本文主要是介绍java中线程同步的几种常用方式。
CountDownLatch
从字面上理解,CountDownLatch是一个同步等待的锁,根据官方的注释可以看出这其实是一个同步执行工具类。
先看一下官方注释的前两段
/**
* A synchronization aid that allows one or more threads to wait until
* a set of operations being performed in other threads completes.
*
* A {@code CountDownLatch} is initialized with a given count.
* The {@link #await await} methods block until the current count reaches
* zero due to invocations of the {@link #countDown} method, after which
* all waiting threads are released and any subsequent invocations of
* {@link #await await} return immediately. This is a one-shot phenomenon
* -- the count cannot be reset. If you need a version that resets the
* count, consider using a {@link CyclicBarrier}.
翻译一下就是:
/**
* CountDownLatch是一个,允许一个或多个线程,
* 等待其他线程中执行的一组操作完成的,同步辅助工具。
*
* CountDownLatch用给定的计数进行初始化。
* 当线程点用await方法后被阻塞,直到当前计数由于其他线程调用countDown()方法而达到零,
* 此后所有等待线程被放,并且任何后续调用await立即返回。
* 这是一次性的操作,计数无法重置。
* 如果您需要重置计数的版本,请考虑使用CyclicBarrier。
* /
解释的很清楚,不在赘述,接着看一下官方提供的伪代码案例
官方案例一
class Driver { // ...
void main() throws InterruptedException {
CountDownLatch startSignal = new CountDownLatch(1);
CountDownLatch doneSignal = new CountDownLatch(N);
for (int i = 0; i < N; ++i) // create and start threads
new Thread(new Worker(startSignal, doneSignal)).start();
doSomethingElse(); // don't let run yet
startSignal.countDown(); // let all threads proceed
doSomethingElse();
doneSignal.await(); // wait for all to finish
}
}
class Worker implements Runnable {
private final CountDownLatch startSignal;
private final CountDownLatch doneSignal;
Worker(CountDownLatch startSignal, CountDownLatch doneSignal) {
this.startSignal = startSignal;
this.doneSignal = doneSignal;
}
public void run() {
try {
startSignal.await();
doWork();
doneSignal.countDown();
} catch (InterruptedException ex) {} // return;
}
void doWork() { ... }
}
这个案例很直白的说明了,CountDownLatch可以让多个线程同时初始化完成后等待,直到主线程要求他们开始执行为止,并且当主线程调用await()之后阻塞直到所有的线程调用countDown()将计数减为0,主线程再次唤醒执行后序操作。
当然这样还有一些其他的注意点,譬如子线程被中断或者子线程的耗时操作很长导致主线程一直阻塞等问题。
官方案例二
class Driver2 { // ...
void main() throws InterruptedException {
CountDownLatch doneSignal = new CountDownLatch(N);
Executor e = ...
for (int i = 0; i < N; ++i) // create and start threads
e.execute(new WorkerRunnable(doneSignal, i));
doneSignal.await(); // wait for all to finish
}
}
class WorkerRunnable implements Runnable {
private final CountDownLatch doneSignal;
private final int i;
WorkerRunnable(CountDownLatch doneSignal, int i) {
this.doneSignal = doneSignal;
this.i = i;
}
public void run() {
try {
doWork(i);
doneSignal.countDown();
} catch (InterruptedException ex) {} // return;
}
void doWork() { ... }
}
}
这个案例是说,当一个问题需要被分成n份进行处理时,将他们用线程池来执行,并让主线程等待。当然官方注释里还说了,如果需要反复用这种形式来执行一些问题时可以考虑使用CyclicBarrier来代替CountDownLatch,因为CountDownLatch是一次性的计数器无法重置。
CyclicBarrier
字面意思:可循环使用的栅栏。主要的作用也是让指定个数的线程到达目标位置后进入等到状态,等所有的线程都到到目标位置后同时开始执行。
构造方法有2个
- CyclicBarrier(int parties),其中parties指等待的线程数目,当await线程数达到parties时,线程同时开始执行。
- CyclicBarrier(int parties, Runnable barrierAction),第二个参数指所有线程达到后执行的操作。
通过第二个构造方法也可以实现CountDownLatch功能,当然这不是CyclicBarrier的目的
再来看一下到达目标位置时的等待方法,有2个重载方法
- await(),这个没什么可说的,到达指定位置后等待
- await(long timeout, TimeUnit unit),这个指到到指定位置后等待一段时间,如果超时则继续执行后序操作。
现在来看2个例子说明一下使用CyclicBarrier可能出现的问题
CyclicBarrier例一
public class CyclicBarrierTest {
public static void main(String[] args) {
try {
final int Num = 5;
CyclicBarrier cyclicBarrier = new CyclicBarrier(Num);
for (int i = 0; i < Num - 1; i++) {
new Thread(new RunnableOne(cyclicBarrier)).start();
}
Thread thread = new Thread(new RunnableTwo(cyclicBarrier));
thread.start();
Thread.sleep(2000);
thread.interrupt();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
private static class RunnableOne implements Runnable {
CyclicBarrier mCyclicBarrier;
RunnableOne(CyclicBarrier cyclicBarrier) {
mCyclicBarrier = cyclicBarrier;
}
@Override
public void run() {
try {
System.out.println("wait in barrier");
mCyclicBarrier.await();
System.out.println("finish");
} catch (InterruptedException | BrokenBarrierException e) {
e.printStackTrace();
}
}
}
private static class RunnableTwo implements Runnable {
CyclicBarrier mCyclicBarrier;
RunnableTwo(CyclicBarrier cyclicBarrier) {
mCyclicBarrier = cyclicBarrier;
}
@Override
public void run() {
try {
System.out.println("wait in barrier");
Thread.sleep(5000);
mCyclicBarrier.await();
System.out.println("finish");
} catch (InterruptedException | BrokenBarrierException e) {
e.printStackTrace();
}
}
}
}
打印结果如下:
wait in barrier
wait in barrier
wait in barrier
wait in barrier
wait in barrier
java.lang.InterruptedException: sleep interrupted
at java.lang.Thread.sleep(Native Method)
at com.jxx.myjavatest.CyclicBarrierTest$RunnableTwo.run(CyclicBarrierTest.java:65)
at java.lang.Thread.run(Thread.java:748)
这个例子的意图也很简单,启动4个RunnableOne,随后启动1个RunnableTwo,在所有线程都await()之前其中一个线程被中断了,因为没有都await()成功,其他4个线程就一直阻塞。
这就提醒我们,要在抛出异常后及时处理,至少也要让其他线程能正常执行下去。
CyclicBarrier例二
public class CyclicBarrierTest {
public static void main(String[] args) {
final int Num = 5;
CyclicBarrier cyclicBarrier = new CyclicBarrier(Num);
for (int i = 0; i < Num - 1; i++) {
new Thread(new RunnableOne(cyclicBarrier)).start();
}
Thread thread = new Thread(new RunnableTwo(cyclicBarrier));
thread.start();
}
private static class RunnableOne implements Runnable {
CyclicBarrier mCyclicBarrier;
RunnableOne(CyclicBarrier cyclicBarrier) {
mCyclicBarrier = cyclicBarrier;
}
@Override
public void run() {
try {
System.out.println("wait in barrier");
Thread.sleep(5000);
mCyclicBarrier.await();
System.out.println("finish");
} catch (InterruptedException | BrokenBarrierException e) {
e.printStackTrace();
}
}
}
private static class RunnableTwo implements Runnable {
CyclicBarrier mCyclicBarrier;
RunnableTwo(CyclicBarrier cyclicBarrier) {
mCyclicBarrier = cyclicBarrier;
}
@Override
public void run() {
try {
System.out.println("wait in barrier");
mCyclicBarrier.await(2000, TimeUnit.MILLISECONDS);
System.out.println("finish");
} catch (InterruptedException | BrokenBarrierException | TimeoutException e) {
e.printStackTrace();
}
}
}
}
打印如下:
wait in barrier
wait in barrier
wait in barrier
wait in barrier
wait in barrier
java.util.concurrent.TimeoutException
at java.util.concurrent.CyclicBarrier.dowait(CyclicBarrier.java:257)
at java.util.concurrent.CyclicBarrier.await(CyclicBarrier.java:435)
at com.jxx.myjavatest.CyclicBarrierTest$RunnableTwo.run(CyclicBarrierTest.java:61)
at java.lang.Thread.run(Thread.java:748)
java.util.concurrent.BrokenBarrierException
at java.util.concurrent.CyclicBarrier.dowait(CyclicBarrier.java:207)
at java.util.concurrent.CyclicBarrier.await(CyclicBarrier.java:362)
at com.jxx.myjavatest.CyclicBarrierTest$RunnableOne.run(CyclicBarrierTest.java:40)
at java.lang.Thread.run(Thread.java:748)
java.util.concurrent.BrokenBarrierException
at java.util.concurrent.CyclicBarrier.dowait(CyclicBarrier.java:207)
at java.util.concurrent.CyclicBarrier.await(CyclicBarrier.java:362)
at com.jxx.myjavatest.CyclicBarrierTest$RunnableOne.run(CyclicBarrierTest.java:40)
at java.lang.Thread.run(Thread.java:748)
java.util.concurrent.BrokenBarrierException
at java.util.concurrent.CyclicBarrier.dowait(CyclicBarrier.java:207)
at java.util.concurrent.CyclicBarrier.await(CyclicBarrier.java:362)
at com.jxx.myjavatest.CyclicBarrierTest$RunnableOne.run(CyclicBarrierTest.java:40)
at java.lang.Thread.run(Thread.java:748)
java.util.concurrent.BrokenBarrierException
at java.util.concurrent.CyclicBarrier.dowait(CyclicBarrier.java:207)
at java.util.concurrent.CyclicBarrier.await(CyclicBarrier.java:362)
at com.jxx.myjavatest.CyclicBarrierTest$RunnableOne.run(CyclicBarrierTest.java:40)
at java.lang.Thread.run(Thread.java:748)
这里模拟了一个await()超时的异常,可以看到在抛出异常后需要我们自己处理后期的事物。同时某一个线程抛出超时异常后,其他线程再次到达会抛出BrokenBarrierException异常,防止继续等待。
Semaphore
其实Semaphore不该放到这里讲,因为Semaphore类似于Lock的存在,是对资源或者线程的一种控制,但是这篇博文主要讲了线程的等待唤起,信号量放这里讲问题也不大。
官方的说法是信号量通常用来限制线程的数量,而不是控制访问一些(物理或逻辑)资源。用法也非常简单,使用前先acquire()获取许可,在获取许可过程中,是线程是被阻塞的,使用完毕release()许可即可。这点类似于Lock,不同的是Semaphore的acquire()可以被允许多次。
Semaphore有两个构造方法,可以指定Semaphore获取是公平的还是非公平的,默认是非公平
看这里,举个栗子:
public class SemaphoreTest {
public static void main(String[] args) {
CountDownLatch startLatch = new CountDownLatch(1);
Semaphore semaphore = new Semaphore(3);
for (int i = 0; i < 10; i++) {
new Thread(new MyRunnable(startLatch, semaphore)).start();
}
startLatch.countDown();
}
private static class MyRunnable implements Runnable {
final CountDownLatch mCountDownLatch;
final Semaphore mSemaphore;
MyRunnable(CountDownLatch countDownLatch, Semaphore semaphore) {
mCountDownLatch = countDownLatch;
mSemaphore = semaphore;
}
@Override
public void run() {
try {
mCountDownLatch.await();
mSemaphore.acquire();
System.out.println(Thread.currentThread().getName() + " acquire success");
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
} finally {
mSemaphore.release();
}
}
}
}
打印如下:
Thread-0 acquire success
Thread-1 acquire success
Thread-9 acquire success
Thread-3 acquire success
Thread-2 acquire success
Thread-4 acquire success
Thread-6 acquire success
Thread-7 acquire success
Thread-5 acquire success
Thread-8 acquire success
可以看出这是默认的非公平锁的情况,再来看一下公平锁的情况
public class SemaphoreTest {
public static void main(String[] args) {
Semaphore semaphore = new Semaphore(3, true);
for (int i = 0; i < 10; i++) {
new Thread(new MyRunnable(semaphore)).start();
}
}
private static class MyRunnable implements Runnable {
final Semaphore mSemaphore;
MyRunnable(Semaphore semaphore) {
mSemaphore = semaphore;
}
@Override
public void run() {
try {
mSemaphore.acquire();
System.out.println(Thread.currentThread().getName() + " acquire success");
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
} finally {
mSemaphore.release();
}
}
}
}
打印如下
Thread-0 acquire success
Thread-1 acquire success
Thread-2 acquire success
Thread-3 acquire success
Thread-4 acquire success
Thread-5 acquire success
Thread-6 acquire success
Thread-7 acquire success
Thread-8 acquire success
Thread-9 acquire success
当然这里肯定有读者想了,直接将Semaphore置为true公平锁的情况就好了,何必去掉CountDownLatch呢。
这里需要注意下,虽然你Semaphore是公平,但是CountDownLatch到点之后唤起线程的顺序是随机的,并不一定就是线程入队的顺序唤起。
线程的join()
jion方法的作用是让主线程阻塞等待子线程完成,当然有几个前提条件,下面细说。
join方法有三个重载的版本
- final void join(); //一直等待到join的线程执行完毕
- final synchronized void join(long millis); //等待指定时间后继续执行
- final synchronized void join(long millis, int nanos); 同上,时间处理了一下
第一个和第三个最后其实调用的都是第二个重载方法,我们来看一下源码
public final synchronized void join(long millis)
throws InterruptedException {
long base = System.currentTimeMillis();
long now = 0;
if (millis < 0) {
throw new IllegalArgumentException("timeout value is negative");
}
if (millis == 0) {
while (isAlive()) {
wait(0);
}
} else {
while (isAlive()) {
long delay = millis - now;
if (delay <= 0) {
break;
}
wait(delay);
now = System.currentTimeMillis() - base;
}
}
}
直接看最后的while循环,可以看到,调用这个方法,其实是调用Object提供的wait(long timeout)让主线程阻塞而已。有几个注意点
- 子线程如果已经销毁,则直接跳过等待
- join(long millis) 是一个同步方位,意味着要想调用此方法需要先获取到子线程的实例对象锁
来看一个例子,验证一下第二点:
public class JoinTest {
public static void main(String[] args) {
final Thread thread = new Thread(new Runnable() {
@Override
public void run() {
try {
Thread.sleep(6000);
System.out.println("4---" + System.currentTimeMillis());
} catch (InterruptedException e) {
e.printStackTrace();
}
}
});
thread.start();
try {
System.out.println("1---" + System.currentTimeMillis());
new Thread(new MyRunnable(thread)).start();
System.out.println("2---" + System.currentTimeMillis());
thread.join(2000);
System.out.println("3---" + System.currentTimeMillis());
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println("finish " + System.currentTimeMillis());
}
private static class MyRunnable implements Runnable {
final Object mObject;
MyRunnable(Object object) {
mObject = object;
}
@Override
public void run() {
synchronized (mObject) {
try {
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
}
}
}
打印如下:
1---1525529803445
2---1525529803446
3---1525529807449
finish 1525529807449
4---1525529809445
可以很清晰的看到,打印完1之后立即打印了2,但是2和3之间打相差了4秒,原因就在join之前需要先获取thread的锁对象,但是需要MyRunnable释放锁之后才能执行。
总结
好了,又到总结的时间了。
- CountDownLatch相对于CyclicBarrier侧重点是,等待其他线程操作完成后主线程在继续后续的操作
- CyclicBarrier相对于CountDownLatch侧重点是,所有的线程操作完成后等待一起继续后续操作。
- CountDownLatch不能重置状态,CyclicBarrier可以重置后多次利用
- CountDownLatch和CyclicBarrier抛出异常后都需要妥善处理
- Semaphore于Lock类似,主要用于线程的访问控制,构造时可以指定是否是公平竞争
- thread.join()主要是让主线程等待子线程执行完毕,有个注意点就是join()执行之前需要获取到子线程的实例对象锁。