随笔- 261 文章- 0 评论- 12 阅读- 45万

synchronized、Lock 为什么要设计成可重入

为什么监视器锁(ObjectMonitor)、Lock 要设计成可重入的目的主要有两个

1、防止死锁

2、面向对象更好的封装代码

先看一段代码

@Slf4j
public class SynchronizedDemo {
    private static final Object lock = new Object();
 
    public static void main(String[] args) {
        synchronized (lock) {
            log.info("enjoy eat");
            synchronized (lock) {
                log.info("enjoy sleep");
            }
        }
    }
}

上述代码中,某个线程第一次成功获取到监视器锁之后,如果 monitor 是不可重入的,当该线程尝试去获取第二个锁的时候阻塞了,那么整个线程就会进入死锁状态,无法继续向下执行了

再看一段代码

@Slf4j
public class SynchronizedDemo {
    private static final Object lock = new Object();
 
    public static void main(String[] args) {
        synchronized (lock) {
            log.info("enjoy eat");
            synchronized (lock) {
                log.info("enjoy sleep");
                synchronized (lock){
                    log.info("do something");
                }
            }
        }
    }
}

上述代码只嵌套了三层锁对象,如果需要 10 次加锁呢,岂不是要写 10 层 synchronized(lock),很不优雅,也不好维护,并且不符合面向对象的编程思想

上述代码可以改造成如下,符合面向对象的编程思想

@Slf4j
public class SynchronizedDemo {
    public static void main(String[] args) {
        LockObject lock = new LockObject();
        lock.eat();
    }
}
 
@Slf4j
class LockObject {
    public synchronized void eat() {
        log.info("enjoy eat");
        sleep();
    }
 
    public synchronized void sleep() {
        log.info("enjoy sleep");
        doSomething();
    }
 
    public synchronized void doSomething() {
        log.info("do something");
    }
}