Java synchronized 关键字和Lock的随笔

rainsilence

浏览: 160780 次
性别:
来自: 上海

最近访客更多访客>>

zhang55090

iop118

cjh326419411

finit

博主相关

博客

微博

相册

留言

关于我

文章分类

社区版块

存档分类

博客分类：

Java

最近客户300个人同时按下一个按钮，在执行到一个业务模块的时候出现了脏读。

package org.test.thread;

public class Worker {

	
	public void executeJob() {
		
		// statement A check()
		....
		
		// statement B
		....
		// 
		dao.save();
		
	}
}

比如上面的代码，第一个线程走到B了，第二个线程走到了A,check的地方。比如重复性check。如果第一个线程和第二个线程的查询主键相同。那么当线程1走到dao.save的地方，线程2刚好跳过check，也就是check无效化了。

于是，考虑用同步来解决。

到JDK5为止，java中实现线程安全起码有3种方法,ThreadLocal,synchronized关键字，Lock。

ThreadLocal是一种用空间换时间的策略。即为每一个线程创建副本。不适用于我们现在碰到的问题。所以不展开。

synchronized关键字:

在java中，他可以是方法修饰符，也可以成为独立的同步块。而加在方法前时，有两种方法，一般的方法和同步方法。两种效果不同。如：

如下这种，

package org.test.thread;

public class Worker {
	public synchronized void executeJob() {
		
	}
}

相当于

package org.test.thread;

public class Worker {

	
	public void executeJob() {
		
		synchronized(this) {
			
		}
		
	}
}

而加在静态方法前

package org.test.thread;

public class Worker {

	public static synchronized void executeJob() {
		
	}
}

则相当于

package org.test.thread;

public class Worker {

	public void executeJob() {
		
		synchronized(this.getClass()) {
			
		}
		
	}
}

虽然我不推荐你这么认为。因为静态方法加同步和一般方法中用同步块锁定类，这两种方法的用法是完全不一样的，虽然他们锁定的对象是一样的(Class)。

要弄清楚他的锁定对象，我们来做个实验：

package org.test.thread;

public class Worker {

	
	public synchronized void executeA(String name) {
		for (int i = 0; i < 10; i++) {
			System.out.println(name + "-executeA-" + i);
			try {
				Thread.sleep(500);
			} catch (InterruptedException e) {
				e.printStackTrace();
			}
		}
	}
	
	
	public synchronized void executeB(String name) {
		for (int i = 0; i < 10; i++) {
			System.out.println(name + "-executeB-" + i);
			try {
				Thread.sleep(500);
			} catch (InterruptedException e) {
				e.printStackTrace();
			}
		}
	}
	
	public synchronized static void executeC(String name) {
		for (int i = 0; i < 10; i++) {
			System.out.println(name + "-executeC-" + i);
			try {
				Thread.sleep(500);
			} catch (InterruptedException e) {
				e.printStackTrace();
			}
		}
	}
	
	private class SynchronizerWorkerA extends Thread {
		
		public void run() {
			executeA("thread1");
		}
	}
	
	private class SynchronizerWorkerB extends Thread {
		public void run() {
			executeB("thread2");
		}
	}
	
	private class SynchronizerWorkerC extends Thread {
		public void run() {
			executeC("thread3");
		}
	}
	
	public static void main(String args[]) {
		
		Worker worker = new Worker();
		
		worker.new SynchronizerWorkerA().start();
		worker.new SynchronizerWorkerB().start();
		worker.new SynchronizerWorkerC().start();
	}
}

结果：

thread1-executeA-0

thread3-executeC-0

thread1-executeA-1

thread3-executeC-1

thread1-executeA-2

thread3-executeC-2

thread1-executeA-3

thread3-executeC-3

thread1-executeA-4

thread3-executeC-4

thread1-executeA-5

thread3-executeC-5

thread1-executeA-6

thread3-executeC-6

thread1-executeA-7

thread3-executeC-7

thread1-executeA-8

thread3-executeC-8

thread1-executeA-9

thread3-executeC-9

thread2-executeB-0

thread2-executeB-1

thread2-executeB-2

thread2-executeB-3

thread2-executeB-4

thread2-executeB-5

thread2-executeB-6

thread2-executeB-7

thread2-executeB-8

thread2-executeB-9

可以看到，A方法和C方法可以同时调用，而B则被锁着，直到A方法的锁释放后才能被调用。原因在于静态方法的锁锁定class，而一般方法的锁锁定对象(效果上可以这么认为)

让我们把代码修改一下：

package org.test.thread;

public class Worker {

	
	public synchronized void executeA(String name) {
		for (int i = 0; i < 10; i++) {
			System.out.println(name + "-" + i);
			try {
				Thread.sleep(500);
			} catch (InterruptedException e) {
				e.printStackTrace();
			}
		}
	}
	
	public synchronized static void executeB(String name) {
		for (int i = 0; i < 10; i++) {
			System.out.println(name + "-" + i);
			try {
				Thread.sleep(500);
			} catch (InterruptedException e) {
				e.printStackTrace();
			}
		}
	}
	
	private class SynchronizerWorkerA extends Thread {
		
		private String name = null;
		
		public SynchronizerWorkerA(String name) {
			this.name = name;
		}
		
		public void run() {
			executeA(name);
		}
	}
	
	private class SynchronizerWorkerB extends Thread {
		
		private String name = null;
		
		public SynchronizerWorkerB(String name) {
			this.name = name;
		}
		
		public void run() {
			executeB(name);
		}
	}
	
	public static void main(String args[]) {
		new Worker().new SynchronizerWorkerA("thread1").start();
		new Worker().new SynchronizerWorkerA("thread2").start();
		new Worker().new SynchronizerWorkerB("thread3").start();
		new Worker().new SynchronizerWorkerB("thread4").start();
	}
}

知道有什么区别吗？我们在一个类的不同对象之间用了同步。

结果：

thread1-0

thread2-0

thread3-0

thread1-1

thread2-1

thread3-1

thread1-2

thread2-2

thread3-2

thread3-3

thread2-3

thread1-3

thread3-4

thread2-4

thread1-4

thread1-5

thread3-5

thread2-5

thread2-6

thread1-6

thread3-6

thread1-7

thread2-7

thread3-7

thread1-8

thread3-8

thread2-8

thread1-9

thread3-9

thread2-9

thread4-0

thread4-1

thread4-2

thread4-3

thread4-4

thread4-5

thread4-6

thread4-7

thread4-8

thread4-9

可以看到，一般方法的同步在不同对象之间没有效果。而静态方法则有作用。(锁定的方式不一样)

理解了这点，我们来看最后一种，Lock

Lock:

JDK5提供了吞吐量比synchronized更好的方法-----Lock接口。他的实现类有两个

让我们看如下代码:

package org.test.thread;

import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;

public class Worker extends Thread {

	private final Lock reentrantLock = new ReentrantLock();
	
	private String name;
	
	public Worker(String name) {
		this.name = name;
	}
	
	public void executeJob() {
		
		try {
			reentrantLock.lock();
			for (int i = 0; i < 10; i++) {
				System.out.println(name + "-" + i);
				try {
					Thread.sleep(500);
				} catch (InterruptedException e) {
					e.printStackTrace();
				}
			}
		} finally {
			reentrantLock.unlock();
		}
		
	}
	
	public void run() {
		executeJob();
	}
	
	public static void main(String args[]) {
		
		new Worker("thread-1").start();
		new Worker("thread-2").start();
	}
}

结果：

thread-1-0

thread-2-0

thread-1-1

thread-2-1

thread-1-2

thread-2-2

thread-1-3

thread-2-3

thread-2-4

thread-1-4

thread-1-5

thread-2-5

thread-1-6

thread-2-6

thread-1-7

thread-2-7

thread-1-8

thread-2-8

thread-1-9

thread-2-9

可以看到，lock没有起作用，这是为啥？

原因在于我们锁定了不同对象。

让我们将代码稍微改改

package org.test.thread;

import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;

public class Worker {

	private final Lock reentrantLock = new ReentrantLock();
	
	public void executeJob(String name) {
		
		try {
			reentrantLock.lock();
			for (int i = 0; i < 10; i++) {
				System.out.println(name + "-" + i);
				try {
					Thread.sleep(500);
				} catch (InterruptedException e) {
					e.printStackTrace();
				}
			}
		} finally {
			reentrantLock.unlock();
		}
		
	}
	
	private class WorkerThread extends Thread {
		
		private String name = null;
		
		public WorkerThread(String name) {
			this.name = name;
		}
		
		public void run() {
			executeJob(name);
		}
	}
	
	public static void main(String args[]) {
		
		Worker worker = new Worker();
		
		worker.new WorkerThread("thread1").start();
		worker.new WorkerThread("thread2").start();
	}
}

结果为:

thread1-0

thread1-1

thread1-2

thread1-3

thread1-4

thread1-5

thread1-6

thread1-7

thread1-8

thread1-9

thread2-0

thread2-1

thread2-2

thread2-3

thread2-4

thread2-5

thread2-6

thread2-7

thread2-8

thread2-9

看起作用了吧。

或者将原代码的lock对象改为static的

package org.test.thread;

import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;

public class Worker extends Thread {

	private final static Lock reentrantLock = new ReentrantLock();
	
	private String name;
	
	public Worker(String name) {
		this.name = name;
	}
	
	public void executeJob() {
		
		try {
			reentrantLock.lock();
			for (int i = 0; i < 10; i++) {
				System.out.println(name + "-" + i);
				try {
					Thread.sleep(500);
				} catch (InterruptedException e) {
					e.printStackTrace();
				}
			}
		} finally {
			reentrantLock.unlock();
		}
		
	}
	
	public void run() {
		executeJob();
	}
	
	public static void main(String args[]) {
		
		new Worker("thread-1").start();
		new Worker("thread-2").start();
	}
}

结果同上

当然，我们也可以用synchronized轻松实现Lock接口所带来的功能：

package org.test.thread;


public class Worker extends Thread {

	private static Object reentrantLock = new Object();
	
	private String name;
	
	public Worker(String name) {
		this.name = name;
	}
	
	public void executeJob() {
		
		synchronized (reentrantLock) {
			for (int i = 0; i < 10; i++) {
				System.out.println(name + "-" + i);
				try {
					Thread.sleep(500);
				} catch (InterruptedException e) {
					e.printStackTrace();
				}
			}
		}
	}
	
	public void run() {
		executeJob();
	}
	
	public static void main(String args[]) {
		
		new Worker("thread-1").start();
		new Worker("thread-2").start();
	}
}

Lock与synchronized的区别（仅列出主要的）

1.当块被同步，如果不想等了，synchronized无法停止等待，也没办法得到锁，而Lock可以用tryLock方法轻松做到.

2.性能上Lock更优

添加了类似锁投票、定时锁等候和可中断锁等候的一些特性。此外，它还提供了在激烈争用情况下更佳的性能

分享到：

NodeJs中的非阻塞方法 | TimeZone构造注意点，及no dsv的解决方案

2012-04-08 12:31
浏览 3820
评论(0)
分类:编程语言
查看更多

发表评论

您还没有登录,请您登录后再发表评论

最近访客更多访客>>

博主相关

文章分类

社区版块

存档分类

最新评论