Overview of synchronization primitives
.NET provides a range of types that you can use to synchronize access to a shared resource or coordinate thread interaction.
Use the same synchronization primitive instance to protect every access to a shared resource. Multiple threads can access a resource concurrently if you use different synchronization primitive instances to protect access to a resource or some parts of code access a resource directly.
WaitHandle class and lightweight synchronization types
Multiple .NET synchronization primitives derive from the System.Threading.WaitHandle class, which encapsulates a native operating system synchronization handle and uses a signaling mechanism for thread interaction. Those classes include:
- System.Threading.Mutex, which grants exclusive access to a shared resource. The state of a mutex is signaled if no thread owns it.
- System.Threading.Semaphore, which limits the number of threads that can access a shared resource or a pool of resources concurrently. The state of a semaphore is set to signaled when its count is greater than zero, and nonsignaled when its count is zero.
- System.Threading.EventWaitHandle, which represents a thread synchronization event and can be either in a signaled or unsignaled state.
- System.Threading.AutoResetEvent, which derives from EventWaitHandle and, when signaled, resets automatically to an unsignaled state after releasing a single waiting thread.
- System.Threading.ManualResetEvent, which derives from EventWaitHandle and, when signaled, stays in a signaled state until the Reset method is called.
In the .NET Framework and .NET Core, some of these types can represent named system synchronization handles, which are visible throughout the operating system and can be used for the inter-process synchronization:
- Mutex (.NET Framework and .NET Core),
- Semaphore (.NET Framework and .NET Core on Windows),
- EventWaitHandle (.NET Framework and .NET Core on Windows).
For more information, see the WaitHandle API reference.
Lightweight synchronization types don't rely on underlying operating system handles and typically provide better performance. However, they cannot be used for the inter-process synchronization. Use those types for thread synchronization within one application.
Synchronization of access to a shared resource
.NET provides a range of synchronization primitives to control access to a shared resource by multiple threads.
The System.Threading.Monitor class grants mutually exclusive access to a shared resource by acquiring or releasing a lock on the object that identifies the resource. While a lock is held, the thread that holds the lock can again acquire and release the lock. Any other thread is blocked from acquiring the lock and the Monitor.Enter method waits until the lock is released. The Enter method acquires a released lock. You can also use the Monitor.TryEnter method to specify the amount of time during which a thread attempts to acquire a lock. Because the Monitor class has thread affinity, the thread that acquired a lock must release the lock by calling the Monitor.Exit method.
For more information, see the Monitor API reference.
Use the lock statement in C# and the SyncLock statement in Visual Basic to synchronize access to a shared resource instead of using the Monitor class directly. Those statements are implemented by using the Enter and Exit methods and a
try…finally block to ensure that the acquired lock is always released.
The System.Threading.Mutex class, like Monitor, grants exclusive access to a shared resource. Use one of the Mutex.WaitOne method overloads to request the ownership of a mutex. Like Monitor, Mutex has thread affinity and the thread that acquired a mutex must release it by calling the Mutex.ReleaseMutex method.
Unlike Monitor, the Mutex class can be used for inter-process synchronization. To do that, use a named mutex, which is visible throughout the operating system. To create a named mutex instance, use a Mutex constructor that specifies a name. You also can call the Mutex.OpenExisting method to open an existing named system mutex.
The System.Threading.SpinLock structure, like Monitor, grants exclusive access to a shared resource based on the availability of a lock. When SpinLock attempts to acquire a lock that is unavailable, it waits in a loop, repeatedly checking until the lock becomes available.
The System.Threading.ReaderWriterLockSlim class grants exclusive access to a shared resource for writing and allows multiple threads to access the resource simultaneously for reading. You might want to use ReaderWriterLockSlim to synchronize access to a shared data structure that supports thread-safe read operations, but requires exclusive access to perform write operation. When a thread requests exclusive access (for example, by calling the ReaderWriterLockSlim.EnterWriteLock method), subsequent reader and writer requests block until all existing readers have exited the lock, and the writer has entered and exited the lock.
For more information, see the ReaderWriterLockSlim API reference.
Semaphore and SemaphoreSlim classes
The System.Threading.Semaphore and System.Threading.SemaphoreSlim classes limit the number of threads that can access a shared resource or a pool of resources concurrently. Additional threads that request the resource wait until any thread releases the semaphore. Because the semaphore doesn't have thread affinity, a thread can acquire the semaphore and another one can release it.
On Windows, you can use Semaphore for the inter-process synchronization. To do that, create a Semaphore instance that represents a named system semaphore by using one of the Semaphore constructors that specifies a name or the Semaphore.OpenExisting method. SemaphoreSlim doesn't support named system semaphores.
Thread interaction, or signaling
Thread interaction (or thread signaling) means that a thread must wait for notification, or a signal, from one or more threads in order to proceed. For example, if thread A calls the Thread.Join method of thread B, thread A is blocked until thread B completes. The synchronization primitives described in the preceding section provide a different mechanism for signaling: by releasing a lock, a thread notifies another thread that it can proceed by acquiring the lock.
This section describes additional signaling constructs provided by .NET.
EventWaitHandle, AutoResetEvent, ManualResetEvent, and ManualResetEventSlim classes
The System.Threading.EventWaitHandle class represents a thread synchronization event.
A synchronization event can be either in an unsignaled or signaled state. When the state of an event is unsignaled, a thread that calls the event's WaitOne overload is blocked until an event is signaled. The EventWaitHandle.Set method sets the state of an event to signaled.
The behavior of an EventWaitHandle that has been signaled depends on its reset mode:
- An EventWaitHandle created with the EventResetMode.AutoReset flag resets automatically after releasing a single waiting thread. It's like a turnstile that allows only one thread through each time it's signaled. The System.Threading.AutoResetEvent class, which derives from EventWaitHandle, represents that behavior.
- An EventWaitHandle created with the EventResetMode.ManualReset flag remains signaled until its Reset method is called. It's like a gate that is closed until signaled and then stays open until someone closes it. The System.Threading.ManualResetEvent class, which derives from EventWaitHandle, represents that behavior. The System.Threading.ManualResetEventSlim class is a lightweight alternative to ManualResetEvent.
On Windows, you can use EventWaitHandle for the inter-process synchronization. To do that, create a EventWaitHandle instance that represents a named system synchronization event by using one of the EventWaitHandle constructors that specifies a name or the EventWaitHandle.OpenExisting method.
The System.Threading.CountdownEvent class represents an event that becomes set when its count is zero. While CountdownEvent.CurrentCount is greater than zero, a thread that calls CountdownEvent.Wait is blocked. Call CountdownEvent.Signal to decrement an event's count.
In contrast to ManualResetEvent or ManualResetEventSlim, which you can use to unblock multiple threads with a signal from one thread, you can use CountdownEvent to unblock one or more threads with signals from multiple threads.
The System.Threading.Barrier class represents a thread execution barrier. A thread that calls the Barrier.SignalAndWait method signals that it reached the barrier and waits until other participant threads reach the barrier. When all participant threads reach the barrier, they proceed and the barrier is reset and can be used again.
You might use Barrier when one or more threads require the results of other threads before proceeding to the next computation phase.
The System.Threading.Interlocked class provides static methods that perform simple atomic operations on a variable. Those atomic operations include addition, increment and decrement, exchange and conditional exchange that depends on a comparison, and read operation of a 64-bit integer value.
For more information, see the Interlocked API reference.
The System.Threading.SpinWait structure provides support for spin-based waiting. You might want to use it when a thread has to wait for an event to be signaled or a condition to be met, but when the actual wait time is expected to be less than the waiting time required by using a wait handle or by otherwise blocking the thread. By using SpinWait, you can specify a short period of time to spin while waiting, and then yield (for example, by waiting or sleeping) only if the condition was not met in the specified time.