C# Lock

原文：http://www.dotnetperls.com/lock

Locking is essential in threaded programs. It restricts code from being executed by more than one thread at the same time. This makes threaded programs reliable. The lockstatement uses a special syntax form to restrict concurrent access.

Note:Lock is compiled into a lower-level implementation based on threading primitives.

Example

Here we see a static method "A" that uses the lock statement on an object. When the method A is called many times on new threads, each invocation of the method accesses the threading primitives implemented by the lock.

Then:Only one method A can call the statements protected by the lock at a single time, regardless of the thread count.

Program that uses lock statement: C#

using System;
using System.Threading;

class Program
{
    static readonly object _object = new object();

    static void A()
    {
	// Lock on the readonly object.
	// ... Inside the lock, sleep for 100 milliseconds.
	// ... This is thread serialization.
	lock (_object)
	{
	    Thread.Sleep(100);
	    Console.WriteLine(Environment.TickCount);
	}
    }

    static void Main()
    {
	// Create ten new threads.
	for (int i = 0; i < 10; i++)
	{
	    ThreadStart start = new ThreadStart(A);
	    new Thread(start).Start();
	}
    }
}

Possible output of the program

28106840
28106949
28107043
28107136
28107246
28107339
28107448
28107542
28107636
28107745

In this example, the Main method creates ten new threads, and then calls Start on each one. The method A is invoked ten times, but the tick count shows the protected method region is executed sequentially—about 100 milliseconds apart.

Note:If you remove the lock statement, the methods will be executed all at once, with no synchronization.

Intermediate representation

Let‘s examine the intermediate representation for the lock statement in the above example method A. In compiler theory, high-level source texts are translated to lower-level streams of instructions.

Tip:The lock statement here is transformed into calls to the static methods Monitor.Enter and Monitor.Exit.

Also:The lock is actually implemented with a try-finally construct. This uses the exception handling control flow.

Intermediate representation for method using lock

.method private hidebysig static void A() cil managed
{
    .maxstack 2
    .locals init (
	[0] object obj2)
    L_0000: ldsfld object Program::_object
    L_0005: dup
    L_0006: stloc.0
    L_0007: call void [mscorlib]System.Threading.Monitor::Enter(object)
    L_000c: ldc.i4.s 100
    L_000e: call void [mscorlib]System.Threading.Thread::Sleep(int32)
    L_0013: call int32 [mscorlib]System.Environment::get_TickCount()
    L_0018: call void [mscorlib]System.Console::WriteLine(int32)
    L_001d: leave.s L_0026
    L_001f: ldloc.0
    L_0020: call void [mscorlib]System.Threading.Monitor::Exit(object)
    L_0025: endfinally
    L_0026: ret
    .try L_000c to L_001f finally handler L_001f to L_0026
}

Relativity

By using the lock statement to synchronize accesses, we are creating a communication between time and state. The state is connected to the concept of time and sequential accesses to the lock.

In the Theory of Relativity, there is also a communication between time and state. This is the speed of light, which is a constant based on the relation of time and space. This connection is present also in locks—in threading constructs.

Tip:For a better description of how relativity mirrors concurrent synchronization, please see the wizard book.

Summary

We examined the lock statement in the C# language, first seeing its usage in an example program, and then describing this synchronization. Next, we stepped into the intermediate representation and its meaning in compiler theory.

Finally:We related the Theory of Relativity and the complexities of the physical universe to the lock statement.