在Java中可以使用synchronized关键字来修饰方法、静态方法和代码块,synchronized能够隐式的获取和释放锁,从而保证在同一时刻,只有一个线程在方法或代码块中。
public class SynchronizedDemo {
public static int num = 0;
public static void main(String[] args) {
SynchronizedDemo demo = new SynchronizedDemo();
new Thread(new DemoRunnable(demo), "T1").start();
new Thread(new DemoRunnable(demo), "T2").start();
}
public synchronized void test(){
for(int i=0; i<3; i++) {
try {
Thread.sleep(100);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println(Thread.currentThread().getName()+" num = "+num++);
}
}
static class DemoRunnable implements Runnable{
private SynchronizedDemo demo;
public DemoRunnable(SynchronizedDemo demo){
this.demo = demo;
}
@Override
public void run() {
demo.test();
}
}
}
对test()使用synchronized修饰时,输出为:
T1 num = 0
T1 num = 1
T1 num = 2
T1 num = 3
T1 num = 4
T2 num = 5
T2 num = 6
T2 num = 7
T2 num = 8
T2 num = 9
当去掉synchronized时,输出为:
T2 num = 1
T1 num = 0
T1 num = 2
T2 num = 3
T1 num = 5
T2 num = 4
T1 num = 6
T2 num = 6
T2 num = 7
T1 num = 7
当synchronized修饰方法时,锁的对象是类的当前实例。
我们对上面的代码中的test()进行改写:
public synchronized void test(){
System.out.println(Thread.currentThread().getName()+" START time= "+new Date());
try {
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println(Thread.currentThread().getName()+" END time= "+new Date());
}
两个线程使用同一个 demo 对象,查看输出可知只有等待其中一个线程结束后,另一个线程才执行test():
public static void main(String[] args) {
SynchronizedDemo demo = new SynchronizedDemo();
new Thread(new DemoRunnable(demo), "T1").start();
new Thread(new DemoRunnable(demo), "T2").start();
}
输出为:
T1 START time= Mon Aug 14 13:02:40 CST 2017
T1 END time= Mon Aug 14 13:02:42 CST 2017
T2 START time= Mon Aug 14 13:02:42 CST 2017
T2 END time= Mon Aug 14 13:02:44 CST 2017
两个线程分别使用一个 demo对象,则两个线程的开始和结束互不影响:
public static void main(String[] args) {
new Thread(new DemoRunnable(new SynchronizedDemo()), "T1").start();
new Thread(new DemoRunnable(new SynchronizedDemo()), "T2").start();
}
输出为:
T1 START time= Mon Aug 14 12:58:58 CST 2017
T2 START time= Mon Aug 14 12:58:58 CST 2017
T2 END time= Mon Aug 14 12:59:00 CST 2017
T1 END time= Mon Aug 14 12:59:00 CST 2017
当synchronized修饰静态方法时,锁的对象是当前类的class对象
我们将上面的test() 改为静态方法,然后执行上面的测试代码,发现两个线程先后执行。
两种测试方法输出都为:
T1 START time= Mon Aug 14 13:04:06 CST 2017
T1 END time= Mon Aug 14 13:04:08 CST 2017
T2 START time= Mon Aug 14 13:04:08 CST 2017
T2 END time= Mon Aug 14 13:04:10 CST 2017
当synchronized修饰代码块时,锁的对象是指定的对象。
修改上面的test():
public Object lock = new Object();
public static void main(String[] args) {
SynchronizedDemo demo = new SynchronizedDemo();
new Thread(new DemoRunnable(demo), "T1").start();
new Thread(new DemoRunnable(demo), "T2").start();
}
public void test(){
System.out.println(Thread.currentThread().getName()+" START time= "+new Date());
synchronized(lock){
System.out.println(Thread.currentThread().getName()+" code block START time= "+new Date());
try {
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println(Thread.currentThread().getName()+" code block end time= "+new Date());
}
System.out.println(Thread.currentThread().getName()+" END time= "+new Date());
}
输出结果为:
T2 START time= Mon Aug 14 15:07:26 CST 2017
T2 code block START time= Mon Aug 14 15:07:26 CST 2017
T1 START time= Mon Aug 14 15:07:26 CST 2017
T2 code block end time= Mon Aug 14 15:07:28 CST 2017
T1 code block START time= Mon Aug 14 15:07:28 CST 2017
T2 END time= Mon Aug 14 15:07:28 CST 2017
T1 code block end time= Mon Aug 14 15:07:30 CST 2017
T1 END time= Mon Aug 14 15:07:30 CST 2017
由输出可以看出对于同步代码块,同一时间只能有一个线程访问。
由synchronized修饰的同步块,如果抛出异常,则锁自动释放。
修改上面的test方法:
public static void main(String[] args) throws InterruptedException {
SynchronizedDemo demo = new SynchronizedDemo();
new Thread(new DemoRunnable(demo), "T1").start();
Thread.sleep(50);//延时创建线程2,使线程1先执行
new Thread(new DemoRunnable(demo), "T2").start();
}
public void test(){
System.out.println(Thread.currentThread().getName()+" START time= "+new Date());
synchronized(lock){
System.out.println(Thread.currentThread().getName()+" code block START time= "+new Date());
try {
//使线程1抛出异常
if("T1".equals(Thread.currentThread().getName())){
throw new RuntimeException();
}
Thread.sleep(2000);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println(Thread.currentThread().getName()+" code block end time= "+new Date());
}
System.out.println(Thread.currentThread().getName()+" END time= "+new Date());
}
由输出结果可以看出,线程T1抛出异常,没有继续执行,而线程T2直接获取了锁。
synchroniced同步块对于同一个线程来说是可重入的,不回出现自己把自己锁死的问题。
下面的例子说明了这个问题:
public class SynchronizedReentrantTest {
public static void main(String[] args) {
test();
}
public static synchronized void test(){
System.out.println("abc");
demo();
System.out.println("ghi");
}
public static synchronized void demo(){
System.out.println("def");
}
}
输出结果:
abc
def
ghi
在代码中,test方法已经持有了class对象的锁,在没有释放的情况下,依然执行了demo方法。