On AbstractQueuedSynchronizer
State
- Internally, maintains a
volatile int state
- state = 0 is free, 1 acquired, > 1 num of reentrants
- State is updated by
unsafe.compareAndSwapInt
Node and CLH lock
In CLH lock, applying thread can only self-spin on local variable, and keep polling the previous node’s status, if it discovers that the prev node released lock, it will finish self spin
static final class Node{
volatile int waitStatus; //control blocking
volatile Node prev;
volatile Node next;
volatile Thread thread;
Node nextWaiter; //Uses sentinel value to detect if it is shared or exclusive
}
- if
prev
is cancelled, this node can use prev node’s status - head inside AQS is a sentinel value
- Custom synchroinzer need to implement how
state
is acquired and released, q maintenance is done by the AQS already - WaitStatus
- 1 - cancelled
- -1 - signal, i.e., current thread need to wake up the
next
when release or cancel - -2 - condition, waiting to be waken up by the condition
- -3 - propergate, shared lock
Acquire
- You need to implement your own
tryAcquire
ortryAcquireShared
private Node enq(final Node node) {
for (;;) {
Node t = tail;
if (t == null) { // Must initialize
if (compareAndSetHead(new Node()))
tail = head;
} else {
node.prev = t;
if (compareAndSetTail(t, node)) {
t.next = node;
return t;
}
}
}
}
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;
if (compareAndSetTail(pred, node)) { //uses unsafe
pred.next = node;
return node;
}
}
enq(node);
return node;
}
private static boolean shouldParkAfterFailedAcquire(Node pred, Node node) {
int ws = pred.waitStatus;
if (ws == Node.SIGNAL)
/*
* This node has already set status asking a release
* to signal it, so it can safely park.
*/
return true;
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 {
/*
* 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;
}
private final boolean parkAndCheckInterrupt() {
LockSupport.park(this); //to wake up thread, use unpark() or interrupted()
return Thread.interrupted(); //ignores inturrptions during the park(), will add a self-interruption only after resource is acquired
}
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)) {//it is possible to get lock now, e.g., head has been released
setHead(node);
p.next = null; // help GC
failed = false;
return interrupted;
}
if (shouldParkAfterFailedAcquire(p, node) &&
parkAndCheckInterrupt())
interrupted = true;
}
} finally {
if (failed)
cancelAcquire(node);
}
}
public void acquire(int arg) {
if(!tryAcquire(arg) && acquireQueued(addWaiter(Node.EXCLUSIVE), arg))
selfInterrupt(); //if the thread has been interrupted during the wait, it won't respond, here we do a self-interrupt to make up for it
}
public void acquireShared(int arg){
if(tryAcquireShared(arg) < 0)
doAcquireShared(arg);
}
- Difference with fair and biased: on biased, CAS to acq first, and acq only when failed, and won’t detect if CLH is idle
- Shared lock: when it becomes head, it will wake up the next thread, upon releasing, S lock will wake up other threads regardless of state, but X lock will wake up only when state = 0
- for X lock, each node CAS checks if the previosu node is header, if so, try to acquire lock
- CountDownLatch: state set to N, at each countDown(), state will CAS to dec by 1, when state= 0, unpark the main thread