ReentrantLock 公平锁源码 第2篇

Reentrant 2

前两篇写完了后我自己研究了下,还有有很多疑惑和问题,这篇就继续以自问自答的方式写

如果没看过第1篇的可以先看看那个https://www.cnblogs.com/sunankang/p/16458795.html

public final void acquire(int arg) {
    if (!tryAcquire(arg) &&
        acquireQueued(addWaiter(Node.EXCLUSIVE), arg))
        selfInterrupt();
}

进入 acquireQueued方法

final boolean acquireQueued(final Node node, int arg) {
    boolean failed = true;
    try {
        //这个属性的作用是啥???

        boolean interrupted = false;
        for (;;) {
            final Node p = node.predecessor();
            if (p == head && tryAcquire(arg)) {
                setHead(node);
                p.next = null; // help GC
                failed = false;
                return interrupted;
            }
            if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())
                interrupted = true;
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

第一个问题

interrupted这个变量的作用

private final boolean parkAndCheckInterrupt() {
    LockSupport.park(this);
    return Thread.interrupted();
}

在 parkAndCheckInterrupt方法中最后return的是这个线程是否被打断,它的作用是啥?

先来回顾 interrupt()， interrupted() 和 isInterrupted()三者区别,长得很像,注意区分

interrupt()的作用是中断线程,如果被中断的线程处于阻塞状态下,例如调用 wait(), join() sleep(),则抛出异常,否则只是设置一个中断标记为true, 注意:仅仅是设置中断状态为true,并不会去 “中断” 线程

interrupted() 获取线程的中断状态并且清空中断状态(将中断状态设置为false)

isInterrupted() 获取线程的中断状态并不会清除中断状态

调用 interrupt 会使park方法立即结束,可以理解为唤醒

继续代码,看这个变量最后到了哪里

情况1 没有被打断过

假设线程没有被中断过,那么 parkAndCheckInterrupt返回就是false

final boolean acquireQueued(final Node node, int arg) {
    boolean failed = true;
    try {
        boolean interrupted = false;
        for (;;) {
            final Node p = node.predecessor();
            if (p == head && tryAcquire(arg)) {
                setHead(node);
                p.next = null; // help GC
                failed = false;
                return interrupted;
            }
            if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())
                interrupted = true;
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

那么不进入 if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())这个if,获取到锁后返回false,回到 acquire方法

public final void acquire(int arg) {
    if (!tryAcquire(arg) && acquireQueued(addWaiter(Node.EXCLUSIVE), arg))
        selfInterrupt();
}

因为false,所以不进入 selfInterrupt(),方法结束

情况2 park或准备park,被唤醒后直接获取到了锁

先证明一下打断是会唤醒park中的线程的

我就再重复粘一下代码了,方便看

private final boolean parkAndCheckInterrupt() {
    LockSupport.park(this);
    return Thread.interrupted();
}

那么返回的就是true,回到上级 acquireQueued方法

final boolean acquireQueued(final Node node, int arg) {
    boolean failed = true;
    try {
        boolean interrupted = false;
        for (;;) {
            final Node p = node.predecessor();
            if (p == head && tryAcquire(arg)) {
                setHead(node);
                p.next = null; // help GC
                failed = false;
                return interrupted;
            }
            //返回到这里
            if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())
                interrupted = true;
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

因为返回true,所以进入 if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt()) 将interrupted返回true

假设循环获取到锁,那么再返回上一级 acquire()

public final void acquire(int arg) {
    if (!tryAcquire(arg) && acquireQueued(addWaiter(Node.EXCLUSIVE), arg))
        selfInterrupt();
}

那么进入 selfInterrupt()

static void selfInterrupt() {
    Thread.currentThread().interrupt();
}

是不是有点疑惑?我如果没有调用过 interrupt() 那ReentrantLock就不做任何操作,我如果调用了,那它再给我调用一次 ???? 还有情况3

情况3 park或准备park,被唤醒后没有获取到锁

final boolean acquireQueued(final Node node, int arg) {
    boolean failed = true;
    try {
        boolean interrupted = false;
        for (;;) {
            final Node p = node.predecessor();
            if (p == head && tryAcquire(arg)) {
                setHead(node);
                p.next = null; // help GC
                failed = false;
                return interrupted;
            }
            //假设在调用shouldParkAfterFailedAcquire成功后,马上就要调用parkAndCheckInterrupt 时间片用完了
            if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())
                interrupted = true;
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

那么这个时候interrupted属性就有用了

首先要知道一点, 一个被中断的线程是无法park的,除非清除了中断状态,即设置为将中断状态设置为false, 口说无凭,直接上图

第二张图还是在park状态,证明了被打断的线程是无法park的,除非将它中断状态设置为false

那么回到代码中就能知道这个的作用

final boolean acquireQueued(final Node node, int arg) {
    boolean failed = true;
    try {
        boolean interrupted = false;
        for (;;) {
            final Node p = node.predecessor();
            if (p == head && tryAcquire(arg)) {
                setHead(node);
                p.next = null; // help GC
                failed = false;
                return interrupted;
            }
            if (shouldParkAfterFailedAcquire(p, node) &&
                parkAndCheckInterrupt())
                interrupted = true;
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

如果线程被打断唤醒,还是在 for(;;)中,还是去获取锁,假设没有获取到呢?那么就一直在for循环中嘎嘎跑,因为线程的状态是被中断的,无法再次park了

private final boolean parkAndCheckInterrupt() {
    LockSupport.park(this);
    return Thread.interrupted();
}

那么现在懂了最后的 Thread.interrupted()作用了吗,就是将中断状态设置回false,好让线程没有获取到锁继续park

那这时候可能就问了:那你ReentrantLock把中断状态给我清空了,我自己如果有需要根据中断状态来判断的代码咋办啊?

好,咱们从park先被打断来捋一下

private final boolean parkAndCheckInterrupt() {
    LockSupport.park(this);
    return Thread.interrupted();
}

因为被打断,线程醒来,执行 Thread.interrupted()并清空中断状态,返回true

final boolean acquireQueued(final Node node, int arg) {
    boolean failed = true;
    try {
        boolean interrupted = false;
        for (;;) {
            final Node p = node.predecessor();
            if (p == head && tryAcquire(arg)) {
                setHead(node);
                p.next = null; // help GC
                failed = false;
                return interrupted;
            }
            if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())
                 //进入这里
                interrupted = true;
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

因为返回的是true,所以进入 if (shouldParkAfterFailedAcquire(p, node) && parkAndCheckInterrupt())的代码块,将 interrupted属性设置为true

那么 for(;;)循环再来一次,如果没有获取到锁.继续park,直到被唤醒,走 tryAcquire()获取到为止,那么此时 interrupted变量就为true了

public final void acquire(int arg) {
    if (!tryAcquire(arg) && acquireQueued(addWaiter(Node.EXCLUSIVE), arg))
        selfInterrupt();
}

那么退出 acquireQueued()方法回到 acquire()中,因为 acquireQueued()返回的是true,所以进入 selfInterrupt()

static void selfInterrupt() {
    Thread.currentThread().interrupt();
}

所以懂了吗?

Original: https://www.cnblogs.com/sunankang/p/16461261.html
Author: Jame!
Title: ReentrantLock 公平锁源码 第2篇