why ConcurrentLinkedQueue#addAll need set tail twice?
Liu via Concurrency-interest <[email protected]> Thu, 23 Jul 2020 21:38:51 +0800 (GMT+08:00)
| Newsgroups | gmane.comp.java.jsr.166-concurrency |
|---|---|
| Message-ID | <[email protected]> |
In JDK8's ConcurrentLinkedQueue, there is an addAll function, and it will set tail twice, why?
public boolean addAll(Collection<? extends E> c) {
if (c == this)
// As historically specified in AbstractQueue#addAll
throw new IllegalArgumentException();
// Copy c into a private chain of Nodes
Node<E> beginningOfTheEnd = null, last = null;
for (E e : c) {
checkNotNull(e);
Node<E> newNode = new Node<E>(e);
if (beginningOfTheEnd == null)
beginningOfTheEnd = last = newNode;
else {
last.lazySetNext(newNode);
last = newNode;
}
}
if (beginningOfTheEnd == null)
return false;
// Atomically append the chain at the tail of this collection
for (Node<E> t = tail, p = t;;) {
Node<E> q = p.next;
if (q == null) {
// p is last node
if (p.casNext(null, beginningOfTheEnd)) {
// Successful CAS is the linearization point
// for all elements to be added to this queue.
if (!casTail(t, last)) { // first set tail
// Try a little harder to update tail,
// since we may be adding many elements.
t = tail;
if (last.next == null)
casTail(t, last); // second set tail
}
return true;
}
// Lost CAS race to another thread; re-read next
}
else if (p == q)
// We have fallen off list. If tail is unchanged, it
// will also be off-list, in which case we need to
// jump to head, from which all live nodes are always
// reachable. Else the new tail is a better bet.
p = (t != (t = tail)) ? t : head;
else
// Check for tail updates after two hops.
p = (p != t && t != (t = tail)) ? t : q;
}
}
I don't quite understand, I just think the second set operation is not necessary.
--------------------------------------------------------------------------------
Regards
Liu
_______________________________________________
Concurrency-interest mailing list
[email protected]
http://cs.oswego.edu/mailman/listinfo/concurrency-interest