leveldb/table/format.cc
leveldb Team 1d6e8d64ee Add support for Zstd-based compression in LevelDB.
This change implements support for Zstd-based compression in LevelDB. Building
up from the Snappy compression (which has been supported since inception), this
change adds Zstd as an alternate compression algorithm.

We are implementing this to provide alternative options for users who might
have different performance and efficiency requirements. For instance, the
Zstandard website (https://facebook.github.io/zstd/) claims that the Zstd
algorithm can achieve around 30% higher compression ratios than Snappy, with
relatively smaller (~10%) slowdowns in de/compression speeds.

Benchmarking results:

$ blaze-bin/third_party/leveldb/db_bench
LevelDB:    version 1.23
Date:       Thu Feb  2 18:50:06 2023
CPU:        56 * Intel(R) Xeon(R) CPU E5-2690 v4 @ 2.60GHz
CPUCache:   35840 KB
Keys:       16 bytes each
Values:     100 bytes each (50 bytes after compression)
Entries:    1000000
RawSize:    110.6 MB (estimated)
FileSize:   62.9 MB (estimated)
------------------------------------------------
fillseq      :       2.613 micros/op;   42.3 MB/s
fillsync     :    3924.432 micros/op;    0.0 MB/s (1000 ops)
fillrandom   :       3.609 micros/op;   30.7 MB/s
overwrite    :       4.508 micros/op;   24.5 MB/s
readrandom   :       6.136 micros/op; (864322 of 1000000 found)
readrandom   :       5.446 micros/op; (864083 of 1000000 found)
readseq      :       0.180 micros/op;  613.3 MB/s
readreverse  :       0.321 micros/op;  344.7 MB/s
compact      :  827043.000 micros/op;
readrandom   :       4.603 micros/op; (864105 of 1000000 found)
readseq      :       0.169 micros/op;  656.3 MB/s
readreverse  :       0.315 micros/op;  350.8 MB/s
fill100K     :     854.009 micros/op;  111.7 MB/s (1000 ops)
crc32c       :       1.227 micros/op; 3184.0 MB/s (4K per op)
snappycomp   :       3.610 micros/op; 1081.9 MB/s (output: 55.2%)
snappyuncomp :       0.691 micros/op; 5656.3 MB/s
zstdcomp     :      15.731 micros/op;  248.3 MB/s (output: 44.1%)
zstduncomp   :       4.218 micros/op;  926.2 MB/s
PiperOrigin-RevId: 509957778
2023-03-28 16:49:13 -07:00

161 lines
5.1 KiB
C++

// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "table/format.h"
#include "leveldb/env.h"
#include "leveldb/options.h"
#include "port/port.h"
#include "table/block.h"
#include "util/coding.h"
#include "util/crc32c.h"
namespace leveldb {
void BlockHandle::EncodeTo(std::string* dst) const {
// Sanity check that all fields have been set
assert(offset_ != ~static_cast<uint64_t>(0));
assert(size_ != ~static_cast<uint64_t>(0));
PutVarint64(dst, offset_);
PutVarint64(dst, size_);
}
Status BlockHandle::DecodeFrom(Slice* input) {
if (GetVarint64(input, &offset_) && GetVarint64(input, &size_)) {
return Status::OK();
} else {
return Status::Corruption("bad block handle");
}
}
void Footer::EncodeTo(std::string* dst) const {
const size_t original_size = dst->size();
metaindex_handle_.EncodeTo(dst);
index_handle_.EncodeTo(dst);
dst->resize(2 * BlockHandle::kMaxEncodedLength); // Padding
PutFixed32(dst, static_cast<uint32_t>(kTableMagicNumber & 0xffffffffu));
PutFixed32(dst, static_cast<uint32_t>(kTableMagicNumber >> 32));
assert(dst->size() == original_size + kEncodedLength);
(void)original_size; // Disable unused variable warning.
}
Status Footer::DecodeFrom(Slice* input) {
const char* magic_ptr = input->data() + kEncodedLength - 8;
const uint32_t magic_lo = DecodeFixed32(magic_ptr);
const uint32_t magic_hi = DecodeFixed32(magic_ptr + 4);
const uint64_t magic = ((static_cast<uint64_t>(magic_hi) << 32) |
(static_cast<uint64_t>(magic_lo)));
if (magic != kTableMagicNumber) {
return Status::Corruption("not an sstable (bad magic number)");
}
Status result = metaindex_handle_.DecodeFrom(input);
if (result.ok()) {
result = index_handle_.DecodeFrom(input);
}
if (result.ok()) {
// We skip over any leftover data (just padding for now) in "input"
const char* end = magic_ptr + 8;
*input = Slice(end, input->data() + input->size() - end);
}
return result;
}
Status ReadBlock(RandomAccessFile* file, const ReadOptions& options,
const BlockHandle& handle, BlockContents* result) {
result->data = Slice();
result->cachable = false;
result->heap_allocated = false;
// Read the block contents as well as the type/crc footer.
// See table_builder.cc for the code that built this structure.
size_t n = static_cast<size_t>(handle.size());
char* buf = new char[n + kBlockTrailerSize];
Slice contents;
Status s = file->Read(handle.offset(), n + kBlockTrailerSize, &contents, buf);
if (!s.ok()) {
delete[] buf;
return s;
}
if (contents.size() != n + kBlockTrailerSize) {
delete[] buf;
return Status::Corruption("truncated block read");
}
// Check the crc of the type and the block contents
const char* data = contents.data(); // Pointer to where Read put the data
if (options.verify_checksums) {
const uint32_t crc = crc32c::Unmask(DecodeFixed32(data + n + 1));
const uint32_t actual = crc32c::Value(data, n + 1);
if (actual != crc) {
delete[] buf;
s = Status::Corruption("block checksum mismatch");
return s;
}
}
switch (data[n]) {
case kNoCompression:
if (data != buf) {
// File implementation gave us pointer to some other data.
// Use it directly under the assumption that it will be live
// while the file is open.
delete[] buf;
result->data = Slice(data, n);
result->heap_allocated = false;
result->cachable = false; // Do not double-cache
} else {
result->data = Slice(buf, n);
result->heap_allocated = true;
result->cachable = true;
}
// Ok
break;
case kSnappyCompression: {
size_t ulength = 0;
if (!port::Snappy_GetUncompressedLength(data, n, &ulength)) {
delete[] buf;
return Status::Corruption("corrupted snappy compressed block length");
}
char* ubuf = new char[ulength];
if (!port::Snappy_Uncompress(data, n, ubuf)) {
delete[] buf;
delete[] ubuf;
return Status::Corruption("corrupted snappy compressed block contents");
}
delete[] buf;
result->data = Slice(ubuf, ulength);
result->heap_allocated = true;
result->cachable = true;
break;
}
case kZstdCompression: {
size_t ulength = 0;
if (!port::Zstd_GetUncompressedLength(data, n, &ulength)) {
delete[] buf;
return Status::Corruption("corrupted zstd compressed block length");
}
char* ubuf = new char[ulength];
if (!port::Zstd_Uncompress(data, n, ubuf)) {
delete[] buf;
delete[] ubuf;
return Status::Corruption("corrupted zstd compressed block contents");
}
delete[] buf;
result->data = Slice(ubuf, ulength);
result->heap_allocated = true;
result->cachable = true;
break;
}
default:
delete[] buf;
return Status::Corruption("bad block type");
}
return Status::OK();
}
} // namespace leveldb