Java多线程 ReentrantLock互斥锁详解

时间:2021-04-02 人气:0

加锁和解锁

我们来看下ReentrantLock的基本用法

ThreadDomain35类

public class ThreadDomain35 {
  private Lock lock = new ReentrantLock();
  public void testMethod()
  {
    try
    {
      lock.lock();
      for (int i = 0; i < 2; i++)
      {
        System.out.println("ThreadName = " + Thread.currentThread().getName() + ", i = " + i);
      }
    }
    finally
    {
      lock.unlock();
    }
  }
}

线程和main方法

public class MyThread35 extends Thread {

  private ThreadDomain35 td;

  public MyThread35(ThreadDomain35 td)
  {
    this.td = td;
  }

  public void run()
  {
    td.testMethod();
  }

  public static void main(String[] args)
  {
    ThreadDomain35 td = new ThreadDomain35();
    MyThread35 mt0 = new MyThread35(td);
    MyThread35 mt1 = new MyThread35(td);
    MyThread35 mt2 = new MyThread35(td);
    mt0.start();
    mt1.start();
    mt2.start();
  }
}

输出结果

ThreadName = Thread-2, i = 0
ThreadName = Thread-2, i = 1
ThreadName = Thread-0, i = 0
ThreadName = Thread-0, i = 1
ThreadName = Thread-1, i = 0
ThreadName = Thread-1, i = 1

一个线程必须执行完才能执行下一个线程，说明ReentrantLock可以加锁。

ReentrantLock持有的对象监视器和synchronized不同

ThreadDomain37类,methodB用synchronized修饰

public class ThreadDomain37 {
  private Lock lock = new ReentrantLock();
  public void methodA()
  {
    try
    {
      lock.lock();
      System.out.println("MethodA begin ThreadName = " + Thread.currentThread().getName());
      Thread.sleep(5000);
      System.out.println("MethodA end ThreadName = " + Thread.currentThread().getName());
    }
    catch (InterruptedException e)
    {
      e.printStackTrace();
    }
    finally
    {
      lock.unlock();
    }
  }
  public synchronized void methodB()
  {
    System.out.println("MethodB begin ThreadName = " + Thread.currentThread().getName());
    System.out.println("MethodB begin ThreadName = " + Thread.currentThread().getName());
  }
}

MyThread37_0类

public class MyThread37_0 extends Thread {
  private ThreadDomain37 td;
  public MyThread37_0(ThreadDomain37 td)
  {
    this.td = td;
  }
  public void run()
  {
    td.methodA();
  }
}

MyThread37_1类

public class MyThread37_1 extends Thread {
  private ThreadDomain37 td;
  public MyThread37_1(ThreadDomain37 td)
  {
    this.td = td;
  }

  public void run()
  {
    td.methodB();
  }
}

MyThread37_main方法

public class MyThread37_main {
  public static void main(String[] args)
  {
    ThreadDomain37 td = new ThreadDomain37();
    MyThread37_0 mt0 = new MyThread37_0(td);
    MyThread37_1 mt1 = new MyThread37_1(td);
    mt0.start();
    mt1.start();
  }
}

运行结果如下

MethodA begin ThreadName = Thread-0
MethodB begin ThreadName = Thread-1
MethodB begin ThreadName = Thread-1
MethodA end ThreadName = Thread-0

加了synchronized依然是异步执行，说明ReentrantLock和synchronized持有的对象监视器不同。ReentrantLock需要手动加锁和释放锁。

Condition

基本用法

synchronized与wait()和nitofy()/notifyAll()方法可以实现等待/唤醒模型，ReentrantLock同样可以，需要借助Condition的await()和signal/signalAll()，await()释放锁。

ThreadDomain38类

public class ThreadDomain38 {
  private Lock lock = new ReentrantLock();
  private Condition condition = lock.newCondition();
  public void await()
  {
    try
    {
      lock.lock();
      System.out.println("await时间为：" + System.currentTimeMillis());
      condition.await();
      System.out.println("await等待结束");
    }
    catch (InterruptedException e)
    {
      e.printStackTrace();
    }
    finally
    {
      lock.unlock();
    }
  }
  public void signal()
  {
    try
    {
      lock.lock();
      System.out.println("signal时间为：" + System.currentTimeMillis());
      condition.signal();
      System.out.println("signal等待结束");
    }
    finally
    {
      lock.unlock();
    }
  }
}

MyThread38类，线程和main方法

public class MyThread38 extends Thread
{
  private ThreadDomain38 td;

  public MyThread38(ThreadDomain38 td)
  {
    this.td = td;
  }
  public void run()
  {
    td.await();
  }
  public static void main(String[] args) throws Exception
  {
    ThreadDomain38 td = new ThreadDomain38();
    MyThread38 mt = new MyThread38(td);
    mt.start();
    Thread.sleep(3000);
    td.signal();
  }
}

运行结果如下

await时间为：1563505465346
signal时间为：1563505468345
signal等待结束
await等待结束

可以看到，ReentrantLock和Condition实现了等待/通知模型。

一个Lock可以创建多个Condition;

notify()唤醒的线程是随机的，signal()可以有选择性地唤醒。

Condition选择唤醒/等待

现在看一个利用Condition选择等待和唤醒的例子

ThreadDomain47类，定义add和sub方法

public class ThreadDomain47 {
  private final Lock lock = new ReentrantLock();

  private final Condition addCondition = lock.newCondition();

  private final Condition subCondition = lock.newCondition();


  private static int num = 0;
  private List<String> lists = new LinkedList<String>();

  public void add() {
    lock.lock();

    try {
      while(lists.size() == 10) {//当集合已满,则"添加"线程等待
        addCondition.await();
      }

      num++;
      lists.add("add Banana" + num);
      System.out.println("The Lists Size is " + lists.size());
      System.out.println("The Current Thread is " + "增加线程");
      System.out.println("==============================");
      this.subCondition.signal();

    } catch (InterruptedException e) {
      e.printStackTrace();
    } finally {//释放锁
      lock.unlock();
    }
  }
  public void sub() {
    lock.lock();

    try {
      while(lists.size() == 0) {//当集合为空时,"减少"线程等待
        subCondition.await();
      }

      String str = lists.get(0);
      lists.remove(0);
      System.out.println("The Token Banana is [" + str + "]");
      System.out.println("The Current Thread is " + "减少线程");
      System.out.println("==============================");
      num--;
      addCondition.signal();

    } catch (InterruptedException e) {
      e.printStackTrace();
    } finally {
      lock.unlock();
    }
  }
}

MyThread40_0类，增加线程

public class MyThread40_0 implements Runnable {
  private ThreadDomain47 task;
  public MyThread40_0(ThreadDomain47 task) {
    this.task = task;
  }
  @Override
  public void run() {
    task.add();
  }
}

MyThread40_1类，减少线程

public class MyThread40_1 implements Runnable {
  private ThreadDomain47 task;

  public MyThread40_1(ThreadDomain47 task) {
    this.task = task;
  }
  @Override
  public void run() {
    task.sub();
  }
}

main方法，启动线程

public class MyThread40_main {
  public static void main(String[] args) {
    ThreadDomain47 task = new ThreadDomain47();
    Thread t1=new Thread(new MyThread40_0(task));
    Thread t3=new Thread(new MyThread40_0(task));
    Thread t7=new Thread(new MyThread40_0(task));
    Thread t8=new Thread(new MyThread40_0(task));
    Thread t2 = new Thread(new MyThread40_1(task));
    Thread t4 = new Thread(new MyThread40_1(task));
    Thread t5 = new Thread(new MyThread40_1(task));
    Thread t6 = new Thread(new MyThread40_1(task));
    t1.start();
    t2.start();
    t3.start();
    t4.start();
    t5.start();
    t6.start();
    t7.start();
    t8.start();
  }
}

输出结果如下

The Lists Size is 1
The Current Thread is 增加线程
==============================
The Lists Size is 2
The Current Thread is 增加线程
==============================
The Token Banana is [add Banana1]
The Current Thread is 减少线程
==============================
The Token Banana is [add Banana2]
The Current Thread is 减少线程
==============================
The Lists Size is 1
The Current Thread is 增加线程
==============================
The Token Banana is [add Banana1]
The Current Thread is 减少线程
==============================
The Lists Size is 1
The Current Thread is 增加线程
==============================
The Token Banana is [add Banana1]
The Current Thread is 减少线程
==============================

可以看到，lists的数量不会增加太多，也不会减少太多。当集合满，使增加线程等待，唤醒减少线程；当集合空，使减少线程等待，唤醒增加线程。我们用wait()/notify()机制无法实现该效果，这里体现了Condition的强大之处。

ReentrantLock中的方法

公平锁和非公平锁

ReentrantLock可以指定公平锁和非公平锁，公平锁根据线程运行的顺序获取锁，非公平锁则通过抢占获得锁，不按线程运行顺序。synchronized是非公平锁。在ReentrantLock(boolean fair)构造函数传入true/false来指定公平锁/非公平锁。
看个例子

ThreadDomain39类和main方法

public class ThreadDomain39 {
  private Lock lock = new ReentrantLock(true);

  public void testMethod()
  {
    try
    {
      lock.lock();
      System.out.println("ThreadName" + Thread.currentThread().getName() + "获得锁");
    }
    finally
    {
      lock.unlock();
    }
  }

  public static void main(String[] args) throws Exception
  {
    final ThreadDomain39 td = new ThreadDomain39();
    Runnable runnable = new Runnable()
    {
      public void run()
      {
        System.out.println("线程" + Thread.currentThread().getName() + "运行了");
        td.testMethod();
      }
    };
    Thread[] threads = new Thread[5];
    for (int i = 0; i < 5; i++)
      threads[i] = new Thread(runnable);
    for (int i = 0; i < 5; i++)
      threads[i].start();
  }
}

输出结果如下

线程Thread-0运行了
ThreadNameThread-0获得锁
线程Thread-1运行了
线程Thread-2运行了
ThreadNameThread-1获得锁
线程Thread-3运行了
线程Thread-4运行了
ThreadNameThread-2获得锁
ThreadNameThread-3获得锁
ThreadNameThread-4获得锁

可以看到公平锁获得锁的顺序和线程运行的顺序相同。公平锁尽可能地让线程获取锁的顺序和线程运行顺序保持一致，再执行几次，可能不一致。

ReentrantLock构造函数传入false，输出结果如下：

线程Thread-0运行了
线程Thread-2运行了
线程Thread-4运行了
线程Thread-3运行了
ThreadNameThread-0获得锁
线程Thread-1运行了
ThreadNameThread-1获得锁
ThreadNameThread-2获得锁
ThreadNameThread-4获得锁
ThreadNameThread-3获得锁

非公平锁获得锁的顺序和线程运行的顺序不同

getHoldCount()

获取当前线程调用lock()的次数，一般debug使用。

看个例子

public class ThreadDomain40 {
  private ReentrantLock lock = new ReentrantLock();
  public void testMethod1()
  {
    try
    {
      lock.lock();
      System.out.println("testMethod1 getHoldCount = " + lock.getHoldCount());
      testMethod2();
    }
    finally
    {
      lock.unlock();
    }
  }
  public void testMethod2()
  {
    try
    {
      lock.lock();
      System.out.println("testMethod2 getHoldCount = " + lock.getHoldCount());
    }
    finally
    {
      lock.unlock();
    }
  }
  public static void main(String[] args)
  {
    ThreadDomain40 td = new ThreadDomain40();
    td.testMethod1();
  }
}

输出结果如下

testMethod1 getHoldCount = 1
testMethod2 getHoldCount = 2

可以看到，testMethod1()被调用了一次，testMethod2()被调用了两次，ReentrantLock和synchronized一样，锁都是可重入的。

getQueueLength()和isFair()

getQueueLength()获取等待的线程数量，isFair()判断是否是公平锁。

ThreadDomain41类和main方法,Thread.sleep(2000)使第一个线程之后的线程都来不及启动，Thread.sleep(Integer.MAX_VALUE)使线程无法unlock()。

public class ThreadDomain41 {
  public ReentrantLock lock = new ReentrantLock();

  public void testMethod()
  {
    try
    {
      lock.lock();
      System.out.println("ThreadName = " + Thread.currentThread().getName() + "进入方法！");
      System.out.println("是否公平锁？" + lock.isFair());
      Thread.sleep(Integer.MAX_VALUE);
    }
    catch (InterruptedException e)
    {
      e.printStackTrace();
    }
    finally
    {
      lock.unlock();
    }
  }

  public static void main(String[] args) throws InterruptedException
  {
    final ThreadDomain41 td = new ThreadDomain41();
    Runnable runnable = new Runnable()
    {
      public void run()
      {
        td.testMethod();
      }
    };
    Thread[] threads = new Thread[10];
    for (int i = 0; i < 10; i++)
      threads[i] = new Thread(runnable);
    for (int i = 0; i < 10; i++)
      threads[i].start();
    Thread.sleep(2000);
    System.out.println("有" + td.lock.getQueueLength() + "个线程正在等待！");
  }
}

输出结果如下

ThreadName = Thread-1进入方法！
是否公平锁？false
有9个线程正在等待！

ReentrantLock默认是非公平锁，只有一个线程lock()，9个线程在等待。

hasQueuedThread()和hasQueuedThreads()

hasQueuedThread(Thread thread)查询指定线程是否在等待锁，hasQueuedThreads()查询是否有线程在等待锁。
看个例子

ThreadDomain41类和main方法，和上面例子类似，Thread.sleep(Integer.MAX_VALUE); 让线程不释放锁，Thread.sleep(2000);让第一个线程之后的线程都无法启动。

public class ThreadDomain42 extends ReentrantLock {
  public void waitMethod()
  {
    try
    {
      lock();
      Thread.sleep(Integer.MAX_VALUE);
    }
    catch (InterruptedException e)
    {
      e.printStackTrace();
    }
    finally
    {
      unlock();
    }
  }

  public static void main(String[] args) throws InterruptedException
  {
    final ThreadDomain42 td = new ThreadDomain42();
    Runnable runnable = new Runnable()
    {
      public void run()
      {
        td.waitMethod();
      }
    };
    Thread t0 = new Thread(runnable);
    t0.start();
    Thread.sleep(500);
    Thread t1 = new Thread(runnable);
    t1.start();
    Thread.sleep(500);
    Thread t2 = new Thread(runnable);
    t2.start();
    Thread.sleep(500);
    System.out.println("t0 is waiting？" + td.hasQueuedThread(t0));
    System.out.println("t1 is waiting？" + td.hasQueuedThread(t1));
    System.out.println("t2 is waiting？" + td.hasQueuedThread(t2));
    System.out.println("Is any thread waiting？" + td.hasQueuedThreads());
  }
}

输出结果如下

t0 is waiting？false
t1 is waiting？true
t2 is waiting？true
Is any thread waiting？true

t0线程获得了锁，t0没有释放锁，导致t1,t2等待锁。

isHeldByCurrentThread()和isLocked()

isHeldByCurrentThread()判断锁是否由当前线程持有，isLocked()判断锁是否由任意线程持有。
请看示例

ThreadDomain43类和main方法

public class ThreadDomain43 extends ReentrantLock {
  public void testMethod()
  {
    try
    {
      lock();
      System.out.println(Thread.currentThread().getName() + "线程持有了锁！");
      System.out.println(Thread.currentThread().getName() + "线程是否持有锁？" +
          isHeldByCurrentThread());
      System.out.println("是否任意线程持有了锁？" + isLocked());
    } finally
    {
      unlock();
    }
  }
  public void testHoldLock()
  {
    System.out.println(Thread.currentThread().getName() + "线程是否持有锁？" +
        isHeldByCurrentThread());
    System.out.println("是否任意线程持有了锁？" + isLocked());
  }

  public static void main(String[] args)
  {
    final ThreadDomain43 td = new ThreadDomain43();
    Runnable runnable0 = new Runnable()
    {
      public void run()
      {
        td.testMethod();
      }
    };
    Runnable runnable1 = new Runnable()
    {
      public void run()
      {
        td.testHoldLock();
      }
    };
    Thread t0 = new Thread(runnable0);
    Thread t1 = new Thread(runnable1);
    t0.start();
    t1.start();
  }
}

输出结果如下

Thread-0线程持有了锁！
Thread-1线程是否持有锁？false
Thread-0线程是否持有锁？true
是否任意线程持有了锁？true
是否任意线程持有了锁？true

Thread-0线程testMethod方法持有锁，Thread-1线程testHoldLock方法没有lock操作，所以不持有锁。

tryLock()和tryLock(long timeout, TimeUnit unit)

tryLock()有加锁的功能，获得了锁且锁没有被另外一个线程持有，此时返回true，否则返回false，可以有效避免死锁。tryLock(long timeout, TimeUnit unit)表示在给定的时间内获得了锁，锁没有被其他线程持有，且不处于中断状态。返回true,否则返回false；

看个例子

public class MyThread39 {
  public static void main(String[] args) {

    System.out.println("开始");
    final Lock lock = new ReentrantLock();
    new Thread() {
      @Override
      public void run() {
        String tName = Thread.currentThread().getName();
        if (lock.tryLock()) {
          System.out.println(tName + "获取到锁！");
        } else {
          System.out.println(tName + "获取不到锁！");
          return;
        }
        try {
          for (int i = 0; i < 5; i++) {
            System.out.println(tName + ":" + i);
          }
          Thread.sleep(5000);
        } catch (Exception e) {
          System.out.println(tName + "出错了！");
        } finally {
          System.out.println(tName + "释放锁！");
          lock.unlock();
        }

      }
    }.start();

    new Thread() {
      @Override
      public void run() {
        String tName = Thread.currentThread().getName();

        try {
          if (lock.tryLock(1,TimeUnit.SECONDS)) {
            System.out.println(tName + "获取到锁！");
          } else {
            System.out.println(tName + "获取不到锁！");
            return;
          }
        } catch (InterruptedException e) {
          e.printStackTrace();
        }

        try {
          for (int i = 0; i < 5; i++) {
            System.out.println(tName + ":" + i);
          }

        } catch (Exception e) {
          System.out.println(tName + "出错");
        } finally {
          System.out.println(tName + "释放锁！");
          lock.unlock();
        }
      }
    }.start();

    System.out.println("结束");
  }
}

输出结果如下

开始
Thread-0获取到锁！
Thread-0:0
Thread-0:1
Thread-0:2
Thread-0:3
Thread-0:4
结束
Thread-1获取不到锁！
Thread-0释放锁！

Thread-0先获得了锁，且sleep了5秒，导致Thread-1获取不到锁，我们给Thread-1的tryLock设置1秒，一秒内获取不到锁就会返回false。

如果Thread.sleep(0)，那么Thread-0和Thread-1都可以获得锁，园友可以自己试下。

synchronized和ReentrantLock的比较

1.synchronized关键字是语法层面的实现，ReentrantLock要手动lock()和unlock();

2.synchronized是不公平锁，ReentrantLock可以指定是公平锁还是非公平锁；

3.synchronized等待/唤醒机制是随机的，ReentrantLock借助Condition的等待/唤醒机制可以自行选择等待/唤醒；

您可能感兴趣的文章:

加载全部内容