C# 线程处理系列 专题二:线程池中的工作者线程

最后更新于:2022-04-02 00:10:39

# [C# 线程处理系列]专题二:线程池中的工作者线程 **目录:** 一、上节补充 二、CLR线程池基础 三、通过线程池的工作者线程实现异步 四、使用委托实现异步 五、任务 **一、上节补充** 对于Thread类还有几个常用方法需要说明的。 **1.1 Suspend和Resume方法** 这两个方法在.net Framework 1.0的时候就支持的方法,他们分别可以挂起线程和恢复挂起的线程。但在.net Framework 2.0以后的版本中这两个方法都过时了,MSDN的解释是这样: 警告: Suspend and Resume methods to synchronize the activities of threads." data-guid="ff8e76fddd4f9d11989c7d25f5e342d1">不要使用 Suspend 和 Resume 方法来同步线程的活动。您无法知道挂起线程时它正在执行什么代码。AppDomain might be blocked." data-guid="7c4587074f4e52d11eadc59e58a09a57">如果您在安全权限评估期间挂起持有锁的线程,则 AppDomain中的其他线程可能被阻止。AppDomain that attempt to use that class are blocked." data-guid="f164549efd95a0a94479d4d1c0d1ceec">如果您在线程正在执行类构造函数时挂起它,则 AppDomain中尝试使用该类的其他线程将被阻止。这样很容易发生死锁。 对于这个解释可能有点抽象吧,让我们来看看一段代码可能会清晰点: ``` class Program { static void Main(string[] args) { // 创建一个线程来测试 Thread thread1 = new Thread(TestMethod); thread1.Name = "Thread1"; thread1.Start(); Thread.Sleep(2000); Console.WriteLine("Main Thread is running"); ////int b = 0; ////int a = 3 / b; ////Console.WriteLine(a); thread1.Resume(); Console.Read(); } private static void TestMethod() { Console.WriteLine("Thread: {0} has been suspended!", Thread.CurrentThread.Name); //将当前线程挂起 Thread.CurrentThread.Suspend(); Console.WriteLine("Thread: {0} has been resumed!", Thread.CurrentThread.Name); } } ``` 在上面这段代码中thread1线程是在主线程中恢复的,但当主线程发生异常时,这时候就thread1一直处于挂起状态,此时thread1所使用的资源就不能释放(除非强制终止进程),当另外线程需要使用这快资源的时候, 这时候就很可能发生死锁现象。 上面一段代码还存在一个隐患,请看下面一小段代码: ``` class Program { static void Main(string[] args) { // 创建一个线程来测试 Thread thread1 = new Thread(TestMethod); thread1.Name = "Thread1"; thread1.Start(); Console.WriteLine("Main Thread is running"); thread1.Resume(); Console.Read(); } private static void TestMethod() { Console.WriteLine("Thread: {0} has been suspended!", Thread.CurrentThread.Name); Thread.Sleep(1000); //将当前线程挂起 Thread.CurrentThread.Suspend(); Console.WriteLine("Thread: {0} has been resumed!", Thread.CurrentThread.Name); } } ``` 当主线程跑(运行)的太快,做完自己的事情去唤醒thread1时,此时thread1还没有挂起而起唤醒thread1,此时就会出现异常了。并且上面使用的Suspend和Resume方法,编译器已经出现警告了,提示这两个方法已经过时, 所以在我们平时使用中应该尽量避免。 **1.2 Abort和 Interrupt方法** Abort方法和Interrupt都是用来终止线程的,但是两者还是有区别的。 1、他们抛出的异常不一样,Abort 方法抛出的异常是ThreadAbortException, Interrupt抛出的异常为ThreadInterruptedException 2、调用interrupt方法的线程之后可以被唤醒,然而调用Abort方法的线程就直接被终止不能被唤醒的。 下面一段代码是掩饰Abort方法的使用 ``` using System; using System.Threading; namespace ConsoleApplication1 { class Program { static void Main(string[] args) { Thread abortThread = new Thread(AbortMethod); abortThread.Name = "Abort Thread"; abortThread.Start(); Thread.Sleep(1000); try { abortThread.Abort(); } catch { Console.WriteLine("{0} Exception happen in Main Thread", Thread.CurrentThread.Name); Console.WriteLine("{0} Status is:{1} In Main Thread ", Thread.CurrentThread.Name, Thread.CurrentThread.ThreadState); } finally { Console.WriteLine("{0} Status is:{1} In Main Thread ", abortThread.Name, abortThread.ThreadState); } abortThread.Join(); Console.WriteLine("{0} Status is:{1} ", abortThread.Name, abortThread.ThreadState); Console.Read(); } private static void AbortMethod() { try { Thread.Sleep(5000); } catch(Exception e) { Console.WriteLine(e.GetType().Name); Console.WriteLine("{0} Exception happen In Abort Thread", Thread.CurrentThread.Name); Console.WriteLine("{0} Status is:{1} In Abort Thread ", Thread.CurrentThread.Name, Thread.CurrentThread.ThreadState); } finally { Console.WriteLine("{0} Status is:{1} In Abort Thread", Thread.CurrentThread.Name, Thread.CurrentThread.ThreadState); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb31d99da.png) 从运行结果可以看出,调用Abort方法的线程引发的异常类型为ThreadAbortException, 以及异常只会在 调用Abort方法的线程中发生,而不会在主线程中抛出,并且调用Abort方法后线程的状态不是立即改变为Aborted状态,而是从AbortRequested->Aborted。 Interrupt方法: ``` using System; using System.Threading; namespace ConsoleApplication1 { class Program { static void Main(string[] args) { Thread interruptThread = new Thread(AbortMethod); interruptThread.Name = "Interrupt Thread"; interruptThread.Start(); interruptThread.Interrupt(); interruptThread.Join(); Console.WriteLine("{0} Status is:{1} ", interruptThread.Name, interruptThread.ThreadState); Console.Read(); } private static void AbortMethod() { try { Thread.Sleep(5000); } catch(Exception e) { Console.WriteLine(e.GetType().Name); Console.WriteLine("{0} Exception happen In Interrupt Thread", Thread.CurrentThread.Name); Console.WriteLine("{0} Status is:{1} In Interrupt Thread ", Thread.CurrentThread.Name, Thread.CurrentThread.ThreadState); } finally { Console.WriteLine("{0} Status is:{1} In Interrupt Thread", Thread.CurrentThread.Name, Thread.CurrentThread.ThreadState); } } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb31e8e39.png) 从结果中可以得到,调用Interrupt方法抛出的异常为:ThreadInterruptException, 以及当调用Interrupt方法后线程的状态应该是中断的, 但是从运行结果看此时的线程因为了Join,Sleep方法而唤醒了线程,为了进一步解释调用Interrupt方法的线程可以被唤醒, 我们可以在线程执行的方法中运用循环,如果线程可以唤醒,则输出结果中就一定会有循环的部分,然而调用Abort方法线程就直接终止,就不会有循环的部分,下面代码相信大家看后肯定会更加理解两个方法的区别的: ``` using System; using System.Threading; namespace ConsoleApplication2 { class Program { static void Main(string[] args) { Thread thread1 = new Thread(TestMethod); thread1.Start(); Thread.Sleep(100); thread1.Interrupt(); Thread.Sleep(3000); Console.WriteLine("after finnally block, the Thread1 status is:{0}", thread1.ThreadState); Console.Read(); } private static void TestMethod() { for (int i = 0; i < 4; i++) { try { Thread.Sleep(2000); Console.WriteLine("Thread is Running"); } catch (Exception e) { if (e != null) { Console.WriteLine("Exception {0} throw ", e.GetType().Name); } } finally { Console.WriteLine("Current Thread status is:{0} ", Thread.CurrentThread.ThreadState); } } } } } ``` 运行结果为: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb3204222.png) 如果把上面的 thread1.Interrupt();改为 thread1.Abort(); 运行结果为: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb32147f4.png) **二、线程池基础** 首先,创建和销毁线程是一个要耗费大量时间的过程,另外,太多的线程也会浪费内存资源,所以通过Thread类来创建过多的线程反而有损于性能,为了改善这样的问题 ,.net中就引入了线程池。 线程池形象的表示就是存放应用程序中使用的线程的一个集合(就是放线程的地方,这样线程都放在一个地方就好管理了)。CLR初始化时,线程池中是没有线程的,在内部, 线程池维护了一个操作请求队列,当应用程序想执行一个异步操作时,就调用一个方法,就将一个任务放到线程池的队列中,线程池中代码从队列中提取任务,将这个任务委派给一个线程池线程去执行,当线程池线程完成任务时,线程不会被销毁,而是返回到线程池中,等待响应另一个请求。由于线程不被销毁, 这样就可以避免因为创建线程所产生的性能损失。 **注意:通过线程池创建的线程默认为后台线程,优先级默认为Normal.** **三、通过线程池的工作者线程实现异步** **3.1 创建工作者线程的方法** public static bool QueueUserWorkItem (WaitCallback callBack); public static bool QueueUserWorkItem(WaitCallback callback, Object state); 这两个方法向线程池的队列添加一个工作项(work item)以及一个可选的状态数据。然后,这两个方法就会立即返回。 工作项其实就是由callback参数标识的一个方法,该方法将由线程池线程执行。同时写的回调方法必须匹配System.Threading.WaitCallback委托类型,定义为: public delegate void WaitCallback(Object state); 下面演示如何通过线程池线程来实现异步调用: ``` using System; using System.Threading; namespace ThreadPoolUse { class Program { static void Main(string[] args) { // 设置线程池中处于活动的线程的最大数目 // 设置线程池中工作者线程数量为1000,I/O线程数量为1000 ThreadPool.SetMaxThreads(1000, 1000); Console.WriteLine("Main Thread: queue an asynchronous method"); PrintMessage("Main Thread Start"); // 把工作项添加到队列中,此时线程池会用工作者线程去执行回调方法 ThreadPool.QueueUserWorkItem(asyncMethod); Console.Read(); } // 方法必须匹配WaitCallback委托 private static void asyncMethod(object state) { Thread.Sleep(1000); PrintMessage("Asynchoronous Method"); Console.WriteLine("Asynchoronous thread has worked "); } // 打印线程池信息 private static void PrintMessage(String data) { int workthreadnumber; int iothreadnumber; // 获得线程池中可用的线程,把获得的可用工作者线程数量赋给workthreadnumber变量 // 获得的可用I/O线程数量给iothreadnumber变量 ThreadPool.GetAvailableThreads(out workthreadnumber, out iothreadnumber); Console.WriteLine("{0}\n CurrentThreadId is {1}\n CurrentThread is background :{2}\n WorkerThreadNumber is:{3}\n IOThreadNumbers is: {4}\n", data, Thread.CurrentThread.ManagedThreadId, Thread.CurrentThread.IsBackground.ToString(), workthreadnumber.ToString(), iothreadnumber.ToString()); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb3224e17.png) 从结果中可以看出,线程池中的可用的工作者线程少了一个,用去执行回调方法了。 ThreadPool.QueueUserWorkItem(WaitCallback callback,Object state) 方法可以把object对象作为参数传送到回调函数中,使用和ThreadPool.QueueUserWorkItem(WaitCallback callback)的使用和类似,这里就不列出了。 **3.2** **协作式取消** .net Framework提供了**取消操作**的模式, 这个模式是协作式的。为了取消一个操作,首先必须创建一个**System.Threading.CancellationTokenSource**对象。 下面代码演示了协作式取消的使用,主要实现当用户在控制台敲下回车键后就停止数数方法。 ``` using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading; namespace ConsoleApplication3 { class Program { static void Main(string[] args) { ThreadPool.SetMaxThreads(1000, 1000); Console.WriteLine("Main thread run"); PrintMessage("Start"); Run(); Console.ReadKey(); } private static void Run() { CancellationTokenSource cts = new CancellationTokenSource(); // 这里用Lambda表达式的方式和使用委托的效果一样的,只是用了Lambda后可以少定义一个方法。 // 这在这里就是让大家明白怎么lambda表达式如何由委托转变的 ////ThreadPool.QueueUserWorkItem(o => Count(cts.Token, 1000)); ThreadPool.QueueUserWorkItem(callback, cts.Token); Console.WriteLine("Press Enter key to cancel the operation\n"); Console.ReadLine(); // 传达取消请求 cts.Cancel(); } private static void callback(object state) { Thread.Sleep(1000); PrintMessage("Asynchoronous Method Start"); CancellationToken token =(CancellationToken)state; Count(token, 1000); } // 执行的操作,当受到取消请求时停止数数 private static void Count(CancellationToken token,int countto) { for (int i = 0; i < countto; i++) { if (token.IsCancellationRequested) { Console.WriteLine("Count is canceled"); break; } Console.WriteLine(i); Thread.Sleep(300); } Console.WriteLine("Cout has done"); } // 打印线程池信息 private static void PrintMessage(String data) { int workthreadnumber; int iothreadnumber; // 获得线程池中可用的线程,把获得的可用工作者线程数量赋给workthreadnumber变量 // 获得的可用I/O线程数量给iothreadnumber变量 ThreadPool.GetAvailableThreads(out workthreadnumber, out iothreadnumber); Console.WriteLine("{0}\n CurrentThreadId is {1}\n CurrentThread is background :{2}\n WorkerThreadNumber is:{3}\n IOThreadNumbers is: {4}\n", data, Thread.CurrentThread.ManagedThreadId, Thread.CurrentThread.IsBackground.ToString(), workthreadnumber.ToString(), iothreadnumber.ToString()); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb3242df9.png) **四、使用委托实现异步** 通过调用ThreadPool的QueueUserWorkItem方法来来启动工作者线程非常方便,但委托WaitCallback指向的是带有一个参数的无返回值的方法,如果我们实际操作中需要有返回值,或者需要带有多个参数, 这时通过这样的方式就难以实现, 为了解决这样的问题,我们可以通过委托来建立工作这线程, 下面代码演示了使用委托如何实现异步: ``` using System; using System.Threading; namespace Delegate { class Program { // 使用委托的实现的方式是使用了异步变成模型APM(Asynchronous Programming Model) // 自定义委托 private delegate string MyTestdelegate(); static void Main(string[] args) { ThreadPool.SetMaxThreads(1000, 1000); PrintMessage("Main Thread Start"); //实例化委托 MyTestdelegate testdelegate = new MyTestdelegate(asyncMethod); // 异步调用委托 IAsyncResult result = testdelegate.BeginInvoke(null, null); // 获取结果并打印出来 string returndata = testdelegate.EndInvoke(result); Console.WriteLine(returndata); Console.ReadLine(); } private static string asyncMethod() { Thread.Sleep(1000); PrintMessage("Asynchoronous Method"); return "Method has completed"; } // 打印线程池信息 private static void PrintMessage(String data) { int workthreadnumber; int iothreadnumber; // 获得线程池中可用的线程,把获得的可用工作者线程数量赋给workthreadnumber变量 // 获得的可用I/O线程数量给iothreadnumber变量 ThreadPool.GetAvailableThreads(out workthreadnumber, out iothreadnumber); Console.WriteLine("{0}\n CurrentThreadId is {1}\n CurrentThread is background :{2}\n WorkerThreadNumber is:{3}\n IOThreadNumbers is: {4}\n", data, Thread.CurrentThread.ManagedThreadId, Thread.CurrentThread.IsBackground.ToString(), workthreadnumber.ToString(), iothreadnumber.ToString()); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb32583be.png) **五、任务** 同样 任务的引入也是为了解决通过ThreadPool.QueueUserWorkItem中限制的问题, 下面代码演示通过任务来实现异步: **5.1 使用任务来实现异步** ``` using System; using System.Threading; using System.Threading.Tasks; namespace TaskUse { class Program { static void Main(string[] args) { ThreadPool.SetMaxThreads(1000, 1000); PrintMessage("Main Thread Start"); // 调用构造函数创建Task对象, Task task = new Task(n => asyncMethod((int)n), 10); // 启动任务 task.Start(); // 等待任务完成 task.Wait(); Console.WriteLine("The Method result is: "+task.Result); Console.ReadLine(); } private static int asyncMethod(int n) { Thread.Sleep(1000); PrintMessage("Asynchoronous Method"); int sum = 0; for (int i = 1; i < n; i++) { // 如果n太大,使用checked使下面代码抛出异常 checked { sum += i; } } return sum; } // 打印线程池信息 private static void PrintMessage(String data) { int workthreadnumber; int iothreadnumber; // 获得线程池中可用的线程,把获得的可用工作者线程数量赋给workthreadnumber变量 // 获得的可用I/O线程数量给iothreadnumber变量 ThreadPool.GetAvailableThreads(out workthreadnumber, out iothreadnumber); Console.WriteLine("{0}\n CurrentThreadId is {1}\n CurrentThread is background :{2}\n WorkerThreadNumber is:{3}\n IOThreadNumbers is: {4}\n", data, Thread.CurrentThread.ManagedThreadId, Thread.CurrentThread.IsBackground.ToString(), workthreadnumber.ToString(), iothreadnumber.ToString()); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb326b304.png) **5.2 取消任务** 如果要取消任务, 同样可以使用一个CancellationTokenSource对象来取消一个Task. 下面代码演示了如何来取消一个任务: ``` using System; using System.Threading; using System.Threading.Tasks; namespace TaskUse { class Program { static void Main(string[] args) { ThreadPool.SetMaxThreads(1000, 1000); PrintMessage("Main Thread Start"); CancellationTokenSource cts = new CancellationTokenSource(); // 调用构造函数创建Task对象,将一个CancellationToken传给Task构造器从而使Task和CancellationToken关联起来 Task task = new Task(n => asyncMethod(cts.Token, (int)n), 10); // 启动任务 task.Start(); // 延迟取消任务 Thread.Sleep(3000); // 取消任务 cts.Cancel(); Console.WriteLine("The Method result is: " + task.Result); Console.ReadLine(); } private static int asyncMethod(CancellationToken ct, int n) { Thread.Sleep(1000); PrintMessage("Asynchoronous Method"); int sum = 0; try { for (int i = 1; i < n; i++) { // 当CancellationTokenSource对象调用Cancel方法时, // 就会引起OperationCanceledException异常 // 通过调用CancellationToken的ThrowIfCancellationRequested方法来定时检查操作是否已经取消, // 这个方法和CancellationToken的IsCancellationRequested属性类似 ct.ThrowIfCancellationRequested(); Thread.Sleep(500); // 如果n太大,使用checked使下面代码抛出异常 checked { sum += i; } } } catch (Exception e) { Console.WriteLine("Exception is:" + e.GetType().Name); Console.WriteLine("Operation is Canceled"); } return sum; } // 打印线程池信息 private static void PrintMessage(String data) { int workthreadnumber; int iothreadnumber; // 获得线程池中可用的线程,把获得的可用工作者线程数量赋给workthreadnumber变量 // 获得的可用I/O线程数量给iothreadnumber变量 ThreadPool.GetAvailableThreads(out workthreadnumber, out iothreadnumber); Console.WriteLine("{0}\n CurrentThreadId is {1}\n CurrentThread is background :{2}\n WorkerThreadNumber is:{3}\n IOThreadNumbers is: {4}\n", data, Thread.CurrentThread.ManagedThreadId, Thread.CurrentThread.IsBackground.ToString(), workthreadnumber.ToString(), iothreadnumber.ToString()); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb327b395.png) **5.3 任务工厂** 同样可以通过任务工厂TaskFactory类型来实现异步操作。 ``` using System; using System.Threading; using System.Threading.Tasks; namespace TaskFactory { class Program { static void Main(string[] args) { ThreadPool.SetMaxThreads(1000, 1000); Task.Factory.StartNew(() => PrintMessage("Main Thread")); Console.Read(); } // 打印线程池信息 private static void PrintMessage(String data) { int workthreadnumber; int iothreadnumber; // 获得线程池中可用的线程,把获得的可用工作者线程数量赋给workthreadnumber变量 // 获得的可用I/O线程数量给iothreadnumber变量 ThreadPool.GetAvailableThreads(out workthreadnumber, out iothreadnumber); Console.WriteLine("{0}\n CurrentThreadId is {1}\n CurrentThread is background :{2}\n WorkerThreadNumber is:{3}\n IOThreadNumbers is: {4}\n", data, Thread.CurrentThread.ManagedThreadId, Thread.CurrentThread.IsBackground.ToString(), workthreadnumber.ToString(), iothreadnumber.ToString()); } } } ``` 运行结果: ![](https://docs.gechiui.com/gc-content/uploads/sites/kancloud/2016-01-23_56a2eb328e308.png) 讲到这里CLR的工作者线程大致讲完了,希望也篇文章可以让大家对线程又有进一步的理解。在后面的一篇线程系列将谈谈CLR线程池的I/O线程。
';