Semaphore 源码分析

// Semaphore.java
public void acquire() throws InterruptedException {
    sync.acquireSharedInterruptibly(1);
}

// AbstractQueueSynchronizer.java
public final void acquireSharedInterruptibly(int arg)
        throws InterruptedException {
    if (Thread.interrupted()) /* 可中断的 */
        throw new InterruptedException();
    if (tryAcquireShared(arg) < 0)
        doAcquireSharedInterruptibly(arg);
}

// AbstractQueueSynchronizer.java
protected int tryAcquireShared(int arg) {
    throw new UnsupportedOperationException();
}

// Semaphore.java NonfairSync
protected int tryAcquireShared(int acquires) {
    return nonfairTryAcquireShared(acquires);
}

// Semaphore.java Sync
final int nonfairTryAcquireShared(int acquires) {
    for (;;) { /* 自旋, 直到成功 */
        int available = getState();
        int remaining = available - acquires;
        if (remaining < 0 ||                          /* 可用数量 < 0, 交由后续判断 */
            compareAndSetState(available, remaining)) /* 可用数量 >= 0, 并且保存最新数量成功 */
            return remaining;
    }
}

// AbstractQueueSynchronizer.java
private void doAcquireSharedInterruptibly(int arg)
    throws InterruptedException {
    final Node node = addWaiter(Node.SHARED); /* 该方法保证将节点插入到队尾, 已在 ReentrantLock 里分析过了 */
    boolean failed = true;
    try {
        for (;;) {
            /* 阻塞, 直到成功返回 */
            final Node p = node.predecessor();
            if (p == head) { /* 只有前驱节点是head时, 才能继续 */
                int r = tryAcquireShared(arg); /* 获取可用数量, >=0表示拿到了许可 */
                if (r >= 0) {
                    setHeadAndPropagate(node, r); /* 修改头节点, 并且向后传递共享状态 */
                    p.next = null; // help GC
                    failed = false;
                    return;
                }
            }
            /* 同独占模式, 可以 park 时, 则 park */
            if (shouldParkAfterFailedAcquire(p, node) &&
                parkAndCheckInterrupt())
                throw new InterruptedException();
        }
    } finally {
        if (failed)
            cancelAcquire(node);
    }
}

private void setHeadAndPropagate(Node node, int propagate) {
    Node h = head; // Record old head for check below
    setHead(node); /* 设置头节点 */
    /*
     * Try to signal next queued node if:
     *   Propagation was indicated by caller,
     *     or was recorded (as h.waitStatus either before
     *     or after setHead) by a previous operation
     *     (note: this uses sign-check of waitStatus because
     *      PROPAGATE status may transition to SIGNAL.)
     * and
     *   The next node is waiting in shared mode,
     *     or we don't know, because it appears null
     *
     * The conservatism in both of these checks may cause
     * unnecessary wake-ups, but only when there are multiple
     * racing acquires/releases, so most need signals now or soon
     * anyway.
     */
    if (propagate > 0 || h == null || h.waitStatus < 0 ||
        (h = head) == null || h.waitStatus < 0) {
        Node s = node.next;
        if (s == null || s.isShared())
            doReleaseShared(); /* 我们在 release 小节分析 */
    }
}

// Semaphore.java
public void release() {
    sync.releaseShared(1);
}

// AbstractQueuedSynchronizer.java
public final boolean releaseShared(int arg) {
    if (tryReleaseShared(arg)) {
        doReleaseShared();
        return true;
    }
    return false;
}

protected boolean tryReleaseShared(int arg) {
    throw new UnsupportedOperationException();
}

// Semaphore.java Sync
protected final boolean tryReleaseShared(int releases) {
    for (;;) { /* 不断尝试, 直到设置成功, 或越界 */
        int current = getState();
        int next = current + releases;
        if (next < current) // overflow
            throw new Error("Maximum permit count exceeded");
        if (compareAndSetState(current, next))
            return true;
    }
}

// AbstractQueuedSynchronizer.java
private void doReleaseShared() {
    /*
     * Ensure that a release propagates, even if there are other
     * in-progress acquires/releases.  This proceeds in the usual
     * way of trying to unparkSuccessor of head if it needs
     * signal. But if it does not, status is set to PROPAGATE to
     * ensure that upon release, propagation continues.
     * Additionally, we must loop in case a new node is added
     * while we are doing this. Also, unlike other uses of
     * unparkSuccessor, we need to know if CAS to reset status
     * fails, if so rechecking.
     */
    for (;;) {
        Node h = head;
        if (h != null && h != tail) {
            int ws = h.waitStatus;
            if (ws == Node.SIGNAL) { /* 头节点的 waitStatus 是 SIGNAL 时, CAS算法保证设置成0时, 才能唤醒后续节点 */
                if (!compareAndSetWaitStatus(h, Node.SIGNAL, 0))
                    continue;            // loop to recheck cases
                unparkSuccessor(h);
            }
            else if (ws == 0 &&      /* 头节点的 waitStatus 是 0 时, CAS算法设置成 PROPAGATE, 表示共享状态向后传播 */
                     !compareAndSetWaitStatus(h, 0, Node.PROPAGATE))
                continue;                // loop on failed CAS
        }
        if (h == head)                   // loop if head changed
            break;
    }
}

protected int tryAcquireShared(int acquires) {
    for (;;) {
        if (hasQueuedPredecessors()) /* 如有有节点在排队, 则返回负数. 后续直接入队, 防止插队. */
            return -1;
        int available = getState();
        int remaining = available - acquires;
        if (remaining < 0 ||
            compareAndSetState(available, remaining))
            return remaining;
    }
}