AbstractQueuedSynchronizer详解（二）——CountDownLatch源码分析

在AbstractQueuedSynchronizer详解（一）——分析ReentrantLock源码中介绍了其内部结构以及独占模式的一个例子：ReentrantLock。本篇文章介绍其共享模式的一个例子：CountDownLatch。

CountDownLatch的使用

CountDownLatch是一个工具类，用于使一个或多个线程等待另一系列线程完成操作，也就是说一些线程在另外一些完成之后才能继续执行，类似线性。
初始CountDownLatch时需要提供一个count参数，await方法将会使线程阻塞，直到这个值变为0，countDown()方法可以使count值减1，一旦到达0后，调用await阻塞的线程将会得到释放。
CountDown一般有两个用法。

一个线程等待多个线程

一个线程等待多个线程完成后再执行，比如一个不限时的考试，10个学生考试，交卷时间不同，当10个学生都交卷后，老师才认为考试结束。下面是该例子：

public class ExamDemo {

    public static class Student implements Runnable {

        private int num;//学号
        private CountDownLatch countDownLatch;

        public Student(int num, CountDownLatch countDownLatch) {
            this.num = num;
            this.countDownLatch = countDownLatch;
        }

        @Override
        public void run() {

            Random random = new Random();

            try {
                TimeUnit.SECONDS.sleep(random.nextInt(10) + 1);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }

            System.out.println("学生" + num + "交卷了");
            //完成工作，将count-1
            countDownLatch.countDown();

        }
    }

    public static class Teacher implements Runnable {

        private CountDownLatch countDownLatch;

        public Teacher(CountDownLatch countDownLatch) {
            this.countDownLatch = countDownLatch;
        }

        @Override
        public void run() {

            System.out.println("考试开始，不限时间!");

            try {
                //等待count变为0
                countDownLatch.await();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }

            System.out.println("考试结束");

        }
    }

    public static void main(String[] args) {

        Executor executor = Executors.newCachedThreadPool();

        CountDownLatch countDownLatch = new CountDownLatch(10);

        executor.execute(new Teacher(countDownLatch));

        for (int i = 1; i <= 10; i++)
            executor.execute(new Student(i, countDownLatch));

    }
}

多个线程等待一个线程

再举个例子，比如几个司机在等红灯，而一旦变成了绿灯，几个司机就可以都通过了。

public class TrafficDemo {

    public static class Driver implements Runnable {

        private int num;
        private CountDownLatch countDownLatch;

        public Driver(int num, CountDownLatch countDownLatch) {
            this.num = num;
            this.countDownLatch = countDownLatch;
        }

        @Override
        public void run() {

            System.out.println("司机" + num + "于红灯前等待");
            try {
                countDownLatch.await();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
            System.out.println("司机" + num + "于绿灯通过");

        }
    }

    public static class TrafficLight implements Runnable {

        private CountDownLatch countDownLatch;

        public TrafficLight(CountDownLatch countDownLatch) {
            this.countDownLatch = countDownLatch;
        }

        @Override
        public void run() {

            try {
                TimeUnit.SECONDS.sleep(5);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }

            System.out.println("绿灯行");
            countDownLatch.countDown();


        }
    }

    public static void main(String[] args) {

        Executor executor = Executors.newCachedThreadPool();
        CountDownLatch countDownLatch = new CountDownLatch(1);
        for (int i = 1; i < 10; i++)
            executor.execute(new Driver(i, countDownLatch));
        executor.execute(new TrafficLight(countDownLatch));

    }

}

CountDownLatch源码解析

CountDownLatch初始化

CountDownLatch只有一个构造方法，如下：

 public CountDownLatch(int count) {
        if (count < 0) throw new IllegalArgumentException("count < 0");
        this.sync = new Sync(count);
    }

我们很容易知道Sync是一个继承自AQS的内部类，它负责CountDownLatch的同步事件。下面是Sync的定义：

/**
     * Synchronization control For CountDownLatch.
     * Uses AQS state to represent count.
     */
    private static final class Sync extends AbstractQueuedSynchronizer {
        private static final long serialVersionUID = 4982264981922014374L;

        //AQS的资源即设定的count
        Sync(int count) {
            setState(count);
        }

        int getCount() {
            return getState();
        }

        protected int tryAcquireShared(int acquires) {
            return (getState() == 0) ? 1 : -1;
        }

        protected boolean tryReleaseShared(int releases) {
            // Decrement count; signal when transition to zero
            for (;;) {
                int c = getState();
                if (c == 0)
                    return false;
                int nextc = c-1;
                if (compareAndSetState(c, nextc))
                    return nextc == 0;
            }
        }
    }

await方法

await方法用于将线程阻塞，实现如下：

public void await() throws InterruptedException {
        sync.acquireSharedInterruptibly(1);
    }

可以看到调用了acquireSharedInterruptibly方法，该方法在AQS中，实现如下：

public final void acquireSharedInterruptibly(int arg)
            throws InterruptedException {
        //线程被interrupt了，抛出异常
        if (Thread.interrupted())
            throw new InterruptedException();
        //尝试获取arg个资源，如果小于0表示获取失败
        if (tryAcquireShared(arg) < 0)
            doAcquireSharedInterruptibly(arg);
    }

从上面的方法可以看到，首先尝试获取资源，如果有资源，那么调用doAcquireSharedInterruptibly方法，tryAcquireShared方法是在Sync中的。从上面可以看到，如果getState为0则该线程不需要加入到等待队列中，而如果不等于0，则需要将该线程加入到等待队里中。
如果该线程需要加入到等待队列中，那么就调用doAcquireSharedInterruptibly方法，如下：

private void doAcquireSharedInterruptibly(int arg)
        throws InterruptedException {
        //将该线程加入到等待队列中
        final Node node = addWaiter(Node.SHARED);
        boolean failed = true;
        try {
            for (;;) {
                //得到前驱节点
                final Node p = node.predecessor();
                //如果前驱节点是头节点
                if (p == head) {
                    //尝试获取资源
                    int r = tryAcquireShared(arg);
                    //getState为0，说明count为0了，等待的线程可以执行了
                    if (r >= 0) {
                        setHeadAndPropagate(node, r);
                        p.next = null; // help GC
                        failed = false;
                        return;
                    }
                }
                //检查是否需要将当前线程挂起
                if (shouldParkAfterFailedAcquire(p, node) &&
                    parkAndCheckInterrupt())
                    throw new InterruptedException();
            }
        } finally {
            if (failed)
                cancelAcquire(node);
        }
    }

在前一篇文章中，我们知道队列的结构如下：

其中Head是个空节点，所以一旦前驱节点是头节点，并且要是count变为了0，那么就调用setHeadAndPropagate方法，然后释放头节点。setHeadAndPropagate方法主要完成两件事，第一更改头节点，第二由于此时等待线程可以执行了，将该事件传播给后面的节点，实现如下：

private void setHeadAndPropagate(Node node, int propagate) {

        Node h = head; // Record old head for check below
        //更改头节点
        setHead(node);
        /*
         * Try to signal next queued node if:
         *   Propagation was indicated by caller,
         *     or was recorded (as h.waitStatus either before
         *     or after setHead) by a previous operation
         *     (note: this uses sign-check of waitStatus because
         *      PROPAGATE status may transition to SIGNAL.)
         * and
         *   The next node is waiting in shared mode,
         *     or we don't know, because it appears null
         *
         * The conservatism in both of these checks may cause
         * unnecessary wake-ups, but only when there are multiple
         * racing acquires/releases, so most need signals now or soon
         * anyway.
         */
         //如果需要传播
        if (propagate > 0 || h == null || h.waitStatus < 0 ||
            (h = head) == null || h.waitStatus < 0) {
            Node s = node.next;
            if (s == null || s.isShared())
                doReleaseShared();
        }
    }

可以看到如果需要传播的话，将会调用doReleasShared方法,方法如下：

private void doReleaseShared() {
        for (;;) {
            Node h = head;
            if (h != null && h != tail) {
                int ws = h.waitStatus;
                if (ws == Node.SIGNAL) {
                    if (!compareAndSetWaitStatus(h, Node.SIGNAL, 0)) //重置状态，成功后，将后继节点唤醒
                        continue;            // loop to recheck cases
                    unparkSuccessor(h);
                }
                else if (ws == 0 &&
                         !compareAndSetWaitStatus(h, 0, Node.PROPAGATE))
                    continue;                // loop on failed CAS
            }
            if (h == head)                   // loop if head changed
                break;
        }
    }

至此，await方法就使线程挂起了，下面再分析countDown方法。

countDown方法

countDown方法用于将count值减1,如果count变为0则释放所有等待线程，方法的实现如下：

public void countDown() {
        sync.releaseShared(1);
    }

releaseShared方法如下:

 public final boolean releaseShared(int arg) {
        //尝试释放共享资源，如果需要释放节点，则释放节点
        if (tryReleaseShared(arg)) {
            doReleaseShared();
            return true;
        }
        return false;
    }

tryReleaseShared需要在子类中实现，Sync中的实现如下：

protected boolean tryReleaseShared(int releases) {
            // Decrement count; signal when transition to zero
            for (;;) {
                int c = getState();
                if (c == 0)
                    return false;
                int nextc = c-1;
                if (compareAndSetState(c, nextc))
                    return nextc == 0;
            }
        }

一开始设置了count，每次countDown会使AQS的state-1，一旦变成0，则返回true。而如果变成了0后，又调用countDown则无效。doReleaseShared方法用于释放节点，这个在上面已经介绍了。

总结

CountDownLatch中使用了AQS，AQS的资源就是CountDownLatch的初始容量设置，即限定放行的线程数，采用了AQS的共享模式。
与CycilBarrier的区别在于CountDownLatch在state为0后放行的线程只会执行一次，而CycliBarrier放行后的线程可以多次执行。