Java Concurrency - Semaphore 信号量

Semaphore 是一个控制访问多个共享资源的计数器。

当一个线程想要访问某个共享资源，首先，它必须获得 semaphore。如果 semaphore 的内部计数器的值大于 0，那么 semaphore 减少计数器的值并允许访问共享的资源。计数器的值大于 0 表示，有可以自由使用的资源，所以线程可以访问并使用它们。另一种情况，如果 semaphore 的计数器的值等于 0，那么 semaphore 让线程进入休眠状态一直到计数器大于 0。计数器的值等于 0 表示全部的共享资源都正被线程们使用，所以此线程想要访问就必须等到某个资源成为自由的。当线程使用完共享资源时，它必须释放 semaphore 以便其他线程可以访问共享资源。这个操作会增加 semaphore 的内部计数器的值。

二进制信号量

二进制信号量（binary semaphores）是一种比较特殊的信号量，这种信号量只保护访问唯一的共享资源，它的内部计数器值只能是 1 或者 0。

假设有若干个线程排队执行打印任务，打印机在同一时间只能执行单个任务，打印过程中其他线程只能挂起等待。

public class Printer {

    private final Semaphore semaphore;

    public Printer() {
        semaphore = new Semaphore(1);
    }

    public void print(Object doc) {
        try {
            semaphore.acquire();
            Long duration = (long) (Math.random() * 10);
            System.out.printf("%s: PrintQueue: Printing a Job during %d seconds\n", Thread.currentThread().getName(), duration);
            TimeUnit.SECONDS.sleep(duration);
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            semaphore.release();
        }
    }

}

public class PrintJob implements Runnable {

    private Printer printer;

    public PrintJob(Printer printer) {
        this.printer = printer;
    }

    @Override
    public void run() {
        System.out.printf("%s: Going to print a job\n", Thread.currentThread().getName());
        printer.print(new Object());
        System.out.printf("%s: The document has been printed\n", Thread.currentThread().getName());
    }

}

public class Main {

    public static void main (String args[]){
        Printer printer = new Printer();

        Thread[] threads = new Thread[10];
        for (int i = 0; i < 10; i++){
            threads[i] = new Thread(new PrintJob(printer), "Thread" + i);
            threads[i].start();
        }
    }

}

运行结果：

Thread1: Going to print a job
Thread6: Going to print a job
Thread9: Going to print a job
Thread7: Going to print a job
Thread8: Going to print a job
Thread4: Going to print a job
Thread5: Going to print a job
Thread3: Going to print a job
Thread0: Going to print a job
Thread2: Going to print a job
Thread1: PrintQueue: Printing a Job during 7 seconds
Thread1: The document has been printed
Thread6: PrintQueue: Printing a Job during 5 seconds
Thread6: The document has been printed
Thread9: PrintQueue: Printing a Job during 4 seconds
Thread9: The document has been printed
Thread7: PrintQueue: Printing a Job during 8 seconds
Thread7: The document has been printed
Thread8: PrintQueue: Printing a Job during 6 seconds
Thread8: The document has been printed
Thread4: PrintQueue: Printing a Job during 2 seconds
Thread4: The document has been printed
Thread5: PrintQueue: Printing a Job during 0 seconds
Thread5: The document has been printed
Thread3: PrintQueue: Printing a Job during 7 seconds
Thread3: The document has been printed
Thread0: PrintQueue: Printing a Job during 0 seconds
Thread0: The document has been printed
Thread2: PrintQueue: Printing a Job during 5 seconds
Thread2: The document has been printed

这个例子的关键是 Printer 类的 print() 方法。此方法展示了当你使用 semaphore 来实现临界区以保护共享资源时必须遵守的三个步骤：

1. 首先，调用 acquire() 方法获得信号量；
2. 然后，操作共享资源；
3. 最后，调用 release()  方法释放信号量。

另一个重点是 Printer 类的构造方法和初始化 Semaphore 对象。将信号量初始化为 1，就是创建了一个二进制信号量。二进制信号量只保护一个共享资源，这个例子中就是 printer。当开始 10 个线程时，第一个获得 semaphore 的线程就获得临界区的访问权，其他线程都会被 semaphore 阻塞，直到获得 semaphore 的线程释放它。当 semaphore 被释放时，它会在等待的线程中选择其中一个并赋予其临界区的访问权。所有的打印任务都会被执行，只是它们需要排队，一个接一个地执行。

控制并发访问多个共享资源

在上个例子，使用二进制信号量来保护一个共享资源。但是，信号量也可以用来保护多个共享资源。下面的类使用信号量控制对内容池的访问：

class Pool {
    private static final int MAX_AVAILABLE = 100;
    private final Semaphore available = new Semaphore(MAX_AVAILABLE, true);

    public Object getItem() throws InterruptedException {
        available.acquire();
        return getNextAvailableItem();
    }

    public void putItem(Object x) {
        if (markAsUnused(x))
            available.release();
    }

    // Not a particularly efficient data structure; just for demo

    protected Object[] items = ... whatever kinds of items being managed
    protected boolean[] used = new boolean[MAX_AVAILABLE];

    protected synchronized Object getNextAvailableItem() {
        for (int i = 0; i < MAX_AVAILABLE; ++i) {
            if (!used[i]) {
                used[i] = true;
                return items[i];
            }
        }
        return null; // not reached
    }

    protected synchronized boolean markAsUnused(Object item) {
        for (int i = 0; i < MAX_AVAILABLE; ++i) {
            if (item == items[i]) {
                if (used[i]) {
                    used[i] = false;
                    return true;
                } else
                    return false;
            }
        }
        return false;
    }

}

公平模式

与 ReentrantLock 一样，Semaphore 也提供一个接收 fair 参数的构造方法。当此参数置为 true 时，Semaphore 保证等待时间最久的线程优先获得信号量。

Semaphore 的方法

Semaphore 类有另外 2 个版本的 acquire() 方法：

• acquireUninterruptibly()：acquire()方法是当 semaphore 的内部计数器的值为 0 时，阻塞线程直到 semaphore 被释放。在阻塞期间，线程可能会被中断，然后此方法抛出 InterruptedException 异常。而此版本的 acquire 方法会忽略线程的中断而且不会抛出任何异常。
• tryAcquire()：此方法会尝试获取 semaphore。如果成功，返回true。如果不成功，返回 false 值，并不会被阻塞和等待 semaphore 的释放。

acquire、acquireUninterruptibly、tryAcquire 和 release 方法有一个外加的包含一个 int 参数的版本。这个参数表示线程想要获取或者释放 semaphore 的许可数。