十三、 AbstractQueuedSynchronizer介绍
AbstractQueuedSynchronizer提供了一个队列，大多数开发者可能从来不会直接用到AQS，AQS中刮泥这个一个单一的状态信息 state,可以通过protected的getState,setState,compareAndSetState函数进行调用。对于ReentrantLock来说，state可以用来表示该线程获可重入锁的次数，semaphore来说state用来表示当前可用信号的个数，FutuerTask用来表示任务状态（例如还没开始，运行，完成，取消）。
十四、CountDownLatch原理
14.1 一个例子
public class Test { private static final int ThreadNum = 10; public static void main(String[] args) { //创建一个CountDownLatch实例，管理计数为ThreadNum
CountDownLatch countDownLatch = new CountDownLatch(ThreadNum); //创建一个固定大小的线程池
ExecutorService executor = Executors.newFixedThreadPool(ThreadNum); //添加线程到线程池
for(int i =0;i<ThreadNum;++i){
executor.execute(new Person(countDownLatch, i+1));
}
System.out.println("开始等待全员签到..."); try { //等待所有线程执行完毕
countDownLatch.await();
System.out.println("签到完毕，开始吃饭");
} catch (InterruptedException e) {
e.printStackTrace();
}finally {
executor.shutdown();
}
} static class Person implements Runnable{ private CountDownLatch countDownLatch; private int index; public Person(CountDownLatch cdl,int index){ this.countDownLatch = cdl; this.index = index;
} @Override
public void run() { try {
Thread.sleep(1000);
} catch (InterruptedException e) { // TODO Auto-generated catch block
e.printStackTrace();
}
System.out.println("person " + index +"签到"); //线程执行完毕，计数器减一
countDownLatch.countDown();
}
}
}如上代码，创建一个线程池和CountDownLatch实例，每个线程通过构造函数传入CountDownLatch的实例，主线程通过await等待线程池里面线程任务全部执行完毕，子线程则执行完毕后调用countDown计数器减一，等所有子线程执行完毕后，主线程的await才会返回。
14.2 原理
先看下类图：
可知CountDownLatch内部还是使用AQS实现的。
public CountDownLatch(int count) { if (count < 0) throw new IllegalArgumentException("count < 0"); this.sync = new Sync(count);
}
Sync(int count) {
setState(count);
}然后看下await方法：
public final void acquireSharedInterruptibly(int arg)
throws InterruptedException { //如果线程被中断则抛异常
if (Thread.interrupted()) throw new InterruptedException(); //尝试看当前是否计数值为0，为0则直接返回，否者进入队列等待
if (tryAcquireShared(arg) < 0)
doAcquireSharedInterruptibly(arg);
} protected int tryAcquireShared(int acquires) { return (getState() == 0) ? 1 : -1;
}如果tryAcquireShared返回-1则 进入doAcquireSharedInterruptibly
private void doAcquireSharedInterruptibly(int arg)
throws InterruptedException { //加入队列状态为共享节点
final Node node = addWaiter(Node.SHARED); boolean failed = true; try { for (;;) { final Node p = node.predecessor(); if (p == head) { int r = tryAcquireShared(arg); if (r >= 0) { //如果多个线程调用了await被放入队列则一个个返回。
setHeadAndPropagate(node, r);
p.next = null; // help GC
failed = false; return;
}
} //shouldParkAfterFailedAcquire会把当前节点状态变为SIGNAL类型，然后调用park方法把当先线程挂起，
if (shouldParkAfterFailedAcquire(p, node) &&
parkAndCheckInterrupt()) throw new InterruptedException();
}
} finally { if (failed)
cancelAcquire(node);
}
}调用await后，当前线程会被阻塞主，知道所有子线程调用了countdown方法，并在在计数为0时候调用该线程unpark方法激活线程，然后该线程重新tryAcquireShared会返回1。
然后看下 countDown方法：
委托给sync public void countDown() {
sync.releaseShared(1);
} public final boolean releaseShared(int arg) { if (tryReleaseShared(arg)) {
doReleaseShared(); return true;
} return false;
}首先看下tryReleaseShared
protected boolean tryReleaseShared(int releases) { //循环进行cas，直到当前线程成功完成cas使计数值（状态值state）减一更新到state
for (;;) { int c = getState(); if (c == 0) return false; int nextc = c-1; if (compareAndSetState(c, nextc)) return nextc == 0;
}
}该函数一直返回false直到当前计数器为0时候才返回true。
private void doReleaseShared() { for (;;) {
Node h = head; if (h != null && h != tail) { int ws = h.waitStatus; //节点类型为SIGNAL，把类型在通过cas设置回去，然后调用unpark激活调用await的线程
if (ws == Node.SIGNAL) { if (!compareAndSetWaitStatus(h, Node.SIGNAL, 0)) continue; // loop to recheck cases
unparkSuccessor(h);
} else if (ws == 0 &&
!compareAndSetWaitStatus(h, 0, Node.PROPAGATE)) continue; // loop on failed CAS
} if (h == head) // loop if head changed
break;
}
}激活主线程后，主线程会在调用tryAcquireShared获取锁。
十五、ReentrantLock独占锁原理
15.1 ReentrantLock结构
先上类图：
可知ReentrantLock最终还是使用AQS来实现，并且根据参数决定内部是公平还是非公平锁，默认是非公平锁
public ReentrantLock() {
sync = new NonfairSync();
} public ReentrantLock(boolean fair) {
sync = fair ? new FairSync() : new NonfairSync();
}加锁代码：
public void lock() {
sync.lock();
}15.2 公平锁原理
先看Lock方法：
protected final boolean tryAcquire(int acquires) { //获取当前线程和状态值
final Thread current = Thread.currentThread(); int c = getState(); //状态为0说明该锁未被任何线程持有
if (c == 0) { //为了实现公平，首先看队列里面是否有节点，有的话再看节点所属线程是不是当前线程，是的话hasQueuedPredecessors返回false,然后使用原子操作compareAndSetState保证一个线程更新状态为1，设置排他锁归属与当前线程。其他线程通过cass则返回false.
if (!hasQueuedPredecessors() &&
compareAndSetState(0, acquires)) {
setExclusiveOwnerThread(current); return true;
}
}//状态不为0说明该锁已经被线程持有，则看是否是当前线程持有，是则重入锁次数+1.
else if (current == getExclusiveOwnerThread()) { int nextc = c + acquires; if (nextc < 0) throw new Error("Maximum lock count exceeded");
setState(nextc); return true;
} return false;
}
}公平性保证代码：
public final boolean hasQueuedPredecessors() {
Node t = tail; // Read fields in reverse initialization order
Node h = head;
Node s; return h != t &&
((s = h.next) == null || s.thread != Thread.currentThread());
}再看看unLock方法，最终调用了Sync的tryRelease方法：
protected final boolean tryRelease(int releases) { //如果不是锁持有者调用UNlock则抛出异常。
int c = getState() - releases; if (Thread.currentThread() != getExclusiveOwnerThread()) throw new IllegalMonitorStateException(); boolean free = false; //如果当前可重入次数为0，则清空锁持有线程
if (c == 0) {
free = true;
setExclusiveOwnerThread(null);
} //设置可重入次数为原始值-1
setState(c); return free;
}15.3 非公平锁原理
final void lock() { //如果当前锁空闲0，则设置状态为1，并且设置当前线程为锁持有者
if (compareAndSetState(0, 1))
setExclusiveOwnerThread(Thread.currentThread()); else
acquire(1);//调用重写的tryAcquire方法-》nonfairTryAcquire方法
} final boolean nonfairTryAcquire(int acquires) { final Thread current = Thread.currentThread(); int c = getState(); if (c == 0) {//状态为0说明没有线程持有该锁
if (compareAndSetState(0, acquires)) {//cass原子性操作，保证只有一个线程可以设置状态
setExclusiveOwnerThread(current);//设置锁所有者
return true;
}
}//如果当前线程是锁持有者则可重入锁计数+1
else if (current == getExclusiveOwnerThread()) { int nextc = c + acquires; if (nextc < 0) // overflow
throw new Error("Maximum lock count exceeded");
setState(nextc); return true;
} return false;
}15.3 总结
可知公平与非公平都是先执行tryAcquire尝试获取锁，如果成功则直接获取锁，如果不成功则把当前线程放入队列。对于放入队列里面的第一个线程A在unpark后会进行自旋调用tryAcquire尝试获取锁，假如这时候有一个线程B执行了lock操作，那么也会调用tryAcquire方法尝试获取锁，但是线程B并不在队列里面，但是线程B有可能比线程A优先获取到锁，也就是说虽然线程A先请求的锁，但是却有可能没有B先获取锁，这是非公平锁实现。而公平锁要保证线程A要比线程B先获取锁。所以公平锁相比非公平锁在tryAcquire里面添加了hasQueuedPredecessors方法用来保证公平性。
十六、ReentrantReadWriteLock原理
如图读写锁内部维护了一个ReadLock和WriteLock，并且也提供了公平和非公平的实现，下面只介绍下非公平的读写锁实现。我们知道AQS里面只维护了一个state状态，而ReentrantReadWriteLock则需要维护读状态和写状态，一个state是无法表示写和读状态的。所以ReentrantReadWriteLock使用state的高16位表示读状态也就是读线程的个数，低16位表示写锁可重入量。
static final int SHARED_SHIFT = 16;
共享锁（读锁）状态单位值65536
static final int SHARED_UNIT = (1 << SHARED_SHIFT);
共享锁线程最大个数65535
static final int MAX_COUNT = (1 << SHARED_SHIFT) - 1;
排它锁(写锁)掩码 二进制 15个1
static final int EXCLUSIVE_MASK = (1 << SHARED_SHIFT) - 1;
/** 返回读锁线程数 */
static int sharedCount(int c) { return c >>> SHARED_SHIFT; }
/** 返回写锁可重入个数 */
static int exclusiveCount(int c) { return c & EXCLUSIVE_MASK; }16.1 WriteLock
lock 获取锁
protected final boolean tryAcquire(int acquires) {
Thread current = Thread.currentThread();
int c = getState();
int w = exclusiveCount(c);
//c!=0说明读锁或者写锁已经被某线程获取
if (c != 0) {
//w=0说明已经有线程获取了读锁或者w!=0并且当前线程不是写锁拥有者，则返回false
if (w == 0 || current != getExclusiveOwnerThread())
return false;
//说明某线程获取了写锁，判断可重入个数
if (w + exclusiveCount(acquires) > MAX_COUNT)
throw new Error("Maximum lock count exceeded");
// 设置可重入数量(1)
setState(c + acquires);
return true;
}
//第一个写线程获取写锁
if (writerShouldBlock() ||
!compareAndSetState(c, c + acquires))
return false;
setExclusiveOwnerThread(current);
return true;
}- unlock 释放锁
```Java
protected final boolean tryRelease(int releases) {
// 看是否是写锁拥有者调用的unlock
if (!isHeldExclusively())
throw new IllegalMonitorStateException();
//获取可重入值，这里没有考虑高16位，因为写锁时候读锁状态值肯定为0
int nextc = getState() - releases;
boolean free = exclusiveCount(nextc) == 0;
//如果写锁可重入值为0则释放锁，否者只是简单更新状态值。
if (free)
setExclusiveOwnerThread(null);
setState(nextc);
return free;
}16.2 ReadLock
对应读锁只需要分析下Sync的tryAcquireShared和tryReleaseShared
lock 获取锁
protected final int tryAcquireShared(int unused) { //获取当前状态值 Thread current = Thread.currentThread(); int c = getState(); //如果写锁计数不为0说明已经有线程获取了写锁，然后看是不是当前线程获取的写锁。 if (exclusiveCount(c) != 0 && getExclusiveOwnerThread() != current) return -1; //获取读锁计数 int r = sharedCount(c); //尝试获取锁，多个读线程只有一个会成功，不成功的进入下面fullTryAcquireShared进行重试 if (!readerShouldBlock() && r < MAX_COUNT && compareAndSetState(c, c + SHARED_UNIT)) { if (r == 0) { firstReader = current; firstReaderHoldCount = 1; } else if (firstReader == current) { firstReaderHoldCount++; } else { HoldCounter rh = cachedHoldCounter; if (rh == null || rh.tid != current.getId()) cachedHoldCounter = rh = readHolds.get(); else if (rh.count == 0) readHolds.set(rh); rh.count++; } return 1; } return fullTryAcquireShared(current); }unlock 释放锁
protected final boolean tryReleaseShared(int unused) { Thread current = Thread.currentThread(); if (firstReader == current) { // assert firstReaderHoldCount > 0; if (firstReaderHoldCount == 1) firstReader = null; else firstReaderHoldCount--; } else { HoldCounter rh = cachedHoldCounter; if (rh == null || rh.tid != current.getId()) rh = readHolds.get(); int count = rh.count; if (count <= 1) { readHolds.remove(); if (count <= 0) throw unmatchedUnlockException(); } --rh.count; } //循环直到自己的读计数-1 cas更新成功 for (;;) { int c = getState(); int nextc = c - SHARED_UNIT; if (compareAndSetState(c, nextc)) return nextc == 0; } }
十七、参考
Java并发编程实践
http://www.hollischuang.com/archives/934
http://ifeve.com/juc-reentrantreadwritelock/
扫一扫
1万+
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
