ViewVC Help

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

root/jsr166/jsr166/src/jsr166y/Phaser.java

Comparing jsr166/src/jsr166y/Phaser.java (file contents):
Revision 1.9 by jsr166, Mon Jan 5 09:11:26 2009 UTC vs.
Revision 1.49 by dl, Fri Nov 5 23:01:47 2010 UTC

#	Line 7 | Line 7
7		package jsr166y;
8
9		import java.util.concurrent.*;
10	<	import java.util.concurrent.atomic.*;
10	>	import java.util.concurrent.atomic.AtomicReference;
11		import java.util.concurrent.locks.LockSupport;
12	–	import sun.misc.Unsafe;
13	–	import java.lang.reflect.*;
12
13		/**
14	<	* A reusable synchronization barrier, similar in functionality to a
14	>	* A reusable synchronization barrier, similar in functionality to
15		* {@link java.util.concurrent.CyclicBarrier CyclicBarrier} and
16		* {@link java.util.concurrent.CountDownLatch CountDownLatch}
17		* but supporting more flexible usage.
18		*
19	<	* <ul>
20	<	*
21	<	* <li> The number of parties synchronizing on a phaser may vary over
22	<	* time.  A task may register to be a party at any time, and may
23	<	* deregister upon arriving at the barrier.  As is the case with most
24	<	* basic synchronization constructs, registration and deregistration
25	<	* affect only internal counts; they do not establish any further
26	<	* internal bookkeeping, so tasks cannot query whether they are
27	<	* registered. (However, you can introduce such bookkeeping by
28	<	* subclassing this class.)
29	<	*
30	<	* <li> Each generation has an associated phase value, starting at
31	<	* zero, and advancing when all parties reach the barrier (wrapping
32	<	* around to zero after reaching {@code Integer.MAX_VALUE}).
33	<	*
34	<	* <li> Like a CyclicBarrier, a Phaser may be repeatedly awaited.
35	<	* Method {@code arriveAndAwaitAdvance} has effect analogous to
36	<	* {@code CyclicBarrier.await}.  However, Phasers separate two
37	<	* aspects of coordination, that may also be invoked independently:
19	>	* <p> <b>Registration.</b> Unlike the case for other barriers, the
20	>	* number of parties <em>registered</em> to synchronize on a phaser
21	>	* may vary over time.  Tasks may be registered at any time (using
22	>	* methods {@link #register}, {@link #bulkRegister}, or forms of
23	>	* constructors establishing initial numbers of parties), and
24	>	* optionally deregistered upon any arrival (using {@link
25	>	* #arriveAndDeregister}).  As is the case with most basic
26	>	* synchronization constructs, registration and deregistration affect
27	>	* only internal counts; they do not establish any further internal
28	>	* bookkeeping, so tasks cannot query whether they are registered.
29	>	* (However, you can introduce such bookkeeping by subclassing this
30	>	* class.)
31	>	*
32	>	* <p> <b>Synchronization.</b> Like a {@code CyclicBarrier}, a {@code
33	>	* Phaser} may be repeatedly awaited.  Method {@link
34	>	* #arriveAndAwaitAdvance} has effect analogous to {@link
35	>	* java.util.concurrent.CyclicBarrier#await CyclicBarrier.await}. Each
36	>	* generation of a {@code Phaser} has an associated phase number. The
37	>	* phase number starts at zero, and advances when all parties arrive
38	>	* at the barrier, wrapping around to zero after reaching {@code
39	>	* Integer.MAX_VALUE}. The use of phase numbers enables independent
40	>	* control of actions upon arrival at a barrier and upon awaiting
41	>	* others, via two kinds of methods that may be invoked by any
42	>	* registered party:
43		*
44		* <ul>
45		*
46	<	*   <li> Arriving at a barrier. Methods {@code arrive} and
47	<	*       {@code arriveAndDeregister} do not block, but return
48	<	*       the phase value current upon entry to the method.
49	<	*
50	<	*   <li> Awaiting others. Method {@code awaitAdvance} requires an
51	<	*       argument indicating the entry phase, and returns when the
52	<	*       barrier advances to a new phase.
53	<	* </ul>
46	>	*   <li> <b>Arrival.</b> Methods {@link #arrive} and
47	>	*       {@link #arriveAndDeregister} record arrival at a
48	>	*       barrier. These methods do not block, but return an associated
49	>	*       <em>arrival phase number</em>; that is, the phase number of
50	>	*       the barrier to which the arrival applied. When the final
51	>	*       party for a given phase arrives, an optional barrier action
52	>	*       is performed and the phase advances.  Barrier actions,
53	>	*       performed by the party triggering a phase advance, are
54	>	*       arranged by overriding method {@link #onAdvance(int, int)},
55	>	*       which also controls termination. Overriding this method is
56	>	*       similar to, but more flexible than, providing a barrier
57	>	*       action to a {@code CyclicBarrier}.
58	>	*
59	>	*   <li> <b>Waiting.</b> Method {@link #awaitAdvance} requires an
60	>	*       argument indicating an arrival phase number, and returns when
61	>	*       the barrier advances to (or is already at) a different phase.
62	>	*       Unlike similar constructions using {@code CyclicBarrier},
63	>	*       method {@code awaitAdvance} continues to wait even if the
64	>	*       waiting thread is interrupted. Interruptible and timeout
65	>	*       versions are also available, but exceptions encountered while
66	>	*       tasks wait interruptibly or with timeout do not change the
67	>	*       state of the barrier. If necessary, you can perform any
68	>	*       associated recovery within handlers of those exceptions,
69	>	*       often after invoking {@code forceTermination}.  Phasers may
70	>	*       also be used by tasks executing in a {@link ForkJoinPool},
71	>	*       which will ensure sufficient parallelism to execute tasks
72	>	*       when others are blocked waiting for a phase to advance.
73		*
74	+	* </ul>
75		*
76	<	* <li> Barrier actions, performed by the task triggering a phase
77	<	* advance while others may be waiting, are arranged by overriding
78	<	* method {@code onAdvance}, that also controls termination.
79	<	* Overriding this method may be used to similar but more flexible
80	<	* effect as providing a barrier action to a CyclicBarrier.
81	<	*
82	<	* <li> Phasers may enter a <em>termination</em> state in which all
60	<	* await actions immediately return, indicating (via a negative phase
61	<	* value) that execution is complete.  Termination is triggered by
62	<	* executing the overridable {@code onAdvance} method that is invoked
63	<	* each time the barrier is about to be tripped. When a Phaser is
64	<	* controlling an action with a fixed number of iterations, it is
76	>	* <p> <b>Termination.</b> A {@code Phaser} may enter a
77	>	* <em>termination</em> state in which all synchronization methods
78	>	* immediately return without updating phaser state or waiting for
79	>	* advance, and indicating (via a negative phase value) that execution
80	>	* is complete.  Termination is triggered when an invocation of {@code
81	>	* onAdvance} returns {@code true}.  As illustrated below, when
82	>	* phasers control actions with a fixed number of iterations, it is
83		* often convenient to override this method to cause termination when
84	<	* the current phase number reaches a threshold. Method
85	<	* {@code forceTermination} is also available to abruptly release
86	<	* waiting threads and allow them to terminate.
84	>	* the current phase number reaches a threshold. Method {@link
85	>	* #forceTermination} is also available to abruptly release waiting
86	>	* threads and allow them to terminate.
87		*
88	<	* <li> Phasers may be tiered to reduce contention. Phasers with large
88	>	* <p> <b>Tiering.</b> Phasers may be <em>tiered</em> (i.e., arranged
89	>	* in tree structures) to reduce contention. Phasers with large
90		* numbers of parties that would otherwise experience heavy
91	<	* synchronization contention costs may instead be arranged in trees.
92	<	* This will typically greatly increase throughput even though it
93	<	* incurs somewhat greater per-operation overhead.
94	<	*
95	<	* <li> By default, {@code awaitAdvance} continues to wait even if
96	<	* the waiting thread is interrupted. And unlike the case in
97	<	* CyclicBarriers, exceptions encountered while tasks wait
98	<	* interruptibly or with timeout do not change the state of the
99	<	* barrier. If necessary, you can perform any associated recovery
100	<	* within handlers of those exceptions, often after invoking
101	<	* {@code forceTermination}.
102	<	*
103	<	* </ul>
91	>	* synchronization contention costs may instead be set up so that
92	>	* groups of sub-phasers share a common parent.  This may greatly
93	>	* increase throughput even though it incurs greater per-operation
94	>	* overhead.
95	>	*
96	>	* <p><b>Monitoring.</b> While synchronization methods may be invoked
97	>	* only by registered parties, the current state of a phaser may be
98	>	* monitored by any caller.  At any given moment there are {@link
99	>	* #getRegisteredParties} parties in total, of which {@link
100	>	* #getArrivedParties} have arrived at the current phase ({@link
101	>	* #getPhase}).  When the remaining ({@link #getUnarrivedParties})
102	>	* parties arrive, the phase advances.  The values returned by these
103	>	* methods may reflect transient states and so are not in general
104	>	* useful for synchronization control.  Method {@link #toString}
105	>	* returns snapshots of these state queries in a form convenient for
106	>	* informal monitoring.
107		*
108		* <p><b>Sample usages:</b>
109		*
110	<	* <p>A Phaser may be used instead of a {@code CountDownLatch} to control
111	<	* a one-shot action serving a variable number of parties. The typical
112	<	* idiom is for the method setting this up to first register, then
113	<	* start the actions, then deregister, as in:
114	<	*
115	<	* <pre>
116	<	*  void runTasks(List&lt;Runnable&gt; list) {
117	<	*    final Phaser phaser = new Phaser(1); // "1" to register self
118	<	*    for (Runnable r : list) {
119	<	*      phaser.register();
120	<	*      new Thread() {
121	<	*        public void run() {
122	<	*          phaser.arriveAndAwaitAdvance(); // await all creation
123	<	*          r.run();
124	<	*          phaser.arriveAndDeregister();   // signal completion
125	<	*        }
126	<	*      }.start();
110	>	* <p>A {@code Phaser} may be used instead of a {@code CountDownLatch}
111	>	* to control a one-shot action serving a variable number of parties.
112	>	* The typical idiom is for the method setting this up to first
113	>	* register, then start the actions, then deregister, as in:
114	>	*
115	>	*  <pre> {@code
116	>	* void runTasks(List<Runnable> tasks) {
117	>	*   final Phaser phaser = new Phaser(1); // "1" to register self
118	>	*   // create and start threads
119	>	*   for (Runnable task : tasks) {
120	>	*     phaser.register();
121	>	*     new Thread() {
122	>	*       public void run() {
123	>	*         phaser.arriveAndAwaitAdvance(); // await all creation
124	>	*         task.run();
125	>	*       }
126	>	*     }.start();
127		*   }
128		*
129	<	*   doSomethingOnBehalfOfWorkers();
130	<	*   phaser.arrive(); // allow threads to start
131	<	*   int p = phaser.arriveAndDeregister(); // deregister self  ...
110	<	*   p = phaser.awaitAdvance(p); // ... and await arrival
111	<	*   otherActions(); // do other things while tasks execute
112	<	*   phaser.awaitAdvance(p); // await final completion
113	<	* }
114	<	* </pre>
129	>	*   // allow threads to start and deregister self
130	>	*   phaser.arriveAndDeregister();
131	>	* }}</pre>
132		*
133		* <p>One way to cause a set of threads to repeatedly perform actions
134		* for a given number of iterations is to override {@code onAdvance}:
135		*
136	<	* <pre>
137	<	*  void startTasks(List&lt;Runnable&gt; list, final int iterations) {
138	<	*    final Phaser phaser = new Phaser() {
139	<	*       public boolean onAdvance(int phase, int registeredParties) {
140	<	*         return phase &gt;= iterations || registeredParties == 0;
136	>	*  <pre> {@code
137	>	* void startTasks(List<Runnable> tasks, final int iterations) {
138	>	*   final Phaser phaser = new Phaser() {
139	>	*     protected boolean onAdvance(int phase, int registeredParties) {
140	>	*       return phase >= iterations || registeredParties == 0;
141	>	*     }
142	>	*   };
143	>	*   phaser.register();
144	>	*   for (final Runnable task : tasks) {
145	>	*     phaser.register();
146	>	*     new Thread() {
147	>	*       public void run() {
148	>	*         do {
149	>	*           task.run();
150	>	*           phaser.arriveAndAwaitAdvance();
151	>	*         } while (!phaser.isTerminated());
152		*       }
153	<	*    };
126	<	*    phaser.register();
127	<	*    for (Runnable r : list) {
128	<	*      phaser.register();
129	<	*      new Thread() {
130	<	*        public void run() {
131	<	*           do {
132	<	*             r.run();
133	<	*             phaser.arriveAndAwaitAdvance();
134	<	*           } while(!phaser.isTerminated();
135	<	*        }
136	<	*      }.start();
153	>	*     }.start();
154		*   }
155		*   phaser.arriveAndDeregister(); // deregister self, don't wait
156	<	* }
140	<	* </pre>
156	>	* }}</pre>
157		*
158	<	* <p> To create a set of tasks using a tree of Phasers,
158	>	* If the main task must later await termination, it
159	>	* may re-register and then execute a similar loop:
160	>	*  <pre> {@code
161	>	*   // ...
162	>	*   phaser.register();
163	>	*   while (!phaser.isTerminated())
164	>	*     phaser.arriveAndAwaitAdvance();}</pre>
165	>	*
166	>	* <p>Related constructions may be used to await particular phase numbers
167	>	* in contexts where you are sure that the phase will never wrap around
168	>	* {@code Integer.MAX_VALUE}. For example:
169	>	*
170	>	*  <pre> {@code
171	>	* void awaitPhase(Phaser phaser, int phase) {
172	>	*   int p = phaser.register(); // assumes caller not already registered
173	>	*   while (p < phase) {
174	>	*     if (phaser.isTerminated())
175	>	*       // ... deal with unexpected termination
176	>	*     else
177	>	*       p = phaser.arriveAndAwaitAdvance();
178	>	*   }
179	>	*   phaser.arriveAndDeregister();
180	>	* }}</pre>
181	>	*
182	>	*
183	>	* <p>To create a set of tasks using a tree of phasers,
184		* you could use code of the following form, assuming a
185	<	* Task class with a constructor accepting a Phaser that
186	<	* it registers for upon construction:
187	<	* <pre>
188	<	*  void build(Task[] actions, int lo, int hi, Phaser b) {
189	<	*    int step = (hi - lo) / TASKS_PER_PHASER;
190	<	*    if (step &gt; 1) {
191	<	*       int i = lo;
192	<	*       while (i &lt; hi) {
193	<	*         int r = Math.min(i + step, hi);
194	<	*         build(actions, i, r, new Phaser(b));
195	<	*         i = r;
196	<	*       }
197	<	*    }
198	<	*    else {
199	<	*      for (int i = lo; i &lt; hi; ++i)
200	<	*        actions[i] = new Task(b);
201	<	*        // assumes new Task(b) performs b.register()
202	<	*    }
162	<	*  }
163	<	*  // .. initially called, for n tasks via
164	<	*  build(new Task[n], 0, n, new Phaser());
165	<	* </pre>
185	>	* Task class with a constructor accepting a phaser that
186	>	* it registers with upon construction:
187	>	*
188	>	*  <pre> {@code
189	>	* void build(Task[] actions, int lo, int hi, Phaser ph) {
190	>	*   if (hi - lo > TASKS_PER_PHASER) {
191	>	*     for (int i = lo; i < hi; i += TASKS_PER_PHASER) {
192	>	*       int j = Math.min(i + TASKS_PER_PHASER, hi);
193	>	*       build(actions, i, j, new Phaser(ph));
194	>	*     }
195	>	*   } else {
196	>	*     for (int i = lo; i < hi; ++i)
197	>	*       actions[i] = new Task(ph);
198	>	*       // assumes new Task(ph) performs ph.register()
199	>	*   }
200	>	* }
201	>	* // .. initially called, for n tasks via
202	>	* build(new Task[n], 0, n, new Phaser());}</pre>
203		*
204		* The best value of {@code TASKS_PER_PHASER} depends mainly on
205		* expected barrier synchronization rates. A value as low as four may
206		* be appropriate for extremely small per-barrier task bodies (thus
207		* high rates), or up to hundreds for extremely large ones.
208		*
172	–	* </pre>
173	–	*
209		* <p><b>Implementation notes</b>: This implementation restricts the
210		* maximum number of parties to 65535. Attempts to register additional
211	<	* parties result in IllegalStateExceptions. However, you can and
211	>	* parties result in {@code IllegalStateException}. However, you can and
212		* should create tiered phasers to accommodate arbitrarily large sets
213		* of participants.
214	+	*
215	+	* @since 1.7
216	+	* @author Doug Lea
217		*/
218		public class Phaser {
219		/*
#	Line 200 | Line 238 | public class Phaser {
238		* and encoding simple, and keeping race windows short.
239		*
240		* Note: there are some cheats in arrive() that rely on unarrived
241	<	* being lowest 16 bits.
241	>	* count being lowest 16 bits.
242		*/
243		private volatile long state;
244
245	<	private static final int ushortBits = 16;
246	<	private static final int ushortMask =  (1 << ushortBits) - 1;
209	<	private static final int phaseMask = 0x7fffffff;
245	>	private static final int ushortMask = 0xffff;
246	>	private static final int phaseMask  = 0x7fffffff;
247
248		private static int unarrivedOf(long s) {
249	<	return (int)(s & ushortMask);
249	>	return (int) (s & ushortMask);
250		}
251
252		private static int partiesOf(long s) {
253	<	return (int)(s & (ushortMask << 16)) >>> 16;
253	>	return ((int) s) >>> 16;
254		}
255
256		private static int phaseOf(long s) {
257	<	return (int)(s >>> 32);
257	>	return (int) (s >>> 32);
258		}
259
260		private static int arrivedOf(long s) {
#	Line 225 | Line 262 | public class Phaser {
262		}
263
264		private static long stateFor(int phase, int parties, int unarrived) {
265	<	return (((long)phase) << 32) | ((parties << 16) | unarrived);
265	>	return ((((long) phase) << 32) | (((long) parties) << 16) |
266	>	(long) unarrived);
267		}
268
269		private static long trippedStateFor(int phase, int parties) {
270	<	return (((long)phase) << 32) | ((parties << 16) | parties);
270	>	long lp = (long) parties;
271	>	return (((long) phase) << 32) | (lp << 16) | lp;
272		}
273
274	<	private static IllegalStateException badBounds(int parties, int unarrived) {
275	<	return new IllegalStateException
276	<	("Attempt to set " + unarrived +
277	<	" unarrived of " + parties + " parties");
274	>	/**
275	>	* Returns message string for bad bounds exceptions.
276	>	*/
277	>	private static String badBounds(int parties, int unarrived) {
278	>	return ("Attempt to set " + unarrived +
279	>	" unarrived of " + parties + " parties");
280		}
281
282		/**
#	Line 244 | Line 285 | public class Phaser {
285		private final Phaser parent;
286
287		/**
288	<	* The root of Phaser tree. Equals this if not in a tree.  Used to
288	>	* The root of phaser tree. Equals this if not in a tree.  Used to
289		* support faster state push-down.
290		*/
291		private final Phaser root;
#	Line 252 | Line 293 | public class Phaser {
293		// Wait queues
294
295		/**
296	<	* Heads of Treiber stacks waiting for nonFJ threads. To eliminate
297	<	* contention while releasing some threads while adding others, we
296	>	* Heads of Treiber stacks for waiting threads. To eliminate
297	>	* contention when releasing some threads while adding others, we
298		* use two of them, alternating across even and odd phases.
299	+	* Subphasers share queues with root to speed up releases.
300		*/
301		private final AtomicReference<QNode> evenQ = new AtomicReference<QNode>();
302		private final AtomicReference<QNode> oddQ  = new AtomicReference<QNode>();
303
304		private AtomicReference<QNode> queueFor(int phase) {
305	<	return (phase & 1) == 0? evenQ : oddQ;
305	>	Phaser r = root;
306	>	return ((phase & 1) == 0) ? r.evenQ : r.oddQ;
307		}
308
309		/**
#	Line 268 | Line 311 | public class Phaser {
311		* root if necessary.
312		*/
313		private long getReconciledState() {
314	<	return parent == null? state : reconcileState();
314	>	return (parent == null) ? state : reconcileState();
315		}
316
317		/**
318		* Recursively resolves state.
319		*/
320		private long reconcileState() {
321	<	Phaser p = parent;
321	>	Phaser par = parent;
322		long s = state;
323	<	if (p != null) {
324	<	while (unarrivedOf(s) == 0 && phaseOf(s) != phaseOf(root.state)) {
325	<	long parentState = p.getReconciledState();
323	>	if (par != null) {
324	>	int phase, rootPhase;
325	>	while ((phase = phaseOf(s)) >= 0 &&
326	>	(rootPhase = phaseOf(root.state)) != phase &&
327	>	(rootPhase < 0 || unarrivedOf(s) == 0)) {
328	>	long parentState = par.getReconciledState();
329		int parentPhase = phaseOf(parentState);
330	<	int phase = phaseOf(s = state);
331	<	if (phase != parentPhase) {
332	<	long next = trippedStateFor(parentPhase, partiesOf(s));
333	<	if (casState(s, next)) {
334	<	releaseWaiters(phase);
335	<	s = next;
336	<	}
330	>	int parties = partiesOf(s);
331	>	long next = trippedStateFor(parentPhase, parties);
332	>	if (phaseOf(root.state) == rootPhase &&
333	>	parentPhase != phase &&
334	>	state == s && casState(s, next)) {
335	>	releaseWaiters(phase);
336	>	if (parties == 0) // exit if the final deregistration
337	>	break;
338		}
339	+	s = state;
340		}
341		}
342		return s;
343		}
344
345		/**
346	<	* Creates a new Phaser without any initially registered parties,
347	<	* initial phase number 0, and no parent.
346	>	* Creates a new phaser without any initially registered parties,
347	>	* initial phase number 0, and no parent. Any thread using this
348	>	* phaser will need to first register for it.
349		*/
350		public Phaser() {
351		this(null);
352		}
353
354		/**
355	<	* Creates a new Phaser with the given numbers of registered
355	>	* Creates a new phaser with the given number of registered
356		* unarrived parties, initial phase number 0, and no parent.
357	<	* @param parties the number of parties required to trip barrier.
357	>	*
358	>	* @param parties the number of parties required to trip barrier
359		* @throws IllegalArgumentException if parties less than zero
360	<	* or greater than the maximum number of parties supported.
360	>	* or greater than the maximum number of parties supported
361		*/
362		public Phaser(int parties) {
363		this(null, parties);
364		}
365
366		/**
367	<	* Creates a new Phaser with the given parent, without any
367	>	* Creates a new phaser with the given parent, without any
368		* initially registered parties. If parent is non-null this phaser
369		* is registered with the parent and its initial phase number is
370		* the same as that of parent phaser.
371	<	* @param parent the parent phaser.
371	>	*
372	>	* @param parent the parent phaser
373		*/
374		public Phaser(Phaser parent) {
375		int phase = 0;
#	Line 333 | Line 384 | public class Phaser {
384		}
385
386		/**
387	<	* Creates a new Phaser with the given parent and numbers of
388	<	* registered unarrived parties. If parent is non-null this phaser
387	>	* Creates a new phaser with the given parent and number of
388	>	* registered unarrived parties. If parent is non-null, this phaser
389		* is registered with the parent and its initial phase number is
390		* the same as that of parent phaser.
391	<	* @param parent the parent phaser.
392	<	* @param parties the number of parties required to trip barrier.
391	>	*
392	>	* @param parent the parent phaser
393	>	* @param parties the number of parties required to trip barrier
394		* @throws IllegalArgumentException if parties less than zero
395	<	* or greater than the maximum number of parties supported.
395	>	* or greater than the maximum number of parties supported
396		*/
397		public Phaser(Phaser parent, int parties) {
398		if (parties < 0 || parties > ushortMask)
#	Line 358 | Line 410 | public class Phaser {
410
411		/**
412		* Adds a new unarrived party to this phaser.
413	<	* @return the current barrier phase number upon registration
413	>	* If an ongoing invocation of {@link #onAdvance} is in progress,
414	>	* this method waits until its completion before registering.
415	>	*
416	>	* @return the arrival phase number to which this registration applied
417		* @throws IllegalStateException if attempting to register more
418	<	* than the maximum supported number of parties.
418	>	* than the maximum supported number of parties
419		*/
420		public int register() {
421		return doRegister(1);
#	Line 368 | Line 423 | public class Phaser {
423
424		/**
425		* Adds the given number of new unarrived parties to this phaser.
426	<	* @param parties the number of parties required to trip barrier.
427	<	* @return the current barrier phase number upon registration
426	>	* If an ongoing invocation of {@link #onAdvance} is in progress,
427	>	* this method waits until its completion before registering.
428	>	*
429	>	* @param parties the number of additional parties required to trip barrier
430	>	* @return the arrival phase number to which this registration applied
431		* @throws IllegalStateException if attempting to register more
432	<	* than the maximum supported number of parties.
432	>	* than the maximum supported number of parties
433	>	* @throws IllegalArgumentException if {@code parties < 0}
434		*/
435		public int bulkRegister(int parties) {
436		if (parties < 0)
#	Line 385 | Line 444 | public class Phaser {
444		* Shared code for register, bulkRegister
445		*/
446		private int doRegister(int registrations) {
447	+	Phaser par = parent;
448	+	long s;
449		int phase;
450	<	for (;;) {
451	<	long s = getReconciledState();
452	<	phase = phaseOf(s);
453	<	int unarrived = unarrivedOf(s) + registrations;
454	<	int parties = partiesOf(s) + registrations;
455	<	if (phase < 0)
456	<	break;
457	<	if (parties > ushortMask || unarrived > ushortMask)
458	<	throw badBounds(parties, unarrived);
459	<	if (phase == phaseOf(root.state) &&
460	<	casState(s, stateFor(phase, parties, unarrived)))
461	<	break;
450	>	while ((phase = phaseOf(s = par==null? state:reconcileState())) >= 0) {
451	>	int p = partiesOf(s);
452	>	int u = unarrivedOf(s);
453	>	int unarrived = u + registrations;
454	>	int parties = p + registrations;
455	>	if (par == null || phase == phaseOf(root.state)) {
456	>	if (parties > ushortMask || unarrived > ushortMask)
457	>	throw new IllegalStateException(badBounds(parties,
458	>	unarrived));
459	>	else if (p != 0 && u == 0)       // back off if advancing
460	>	Thread.yield();              // not worth actually blocking
461	>	else if (casState(s, stateFor(phase, parties, unarrived)))
462	>	break;
463	>	}
464		}
465		return phase;
466		}
467
468		/**
469		* Arrives at the barrier, but does not wait for others.  (You can
470	<	* in turn wait for others via {@link #awaitAdvance}).
470	>	* in turn wait for others via {@link #awaitAdvance}).  It is an
471	>	* unenforced usage error for an unregistered party to invoke this
472	>	* method.
473		*
474	<	* @return the barrier phase number upon entry to this method, or a
410	<	* negative value if terminated;
474	>	* @return the arrival phase number, or a negative value if terminated
475		* @throws IllegalStateException if not terminated and the number
476	<	* of unarrived parties would become negative.
476	>	* of unarrived parties would become negative
477		*/
478		public int arrive() {
479	+	Phaser par = parent;
480	+	long s;
481		int phase;
482	<	for (;;) {
417	<	long s = state;
418	<	phase = phaseOf(s);
482	>	while ((phase = phaseOf(s = par==null? state:reconcileState())) >= 0) {
483		int parties = partiesOf(s);
484		int unarrived = unarrivedOf(s) - 1;
485	<	if (unarrived > 0) {        // Not the last arrival
486	<	if (casState(s, s - 1)) // s-1 adds one arrival
485	>	if (parties == 0 || unarrived < 0)
486	>	throw new IllegalStateException(badBounds(parties,
487	>	unarrived));
488	>	else if (unarrived > 0) {           // Not the last arrival
489	>	if (casState(s, s - 1))         // s-1 adds one arrival
490		break;
491		}
492	<	else if (unarrived == 0) {  // the last arrival
493	<	Phaser par = parent;
494	<	if (par == null) {      // directly trip
495	<	if (casState
496	<	(s,
497	<	trippedStateFor(onAdvance(phase, parties)? -1 :
431	<	((phase + 1) & phaseMask), parties))) {
432	<	releaseWaiters(phase);
433	<	break;
434	<	}
435	<	}
436	<	else {                  // cascade to parent
437	<	if (casState(s, s - 1)) { // zeroes unarrived
438	<	par.arrive();
439	<	reconcileState();
440	<	break;
441	<	}
492	>	else if (par == null) {             // directly trip
493	>	if (casState(s, trippedStateFor(onAdvance(phase, parties) ? -1 :
494	>	((phase + 1) & phaseMask),
495	>	parties))) {
496	>	releaseWaiters(phase);
497	>	break;
498		}
499		}
500	<	else if (phase < 0) // Don't throw exception if terminated
501	<	break;
446	<	else if (phase != phaseOf(root.state)) // or if unreconciled
500	>	else if (phaseOf(root.state) == phase && casState(s, s - 1)) {
501	>	par.arrive();                   // cascade to parent
502		reconcileState();
503	<	else
504	<	throw badBounds(parties, unarrived);
503	>	break;
504	>	}
505		}
506		return phase;
507		}
508
509		/**
510	<	* Arrives at the barrier, and deregisters from it, without
511	<	* waiting for others. Deregistration reduces number of parties
510	>	* Arrives at the barrier and deregisters from it without waiting
511	>	* for others. Deregistration reduces the number of parties
512		* required to trip the barrier in future phases.  If this phaser
513		* has a parent, and deregistration causes this phaser to have
514	<	* zero parties, this phaser is also deregistered from its parent.
514	>	* zero parties, this phaser also arrives at and is deregistered
515	>	* from its parent.  It is an unenforced usage error for an
516	>	* unregistered party to invoke this method.
517		*
518	<	* @return the current barrier phase number upon entry to
462	<	* this method, or a negative value if terminated;
518	>	* @return the arrival phase number, or a negative value if terminated
519		* @throws IllegalStateException if not terminated and the number
520	<	* of registered or unarrived parties would become negative.
520	>	* of registered or unarrived parties would become negative
521		*/
522		public int arriveAndDeregister() {
523	<	// similar code to arrive, but too different to merge
523	>	// similar to arrive, but too different to merge
524		Phaser par = parent;
525	+	long s;
526		int phase;
527	<	for (;;) {
471	<	long s = state;
472	<	phase = phaseOf(s);
527	>	while ((phase = phaseOf(s = par==null? state:reconcileState())) >= 0) {
528		int parties = partiesOf(s) - 1;
529		int unarrived = unarrivedOf(s) - 1;
530	<	if (parties >= 0) {
531	<	if (unarrived > 0 || (unarrived == 0 && par != null)) {
532	<	if (casState
533	<	(s,
534	<	stateFor(phase, parties, unarrived))) {
535	<	if (unarrived == 0) {
536	<	par.arriveAndDeregister();
537	<	reconcileState();
538	<	}
539	<	break;
540	<	}
541	<	continue;
487	<	}
488	<	if (unarrived == 0) {
489	<	if (casState
490	<	(s,
491	<	trippedStateFor(onAdvance(phase, parties)? -1 :
492	<	((phase + 1) & phaseMask), parties))) {
493	<	releaseWaiters(phase);
494	<	break;
495	<	}
496	<	continue;
497	<	}
498	<	if (phase < 0)
530	>	if (parties < 0 || unarrived < 0)
531	>	throw new IllegalStateException(badBounds(parties,
532	>	unarrived));
533	>	else if (unarrived > 0) {
534	>	if (casState(s, stateFor(phase, parties, unarrived)))
535	>	break;
536	>	}
537	>	else if (par == null) {
538	>	if (casState(s, trippedStateFor(onAdvance(phase, parties)? -1:
539	>	(phase + 1) & phaseMask,
540	>	parties))) {
541	>	releaseWaiters(phase);
542		break;
500	–	if (par != null && phase != phaseOf(root.state)) {
501	–	reconcileState();
502	–	continue;
543		}
544		}
545	<	throw badBounds(parties, unarrived);
545	>	else if (phaseOf(root.state) == phase &&
546	>	casState(s, stateFor(phase, parties, 0))) {
547	>	if (parties == 0)
548	>	par.arriveAndDeregister();
549	>	else
550	>	par.arrive();
551	>	reconcileState();
552	>	break;
553	>	}
554		}
555		return phase;
556		}
557
558		/**
559		* Arrives at the barrier and awaits others. Equivalent in effect
560	<	* to {@code awaitAdvance(arrive())}.  If you instead need to
561	<	* await with interruption of timeout, and/or deregister upon
562	<	* arrival, you can arrange them using analogous constructions.
563	<	* @return the phase on entry to this method
560	>	* to {@code awaitAdvance(arrive())}.  If you need to await with
561	>	* interruption or timeout, you can arrange this with an analogous
562	>	* construction using one of the other forms of the {@code
563	>	* awaitAdvance} method.  If instead you need to deregister upon
564	>	* arrival, use {@link #arriveAndDeregister}. It is an unenforced
565	>	* usage error for an unregistered party to invoke this method.
566	>	*
567	>	* @return the arrival phase number, or a negative number if terminated
568		* @throws IllegalStateException if not terminated and the number
569	<	* of unarrived parties would become negative.
569	>	* of unarrived parties would become negative
570		*/
571		public int arriveAndAwaitAdvance() {
572		return awaitAdvance(arrive());
573		}
574
575		/**
576	<	* Awaits the phase of the barrier to advance from the given
577	<	* value, or returns immediately if argument is negative or this
578	<	* barrier is terminated.
579	<	* @param phase the phase on entry to this method
580	<	* @return the phase on exit from this method
576	>	* Awaits the phase of the barrier to advance from the given phase
577	>	* value, returning immediately if the current phase of the
578	>	* barrier is not equal to the given phase value or this barrier
579	>	* is terminated.  It is an unenforced usage error for an
580	>	* unregistered party to invoke this method.
581	>	*
582	>	* @param phase an arrival phase number, or negative value if
583	>	* terminated; this argument is normally the value returned by a
584	>	* previous call to {@code arrive} or its variants
585	>	* @return the next arrival phase number, or a negative value
586	>	* if terminated or argument is negative
587		*/
588		public int awaitAdvance(int phase) {
589		if (phase < 0)
590		return phase;
591	<	long s = getReconciledState();
534	<	int p = phaseOf(s);
591	>	int p = getPhase();
592		if (p != phase)
593		return p;
537	–	if (unarrivedOf(s) == 0)
538	–	parent.awaitAdvance(phase);
539	–	// Fall here even if parent waited, to reconcile and help release
594		return untimedWait(phase);
595		}
596
597		/**
598	<	* Awaits the phase of the barrier to advance from the given
599	<	* value, or returns immediately if argument is negative or this
600	<	* barrier is terminated, or throws InterruptedException if
601	<	* interrupted while waiting.
602	<	* @param phase the phase on entry to this method
603	<	* @return the phase on exit from this method
598	>	* Awaits the phase of the barrier to advance from the given phase
599	>	* value, throwing {@code InterruptedException} if interrupted
600	>	* while waiting, or returning immediately if the current phase of
601	>	* the barrier is not equal to the given phase value or this
602	>	* barrier is terminated. It is an unenforced usage error for an
603	>	* unregistered party to invoke this method.
604	>	*
605	>	* @param phase an arrival phase number, or negative value if
606	>	* terminated; this argument is normally the value returned by a
607	>	* previous call to {@code arrive} or its variants
608	>	* @return the next arrival phase number, or a negative value
609	>	* if terminated or argument is negative
610		* @throws InterruptedException if thread interrupted while waiting
611		*/
612	<	public int awaitAdvanceInterruptibly(int phase) throws InterruptedException {
612	>	public int awaitAdvanceInterruptibly(int phase)
613	>	throws InterruptedException {
614		if (phase < 0)
615		return phase;
616	<	long s = getReconciledState();
556	<	int p = phaseOf(s);
616	>	int p = getPhase();
617		if (p != phase)
618		return p;
559	–	if (unarrivedOf(s) != 0)
560	–	parent.awaitAdvanceInterruptibly(phase);
619		return interruptibleWait(phase);
620		}
621
622		/**
623	<	* Awaits the phase of the barrier to advance from the given value
624	<	* or the given timeout elapses, or returns immediately if
625	<	* argument is negative or this barrier is terminated.
626	<	* @param phase the phase on entry to this method
627	<	* @return the phase on exit from this method
623	>	* Awaits the phase of the barrier to advance from the given phase
624	>	* value or the given timeout to elapse, throwing {@code
625	>	* InterruptedException} if interrupted while waiting, or
626	>	* returning immediately if the current phase of the barrier is
627	>	* not equal to the given phase value or this barrier is
628	>	* terminated.  It is an unenforced usage error for an
629	>	* unregistered party to invoke this method.
630	>	*
631	>	* @param phase an arrival phase number, or negative value if
632	>	* terminated; this argument is normally the value returned by a
633	>	* previous call to {@code arrive} or its variants
634	>	* @param timeout how long to wait before giving up, in units of
635	>	*        {@code unit}
636	>	* @param unit a {@code TimeUnit} determining how to interpret the
637	>	*        {@code timeout} parameter
638	>	* @return the next arrival phase number, or a negative value
639	>	* if terminated or argument is negative
640		* @throws InterruptedException if thread interrupted while waiting
641		* @throws TimeoutException if timed out while waiting
642		*/
643	<	public int awaitAdvanceInterruptibly(int phase, long timeout, TimeUnit unit)
643	>	public int awaitAdvanceInterruptibly(int phase,
644	>	long timeout, TimeUnit unit)
645		throws InterruptedException, TimeoutException {
646	+	long nanos = unit.toNanos(timeout);
647		if (phase < 0)
648		return phase;
649	<	long s = getReconciledState();
578	<	int p = phaseOf(s);
649	>	int p = getPhase();
650		if (p != phase)
651		return p;
652	<	if (unarrivedOf(s) == 0)
582	<	parent.awaitAdvanceInterruptibly(phase, timeout, unit);
583	<	return timedWait(phase, unit.toNanos(timeout));
652	>	return timedWait(phase, nanos);
653		}
654
655		/**
#	Line 591 | Line 660 | public class Phaser {
660		* unexpected exceptions.
661		*/
662		public void forceTermination() {
663	<	for (;;) {
664	<	long s = getReconciledState();
665	<	int phase = phaseOf(s);
666	<	int parties = partiesOf(s);
667	<	int unarrived = unarrivedOf(s);
668	<	if (phase < 0 ||
669	<	casState(s, stateFor(-1, parties, unarrived))) {
601	<	releaseWaiters(0);
602	<	releaseWaiters(1);
603	<	if (parent != null)
604	<	parent.forceTermination();
605	<	return;
606	<	}
607	<	}
663	>	Phaser r = root;    // force at root then reconcile
664	>	long s;
665	>	while (phaseOf(s = r.state) >= 0)
666	>	r.casState(s, stateFor(-1, partiesOf(s), unarrivedOf(s)));
667	>	reconcileState();
668	>	releaseWaiters(0);  // ensure wakeups on both queues
669	>	releaseWaiters(1);
670		}
671
672		/**
673		* Returns the current phase number. The maximum phase number is
674		* {@code Integer.MAX_VALUE}, after which it restarts at
675		* zero. Upon termination, the phase number is negative.
676	+	*
677		* @return the phase number, or a negative value if terminated
678		*/
679		public final int getPhase() {
#	Line 618 | Line 681 | public class Phaser {
681		}
682
683		/**
621	–	* Returns {@code true} if the current phase number equals the given phase.
622	–	* @param phase the phase
623	–	* @return {@code true} if the current phase number equals the given phase
624	–	*/
625	–	public final boolean hasPhase(int phase) {
626	–	return phaseOf(getReconciledState()) == phase;
627	–	}
628	–
629	–	/**
684		* Returns the number of parties registered at this barrier.
685	+	*
686		* @return the number of parties
687		*/
688		public int getRegisteredParties() {
689	<	return partiesOf(state);
689	>	return partiesOf(getReconciledState());
690		}
691
692		/**
693	<	* Returns the number of parties that have arrived at the current
694	<	* phase of this barrier.
693	>	* Returns the number of registered parties that have arrived at
694	>	* the current phase of this barrier.
695	>	*
696		* @return the number of arrived parties
697		*/
698		public int getArrivedParties() {
699	<	return arrivedOf(state);
699	>	return arrivedOf(getReconciledState());
700		}
701
702		/**
703		* Returns the number of registered parties that have not yet
704		* arrived at the current phase of this barrier.
705	+	*
706		* @return the number of unarrived parties
707		*/
708		public int getUnarrivedParties() {
709	<	return unarrivedOf(state);
709	>	return unarrivedOf(getReconciledState());
710		}
711
712		/**
713	<	* Returns the parent of this phaser, or null if none.
714	<	* @return the parent of this phaser, or null if none
713	>	* Returns the parent of this phaser, or {@code null} if none.
714	>	*
715	>	* @return the parent of this phaser, or {@code null} if none
716		*/
717		public Phaser getParent() {
718		return parent;
#	Line 663 | Line 721 | public class Phaser {
721		/**
722		* Returns the root ancestor of this phaser, which is the same as
723		* this phaser if it has no parent.
724	+	*
725		* @return the root ancestor of this phaser
726		*/
727		public Phaser getRoot() {
#	Line 671 | Line 730 | public class Phaser {
730
731		/**
732		* Returns {@code true} if this barrier has been terminated.
733	+	*
734		* @return {@code true} if this barrier has been terminated
735		*/
736		public boolean isTerminated() {
#	Line 678 | Line 738 | public class Phaser {
738		}
739
740		/**
741	<	* Overridable method to perform an action upon phase advance, and
742	<	* to control termination. This method is invoked whenever the
743	<	* barrier is tripped (and thus all other waiting parties are
744	<	* dormant). If it returns true, then, rather than advance the
745	<	* phase number, this barrier will be set to a final termination
746	<	* state, and subsequent calls to {@code isTerminated} will
747	<	* return true.
741	>	* Overridable method to perform an action upon impending phase
742	>	* advance, and to control termination. This method is invoked
743	>	* upon arrival of the party tripping the barrier (when all other
744	>	* waiting parties are dormant).  If this method returns {@code
745	>	* true}, then, rather than advance the phase number, this barrier
746	>	* will be set to a final termination state, and subsequent calls
747	>	* to {@link #isTerminated} will return true. Any (unchecked)
748	>	* Exception or Error thrown by an invocation of this method is
749	>	* propagated to the party attempting to trip the barrier, in
750	>	* which case no advance occurs.
751	>	*
752	>	* <p>The arguments to this method provide the state of the phaser
753	>	* prevailing for the current transition. (When called from within
754	>	* an implementation of {@code onAdvance} the values returned by
755	>	* methods such as {@code getPhase} may or may not reliably
756	>	* indicate the state to which this transition applies.)
757		*
758	<	* <p> The default version returns true when the number of
758	>	* <p>The default version returns {@code true} when the number of
759		* registered parties is zero. Normally, overrides that arrange
760		* termination for other reasons should also preserve this
761		* property.
762		*
694	–	* <p> You may override this method to perform an action with side
695	–	* effects visible to participating tasks, but it is in general
696	–	* only sensible to do so in designs where all parties register
697	–	* before any arrive, and all {@code awaitAdvance} at each phase.
698	–	* Otherwise, you cannot ensure lack of interference. In
699	–	* particular, this method may be invoked more than once per
700	–	* transition if other parties successfully register while the
701	–	* invocation of this method is in progress, thus postponing the
702	–	* transition until those parties also arrive, re-triggering this
703	–	* method.
704	–	*
763		* @param phase the phase number on entering the barrier
764		* @param registeredParties the current number of registered parties
765		* @return {@code true} if this barrier should terminate
#	Line 729 | Line 787 | public class Phaser {
787
788		// methods for waiting
789
732	–	/** The number of CPUs, for spin control */
733	–	static final int NCPUS = Runtime.getRuntime().availableProcessors();
734	–
790		/**
791	<	* The number of times to spin before blocking in timed waits.
737	<	* The value is empirically derived.
791	>	* Wait nodes for Treiber stack representing wait queue
792		*/
793	<	static final int maxTimedSpins = (NCPUS < 2)? 0 : 32;
793	>	static final class QNode implements ForkJoinPool.ManagedBlocker {
794	>	final Phaser phaser;
795	>	final int phase;
796	>	final long startTime;
797	>	final long nanos;
798	>	final boolean timed;
799	>	final boolean interruptible;
800	>	volatile boolean wasInterrupted = false;
801	>	volatile Thread thread; // nulled to cancel wait
802	>	QNode next;
803
804	<	/**
805	<	* The number of times to spin before blocking in untimed waits.
806	<	* This is greater than timed value because untimed waits spin
807	<	* faster since they don't need to check times on each spin.
808	<	*/
809	<	static final int maxUntimedSpins = maxTimedSpins * 32;
804	>	QNode(Phaser phaser, int phase, boolean interruptible,
805	>	boolean timed, long startTime, long nanos) {
806	>	this.phaser = phaser;
807	>	this.phase = phase;
808	>	this.timed = timed;
809	>	this.interruptible = interruptible;
810	>	this.startTime = startTime;
811	>	this.nanos = nanos;
812	>	thread = Thread.currentThread();
813	>	}
814
815	<	/**
816	<	* The number of nanoseconds for which it is faster to spin
817	<	* rather than to use timed park. A rough estimate suffices.
818	<	*/
819	<	static final long spinForTimeoutThreshold = 1000L;
815	>	public boolean isReleasable() {
816	>	return (thread == null ||
817	>	phaser.getPhase() != phase ||
818	>	(interruptible && wasInterrupted) ||
819	>	(timed && (nanos - (System.nanoTime() - startTime)) <= 0));
820	>	}
821
822	<	/**
823	<	* Wait nodes for Treiber stack representing wait queue for non-FJ
824	<	* tasks.
825	<	*/
826	<	static final class QNode {
827	<	QNode next;
828	<	volatile Thread thread; // nulled to cancel wait
829	<	QNode() {
830	<	thread = Thread.currentThread();
822	>	public boolean block() {
823	>	if (Thread.interrupted()) {
824	>	wasInterrupted = true;
825	>	if (interruptible)
826	>	return true;
827	>	}
828	>	if (!timed)
829	>	LockSupport.park(this);
830	>	else {
831	>	long waitTime = nanos - (System.nanoTime() - startTime);
832	>	if (waitTime <= 0)
833	>	return true;
834	>	LockSupport.parkNanos(this, waitTime);
835	>	}
836	>	return isReleasable();
837		}
838	+
839		void signal() {
840		Thread t = thread;
841		if (t != null) {
#	Line 768 | Line 843 | public class Phaser {
843		LockSupport.unpark(t);
844		}
845		}
846	+
847	+	boolean doWait() {
848	+	if (thread != null) {
849	+	try {
850	+	ForkJoinPool.managedBlock(this);
851	+	} catch (InterruptedException ie) {
852	+	wasInterrupted = true; // can't currently happen
853	+	}
854	+	}
855	+	return wasInterrupted;
856	+	}
857		}
858
859		/**
860	<	* Removes and signals waiting threads from wait queue
860	>	* Removes and signals waiting threads from wait queue.
861		*/
862		private void releaseWaiters(int phase) {
863		AtomicReference<QNode> head = queueFor(phase);
#	Line 783 | Line 869 | public class Phaser {
869		}
870
871		/**
872	+	* Tries to enqueue given node in the appropriate wait queue.
873	+	*
874	+	* @return true if successful
875	+	*/
876	+	private boolean tryEnqueue(QNode node) {
877	+	AtomicReference<QNode> head = queueFor(node.phase);
878	+	return head.compareAndSet(node.next = head.get(), node);
879	+	}
880	+
881	+	/**
882		* Enqueues node and waits unless aborted or signalled.
883	+	*
884	+	* @return current phase
885		*/
886		private int untimedWait(int phase) {
789	–	int spins = maxUntimedSpins;
887		QNode node = null;
791	–	boolean interrupted = false;
888		boolean queued = false;
889	+	boolean interrupted = false;
890		int p;
891		while ((p = getPhase()) == phase) {
892	<	interrupted = Thread.interrupted();
893	<	if (node != null) {
894	<	if (!queued) {
895	<	AtomicReference<QNode> head = queueFor(phase);
896	<	queued = head.compareAndSet(node.next = head.get(), node);
897	<	}
898	<	else if (node.thread != null)
899	<	LockSupport.park(this);
803	<	}
804	<	else if (spins <= 0)
805	<	node = new QNode();
806	<	else
807	<	--spins;
892	>	if (Thread.interrupted())
893	>	interrupted = true;
894	>	else if (node == null)
895	>	node = new QNode(this, phase, false, false, 0, 0);
896	>	else if (!queued)
897	>	queued = tryEnqueue(node);
898	>	else if (node.doWait())
899	>	interrupted = true;
900		}
901		if (node != null)
902		node.thread = null;
903	+	releaseWaiters(phase);
904		if (interrupted)
905		Thread.currentThread().interrupt();
813	–	releaseWaiters(phase);
906		return p;
907		}
908
909		/**
910	<	* Messier interruptible version
910	>	* Interruptible version
911	>	* @return current phase
912		*/
913		private int interruptibleWait(int phase) throws InterruptedException {
821	–	int spins = maxUntimedSpins;
914		QNode node = null;
915		boolean queued = false;
916		boolean interrupted = false;
917		int p;
918	<	while ((p = getPhase()) == phase) {
919	<	if (interrupted = Thread.interrupted())
920	<	break;
921	<	if (node != null) {
922	<	if (!queued) {
923	<	AtomicReference<QNode> head = queueFor(phase);
924	<	queued = head.compareAndSet(node.next = head.get(), node);
925	<	}
926	<	else if (node.thread != null)
835	<	LockSupport.park(this);
836	<	}
837	<	else if (spins <= 0)
838	<	node = new QNode();
839	<	else
840	<	--spins;
918	>	while ((p = getPhase()) == phase && !interrupted) {
919	>	if (Thread.interrupted())
920	>	interrupted = true;
921	>	else if (node == null)
922	>	node = new QNode(this, phase, true, false, 0, 0);
923	>	else if (!queued)
924	>	queued = tryEnqueue(node);
925	>	else if (node.doWait())
926	>	interrupted = true;
927		}
928		if (node != null)
929		node.thread = null;
930	+	if (p != phase || (p = getPhase()) != phase)
931	+	releaseWaiters(phase);
932		if (interrupted)
933		throw new InterruptedException();
846	–	releaseWaiters(phase);
934		return p;
935		}
936
937		/**
938	<	* Even messier timeout version.
938	>	* Timeout version.
939	>	* @return current phase
940		*/
941		private int timedWait(int phase, long nanos)
942		throws InterruptedException, TimeoutException {
943	+	long startTime = System.nanoTime();
944	+	QNode node = null;
945	+	boolean queued = false;
946	+	boolean interrupted = false;
947		int p;
948	<	if ((p = getPhase()) == phase) {
949	<	long lastTime = System.nanoTime();
950	<	int spins = maxTimedSpins;
951	<	QNode node = null;
952	<	boolean queued = false;
953	<	boolean interrupted = false;
954	<	while ((p = getPhase()) == phase) {
955	<	if (interrupted = Thread.interrupted())
956	<	break;
957	<	long now = System.nanoTime();
958	<	if ((nanos -= now - lastTime) <= 0)
867	<	break;
868	<	lastTime = now;
869	<	if (node != null) {
870	<	if (!queued) {
871	<	AtomicReference<QNode> head = queueFor(phase);
872	<	queued = head.compareAndSet(node.next = head.get(), node);
873	<	}
874	<	else if (node.thread != null &&
875	<	nanos > spinForTimeoutThreshold) {
876	<	LockSupport.parkNanos(this, nanos);
877	<	}
878	<	}
879	<	else if (spins <= 0)
880	<	node = new QNode();
881	<	else
882	<	--spins;
883	<	}
884	<	if (node != null)
885	<	node.thread = null;
886	<	if (interrupted)
887	<	throw new InterruptedException();
888	<	if (p == phase && (p = getPhase()) == phase)
889	<	throw new TimeoutException();
948	>	while ((p = getPhase()) == phase && !interrupted) {
949	>	if (Thread.interrupted())
950	>	interrupted = true;
951	>	else if (nanos - (System.nanoTime() - startTime) <= 0)
952	>	break;
953	>	else if (node == null)
954	>	node = new QNode(this, phase, true, true, startTime, nanos);
955	>	else if (!queued)
956	>	queued = tryEnqueue(node);
957	>	else if (node.doWait())
958	>	interrupted = true;
959		}
960	<	releaseWaiters(phase);
960	>	if (node != null)
961	>	node.thread = null;
962	>	if (p != phase || (p = getPhase()) != phase)
963	>	releaseWaiters(phase);
964	>	if (interrupted)
965	>	throw new InterruptedException();
966	>	if (p == phase)
967	>	throw new TimeoutException();
968		return p;
969		}
970
971	<	// Temporary Unsafe mechanics for preliminary release
971	>	// Unsafe mechanics
972
973	<	static final Unsafe _unsafe;
974	<	static final long stateOffset;
973	>	private static final sun.misc.Unsafe UNSAFE = getUnsafe();
974	>	private static final long stateOffset =
975	>	objectFieldOffset("state", Phaser.class);
976
977	<	static {
977	>	private final boolean casState(long cmp, long val) {
978	>	return UNSAFE.compareAndSwapLong(this, stateOffset, cmp, val);
979	>	}
980	>
981	>	private static long objectFieldOffset(String field, Class<?> klazz) {
982		try {
983	<	if (Phaser.class.getClassLoader() != null) {
984	<	Field f = Unsafe.class.getDeclaredField("theUnsafe");
985	<	f.setAccessible(true);
986	<	_unsafe = (Unsafe)f.get(null);
987	<	}
988	<	else
908	<	_unsafe = Unsafe.getUnsafe();
909	<	stateOffset = _unsafe.objectFieldOffset
910	<	(Phaser.class.getDeclaredField("state"));
911	<	} catch (Exception e) {
912	<	throw new RuntimeException("Could not initialize intrinsics", e);
983	>	return UNSAFE.objectFieldOffset(klazz.getDeclaredField(field));
984	>	} catch (NoSuchFieldException e) {
985	>	// Convert Exception to corresponding Error
986	>	NoSuchFieldError error = new NoSuchFieldError(field);
987	>	error.initCause(e);
988	>	throw error;
989		}
990		}
991
992	<	final boolean casState(long cmp, long val) {
993	<	return _unsafe.compareAndSwapLong(this, stateOffset, cmp, val);
992	>	/**
993	>	* Returns a sun.misc.Unsafe.  Suitable for use in a 3rd party package.
994	>	* Replace with a simple call to Unsafe.getUnsafe when integrating
995	>	* into a jdk.
996	>	*
997	>	* @return a sun.misc.Unsafe
998	>	*/
999	>	private static sun.misc.Unsafe getUnsafe() {
1000	>	try {
1001	>	return sun.misc.Unsafe.getUnsafe();
1002	>	} catch (SecurityException se) {
1003	>	try {
1004	>	return java.security.AccessController.doPrivileged
1005	>	(new java.security
1006	>	.PrivilegedExceptionAction<sun.misc.Unsafe>() {
1007	>	public sun.misc.Unsafe run() throws Exception {
1008	>	java.lang.reflect.Field f = sun.misc
1009	>	.Unsafe.class.getDeclaredField("theUnsafe");
1010	>	f.setAccessible(true);
1011	>	return (sun.misc.Unsafe) f.get(null);
1012	>	}});
1013	>	} catch (java.security.PrivilegedActionException e) {
1014	>	throw new RuntimeException("Could not initialize intrinsics",
1015	>	e.getCause());
1016	>	}
1017	>	}
1018		}
1019		}

Diff Legend

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