dash/src/checkqueue.h
MarcoFalke 22838d1b8c Merge bitcoin/bitcoin#23397: Avoid excessive lock contention in CCheckQueue::Add
459e208276a4d1457d37bf41c977e62caf05456d Exit early for an empty vChecks in CCheckQueue::Add (Hennadii Stepanov)
c43aa623435a277a692dbde784e8a7146f5573e9 Avoid excessive lock contention in CCheckQueue::Add (Hennadii Stepanov)

Pull request description:

  This PR significantly reduces lock contention in the `CCheckQueue` class by releasing a mutex before calling `std::condition_variable::notify_one` and `std::condition_variable::notify_all`.

  From C++ [docs](https://en.cppreference.com/w/cpp/thread/condition_variable/notify_one):
  > The notifying thread does not need to hold the lock on the same mutex as the one held by the waiting thread(s); in fact doing so is a pessimization, since the notified thread would immediately block again, waiting for the notifying thread to release the lock.

  Related to:
  - #23167
  - #23223

ACKs for top commit:
  martinus:
    ACK 459e208, codereview and tested. I first thought this introduced a segfault in `psbt_wallet_tests/psbt_updater_test` because that test failed for me, but thats a different issue fixed in #23403.
  vasild:
    ACK 459e208276a4d1457d37bf41c977e62caf05456d
  theStack:
    Code-review ACK 459e208276a4d1457d37bf41c977e62caf05456d

Tree-SHA512: c197858656392ba3ebcd638d713cf93c9fb48b7b3bad193209490d2828f9c7e3ae4dee6f84674f2f34dceed894139562e29579ee7299e06756c8c990caddc5ed
2022-04-11 09:41:11 -07:00

256 lines
7.7 KiB
C++

// Copyright (c) 2012-2015 The Bitcoin Core developers
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#ifndef BITCOIN_CHECKQUEUE_H
#define BITCOIN_CHECKQUEUE_H
#include <sync.h>
#include <tinyformat.h>
#include <util/threadnames.h>
#include <algorithm>
#include <vector>
template <typename T>
class CCheckQueueControl;
/**
* Queue for verifications that have to be performed.
* The verifications are represented by a type T, which must provide an
* operator(), returning a bool.
*
* One thread (the master) is assumed to push batches of verifications
* onto the queue, where they are processed by N-1 worker threads. When
* the master is done adding work, it temporarily joins the worker pool
* as an N'th worker, until all jobs are done.
*/
template <typename T>
class CCheckQueue
{
private:
//! Mutex to protect the inner state
Mutex m_mutex;
//! Worker threads block on this when out of work
std::condition_variable m_worker_cv;
//! Master thread blocks on this when out of work
std::condition_variable m_master_cv;
//! The queue of elements to be processed.
//! As the order of booleans doesn't matter, it is used as a LIFO (stack)
std::vector<T> queue GUARDED_BY(m_mutex);
//! The number of workers (including the master) that are idle.
int nIdle GUARDED_BY(m_mutex){0};
//! The total number of workers (including the master).
int nTotal GUARDED_BY(m_mutex){0};
//! The temporary evaluation result.
bool fAllOk GUARDED_BY(m_mutex){true};
/**
* Number of verifications that haven't completed yet.
* This includes elements that are no longer queued, but still in the
* worker's own batches.
*/
unsigned int nTodo GUARDED_BY(m_mutex){0};
//! The maximum number of elements to be processed in one batch
const unsigned int nBatchSize;
std::vector<std::thread> m_worker_threads;
bool m_request_stop GUARDED_BY(m_mutex){false};
/** Internal function that does bulk of the verification work. */
bool Loop(bool fMaster)
{
std::condition_variable& cond = fMaster ? m_master_cv : m_worker_cv;
std::vector<T> vChecks;
vChecks.reserve(nBatchSize);
unsigned int nNow = 0;
bool fOk = true;
do {
{
WAIT_LOCK(m_mutex, lock);
// first do the clean-up of the previous loop run (allowing us to do it in the same critsect)
if (nNow) {
fAllOk &= fOk;
nTodo -= nNow;
if (nTodo == 0 && !fMaster)
// We processed the last element; inform the master it can exit and return the result
m_master_cv.notify_one();
} else {
// first iteration
nTotal++;
}
// logically, the do loop starts here
while (queue.empty() && !m_request_stop) {
if (fMaster && nTodo == 0) {
nTotal--;
bool fRet = fAllOk;
// reset the status for new work later
fAllOk = true;
// return the current status
return fRet;
}
nIdle++;
cond.wait(lock); // wait
nIdle--;
}
if (m_request_stop) {
return false;
}
// Decide how many work units to process now.
// * Do not try to do everything at once, but aim for increasingly smaller batches so
// all workers finish approximately simultaneously.
// * Try to account for idle jobs which will instantly start helping.
// * Don't do batches smaller than 1 (duh), or larger than nBatchSize.
nNow = std::max(1U, std::min(nBatchSize, (unsigned int)queue.size() / (nTotal + nIdle + 1)));
vChecks.resize(nNow);
for (unsigned int i = 0; i < nNow; i++) {
// We want the lock on the m_mutex to be as short as possible, so swap jobs from the global
// queue to the local batch vector instead of copying.
vChecks[i].swap(queue.back());
queue.pop_back();
}
// Check whether we need to do work at all
fOk = fAllOk;
}
// execute work
for (T& check : vChecks)
if (fOk)
fOk = check();
vChecks.clear();
} while (true);
}
public:
//! Mutex to ensure only one concurrent CCheckQueueControl
Mutex m_control_mutex;
//! Create a new check queue
explicit CCheckQueue(unsigned int nBatchSizeIn)
: nBatchSize(nBatchSizeIn)
{
}
//! Create a pool of new worker threads.
void StartWorkerThreads(const int threads_num)
{
{
LOCK(m_mutex);
nIdle = 0;
nTotal = 0;
fAllOk = true;
}
assert(m_worker_threads.empty());
for (int n = 0; n < threads_num; ++n) {
m_worker_threads.emplace_back([this, n]() {
util::ThreadRename(strprintf("scriptch.%i", n));
Loop(false /* worker thread */);
});
}
}
//! Wait until execution finishes, and return whether all evaluations were successful.
bool Wait()
{
return Loop(true /* master thread */);
}
//! Add a batch of checks to the queue
void Add(std::vector<T>& vChecks)
{
if (vChecks.empty()) {
return;
}
{
LOCK(m_mutex);
for (T& check : vChecks) {
queue.emplace_back();
check.swap(queue.back());
}
nTodo += vChecks.size();
}
if (vChecks.size() == 1) {
m_worker_cv.notify_one();
} else {
m_worker_cv.notify_all();
}
}
//! Stop all of the worker threads.
void StopWorkerThreads()
{
WITH_LOCK(m_mutex, m_request_stop = true);
m_worker_cv.notify_all();
for (std::thread& t : m_worker_threads) {
t.join();
}
m_worker_threads.clear();
WITH_LOCK(m_mutex, m_request_stop = false);
}
~CCheckQueue()
{
assert(m_worker_threads.empty());
}
};
/**
* RAII-style controller object for a CCheckQueue that guarantees the passed
* queue is finished before continuing.
*/
template <typename T>
class CCheckQueueControl
{
private:
CCheckQueue<T> * const pqueue;
bool fDone;
public:
CCheckQueueControl() = delete;
CCheckQueueControl(const CCheckQueueControl&) = delete;
CCheckQueueControl& operator=(const CCheckQueueControl&) = delete;
explicit CCheckQueueControl(CCheckQueue<T> * const pqueueIn) : pqueue(pqueueIn), fDone(false)
{
// passed queue is supposed to be unused, or nullptr
if (pqueue != nullptr) {
ENTER_CRITICAL_SECTION(pqueue->m_control_mutex);
}
}
bool Wait()
{
if (pqueue == nullptr)
return true;
bool fRet = pqueue->Wait();
fDone = true;
return fRet;
}
void Add(std::vector<T>& vChecks)
{
if (pqueue != nullptr)
pqueue->Add(vChecks);
}
~CCheckQueueControl()
{
if (!fDone)
Wait();
if (pqueue != nullptr) {
LEAVE_CRITICAL_SECTION(pqueue->m_control_mutex);
}
}
};
#endif // BITCOIN_CHECKQUEUE_H