I have this code:
mainwindow.h:
namespace Ui {
class MainWindow;
}
class MainWindow : public QMainWindow {
private:
QMutex mutex;
}
mainwindow.cpp:
void MainWindow::on_calculateBtn_clicked() {
QMutexLocker locker(&mutex);
qDebug() << "mutex has been locked" << endl;
ui->calculateBtn->setEnabled(false);
startProcess(); // huge calcutations
ui->calculateBtn->setEnabled(true); // performed before startProcess() has finished (why?)
qDebug() << "mutex will be unlocked" << endl;
}
If I click calculateBtn again while startProcess() has not finished, my program crashed:
pure virtual method called
The program has unexpectedly finished.
I tried:
void MainWindow::on_calculateBtn_clicked() {
if (!processing) {
processing = true;
ui->calculateBtn->setEnabled(false);
startProcess();
ui->calculateBtn->setEnabled(true); // performed before startProcess() has finished (why?)
processing = false;
}
}
There is no shared data, I just want one startProcess() will not be started before other startProcess() finished.
Why did it happen? I think that mutex have to lock function startProcess() in on_calculateBtn_clicked() and nothing should happens. It seems I don't know any important things. Thanks in advance for any advice.
The same mutex is locked twice from the same thread (the main thread, which contains the event loop), which is invalid for a non-recursive mutex.
But even a recursive mutex will not solve the basic problem of your code; you need a flag to indicate that you are already doing the calculations, and return from all subsequent calls to your method while they are running, else you'll start them multiple times in the same thread, one interrupting the other, probably with bad results. Even better, disable the button while the method is running and take care it isn't called by other ways.
However, if calling startProcess() multiple times and run it simultaneously is intended, you'll have to start a thread for each button press and take care for access to shared data (using mutexes, most probably) - that's where the real fun begins.
I think that you (by default) have a Qt::DirectConnection with this button press, right? i.e.:
connect(..., SIGNAL(...),
..., SLOT(:on_calculateBtn_clicked()), <by-default-Qt::DirectConnection>);
The issue I see here is that the first button press will run the function void MainWindow::on_calculateBtn_clicked() immediately.... which is all good so far, the mutex is locked and huge calcs are running.
However when you press the button again, void MainWindow::on_calculateBtn_clicked() is again immediate run (like an interrupt). The first thing it does is try to lock the mutex and it must hang here.
If you make connection to the slot void MainWindow::on_calculateBtn_clicked() Qt::QueuedConnection then it won't trigger the button press until it has clear the other tasks on its task queue.
but.... weather or not your design here is good is questionable, I think over you should re-think your strategy (as some comments have suggested)
EDIT
Oh yeah, meant to add..... to answer your question, therefore I don't think the mutex is begin unlocked twice... its just the nature of the direct connection
Related
I am currently working on a game in C++ which is supposed to use multithreading to handle bot instances.
The Bot class has a member function start(), which calls the bot's function walk() in a new thread. This is my attempt (in excerpts:)
void Bot::walk(Physics& physics)
{
// do stuff
std::cout << "call to walk() function" << std::endl;
}
void Bot::start(Physics& physics)
{
std::thread m_thread(&Bot::walk, this, std::ref(physics));
m_thread.join();
}
In another class which handles the initialization of the game, all bot instances are started like this: (m_bots is a vector of Bots)
void Level::start()
{
// Start all Bots
for(auto it: m_bots)
{
it->start(*m_physics);
}
}
On startup the game window opens as usual but then continues to freeze. The console output shows that the walk() function is called though.
What thread-specific catch am I missing?
I hope these excerpts cover all the relevant information. If this is not the case, please let me know!
Joining a thread blocks until the thread finishes. As a result, Bot::start won't return until the thread that it spawns finishes, so only one thread will run at a time.
Someone will probably tell you that the solution is to change join() to detach(). That's not really the solution, unless your design calls for a bunch of free-running threads; if you care about when they finish, join() is the way to do it, just not there.
One approach is to create an object of type std::vector<std::thread> and launch all of the thread with calls to Bot::start. After the threads are running you can join all the spawned threads.
I have a multithreaded application written in C++ with Qt. Currently, my application works by having a QThread instance (I have NOT subclassed QThread) which uses the default run() implementation that just calls QThread's exec() method, which provides an event dispatch loop.
I call moveToThread on certain QObject subclasses which actually perform the work I want done in a separate thread. I tell these objects to do work using Qt's signals/slots mechanism. I stop the thread gracefully by informing my worker objects to stop their work, and then calling quit() and wait() on my thread. This all works very nicely.
However, I now want to implement the following scenario:
The user clicks e.g. the "X" button on my application, because they want to close it.
I don't want any new work to be started, so I pause the event dispatch thread. If the current event being dispatched continues running, that's fine.
I prompt the user, allowing them to either a) discard all remaining jobs and exit (using quit() andwait()` - this already works), or b) don't exit the application, but instead continue working (resume the thread).
My problem is that QThread doesn't seem to have a pause() method. I've seen various examples online which add one (like the answers to this question). The problem is that these examples depend on having a custom run() implementation, and implementing pause there. Since I'm relying on QThread's event dispatch loop, these solutions won't work. I've considered doing something like reimplementing exec() or creating my own subclass of QAbstractEventDispatcher, but these solutions seem like a whole lot of work to get simple pause / resume functionality.
What's the easiest way to pause QThread's event dispatch loop (preventing it from dispatching any new events, but letting the current event continue)?
I tried out the method suggested in the comments, but it took a bit of screwing around to get it to work totally correctly, so here's what I ended up with:
I subclassed QThread to add two new methods: pause and resume. There were a few things that needed to be dealt with delicately:
Calling start() while the thread is still running does nothing. Since resume() might be called before the thread's existing job stops running, we need to do the actual resume in a slot connected to the thread's finished() signal.
The finished() signal may be emitted just before the thread actually stops. Because of this, we need to call wait() in our slot before calling start().
If resume() is called after the thread is already stopped, simply setting state variables won't work, because finished() will never be emitted. Because of this, we need to deal with that case by having non-signal-related resume code in the resume() method as well.
Here's the final product.
pausablethread.h:
#ifndef INCLUDE_PAUSABLE_THREAD_H
#define INCLUDE_PAUSABLE_THREAD_H
#include <QThread>
class QMutex;
class PausableThread : public QThread
{
Q_OBJECT
public:
PausableThread(QObject *parent = 0);
virtual ~PausableThread();
void pause();
void resume();
private:
QMutex *controlMutex;
bool paused;
bool resumeScheduled;
private Q_SLOTS:
void doResume();
};
#endif
pausablethread.cpp:
#include "pausablethread.h"
#include <QMutex>
#include <QMutexLocker>
PausableThread::PausableThread(QObject *parent)
: QThread(parent), paused(false), resumeScheduled(false)
{
controlMutex = new QMutex(QMutex::NonRecursive);
QObject::connect(this, SIGNAL(finished()),
this, SLOT(doResume()));
}
PausableThread::~PausableThread()
{
delete controlMutex;
}
void PausableThread::pause()
{
QMutexLocker locker(controlMutex);
if(paused)
return;
paused = true;
quit();
}
void PausableThread::resume()
{
QMutexLocker locker(controlMutex);
if(!paused)
return;
if(resumeScheduled)
return;
if(isFinished())
{
start();
paused = false;
resumeScheduled = false;
}
else
{
resumeScheduled = true;
}
}
void PausableThread::doResume()
{ /* SLOT */
QMutexLocker locker(controlMutex);
if(!resumeScheduled)
return;
paused = false;
resumeScheduled = false;
wait();
start();
}
This seems to work, mostly. I believe there are some potential race conditions if the thread happens to finish or start at the same time execution is inside resume() or doResume() in a different thread. It's not exactly clear to me how to solve this.
I tried something like overriding the superclass's start() slot with the following:
void start(Priority priority)
{
QMutexLocker locker(controlMutex);
QThread::start(priority);
}
The problem with this is that start() never actually returns until the thread finishes, so it never releases its lock on the mutex. Thus, when doResume() tries to acquire a lock, a deadlock is encountered, and the thread isn't successfully resumed.
I think what is really needed is a mutex which is exclusively locked any time the thread's running state is changed, but it isn't clear to me how to implement this.
Regardless, the window for this race to occur is very small,and this works "well enough" to answer my question. Thanks to #Lol4t0 for the suggestion!
I have a situation where a notify() 'can' be called before a wait().
I am trying to make a simulator to schedule its next event when I 'notify' him by sending him messages. So I have devised a wait->notify->scedule chain
void Broker::pause()
{
boost::unique_lock<boost::mutex> lock(m_pause_mutex);
{
std::cout << "pausing the simulation" << std::endl;
m_cond_cnn.wait(lock);
std::cout << "Simulation UNpaused" << std::endl;
// the following line causes the current function to be called at
// a later time, and a notify() can happen before the current function
// is called again
Simulator::Schedule(MilliSeconds(xxx), &Broker::pause, this);
}
}
void Broker::messageReceiveCallback(std::string message) {
boost::unique_lock<boost::mutex> lock(m_pause_mutex);
{
m_cond_cnn.notify_one();
}
}
the problem here is that: there can be situations that a notify() is called before its wait() is called.
Is there a solution for such situation?
thank you
Condition variables can hardly be used alone, if only because, as you noticed, they only wake the currently waiting threads. There's also the matter of spurious wake-ups (ie. the condition variable can sometimes wake up a thread without any corresponding notify having been called). To work properly, condition variables usually need another variable to maintain a more reliable state.
To solve both those problems, in your case you just need to add a boolean flag:
boost::unique_lock<boost::mutex> lock(m_pause_mutex);
while (!someFlag)
m_cond_cnn.wait(lock);
someFlag = false;
//...
boost::unique_lock<boost::mutex> lock(m_pause_mutex);
someFlag = true;
m_cond_cnn.notify_one();
I think that syam's answer is fine in general but in your specific case where you seem to be using ns-3, I would suggest instead that you restructure your code to use the right primitives in ns-3:
I suspect that you use one of the ns-3 realtime simulator implementations. Good.
Schedule a keeplive event for the 0.1s to make sure that the simulator keeps running (it will top running when there are no events left).
Optionally, use a boolean in this keepalive event to check if you should reschedule the keepalive event or call Simulator::Stop.
Create a thread to run the simulator mainloop with Simulator::Run(). The simulator will sleep until the next scheduled event is supposed to expire or until a new event is externally scheduled
Use Simulator::ScheduleWithContext to schedule an event externally from another thread.
Keep in mind that the ns-3 API is not thread safe in general. The only ns-3 API that is thread-safe is ns3::Simulator::ScheduleWithContext. I can't stress out how important it is to not use any other API available in the ns-3:: namespace from a thread that is not the main thread.
I have been learning Qt for a while and a few days ago I decided to go on multi-threading by QThread. I have been following this person. Since he says that subclassing QThread is not the appropriate way but by using QObject is. I am following his method. But he has created another class which inherits from QObject. I decided to make a few changes:
class myclass
{
private:
dowork();
private slots:
slota();
slotb();
slotc();
}
myclass::dowork()
{
QThread lett;
QThread latt;
QThread pltt;
QObject lethread;
connect(&lett, SIGNAL(started()), this, SLOT(slota()));
lethread.moveToThread(&lett);
QObject lathread;
connect(&latt, SIGNAL(started()), this, SLOT(slotb()));
lathread.moveToThread(&latt);
QObject plthread;
connect(&pltt, SIGNAL(started()), this, SLOT(slotc()));
plthread.moveToThread(&pltt);
lett.start();
latt.start();
pltt.start();/*
lett.wait();
latt.wait();
pltt.wait();*/
while(lett.isRunning())
{
//qDebug() << "Still Running!";
}
lett.quit();
}
If I run the program it keeps running while loop infinitely! It doesn't come out of that at all.
I used qdebug inside the implementation of the 3 slots to check whether it is entering them or not, and surprisingly none of the slots gets called. Why is that so?
If I remove the comments for .wait then the program pauses.
What is wrong in this method? I want to avoid that extra class as my program is very small.
while(lett.isRunning())
{
//qDebug() << "Still Running!";
}
lett.quit();
Terribly wrong. The current thread is busy waiting and will be eating all the time. But this is not your problem. Your problem is certainly that the main event loop has probably not started yet, ie you didn't do QApplication.exec() yet.
Qt uses events to start threads, and when you do use movetoThread, start, specific events are queued in the current thread (ie the thread executing these methods). If the main event loop has already started, then will be processed as soon as they can. But in your case I bet that you main function is :
int main(){
QApplication app();
myclass onlyclassIwanttomake;
onlyclassIwanttomake.dowork();
app.exec(); <----HERE!!!
}
The thread will be started when the main thread executes app.exec();. When you are busy waiting, you are preventing the main thread to process the events required to start your
events.
EDIT: Well, another issue is that
lett.start();
latt.start();
pltt.start();
start 3 threads which are started, and stay idle forever. They are not doing anything, but they are running (even if you remove your loop).
This is a entry written about Qt threading. Please take the time to read it.
Greetings, everyone!
I have a class (say, "Switcher" ) that executes some very-very long operation and notifies its listener, that operation is complete. The operation is long, and I isolate actual switching into separate thread:
class Switcher
{
public:
// this is what other users call:
void StartSwitching()
{
// another switch is initiated, I must terminate previous switching operation:
if ( m_Thread != NULL )
{
if ( WaitForThread(m_Thread, 3000) != OK )
{
TerminateThread(m_Thread);
}
}
// start new switching thread:
m_Thread = StartNewThread( ThreadProc );
}
// this is a thread procedure:
static void ThreadProc()
{
DoActualSwitching();
NotifyListener();
}
private:
Thread m_Thread;
};
The logic is rather simple - if user initiates new switching before the previous one is complete, I terminate previous switching (don't care of what happens inside "DoActualSwitching()") and start the new one. The problem is that sometimes, when terminating thread, I loose the "NotifyListener()" call.
I would like to introduce some improvements to ensure, that NotifyListener() is called every time, even if thread is terminated. Is there any pattern to do this? I can only think of another thread, that infinitely waits for the switcher and if the switcher is done (correctly or by termination), it can emit notification. But introducing another thread seems an overplay for me. Can you think of any other solution (p.s. the platform is win32)?
Thank you!
First, you should never call TerminateThread. You cannot know which operation is terminated when calling TerminateThread and so that could lead to memory leaks/resource leaks/state corruption.
To get your thread to be interruptable/cancelable, you supply a 'cancel' state, which is checked by the thread itself. Then your notify end will always work.
TerminateThread() here whacks the thread, and if it was inside DoActualSwitching(), that's where it'll die, and NotifyListener() will not be called on that thread. This is what TerminateThread() does, and there is no way to make it behave differently.
What you are looking for is a bit more graceful way to terminate the thread. Without more info about your application it's difficult to suggest an optimal approach, but if you can edit DoActualSwitching(), then I'd add
if (WAIT_OBJECT_0 == WaitForSingleObject(m_ExitThreadEvent, 0))
break;
into the loop there, and call SetEvent(m_ExitThreadEvent) instead of TerminateThread(). Of course you'll need to create the event and add the handle to the class. If your model suggest that there is only one switching thread at a time, I'd use autoreset event here, otherwise some more code is needed.
Good luck!