实现生产者消费者模式的四种方式（Synchronized、Lock、Semaphore、BlockingQueue）

所谓生产者消费者模式，即N个线程进行生产，同时N个线程进行消费，两种角色通过内存缓冲区进行通信

图片来源https://www.cnblogs.com/chent...

下面我们通过四种方式，来实现生产者消费者模式。

首先是最原始的synchronized方式

定义库存类（即图中缓存区）

class Stock {
    private String name;
    // 标记库存是否有内容
    private boolean hasComputer = false;

    public synchronized void putOne(String name) {
        // 若库存中已有内容，则生产线程阻塞等待
        while (hasComputer) {
            try {
                this.wait();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
        this.name = name;
        System.out.println("生产者...生产了 " + name);
        // 更新标记
        this.hasComputer = true;
        // 这里用notify的话，假设p0执行完毕，此时c0,c1都在wait, 同时唤醒另一个provider:p1，
        // p1判断标记后休眠，造成所有线程都wait的局面，即死锁；
        // 因此使用notifyAll解决死锁问题
        this.notifyAll();
    }

    public synchronized void takeOne() {
        // 若库存中没有内容，则消费线程阻塞等待生产完毕后继续
        while (!hasComputer) {
            try {
                this.wait();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
        System.out.println("消费者...消费了 " + name);
        this.hasComputer = false;
        this.notifyAll();
    }
}

定义生产者和消费者（为了节省空间和方便阅读，这里将生产者和消费者定义成了匿名内部类）

public static void main(String[] args) {
    // 用于通信的库存类
    Stock computer = new Stock();
    // 定义两个生产者和两个消费者
    Thread p1 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.putOne("Dell");
            }
        }
    });
    Thread p2 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.putOne("Mac");
            }
        }
    });
    
    Thread c1 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.takeOne();
            }
        }
    });
    Thread c2 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.takeOne();
            }
        }
    });
    p1.start();
    p2.start();
    c1.start();
    c2.start();
}

运行结果图

---

第二种方式：Lock

Jdk1.5之后加入了Lock接口，一个lock对象可以有多个Condition类,Condition类负责对lock对象进行wait,notify,notifyall操作

定义库存类

class LockStock {
    final Lock lock = new ReentrantLock();
    final Condition notFull = lock.newCondition();
    final Condition notEmpty = lock.newCondition();

    // 加入库存概念，可批量生产和消费
    // 定义最大库存为10
    final String[] stock = new String[10];
    // 写入标记、读取标记、已有商品数量
    int putptr, takeptr, count;

    public void put(String computer) {
        // lock代替synchronized
        lock.lock();
        try {
            // 若库存已满则生产者线程阻塞
            while (count == stock.length)
                notFull.await();
            // 库存中加入商品
            stock[putptr] = computer;
            // 库存已满，指针置零，方便下次重新写入
            if (++putptr == stock.length) putptr = 0;
            ++count;
            System.out.println(computer + " 正在生产数据： -- 库存剩余：" + count);
            notEmpty.signal();
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            lock.unlock();
        }
        try {
            Thread.sleep(1000);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }

    public String take(String consumerName) {
        lock.lock();
        try {
            while (count == 0)
                notEmpty.await();
            // 从库存中获取商品
            String computer = stock[takeptr];
            if (++takeptr == stock.length) takeptr = 0;
            --count;
            System.out.println(consumerName + " 正在消费数据：" + computer + " -- 库存剩余：" + count);
            notFull.signal();
            return computer;
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            lock.unlock();
        }

        // 无逻辑作用，放慢速度
        try {
            Thread.sleep(1000);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        return "";
    }
}

以上部分代码摘自java7 API中Condition接口的官方示例

接着还是定义生产者和消费者

public static void main(String[] args) {
    LockStock computer = new LockStock();
    Thread p1 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.put("Dell");
            }
        }
    });
    Thread p2 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.put("Mac");
            }
        }
    });

    Thread c1 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.take("zhangsan");
            }
        }
    });
    Thread c2 = new Thread(new Runnable() {
        @Override
        public void run() {
            while (true) {
                computer.take("李四");
            }
        }
    });
    // 两个生产者两个消费者同时运行
    p1.start();
    p2.start();
    c1.start();
    c2.start();
}

运行结果图：

---

第三种方式：Semaphore
首先依旧是库存类：

class Stock {
    List<String> stock = new LinkedList();
    // 互斥量，控制共享数据的互斥访问
    private Semaphore mutex = new Semaphore(1);

    // canProduceCount可以生产的总数量。 通过生产者调用acquire，减少permit数目
    private Semaphore canProduceCount = new Semaphore(10);

    // canConsumerCount可以消费的数量。通过生产者调用release，增加permit数目
    private Semaphore canConsumerCount = new Semaphore(0);

    public void put(String computer) {
        try {
            // 可生产数量 -1
            canProduceCount.acquire();
            mutex.acquire();
            // 生产一台电脑
            stock.add(computer);
            System.out.println(computer + " 正在生产数据" + " -- 库存剩余：" + stock.size());
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            // 释放互斥锁
            mutex.release();
            // 释放canConsumerCount，增加可以消费的数量
            canConsumerCount.release();
        }
        // 无逻辑作用，放慢速度
        try {
            Thread.sleep(500);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }

    public void get(String consumerName) {
        try {
            // 可消费数量 -1
            canConsumerCount.acquire();
            mutex.acquire();
            // 从库存消费一台电脑
            String removedVal = stock.remove(0);
            System.out.println(consumerName + " 正在消费数据：" + removedVal + " -- 库存剩余：" + stock.size());
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            mutex.release();
            // 消费后释放canProduceCount，增加可以生产的数量
            canProduceCount.release();
        }
    }
}

还是生产消费者：

public class SemaphoreTest {
    public static void main(String[] args) {
        // 用于多线程操作的库存变量
        final Stock stock = new Stock();
        // 定义两个生产者和两个消费者
        Thread dellProducer = new Thread(new Runnable() {
            @Override
            public void run() {
                while (true) {
                    stock.put("Del");
                }
            }
        });
        Thread macProducer = new Thread(new Runnable() {
            @Override
            public void run() {
                while (true) {
                    stock.put("Mac");
                }
            }
        });
        Thread consumer1 = new Thread(new Runnable() {
            @Override
            public void run() {
                while (true) {
                    stock.get("zhangsan");
                }
            }
        });
        Thread consumer2 = new Thread(new Runnable() {
            @Override
            public void run() {
                while (true) {
                    stock.get("李四");
                }
            }
        });
        dellProducer.start();
        macProducer.start();
        consumer1.start();
        consumer2.start();
    }
}

运行结果图：

---

第四种方式：BlockingQueue
BlockingQueue的put和take底层实现其实也是使用了第二种方式中的ReentrantLock+Condition，并且帮我们实现了库存队列，方便简洁
1、定义生产者

class Producer implements Runnable {
    // 库存队列
    private BlockingQueue<String> stock;
    // 生产/消费延迟
    private int timeOut;
    private String name;

    public Producer(BlockingQueue<String> stock, int timeout, String name) {
        this.stock = stock;
        this.timeOut = timeout;
        this.name = name;
    }

    @Override
    public void run() {
        while (true) {
            try {
                stock.put(name);
                System.out.println(name + " 正在生产数据" + " -- 库存剩余：" + stock.size());
                TimeUnit.MILLISECONDS.sleep(timeOut);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }
}

2、定义消费者

class Consumer implements Runnable {
    // 库存队列
    private BlockingQueue<String> stock;
    private String consumerName;

    public Consumer(BlockingQueue<String> stock, String name) {
        this.stock = stock;
        this.consumerName = name;
    }

    @Override
    public void run() {
        while (true) {
            try {
                // 从库存消费一台电脑
                String takeName = stock.take();
                System.out.println(consumerName + " 正在消费数据：" + takeName + " -- 库存剩余：" + stock.size());
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }
}

3、定义库存并运行

public static void main(String[] args) {
        // 定义最大库存为10
        BlockingQueue<String> stock = new ArrayBlockingQueue<>(10);
        Thread p1 = new Thread(new Producer(stock, 500, "Mac"));
        Thread p2 = new Thread(new Producer(stock, 500, "Dell"));
        Thread c1 = new Thread(new Consumer(stock,"zhangsan"));
        Thread c2 = new Thread(new Consumer(stock, "李四"));

        p1.start();
        p2.start();
        c1.start();
        c2.start();

    }

运行结果图：

感谢阅读~欢迎指正和补充~~~