一. 简介
CyclicBarrier ,回环栅栏(循环屏障),通过它可以实现让一组线程等待至某个状态(屏障点)之后再全部同时执行。叫做回环是因为当所有等待线程都被释放以后,CyclicBarrier可以被重用。
二. CyclicBarrier的使用
构造方法
//parties表示屏障拦截的线程数量,每个线程调用await方法告诉CyclicBarrier我已经到达了屏障,然后当前线程被阻塞。
public CyclicBarrier(int parties) {
this(parties, null);
}
//用于在线程到达屏障时,优先执行 barrierAction,
//方便处理更复杂的业务场景(该线程的执行时机是在到达屏障之后再执行)
public CyclicBarrier(int parties, Runnable barrierAction) {
if (parties <= 0) throw new IllegalArgumentException();
this.parties = parties;
this.count = parties;
this.barrierCommand = barrierAction;
}
重要方法
//指定数量的线程全部调用await()方法时,这些线程不再阻塞
// BrokenBarrierException 表示栅栏已经被破坏,破坏的原因可能是其中一个线程 await() 时被中断或者超时
public int await() throws InterruptedException, BrokenBarrierException
public int await(long timeout, TimeUnit unit) throws InterruptedException, BrokenBarrierException, TimeoutException
//循环重置
public void reset() {}
三. CyclicBarrier的应用场景
CyclicBarrier 可以用于多线程计算数据,最后合并计算结果的场景。
例子一 ,多个线程调用await之后阻塞,等到达到屏障拦截的线程数量之后,再一起执行
public static void main(String[] args) {
CyclicBarrier cyclicBarrier = new CyclicBarrier(3);
for (int i = 0; i < 5; i++) {
new Thread(new Runnable() {
@Override
public void run() {
try {
System.out.println(Thread.currentThread().getName()
+ "开始等待其他线程");
cyclicBarrier.await();
System.out.println(Thread.currentThread().getName() + "开始执行");
//TODO 模拟业务处理
Thread.sleep(5000);
System.out.println(Thread.currentThread().getName() + "执行完毕");
} catch (Exception e) {
e.printStackTrace();
}
}
}).start();
}
}
Thread-0开始等待其他线程
Thread-2开始等待其他线程
Thread-1开始等待其他线程
Thread-4开始等待其他线程
Thread-3开始等待其他线程
Thread-2开始执行
Thread-1开始执行
Thread-0开始执行
例子二 用于多线程计算数据,最后合并计算结果的场景
//保存每个学生的平均成绩
private ConcurrentHashMap<String, Integer> map=new ConcurrentHashMap<String,Integer>();
private ExecutorService threadPool= Executors.newFixedThreadPool(3);
private CyclicBarrier cb=new CyclicBarrier(3,()->{
int result=0;
Set<String> set = map.keySet();
for(String s:set){
result+=map.get(s);
}
System.out.println("三人平均成绩为:"+(result/3)+"分");
});
public void count(){
for(int i=0;i<3;i++){
threadPool.execute(new Runnable(){
@Override
public void run() {
//获取学生平均成绩
int score=(int)(Math.random()*40+60);
map.put(Thread.currentThread().getName(), score);
System.out.println(Thread.currentThread().getName()
+"同学的平均成绩为:"+score);
try {
//执行完运行await(),等待所有学生平均成绩都计算完毕
cb.await();
} catch (InterruptedException | BrokenBarrierException e) {
e.printStackTrace();
}
}
});
}
}
public static void main(String[] args) {
CyclicBarrierTest2 cb=new CyclicBarrierTest2();
cb.count();
}
pool-1-thread-3同学的平均成绩为:62
pool-1-thread-1同学的平均成绩为:74
pool-1-thread-2同学的平均成绩为:85
三人平均成绩为:73分
例子三 利用CyclicBarrier的计数器能够重置,屏障可以重复使用的特性
public static void main(String[] args) {
AtomicInteger counter = new AtomicInteger();
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
5, 5, 1000, TimeUnit.SECONDS,
new ArrayBlockingQueue<>(100),
(r) -> new Thread(r, counter.addAndGet(1) + " 号 "),
new ThreadPoolExecutor.AbortPolicy());
CyclicBarrier cyclicBarrier = new CyclicBarrier(5,
() -> System.out.println("裁判:比赛开始~~"));
for (int i = 0; i < 10; i++) {
threadPoolExecutor.submit(new Runner(cyclicBarrier));
}
}
static class Runner extends Thread{
private CyclicBarrier cyclicBarrier;
public Runner (CyclicBarrier cyclicBarrier) {
this.cyclicBarrier = cyclicBarrier;
}
@Override
public void run() {
try {
int sleepMills = ThreadLocalRandom.current().nextInt(1000);
Thread.sleep(sleepMills);
System.out.println(Thread.currentThread().getName() + " 选手已就位, 准备共用时: " + sleepMills + "ms" + cyclicBarrier.getNumberWaiting());
cyclicBarrier.await();
} catch (InterruptedException e) {
e.printStackTrace();
}catch(BrokenBarrierException e){
e.printStackTrace();
}
}
}
3 号 选手已就位, 准备共用时: 223ms0
2 号 选手已就位, 准备共用时: 315ms1
5 号 选手已就位, 准备共用时: 471ms2
1 号 选手已就位, 准备共用时: 556ms3
4 号 选手已就位, 准备共用时: 923ms4
裁判:比赛开始~~
3 号 选手已就位, 准备共用时: 285ms0
2 号 选手已就位, 准备共用时: 413ms1
1 号 选手已就位, 准备共用时: 533ms2
5 号 选手已就位, 准备共用时: 661ms3
4 号 选手已就位, 准备共用时: 810ms4
裁判:比赛开始~~
四. CyclicBarrier与CountDownLatch的区别
- CountDownLatch的计数器只能使用一次,而CyclicBarrier的计数器可以使用reset() 方法重置。所以CyclicBarrier能处理更为复杂的业务场景,比如如果计算发生错误,可以重置计数器,并让线程们重新执行一次
- CyclicBarrier还提供getNumberWaiting(可以获得CyclicBarrier阻塞的线程数量)、isBroken(用来知道阻塞的线程是否被中断)等方法。
- CountDownLatch会阻塞主线程,CyclicBarrier不会阻塞主线程,只会阻塞子线程。
- CountDownLatch和CyclicBarrier都能够实现线程之间的等待,只不过它们侧重点不同。CountDownLatch一般用于一个或多个线程,等待其他线程执行完任务后,再执行。CyclicBarrier一般用于一组线程互相等待至某个状态,然后这一组线程再同时执行。
- CyclicBarrier 还可以提供一个 barrierAction,合并多线程计算结果。
- CyclicBarrier是通过ReentrantLock的"独占锁"和Conditon来实现一组线程的阻塞唤醒的,而CountDownLatch则是通过AQS的“共享锁”实现
五. CyclicBarrier源码解析
以例子以为例,thread0 从await方法开始,await会调用dowait
public int await() throws InterruptedException, BrokenBarrierException {
try {
return dowait(false, 0L);
} catch (TimeoutException toe) {
throw new Error(toe); // cannot happen
}
}
那么dowait里又做了什么
private int dowait(boolean timed, long nanos)
throws InterruptedException, BrokenBarrierException,
TimeoutException {
final ReentrantLock lock = this.lock;
//如果状态为0就将其修改为1,并设置当前线程,调用await时外面肯定是lock
lock.lock();
try {
//每一个栈栏算是一代
final Generation g = generation;
if (g.broken)
throw new BrokenBarrierException();
if (Thread.interrupted()) {
breakBarrier();
throw new InterruptedException();
}
//每个线程执行一次,count自减一
int index = --count;
// count减到0,说明几个线程都到达屏障,就会重置 进入下一个屏障
if (index == 0) {
boolean ranAction = false;
try {
final Runnable command = barrierCommand;
if (command != null)
command.run();
ranAction = true;
nextGeneration();
return 0;
} finally {
if (!ranAction)
breakBarrier();
}
}
// loop until tripped, broken, interrupted, or timed out
for (;;) {
try {
if (!timed)
//trip就是一个条件队列condition,入条件等待队列,单向链表结构
trip.await();
else if (nanos > 0L)
nanos = trip.awaitNanos(nanos);
} catch (InterruptedException ie) {
if (g == generation && ! g.broken) {
breakBarrier();
throw ie;
} else {
Thread.currentThread().interrupt();
}
}
if (g.broken)
throw new BrokenBarrierException();
if (g != generation)
return index;
if (timed && nanos <= 0L) {
breakBarrier();
throw new TimeoutException();
}
}
} finally {
lock.unlock();
}
}
入队阻塞的方法在await()中,下面看一下
public final void await() throws InterruptedException {
if (Thread.interrupted())
throw new InterruptedException();
//入队 创建节点
Node node = addConditionWaiter();
// 释放锁,这样其他线程才能获取锁,执行
int savedState = fullyRelease(node);
int interruptMode = 0;
while (!isOnSyncQueue(node)) {
//不在当前队列中就阻塞
LockSupport.park(this);
if ((interruptMode = checkInterruptWhileWaiting(node)) != 0)
break;
}
if (acquireQueued(node, savedState) && interruptMode != THROW_IE)
interruptMode = REINTERRUPT;
if (node.nextWaiter != null) // clean up if cancelled
unlinkCancelledWaiters();
if (interruptMode != 0)
reportInterruptAfterWait(interruptMode);
}
addConditionWaiter中创建节点,对于thread0,头节点是自己,lastWaiter也是自己
/**
* 添加条件等待节点
*/
private Node addConditionWaiter() {
Node t = lastWaiter;
//节点取消 则移除.
if (t != null && t.waitStatus != Node.CONDITION) {
unlinkCancelledWaiters();
t = lastWaiter;
}
//创建一个节点,thread=Thread.currentThread(),waitstatus= Node.CONDITION =-2
Node node = new Node(Thread.currentThread(), Node.CONDITION);
//如果上一个节点为null,则将该节点设置为头节点
if (t == null)
firstWaiter = node;
else
t.nextWaiter = node;
lastWaiter = node;
return node;
}
释放锁
final int fullyRelease(Node node) {
boolean failed = true;
try {
int savedState = getState();
if (release(savedState)) {
failed = false;
return savedState;
} else {
throw new IllegalMonitorStateException();
}
} finally {
if (failed)
node.waitStatus = Node.CANCELLED;
}
}
public final boolean release(int arg) {
if (tryRelease(arg)) {
Node h = head;
if (h != null && h.waitStatus != 0)
unparkSuccessor(h);
return true;
}
return false;
}
protected final boolean tryRelease(int releases) {
int c = getState() - releases;
if (Thread.currentThread() != getExclusiveOwnerThread())
throw new IllegalMonitorStateException();
boolean free = false;
//释放锁时state变为0了
if (c == 0) {
free = true;
//将当前线程设置为null
setExclusiveOwnerThread(null);
}
//修改state状态值
setState(c);
return free;
}
thread0的大致流程
首先会调用lock.lock进行加锁,加锁之后调用trip.await方法进行入队阻塞,入队是通过addConditionWaiter添加进条件等待队列,然后通过fullyRelease释放锁,设置当前线程为null,然后修改state状态值,最后调用LockSupport.park(this);进行阻塞。
threa1的流程大致和thread0一样,还没有达到屏障数量,入队的地方和thread0不一样
private Node addConditionWaiter() {
Node t = lastWaiter;
//节点取消 则移除.
if (t != null && t.waitStatus != Node.CONDITION) {
unlinkCancelledWaiters();
t = lastWaiter;
}
//创建一个节点,thread=Thread.currentThread(),waitstatus= Node.CONDITION =-2
Node node = new Node(Thread.currentThread(), Node.CONDITION);
//如果上一个节点为null,则将该节点设置为头节点
if (t == null)
firstWaiter = node;
else
//thread1执行时,t已经不为null了
t.nextWaiter = node;
//lastWaiter指向当前
lastWaiter = node;
return node;
}
thread1流程
thread2执行的时候count已经减到0,会执行nextGeneration方法
//每个线程执行一次,count自减一
int index = --count;
// count减到0,说明几个线程都到达屏障,就会重置 进入下一个屏障
if (index == 0) {
boolean ranAction = false;
try {
final Runnable command = barrierCommand;
if (command != null)
command.run();
ranAction = true;
//开启下一代屏障
nextGeneration();
return 0;
} finally {
if (!ranAction)
breakBarrier();
}
}
private void nextGeneration() {
// 唤醒所有线程,唤醒操作是在同步等待队列中,所以要将条件等待队列转换为同步等待队列
trip.signalAll();
// 重置count
count = parties;
//创建下一代屏障
generation = new Generation();
}
public final void signalAll() {
if (!isHeldExclusively())
throw new IllegalMonitorStateException();
Node first = firstWaiter;
if (first != null)
doSignalAll(first);
}
private void doSignalAll(Node first) {
//将firstWaiter和lastWaiter都置为null,就没有首尾节点了
lastWaiter = firstWaiter = null;
//条件队列的出队
do {
//循环将条件队列转同步队列
Node next = first.nextWaiter;
//first的nextWaiter置为null
first.nextWaiter = null;
//条件队列转同步队列,因为唤醒是在同步队列中
transferForSignal(first);
first = next;
} while (first != null);
}
条件队列的出队
/**
* 条件队列转同步队列
*/
final boolean transferForSignal(Node node) {
//将同步状态改为0
if (!compareAndSetWaitStatus(node, Node.CONDITION, 0))
return false;
Node p = enq(node);
int ws = p.waitStatus;
//将ws设置为-1
if (ws > 0 || !compareAndSetWaitStatus(p, ws, Node.SIGNAL))
LockSupport.unpark(node.thread);
return true;
}
private Node enq(final Node node) {
for (;;) {
Node t = tail;
if (t == null) { // Must initialize
if (compareAndSetHead(new Node()))
tail = head;
} else {
node.prev = t;
if (compareAndSetTail(t, node)) {
t.next = node;
return t;
}
}
}
}
同步队列入队
thread2开始解锁
private int dowait(boolean timed, long nanos){
//....
} finally {
lock.unlock();
}
}
public void unlock() {
sync.release(1);
}
public final boolean release(int arg) {
if (tryRelease(arg)) {
Node h = head;
if (h != null && h.waitStatus != 0)
//唤醒线程0
unparkSuccessor(h);
return true;
}
return false;
}
private void unparkSuccessor(Node node) {
int ws = node.waitStatus;
if (ws < 0)
//将ws设置为0
compareAndSetWaitStatus(node, ws, 0);
Node s = node.next;
if (s == null || s.waitStatus > 0) {
s = null;
for (Node t = tail; t != null && t != node; t = t.prev)
if (t.waitStatus <= 0)
s = t;
}
if (s != null)
//唤醒线程0
LockSupport.unpark(s.thread);
}
Thread0唤醒之后就会获取锁,执行业务逻辑然后再释放锁
public final void await() throws InterruptedException {
if (Thread.interrupted())
throw new InterruptedException();
Node node = addConditionWaiter();
int savedState = fullyRelease(node);
int interruptMode = 0;
while (!isOnSyncQueue(node)) {
LockSupport.park(this);
if ((interruptMode = checkInterruptWhileWaiting(node)) != 0)
break;
}
// 在这里获取锁,再释放锁
if (acquireQueued(node, savedState) && interruptMode != THROW_IE)
interruptMode = REINTERRUPT;
if (node.nextWaiter != null)
unlinkCancelledWaiters();
if (interruptMode != 0)
reportInterruptAfterWait(interruptMode);
}
//CAS尝试获取锁
final boolean acquireQueued(final Node node, int arg) {
boolean failed = true;
try {
boolean interrupted = false;
for (;;) {
final Node p = node.predecessor();
//cas获取锁
if (p == head && tryAcquire(arg)) {
setHead(node);
p.next = null; // help GC
failed = false;
return interrupted;
}
if (shouldParkAfterFailedAcquire(p, node) &&
parkAndCheckInterrupt())
interrupted = true;
}
} finally {
if (failed)
cancelAcquire(node);
}
}
同步队列就会出队
Thread0唤醒之后,执行业务逻辑之后会unlock,又会唤醒Thread1,Thread1唤醒之后,又会唤醒Thread2。
总结:
await方法,
前半段 释放锁 进入条件队列,阻塞线程(Thread0 Thread1),
过渡阶段 其他线程调用singnal/signalAll唤醒(Thread2),条件队列转同步队列,可以在释放锁的时候唤醒head的后续节点所在的线程
后半段 (Thread0)被唤醒的线程获取锁(如果有竞争,CAS获取锁失败,还会阻塞),Thread0释放锁,唤醒同步队列中head的后续节点所在的线程(独占锁的逻辑)