关于System.out.println的多线程并发问题

本文转载自
IT江湖，原文链接：Click me

如果println函数的参数为常量则不会出现线程并发问题，但是如果参数为表达式形式，则JVM在执行println函数的时候会分为几步来执行，从而造成并发问题。

如下例子所示：

Java

public class Test
{
public static void main(String[] args)
{
ExecutorService pool = Executors.newFixedThreadPool(2);
Runnable t1 = new MyRunnable("张三", 2000);
Runnable t2 = new MyRunnable("李四", 3600);
Runnable t3 = new MyRunnable("王五", 2700);
Runnable t4 = new MyRunnable("老张", 600);
Runnable t5 = new MyRunnable("老牛", 1300);
Runnable t6 = new MyRunnable("老朱", 800);
//执行各个线程
pool.execute(t1);
pool.execute(t2);
pool.execute(t3);
pool.execute(t4);
pool.execute(t5);
pool.execute(t6);
//关闭线程池
pool.shutdown();
}
}

class MyRunnable implements Runnable
{
private static AtomicLong aLong = new AtomicLong(10000); //原子量，每个线程都可以自由操作
private String name; //操作人
private int data; //操作数

MyRunnable(String name, int data)
{
this.name = name;
this.data = data;
}
public void run()
{
System.out.println(name + "执行了" + data + "，当前余额：" + aLong.addAndGet(data));
}
}

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

public class Test { public static void main(String[] args) { ExecutorService pool = Executors.newFixedThreadPool(2); Runnable t1 = new MyRunnable("张三", 2000); Runnable t2 = new MyRunnable("李四", 3600); Runnable t3 = new MyRunnable("王五", 2700); Runnable t4 = new MyRunnable("老张", 600); Runnable t5 = new MyRunnable("老牛", 1300); Runnable t6 = new MyRunnable("老朱", 800); //执行各个线程 pool.execute(t1); pool.execute(t2); pool.execute(t3); pool.execute(t4); pool.execute(t5); pool.execute(t6); //关闭线程池 pool.shutdown(); } } class MyRunnable implements Runnable { private static AtomicLong aLong = new AtomicLong(10000); //原子量，每个线程都可以自由操作 private String name; //操作人 private int data; //操作数 MyRunnable(String name, int data) { this.name = name; this.data = data; } public void run() { System.out.println(name + "执行了" + data + "，当前余额：" + aLong.addAndGet(data)); } }

执行的结果为：

张三执行了2000，当前余额：12000

李四执行了3600，当前余额：15600

王五执行了2700，当前余额：18300

老牛执行了1300，当前余额：20200

老朱执行了800，当前余额：21000

老张执行了600，当前余额：18900

通过反编译，上面println函数的主要处理过程分为下面三步：

Java

long l = aLong.addAndGet(data);
String sb =new StringBuilder(String.valueOf(name))).append("执行了").append(data).append("，当前余额：").append(l);
synchronized (this)
{
print(sb.toString);
newLine();
}

1 2 3 4 5 6 7

long l = aLong.addAndGet(data); String sb =new StringBuilder(String.valueOf(name))).append("执行了").append(data).append("，当前余额：").append(l); synchronized (this) { print(sb.toString); newLine(); }

所以可以看出上面程序的执行顺序为：、

张三执行完->李四执行完->王五执行完->老张只执行到加操作，被老牛打断->老牛执行完->老朱执行完->老张执行剩下的输出操作

要想有一个正常的输出顺序，应该修改代码为：

Java

import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.atomic.AtomicLong;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;

public class Test
{
public static void main(String[] args)
{
ExecutorService pool = Executors.newFixedThreadPool(2);
Lock lock = new ReentrantLock(false);
Runnable t1 = new MyRunnable("张三", 2000, lock);
Runnable t2 = new MyRunnable("李四", 3600, lock);
Runnable t3 = new MyRunnable("王五", 2700, lock);
Runnable t4 = new MyRunnable("老张", 600, lock);
Runnable t5 = new MyRunnable("老牛", 1300, lock);
Runnable t6 = new MyRunnable("老朱", 800, lock);
//执行各个线程
pool.execute(t1);
pool.execute(t2);
pool.execute(t3);
pool.execute(t4);
pool.execute(t5);
pool.execute(t6);
//关闭线程池
pool.shutdown();
}
}

class MyRunnable implements Runnable
{
private static AtomicLong aLong = new AtomicLong(10000); //原子量，每个线程都可以自由操作
private String name; //操作人
private int data; //操作数
private Lock lock;

MyRunnable(String name, int data, Lock lock)
{
this.name = name;
this.data = data;
this.lock = lock;
}

public void run()
{
lock.lock();
System.out.println(name + "执行了" + data + "，当前余额：" + aLong.addAndGet(data));
lock.unlock();
}
}

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

import java.util.concurrent.ExecutorService; import java.util.concurrent.Executors; import java.util.concurrent.atomic.AtomicLong; import java.util.concurrent.locks.Lock; import java.util.concurrent.locks.ReentrantLock; public class Test { public static void main(String[] args) { ExecutorService pool = Executors.newFixedThreadPool(2); Lock lock = new ReentrantLock(false); Runnable t1 = new MyRunnable("张三", 2000, lock); Runnable t2 = new MyRunnable("李四", 3600, lock); Runnable t3 = new MyRunnable("王五", 2700, lock); Runnable t4 = new MyRunnable("老张", 600, lock); Runnable t5 = new MyRunnable("老牛", 1300, lock); Runnable t6 = new MyRunnable("老朱", 800, lock); //执行各个线程 pool.execute(t1); pool.execute(t2); pool.execute(t3); pool.execute(t4); pool.execute(t5); pool.execute(t6); //关闭线程池 pool.shutdown(); } } class MyRunnable implements Runnable { private static AtomicLong aLong = new AtomicLong(10000); //原子量，每个线程都可以自由操作 private String name; //操作人 private int data; //操作数 private Lock lock; MyRunnable(String name, int data, Lock lock) { this.name = name; this.data = data; this.lock = lock; } public void run() { lock.lock(); System.out.println(name + "执行了" + data + "，当前余额：" + aLong.addAndGet(data)); lock.unlock(); } }

结果为：

张三执行了2000，当前余额：12000

王五执行了2700，当前余额：14700

老张执行了600，当前余额：15300

老牛执行了1300，当前余额：16600

老朱执行了800，当前余额：17400

李四执行了3600，当前余额：21000

【注意】这里使用了一个对象锁，来控制对并发代码的访问。不管运行多少次，执行次序如何，最终余额均为21000，这个结果是正确的。有关原子量的用法很简单，关键是对原子量的认识，原子仅仅是保证变量操作的原子性，但整个程序还需要考虑线程安全的。