在java多线程中，为了提高效率有些共享资源允许同时进行多个读的操作，但只允许一个写的操作，比如一个文件，只要其内容不变可以让多个线程同时读，不必做排他的锁定，排他的锁定只有在写的时候需要，以保证别的线程不会看到数据不完整的文件。
下面是个关于多线程读写锁的例子，我稍微做了下修改，蛮容易理解的，来至于http://www.highya.com/redirect.php?fid=113&tid=7180&goto=nextoldset。
这里模拟了这样一个场景: 在ReadWriteLockOperator对象里设置一个共享资源 shareResources 。
有3个读者(A, B, C)一直连续的从 shareResources 获取信息, 然后输出到控制台 ;有一个作者每隔60秒往shareResources 加入信息, 加信息的过程相对耗时, 在这段时间, 任何读者都不能访问 shareResources。
写了4个类来验证这种情况，只在windows下做了测试。
ReadTask.java 读任务
WriteTask.java 写任务
ReadWriteLockLogic.java 读写操作的逻辑
ReadWriteLockTest.java 带有main方法的测试类
---------------------------------------混哥线-----------------------------------------------
public class ReadTask extends Thread {
//logic bean
private ReadWriteLockLogic readWriteLockOperator;
//读者
private String reader;
public ReadTask(ReadWriteLockLogic readWriteLockOperator, String reader) {
this .readWriteLockOperator = readWriteLockOperator;
this .reader = reader;
}
private ReadTask(){}
// 执行任务
public void run() {
if ( this .readWriteLockOperator != null ){
try {
while (!isInterrupted()){
Thread.sleep( 200 );
System.out.println(reader + " read:"
+ Thread.currentThread().toString() + " : " + this .readWriteLockOperator.read());
}
} catch (Exception e) {
// TODO: handle exception
}
}
}
}-------------------------------------------------------------------------------------
public class WriteTask extends Thread{
//logic bean
private ReadWriteLockLogic readWriteLockOperator;
//作者
private String writer;
public WriteTask(ReadWriteLockLogic readWriteLockOperator, String writer) {
this .readWriteLockOperator = readWriteLockOperator;
this .writer = writer;
}
private WriteTask(){}
// 一个很耗时的写任务
public void run() {
try {
while (!isInterrupted()){
Thread.sleep( 100 );
this .readWriteLockOperator.write( this .writer, "hehehhe" );
}
} catch (Exception e) {
// TODO: handle exception
}
}
}----------------------------------------------------------------------------------
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReadWriteLock;
import java.util.concurrent.locks.ReentrantReadWriteLock;
//读写操作的逻辑
public class ReadWriteLockLogic {
// 初始化一个 ReadWriteLock
private ReadWriteLock lock = new ReentrantReadWriteLock();
//共享资源
private List<String> shareResources = new ArrayList<String>( 0 );
//读
public String read() {
// 得到 readLock 并锁定
Lock readLock = lock.readLock();
readLock.lock();
try {
// 读相对省时，做空循环 大约0.5second
for ( int i= 0 ;i< 2500000 ; i++){
System.out.print( "" );
}
// 做读的工作
StringBuffer buffer = new StringBuffer();
for (String shareResource : shareResources) {
buffer.append(shareResource);
buffer.append( "\t" );
}
return buffer.toString();
} finally {
readLock.unlock(); //一定要保证锁的释放
}
}
//写
public void write(String writer, String content) {
// 得到 writeLock 并锁定
Lock writeLock = lock.writeLock();
writeLock.lock();
try {
System.out.println(writer + " write ===" + Thread.currentThread().toString());
// 写比较费时，所以做空循环 大约13second
for ( int i= 0 ;i< 10000000 ; i++){
System.out.print( "" );
System.out.print( "" );
}
// 做写的工作
int count = shareResources.size();
for ( int i=count; i < count + 1 ; i++) {
shareResources.add(content + "_" + i);
}
} finally {
writeLock.unlock(); //一定要保证锁的释放
}
}
}------------------------------------------------------------------------------------
import java.util.concurrent.ExecutionException;
import java.util.concurrent.Executors;
import java.util.concurrent.ScheduledExecutorService;
import java.util.concurrent.TimeUnit;
public class ReadWriteLockTest {
public static void main(String[] args) throws InterruptedException, ExecutionException {
//1 创建一个具有排程功能的线程池
ScheduledExecutorService service = Executors.newScheduledThreadPool( 5 );
//2 读写锁的logic bean
ReadWriteLockLogic lockOperator = new ReadWriteLockLogic();
//3 生成一个可执行任务(该任务执行完毕可以返回结果 或者 抛出异常；而Runnable接口的run方法则不行)
Runnable writeTask1 = new WriteTask(lockOperator, "作者A" );
//4 延时0秒后每2秒重复执行writeTask1;
service.scheduleAtFixedRate(writeTask1, 0 , 60 , TimeUnit.SECONDS);
//5 创建3个读任务
Runnable readTask1 = new WriteTask(lockOperator, "作者B" );
Runnable readTask2 = new ReadTask(lockOperator, "读者B" );
Runnable readTask3 = new ReadTask(lockOperator, "读者C" );
//6 延时0秒后每秒执行一次task1;
service.scheduleAtFixedRate(readTask1, 1 , 1 , TimeUnit.SECONDS);
service.scheduleAtFixedRate(readTask2, 2 , 1 , TimeUnit.SECONDS);
service.scheduleAtFixedRate(readTask3, 3 , 1 , TimeUnit.SECONDS);
}
}----------------------------------------------------------------------------------------
作者A write ===Thread[pool-1-thread-1,5,main]
作者B write ===Thread[pool-1-thread-4,5,main]
读者C read:Thread[pool-1-thread-3,5,main] : hehehhe_0 hehehhe_1
读者B read:Thread[pool-1-thread-2,5,main] : hehehhe_0 hehehhe_1
作者A write ===Thread[pool-1-thread-1,5,main]
................
通过观察控制台，可以看到作者a出现后，大约5秒作者b才会出现，而又过了5秒后，读者c和读者b同时会出现，接着5秒后，作者a又出现了。这说明了，读锁之间没有排斥，可以多线程持有并且排斥WriteLock的持有线程。而WriteLock是全部排斥的，是独占的,比较独!
下面是附赠的读写锁的小知识，来至http://www.txdnet.cn/essay/view.jsp?tid=1288670091703&cid=2
(a).重入方面其内部的WriteLock可以获取ReadLock,但是反过来ReadLock想要获得WriteLock则永远都不要想.
(b).WriteLock可以降级为ReadLock,顺序是:先获得WriteLock再获得ReadLock,然后释放WriteLock,这时候线程将保持Readlock的持有.反过来ReadLock想要升级为WriteLock则不可能,为什么?参看(a),呵呵.
(c).ReadLock可以被多个线程持有并且在作用时排斥任何的WriteLock,而WriteLock则是完全的互斥.这一特性最为重要,因为对于高读取频率而相对较低写入的数据结构,使用此类锁同步机制则可以提高并发量.
(d).不管是ReadLock还是WriteLock都支持Interrupt,语义与ReentrantLock一致.
(e).WriteLock支持Condition并且与ReentrantLock语义一致,而ReadLock则不能使用Condition,否则抛出UnsupportedOperationException异常.
