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 1.4 by dl, Sat Sep 6 13:19:17 2008 UTC vs.
Revision 1.15 by jsr166, Tue Jul 21 18:11:44 2009 UTC

#	Line 5 | Line 5
5		*/
6
7		package jsr166y;
8	+
9		import java.util.concurrent.*;
10		import java.util.concurrent.atomic.*;
11		import java.util.concurrent.locks.LockSupport;
#	Line 13 | Line 14 | import java.lang.reflect.*;
14
15		/**
16		* A reusable synchronization barrier, similar in functionality to a
17	<	* {@link java.util.concurrent.CyclicBarrier} and {@link
18	<	* java.util.concurrent.CountDownLatch} but supporting more flexible
19	<	* usage.
17	>	* {@link java.util.concurrent.CyclicBarrier CyclicBarrier} and
18	>	* {@link java.util.concurrent.CountDownLatch CountDownLatch}
19	>	* but supporting more flexible usage.
20		*
21		* <ul>
22		*
#	Line 25 | Line 26 | import java.lang.reflect.*;
26		* basic synchronization constructs, registration and deregistration
27		* affect only internal counts; they do not establish any further
28		* internal bookkeeping, so tasks cannot query whether they are
29	<	* registered. (However, you can introduce such bookkeeping in by
29	>	* registered. (However, you can introduce such bookkeeping by
30		* subclassing this class.)
31		*
32		* <li> Each generation has an associated phase value, starting at
33		* zero, and advancing when all parties reach the barrier (wrapping
34	<	* around to zero after reaching <tt>Integer.MAX_VALUE</tt>).
34	>	* around to zero after reaching {@code Integer.MAX_VALUE}).
35		*
36		* <li> Like a CyclicBarrier, a Phaser may be repeatedly awaited.
37	<	* Method <tt>arriveAndAwaitAdvance</tt> has effect analogous to
38	<	* <tt>CyclicBarrier.await</tt>.  However, Phasers separate two
37	>	* Method {@code arriveAndAwaitAdvance} has effect analogous to
38	>	* {@code CyclicBarrier.await}.  However, Phasers separate two
39		* aspects of coordination, that may also be invoked independently:
40		*
41		* <ul>
42		*
43	<	*   <li> Arriving at a barrier. Methods <tt>arrive</tt> and
44	<	*       <tt>arriveAndDeregister</tt> do not block, but return
43	>	*   <li> Arriving at a barrier. Methods {@code arrive} and
44	>	*       {@code arriveAndDeregister} do not block, but return
45		*       the phase value current upon entry to the method.
46		*
47	<	*   <li> Awaiting others. Method <tt>awaitAdvance</tt> requires an
47	>	*   <li> Awaiting others. Method {@code awaitAdvance} requires an
48		*       argument indicating the entry phase, and returns when the
49		*       barrier advances to a new phase.
50		* </ul>
#	Line 51 | Line 52 | import java.lang.reflect.*;
52		*
53		* <li> Barrier actions, performed by the task triggering a phase
54		* advance while others may be waiting, are arranged by overriding
55	<	* method <tt>onAdvance</tt>, that also controls termination.
56	<	* Overriding this method may be used to similar but more flecible
55	>	* method {@code onAdvance}, that also controls termination.
56	>	* Overriding this method may be used to similar but more flexible
57		* effect as providing a barrier action to a CyclicBarrier.
58		*
59		* <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 <tt>onAdvance</tt> 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
65	<	* often convenient to override this method to cause termination when
66	<	* the current phase number reaches a threshold. Method
67	<	* <tt>forceTermination</tt> is also available to abruptly release
68	<	* waiting threads and allow them to terminate.
60	>	* actions immediately return without updating phaser state or waiting
61	>	* for advance, and indicating (via a negative phase value) that
62	>	* execution is complete.  Termination is triggered by executing the
63	>	* overridable {@code onAdvance} method that is invoked each time the
64	>	* barrier is about to be tripped. When a Phaser is controlling an
65	>	* action with a fixed number of iterations, it is often convenient to
66	>	* override this method to cause termination when the current phase
67	>	* number reaches a threshold. Method {@code forceTermination} is also
68	>	* available to abruptly release waiting threads and allow them to
69	>	* terminate.
70		*
71		* <li> Phasers may be tiered to reduce contention. Phasers with large
72		* numbers of parties that would otherwise experience heavy
#	Line 72 | Line 74 | import java.lang.reflect.*;
74		* This will typically greatly increase throughput even though it
75		* incurs somewhat greater per-operation overhead.
76		*
77	<	* <li> By default, <tt>awaitAdvance</tt> continues to wait even if
77	>	* <li> By default, {@code awaitAdvance} continues to wait even if
78		* the waiting thread is interrupted. And unlike the case in
79		* CyclicBarriers, exceptions encountered while tasks wait
80		* interruptibly or with timeout do not change the state of the
81		* barrier. If necessary, you can perform any associated recovery
82		* within handlers of those exceptions, often after invoking
83	<	* <tt>forceTermination</tt>.
83	>	* {@code forceTermination}.
84	>	*
85	>	* <li>Phasers ensure lack of starvation when used by ForkJoinTasks.
86		*
87		* </ul>
88		*
89		* <p><b>Sample usages:</b>
90		*
91	<	* <p>A Phaser may be used instead of a <tt>CountdownLatch</tt> to control
91	>	* <p>A Phaser may be used instead of a {@code CountDownLatch} to control
92		* a one-shot action serving a variable number of parties. The typical
93		* idiom is for the method setting this up to first register, then
94		* start the actions, then deregister, as in:
95		*
96	<	* <pre>
97	<	*  void runTasks(List&lt;Runnable&gt; list) {
98	<	*    final Phaser phaser = new Phaser(1); // "1" to register self
99	<	*    for (Runnable r : list) {
100	<	*      phaser.register();
101	<	*      new Thread() {
102	<	*        public void run() {
103	<	*          phaser.arriveAndAwaitAdvance(); // await all creation
104	<	*          r.run();
105	<	*          phaser.arriveAndDeregister();   // signal completion
106	<	*        }
107	<	*      }.start();
96	>	*  <pre> {@code
97	>	* void runTasks(List<Runnable> list) {
98	>	*   final Phaser phaser = new Phaser(1); // "1" to register self
99	>	*   for (Runnable r : list) {
100	>	*     phaser.register();
101	>	*     new Thread() {
102	>	*       public void run() {
103	>	*         phaser.arriveAndAwaitAdvance(); // await all creation
104	>	*         r.run();
105	>	*         phaser.arriveAndDeregister();   // signal completion
106	>	*       }
107	>	*     }.start();
108		*   }
109	+	*
110	+	*   doSomethingOnBehalfOfWorkers();
111		*   phaser.arrive(); // allow threads to start
112	<	*   int p = phaser.arriveAndDeregister(); // deregister self
112	>	*   int p = phaser.arriveAndDeregister(); // deregister self  ...
113	>	*   p = phaser.awaitAdvance(p); // ... and await arrival
114		*   otherActions(); // do other things while tasks execute
115	<	*   phaser.awaitAdvance(p); // wait for all tasks to arrive
116	<	* }
110	<	* </pre>
115	>	*   phaser.awaitAdvance(p); // await final completion
116	>	* }}</pre>
117		*
118		* <p>One way to cause a set of threads to repeatedly perform actions
119	<	* for a given number of iterations is to override <tt>onAdvance</tt>:
119	>	* for a given number of iterations is to override {@code onAdvance}:
120		*
121	<	* <pre>
122	<	*  void startTasks(List&lt;Runnable&gt; list, final int iterations) {
123	<	*    final Phaser phaser = new Phaser() {
124	<	*       public boolean onAdvance(int phase, int registeredParties) {
125	<	*         return phase &gt;= iterations || registeredParties == 0;
121	>	*  <pre> {@code
122	>	* void startTasks(List<Runnable> list, final int iterations) {
123	>	*   final Phaser phaser = new Phaser() {
124	>	*     public boolean onAdvance(int phase, int registeredParties) {
125	>	*       return phase >= iterations || registeredParties == 0;
126	>	*     }
127	>	*   };
128	>	*   phaser.register();
129	>	*   for (Runnable r : list) {
130	>	*     phaser.register();
131	>	*     new Thread() {
132	>	*       public void run() {
133	>	*         do {
134	>	*           r.run();
135	>	*           phaser.arriveAndAwaitAdvance();
136	>	*         } while(!phaser.isTerminated();
137		*       }
138	<	*    };
122	<	*    phaser.register();
123	<	*    for (Runnable r : list) {
124	<	*      phaser.register();
125	<	*      new Thread() {
126	<	*        public void run() {
127	<	*           do {
128	<	*             r.run();
129	<	*             phaser.arriveAndAwaitAdvance();
130	<	*           } while(!phaser.isTerminated();
131	<	*        }
132	<	*      }.start();
138	>	*     }.start();
139		*   }
140		*   phaser.arriveAndDeregister(); // deregister self, don't wait
141	<	* }
136	<	* </pre>
141	>	* }}</pre>
142		*
143		* <p> To create a set of tasks using a tree of Phasers,
144		* you could use code of the following form, assuming a
145		* Task class with a constructor accepting a Phaser that
146		* it registers for upon construction:
147	<	* <pre>
148	<	*  void build(Task[] actions, int lo, int hi, Phaser b) {
149	<	*    int step = (hi - lo) / TASKS_PER_PHASER;
150	<	*    if (step &gt; 1) {
151	<	*       int i = lo;
152	<	*       while (i &lt; hi) {
153	<	*         int r = Math.min(i + step, hi);
154	<	*         build(actions, i, r, new Phaser(b));
155	<	*         i = r;
156	<	*       }
157	<	*    }
158	<	*    else {
159	<	*      for (int i = lo; i &lt; hi; ++i)
160	<	*        actions[i] = new Task(b);
161	<	*        // assumes new Task(b) performs b.register()
162	<	*    }
163	<	*  }
164	<	*  // .. initially called, for n tasks via
160	<	*  build(new Task[n], 0, n, new Phaser());
161	<	* </pre>
147	>	*  <pre> {@code
148	>	* void build(Task[] actions, int lo, int hi, Phaser b) {
149	>	*   int step = (hi - lo) / TASKS_PER_PHASER;
150	>	*   if (step > 1) {
151	>	*     int i = lo;
152	>	*     while (i < hi) {
153	>	*       int r = Math.min(i + step, hi);
154	>	*       build(actions, i, r, new Phaser(b));
155	>	*       i = r;
156	>	*     }
157	>	*   } else {
158	>	*     for (int i = lo; i < hi; ++i)
159	>	*       actions[i] = new Task(b);
160	>	*       // assumes new Task(b) performs b.register()
161	>	*   }
162	>	* }
163	>	* // .. initially called, for n tasks via
164	>	* build(new Task[n], 0, n, new Phaser());}</pre>
165		*
166	<	* The best value of <tt>TASKS_PER_PHASER</tt> depends mainly on
166	>	* The best value of {@code TASKS_PER_PHASER} depends mainly on
167		* expected barrier synchronization rates. A value as low as four may
168		* be appropriate for extremely small per-barrier task bodies (thus
169		* high rates), or up to hundreds for extremely large ones.
#	Line 192 | Line 195 | public class Phaser {
195		* However, to efficiently maintain atomicity, these values are
196		* packed into a single (atomic) long. Termination uses the sign
197		* bit of 32 bit representation of phase, so phase is set to -1 on
198	<	* termination. Good performace relies on keeping state decoding
198	>	* termination. Good performance relies on keeping state decoding
199		* and encoding simple, and keeping race windows short.
200		*
201		* Note: there are some cheats in arrive() that rely on unarrived
202	<	* being lowest 16 bits.
202	>	* count being lowest 16 bits.
203		*/
204		private volatile long state;
205
206		private static final int ushortBits = 16;
207	<	private static final int ushortMask =  (1 << ushortBits) - 1;
208	<	private static final int phaseMask = 0x7fffffff;
207	>	private static final int ushortMask = 0xffff;
208	>	private static final int phaseMask  = 0x7fffffff;
209
210		private static int unarrivedOf(long s) {
211		return (int)(s & ushortMask);
212		}
213
214		private static int partiesOf(long s) {
215	<	return (int)(s & (ushortMask << 16)) >>> 16;
215	>	return ((int)s) >>> 16;
216		}
217
218		private static int phaseOf(long s) {
#	Line 221 | Line 224 | public class Phaser {
224		}
225
226		private static long stateFor(int phase, int parties, int unarrived) {
227	<	return (((long)phase) << 32) | ((parties << 16) | unarrived);
227	>	return ((((long)phase) << 32) | (((long)parties) << 16) |
228	>	(long)unarrived);
229		}
230
231		private static long trippedStateFor(int phase, int parties) {
232	<	return (((long)phase) << 32) | ((parties << 16) | parties);
232	>	long lp = (long)parties;
233	>	return (((long)phase) << 32) | (lp << 16) | lp;
234		}
235
236	<	private static IllegalStateException badBounds(int parties, int unarrived) {
237	<	return new IllegalStateException
238	<	("Attempt to set " + unarrived +
239	<	" unarrived of " + parties + " parties");
236	>	/**
237	>	* Returns message string for bad bounds exceptions.
238	>	*/
239	>	private static String badBounds(int parties, int unarrived) {
240	>	return ("Attempt to set " + unarrived +
241	>	" unarrived of " + parties + " parties");
242		}
243
244		/**
#	Line 248 | Line 255 | public class Phaser {
255		// Wait queues
256
257		/**
258	<	* Heads of Treiber stacks waiting for nonFJ threads. To eliminate
258	>	* Heads of Treiber stacks for waiting threads. To eliminate
259		* contention while releasing some threads while adding others, we
260		* use two of them, alternating across even and odd phases.
261		*/
#	Line 292 | Line 299 | public class Phaser {
299
300		/**
301		* Creates a new Phaser without any initially registered parties,
302	<	* initial phase number 0, and no parent.
302	>	* initial phase number 0, and no parent. Any thread using this
303	>	* Phaser will need to first register for it.
304		*/
305		public Phaser() {
306		this(null);
#	Line 301 | Line 309 | public class Phaser {
309		/**
310		* Creates a new Phaser with the given numbers of registered
311		* unarrived parties, initial phase number 0, and no parent.
312	<	* @param parties the number of parties required to trip barrier.
312	>	*
313	>	* @param parties the number of parties required to trip barrier
314		* @throws IllegalArgumentException if parties less than zero
315	<	* or greater than the maximum number of parties supported.
315	>	* or greater than the maximum number of parties supported
316		*/
317		public Phaser(int parties) {
318		this(null, parties);
#	Line 314 | Line 323 | public class Phaser {
323		* initially registered parties. If parent is non-null this phaser
324		* is registered with the parent and its initial phase number is
325		* the same as that of parent phaser.
326	<	* @param parent the parent phaser.
326	>	*
327	>	* @param parent the parent phaser
328		*/
329		public Phaser(Phaser parent) {
330		int phase = 0;
#	Line 330 | Line 340 | public class Phaser {
340
341		/**
342		* Creates a new Phaser with the given parent and numbers of
343	<	* registered unarrived parties. If parent is non-null this phaser
343	>	* registered unarrived parties. If parent is non-null, this phaser
344		* is registered with the parent and its initial phase number is
345		* the same as that of parent phaser.
346	<	* @param parent the parent phaser.
347	<	* @param parties the number of parties required to trip barrier.
346	>	*
347	>	* @param parent the parent phaser
348	>	* @param parties the number of parties required to trip barrier
349		* @throws IllegalArgumentException if parties less than zero
350	<	* or greater than the maximum number of parties supported.
350	>	* or greater than the maximum number of parties supported
351		*/
352		public Phaser(Phaser parent, int parties) {
353		if (parties < 0 || parties > ushortMask)
#	Line 354 | Line 365 | public class Phaser {
365
366		/**
367		* Adds a new unarrived party to this phaser.
368	+	*
369		* @return the current barrier phase number upon registration
370		* @throws IllegalStateException if attempting to register more
371	<	* than the maximum supported number of parties.
371	>	* than the maximum supported number of parties
372		*/
373		public int register() {
374		return doRegister(1);
#	Line 364 | Line 376 | public class Phaser {
376
377		/**
378		* Adds the given number of new unarrived parties to this phaser.
379	<	* @param parties the number of parties required to trip barrier.
379	>	*
380	>	* @param parties the number of parties required to trip barrier
381		* @return the current barrier phase number upon registration
382		* @throws IllegalStateException if attempting to register more
383	<	* than the maximum supported number of parties.
383	>	* than the maximum supported number of parties
384		*/
385		public int bulkRegister(int parties) {
386		if (parties < 0)
#	Line 390 | Line 403 | public class Phaser {
403		if (phase < 0)
404		break;
405		if (parties > ushortMask || unarrived > ushortMask)
406	<	throw badBounds(parties, unarrived);
406	>	throw new IllegalStateException(badBounds(parties, unarrived));
407		if (phase == phaseOf(root.state) &&
408		casState(s, stateFor(phase, parties, unarrived)))
409		break;
#	Line 403 | Line 416 | public class Phaser {
416		* in turn wait for others via {@link #awaitAdvance}).
417		*
418		* @return the barrier phase number upon entry to this method, or a
419	<	* negative value if terminated;
419	>	* negative value if terminated
420		* @throws IllegalStateException if not terminated and the number
421	<	* of unarrived parties would become negative.
421	>	* of unarrived parties would become negative
422		*/
423		public int arrive() {
424		int phase;
425		for (;;) {
426		long s = state;
427		phase = phaseOf(s);
428	+	if (phase < 0)
429	+	break;
430		int parties = partiesOf(s);
431		int unarrived = unarrivedOf(s) - 1;
432		if (unarrived > 0) {        // Not the last arrival
#	Line 437 | Line 452 | public class Phaser {
452		}
453		}
454		}
440	–	else if (phase < 0) // Don't throw exception if terminated
441	–	break;
455		else if (phase != phaseOf(root.state)) // or if unreconciled
456		reconcileState();
457		else
458	<	throw badBounds(parties, unarrived);
458	>	throw new IllegalStateException(badBounds(parties, unarrived));
459		}
460		return phase;
461		}
#	Line 455 | Line 468 | public class Phaser {
468		* zero parties, this phaser is also deregistered from its parent.
469		*
470		* @return the current barrier phase number upon entry to
471	<	* this method, or a negative value if terminated;
471	>	* this method, or a negative value if terminated
472		* @throws IllegalStateException if not terminated and the number
473	<	* of registered or unarrived parties would become negative.
473	>	* of registered or unarrived parties would become negative
474		*/
475		public int arriveAndDeregister() {
476		// similar code to arrive, but too different to merge
#	Line 466 | Line 479 | public class Phaser {
479		for (;;) {
480		long s = state;
481		phase = phaseOf(s);
482	+	if (phase < 0)
483	+	break;
484		int parties = partiesOf(s) - 1;
485		int unarrived = unarrivedOf(s) - 1;
486		if (parties >= 0) {
#	Line 491 | Line 506 | public class Phaser {
506		}
507		continue;
508		}
494	–	if (phase < 0)
495	–	break;
509		if (par != null && phase != phaseOf(root.state)) {
510		reconcileState();
511		continue;
512		}
513		}
514	<	throw badBounds(parties, unarrived);
514	>	throw new IllegalStateException(badBounds(parties, unarrived));
515		}
516		return phase;
517		}
518
519		/**
520		* Arrives at the barrier and awaits others. Equivalent in effect
521	<	* to <tt>awaitAdvance(arrive())</tt>.  If you instead need to
521	>	* to {@code awaitAdvance(arrive())}.  If you instead need to
522		* await with interruption of timeout, and/or deregister upon
523		* arrival, you can arrange them using analogous constructions.
524	+	*
525		* @return the phase on entry to this method
526		* @throws IllegalStateException if not terminated and the number
527	<	* of unarrived parties would become negative.
527	>	* of unarrived parties would become negative
528		*/
529		public int arriveAndAwaitAdvance() {
530		return awaitAdvance(arrive());
#	Line 520 | Line 534 | public class Phaser {
534		* Awaits the phase of the barrier to advance from the given
535		* value, or returns immediately if argument is negative or this
536		* barrier is terminated.
537	+	*
538		* @param phase the phase on entry to this method
539		* @return the phase on exit from this method
540		*/
#	Line 530 | Line 545 | public class Phaser {
545		int p = phaseOf(s);
546		if (p != phase)
547		return p;
548	<	if (unarrivedOf(s) == 0)
548	>	if (unarrivedOf(s) == 0 && parent != null)
549		parent.awaitAdvance(phase);
550		// Fall here even if parent waited, to reconcile and help release
551		return untimedWait(phase);
#	Line 538 | Line 553 | public class Phaser {
553
554		/**
555		* Awaits the phase of the barrier to advance from the given
556	<	* value, or returns immediately if argumet is negative or this
556	>	* value, or returns immediately if argument is negative or this
557		* barrier is terminated, or throws InterruptedException if
558		* interrupted while waiting.
559	+	*
560		* @param phase the phase on entry to this method
561		* @return the phase on exit from this method
562		* @throws InterruptedException if thread interrupted while waiting
563		*/
564	<	public int awaitAdvanceInterruptibly(int phase) throws InterruptedException {
564	>	public int awaitAdvanceInterruptibly(int phase)
565	>	throws InterruptedException {
566		if (phase < 0)
567		return phase;
568		long s = getReconciledState();
569		int p = phaseOf(s);
570		if (p != phase)
571		return p;
572	<	if (unarrivedOf(s) != 0)
572	>	if (unarrivedOf(s) == 0 && parent != null)
573		parent.awaitAdvanceInterruptibly(phase);
574		return interruptibleWait(phase);
575		}
#	Line 561 | Line 578 | public class Phaser {
578		* Awaits the phase of the barrier to advance from the given value
579		* or the given timeout elapses, or returns immediately if
580		* argument is negative or this barrier is terminated.
581	+	*
582		* @param phase the phase on entry to this method
583		* @return the phase on exit from this method
584		* @throws InterruptedException if thread interrupted while waiting
#	Line 574 | Line 592 | public class Phaser {
592		int p = phaseOf(s);
593		if (p != phase)
594		return p;
595	<	if (unarrivedOf(s) == 0)
595	>	if (unarrivedOf(s) == 0 && parent != null)
596		parent.awaitAdvanceInterruptibly(phase, timeout, unit);
597		return timedWait(phase, unit.toNanos(timeout));
598		}
#	Line 605 | Line 623 | public class Phaser {
623
624		/**
625		* Returns the current phase number. The maximum phase number is
626	<	* <tt>Integer.MAX_VALUE</tt>, after which it restarts at
626	>	* {@code Integer.MAX_VALUE}, after which it restarts at
627		* zero. Upon termination, the phase number is negative.
628	+	*
629		* @return the phase number, or a negative value if terminated
630		*/
631		public final int getPhase() {
#	Line 614 | Line 633 | public class Phaser {
633		}
634
635		/**
636	<	* Returns true if the current phase number equals the given phase.
636	>	* Returns {@code true} if the current phase number equals the given phase.
637	>	*
638		* @param phase the phase
639	<	* @return true if the current phase number equals the given phase.
639	>	* @return {@code true} if the current phase number equals the given phase
640		*/
641		public final boolean hasPhase(int phase) {
642		return phaseOf(getReconciledState()) == phase;
#	Line 624 | Line 644 | public class Phaser {
644
645		/**
646		* Returns the number of parties registered at this barrier.
647	+	*
648		* @return the number of parties
649		*/
650		public int getRegisteredParties() {
#	Line 633 | Line 654 | public class Phaser {
654		/**
655		* Returns the number of parties that have arrived at the current
656		* phase of this barrier.
657	+	*
658		* @return the number of arrived parties
659		*/
660		public int getArrivedParties() {
#	Line 642 | Line 664 | public class Phaser {
664		/**
665		* Returns the number of registered parties that have not yet
666		* arrived at the current phase of this barrier.
667	+	*
668		* @return the number of unarrived parties
669		*/
670		public int getUnarrivedParties() {
#	Line 650 | Line 673 | public class Phaser {
673
674		/**
675		* Returns the parent of this phaser, or null if none.
676	<	* @return the parent of this phaser, or null if none.
676	>	*
677	>	* @return the parent of this phaser, or null if none
678		*/
679		public Phaser getParent() {
680		return parent;
#	Line 659 | Line 683 | public class Phaser {
683		/**
684		* Returns the root ancestor of this phaser, which is the same as
685		* this phaser if it has no parent.
686	<	* @return the root ancestor of this phaser.
686	>	*
687	>	* @return the root ancestor of this phaser
688		*/
689		public Phaser getRoot() {
690		return root;
691		}
692
693		/**
694	<	* Returns true if this barrier has been terminated.
695	<	* @return true if this barrier has been terminated
694	>	* Returns {@code true} if this barrier has been terminated.
695	>	*
696	>	* @return {@code true} if this barrier has been terminated
697		*/
698		public boolean isTerminated() {
699		return getPhase() < 0;
#	Line 679 | Line 705 | public class Phaser {
705		* barrier is tripped (and thus all other waiting parties are
706		* dormant). If it returns true, then, rather than advance the
707		* phase number, this barrier will be set to a final termination
708	<	* state, and subsequent calls to <tt>isTerminated</tt> will
708	>	* state, and subsequent calls to {@code isTerminated} will
709		* return true.
710		*
711		* <p> The default version returns true when the number of
#	Line 690 | Line 716 | public class Phaser {
716		* <p> You may override this method to perform an action with side
717		* effects visible to participating tasks, but it is in general
718		* only sensible to do so in designs where all parties register
719	<	* before any arrive, and all <tt>awaitAdvance</tt> at each phase.
719	>	* before any arrive, and all {@code awaitAdvance} at each phase.
720		* Otherwise, you cannot ensure lack of interference. In
721		* particular, this method may be invoked more than once per
722		* transition if other parties successfully register while the
#	Line 699 | Line 725 | public class Phaser {
725		* method.
726		*
727		* @param phase the phase number on entering the barrier
728	<	* @param registeredParties the current number of registered
729	<	* parties.
704	<	* @return true if this barrier should terminate
728	>	* @param registeredParties the current number of registered parties
729	>	* @return {@code true} if this barrier should terminate
730		*/
731		protected boolean onAdvance(int phase, int registeredParties) {
732		return registeredParties <= 0;
#	Line 710 | Line 735 | public class Phaser {
735		/**
736		* Returns a string identifying this phaser, as well as its
737		* state.  The state, in brackets, includes the String {@code
738	<	* "phase ="} followed by the phase number, {@code "parties ="}
738	>	* "phase = "} followed by the phase number, {@code "parties = "}
739		* followed by the number of registered parties, and {@code
740	<	* "arrived ="} followed by the number of arrived parties
740	>	* "arrived = "} followed by the number of arrived parties.
741		*
742		* @return a string identifying this barrier, as well as its state
743		*/
744		public String toString() {
745		long s = getReconciledState();
746	<	return super.toString() + "[phase = " + phaseOf(s) + " parties = " + partiesOf(s) + " arrived = " + arrivedOf(s) + "]";
746	>	return super.toString() +
747	>	"[phase = " + phaseOf(s) +
748	>	" parties = " + partiesOf(s) +
749	>	" arrived = " + arrivedOf(s) + "]";
750		}
751
752		// methods for waiting
753
726	–	/** The number of CPUs, for spin control */
727	–	static final int NCPUS = Runtime.getRuntime().availableProcessors();
728	–
729	–	/**
730	–	* The number of times to spin before blocking in timed waits.
731	–	* The value is empirically derived.
732	–	*/
733	–	static final int maxTimedSpins = (NCPUS < 2)? 0 : 32;
734	–
735	–	/**
736	–	* The number of times to spin before blocking in untimed waits.
737	–	* This is greater than timed value because untimed waits spin
738	–	* faster since they don't need to check times on each spin.
739	–	*/
740	–	static final int maxUntimedSpins = maxTimedSpins * 32;
741	–
742	–	/**
743	–	* The number of nanoseconds for which it is faster to spin
744	–	* rather than to use timed park. A rough estimate suffices.
745	–	*/
746	–	static final long spinForTimeoutThreshold = 1000L;
747	–
754		/**
755	<	* Wait nodes for Treiber stack representing wait queue for non-FJ
750	<	* tasks.
755	>	* Wait nodes for Treiber stack representing wait queue
756		*/
757	<	static final class QNode {
758	<	QNode next;
757	>	static final class QNode implements ForkJoinPool.ManagedBlocker {
758	>	final Phaser phaser;
759	>	final int phase;
760	>	final long startTime;
761	>	final long nanos;
762	>	final boolean timed;
763	>	final boolean interruptible;
764	>	volatile boolean wasInterrupted = false;
765		volatile Thread thread; // nulled to cancel wait
766	<	QNode() {
766	>	QNode next;
767	>	QNode(Phaser phaser, int phase, boolean interruptible,
768	>	boolean timed, long startTime, long nanos) {
769	>	this.phaser = phaser;
770	>	this.phase = phase;
771	>	this.timed = timed;
772	>	this.interruptible = interruptible;
773	>	this.startTime = startTime;
774	>	this.nanos = nanos;
775		thread = Thread.currentThread();
776		}
777	+	public boolean isReleasable() {
778	+	return (thread == null ||
779	+	phaser.getPhase() != phase ||
780	+	(interruptible && wasInterrupted) ||
781	+	(timed && (nanos - (System.nanoTime() - startTime)) <= 0));
782	+	}
783	+	public boolean block() {
784	+	if (Thread.interrupted()) {
785	+	wasInterrupted = true;
786	+	if (interruptible)
787	+	return true;
788	+	}
789	+	if (!timed)
790	+	LockSupport.park(this);
791	+	else {
792	+	long waitTime = nanos - (System.nanoTime() - startTime);
793	+	if (waitTime <= 0)
794	+	return true;
795	+	LockSupport.parkNanos(this, waitTime);
796	+	}
797	+	return isReleasable();
798	+	}
799		void signal() {
800		Thread t = thread;
801		if (t != null) {
#	Line 762 | Line 803 | public class Phaser {
803		LockSupport.unpark(t);
804		}
805		}
806	+	boolean doWait() {
807	+	if (thread != null) {
808	+	try {
809	+	ForkJoinPool.managedBlock(this, false);
810	+	} catch (InterruptedException ie) {
811	+	}
812	+	}
813	+	return wasInterrupted;
814	+	}
815	+
816		}
817
818		/**
819	<	* Removes and signals waiting threads from wait queue
819	>	* Removes and signals waiting threads from wait queue.
820		*/
821		private void releaseWaiters(int phase) {
822		AtomicReference<QNode> head = queueFor(phase);
#	Line 777 | Line 828 | public class Phaser {
828		}
829
830		/**
831	+	* Tries to enqueue given node in the appropriate wait queue.
832	+	*
833	+	* @return true if successful
834	+	*/
835	+	private boolean tryEnqueue(QNode node) {
836	+	AtomicReference<QNode> head = queueFor(node.phase);
837	+	return head.compareAndSet(node.next = head.get(), node);
838	+	}
839	+
840	+	/**
841		* Enqueues node and waits unless aborted or signalled.
842	+	*
843	+	* @return current phase
844		*/
845		private int untimedWait(int phase) {
783	–	int spins = maxUntimedSpins;
846		QNode node = null;
785	–	boolean interrupted = false;
847		boolean queued = false;
848	+	boolean interrupted = false;
849		int p;
850		while ((p = getPhase()) == phase) {
851	<	interrupted = Thread.interrupted();
852	<	if (node != null) {
853	<	if (!queued) {
854	<	AtomicReference<QNode> head = queueFor(phase);
855	<	queued = head.compareAndSet(node.next = head.get(), node);
856	<	}
795	<	else if (node.thread != null)
796	<	LockSupport.park(this);
797	<	}
798	<	else if (spins <= 0)
799	<	node = new QNode();
851	>	if (Thread.interrupted())
852	>	interrupted = true;
853	>	else if (node == null)
854	>	node = new QNode(this, phase, false, false, 0, 0);
855	>	else if (!queued)
856	>	queued = tryEnqueue(node);
857		else
858	<	--spins;
858	>	interrupted = node.doWait();
859		}
860		if (node != null)
861		node.thread = null;
862	+	releaseWaiters(phase);
863		if (interrupted)
864		Thread.currentThread().interrupt();
807	–	releaseWaiters(phase);
865		return p;
866		}
867
868		/**
869	<	* Messier interruptible version
869	>	* Interruptible version
870	>	* @return current phase
871		*/
872		private int interruptibleWait(int phase) throws InterruptedException {
815	–	int spins = maxUntimedSpins;
873		QNode node = null;
874		boolean queued = false;
875		boolean interrupted = false;
876		int p;
877	<	while ((p = getPhase()) == phase) {
878	<	if (interrupted = Thread.interrupted())
879	<	break;
880	<	if (node != null) {
881	<	if (!queued) {
882	<	AtomicReference<QNode> head = queueFor(phase);
883	<	queued = head.compareAndSet(node.next = head.get(), node);
827	<	}
828	<	else if (node.thread != null)
829	<	LockSupport.park(this);
830	<	}
831	<	else if (spins <= 0)
832	<	node = new QNode();
877	>	while ((p = getPhase()) == phase && !interrupted) {
878	>	if (Thread.interrupted())
879	>	interrupted = true;
880	>	else if (node == null)
881	>	node = new QNode(this, phase, true, false, 0, 0);
882	>	else if (!queued)
883	>	queued = tryEnqueue(node);
884		else
885	<	--spins;
885	>	interrupted = node.doWait();
886		}
887		if (node != null)
888		node.thread = null;
889	+	if (p != phase || (p = getPhase()) != phase)
890	+	releaseWaiters(phase);
891		if (interrupted)
892		throw new InterruptedException();
840	–	releaseWaiters(phase);
893		return p;
894		}
895
896		/**
897	<	* Even messier timeout version.
897	>	* Timeout version.
898	>	* @return current phase
899		*/
900		private int timedWait(int phase, long nanos)
901		throws InterruptedException, TimeoutException {
902	+	long startTime = System.nanoTime();
903	+	QNode node = null;
904	+	boolean queued = false;
905	+	boolean interrupted = false;
906		int p;
907	<	if ((p = getPhase()) == phase) {
908	<	long lastTime = System.nanoTime();
909	<	int spins = maxTimedSpins;
910	<	QNode node = null;
911	<	boolean queued = false;
912	<	boolean interrupted = false;
913	<	while ((p = getPhase()) == phase) {
914	<	if (interrupted = Thread.interrupted())
915	<	break;
916	<	long now = System.nanoTime();
917	<	if ((nanos -= now - lastTime) <= 0)
861	<	break;
862	<	lastTime = now;
863	<	if (node != null) {
864	<	if (!queued) {
865	<	AtomicReference<QNode> head = queueFor(phase);
866	<	queued = head.compareAndSet(node.next = head.get(), node);
867	<	}
868	<	else if (node.thread != null &&
869	<	nanos > spinForTimeoutThreshold) {
870	<	LockSupport.parkNanos(this, nanos);
871	<	}
872	<	}
873	<	else if (spins <= 0)
874	<	node = new QNode();
875	<	else
876	<	--spins;
877	<	}
878	<	if (node != null)
879	<	node.thread = null;
880	<	if (interrupted)
881	<	throw new InterruptedException();
882	<	if (p == phase && (p = getPhase()) == phase)
883	<	throw new TimeoutException();
907	>	while ((p = getPhase()) == phase && !interrupted) {
908	>	if (Thread.interrupted())
909	>	interrupted = true;
910	>	else if (nanos - (System.nanoTime() - startTime) <= 0)
911	>	break;
912	>	else if (node == null)
913	>	node = new QNode(this, phase, true, true, startTime, nanos);
914	>	else if (!queued)
915	>	queued = tryEnqueue(node);
916	>	else
917	>	interrupted = node.doWait();
918		}
919	<	releaseWaiters(phase);
919	>	if (node != null)
920	>	node.thread = null;
921	>	if (p != phase || (p = getPhase()) != phase)
922	>	releaseWaiters(phase);
923	>	if (interrupted)
924	>	throw new InterruptedException();
925	>	if (p == phase)
926	>	throw new TimeoutException();
927		return p;
928		}
929
930		// Temporary Unsafe mechanics for preliminary release
931	+	private static Unsafe getUnsafe() throws Throwable {
932	+	try {
933	+	return Unsafe.getUnsafe();
934	+	} catch (SecurityException se) {
935	+	try {
936	+	return java.security.AccessController.doPrivileged
937	+	(new java.security.PrivilegedExceptionAction<Unsafe>() {
938	+	public Unsafe run() throws Exception {
939	+	return getUnsafePrivileged();
940	+	}});
941	+	} catch (java.security.PrivilegedActionException e) {
942	+	throw e.getCause();
943	+	}
944	+	}
945	+	}
946
947	<	static final Unsafe _unsafe;
947	>	private static Unsafe getUnsafePrivileged()
948	>	throws NoSuchFieldException, IllegalAccessException {
949	>	Field f = Unsafe.class.getDeclaredField("theUnsafe");
950	>	f.setAccessible(true);
951	>	return (Unsafe) f.get(null);
952	>	}
953	>
954	>	private static long fieldOffset(String fieldName)
955	>	throws NoSuchFieldException {
956	>	return UNSAFE.objectFieldOffset
957	>	(Phaser.class.getDeclaredField(fieldName));
958	>	}
959	>
960	>	static final Unsafe UNSAFE;
961		static final long stateOffset;
962
963		static {
964		try {
965	<	if (Phaser.class.getClassLoader() != null) {
966	<	Field f = Unsafe.class.getDeclaredField("theUnsafe");
967	<	f.setAccessible(true);
899	<	_unsafe = (Unsafe)f.get(null);
900	<	}
901	<	else
902	<	_unsafe = Unsafe.getUnsafe();
903	<	stateOffset = _unsafe.objectFieldOffset
904	<	(Phaser.class.getDeclaredField("state"));
905	<	} catch (Exception e) {
965	>	UNSAFE = getUnsafe();
966	>	stateOffset = fieldOffset("state");
967	>	} catch (Throwable e) {
968		throw new RuntimeException("Could not initialize intrinsics", e);
969		}
970		}
971
972		final boolean casState(long cmp, long val) {
973	<	return _unsafe.compareAndSwapLong(this, stateOffset, cmp, val);
973	>	return UNSAFE.compareAndSwapLong(this, stateOffset, cmp, val);
974		}
975		}

Diff Legend

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