美文网首页
JAVA并发(11)——ReentrantReadWriteLo

JAVA并发(11)——ReentrantReadWriteLo

作者: alexwu59 | 来源:发表于2018-07-03 15:05 被阅读0次

    ReentrantReadWriteLock的使用

    ReentrantReadWriteLock是JDK提供的读写锁机制,写锁是排他锁,读锁是共享锁。

    //创建一个读写锁
    ReentrantReadWriteLock reentrantLock = new ReentrantReadWriteLock();
    //根据读写锁对象分别创建写锁和读锁
    Lock w = reentrantLock.writeLock();
    Lock r = reentrantLock.readLock();
    

    ReentrantReadWriteLock把AQS表示锁状态的字段state逻辑上分为了两个部分:高16位是同一个锁被获取共享锁(读锁)的次数,低16位是同一个锁被获取排他锁(写锁)的次数。
    默认情况下,创建ReentrantReadWriteLock对象也是按照非公平策略:

    public ReentrantReadWriteLock(boolean fair) {
            //sync 的实现与ReentrantLock一样
            sync = fair ? new FairSync() : new NonfairSync();
            readerLock = new ReadLock(this);
            writerLock = new WriteLock(this);
        }
    

    由下面代码可以知道,ReadLock和WriteLock使用的相同的sync变量:

    readerLock = new ReadLock(this);
    writerLock = new WriteLock(this);
    
    protected WriteLock(ReentrantReadWriteLock lock) {
                sync = lock.sync;
    }
    protected ReadLock(ReentrantReadWriteLock lock) {
                sync = lock.sync;
    }
    

    写锁加锁

    执行写锁w.lock()方法,来获取写锁,w.lock方法的实现最终会调用Sync类的tryAcquire方法
    写锁的tryAcquire方法实现如下:

    protected final boolean tryAcquire(int acquires) {
                Thread current = Thread.currentThread();
                //获取锁状态
                int c = getState();
                //获取排他锁的个数,也就是写锁的个数
                int w = exclusiveCount(c);
                //锁状态个数不为0,且写锁为0或者持有锁的线程不是当前线程,那么失败
                if (c != 0) {
                    // (Note: if c != 0 and w == 0 then shared count != 0)
                    if (w == 0 || current != getExclusiveOwnerThread())
                        return false;
                    if (w + exclusiveCount(acquires) > MAX_COUNT)
                        throw new Error("Maximum lock count exceeded");
                    // 如果是当前线程之前已经获取该锁
                    setState(c + acquires);
                    return true;
                }
                if (writerShouldBlock() ||
                    !compareAndSetState(c, c + acquires))
                    return false;
                setExclusiveOwnerThread(current);
                return true;
    }
    

    主要逻辑:
    1.判断锁状态
    2.计算获取写锁的次数
    (1)当前锁状态为0,执行 (2),否则执行下面
    a 如果当前锁状态是读锁或者 b当前锁状态是写锁,但是持有锁的线程不是当前线程,那么获取失败,返回false
    (2)设置锁的装状态,如果成功这设置获取锁的线程为当前线程,返回true。否则获取锁失败,返回false

    如果写锁的tryAcquire方法返回true
    那么会把当前线程封装为一个Node节点对象,把该node节点对象加入AQS的阻塞队列中。

    写锁释放

    写锁的释放,会调用ReentrantReadWriteLock内部静态类Sync的tryRelease方法,下面是tryRelease具体实现:

    protected final boolean tryRelease(int releases) {
               //判断当前申请释放锁的线程是不是持有锁的线程
                if (!isHeldExclusively())
                    throw new IllegalMonitorStateException();
                int nextc = getState() - releases;
                //判断排它锁值状态是不是0
                boolean free = exclusiveCount(nextc) == 0;
                if (free)
                    setExclusiveOwnerThread(null);
                setState(nextc);
                return free;
     }
    

    当tryRelease返回true之后,此时没有线程持有锁,接着就可以从AQS的阻塞队列中释放一个线程去执行:

     public final boolean release(int arg) {
            if (tryRelease(arg)) {
                Node h = head;
                if (h != null && h.waitStatus != 0)
                    unparkSuccessor(h); //从阻塞对列中获取一个线程
                return true;
            }
            return false;
    }
    

    读锁获取

    读锁也被称为共享锁,获取共享锁的方法tryAcquireShared实现如下

    protected final int tryAcquireShared(int unused) {
                Thread current = Thread.currentThread();
                int c = getState();
               //排它锁不等于0并且拥有排他锁的线程不是当前线程,那么返回-1
                if (exclusiveCount(c) != 0 &&
                    getExclusiveOwnerThread() != current)
                    return -1;
                //获取共享锁数量
                int r = sharedCount(c);
                //这里有个readerShouldBlock()是为了避免在非公平策略下写锁一直处于饥饿状态,
               //readerShouldBlock实现主要是判断如果当阻塞队列的第一个等待锁的线程是写操作
               //那么就要阻塞当前读线程,使得让写线程能有机会获取到写锁
                if (!readerShouldBlock() &&
                    r < MAX_COUNT &&
                    compareAndSetState(c, c + SHARED_UNIT)) {
                    if (r == 0) {
                        firstReader = current;
                        firstReaderHoldCount = 1;
                    } else if (firstReader == current) {
                        firstReaderHoldCount++;
                    } else {
                        //如果当前锁是读锁,但是持有锁的线程不是当前线程,执行下面的操作
                       //给获取一个存储在ThreadLocal对象中的缓存对象HoldCounter,该值主要是存储
                       //每个线程获取读锁的个数
                        HoldCounter rh = cachedHoldCounter;
                        if (rh == null || rh.tid != getThreadId(current))
                            cachedHoldCounter = rh = readHolds.get();//readHolds 是Threadlocal对象
                        else if (rh.count == 0)
                            readHolds.set(rh);
                        rh.count++;
                    }
                    return 1;
                }
                //下面的方法是处理CAS失败或者是在tryAcquireShared没有处理的重入读锁
                return fullTryAcquireShared(current);
    }
    

    线程从tryAcquireShared方法中返回大于0表示成功获取到共享锁,如果返回小于0表示获取共享锁失败,那么就需要执行doAcquireShared()方法,把当前线程加入到等待队列中。

    读锁释放

    读锁释放的主要逻辑:
    1.判断获取读锁的第一个线程是不是当前线程,如果是修改锁数量,然后执行3,如果不是当前线程,执行2
    2.获取当前线程自己的HoldCounter 对象,修改HoldCounter 中读锁数量值,执行3
    3.使用CAS修改AQS的锁状态信息,并返回AQS的缩状态

    protected final boolean tryReleaseShared(int unused) {
                Thread current = Thread.currentThread();
                if (firstReader == current) {
                    // assert firstReaderHoldCount > 0;
                    if (firstReaderHoldCount == 1)
                        firstReader = null;
                    else
                        firstReaderHoldCount--;
                } else {
                    HoldCounter rh = cachedHoldCounter;
                    if (rh == null || rh.tid != getThreadId(current))
                        rh = readHolds.get();
                    int count = rh.count;
                    if (count <= 1) {
                        readHolds.remove();
                        if (count <= 0)
                            throw unmatchedUnlockException();
                    }
                    --rh.count;
                }
                for (;;) {
                    int c = getState();
                    int nextc = c - SHARED_UNIT;
                    if (compareAndSetState(c, nextc))
                        // Releasing the read lock has no effect on readers,
                        // but it may allow waiting writers to proceed if
                        // both read and write locks are now free.
                        return nextc == 0;
                }
    }
    

    相关文章

      网友评论

          本文标题:JAVA并发(11)——ReentrantReadWriteLo

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