使用递增计数器的线程同步工具 —— 信号量，它的原理是什么样子的？

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"前言"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在 JUC 中线程同步器除了 CountDownLatch 和 CycleBarrier ，还有一个叫做 Semaphore （信号量），同样是基于 AQS 实现的。下面来看看信号量的内部原理。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"公众号：liuzhihangs，记录工作学习中的技术、开发及源码笔记；时不时分享一些生活中的见闻感悟。欢迎大佬来指导！"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"介绍"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"一个计数信号量。 从概念上讲，信号量维护了一组许可。 如果有必要，在许可可用之前调用 acquire 方法会被阻塞，直到许可证可用。 调用 release 方法会增加了一个许可证，从而释放被阻塞的线程。 "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"numberedlist","attrs":{"start":"1","normalizeStart":1},"content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":1,"align":null,"origin":null},"content":[{"type":"text","text":"声明时指定初始许可数量。"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":2,"align":null,"origin":null},"content":[{"type":"text","text":"调用 acquire(int permits) 方法，指定目标许可数量。"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":3,"align":null,"origin":null},"content":[{"type":"text","text":"调用 release(int permits) 方法，发布指定的许可数量。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在许可数量没有到达指定目标数量时，调用 acquire 方法的线程会被阻塞。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"基本使用"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"public class SemaphoreTest1 {\n\n private static final Semaphore SEMAPHORE = new Semaphore(0);\n\n public static void main(String[] args) throws InterruptedException {\n\n ExecutorService pool = new ThreadPoolExecutor(5, 5, 0L, TimeUnit.MILLISECONDS,\n new LinkedBlockingQueue<>(1024),\n new ThreadFactoryBuilder().setNameFormat(\"Thread-pool-%d\").build(),\n new ThreadPoolExecutor.AbortPolicy());\n\n\n for (int i = 0; i < 5; i++) {\n\n pool.submit(() -> {\n\n try {\n Thread.sleep(1000 + new Random().nextInt(1000));\n } catch (InterruptedException ignored) {\n }\n\n System.out.println(\"当前线程: \" + Thread.currentThread().getName() + \" 发布一个许可\");\n SEMAPHORE.release(1);\n\n });\n }\n\n System.out.println(\"-----> 这里是主线程\");\n\n SEMAPHORE.acquire(5);\n\n System.out.println(\"-----> 主线程执行完毕\");\n\n pool.shutdown();\n }\n\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"bash"},"content":[{"type":"text","text":"-----> 这里是主线程\n当前线程: Thread-pool-2 发布一个许可\n当前线程: Thread-pool-4 发布一个许可\n当前线程: Thread-pool-1 发布一个许可\n当前线程: Thread-pool-0 发布一个许可\n当前线程: Thread-pool-3 发布一个许可\n-----> 主线程执行完毕"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"上面这个方法也是模拟了类似 CountDownLatch 的用法， 在子线程执行完毕之后，主线程继续执行。只不过 Semaphore 和 CountDownLatch 区别最大的是："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Semaphore 是从指定数值开始增加，直到到达许可数量，然后被阻塞线程开始继续执行。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"CountDownLatch 是从指定数量的线程开始减少，直到为 0 时，被阻塞的线程开始继续执行。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"当然这只是最简单的用法，除此让主线程等待，同样也可以让其他线程等待，然后再开始执行。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"问题疑问"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"numberedlist","attrs":{"start":"1","normalizeStart":1},"content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":1,"align":null,"origin":null},"content":[{"type":"text","text":"Semaphore 和 AQS 有什么关系？"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":2,"align":null,"origin":null},"content":[{"type":"text","text":"Semaphore 和 CountDownLatch 有什么区别？"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"源码分析"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"基本结构"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/6f/6f4faaa831bd71d626b257aaf6b06300.png","alt":"Semaphore-cover-iGaTzJ","title":null,"style":null,"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通过类图可以看出在 Semaphore 里面有一个静态内部类 Sync 继承了 AQS，同时为了区分公平和非公平的情况，Sync 分别有两个子类：NonfairSync 、FairSync。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"下面根据案例分别从构造函数、acquire()、release() 入手，从而了解内部实现原理。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"初始化"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"public Semaphore(int permits) {\n sync = new NonfairSync(permits);\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"初始化默认非公平锁， 同时需要传入指定许可数， 可以看到这块代码是调用的 AQS 的 setState(permits) 方法。代码如下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"static final class NonfairSync extends Sync {\n private static final long serialVersionUID = -2694183684443567898L;\n\n NonfairSync(int permits) {\n super(permits);\n }\n}\n\nabstract static class Sync extends AbstractQueuedSynchronizer {\n private static final long serialVersionUID = 1192457210091910933L;\n\n Sync(int permits) {\n setState(permits);\n }\n }\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"setState 方法其实就是对 AQS 的 state 进行赋值。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"补充"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"1. 在 ReentrantLock 中 state 代表加锁状态，0 没有线程获得锁，大于等于 1 已经有线程获得锁，大于 1 说明该获得锁的线程多次重入。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"2. 在 ReentrantReadWriteLock 中 state 代表锁的状态。state 为 0 ，没有线程持有锁，state 的高 16 为代表读锁状态，低 16 为代表写锁状态。通过位运算可以获取读写锁的实际值。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"3. 而在这里 （CountDownLatch）则代表门闩或者说计数的值。"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如果对 state 有所遗忘，可以阅读前面的 AQS 、CAS 相关代码。 state 在这里代表的是信号量的许可数量。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"acquire()"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"public void acquire() throws InterruptedException {\n sync.acquireSharedInterruptibly(1);\n}\n\npublic void acquire(int permits) throws InterruptedException {\n if (permits < 0) throw new IllegalArgumentException();\n sync.acquireSharedInterruptibly(permits);\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"acquire() 和 acquire(int permits) 调用的都是 sync.acquireSharedInterruptibly(permits) 方法，只不过一个支持传递参数，一个默认为 1。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"acquireSharedInterruptibly 方法，其实就是 Sync 继承自 AQS 的。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"这块可以阅读 AQS 的文章，这里简单介绍下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"private void doAcquireSharedInterruptibly(int arg)\n throws InterruptedException {\n final Node node = addWaiter(Node.SHARED);\n boolean failed = true;\n try {\n for (;;) {\n final Node p = node.predecessor();\n if (p == head) {\n int r = tryAcquireShared(arg);\n if (r >= 0) {\n setHeadAndPropagate(node, r);\n p.next = null; // help GC\n failed = false;\n return;\n }\n }\n if (shouldParkAfterFailedAcquire(p, node) &&\n parkAndCheckInterrupt())\n throw new InterruptedException();\n }\n } finally {\n if (failed)\n cancelAcquire(node);\n }\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"numberedlist","attrs":{"start":"1","normalizeStart":1},"content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":1,"align":null,"origin":null},"content":[{"type":"text","text":"在失败后会使用 "},{"type":"codeinline","content":[{"type":"text","text":"doAcquireSharedInterruptibly(arg);"}]},{"type":"text","text":" 不断获取资源；"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":2,"align":null,"origin":null},"content":[{"type":"codeinline","content":[{"type":"text","text":"final Node node = addWaiter(Node.SHARED);"}]},{"type":"text","text":" 会创建节点以共享模式放到队列里；"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":3,"align":null,"origin":null},"content":[{"type":"text","text":"在循环中不断判断前一个节点，如果是 head，则尝试获取共享资源；"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":4,"align":null,"origin":null},"content":[{"type":"text","text":"在共享模式下获取到资源后会使用 "},{"type":"codeinline","content":[{"type":"text","text":"setHeadAndPropagate(node, r);"}]},{"type":"text","text":" 设置头节点，同时唤醒后续节点。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"tryAcquireShared 是需要子类实现，也就是在 Semaphore.Sync 的实现类中实现了，这里以 FairSync 做讲解："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"\nstatic final class FairSync extends Sync {\n private static final long serialVersionUID = 2014338818796000944L;\n\n FairSync(int permits) {\n super(permits);\n }\n\n protected int tryAcquireShared(int acquires) {\n for (;;) {\n // 如果前面有节点，则直接返回 -1 表示失败\n if (hasQueuedPredecessors())\n return -1;\n // 获取当前信号量\n int available = getState();\n // 获取当前剩余量\n int remaining = available - acquires;\n // 如果小于 0 或者 CAS 设置信号量成功 则直接返回\n if (remaining < 0 ||\n compareAndSetState(available, remaining))\n return remaining;\n }\n }\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"而这段代码的含义："}]},{"type":"numberedlist","attrs":{"start":"1","normalizeStart":1},"content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":1,"align":null,"origin":null},"content":[{"type":"text","text":"如果前面有节点，则直接阻塞；"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":2,"align":null,"origin":null},"content":[{"type":"text","text":"如果当前剩余信号量小于 0 ，则返回负值，直接阻塞；"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":3,"align":null,"origin":null},"content":[{"type":"text","text":"如果当前剩余量大于等于 0 ，会 CAS 更新信号量，并返回非负数。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":">这块数值的含义，在 AQS 中定义了，含义如下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":">1. 小于 0: 表示失败；"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":">2. 等于 0: 表示共享模式获取资源成功，但后续的节点不能以共享模式获取成功; "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":">3. 大于 0: 表示共享模式获取资源成功，后续节点在共享模式获取也可能会成功，在这种情况下，后续等待线程必须检查可用性。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"release()"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"public void release() {\n sync.releaseShared(1);\n}\npublic void release(int permits) {\n if (permits < 0) throw new IllegalArgumentException();\n sync.releaseShared(permits);\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"发布许可证的给定数量，该数量增加可用的许可数量。 看其内部调用的是 Sync 的 releaseShared， 其实就是 AQS 的对应方法："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"\npublic final boolean releaseShared(int arg) {\n if (tryReleaseShared(arg)) {\n doReleaseShared();\n return true;\n }\n return false;\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如果实现tryReleaseShared返回true，以共享模式释放资源。 其中的 tryReleaseShared 部分由 Semaphore.Sync 中实现，逻辑如下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"java"},"content":[{"type":"text","text":"\nprotected final boolean tryReleaseShared(int releases) {\n for (;;) {\n // 获取当前 state\n int current = getState();\n // 对 state 进行增加\n int next = current + releases;\n if (next < current) // overflow\n throw new Error(\"Maximum permit count exceeded\");\n // 使用 CAS 赋值\n if (compareAndSetState(current, next))\n return true;\n }\n}"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通过上面代码可以看出，在 Semaphore 的 release 方法中主要就是对 state 进行增加，增加成功后会调用 AQS 的 doReleaseShared 方法唤醒头节点。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"总结"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"Q&A"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"Q: 既然 Semaphore 也是基于 AQS， 那在 Semaphore 中 state 的含义代表什么？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"A:"},{"type":"text","text":" 在 Semaphore 中 state 代表许可数量，acquire 方法当许可小于指定数量会阻塞线程，release 方法增加许可当许可增加成功则唤醒阻塞节点。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"Q: Semaphore 基于 AQS 具体是怎么实现的呢？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"A:"},{"type":"text","text":" "}]},{"type":"numberedlist","attrs":{"start":"1","normalizeStart":1},"content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":1,"align":null,"origin":null},"content":[{"type":"text","text":"初始设置 state 的初始值，即初始许可数量。"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":2,"align":null,"origin":null},"content":[{"type":"text","text":"acquire 方法设置目标数量，当目标数量大于当前数量时，会阻塞线程并将其放到阻塞队列中。此处基于 AQS 实现。"}]}]},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":3,"align":null,"origin":null},"content":[{"type":"text","text":"release 对 state 进行增加，成功后会调用 AQS 的 doReleaseShared 唤醒头结点。同样是基于 AQS 实现。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"Q: Semaphore 和 CountDownLatch 有什么区别？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"A:"},{"type":"text","text":" Semaphore 的计数器是递加的，而 CountDownLatch 是递减的。相同点就是计数器都不可以重置。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"结束语"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在阅读 Semaphore 源码过程中，发现其主要功能都是基于 AQS 实现的，可以回顾阅读 AQS 的相关笔记。同样 Semaphore 也支持公平和非公平模式，这块就需要小伙伴自己去阅读啦。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}}]}