共计 5219 个字符,预计需要花费 14 分钟才能阅读完成。
之前 ReentrantReadWriteLock 讲了读写锁的场景,这边来讲他的源码,以非公平锁为例,其实和公平锁主要代码是一致的。
Sync 类
static final int SHARED_SHIFT = 16;// 高 16 位是共享,用于读,低 16 位是独占,用于写,用一个字段保证原子性
static final int SHARED_UNIT = (1 << SHARED_SHIFT);// 左移 16 位,也就是高位的最后一个是 0000 0000 0000 0001 0000 0000 0000 0000
static final int MAX_COUNT = (1 << SHARED_SHIFT) - 1;// 最大读的数量,正常不会这么多
static final int EXCLUSIVE_MASK = (1 << SHARED_SHIFT) - 1;// 左移 16 位,再减 1,也就是 0000 0000 0000 0000 1111 1111 1111 1111
static int sharedCount(int c) {return c >>> SHARED_SHIFT;}// 无符号右移 16 位
static int exclusiveCount(int c) {return c & EXCLUSIVE_MASK;}// 返回的不为 1,说明有写锁,因为低 16 位都是 1,1 与 1 为 1,如果有些,肯定有个为 1
static final class HoldCounter {// 每个线程持有的锁的数量
int count = 0;
// Use id, not reference, to avoid garbage retention
final long tid = getThreadId(Thread.currentThread());
}
static final class ThreadLocalHoldCounter
extends ThreadLocal<HoldCounter> {// 本地线程
public HoldCounter initialValue() {return new HoldCounter();
}
}
private transient ThreadLocalHoldCounter readHolds;// 本地线程,记录持有的锁的数量信息
private transient HoldCounter cachedHoldCounter;// 缓存 HoldCounter 的数据
private transient Thread firstReader = null;// 第一个获取读锁的线程
private transient int firstReaderHoldCount;// 第一个获取读锁的线程持有的数量
读锁的 lock 方法
public void lock() {sync.acquireShared(1);
}
public final void acquireShared(int arg) {if (tryAcquireShared(arg) < 0)// 小于 0 没获取到锁
doAcquireShared(arg);
}
protected final int tryAcquireShared(int unused) {Thread current = Thread.currentThread();// 获取本地线程
int c = getState();// 获取 state 的值
if (exclusiveCount(c) != 0 &&// 不为 0 说明有写锁,原因上面分析了
getExclusiveOwnerThread() != current)// 不是当前线程,说明不是重入
return -1;
int r = sharedCount(c);// 获取读锁的个数
if (!readerShouldBlock() &&// 读锁无堵塞
r < MAX_COUNT &&// 读锁没到最大值
compareAndSetState(c, c + SHARED_UNIT)) {//cas 操作,高位加 1 成功说明获取到了读锁
if (r == 0) {// 等于 0 说明第一个获取读锁
firstReader = current;// 当前线程就是第一个
firstReaderHoldCount = 1;// 数量为 1
} else if (firstReader == current) {// 如果不是第一个,但是是当前线程
firstReaderHoldCount++;// 数量加 1
} else {// 既不是第一个,也不是当前线程
HoldCounter rh = cachedHoldCounter;// 获取缓存 HoldCounter
if (rh == null || rh.tid != getThreadId(current))// 如果不为空,或者通过线程 id 对比不是当前线程
cachedHoldCounter = rh = readHolds.get();// 缓存设置为当前线程
else if (rh.count == 0)// 缓存的是当前线程,而且锁的数量为 0,加入到本地缓存,如果数量不为 0,说明已经在本地缓存了
readHolds.set(rh);
rh.count++;// 锁的数量加 1
}
return 1;
}
return fullTryAcquireShared(current);
}
// 如果阻塞或者 cas 失败的情况,再重试获取锁
final int fullTryAcquireShared(Thread current) {
HoldCounter rh = null;
for (;;) {//
int c = getState();
if (exclusiveCount(c) != 0) {// 上面分析了,如果是写锁,并且不是当前线程,放弃
if (getExclusiveOwnerThread() != current)
return -1;
// else we hold the exclusive lock; blocking here
// would cause deadlock.
} else if (readerShouldBlock()) {// 阻塞的情况
// Make sure we're not acquiring read lock reentrantly
if (firstReader == current) {// 当前线程是第一个不处理
// 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();// 如果是 0,移除本地缓存}
}
if (rh.count == 0)
return -1;//
}
}
if (sharedCount(c) == MAX_COUNT)// 读锁数量太大,抛异常
throw new Error("Maximum lock count exceeded");
if (compareAndSetState(c, c + SHARED_UNIT)) {if (sharedCount(c) == 0) {// 没有读锁
firstReader = current;// 当前线程就是第一个
firstReaderHoldCount = 1;// 数量为 1
} else if (firstReader == current) {// 如果不是第一个,但是是当前线程
firstReaderHoldCount++;// 数量加 1
} else {// 既不是第一个,也不是当前线程
if (rh == null)
rh = cachedHoldCounter;// 获取缓存 HoldCounter
if (rh == null || rh.tid != getThreadId(current))// 如果不为空,或者通过线程 id 对比不是当前线程
rh = readHolds.get();// 缓存设置为当前线程
else if (rh.count == 0)// 缓存的是当前线程,而且锁的数量为 0,加入到本地缓存,如果数量不为 0,说明已经在本地缓存了
readHolds.set(rh);
rh.count++;// 锁的数量加 1
cachedHoldCounter = rh; // cache for release
}
return 1;
}
}
}
读锁的 unlock 方法
public void unlock() {sync.releaseShared(1);
}
public final boolean releaseShared(int arg) {if (tryReleaseShared(arg)) {doReleaseShared();
return true;
}
return false;
}
protected final boolean tryReleaseShared(int unused) {Thread current = Thread.currentThread();
if (firstReader == current) {// 如果第一个是当前线程
// assert firstReaderHoldCount > 0;
if (firstReaderHoldCount == 1)// 如果数量为 1,就是直接设为空
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) {// 数量小于等于 1,移除
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;// 写锁和读锁为 0,无锁
}
}
写锁的 lock 方法
public void lock() {sync.acquire(1);
}
public final void acquire(int arg) {if (!tryAcquire(arg) &&
acquireQueued(addWaiter(Node.EXCLUSIVE), arg))// 失败了就进入阻塞队列
selfInterrupt();}
protected final boolean tryAcquire(int acquires) {Thread current = Thread.currentThread();// 获取当前线程
int c = getState();// 获取 state
int w = exclusiveCount(c);// 不为 0 说明有写锁
if (c != 0) {// 有读或者写锁
// 无写锁或者读锁被占
if (w == 0 || current != getExclusiveOwnerThread())
return false;
if (w + exclusiveCount(acquires) > MAX_COUNT)
throw new Error("Maximum lock count exceeded");
// Reentrant acquire
setState(c + acquires);
return true;
}
if (writerShouldBlock() ||// 无阻塞
!compareAndSetState(c, c + acquires))// 设置成功
return false;
setExclusiveOwnerThread(current);
return true;
}
写锁的 unlock 方法
public void unlock() {sync.release(1);
}
public final boolean release(int arg) {if (tryRelease(arg)) {
Node h = head;
if (h != null && h.waitStatus != 0)
unparkSuccessor(h);// 唤醒
return true;
}
return false;
}
protected final boolean tryRelease(int releases) {if (!isHeldExclusively())// 不是独占锁,抛异常
throw new IllegalMonitorStateException();
int nextc = getState() - releases;
boolean free = exclusiveCount(nextc) == 0;// 写锁都释放了
if (free)
setExclusiveOwnerThread(null);
setState(nextc);
return free;
}
正文完