
> lib/zstd/common/fse_decompress.c | 390 ++
> lib/zstd/common/huf.h | 356 ++
> lib/zstd/common/mem.h | 259 +
> lib/zstd/common/zstd_common.c | 83 +
> lib/zstd/common/zstd_deps.h | 125 +
> lib/zstd/common/zstd_internal.h | 450 ++
> lib/zstd/compress.c | 3485 -----------
> lib/zstd/compress/fse_compress.c | 625 ++
> lib/zstd/compress/hist.c | 165 +
> lib/zstd/compress/hist.h | 75 +
> lib/zstd/compress/huf_compress.c | 902 +++
> lib/zstd/compress/zstd_compress.c | 5105 +++++++++++++++++
> lib/zstd/compress/zstd_compress_internal.h | 1188 ++++
> lib/zstd/compress/zstd_compress_literals.c | 158 +
> lib/zstd/compress/zstd_compress_literals.h | 29 +
> lib/zstd/compress/zstd_compress_sequences.c | 439 ++
> lib/zstd/compress/zstd_compress_sequences.h | 54 +
> lib/zstd/compress/zstd_compress_superblock.c | 850 +++
> lib/zstd/compress/zstd_compress_superblock.h | 32 +
> lib/zstd/compress/zstd_cwksp.h | 482 ++
> lib/zstd/compress/zstd_double_fast.c | 521 ++
> lib/zstd/compress/zstd_double_fast.h | 32 +
> lib/zstd/compress/zstd_fast.c | 496 ++
> lib/zstd/compress/zstd_fast.h | 31 +
> lib/zstd/compress/zstd_lazy.c | 1412 +++++
> lib/zstd/compress/zstd_lazy.h | 81 +
> lib/zstd/compress/zstd_ldm.c | 686 +++
> lib/zstd/compress/zstd_ldm.h | 110 +
> lib/zstd/compress/zstd_ldm_geartab.h | 103 +
> lib/zstd/compress/zstd_opt.c | 1345 +++++
> lib/zstd/compress/zstd_opt.h | 50 +
> lib/zstd/decompress.c | 2531 --------
> lib/zstd/decompress/huf_decompress.c | 1206 ++++
> lib/zstd/decompress/zstd_ddict.c | 241 +
> lib/zstd/decompress/zstd_ddict.h | 44 +
> lib/zstd/decompress/zstd_decompress.c | 2075 +++++++
> lib/zstd/decompress/zstd_decompress_block.c | 1540 +++++
> lib/zstd/decompress/zstd_decompress_block.h | 62 +
> .../decompress/zstd_decompress_internal.h | 202 +
> lib/zstd/decompress_sources.h | 28 +
> lib/zstd/entropy_common.c | 243 -
> lib/zstd/error_private.h | 53 -
> lib/zstd/fse.h | 575 --
> lib/zstd/fse_compress.c | 795 ---
> lib/zstd/fse_decompress.c | 325 --
> lib/zstd/huf.h | 212 -
> lib/zstd/huf_compress.c | 773 ---
> lib/zstd/huf_decompress.c | 960 ----
> lib/zstd/mem.h | 151 -
> lib/zstd/zstd_common.c | 75 -
> lib/zstd/zstd_compress_module.c | 124 +
> lib/zstd/zstd_decompress_module.c | 105 +
> lib/zstd/zstd_internal.h | 273 -
> lib/zstd/zstd_opt.h | 1014 ----
> 75 files changed, 27287 insertions(+), 12940 deletions(-)
> create mode 100644 include/linux/zstd_errors.h
> create mode 100644 include/linux/zstd_lib.h
> delete mode 100644 lib/zstd/bitstream.h
> create mode 100644 lib/zstd/common/bitstream.h
> create mode 100644 lib/zstd/common/compiler.h
> create mode 100644 lib/zstd/common/cpu.h
> create mode 100644 lib/zstd/common/debug.c
> create mode 100644 lib/zstd/common/debug.h
> create mode 100644 lib/zstd/common/entropy_common.c
> create mode 100644 lib/zstd/common/error_private.c
> create mode 100644 lib/zstd/common/error_private.h
> create mode 100644 lib/zstd/common/fse.h
> create mode 100644 lib/zstd/common/fse_decompress.c
> create mode 100644 lib/zstd/common/huf.h
> create mode 100644 lib/zstd/common/mem.h
> create mode 100644 lib/zstd/common/zstd_common.c
> create mode 100644 lib/zstd/common/zstd_deps.h
> create mode 100644 lib/zstd/common/zstd_internal.h
> delete mode 100644 lib/zstd/compress.c
> create mode 100644 lib/zstd/compress/fse_compress.c
> create mode 100644 lib/zstd/compress/hist.c
> create mode 100644 lib/zstd/compress/hist.h
> create mode 100644 lib/zstd/compress/huf_compress.c
> create mode 100644 lib/zstd/compress/zstd_compress.c
> create mode 100644 lib/zstd/compress/zstd_compress_internal.h
> create mode 100644 lib/zstd/compress/zstd_compress_literals.c
> create mode 100644 lib/zstd/compress/zstd_compress_literals.h
> create mode 100644 lib/zstd/compress/zstd_compress_sequences.c
> create mode 100644 lib/zstd/compress/zstd_compress_sequences.h
> create mode 100644 lib/zstd/compress/zstd_compress_superblock.c
> create mode 100644 lib/zstd/compress/zstd_compress_superblock.h
> create mode 100644 lib/zstd/compress/zstd_cwksp.h
> create mode 100644 lib/zstd/compress/zstd_double_fast.c
> create mode 100644 lib/zstd/compress/zstd_double_fast.h
> create mode 100644 lib/zstd/compress/zstd_fast.c
> create mode 100644 lib/zstd/compress/zstd_fast.h
> create mode 100644 lib/zstd/compress/zstd_lazy.c
> create mode 100644 lib/zstd/compress/zstd_lazy.h
> create mode 100644 lib/zstd/compress/zstd_ldm.c
> create mode 100644 lib/zstd/compress/zstd_ldm.h
> create mode 100644 lib/zstd/compress/zstd_ldm_geartab.h
> create mode 100644 lib/zstd/compress/zstd_opt.c
> create mode 100644 lib/zstd/compress/zstd_opt.h
> delete mode 100644 lib/zstd/decompress.c
> create mode 100644 lib/zstd/decompress/huf_decompress.c
> create mode 100644 lib/zstd/decompress/zstd_ddict.c
> create mode 100644 lib/zstd/decompress/zstd_ddict.h
> create mode 100644 lib/zstd/decompress/zstd_decompress.c
> create mode 100644 lib/zstd/decompress/zstd_decompress_block.c
> create mode 100644 lib/zstd/decompress/zstd_decompress_block.h
> create mode 100644 lib/zstd/decompress/zstd_decompress_internal.h
> create mode 100644 lib/zstd/decompress_sources.h
> delete mode 100644 lib/zstd/entropy_common.c
> delete mode 100644 lib/zstd/error_private.h
> delete mode 100644 lib/zstd/fse.h
> delete mode 100644 lib/zstd/fse_compress.c
> delete mode 100644 lib/zstd/fse_decompress.c
> delete mode 100644 lib/zstd/huf.h
> delete mode 100644 lib/zstd/huf_compress.c
> delete mode 100644 lib/zstd/huf_decompress.c
> delete mode 100644 lib/zstd/mem.h
> delete mode 100644 lib/zstd/zstd_common.c
> create mode 100644 lib/zstd/zstd_compress_module.c
> create mode 100644 lib/zstd/zstd_decompress_module.c
> delete mode 100644 lib/zstd/zstd_internal.h
> delete mode 100644 lib/zstd/zstd_opt.h
>
> --
> 2.31.0
>

Running this series for quite some time already (v9 as well as previous
submissions) on different machines without any issues.

Hence, for the series:

Tested-by: Oleksandr Natalenko <oleksandr@xxxxxxxxxxxxxx>

Thanks!

--
Oleksandr Natalenko (post-factum)