ViewVC Help

View File | Revision Log | Show Annotations | Download File | Root Listing

root/jsr166/jsr166/src/jsr166y/LinkedTransferQueue.java

Comparing jsr166/src/jsr166y/LinkedTransferQueue.java (file contents):
Revision 1.62 by jsr166, Mon Nov 2 03:01:10 2009 UTC vs.
Revision 1.79 by dl, Fri Sep 10 10:43:23 2010 UTC

#	Line 15 | Line 15 | import java.util.Iterator;
15		import java.util.NoSuchElementException;
16		import java.util.Queue;
17		import java.util.concurrent.locks.LockSupport;
18	+
19		/**
20		* An unbounded {@link TransferQueue} based on linked nodes.
21		* This queue orders elements FIFO (first-in-first-out) with respect
#	Line 206 | Line 207 | public class LinkedTransferQueue<E> exte
207		* additional GC bookkeeping ("write barriers") that are sometimes
208		* more costly than the writes themselves because of contention).
209		*
209	–	* Removal of interior nodes (due to timed out or interrupted
210	–	* waits, or calls to remove(x) or Iterator.remove) can use a
211	–	* scheme roughly similar to that described in Scherer, Lea, and
212	–	* Scott's SynchronousQueue. Given a predecessor, we can unsplice
213	–	* any node except the (actual) tail of the queue. To avoid
214	–	* build-up of cancelled trailing nodes, upon a request to remove
215	–	* a trailing node, it is placed in field "cleanMe" to be
216	–	* unspliced upon the next call to unsplice any other node.
217	–	* Situations needing such mechanics are not common but do occur
218	–	* in practice; for example when an unbounded series of short
219	–	* timed calls to poll repeatedly time out but never otherwise
220	–	* fall off the list because of an untimed call to take at the
221	–	* front of the queue. Note that maintaining field cleanMe does
222	–	* not otherwise much impact garbage retention even if never
223	–	* cleared by some other call because the held node will
224	–	* eventually either directly or indirectly lead to a self-link
225	–	* once off the list.
226	–	*
210		* *** Overview of implementation ***
211		*
212		* We use a threshold-based approach to updates, with a slack
#	Line 239 | Line 222 | public class LinkedTransferQueue<E> exte
222		* per-thread one available, but even ThreadLocalRandom is too
223		* heavy for these purposes.
224		*
225	<	* With such a small slack threshold value, it is rarely
226	<	* worthwhile to augment this with path short-circuiting; i.e.,
227	<	* unsplicing nodes between head and the first unmatched node, or
228	<	* similarly for tail, rather than advancing head or tail
246	<	* proper. However, it is used (in awaitMatch) immediately before
247	<	* a waiting thread starts to block, as a final bit of helping at
248	<	* a point when contention with others is extremely unlikely
249	<	* (since if other threads that could release it are operating,
250	<	* then the current thread wouldn't be blocking).
225	>	* With such a small slack threshold value, it is not worthwhile
226	>	* to augment this with path short-circuiting (i.e., unsplicing
227	>	* interior nodes) except in the case of cancellation/removal (see
228	>	* below).
229		*
230		* We allow both the head and tail fields to be null before any
231		* nodes are enqueued; initializing upon first append.  This
#	Line 329 | Line 307 | public class LinkedTransferQueue<E> exte
307		*    versa) compared to their predecessors receive additional
308		*    chained spins, reflecting longer paths typically required to
309		*    unblock threads during phase changes.
310	+	*
311	+	*
312	+	* ** Unlinking removed interior nodes **
313	+	*
314	+	* In addition to minimizing garbage retention via self-linking
315	+	* described above, we also unlink removed interior nodes. These
316	+	* may arise due to timed out or interrupted waits, or calls to
317	+	* remove(x) or Iterator.remove.  Normally, given a node that was
318	+	* at one time known to be the predecessor of some node s that is
319	+	* to be removed, we can unsplice s by CASing the next field of
320	+	* its predecessor if it still points to s (otherwise s must
321	+	* already have been removed or is now offlist). But there are two
322	+	* situations in which we cannot guarantee to make node s
323	+	* unreachable in this way: (1) If s is the trailing node of list
324	+	* (i.e., with null next), then it is pinned as the target node
325	+	* for appends, so can only be removed later after other nodes are
326	+	* appended. (2) We cannot necessarily unlink s given a
327	+	* predecessor node that is matched (including the case of being
328	+	* cancelled): the predecessor may already be unspliced, in which
329	+	* case some previous reachable node may still point to s.
330	+	* (For further explanation see Herlihy & Shavit "The Art of
331	+	* Multiprocessor Programming" chapter 9).  Although, in both
332	+	* cases, we can rule out the need for further action if either s
333	+	* or its predecessor are (or can be made to be) at, or fall off
334	+	* from, the head of list.
335	+	*
336	+	* Without taking these into account, it would be possible for an
337	+	* unbounded number of supposedly removed nodes to remain
338	+	* reachable.  Situations leading to such buildup are uncommon but
339	+	* can occur in practice; for example when a series of short timed
340	+	* calls to poll repeatedly time out but never otherwise fall off
341	+	* the list because of an untimed call to take at the front of the
342	+	* queue.
343	+	*
344	+	* When these cases arise, rather than always retraversing the
345	+	* entire list to find an actual predecessor to unlink (which
346	+	* won't help for case (1) anyway), we record a conservative
347	+	* estimate of possible unsplice failures (in "sweepVotes").
348	+	* We trigger a full sweep when the estimate exceeds a threshold
349	+	* ("SWEEP_THRESHOLD") indicating the maximum number of estimated
350	+	* removal failures to tolerate before sweeping through, unlinking
351	+	* cancelled nodes that were not unlinked upon initial removal.
352	+	* We perform sweeps by the thread hitting threshold (rather than
353	+	* background threads or by spreading work to other threads)
354	+	* because in the main contexts in which removal occurs, the
355	+	* caller is already timed-out, cancelled, or performing a
356	+	* potentially O(n) operation (e.g. remove(x)), none of which are
357	+	* time-critical enough to warrant the overhead that alternatives
358	+	* would impose on other threads.
359	+	*
360	+	* Because the sweepVotes estimate is conservative, and because
361	+	* nodes become unlinked "naturally" as they fall off the head of
362	+	* the queue, and because we allow votes to accumulate even while
363	+	* sweeps are in progress, there are typically significantly fewer
364	+	* such nodes than estimated.  Choice of a threshold value
365	+	* balances the likelihood of wasted effort and contention, versus
366	+	* providing a worst-case bound on retention of interior nodes in
367	+	* quiescent queues. The value defined below was chosen
368	+	* empirically to balance these under various timeout scenarios.
369	+	*
370	+	* Note that we cannot self-link unlinked interior nodes during
371	+	* sweeps. However, the associated garbage chains terminate when
372	+	* some successor ultimately falls off the head of the list and is
373	+	* self-linked.
374		*/
375
376		/** True if on multiprocessor */
#	Line 355 | Line 397 | public class LinkedTransferQueue<E> exte
397		private static final int CHAINED_SPINS = FRONT_SPINS >>> 1;
398
399		/**
400	+	* The maximum number of estimated removal failures (sweepVotes)
401	+	* to tolerate before sweeping through the queue unlinking
402	+	* cancelled nodes that were not unlinked upon initial
403	+	* removal. See above for explanation. The value must be at least
404	+	* two to avoid useless sweeps when removing trailing nodes.
405	+	*/
406	+	static final int SWEEP_THRESHOLD = 32;
407	+
408	+	/**
409		* Queue nodes. Uses Object, not E, for items to allow forgetting
410		* them after use.  Relies heavily on Unsafe mechanics to minimize
411	<	* unnecessary ordering constraints: Writes that intrinsically
412	<	* precede or follow CASes use simple relaxed forms.  Other
362	<	* cleanups use releasing/lazy writes.
411	>	* unnecessary ordering constraints: Writes that are intrinsically
412	>	* ordered wrt other accesses or CASes use simple relaxed forms.
413		*/
414		static final class Node {
415		final boolean isData;   // false if this is a request node
#	Line 373 | Line 423 | public class LinkedTransferQueue<E> exte
423		}
424
425		final boolean casItem(Object cmp, Object val) {
426	<	assert cmp == null || cmp.getClass() != Node.class;
426	>	//            assert cmp == null || cmp.getClass() != Node.class;
427		return UNSAFE.compareAndSwapObject(this, itemOffset, cmp, val);
428		}
429
430		/**
431	<	* Creates a new node. Uses relaxed write because item can only
432	<	* be seen if followed by CAS.
431	>	* Constructs a new node.  Uses relaxed write because item can
432	>	* only be seen after publication via casNext.
433		*/
434		Node(Object item, boolean isData) {
435		UNSAFE.putObject(this, itemOffset, item); // relaxed write
#	Line 395 | Line 445 | public class LinkedTransferQueue<E> exte
445		}
446
447		/**
448	<	* Sets item to self (using a releasing/lazy write) and waiter
449	<	* to null, to avoid garbage retention after extracting or
450	<	* cancelling.
448	>	* Sets item to self and waiter to null, to avoid garbage
449	>	* retention after matching or cancelling. Uses relaxed writes
450	>	* because order is already constrained in the only calling
451	>	* contexts: item is forgotten only after volatile/atomic
452	>	* mechanics that extract items.  Similarly, clearing waiter
453	>	* follows either CAS or return from park (if ever parked;
454	>	* else we don't care).
455		*/
456		final void forgetContents() {
457	<	UNSAFE.putOrderedObject(this, itemOffset, this);
458	<	UNSAFE.putOrderedObject(this, waiterOffset, null);
457	>	UNSAFE.putObject(this, itemOffset, this);
458	>	UNSAFE.putObject(this, waiterOffset, null);
459		}
460
461		/**
#	Line 435 | Line 489 | public class LinkedTransferQueue<E> exte
489		* Tries to artificially match a data node -- used by remove.
490		*/
491		final boolean tryMatchData() {
492	<	assert isData;
492	>	//            assert isData;
493		Object x = item;
494		if (x != null && x != this && casItem(x, null)) {
495		LockSupport.unpark(waiter);
#	Line 459 | Line 513 | public class LinkedTransferQueue<E> exte
513		/** head of the queue; null until first enqueue */
514		transient volatile Node head;
515
462	–	/** predecessor of dangling unspliceable node */
463	–	private transient volatile Node cleanMe; // decl here reduces contention
464	–
516		/** tail of the queue; null until first append */
517		private transient volatile Node tail;
518
519	+	/** The number of apparent failures to unsplice removed nodes */
520	+	private transient volatile int sweepVotes;
521	+
522		// CAS methods for fields
523		private boolean casTail(Node cmp, Node val) {
524		return UNSAFE.compareAndSwapObject(this, tailOffset, cmp, val);
#	Line 474 | Line 528 | public class LinkedTransferQueue<E> exte
528		return UNSAFE.compareAndSwapObject(this, headOffset, cmp, val);
529		}
530
531	<	private boolean casCleanMe(Node cmp, Node val) {
532	<	return UNSAFE.compareAndSwapObject(this, cleanMeOffset, cmp, val);
531	>	private boolean casSweepVotes(int cmp, int val) {
532	>	return UNSAFE.compareAndSwapInt(this, sweepVotesOffset, cmp, val);
533		}
534
535		/*
536	<	* Possible values for "how" argument in xfer method. Beware that
483	<	* the order of assigned numerical values matters.
536	>	* Possible values for "how" argument in xfer method.
537		*/
538	<	private static final int NOW     = 0; // for untimed poll, tryTransfer
539	<	private static final int ASYNC   = 1; // for offer, put, add
540	<	private static final int SYNC    = 2; // for transfer, take
541	<	private static final int TIMEOUT = 3; // for timed poll, tryTransfer
538	>	private static final int NOW   = 0; // for untimed poll, tryTransfer
539	>	private static final int ASYNC = 1; // for offer, put, add
540	>	private static final int SYNC  = 2; // for transfer, take
541	>	private static final int TIMED = 3; // for timed poll, tryTransfer
542
543		@SuppressWarnings("unchecked")
544		static <E> E cast(Object item) {
545	<	assert item == null || item.getClass() != Node.class;
545	>	//        assert item == null || item.getClass() != Node.class;
546		return (E) item;
547		}
548
#	Line 498 | Line 551 | public class LinkedTransferQueue<E> exte
551		*
552		* @param e the item or null for take
553		* @param haveData true if this is a put, else a take
554	<	* @param how NOW, ASYNC, SYNC, or TIMEOUT
555	<	* @param nanos timeout in nanosecs, used only if mode is TIMEOUT
554	>	* @param how NOW, ASYNC, SYNC, or TIMED
555	>	* @param nanos timeout in nanosecs, used only if mode is TIMED
556		* @return an item if matched, else e
557		* @throws NullPointerException if haveData mode but e is null
558		*/
#	Line 518 | Line 571 | public class LinkedTransferQueue<E> exte
571		break;
572		if (p.casItem(item, e)) { // match
573		for (Node q = p; q != h;) {
574	<	Node n = q.next;  // update head by 2
575	<	if (n != null)    // unless singleton
523	<	q = n;
524	<	if (head == h && casHead(h, q)) {
574	>	Node n = q.next;  // update by 2 unless singleton
575	>	if (head == h && casHead(h, n == null? q : n)) {
576		h.forgetNext();
577		break;
578		}                 // advance and retry
#	Line 537 | Line 588 | public class LinkedTransferQueue<E> exte
588		p = (p != n) ? n : (h = head); // Use head if p offlist
589		}
590
591	<	if (how >= ASYNC) {               // No matches available
591	>	if (how != NOW) {                 // No matches available
592		if (s == null)
593		s = new Node(e, haveData);
594		Node pred = tryAppend(s, haveData);
595		if (pred == null)
596		continue retry;           // lost race vs opposite mode
597	<	if (how >= SYNC)
598	<	return awaitMatch(s, pred, e, how, nanos);
597	>	if (how != ASYNC)
598	>	return awaitMatch(s, pred, e, (how == TIMED), nanos);
599		}
600		return e; // not waiting
601		}
#	Line 593 | Line 644 | public class LinkedTransferQueue<E> exte
644		* predecessor, or null if unknown (the null case does not occur
645		* in any current calls but may in possible future extensions)
646		* @param e the comparison value for checking match
647	<	* @param how either SYNC or TIMEOUT
648	<	* @param nanos timeout value
647	>	* @param timed if true, wait only until timeout elapses
648	>	* @param nanos timeout in nanosecs, used only if timed is true
649		* @return matched item, or e if unmatched on interrupt or timeout
650		*/
651	<	private E awaitMatch(Node s, Node pred, E e, int how, long nanos) {
652	<	long lastTime = (how == TIMEOUT) ? System.nanoTime() : 0L;
651	>	private E awaitMatch(Node s, Node pred, E e, boolean timed, long nanos) {
652	>	long lastTime = timed ? System.nanoTime() : 0L;
653		Thread w = Thread.currentThread();
654		int spins = -1; // initialized after first item and cancel checks
655		ThreadLocalRandom randomYields = null; // bound if needed
#	Line 606 | Line 657 | public class LinkedTransferQueue<E> exte
657		for (;;) {
658		Object item = s.item;
659		if (item != e) {                  // matched
660	<	assert item != s;
660	>	//                assert item != s;
661		s.forgetContents();           // avoid garbage
662		return this.<E>cast(item);
663		}
664	<	if ((w.isInterrupted() || (how == TIMEOUT && nanos <= 0)) &&
665	<	s.casItem(e, s)) {       // cancel
664	>	if ((w.isInterrupted() || (timed && nanos <= 0)) &&
665	>	s.casItem(e, s)) {        // cancel
666		unsplice(pred, s);
667		return e;
668		}
#	Line 621 | Line 672 | public class LinkedTransferQueue<E> exte
672		randomYields = ThreadLocalRandom.current();
673		}
674		else if (spins > 0) {             // spin
675	<	if (--spins == 0)
676	<	shortenHeadPath();        // reduce slack before blocking
626	<	else if (randomYields.nextInt(CHAINED_SPINS) == 0)
675	>	--spins;
676	>	if (randomYields.nextInt(CHAINED_SPINS) == 0)
677		Thread.yield();           // occasionally yield
678		}
679		else if (s.waiter == null) {
680		s.waiter = w;                 // request unpark then recheck
681		}
682	<	else if (how == TIMEOUT) {
682	>	else if (timed) {
683		long now = System.nanoTime();
684		if ((nanos -= now - lastTime) > 0)
685		LockSupport.parkNanos(this, nanos);
#	Line 637 | Line 687 | public class LinkedTransferQueue<E> exte
687		}
688		else {
689		LockSupport.park(this);
640	–	s.waiter = null;
641	–	spins = -1;                   // spin if front upon wakeup
690		}
691		}
692		}
#	Line 659 | Line 707 | public class LinkedTransferQueue<E> exte
707		return 0;
708		}
709
662	–	/**
663	–	* Tries (once) to unsplice nodes between head and first unmatched
664	–	* or trailing node; failing on contention.
665	–	*/
666	–	private void shortenHeadPath() {
667	–	Node h, hn, p, q;
668	–	if ((p = h = head) != null && h.isMatched() &&
669	–	(q = hn = h.next) != null) {
670	–	Node n;
671	–	while ((n = q.next) != q) {
672	–	if (n == null || !q.isMatched()) {
673	–	if (hn != q && h.next == hn)
674	–	h.casNext(hn, q);
675	–	break;
676	–	}
677	–	p = q;
678	–	q = n;
679	–	}
680	–	}
681	–	}
682	–
710		/* -------------- Traversal methods -------------- */
711
712		/**
#	Line 792 | Line 819 | public class LinkedTransferQueue<E> exte
819		public final void remove() {
820		Node p = lastRet;
821		if (p == null) throw new IllegalStateException();
822	<	findAndRemoveDataNode(lastPred, p);
822	>	if (p.tryMatchData())
823	>	unsplice(lastPred, p);
824		}
825		}
826
#	Line 802 | Line 830 | public class LinkedTransferQueue<E> exte
830		* Unsplices (now or later) the given deleted/cancelled node with
831		* the given predecessor.
832		*
833	<	* @param pred predecessor of node to be unspliced
833	>	* @param pred a node that was at one time known to be the
834	>	* predecessor of s, or null or s itself if s is/was at head
835		* @param s the node to be unspliced
836		*/
837	<	private void unsplice(Node pred, Node s) {
838	<	s.forgetContents(); // clear unneeded fields
837	>	final void unsplice(Node pred, Node s) {
838	>	s.forgetContents(); // forget unneeded fields
839		/*
840	<	* At any given time, exactly one node on list cannot be
841	<	* unlinked -- the last inserted node. To accommodate this, if
842	<	* we cannot unlink s, we save its predecessor as "cleanMe",
843	<	* processing the previously saved version first. Because only
844	<	* one node in the list can have a null next, at least one of
816	<	* node s or the node previously saved can always be
817	<	* processed, so this always terminates.
840	>	* See above for rationale. Briefly: if pred still points to
841	>	* s, try to unlink s.  If s cannot be unlinked, because it is
842	>	* trailing node or pred might be unlinked, and neither pred
843	>	* nor s are head or offlist, add to sweepVotes, and if enough
844	>	* votes have accumulated, sweep.
845		*/
846	<	if (pred != null && pred != s) {
847	<	while (pred.next == s) {
848	<	Node oldpred = (cleanMe == null) ? null : reclean();
849	<	Node n = s.next;
850	<	if (n != null) {
851	<	if (n != s)
852	<	pred.casNext(s, n);
853	<	break;
846	>	if (pred != null && pred != s && pred.next == s) {
847	>	Node n = s.next;
848	>	if (n == null ||
849	>	(n != s && pred.casNext(s, n) && pred.isMatched())) {
850	>	for (;;) {               // check if at, or could be, head
851	>	Node h = head;
852	>	if (h == pred || h == s || h == null)
853	>	return;          // at head or list empty
854	>	if (!h.isMatched())
855	>	break;
856	>	Node hn = h.next;
857	>	if (hn == null)
858	>	return;          // now empty
859	>	if (hn != h && casHead(h, hn))
860	>	h.forgetNext();  // advance head
861		}
862	<	if (oldpred == pred ||      // Already saved
863	<	((oldpred == null || oldpred.next == s) &&
864	<	casCleanMe(oldpred, pred))) {
865	<	break;
862	>	if (pred.next != pred && s.next != s) { // recheck if offlist
863	>	for (;;) {           // sweep now if enough votes
864	>	int v = sweepVotes;
865	>	if (v < SWEEP_THRESHOLD) {
866	>	if (casSweepVotes(v, v + 1))
867	>	break;
868	>	}
869	>	else if (casSweepVotes(v, 0)) {
870	>	sweep();
871	>	break;
872	>	}
873	>	}
874		}
875		}
876		}
877		}
878
879		/**
880	<	* Tries to unsplice the deleted/cancelled node held in cleanMe
881	<	* that was previously uncleanable because it was at tail.
840	<	*
841	<	* @return current cleanMe node (or null)
880	>	* Unlinks matched (typically cancelled) nodes encountered in a
881	>	* traversal from head.
882		*/
883	<	private Node reclean() {
884	<	/*
885	<	* cleanMe is, or at one time was, predecessor of a cancelled
886	<	* node s that was the tail so could not be unspliced.  If it
887	<	* is no longer the tail, try to unsplice if necessary and
888	<	* make cleanMe slot available.  This differs from similar
849	<	* code in unsplice() because we must check that pred still
850	<	* points to a matched node that can be unspliced -- if not,
851	<	* we can (must) clear cleanMe without unsplicing.  This can
852	<	* loop only due to contention.
853	<	*/
854	<	Node pred;
855	<	while ((pred = cleanMe) != null) {
856	<	Node s = pred.next;
857	<	Node n;
858	<	if (s == null || s == pred || !s.isMatched())
859	<	casCleanMe(pred, null); // already gone
860	<	else if ((n = s.next) != null) {
861	<	if (n != s)
862	<	pred.casNext(s, n);
863	<	casCleanMe(pred, null);
864	<	}
865	<	else
883	>	private void sweep() {
884	>	for (Node p = head, s, n; p != null && (s = p.next) != null; ) {
885	>	if (!s.isMatched())
886	>	// Unmatched nodes are never self-linked
887	>	p = s;
888	>	else if ((n = s.next) == null) // trailing node is pinned
889		break;
890	<	}
891	<	return pred;
892	<	}
893	<
894	<	/**
872	<	* Main implementation of Iterator.remove(). Find
873	<	* and unsplice the given data node.
874	<	* @param possiblePred possible predecessor of s
875	<	* @param s the node to remove
876	<	*/
877	<	final void findAndRemoveDataNode(Node possiblePred, Node s) {
878	<	assert s.isData;
879	<	if (s.tryMatchData()) {
880	<	if (possiblePred != null && possiblePred.next == s)
881	<	unsplice(possiblePred, s); // was actual predecessor
882	<	else {
883	<	for (Node pred = null, p = head; p != null; ) {
884	<	if (p == s) {
885	<	unsplice(pred, p);
886	<	break;
887	<	}
888	<	if (p.isUnmatchedRequest())
889	<	break;
890	<	pred = p;
891	<	if ((p = p.next) == pred) { // stale
892	<	pred = null;
893	<	p = head;
894	<	}
895	<	}
896	<	}
890	>	else if (s == n)    // stale
891	>	// No need to also check for p == s, since that implies s == n
892	>	p = head;
893	>	else
894	>	p.casNext(s, n);
895		}
896		}
897
#	Line 1042 | Line 1040 | public class LinkedTransferQueue<E> exte
1040		*/
1041		public boolean tryTransfer(E e, long timeout, TimeUnit unit)
1042		throws InterruptedException {
1043	<	if (xfer(e, true, TIMEOUT, unit.toNanos(timeout)) == null)
1043	>	if (xfer(e, true, TIMED, unit.toNanos(timeout)) == null)
1044		return true;
1045		if (!Thread.interrupted())
1046		return false;
#	Line 1058 | Line 1056 | public class LinkedTransferQueue<E> exte
1056		}
1057
1058		public E poll(long timeout, TimeUnit unit) throws InterruptedException {
1059	<	E e = xfer(null, false, TIMEOUT, unit.toNanos(timeout));
1059	>	E e = xfer(null, false, TIMED, unit.toNanos(timeout));
1060		if (e != null || !Thread.interrupted())
1061		return e;
1062		throw new InterruptedException();
#	Line 1131 | Line 1129 | public class LinkedTransferQueue<E> exte
1129		* @return {@code true} if this queue contains no elements
1130		*/
1131		public boolean isEmpty() {
1132	<	return firstOfMode(true) == null;
1132	>	for (Node p = head; p != null; p = succ(p)) {
1133	>	if (!p.isMatched())
1134	>	return !p.isData;
1135	>	}
1136	>	return true;
1137		}
1138
1139		public boolean hasWaitingConsumer() {
#	Line 1225 | Line 1227 | public class LinkedTransferQueue<E> exte
1227		objectFieldOffset(UNSAFE, "head", LinkedTransferQueue.class);
1228		private static final long tailOffset =
1229		objectFieldOffset(UNSAFE, "tail", LinkedTransferQueue.class);
1230	<	private static final long cleanMeOffset =
1231	<	objectFieldOffset(UNSAFE, "cleanMe", LinkedTransferQueue.class);
1230	>	private static final long sweepVotesOffset =
1231	>	objectFieldOffset(UNSAFE, "sweepVotes", LinkedTransferQueue.class);
1232
1233		static long objectFieldOffset(sun.misc.Unsafe UNSAFE,
1234		String field, Class<?> klazz) {

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines