Qt Slot Thread-safety

• Qt Public Slots
• Qt Thread Signal Slot

Making multithreading easy. Qt provides thread support in the form of platform-independent threading classes, a thread-safe way of posting events, and signal-slot connections across threads. C: Thread Safety in a Signal/Slot Library. I suspect there is no clearly good answer, but clarity will come from documenting the guarantees you wish to make about concurrent access to an Emitter object.

Signals and slots were one of the distinguishing features that made Qt an exciting and innovative tool back in time. But sometimes you can teach new tricks to an old dog, and QObjects gained a new way to connect between signals and slots in Qt5, plus some extra features to connect to other functions which are not slots. Let’s review how to get the most of that feature. This assumes you are already moderately familiar with signals and slots.

One simple thought about the basics

I am not going to bore you with repeating basic knowledge you already have, but I want you to look at signals and slots from a certain angle, so it will be easier to understand the design of the feature I will cover next. What’s the purpose of signals and slots? It’s a way in which “one object” makes sure that when “something happened”, then “other object” “reacts to something happened”. As simple as that. That can be expressed in pseudocode like this:

Notice that the four phrases that are into quotes in the previous paragraph are the four arguments of the function call in the pseudocode. Notice also that one typical way to write the connect statement is aligning the arguments like this, because then the first column (first and third arguments) are object instances that answer “where?” and the second column (second and fourth arguments) are functions that answer “what?”.

In C++ instead of pseudocode, and using real life objects and classes, this would look like this in Qt4 or earlier:

That could be a typical statement from a “Hello World” tutorial, where a button is created and shown, and when it’s pressed the whole window closes and the application terminates.

Now to the main point that I want you to notice here. This has a very subtle advantage over a typical mechanism used in standard C or C++ 11 like callbacks with function pointers and lambda functions wrapped in std::function, and is subtle only because is so nice we often forget about it when we have used signals and slots for a while. If the sender object is destroyed, it obviously can not emit any signal because it is a member function of its class. But for the sender to call the receiver, it needs a pointer to it, and you as a user, don’t need to worry at all about the receiver being destroyed and becoming invalid (that is done automatically by the library), so you very rarely need to call QObject::disconnect.

So signals and slots are very safe by default, and in an automatic way.

The new versus the old way to use connect

The previous example shows one way that works across old versions of Qt published so far (Qt 1 to 5). Recently a blog post about porting a tutorial application from Qt 1 to Qt 5.11 has been published, and no porting was needed at all for signals, slots, or the connections! That doesn’t mean the feature is perfect, since a new way to make connections was added, keeping all the previous functionality.

The main problem with the example above is that (as you probably knew, or guessed from being all uppercase) is that SIGNAL and SLOT are macros, and what those macros do is convert to a string the argument passed. This is a problem because any typo in what gets passed to those means that the call to connect would fail and return false. So since Qt 5.0, a new overload to QObject::connect exists, and supports passing as second and fourth arguments a function pointer to specify which member function should be called. Ported to the new syntax, the above example is:

Now any typo in the name will produce a compile time error. If you misspelled “click” with “clik” in the first example, that would only fail printing a warning in the console when that function gets called. If you did that in some dialog of an application you would have to navigate to that dialog to confirm that it worked! And it would be even more annoying if you were connecting to some error handling, and is not that easy to trigger said error. But if you did the same typo in the last example, it would be a compile time error, which is clearly much better.

This example is usually said to be using the “new syntax”, and the previous example the “old syntax”. Just remember that the old is still valid, but the new is preferred in most situations.

Since this is an exciting new feature added to a new major version, which has received some extra polishing during the minor releases, many blog posts from other members of the Qt community have been published about it (for example covering implementation details or the issues that could arise when there are arguments involved). I won’t cover those topics again, and instead I will focus on the details that in my experience would be most beneficial for people to read on.

No need to declare members as slots anymore (or almost)

Qt Public Slots

The new syntax allows to call not just a member function declared as slot in the header with public slots: (or with protected or private instead of public), but any kind of function (more on that in the next section). There is still one use case where you would want to declare functions as slots, and that is if you want to make that function usable by any feature that happens at run time. That could be QML, for example.

Connecting to anything callable

Now we can connect to any “callable”, which could be a free standing function, a lambda function or a member function of an object that doesn’t derive from QObject. That looks in code like the following:

But wait, where is that nice symmetry with 2 rows and two columns now?

When you connect to a lambda, there is a receiver object, the lambda itself, but there is no signature to specify since it’s the function call operator (the same would happen to a function object or “functor”, by the way). And when there is a free standing function there is a signature, but there is no instance, so the third and the fourth arguments of the first two calls are somewhat merged. Note that the arguments are still checked at compile time: the signal has no arguments, and the lambda has no arguments either. Both sender and receiver are in agreement.

The example using std::bind requires a bit more explanation if you are not familiar with it. In this case we have the two objects and the two function pointers, which is to be expected for what is wanted. We don’t often think about it like this, but we always need a pointer to call a member function (unless it is static). When it is not used, it is because this is implicit, and this->call() can be shortened to call(). So what std::bind does here is create a callable object that glues together the particular instance that we want with one member function. We could do the same with a lambda:

Note that std::bind is actually much more powerful, and can be very useful when the number of arguments differ. But we will leave that topic to another article.

One common use of the above pattern with std::bind is when you have a class implemented through a data pointer (private implementation or pimpl idiom). If you need a button or a timer to call a member function of the private class that is not going to be a QObject, you can write something like this:

Recovering symmetry, safety and convenience

With the previous examples that nice balance of the four arguments is gone. But we are missing something more important.

What would happen if the lambda of the previous examples would use an invalid pointer? In the very first C++ example we showed a button wanting to close the application. Imagine that the button required to close a dialog, or stop some network request, etc. If the object is destroyed because said dialog is already closed or the request finished long ago, we want to manage that automatically so we don’t use an invalid pointer.

An example. For some reason you show some widget and you need to do some last minute update after it has been shown. It needs to happen soon but not immediately, so you use a timer with a short timeout. And you write

That works, but has a subtle problem. It could be that the widget gets shown and immediately closed. The timer under the scenes doesn’t know that, and it will happily call you, and crash the application. If you made the timer connect to a slot of the widget, that won’t happen: as soon as the dialog goes away, the connection gets broken.

Qt Thread Signal Slot

Since Qt 5.2 we can have the best of both worlds, and recover that nice warm feeling of having a well balanced connect statement with two objects and two functions. 🙂

In that Qt version an additional overload was added to QObject::connect, and now there is the possibility to pass as third argument a so called “context object”, and as fourth argument the same variety of callables shown previously. Then the context object serves the purpose of automatically breaking the connection when the context is destroyed. That warranties the problem mentioned is now gone. You can easily handle that there are no longer invalid captures on a lambda.

The previous example is almost as the previous:

Now it is as if the lambda were a slot in your class, because to the timer, the context of the connection is the same.

The only requirement is that said context object has to be a QObject. This is not usually a problem, since you can create and ad-hoc QObject instance and even do simple but useful tricks with it. For example, say that you want to run a lambda only on the first click:

This will delete the ad-hoc QObject guard on the first invocation, and the connection will be automatically broken. The object also has the button as a parent, so it won’t be leaked if the button is never clicked and goes away (it will be deleted as well). You can use any QObject as context object, but the most common case will be to shut down timers, processes, requests, or anything related to what your user interface is doing when some dialog, window or panel closes.

Tip: There are utility classes in Qt to handle the lifetime of QObjects automatically, like QScopedPointer and QObjectCleanupHandler. If you have some part of the application using Qt classes but no UI tightly related to that, you can surely find a way to leverage those as members of a class not based on QObject. It is often stated as a criticism to Qt, that you can’t put QObjects in containers or smart pointers. Often the alternatives do exist and can be as good, if not better (but admittedly this is a matter of taste).

Bonus point: thread safety by thread affinity

The above section is the main goal of this article. The context object can save you crashes, and having to manually disconnect. But there is one additional important use of it: making the signal be delivered in the thread that you prefer, so you can save from tedious and error prone locking.

Again, there is one killer feature of signals and slots that we often ignore because it happens automatically. When one QObject instance is the receiver of a signal, its thread affinity is checked, and by default the signal is delivered directly as a function call when is the same thread affinity of the sender. But if the thread affinity differs, it will be delivered posting an event to the object. The internals of Qt will convert that event to a function call that will happen in the next run of the event loop of the receiver, so it will be in the “normal” thread for that object, and you often can forget about locks entirely. The locks are inside Qt, because QCoreApplication::postEvent (the function used to add the event to the queue) is thread-safe. In case of need, you can force a direct call from different threads, or a queued call from the same thread. Check the fifth argument in the QObject::connect documentation (it’s an argument which defaults to Qt::AutoConection).

Let’s see it in a very typical example.

This shows a class that derives from QRunnable to reimplement the run() function, and that derives from QObject to provide the finished() signal. An instance is created after the user activates a button, and then we show some progress bar and run the task. But we want to notify the user when the task is done (show some message, hide some progress bar, etc.).

In the above example, the third argument (context object) might be forgotten, and the code will compile and run, but it would be a serious bug. It would mean that you would attempt to call into the UI thread from the thread where the task was run (which is a helper thread pool, not the UI thread). This is wrong, and in some cases Qt will nicely warn you that you are using some function from the wrong thread, but if you are not lucky, you will have a mysterious crash.

Wrap up

Hopefully now you’ve understood why that odd point was made in the introduction section. You don’t have to agree that it is aesthetically pleasing to write the arguments to connect in two rows and two columns, but if you understood the importance of using a context object as a rule of thumb, you probably will find your preferred way to remember if that third argument is needed when you write (or review other’s) code using connect.

Qt provides thread support in the form of platform-independent threading classes, a thread-safe way of posting events, and signal-slot connections across threads. This makes it easy to develop portable multithreaded Qt applications and take advantage of multiprocessor machines. Multithreaded programming is also a useful paradigm for performing time-consuming operations without freezing the user interface of an application.

Earlier versions of Qt offered an option to build the library without thread support. Since Qt 4.0, threads are always enabled.

This document is intended for an audience that has knowledge of, and experience with, multithreaded applications. If you are new to threading see our Recommended Reading list.

Topics:

The Threading Classes

Qt includes the following thread classes:

• QThread provides the means to start a new thread.
• QThreadStorage provides per-thread data storage.
• QMutex provides a mutual exclusion lock, or mutex.
• QMutexLocker is a convenience class that automatically locks and unlocks a QMutex.
• QReadWriteLock provides a lock that allows simultaneous read access.
• QReadLocker and QWriteLocker are convenience classes that automatically lock and unlock a QReadWriteLock.
• QSemaphore provides an integer semaphore (a generalization of a mutex).
• QWaitCondition provides a way for threads to go to sleep until woken up by another thread.

Creating a Thread

To create a thread, subclass QThread and reimplement its run() function. For example:

Then, create an instance of the thread object and call QThread::start(). The code that appears in the run() reimplementation will then be executed in a separate thread. Creating threads is explained in more detail in the QThread documentation.

Note that QCoreApplication::exec() must always be called from the main thread (the thread that executes main()), not from a QThread. In GUI applications, the main thread is also called the GUI thread because it's the only thread that is allowed to perform GUI-related operations.

In addition, you must create the QApplication (or QCoreApplication) object before you can create a QThread.

Synchronizing Threads

The QMutex, QReadWriteLock, QSemaphore, and QWaitCondition classes provide means to synchronize threads. While the main idea with threads is that they should be as concurrent as possible, there are points where threads must stop and wait for other threads. For example, if two threads try to access the same global variable simultaneously, the results are usually undefined.

QMutex provides a mutually exclusive lock, or mutex. At most one thread can hold the mutex at any time. If a thread tries to acquire the mutex while the mutex already locked, the thread will be put to sleep until the thread that current holds the mutex unlocks it. Mutexes are often used to protect accesses to shared data (i.e., data that can be accessed from multiple threads simultaneously). In the Reentrancy and Thread-Safety section below, we will use it to make a class thread-safe.

QReadWriteLock is similar to QMutex, except that it distinguishes between 'read' and 'write' access to shared data and allows multiple readers to access the data simultaneously. Using QReadWriteLock instead of QMutex when it is possible can make multithreaded programs more concurrent.

QSemaphore is a generalization of QMutex that protects a certain number of identical resources. In contrast, a mutex protects exactly one resource. The Semaphores example shows a typical application of semaphores: synchronizing access to a circular buffer between a producer and a consumer.

QWaitCondition allows a thread to wake up other threads when some condition has been met. One or many threads can block waiting for a QWaitCondition to set a condition with wakeOne() or wakeAll(). Use wakeOne() to wake one randomly selected event or wakeAll() to wake them all. The Wait Conditions example shows how to solve the producer-consumer problem using QWaitCondition instead of QSemaphore.

Reentrancy and Thread-Safety

Throughout the Qt documentation, the terms reentrant and thread-safe are used to specify how a function can be used in multithreaded applications:

• A reentrant function can be called simultaneously by multiple threads provided that each invocation of the function references unique data.
• A thread-safe function can be called simultaneously by multiple threads when each invocation references shared data. All access to the shared data is serialized.

By extension, a class is said to be reentrant if any of its functions can be called simultaneously by multiple threads on different instances of the class, and thread-safe if it even works if the different threads operate on the same instance.

Note that the terminology in this domain isn't entirely standardized. POSIX uses a somewhat different definition of reentrancy and thread-safety for its C APIs. When dealing with an object-oriented C++ class library such as Qt, the definitions must be adapted.

Most C++ classes are inherently reentrant, since they typically only reference member data. Any thread can call such a member function on an instance of the class, as long as no other thread is calling a member function on the same instance. For example, the Counter class below is reentrant:

The class isn't thread-safe, because if multiple threads try to modify the data member n, the result is undefined. This is because C++'s ++ and -- operators aren't necessarily atomic. Indeed, they usually expand to three machine instructions:

1. Load the variable's value in a register.
2. Increment or decrement the register's value.
3. Store the register's value back into main memory.

If thread A and thread B load the variable's old value simultaneously, increment their register, and store it back, they end up overwriting each other, and the variable is incremented only once!

Clearly, the access must be serialized: Thread A must perform steps 1, 2, 3 without interruption (atomically) before thread B can perform the same steps; or vice versa. An easy way to make the class thread-safe is to protect all access to the data members with a QMutex:

The QMutexLocker class automatically locks the mutex in its constructor and unlocks it when the destructor is invoked, at the end of the function. Locking the mutex ensures that access from different threads will be serialized. The mutex data member is declared with the mutable qualifier because we need to lock and unlock the mutex in value(), which is a const function.

Most Qt classes are reentrant and not thread-safe, to avoid the overhead of repeatedly locking and unlocking a QMutex. For example, QString is reentrant, meaning that you can use it in different threads, but you can't access the same QString object from different threads simultaneously (unless you protect it with a mutex yourself). A few classes and functions are thread-safe; these are mainly thread-related classes such as QMutex, or fundamental functions such as QCoreApplication::postEvent().

Threads and QObjects

QThread inherits QObject. It emits signals to indicate that the thread started or finished executing, and provides a few slots as well.

More interesting is that QObjects can be used in multiple threads, emit signals that invoke slots in other threads, and post events to objects that 'live' in other threads. This is possible because each thread is allowed to have its own event loop.

QObject Reentrancy

QObject is reentrant. Most of its non-GUI subclasses, such as QTimer, QTcpSocket, QUdpSocket, QHttp, QFtp, and QProcess, are also reentrant, making it possible to use these classes from multiple threads simultaneously. There are two constraints to be aware of:

• The child of a QObject must always be created in the thread where the parent was created. This implies, among other things, that you should never pass the QThread object (this) as the parent of an object created in the thread (since the QThread object itself was created in another thread).
• You must ensure that all objects created in a thread are deleted before you delete the QThread. This can be done easily by creating the objects on the stack in your run() implementation.

Although QObject is reentrant, the GUI classes, notably QWidget and all its subclasses, are not reentrant. They can only be used from the main thread. As noted earlier, QCoreApplication::exec() must also be called from that thread.

In practice, the impossibility of using GUI classes in other threads than the main thread can easily be worked around by putting time-consuming operations in a separate worker thread and displaying the results on screen in the main thread when the worker thread is finished. This is the approach used for implementing the Mandelbrot and the Blocking Fortune Client example.

Per-Thread Event Loop

Each thread can have its own event loop. The initial thread starts its event loops using QCoreApplication::exec(); other threads can start an event loop using QThread::exec(). Like QCoreApplication, QThread provides an exit(int) function and a quit() slot.

An event loop in a thread makes it possible for the thread to use certain non-GUI Qt classes that require the presence of an event loop (such as QTimer, QTcpSocket, and QProcess). It also makes it possible to connect signals from any threads to slots of a specific thread. This is explained in more detail in the Signals and Slots Across Threads section below.

A QObject instance is said to live in the thread in which it is created. Events to that object are dispatched by that thread's event loop. The thread in which a QObject lives is available using QObject::thread().

Note that for QObjects that are created before QApplication, QObject::thread() returns zero. This means that the main thread will only handle posted events for these objects; other event processing is not done at all for objects with no thread. Use the QObject::moveToThread() function to change the thread affinity for an object and its children (the object cannot be moved if it has a parent).

Calling delete on a QObject from another thread than the thread where it is created (or accessing the object in other ways) is unsafe unless you can guarantee that the object isn't processing events at the same moment. Use QObject::deleteLater() instead; it will post a DeferredDelete event, which the event loop of the object's thread will eventually pick up.

If no event loop is running, events won't be delivered to the object. For example, if you create a QTimer object in a thread but never call exec(), the QTimer will never emit its timeout() signal. Calling deleteLater() won't work either. (These restrictions apply to the main thread as well.)

You can manually post events to any object in any thread at any time using the thread-safe function QCoreApplication::postEvent(). The events will automatically be dispatched by the event loop of the thread where the object was created.

Event filters are supported in all threads, with the restriction that the monitoring object must live in the same thread as the monitored object. Similarly, QCoreApplication::sendEvent() (unlike postEvent()) can only be used to dispatch events to objects living in the thread from which the function is called.

Accessing QObject Subclasses from Other Threads

QObject and all of its subclasses are not thread-safe. This includes the entire event delivery system. It is important to keep in mind that the event loop may be delivering events to your QObject subclass while you are accessing the object from another thread.

If you are calling a function on an QObject subclass that doesn't live in the current thread and the object might receive events, you must protect all access to your QObject subclass's internal data with a mutex; otherwise, you may experience crashes or other undesired behavior.

Like other objects, QThread objects live in the thread where the object was created -- not in the thread that is created when QThread::run() is called. It is generally unsafe to provide slots in your QThread subclass, unless you protect the member variables with a mutex.

On the other hand, you can safely emit signals from your QThread::run() implementation, because signal emission is thread-safe.

Signals and Slots Across Threads

Qt supports three types of signal-slot connections:

• With direct connections, the slot gets called immediately when the signal is emitted. The slot is executed in the thread that emitted the signal (which is not necessarily the thread where the receiver object lives).
• With queued connections, the slot is invoked when control returns to the event loop of the thread to which the object belongs. The slot is executed in the thread where the receiver object lives.
• With auto connections (the default), the behavior is the same as with direct connections if the signal is emitted in the thread where the receiver lives; otherwise, the behavior is that of a queued connection.

The connection type can be specified by passing an additional argument to connect(). Be aware that using direct connections when the sender and receiver live in different threads is unsafe if an event loop is running in the receiver's thread, for the same reason that calling any function on an object living in another thread is unsafe.

QObject::connect() itself is thread-safe.

The Mandelbrot example uses a queued connection to communicate between a worker thread and the main thread. To avoid freezing the main thread's event loop (and, as a consequence, the application's user interface), all the Mandelbrot fractal computation is done in a separate worker thread. The thread emits a signal when it is done rendering the fractal.

Similarly, the Blocking Fortune Client example uses a separate thread for communicating with a TCP server asynchronously.

Threads and Implicit Sharing

Qt uses an optimization called implicit data sharing for many of its value class, notably QImage and QString. In many people's minds, implicit sharing and multithreading are incompatible concepts, because of the way the reference counting is typically done. One solution is to protect the internal reference counter with a mutex, but this is prohibitively slow. Earlier versions of Qt didn't provide a satisfactory solution to this problem.

Beginning with Qt 4, implicit shared classes can safely be copied across threads, like any other value classes. They are fully reentrant. The implicit sharing is really implicit. This is implemented using atomic reference counting operations, which are implemented in assembly language for the different platforms supported by Qt. Atomic reference counting is very fast, much faster than using a mutex.

This having been said, if you access the same object in multiple threads simultaneously (as opposed to copies of the same object), you still need a mutex to serialize the accesses, just like with any reentrant class.

To sum it up, implicitly shared classes in Qt 4 are really implicitly shared. Even in multithreaded applications, you can safely use them as if they were plain, non-shared, reentrant classes.

Threads and the SQL Module

A connection can only be used from within the thread that created it. Moving connections between threads or creating queries from a different thread is not supported.

In addition, the third party libraries used by the QSqlDrivers can impose further restrictions on using the SQL Module in a multithreaded program. Consult the manual of your database client for more information

Recommended Reading