美文网首页
Java并发实践

Java并发实践

作者: 落日无风 | 来源:发表于2017-11-16 14:05 被阅读24次

    Java Concurrency

    在多线程环境下,为了保证共享数据的原子和内存可见性,需要进行锁操作。在JAVA中提供了内置锁和显示锁。本文使用用例结合,来介绍以下锁的用法:

    内置锁(synchronized)

    • 内置锁用来锁定代码块,在进入代码的时候获取锁定,在退出(或者异常退出)释放锁定。内置锁是互斥的,意味中同一时刻只能有一个线程获取该锁,其它线程只能等待或者阻塞直到锁的释放。如下面代码中,假如线程1执行addOne操作,当线程2调用getOne时,就需要等待线程1执行完成并释放锁。
        public class ProductPool {
            private Integer product = new Integer(0);
    
            public synchronized Integer getProduct() {
                return product;
            }
    
            public synchronized void addOne() {
                this.product = this.product + 1;
                LOG.info("produce value: {}", this.product);
            }
    
            public synchronized Integer getOne() {
                Integer old = new Integer(this.product);
                this.product = this.product - 1;
                return old;
            }
        }
    
    • 内置锁是可以重入的。当线程A获取锁执行某操作,如果在当前线程A内,某个步骤也需要获取该锁,该步骤是可以获取到锁的。如下例子,当ChildClass的对象执行doPrint时已经获取到了锁,内部继续调用super.doPrint,如果不能重入就会发生死锁。在同一线程内,锁可以重入。
    
        public class SynchronizedDeakLock {
            private static final Logger LOG = LoggerFactory.getLogger(SynchronizedLock.class);
    
            public class BaseClass {
                public synchronized void doPrint() {
                    LOG.info("base class print");
                }
            }
        
            public class ChildClass extends BaseClass {
                @Override
                public synchronized void doPrint() {
                    LOG.info("child class do print");
                    super.doPrint();
                }
            }
        
            @Test
            public void testDeadLock() {
                ChildClass childClass = new ChildClass();
                childClass.doPrint();
            }
        }
    
    • 内置锁使用非常简单,在需要同步的方法、代码块上加入synchronized就行了,不需要显示的获取和释放锁。且内置锁是JVM内置的,它可以执行部分优化,比如在线程封闭锁对象(该对象使用了锁,但是却不是共享对象,只在某一个线程使用)的锁消除,改变锁的粒度来消除内置锁的同步等。
    • 在某些情况下,我们希望获取锁但又不想一直等待,所以我们指定获取到锁的最大时间,如果获取不到就超时。内置锁对这种细粒度的控制是不支持的,JAVA提供了一种新的锁机制:显示锁。下章,我们就对该话题进行讨论。

    ReentrantLock

    ReentrantLock是JAVA 5提供的细粒度的锁,作为内置锁在某些场景的补充。比如:支持线程获取锁的时间设置,支持获取锁线程对interrupt事件响应。但是在使用时必须显示的获取锁,然后在finally中释放。如果不释放,相当于在程序中放置了个定时炸弹,后期很难发现。它实现了Lock的以下API(部分例子为了达到测试效果没有unlock, 实际使用中绝对不能这样):

    1 . void lock() 获取锁,一致等待直到获取。下面的例子中,在主线程中获取锁且不释放, 子线程调用lock方法来获取锁。可以看到,子线程一致处于RUNNABLE状态,即使它被interrupt。

        @Test
        public void testLockWithInterrupt() throws InterruptedException {
            final Lock lock = new ReentrantLock();
            lock.lock();
            Thread childThread = new Thread(() -> {
                   lock.lock();
                }, "t1 thread");
            childThread.start();
            childThread.interrupt();
    
            LOG.info("the child thread state: {}", childThread.getState().name());
            assertFalse(childThread.isInterrupted());
        }
    

    2 . void lockInterruptibly() throws InterruptedException; 获取锁直到线程被interrupt, 线程抛出InterruptedException。下面的例子中,主线程获取锁且不释放,子线程调用lockInterruptibly方法来获取锁。首先在子线程获取不到锁的时候,会处于一直等待状态;当主线程中调用子线程interrupt时,子线程会抛出InterruptedException。

        @Test(expected = InterruptedException.class)
        public void testLockInterruptibly() throws Exception {
            final Lock lock = new ReentrantLock();
            lock.lock();
            Thread.sleep(1000);
            Thread mainThread = Thread.currentThread();
            Thread t1=new Thread(new Runnable(){
                @Override
                public void run() {
                    try {
                        lock.lockInterruptibly();
                    } catch (InterruptedException e) {
                        LOG.error(" thread interrupted: {}", e);
                        mainThread.interrupt();
                    }
                }
            }, "t1 thread");
            t1.start();
            Thread.sleep(1000);
            t1.interrupt();
            Thread.sleep(1000000);
        }
    

    3 . boolean tryLock() 获取锁,如果获取不到则立即返回false。

        @Test
        public void testTryLock() throws InterruptedException {
            CountDownLatch countDownLatch = new CountDownLatch(1);
            ReentrantLock reentrantLock = new ReentrantLock();
            Runnable runnable = () -> {
                reentrantLock.lock();
                try {
                    Thread.sleep(2 * 1000l);
                    countDownLatch.countDown();
                } catch (InterruptedException e) {
                    e.printStackTrace();
                } finally {
                    reentrantLock.unlock();
                }
            };
        
            Runnable interruptRunnable = () -> {
                boolean result = reentrantLock.tryLock();
                if (result) {
                    LOG.info("lock success");
                    reentrantLock.unlock();
                } else {
                    LOG.info("lock failed");
                }
            };
        
            new Thread(runnable).start();
            new Thread(interruptRunnable).start();
            countDownLatch.await();
        }
        
    

    4 . boolean tryLock(long time, TimeUnit unit) throws InterruptedException 在指定的时间内获取锁,且返回结果。

    @Test
    public void testTryLockWithTime() throws InterruptedException, ExecutionException {
        final Lock lock = new ReentrantLock();
        lock.lock();
        CompletableFuture<Boolean> completableFuture = CompletableFuture.supplyAsync(() -> tryLock(lock));
        assertFalse(completableFuture.get().booleanValue());
    }
    
    private boolean tryLock(Lock lock) {
        try {
            boolean result = lock.tryLock(100, TimeUnit.MILLISECONDS);
            LOG.info("lock result: {}", result);
            return result;
        } catch (InterruptedException e) {
            LOG.error("interrupted: {}", e);
        }
        return false;
    }
        
    

    Semaphore

    信号量常常用来控制对某一资源的访问数量。例如,下面的测试中我们设置信号量的permits为5,当其中5个现在获取且没释放,其它访问线程是获取不到permit的。

    @Test
    public void testSemaphore() throws InterruptedException {
        Semaphore semaphore = new Semaphore(5);
        CountDownLatch countDownLatch = new CountDownLatch(2000);
        Executor executor = Executors.newFixedThreadPool(10);
    
        Runnable runnable = () -> {
            boolean isAcquired = semaphore.tryAcquire();
            if (isAcquired) {
                try {
                    LOG.info("semaphore is acquired");
                    TimeUnit.MICROSECONDS.sleep(2);
                } catch (InterruptedException ex) {
                    LOG.error("error: {}", ex);
                } finally {
                    semaphore.release();
                }
            } else {
                LOG.info("semaphore is not acquired");
            }
            countDownLatch.countDown();
        };
        IntStream.range(1, 2001).forEach(i ->
                executor.execute(runnable)
        );
        countDownLatch.await();
    }
    

    线程池(Thread pool)

    线程池中的任务相对独立,才能使它的性能达到最优。在线程池中,如果出现相互依赖的线程,这可能导致线程死锁。比如:我们开启一个只有1个线程的线程池,调用A任务时,A开始了B任务。然后A任务依赖B任务的完成。在实际执行中,A使用了线程池中的线程,B任务不能获取线程执行,导致A任务不停的处于等待,而B任务也在等待A释放线程。

        @Test
        public void testThreadPoolThreadDependency() {
            ExecutorService executor = Executors.newSingleThreadExecutor();
            Callable<String> stringRunnable = () -> {
                return "test";
            };
            Callable<String> runnable = () -> {
                Future<String> result = executor.submit(stringRunnable);
                try {
                    return result.get();
                } catch (InterruptedException e) {
                    return null;
                } catch (ExecutionException e) {
                    return null;
                }
            };
            try {
                LOG.info(executor.submit(runnable).get());
            } catch (InterruptedException e) {
                e.printStackTrace();
            } catch (ExecutionException e) {
                e.printStackTrace();
            }
        }
    
    

    运行上面测试,会发现处于一直等待的情况,查看thread dump:

    "pool-1-thread-1" #11 prio=5 os_prio=31 tid=0x00007fd6d606f000 nid=0x5703 waiting on condition [0x0000000122af2000]
       java.lang.Thread.State: WAITING (parking)
            at sun.misc.Unsafe.park(Native Method)
            - parking to wait for  <0x0000000795f453d8> (a java.util.concurrent.FutureTask)
            at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
            at java.util.concurrent.FutureTask.awaitDone(FutureTask.java:429)
            at java.util.concurrent.FutureTask.get(FutureTask.java:191)
            at com.eyesee.concurrency.threadpool.ThreadPoolExecutorTest.lambda$testThreadPoolThreadDependency$1(ThreadPoolExecutorTest.java:25)
            at com.eyesee.concurrency.threadpool.ThreadPoolExecutorTest$$Lambda$2/183264084.call(Unknown Source)
            at java.util.concurrent.FutureTask.run(FutureTask.java:266)
            at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
            at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
            at java.lang.Thread.run(Thread.java:745)
    
       Locked ownable synchronizers:
            - <0x0000000795f22fd0> (a java.util.concurrent.ThreadPoolExecutor$Worker)
    

    源代码详见:https://github.com/jessepys/javaconcurrency

    相关文章

      网友评论

          本文标题:Java并发实践

          本文链接:https://www.haomeiwen.com/subject/tmulvxtx.html