标签:一个 模板模式 允许 lse 排它锁 nfa nec 线程并发 策略
目录
AQS为依赖FIFO的等待队列的阻塞锁和相关的同步器(信号量,事件等)实现提供了一个框架。设计的目标是为大部分的依赖一个原子int值表示状态的同步器提供一个有用的基础。子类必须实现一个受保护的来改变状态值,并且定义哪种状态值表示资源的获取或者释放。通过这些,类里面的其他的方法就可以实现排队和阻塞的机制。子类也可以维护其他的状态值,但是为了获取同步一般只追踪使用getState、setState、compareAndSetState修改的int值。
模板模式:父类定义好了算法的框架,第一步做什么第二步做什么,同时把某些步骤的实现延迟到子类去实现。
tryAcquireNanos(int arg, long nanosTimeout)
tryAcquireSharedNanos(int arg, long nanosTimeout)
release
releaseShared(int arg)
//同步状态被设置成了volatile,确保了state在不同线程的可见性。
private volatile int state;
略
//1。lock方法,先通过原子操作设置state的值为1。
final void lock() {
//不可重入。如果当前的state是0,则替换为1。同时把当前的线程设置为占有资源的线程
if (compareAndSetState(0, 1))
//1.1如果设置成功,当前线程设置为占有资源的线程。
setExclusiveOwnerThread(Thread.currentThread());
else
//1.2否则调用acquire获取
acquire(1);
}
//2.acquire(1) 获取资源,即使等待,直到获取成功为止。
public final void acquire(int arg) {
//如果tryAcquire返回true直接短路.该方法有子类实现。
//否则在addWaiter把当前线程打包成一个Node节点放到同步队列队尾
//通过
if (!tryAcquire(arg) &&
acquireQueued(addWaiter(Node.EXCLUSIVE), arg))
selfInterrupt();
}
//2.1 tryAcquire(arg)
//子类需要自己实现tryAcquire逻辑。主要通过(getState/setState/compareAndSetState),是否可重入,是否阻塞
//非公平锁资源获取
final boolean nonfairTryAcquire(int acquires) {
final Thread current = Thread.currentThread();
int c = getState();
if (c == 0) {
if (compareAndSetState(0, acquires)) {
setExclusiveOwnerThread(current);
return true;
}
}
else if (current == getExclusiveOwnerThread()) {
int nextc = c + acquires;
if (nextc < 0) // overflow
throw new Error("Maximum lock count exceeded");
setState(nextc);
return true;
}
return false;
}
//2.2 addWaiter(Node.EXCLUSIVE)
private Node addWaiter(Node mode) {
Node node = new Node(Thread.currentThread(), mode);
// Try the fast path of enq; backup to full enq on failure
Node pred = tail;
if (pred != null) {
node.prev = pred;
//先通过原子操作compareAndSetTail,尝试把当前node加到同步队列的尾部。
if (compareAndSetTail(pred, node)) {
pred.next = node;
return node;
}
}
//上一步失败则通过enq入队。
enq(node);
return node;
}
//2.2。1 enq方法
private Node enq(final Node node) {
//使用自旋的方式
for (;;) {
Node t = tail;
if (t == null) { // Must initialize
// 队列为空,创建一个空的标志结点作为head结点,并将tail也指向它。
if (compareAndSetHead(new Node()))
tail = head;
} else {
//队列不为空。则把当前node加入队尾
node.prev = t;
if (compareAndSetTail(t, node)) {
t.next = node;
return t;
}
}
}
}
//2.3 acquireQueued(Node, int) 线程已经加入队尾,处于等待状态。直到其他线程释放线程唤醒自己。自己获取资源后,再处理自己的逻辑。
final boolean acquireQueued(final Node node, int arg) {
//是否获得资源,true表示未获得
boolean failed = true;
try {
//是否被中断过
boolean interrupted = false;
for (;;) {//自旋的方式
//获取前驱节点
final Node p = node.predecessor();
//如果前驱节点是head节点,那么当前节点是第二个节点。
//并且尝试获取资源成功,则设置当前节点为head
if (p == head && tryAcquire(arg)) {
setHead(node);
p.next = null; // help GC
failed = false;
return interrupted;
}
//如果获取失败。将前驱节点的waitStatus改为Node.SIGNAL,就是告诉前驱节点,如果你释放了资源通知我一下。
if (shouldParkAfterFailedAcquire(p, node) &&
parkAndCheckInterrupt())
interrupted = true;
}
} finally {
if (failed)
cancelAcquire(node);
}
}
//2.3.1 shouldParkAfterFailedAcquire(Node,Node) 获取资源失败后,判断前驱节点的waitStatus的值,1)如果前驱节点是取消状态舍弃;2)如果是正常状态设置前驱节点的状态是Node.SIGNAL并且返回true;3)如果前驱节点状态时Node.SIGNAL,直接返回true。
private static boolean shouldParkAfterFailedAcquire(Node pred, Node node) {
int ws = pred.waitStatus;
//如果waitStatus = Node.SIGNAL,返回true
if (ws == Node.SIGNAL)
/*
* This node has already set status asking a release
* to signal it, so it can safely park.
*/
return true;
//如果waitStatus = Node.CANCELLED,则舍弃这个前驱节点。继续判断下一个前驱节点的waitStatus
if (ws > 0) {
/*
* Predecessor was cancelled. Skip over predecessors and
* indicate retry.
*/
do {
node.prev = pred = pred.prev;
} while (pred.waitStatus > 0);
pred.next = node;
} else {
//如果前驱正常,那就把前驱的状态设置成SIGNAL,告诉它拿完号后通知自己一下。有可能失败,人家说不定刚刚释放完呢!
/*
* 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);
}
return false;
}
//2.3.2 parkAndCheckInterrupt() 如果当前节点的前驱节点是Node.SIGNAL,则把当前线程挂起,处于等待状态。
private final boolean parkAndCheckInterrupt() {
//park的县城会处于waiting状态Thread。State。WAITING,通过unpark方法或者interrupt()方法唤醒
LockSupport.park(this);
return Thread.interrupted();
}
//1.独占模式下释放资源 一般都回成功,返回true
public final boolean release(int arg) {
//当state=0时,返回true
if (tryRelease(arg)) {
Node h = head;
if (h != null && h.waitStatus != 0)
unparkSuccessor(h);
return true;
}
return false;
}
//2. tryRelease(int) 子类自己实现的方法,用于释放资源。
//一般会释放成功。因为只有获取了资源的线程才会释放资源,直接减去相应量的资源。不用考虑线程安全问题。不需要原则操作。
protected final boolean tryRelease(int releases) {
int c = getState() - releases;
if (Thread.currentThread() != getExclusiveOwnerThread())
throw new IllegalMonitorStateException();
boolean free = false;
if (c == 0) {
free = true;
setExclusiveOwnerThread(null);
}
setState(c);
return free;
}
//3. unparkSuccessor(Node) 唤醒下一个节点。如果下一个节点null或者cancelled,从队尾往前取节点依次唤醒。
private void unparkSuccessor(Node node) {
/*
* If status is negative (i.e., possibly needing signal) try
* to clear in anticipation of signalling. It is OK if this
* fails or if status is changed by waiting thread.
*/
int ws = node.waitStatus;
if (ws < 0)
compareAndSetWaitStatus(node, ws, 0);
/*
* Thread to unpark is held in successor, which is normally
* just the next node. But if cancelled or apparently null,
* traverse backwards from tail to find the actual
* non-cancelled successor.
*/
Node s = node.next;
if (s == null || s.waitStatus > 0) {
s = null;
for (Node t = tail; t != null && t != node; t = t.prev)
if (t.waitStatus <= 0)
s = t;
}
if (s != null)
LockSupport.unpark(s.thread);
}
//tryAcquireShared方法尝试获取资源,成功则直接返回。
//aqs里面已经定义好,获取失败小于0。
public final void acquireShared(int arg) {
if (tryAcquireShared(arg) < 0)
doAcquireShared(arg);
}
//2.doAcquireShared 如果调用tryAcquireShared获取失败,这里则调用doAcquireShared把节点放入同步队列尾部,并且通过park()让其休息,直到有线程唤醒它。
privateprivate void setHeadAndPropagate(Node node, int propagate) {
Node h = head; // Record old head for check below
setHead(node);
/*
* 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();
}
} void doAcquireShared(int arg) {
final Node node = addWaiter(Node.SHARED);
boolean failed = true;
try {
boolean interrupted = false;
for (;;) {
final Node p = node.predecessor();
if (p == head) {
int r = tryAcquireShared(arg);
if (r >= 0) {
setHeadAndPropagate(node, r);
p.next = null; // help GC
if (interrupted)
selfInterrupt();
failed = false;
return;
}
}
if (shouldParkAfterFailedAcquire(p, node) &&
parkAndCheckInterrupt())
interrupted = true;
}
} finally {
if (failed)
cancelAcquire(node);
}
}
// 3。 setHeadAndPropagate(node, r); 将head指向自己,如果还有剩余量,则继续花唤醒下一个邻居线程。
//1.releaseShared 先调用tryReleaseShared方法尝试释放
public final boolean releaseShared(int arg) {
if (tryReleaseShared(arg)) {
doReleaseShared();
return true;
}
return false;
}
//2。唤醒后继节点
private void doReleaseShared() {
for (;;) {
Node h = head;
if (h != null && h != tail) {
int ws = h.waitStatus;
if (ws == Node.SIGNAL) {
if (!compareAndSetWaitStatus(h, Node.SIGNAL, 0))
continue;
unparkSuccessor(h);//唤醒后继
}
else if (ws == 0 &&
!compareAndSetWaitStatus(h, 0, Node.PROPAGATE))
continue;
}
if (h == head)// head发生变化
break;
}
}
独占模式下的tryRelease()在完全释放掉资源(state=0)后,才会返回true去唤醒其他线程,这主要是基于独占下可重入的考量;而共享模式下的releaseShared()则没有这种要求,共享模式实质就是控制一定量的线程并发执行,那么拥有资源的线程在释放掉部分资源时就可以唤醒后继等待结点。例如,资源总量是13,A(5)和B(7)分别获取到资源并发运行,C(4)来时只剩1个资源就需要等待。A在运行过程中释放掉2个资源量,然后tryReleaseShared(2)返回true唤醒C,C一看只有3个仍不够继续等待;随后B又释放2个,tryReleaseShared(2)返回true唤醒C,C一看有5个够自己用了,然后C就可以跟A和B一起运行。而ReentrantReadWriteLock读锁的tryReleaseShared()只有在完全释放掉资源(state=0)才返回true,所以自定义同步器可以根据需要决定tryReleaseShared()的返回值。
//以公平所为例子
protected final boolean tryAcquire(int acquires) {
final Thread current = Thread.currentThread();
int c = getState();
if (c == 0) {
if (!hasQueuedPredecessors() &&
compareAndSetState(0, acquires)) {
setExclusiveOwnerThread(current);
return true;
}
}
//关键是这里如果当前线程就是目前占有所的线程。则把state增加对应的acquires。因为当前线程持有锁,所以这里使用setState(nextc)来直接设置state的值
else if (current == getExclusiveOwnerThread()) {
int nextc = c + acquires;
if (nextc < 0)
throw new Error("Maximum lock count exceeded");
setState(nextc);
return true;
}
return false;
}
//可以猜测在释放资源的时候会扣减对应的releases。最终state=0,唤醒等待队列里面的下一个节点(如果存在的话)。
protected final boolean tryRelease(int releases) {
int c = getState() - releases;
if (Thread.currentThread() != getExclusiveOwnerThread())
throw new IllegalMonitorStateException();
boolean free = false;
if (c == 0) {
free = true;
setExclusiveOwnerThread(null);
}
setState(c);
return free;
}
//1。公平锁的lock vs 非公平锁的lock
//1.1 fairSync
final void lock() {
acquire(1);
}
//1.2 NonfairSync
final void lock() {
//一上来完全不管同步队列里面的节点,直接判断一把,如果同步器的state=0,立刻用原子操作修改stata值。返回true表示抢占资源成功。
if (compareAndSetState(0, 1))
setExclusiveOwnerThread(Thread.currentThread());
else
//上一步失败了,才尝试获取
acquire(1);
}
//2. 公平锁的lock vs 非公平锁的acquire(1)
//2.1 NonfairSync
final boolean nonfairTryAcquire(int acquires) {
final Thread current = Thread.currentThread();
int c = getState();
if (c == 0) {
//和fairSync不同,不需要判断当前线程是否同步队列的第一个节点。直接进入if里面的逻辑。
if (compareAndSetState(0, acquires)) {
setExclusiveOwnerThread(current);
return true;
}
}
。。。。
}
//2.2 fairSync
protected final boolean tryAcquire(int acquires) {
final Thread current = Thread.currentThread();
int c = getState();
if (c == 0) {
//hasQueuedPredecessors()方法会判断当前节点是否有强制节点,不存在前置节点才走if里面的逻辑。
if (!hasQueuedPredecessors() &&
compareAndSetState(0, acquires)) {
setExclusiveOwnerThread(current);
return true;
}
}
。。。。
}
排它锁一个时刻只允许一个线程访问。而读写锁一个时刻允许多个读线程访问,当写锁被其他线程占有未释放时,所有的读线程和其他读线程均被阻塞。在ReentrantReadWriteLock内部分别维护了一个读锁和一个写锁,通过分离读锁和写锁,使得并发性相比一般的排他锁有了很大的提升。同步器的同步状态通过把状态int值分成高16位和低16位分别表示读锁和写锁的状态。
//1.ReentrantReadWriteLock内部分别定义了一个writerLock和readerLock
private final ReentrantReadWriteLock.ReadLock readerLock;
/** Inner class providing writelock */
private final ReentrantReadWriteLock.WriteLock writerLock;
//2。1 读锁lock 共享模式获取资源
public final void acquireShared(int arg) {
//先调用子类自己实现的tryAcquireShared尝试获取资源
if (tryAcquireShared(arg) < 0)
doAcquireShared(arg);
}
//2.1.1 tryAcquireShared 尝试获取
protected final int tryAcquireShared(int unused) {
Thread current = Thread.currentThread();
//c是锁状态为:高位16位表示共享锁的数量,低位16位表示独占锁的数量
int c = getState();
//exclusiveCount(c) 取低16位的值,也就是写锁状态位:不等于0表示写锁被占用
if (exclusiveCount(c) != 0 &&
getExclusiveOwnerThread() != current)
//写锁被其他线程占用,获取读锁失败
return -1;
//取高16位的值,读锁状态位
int r = sharedCount(c);
//readerShouldBlock()根据读锁获取策略,返回是否阻塞当前读锁获取操作。后面会详细说明此方法
if (!readerShouldBlock() &&
r < MAX_COUNT &&
//cas修改高16位的读锁状态,即获取读锁
compareAndSetState(c, c + SHARED_UNIT)) {
//首次获取读锁
if (r == 0) {
//缓存首次获取读锁的线程,及其读锁重入次数
firstReader = current;
firstReaderHoldCount = 1;
} else if (firstReader == current) {//线程重入
firstReaderHoldCount++;
} else {
//cachedHoldCounter是最后获取锁的线程的读锁重入次数
HoldCounter rh = cachedHoldCounter;
/readHolds是缓存了当前线程的读锁重入次数信息的ThreadLocal
if (rh == null || rh.tid != getThreadId(current))
cachedHoldCounter = rh = readHolds.get();
else if (rh.count == 0)
//缓存当前线程的holdCounter
//fullTryAcquireShared()方法中,
//获取读锁失败的线程会执行:readHolds.remove(),故此时需要重新设置
readHolds.set(rh);
rh.count++;
}
return 1;
}
//首次获取读锁失败后,重试获取
return fullTryAcquireShared(current);
}
//2.1.2。1 fullTryAcquireShared
final int fullTryAcquireShared(Thread current) {
//rh表示当前线程的锁计数器
HoldCounter rh = null;
for (;;) {
int c = getState();
//写锁被占用
if (exclusiveCount(c) != 0) {
//如果其他线程占用,读锁获取失败。如果当前读线程占用,表示“锁降级”。
//如果这里不支持锁降级,这里线程持有写锁(exclusiveCount(c) != 0),这里会出现死锁。
if (getExclusiveOwnerThread() != current)
return -1;
} else if (readerShouldBlock()) {
// Make sure we're not acquiring read lock reentrantly
if (firstReader == current) {
// assert firstReaderHoldCount > 0;
} else {
if (rh == null) {
rh = cachedHoldCounter;
if (rh == null || rh.tid != getThreadId(current)) {
rh = readHolds.get();
if (rh.count == 0)
readHolds.remove();
}
}
if (rh.count == 0)
return -1;
}
}
if (sharedCount(c) == MAX_COUNT)
throw new Error("Maximum lock count exceeded");
if (compareAndSetState(c, c + SHARED_UNIT)) {
if (sharedCount(c) == 0) {
firstReader = current;
firstReaderHoldCount = 1;
} else if (firstReader == current) {
firstReaderHoldCount++;
} else {
if (rh == null)
rh = cachedHoldCounter;
if (rh == null || rh.tid != getThreadId(current))
rh = readHolds.get();
else if (rh.count == 0)
readHolds.set(rh);
rh.count++;
cachedHoldCounter = rh; // cache for release
}
return 1;
}
}
}
//2.1.2 doAcquireShared
//AQS里面doAcquireShared方法,如果当前节点是head.next时,才会尝试获取资源。
private void doAcquireShared(int arg) {
final Node node = addWaiter(Node.SHARED);
boolean failed = true;
try {
boolean interrupted = false;
for (;;) {
final Node p = node.predecessor();
if (p == head) {
//tryAcquireShared(arg);如果获取成功则返回还剩余多少个资源。
int r = tryAcquireShared(arg);
if (r >= 0) {
//如果还有剩余量,继续唤醒下一个邻居线程
setHeadAndPropagate(node, r);
p.next = null; // help GC
if (interrupted)
selfInterrupt();
failed = false;
return;
}
}
if (shouldParkAfterFailedAcquire(p, node) &&
parkAndCheckInterrupt())
interrupted = true;
}
} finally {
if (failed)
cancelAcquire(node);
}
}
//FairSyn
//如果当前线程不是同步队列头结点的next节点(head.next)则阻塞当前线程
final boolean readerShouldBlock() {
return hasQueuedPredecessors();
}
//NonFairSyn
//为了防止写线程饥饿,如果同步队列中的第一个线程是以独占模式获取锁(写锁),那么当前获取读锁的线程需要阻塞,让队列中的第一个线程先执行
final boolean readerShouldBlock() {
return apparentlyFirstQueuedIsExclusive();
}
//2.2 写锁lock
跟独占模式比,还有一点需要注意的是,这里只有线程是head.next时(“老二”),才会去尝试获取资源,有剩余的话还会唤醒之后的队友。
那么问题就来了,假如老大用完后释放了5个资源,而老二需要6个,老三需要1个,老四需要2个。老大先唤醒老二,老二一看资源不够,
他是把资源让给老三呢,还是不让?答案是否定的!老二会继续park()等待其他线程释放资源,也更不会去唤醒老三和老四了。
独占模式,同一时刻只有一个线程去执行,这样做未尝不可;但共享模式下,多个线程是可以同时执行的,现在因为老二的资源需求量大,而把后面量小的老三和老四也都卡住了。
当然,这并不是问题,只是AQS保证严格按照入队顺序唤醒罢了(保证公平,但降低了并发)。
标签:一个 模板模式 允许 lse 排它锁 nfa nec 线程并发 策略
原文地址:https://www.cnblogs.com/codetree/p/10884170.html