xzkjgw 2020-01-17
CountDownLatch允许一个或多个线程等待直到在其他线程中执行的一组操作完成的同步辅助。简单来说,就是CountDownLatch内部维护了一个计数器,每个线程完成自己的操作之后都会将计数器减一,然后会在计数器的值变为0之前一直阻塞,直到计数器的值变为0.
这个例子主要演示了,如何利用CountDownLatch去协调多个线程同时开始运行。这个时候的CountDownLatch中的计数器的现实含义是等待创建的线程个数,每个线程在开始任务之前都会调用await()方法阻塞,直到所有线程都创建好,每当一个线程创建好后,都会提交调用countDown()方法将计数器的值减一(代表待创建的线程数减一)。
public static void main(String[] args) { Test countDownLatchTest=new Test(); countDownLatchTest.runThread(); } //计数器为10,代表有10个线程等待创建 CountDownLatch countDownLatch=new CountDownLatch(10); /** * 创建一个线程 * @return */ private Thread createThread(int i){ Thread thread=new Thread(new Runnable() { @Override public void run() { try { //在此等待,直到计数器变为0 countDownLatch.await(); System.out.println("thread"+Thread.currentThread().getName()+"准备完毕"+System.currentTimeMillis()); }catch (InterruptedException e){ e.printStackTrace(); } } }); thread.setName("thread-"+i); return thread; } public void runThread(){ ExecutorService executorService= Executors.newFixedThreadPool(10); try { for(int i=0;i<10;i++){ Thread.sleep(100); executorService.submit(createThread(i)); //一个线程创建好了,待创建的线程数减一 countDownLatch.countDown(); } }catch (InterruptedException e){ e.printStackTrace(); } }
下面我们就以这个例子,来解释源码:
从锁的分类上来讲,CountDownLatch其实是一个”共享锁“。还有一个需要注意的是CountDownLath是响应中断的,如果线程在对锁进行操作的期间发生了中断,会直接抛出InterruptedException。
刚才我们也提到了,CountDownLatch中一个非常重要的东西就是计数器。那么我们首先需要分析的就是源码中哪个部分充当了计数器的角色。
我们通过构造方法来查看:
我们的代码CountDownLatch countDownLatch=new CountDownLatch(10);
背后实际上是调用了下面这个方法:
public CountDownLatch(int count) { if (count < 0) throw new IllegalArgumentException("count < 0"); this.sync = new Sync(count); }
而这个Sync的实例化又做了什么工作呢?
Sync(int count) { setState(count); //就是修改了AQS中的state值 }
现在已经解决了我们的第一个问题,实际上AQS中的state充当了计数器。
public void await() throws InterruptedException { sync.acquireSharedInterruptibly(1); }
void acquireSharedInterruptibly(int arg)
的实现如下public final void acquireSharedInterruptibly(int arg) throws InterruptedException { if (Thread.interrupted()) //如果线程中断了,则抛异常。 //证明了之前所说的CountDownLatch是会响应中断的 throw new InterruptedException(); if (tryAcquireShared(arg) < 0) doAcquireSharedInterruptibly(arg); }
tryAcquireShared(arg)
protected int tryAcquireShared(int acquires) { return (getState() == 0) ? 1 : -1; }
doAcquireSharedInterruptibly(arg)
private void doAcquireSharedInterruptibly(int arg) throws InterruptedException { final Node node = addWaiter(Node.SHARED); boolean failed = true; try { for (;;) { final Node p = node.predecessor(); //获取当前节点的前驱节点 if (p == head) {//前一个节点是头节点 int r = tryAcquireShared(arg); //去看一看state是否为0,步骤3分析过 if (r >= 0) { //如果state目前为0,就出队 setHeadAndPropagate(node, r); p.next = null; // help GC failed = false; return; } } if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt()) //进入阻塞队列阻塞,如果发生中断,则抛异常 throw new InterruptedException(); } } finally { if (failed) cancelAcquire(node); }
CountDownLatch的await方法比其它几个锁的实现简单得多。不过需要注意的一点就是CountDownLatch是会响应中断的,这一点在源码中也有多处体现。
public void countDown() { sync.releaseShared(1); }
boolean releaseShared(int arg)
的具体实现如下:public final boolean releaseShared(int arg) { if (tryReleaseShared(arg)) { doReleaseShared(); return true; } return false; }
tryReleaseShared(arg)
方法的具体实现如下:protected boolean tryReleaseShared(int releases) { // Decrement count; signal when transition to zero for (;;) {//自旋 int c = getState(); if (c == 0)//计数器已经都是0了,当然会释放失败咯 return false; int nextc = c-1;//释放后,计数器减一 if (compareAndSetState(c, nextc))//CAS修改计数器 return nextc == 0; } }
这个方法就是去尝试直接修改state的值。如果state的修改成功,且修改后的state值为0,就会返回true。就会执行doReleaseShared();
方法。
doReleaseShared();
的实现如下,它的作用就是state为0的时候,去唤醒等待队列中的线程。private void doReleaseShared() { /* * Ensure that a release propagates, even if there are other * in-progress acquires/releases. This proceeds in the usual * way of trying to unparkSuccessor of head if it needs * signal. But if it does not, status is set to PROPAGATE to * ensure that upon release, propagation continues. * Additionally, we must loop in case a new node is added * while we are doing this. Also, unlike other uses of * unparkSuccessor, we need to know if CAS to reset status * fails, if so rechecking. */ for (;;) { //自旋 Node h = head; if (h != null && h != tail) { int ws = h.waitStatus; if (ws == Node.SIGNAL) { if (!compareAndSetWaitStatus(h, Node.SIGNAL, 0)) 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; } }
现在基本源码已经分析完毕了,只要理解了AQS和CountDownLatch的计数器到底是什么,就能够很好的理解CountDownLatch的原理了。