Java并发——ThreadPoolExecutor源码解析

本文总结一下对线程池源码的学习，基于jdk 1.8

什么是线程池

顾名思义线程池就是一个可以提供一组可复用线程的对象。线程池内部有阻塞队列，用来存放等待执行的任务。然后内部的线程来执行这些任务，线程会不断的从阻塞队列中获取任务来执行，而不是执行完一个任务就销毁。

线程池的作用

在高并发场景下，如果给每个任务都去创建一个线程来执行，结果就是大量的线程创建与销毁，系统的开销将会很大，影响应用的执行效率。

同时，线程池可以有效的限制应用程序中同一时刻运行的线程数量，避免CPU资源不足，造成阻塞。

线程池的使用

定义一个线程池

ExecutorService executor = new ThreadPoolExecutor(1, 4, 20,
                TimeUnit.SECONDS, new ArrayBlockingQueue<>(10));

该线程池，核心线程数为1，最大线程数为4，非核心线程空闲存活时间20s，阻塞队列是长度为10的 ArrayBlockingQueue，线程工厂和饱和拒绝策略没有定义，采用默认实现

线程池中添加任务
Executor接口提供了execute方法，传入Runnable接口的实现（任务），线程池将会调度执行这些任务

for (int i = 0; i < 12; i++) {
    executor.execute(() -> 
        System.out.println(Thread.currentThread().getName()));
}

ThreadPoolExecutor源码分析

定义线程池

public ThreadPoolExecutor(int corePoolSize,
                          int maximumPoolSize,
                          long keepAliveTime,
                          TimeUnit unit,
                          BlockingQueue<Runnable> workQueue,
                          ThreadFactory threadFactory,
                          RejectedExecutionHandler handler) {
        if (corePoolSize < 0 ||
            maximumPoolSize <= 0 ||
            maximumPoolSize < corePoolSize ||
            keepAliveTime < 0)
            throw new IllegalArgumentException();
        if (workQueue == null || threadFactory == null || handler == null)
            throw new NullPointerException();
        this.acc = System.getSecurityManager() == null ?
                null :
                AccessController.getContext();
        this.corePoolSize = corePoolSize;
        this.maximumPoolSize = maximumPoolSize;
        this.workQueue = workQueue;
        this.keepAliveTime = unit.toNanos(keepAliveTime);
        this.threadFactory = threadFactory;
        this.handler = handler;
    }

这是线程池最终执行的构造方法，共有7个参数，分别是

• 核心线程数
• 最大线程数(核心线程+非核心线程)
• 非核心线程空闲存活时间
• 空闲存活时间单位
• 阻塞队列
• 线程工厂
• 饱和拒绝策略

在定义时前5个参数是必须传递的，后两个参数不传递表示使用默认提供
注意看第三个参数，默认它是作用在非核心线程上的，如果希望同时作用在核心线程上，可以调用如下方法设置

allowCoreThreadTimeOut(true);

线程池的状态

下面来看一下线程池内部的一些状态，以及工作线程数的封装

    /**
     * The main pool control state, ctl, is an atomic integer packing
     * two conceptual fields
     *   workerCount, indicating the effective number of threads
     *   runState,    indicating whether running, shutting down etc
     *
     * In order to pack them into one int, we limit workerCount to
     * (2^29)-1 (about 500 million) threads rather than (2^31)-1 (2
     * billion) otherwise representable. 
     */
    private final AtomicInteger ctl = new AtomicInteger(ctlOf(RUNNING, 0));
    private static final int COUNT_BITS = Integer.SIZE - 3;
    private static final int CAPACITY   = (1 << COUNT_BITS) - 1;

    // runState is stored in the high-order bits
    private static final int RUNNING    = -1 << COUNT_BITS;
    private static final int SHUTDOWN   =  0 << COUNT_BITS;
    private static final int STOP       =  1 << COUNT_BITS;
    private static final int TIDYING    =  2 << COUNT_BITS;
    private static final int TERMINATED =  3 << COUNT_BITS;

    // Packing and unpacking ctl
    private static int runStateOf(int c)     { return c & ~CAPACITY; }
    private static int workerCountOf(int c)  { return c & CAPACITY; }
    private static int ctlOf(int rs, int wc) { return rs | wc; }

节选了部分关键注释说明

ctl是一个原子的Integer类型，包含了workerCount和runState，为了把这个两个值拼到一个int中，限制了workerCount最大为2^29 -1，大约为500多万，而不是2^31-1。

也就是说作者把工作线程数和状态值拼接到了一个int中，这些属性含义如下

属性	含义	值
COUNT_BITS	2进制计数位数	29
CAPACITY	线程数容量	(2^29)-1
RUNNING	运行状态	-2^29
SHUTDOWN	关闭状态(不接受新任务，把已有任务执行完)	0
STOP	停止(不接受新任务，终止正在执行的)	2^29
TIDYING	所有任务终止，工作线程数为0	2^30
TERMINATED	terminated()方法执行完成	2^29 + 2^30

通过上表可以看到，线程池的5个状态数值是递增的
所以只要状态是>=SHUTDOWN，就代表线程池不会再接受新的任务

三个静态方法解释

• ctlOf(int rs, int wc)
  线程池状态与线程数拼成一个int，高3位为状态，低29位为工作线程数
• runStateOf(int c)
  获取线程池状态
• workerCountOf(int c)
  获取工作线程数

执行任务—execute

public void execute(Runnable command) {
        if (command == null)
            throw new NullPointerException();
        /*
         * Proceed in 3 steps:
         *
         * 1. If fewer than corePoolSize threads are running, try to
         * start a new thread with the given command as its first
         * task.  The call to addWorker atomically checks runState and
         * workerCount, and so prevents false alarms that would add
         * threads when it shouldn't, by returning false.
         *
         * 2. If a task can be successfully queued, then we still need
         * to double-check whether we should have added a thread
         * (because existing ones died since last checking) or that
         * the pool shut down since entry into this method. So we
         * recheck state and if necessary roll back the enqueuing if
         * stopped, or start a new thread if there are none.
         *
         * 3. If we cannot queue task, then we try to add a new
         * thread.  If it fails, we know we are shut down or saturated
         * and so reject the task.
         */
        int c = ctl.get();
        if (workerCountOf(c) < corePoolSize) {
            if (addWorker(command, true))
                return;
            c = ctl.get();
        }
        if (isRunning(c) && workQueue.offer(command)) {
            int recheck = ctl.get();
            if (! isRunning(recheck) && remove(command))
                reject(command);
            else if (workerCountOf(recheck) == 0)
                addWorker(null, false);
        }
        else if (!addWorker(command, false))
            reject(command);
    }

可以看到，Doug Lea老爷子已经将该方法的流程注释的很清晰了，我这里就通俗的描述一下：

• 如果运行的线程数小于核心线程数，那么就新启动一个线程，并将该任务作为此线程的firstTask
• 若线程池的核心线程数已经满了，就将任务添加到阻塞队列中，需要二次检查(因为有可能在上一次检查之后死掉，或者是进入该方法时线程池被关闭)，若线程池不是运行状态，则将该任务从队列中移除，并进行拒绝处理。如果二次检查后没有工作的线程了，那么就新启动一个线程执行该任务
• 如果阻塞队列也满了，就新启动一个非核心线程，如果失败的话，说明线程池被shutdown或者是队列容量和最大线程数都已达到上限，将此任务拒绝掉

添加工作线程—addWorker

private boolean addWorker(Runnable firstTask, boolean core) {
    retry:
    for (;;) {
        int c = ctl.get();
        int rs = runStateOf(c);

        // Check if queue empty only if necessary.
        if (rs >= SHUTDOWN &&
            ! (rs == SHUTDOWN &&
               firstTask == null &&
               ! workQueue.isEmpty()))
            return false;

        for (;;) {
            // 获取工作线程数
            int wc = workerCountOf(c);
            // 如果大于CAPACITY最大容量，或者core为true，与corePoolSize比，
            // 否则与maximumPoolSize比较，如果大于允许的线程数则返回 false
            if (wc >= CAPACITY ||
                wc >= (core ? corePoolSize : maximumPoolSize))
                return false;
            // worker + 1成功，跳出retry外层循环
            if (compareAndIncrementWorkerCount(c))
                break retry;
            // cas操作失败，如果线程池状态改变，跳出内层循环，继续判断状态
            c = ctl.get();  // Re-read ctl
            if (runStateOf(c) != rs)
                continue retry;
            // else CAS failed due to workerCount change; retry inner loop
        }
    }

    boolean workerStarted = false;
    boolean workerAdded = false;
    Worker w = null;
    try {
        w = new Worker(firstTask);
        final Thread t = w.thread;
        if (t != null) {
            final ReentrantLock mainLock = this.mainLock;
            mainLock.lock();
            try {
                // 拿到锁以后二次检查
                int rs = runStateOf(ctl.get());
                
                // 如果在运行状态，或者是SHUTDOWN状态且firstTask为空（取queue中任务）
                if (rs < SHUTDOWN ||
                    (rs == SHUTDOWN && firstTask == null)) {
                    // 线程已经启动，抛出异常
                    if (t.isAlive()) // precheck that t is startable
                        throw new IllegalThreadStateException();
                    // 添加到workers中
                    workers.add(w);
                    // 记录最大的worker数量
                    int s = workers.size();
                    if (s > largestPoolSize)
                        largestPoolSize = s;
                    workerAdded = true;
                }
            } finally {
                mainLock.unlock();
            }
            // 启动线程
            if (workerAdded) {
                t.start();
                workerStarted = true;
            }
        }
    } finally {
        // 添加失败
        if (! workerStarted)
            addWorkerFailed(w);
    }
    return workerStarted;
}

firstTask
addWorker方法的第一个参数是firstTask，firstTask是线程池中Worker对象的一个属性，该对象代表新启动线程的第一个任务。
在execute方法源码中可以看到，只有在新增线程时才会给firstTask赋值，如果任务被添加到queue中，将其置为null，线程会去阻塞队列中获取任务执行。

在添加worker前，会在有必要的情况下检查阻塞队列是否为空

if (rs >= SHUTDOWN &&
    ! (rs == SHUTDOWN &&
       firstTask == null &&
       ! workQueue.isEmpty()))
    return false;

1、如果状态大于SHUTDOWN，不接受新任务，直接返回false；
2、如果状态等于SHUTDOWN，firstTask != null，返回false，不允许新增任务；
2、如果状态等于SHUTDOWN，firstTask == null，说明该线程会去队列中取任务执行，如果此时workQueue.isEmpty()，则返回false；

addWorkerFailed
添加线程失败时，会将刚创建的worker对象移除掉

private void addWorkerFailed(Worker w) {
    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        // HashSet中移除worker
        if (w != null)
            workers.remove(w);
        // 线程数减一
        decrementWorkerCount();
        // 尝试关闭线程池
        tryTerminate();
    } finally {
        mainLock.unlock();
    }
}

内部线程包装对象—Worker

Worker是ThreadPoolExecutor中的内部类，包含了执行任务的线程（节选了部分属性和方法）

private final class Worker extends AbstractQueuedSynchronizer
    implements Runnable {
    /** 执行任务的线程 */
    final Thread thread;
    /** 初始化运行的任务，可能为空 */
    Runnable firstTask;
    /** 每个worker完成任务的计数器 */
    volatile long completedTasks;

    /**
     * Creates with given first task and thread from ThreadFactory.
     * @param firstTask the first task (null if none)
     */
    Worker(Runnable firstTask) {
        // 无法获取锁，从而禁止 interrupt worker
        setState(-1); // inhibit interrupts until runWorker
        this.firstTask = firstTask;
        // 线程工厂初始化线程
        this.thread = getThreadFactory().newThread(this);
    }

    /** Delegates main run loop to outer runWorker  */
    public void run() {
        runWorker(this);
    }
}

Worker对象通过继承AbstractQueuedSynchronizer队列同步器，来控制worker的同步状态，
新建worker时，setState(-1) ，设置状态为 -1 使得其他线程无法获取到worker的锁，禁止interrupt该线程（只有当前状态为 0 时才有机会获得锁）

执行任务——runWorker
线程池真正执行任务的地方就在这里了

final void runWorker(Worker w) {
    Thread wt = Thread.currentThread();
    Runnable task = w.firstTask;
    w.firstTask = null;
    w.unlock(); // allow interrupts
    // 标记线程是否是因为发生异常中断的
    boolean completedAbruptly = true;
    try {
        // 获取要执行的任务, firstTask为空就从队列中取
        while (task != null || (task = getTask()) != null) {
            w.lock();
            // If pool is stopping, ensure thread is interrupted;
            // if not, ensure thread is not interrupted.  This
            // requires a recheck in second case to deal with
            // shutdownNow race while clearing interrupt
            if ((runStateAtLeast(ctl.get(), STOP) ||
                 (Thread.interrupted() &&
                  runStateAtLeast(ctl.get(), STOP))) &&
                !wt.isInterrupted())
                wt.interrupt();
            try {
                // 钩子函数，可以在执行前自定义些操作
                beforeExecute(wt, task);
                Throwable thrown = null;
                try {
                    // 执行任务
                    task.run();
                } catch (RuntimeException x) {
                    thrown = x; throw x;
                } catch (Error x) {
                    thrown = x; throw x;
                } catch (Throwable x) {
                    thrown = x; throw new Error(x);
                } finally {
                    // 钩子函数，任务执行完成后调用的方法
                    afterExecute(task, thrown);
                }
            } finally {
                task = null;
                w.completedTasks++;
                w.unlock();
            }
        }
        // 执行成功，将异常标记置为 false
        completedAbruptly = false;
    } finally {
        // 执行worker退出操作
        processWorkerExit(w, completedAbruptly);
    }
}

线程池状态检查

if ((runStateAtLeast(ctl.get(), STOP) ||
             (Thread.interrupted() &&
              runStateAtLeast(ctl.get(), STOP))) &&
            !wt.isInterrupted())
            wt.interrupt();

1、如果runState >= stop(stop状态线程池要中断正在运行的任务)，且线程未被设置为中断，则interrupt线程
2、如果runState < stop，进行二次检查(有可能在第一次获取状态后，调用了shutdownNow方法)，此时线程如果有中断标记，则清除(Thread.interrupted()返回线程中断状态，并将其清除)，再次查看状态，runSate >= stop 则interrupt线程

获取任务——getTask

private Runnable getTask() {
    // 上一次循环取task是否超时
    boolean timedOut = false; 

    for (;;) {
        int c = ctl.get();
        int rs = runStateOf(c);

        // 检查线程池状态
        if (rs >= SHUTDOWN && (rs >= STOP || workQueue.isEmpty())) {
            decrementWorkerCount();
            return null;
        }

        int wc = workerCountOf(c);

        // 线程空闲了是否需要退出
        boolean timed = allowCoreThreadTimeOut || wc > corePoolSize;

        // 检查线程池workerCount
        if ((wc > maximumPoolSize || (timed && timedOut))
            && (wc > 1 || workQueue.isEmpty())) {
            if (compareAndDecrementWorkerCount(c))
                return null;
            continue;
        }

        try {
            Runnable r = timed ?
                workQueue.poll(keepAliveTime, TimeUnit.NANOSECONDS) :
                workQueue.take();
            if (r != null)
                return r;
            timedOut = true;
        } catch (InterruptedException retry) {
            timedOut = false;
        }
    }
}

1、循环取任务，直到取到任务，或者是不需要返回任务为止；
2、如果线程池是 > stop状态，则workerCount减1，返回null，
如果是shutdown状态，且队列为空，则workerCount减1，返回null
3、(wc > maximumPoolSize || (timed && timedOut))
wc > 最大线程数 或者是 线程空闲了keepAliveTime 且 空闲需被销毁
(wc > 1 || workQueue.isEmpty())
wc > 1 或者 队列为空
同时满足上述两个条件，说明该线程不需要获取任务来执行，则workerCount减1，返回null

• timedOut代表上一次循环中，取task时候是否超时(代表了该线程空闲了keepAliveTime时间)
• timed代表该线程空闲了是否需要销毁

4、如果timed == true，则调用poll方法，等待keepAliveTime时间，
否则调用take方法阻塞直到获取到任务(到达这一步，说明线程池状态为running或者是shutdown且workQueue不为空)

worker退出——processWorkerExit
任务执行完成后，在finally语句中执行worker的退出操作

private void processWorkerExit(Worker w, boolean completedAbruptly) {
    // 如果线程异常退出，则workerCount减 1
    if (completedAbruptly)
        decrementWorkerCount();

    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        completedTaskCount += w.completedTasks;
        // set 中异常执行完的worker对象
        workers.remove(w);
    } finally {
        mainLock.unlock();
    }
    // 尝试停止线程池
    tryTerminate();

    int c = ctl.get();
    // 状态为running或者shutdown
    if (runStateLessThan(c, STOP)) {
        // 异常退出直接新增加一个worker
        if (!completedAbruptly) {
            // 计算最小线程数
            int min = allowCoreThreadTimeOut ? 0 : corePoolSize;
            if (min == 0 && ! workQueue.isEmpty())
                min = 1;
            // 当前工作线程大于min，则无需新增，直接返回
            if (workerCountOf(c) >= min)
                return; // replacement not needed
        }
        addWorker(null, false);
    }
}

该方法中有个很重要的操作就是调用tryTerminate方法，尝试终止线程池
接下来就来分析线程池的关闭操作
tryTerminate + awaitTermination

final void tryTerminate() {
    for (;;) {
        int c = ctl.get();
        // 1、stop状态，则往下执行
        // 2、shutdown且队列为空则往下执行，其余情况直接return
        if (isRunning(c) ||
            runStateAtLeast(c, TIDYING) ||
            (runStateOf(c) == SHUTDOWN && ! workQueue.isEmpty()))
            return;
        // 如果工作线程数不为空，则中断一个worker线程
        if (workerCountOf(c) != 0) { // Eligible to terminate
            interruptIdleWorkers(ONLY_ONE);
            return;
        }

        // 执行到这里，说明worker为0，且没有任务需要执行
        final ReentrantLock mainLock = this.mainLock;
        mainLock.lock();
        try {
            // 设置线程池状态为tidying
            if (ctl.compareAndSet(c, ctlOf(TIDYING, 0))) {
                try {
                    // 调用钩子函数，需继承在子类中实现
                    terminated();
                } finally {
                    // 设置线程池状态为terminated
                    ctl.set(ctlOf(TERMINATED, 0));
                    // 线程池终止完成信号通知，通知awaitTermination方法
                    termination.signalAll();
                }
                return;
            }
        } finally {
            mainLock.unlock();
        }
        // else retry on failed CAS
    }
}

public boolean awaitTermination(long timeout, TimeUnit unit)
    throws InterruptedException {
    // 设置的阻塞超时时间
    long nanos = unit.toNanos(timeout);
    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        for (;;) {
            // 如果线程池已经关闭，直接返回
            if (runStateAtLeast(ctl.get(), TERMINATED))
                return true;
            if (nanos <= 0)
                return false;
            // 阻塞，如果tryTerminate方法关闭成功的话，会唤醒这里
            nanos = termination.awaitNanos(nanos);
        }
    } finally {
        mainLock.unlock();
    }
}

该方法中比较重要的一步操作就是中断空闲线程interruptIdleWorkers(ONLY_ONE)

/**
 * @param onlyOne If true, interrupt at most one worker. This is
 * called only from tryTerminate when termination is otherwise
 * enabled but there are still other workers.  In this case, at
 * most one waiting worker is interrupted to propagate shutdown
 * signals in case all threads are currently waiting.
 * Interrupting any arbitrary thread ensures that newly arriving
 * workers since shutdown began will also eventually exit.
 * To guarantee eventual termination, it suffices to always
 * interrupt only one idle worker, but shutdown() interrupts all
 * idle workers so that redundant workers exit promptly, not
 * waiting for a straggler task to finish.
 */
private void interruptIdleWorkers(boolean onlyOne) {
    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        for (Worker w : workers) {
            Thread t = w.thread;
            // 如果线程未中断，且可以获取到锁，则interrupt
            if (!t.isInterrupted() && w.tryLock()) {
                try {
                    t.interrupt();
                } catch (SecurityException ignore) {
                } finally {
                    w.unlock();
                }
            }
            // 如果该值为true，则跳出循环
            if (onlyOne)
                break;
        }
    } finally {
        mainLock.unlock();
    }
}

粗略翻译一下方法上面的注释，如果onlyOne参数被设置为true的话，该方法最多只会中断一个worker线程，为了把shutdown信号传播下去，保证线程池最终的关闭，最多就只中断一个空闲线程。
线程阻塞的话就是阻塞在getTask方法中，这里中断一个线程后，getTask --> processWorkerExit --> tryTerminate --> interruptIdleWorkers --> getTask
其实tryTerminate方法中，为什么要设置onlyOne为true，如果那个地方是false会是什么结果，没有思考的很明白，后续多查阅些资料实践一下。
其实上面已经涉及到了线程池的关闭流程，下面还有两个比较重要的方法来分析一下

关闭线程池——shutdown

public void shutdown() {
    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        checkShutdownAccess();
        // 将线程池状态设置为shutdown
        advanceRunState(SHUTDOWN);
        // 中断所有空闲的线程
        interruptIdleWorkers();
        onShutdown(); // hook for ScheduledThreadPoolExecutor
    } finally {
        mainLock.unlock();
    }
    // 尝试关闭线程池
    tryTerminate();
}

这个地方的巧妙之处就在于最后的tryTerminate方法，因为线程池shutdown状态时，是要把剩下的任务执行完的，如果调shutdown方法的时候恰好所有线程都在执行任务，那么就无法中断。
关闭线程池——shutdownNow

public List<Runnable> shutdownNow() {
    List<Runnable> tasks;
    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        checkShutdownAccess();
        // 设置线程池状态为stop
        advanceRunState(STOP);
        // 中断所有worker
        interruptWorkers();
        // 取出队列中的任务并返回
        tasks = drainQueue();
    } finally {
        mainLock.unlock();
    }
    // 尝试关闭线程池
    tryTerminate();
    return tasks;
}

stop状态需要把所有线程中断，任务也放弃，所有shutdownNow会中断所有worker线程

private void interruptWorkers() {
    final ReentrantLock mainLock = this.mainLock;
    mainLock.lock();
    try {
        for (Worker w : workers)
            w.interruptIfStarted();
    } finally {
        mainLock.unlock();
    }
}

// 此方法是内部类Worker中的方法，提到这里来便于阅读
    void interruptIfStarted() {
        Thread t;
        if (getState() >= 0 && (t = thread) != null && !t.isInterrupted()) {
            try {
                t.interrupt();
            } catch (SecurityException ignore) {
            }
        }
    }

持有worker锁时，state 为 1，未持有锁时为 0，所以这里就可以看出区别，shutdown方法是只能中断空闲的worker线程，而shutdownNow则是把所有worker线程都中断。

线程池的基本流程就到这里了，如果有理解的不对的地方，或者需要补充的地方，还望各位小伙伴不吝赐教 ^-^ ~