多線程（十、AQS原理-ReentrantLock公平鎖）

ReentrantLock介紹

ReentrantLock 基于AQS實現(xiàn)了公平和非公平的獨占鎖功能。

ReentrantLock定義AQS的同步狀態(tài)（synchronization state）如下：

State為0表示鎖可用；為1表示被占用；為N表示鎖重入的次數，是獨占資源。

ReentrantLock實現(xiàn)公平鎖原理

案例代碼如下：

1、啟動文件

public class Main {

    public static void main(String[] args) throws ParseException {

        ReentrantLock lock = new ReentrantLock(true);
        Thread t1 = new Thread(new Task(lock),"Thread-1");
        Thread t2 = new Thread(new Task(lock),"Thread-2");
        Thread t3 = new Thread(new Task(lock),"Thread-3");

        t1.start();
        t2.start();
        t3.start();
    }
}

2、Task

import java.util.concurrent.locks.ReentrantLock;

public class Task implements Runnable{

    private ReentrantLock lock;

    public Task(ReentrantLock lock) {
        this.lock = lock;
    }

    @Override
    public void run() {

        try {
            lock.lock();
            System.out.println(Thread.currentThread().getName() + "獲取到鎖....");
            Thread.sleep(2000);
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            System.out.println(Thread.currentThread().getName() + "釋放鎖....");
            lock.unlock();
        }
    }
}

3、執(zhí)行結果：

案例分析

1、Thread-1調用lock方法

ReentrantLock內部繼承AQS實現(xiàn)了1個抽象類Sync

繼承于Sync實現(xiàn)了2個內部類FairSync（公平的）和NonfairSync（非公平的）

此時調用FairSync的lock方法

acquire方法來自AQS，注意參數是1

tryAcquire是ReentrantLock自己實現(xiàn)的，嘗試獲取鎖

protected final boolean tryAcquire(int acquires) { //參數是1
        final Thread current = Thread.currentThread(); //當前線程
        int c = getState(); //獲取當前同步狀態(tài)
        if (c == 0) { //如果是0，則鎖沒有被占用
            //等待隊列中，前面沒有其他等待的線程，則用CAS的方法更新同步狀態(tài)state
            if (!hasQueuedPredecessors() &&
                    compareAndSetState(0, acquires)) {
                setExclusiveOwnerThread(current); //成功的話，則設置鎖的占有線程為當前線程
                return true; //返回獲取資源成功
            }
        }
        //如果鎖已經被占用，則判斷是不是自己占用的
        else if (current == getExclusiveOwnerThread()) {
            int nextc = c + acquires;//如果是自己占用的，則是重入，增加state值，累加1
            if (nextc < 0) //重入次數過大，拋出異常
                throw new Error("Maximum lock count exceeded");
            setState(nextc); //設置state值
            return true; //重入返回ture
        }
        return false;//沒有獲取資源返回false
    }

Thread-1獲取了鎖資源，沒有釋放。

2、Thread-2，開始請求資源，調用lock，此時鎖資源還被Thread-1占用

addWaiter方法：

private Node addWaiter(Node mode) {
        //把當前線程包裝成節(jié)點，準備放入等待隊列
        Node node = new Node(Thread.currentThread(), mode);
        // Try the fast path of enq; backup to full enq on failure
        //嘗試直接把節(jié)點設置成隊尾，否則執(zhí)行enq
        Node pred = tail;
        if (pred != null) {
            node.prev = pred;//當前節(jié)點的上一個節(jié)點是之前的隊尾節(jié)點
            if (compareAndSetTail(pred, node)) {
                pred.next = node;
                return node;
            }
        }
        //當前節(jié)點插入隊尾
        enq(node);
        return node;
    }

enq自旋+初始化等待隊列，并返回Thread-2節(jié)點

private Node enq(final Node node) {
        //采用自旋，保證節(jié)點插入
        for (;;) {
            Node t = tail;
            if (t == null) { // Must initialize 如果隊列為空，則創(chuàng)建一個空的節(jié)點，設置為頭尾節(jié)點
                if (compareAndSetHead(new Node()))
                    tail = head;
            } else {
                node.prev = t;
                if (compareAndSetTail(t, node)) { //隊列不為空，追加到隊尾
                    t.next = node;
                    return t;
                }
            }
        }
    }

然后對Thread-2包裝節(jié)點執(zhí)行acquireQueued

final boolean acquireQueued(final Node node, int arg) {
        boolean failed = true;
        try {
            boolean interrupted = false;
            for (;;) {
                //判斷節(jié)點的前任節(jié)點是不是頭節(jié)點，頭節(jié)點是一個空節(jié)點
                final Node p = node.predecessor();
                //如果是頭節(jié)點，則說明當前節(jié)點是隊列里的第一個節(jié)點，首節(jié)點。
                //則嘗試獲取鎖資源，此處因為Thread-1占用著資源，則失敗
                if (p == head && tryAcquire(arg)) {
                    setHead(node);
                    p.next = null; // help GC
                    failed = false;
                    return interrupted;
                }
                //失敗之后，則判斷當前節(jié)點線程Thread-2是不是可以阻塞
                if (shouldParkAfterFailedAcquire(p, node) &&
                        parkAndCheckInterrupt())
                    interrupted = true;
            }
        } finally {
            if (failed)
                cancelAcquire(node);
        }
    }

是否阻塞shouldParkAfterFailedAcquire

private static boolean shouldParkAfterFailedAcquire(Node pred, Node node) {
        int ws = pred.waitStatus;//前驅節(jié)點的狀態(tài)
        if (ws == Node.SIGNAL) //如果是SIGNAL，則說明前驅節(jié)點狀態(tài)可以喚醒后繼節(jié)點，可以阻塞
            /*
             * This node has already set status asking a release
             * to signal it, so it can safely park.
             */
            return true;
        if (ws > 0) {
            /*
             * Predecessor was cancelled. Skip over predecessors and
             * indicate retry.
             */
            do {
                node.prev = pred = pred.prev; //只有CANCELLED狀態(tài)大于0，則把取消狀態(tài)的節(jié)點從隊列刪除
            } while (pred.waitStatus > 0);
            pred.next = node;
        } else {
            /*
             * waitStatus must be 0 or PROPAGATE.  Indicate that we
             * need a signal, but don't park yet.  Caller will need to
             * retry to make sure it cannot acquire before parking.
             */
            compareAndSetWaitStatus(pred, ws, Node.SIGNAL);//設置前驅節(jié)點為SIGNAL狀態(tài)
        }
        return false;
    }

如果可以阻塞，則調用parkAndCheckInterrupt，阻塞線程，至此Thread-2進入隊列，并阻塞了，耐心等待

3、Thread-3同Thread-2，略過

4、Thread-1釋放鎖資源

release方法：

tryRelease

然后通過unparkSuccessor，喚醒首節(jié)點，保證公平策略。

至此Thread-1釋放完成，Thread-2可以獲得資源，依次類推。