rocksdb/file/random_access_file_reader.h
Cheng Chang 4fc216649d Support direct IO in RandomAccessFileReader::MultiRead (#6446)
Summary:
By supporting direct IO in RandomAccessFileReader::MultiRead, the benefits of parallel IO (IO uring) and direct IO can be combined.

In direct IO mode, read requests are aligned and merged together before being issued to RandomAccessFile::MultiRead, so blocks in the original requests might share the same underlying buffer, the shared buffers are returned in `aligned_bufs`, which is a new parameter of the `MultiRead` API.

For example, suppose alignment requirement for direct IO is 4KB, one request is (offset: 1KB, len: 1KB), another request is (offset: 3KB, len: 1KB), then since they all belong to page (offset: 0, len: 4KB), `MultiRead` only reads the page with direct IO into a buffer on heap, and returns 2 Slices referencing regions in that same buffer. See `random_access_file_reader_test.cc` for more examples.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6446

Test Plan: Added a new test `random_access_file_reader_test.cc`.

Reviewed By: anand1976

Differential Revision: D20097518

Pulled By: cheng-chang

fbshipit-source-id: ca48a8faf9c3af146465c102ef6b266a363e78d1
2020-03-20 16:33:26 -07:00

139 lines
4.9 KiB
C++

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// 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.
#pragma once
#include <atomic>
#include <sstream>
#include <string>
#include "port/port.h"
#include "rocksdb/env.h"
#include "rocksdb/file_system.h"
#include "rocksdb/listener.h"
#include "rocksdb/rate_limiter.h"
#include "util/aligned_buffer.h"
namespace ROCKSDB_NAMESPACE {
class Statistics;
class HistogramImpl;
using AlignedBuf = std::unique_ptr<const char[]>;
// RandomAccessFileReader is a wrapper on top of Env::RnadomAccessFile. It is
// responsible for:
// - Handling Buffered and Direct reads appropriately.
// - Rate limiting compaction reads.
// - Notifying any interested listeners on the completion of a read.
// - Updating IO stats.
class RandomAccessFileReader {
private:
#ifndef ROCKSDB_LITE
void NotifyOnFileReadFinish(uint64_t offset, size_t length,
const FileOperationInfo::TimePoint& start_ts,
const FileOperationInfo::TimePoint& finish_ts,
const Status& status) const {
FileOperationInfo info(file_name_, start_ts, finish_ts);
info.offset = offset;
info.length = length;
info.status = status;
for (auto& listener : listeners_) {
listener->OnFileReadFinish(info);
}
}
#endif // ROCKSDB_LITE
bool ShouldNotifyListeners() const { return !listeners_.empty(); }
std::unique_ptr<FSRandomAccessFile> file_;
std::string file_name_;
Env* env_;
Statistics* stats_;
uint32_t hist_type_;
HistogramImpl* file_read_hist_;
RateLimiter* rate_limiter_;
std::vector<std::shared_ptr<EventListener>> listeners_;
public:
explicit RandomAccessFileReader(
std::unique_ptr<FSRandomAccessFile>&& raf, const std::string& _file_name,
Env* env = nullptr, Statistics* stats = nullptr, uint32_t hist_type = 0,
HistogramImpl* file_read_hist = nullptr,
RateLimiter* rate_limiter = nullptr,
const std::vector<std::shared_ptr<EventListener>>& listeners = {})
: file_(std::move(raf)),
file_name_(std::move(_file_name)),
env_(env),
stats_(stats),
hist_type_(hist_type),
file_read_hist_(file_read_hist),
rate_limiter_(rate_limiter),
listeners_() {
#ifndef ROCKSDB_LITE
std::for_each(listeners.begin(), listeners.end(),
[this](const std::shared_ptr<EventListener>& e) {
if (e->ShouldBeNotifiedOnFileIO()) {
listeners_.emplace_back(e);
}
});
#else // !ROCKSDB_LITE
(void)listeners;
#endif
}
RandomAccessFileReader(RandomAccessFileReader&& o) ROCKSDB_NOEXCEPT {
*this = std::move(o);
}
RandomAccessFileReader& operator=(RandomAccessFileReader&& o)
ROCKSDB_NOEXCEPT {
file_ = std::move(o.file_);
env_ = std::move(o.env_);
stats_ = std::move(o.stats_);
hist_type_ = std::move(o.hist_type_);
file_read_hist_ = std::move(o.file_read_hist_);
rate_limiter_ = std::move(o.rate_limiter_);
return *this;
}
RandomAccessFileReader(const RandomAccessFileReader&) = delete;
RandomAccessFileReader& operator=(const RandomAccessFileReader&) = delete;
// In non-direct IO mode,
// 1. if using mmap, result is stored in a buffer other than scratch;
// 2. if not using mmap, result is stored in the buffer starting from scratch.
//
// In direct IO mode, an aligned buffer is allocated internally.
// 1. If aligned_buf is null, then results are copied to the buffer
// starting from scratch;
// 2. Otherwise, scratch is not used and can be null, the aligned_buf owns
// the internally allocated buffer on return, and the result refers to a
// region in aligned_buf.
Status Read(uint64_t offset, size_t n, Slice* result, char* scratch,
AlignedBuf* aligned_buf, bool for_compaction = false) const;
// REQUIRES:
// num_reqs > 0, reqs do not overlap, and offsets in reqs are increasing.
// In non-direct IO mode, aligned_buf should be null;
// In direct IO mode, aligned_buf stores the aligned buffer allocated inside
// MultiRead, the result Slices in reqs refer to aligned_buf.
Status MultiRead(FSReadRequest* reqs, size_t num_reqs,
AlignedBuf* aligned_buf) const;
Status Prefetch(uint64_t offset, size_t n) const {
return file_->Prefetch(offset, n, IOOptions(), nullptr);
}
FSRandomAccessFile* file() { return file_.get(); }
std::string file_name() const { return file_name_; }
bool use_direct_io() const { return file_->use_direct_io(); }
};
} // namespace ROCKSDB_NAMESPACE