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 1.6 by dl, Tue Oct 28 23:03:24 2008 UTC vs.
Revision 1.29 by jsr166, Wed Aug 12 04:02:45 2009 UTC

#	Line 5 | Line 5
5		*/
6
7		package jsr166y;
8	+
9		import java.util.concurrent.*;
10	<	import java.util.concurrent.atomic.*;
10	>
11	>	import java.util.concurrent.atomic.AtomicReference;
12		import java.util.concurrent.locks.LockSupport;
11	–	import sun.misc.Unsafe;
12	–	import java.lang.reflect.*;
13
14		/**
15		* A reusable synchronization barrier, similar in functionality to a
16	<	* {@link java.util.concurrent.CyclicBarrier} and {@link
17	<	* java.util.concurrent.CountDownLatch} but supporting more flexible
18	<	* usage.
16	>	* {@link java.util.concurrent.CyclicBarrier CyclicBarrier} and
17	>	* {@link java.util.concurrent.CountDownLatch CountDownLatch}
18	>	* but supporting more flexible usage.
19		*
20		* <ul>
21		*
#	Line 25 | Line 25 | import java.lang.reflect.*;
25		* basic synchronization constructs, registration and deregistration
26		* affect only internal counts; they do not establish any further
27		* internal bookkeeping, so tasks cannot query whether they are
28	<	* registered. (However, you can introduce such bookkeeping in by
28	>	* registered. (However, you can introduce such bookkeeping by
29		* subclassing this class.)
30		*
31		* <li> Each generation has an associated phase value, starting at
32		* zero, and advancing when all parties reach the barrier (wrapping
33	<	* around to zero after reaching <tt>Integer.MAX_VALUE</tt>).
33	>	* around to zero after reaching {@code Integer.MAX_VALUE}).
34		*
35	<	* <li> Like a CyclicBarrier, a Phaser may be repeatedly awaited.
36	<	* Method <tt>arriveAndAwaitAdvance</tt> has effect analogous to
37	<	* <tt>CyclicBarrier.await</tt>.  However, Phasers separate two
38	<	* aspects of coordination, that may also be invoked independently:
35	>	* <li> Like a {@code CyclicBarrier}, a phaser may be repeatedly
36	>	* awaited.  Method {@link #arriveAndAwaitAdvance} has effect
37	>	* analogous to {@link java.util.concurrent.CyclicBarrier#await
38	>	* CyclicBarrier.await}.  However, phasers separate two aspects of
39	>	* coordination, which 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 {@link #arrive} and
44	>	*       {@link #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 {@link #awaitAdvance} requires an
48		*       argument indicating the entry phase, and returns when the
49		*       barrier advances to a new phase.
50		* </ul>
51		*
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 flexible
57	<	* effect as providing a barrier action to a CyclicBarrier.
54	>	* advance, are arranged by overriding method {@link #onAdvance(int,
55	>	* int)}, which also controls termination. Overriding this method is
56	>	* similar to, but more flexible than, providing a barrier action to a
57	>	* {@code 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
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 when an invocation
63	>	* of {@code onAdvance} returns {@code true}.  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.
66	>	* the current phase number reaches a threshold. Method {@link
67	>	* #forceTermination} is also available to abruptly release waiting
68	>	* threads and allow them to terminate.
69		*
70		* <li> Phasers may be tiered to reduce contention. Phasers with large
71		* numbers of parties that would otherwise experience heavy
#	Line 72 | Line 73 | import java.lang.reflect.*;
73		* This will typically greatly increase throughput even though it
74		* incurs somewhat greater per-operation overhead.
75		*
76	<	* <li> By default, <tt>awaitAdvance</tt> continues to wait even if
76	>	* <li> By default, {@code awaitAdvance} continues to wait even if
77		* the waiting thread is interrupted. And unlike the case in
78	<	* CyclicBarriers, exceptions encountered while tasks wait
78	>	* {@code CyclicBarrier}, exceptions encountered while tasks wait
79		* interruptibly or with timeout do not change the state of the
80		* barrier. If necessary, you can perform any associated recovery
81		* within handlers of those exceptions, often after invoking
82	<	* <tt>forceTermination</tt>.
82	>	* {@code forceTermination}.
83	>	*
84	>	* <li>Phasers may be used to coordinate tasks executing in a {@link
85	>	* ForkJoinPool}, which will ensure sufficient parallelism to execute
86	>	* tasks when others are blocked waiting for a phase to advance.
87		*
88		* </ul>
89		*
90		* <p><b>Sample usages:</b>
91		*
92	<	* <p>A Phaser may be used instead of a <tt>CountdownLatch</tt> to control
93	<	* a one-shot action serving a variable number of parties. The typical
94	<	* idiom is for the method setting this up to first register, then
95	<	* start the actions, then deregister, as in:
96	<	*
97	<	* <pre>
98	<	*  void runTasks(List&lt;Runnable&gt; list) {
99	<	*    final Phaser phaser = new Phaser(1); // "1" to register self
100	<	*    for (Runnable r : list) {
101	<	*      phaser.register();
102	<	*      new Thread() {
103	<	*        public void run() {
104	<	*          phaser.arriveAndAwaitAdvance(); // await all creation
105	<	*          r.run();
106	<	*          phaser.arriveAndDeregister();   // signal completion
107	<	*        }
108	<	*      }.start();
92	>	* <p>A {@code Phaser} may be used instead of a {@code CountDownLatch}
93	>	* to control a one-shot action serving a variable number of
94	>	* parties. The typical idiom is for the method setting this up to
95	>	* first register, then start the actions, then deregister, as in:
96	>	*
97	>	*  <pre> {@code
98	>	* void runTasks(List<Runnable> list) {
99	>	*   final Phaser phaser = new Phaser(1); // "1" to register self
100	>	*   // create and start threads
101	>	*   for (Runnable r : list) {
102	>	*     phaser.register();
103	>	*     new Thread() {
104	>	*       public void run() {
105	>	*         phaser.arriveAndAwaitAdvance(); // await all creation
106	>	*         r.run();
107	>	*       }
108	>	*     }.start();
109		*   }
110		*
111	<	*   doSomethingOnBehalfOfWorkers();
112	<	*   phaser.arrive(); // allow threads to start
113	<	*   int p = phaser.arriveAndDeregister(); // deregister self  ...
109	<	*   p = phaser.awaitAdvance(p); // ... and await arrival
110	<	*   otherActions(); // do other things while tasks execute
111	<	*   phaser.awaitAdvance(p); // awit final completion
112	<	* }
113	<	* </pre>
111	>	*   // allow threads to start and deregister self
112	>	*   phaser.arriveAndDeregister();
113	>	* }}</pre>
114		*
115		* <p>One way to cause a set of threads to repeatedly perform actions
116	<	* for a given number of iterations is to override <tt>onAdvance</tt>:
116	>	* for a given number of iterations is to override {@code onAdvance}:
117		*
118	<	* <pre>
119	<	*  void startTasks(List&lt;Runnable&gt; list, final int iterations) {
120	<	*    final Phaser phaser = new Phaser() {
121	<	*       public boolean onAdvance(int phase, int registeredParties) {
122	<	*         return phase &gt;= iterations || registeredParties == 0;
118	>	*  <pre> {@code
119	>	* void startTasks(List<Runnable> list, final int iterations) {
120	>	*   final Phaser phaser = new Phaser() {
121	>	*     public boolean onAdvance(int phase, int registeredParties) {
122	>	*       return phase >= iterations || registeredParties == 0;
123	>	*     }
124	>	*   };
125	>	*   phaser.register();
126	>	*   for (Runnable r : list) {
127	>	*     phaser.register();
128	>	*     new Thread() {
129	>	*       public void run() {
130	>	*         do {
131	>	*           r.run();
132	>	*           phaser.arriveAndAwaitAdvance();
133	>	*         } while(!phaser.isTerminated();
134		*       }
135	<	*    };
125	<	*    phaser.register();
126	<	*    for (Runnable r : list) {
127	<	*      phaser.register();
128	<	*      new Thread() {
129	<	*        public void run() {
130	<	*           do {
131	<	*             r.run();
132	<	*             phaser.arriveAndAwaitAdvance();
133	<	*           } while(!phaser.isTerminated();
134	<	*        }
135	<	*      }.start();
135	>	*     }.start();
136		*   }
137		*   phaser.arriveAndDeregister(); // deregister self, don't wait
138	<	* }
139	<	* </pre>
138	>	* }}</pre>
139		*
140	<	* <p> To create a set of tasks using a tree of Phasers,
140	>	* <p>To create a set of tasks using a tree of phasers,
141		* you could use code of the following form, assuming a
142	<	* Task class with a constructor accepting a Phaser that
142	>	* Task class with a constructor accepting a phaser that
143		* it registers for upon construction:
144	<	* <pre>
145	<	*  void build(Task[] actions, int lo, int hi, Phaser b) {
146	<	*    int step = (hi - lo) / TASKS_PER_PHASER;
147	<	*    if (step &gt; 1) {
148	<	*       int i = lo;
149	<	*       while (i &lt; hi) {
150	<	*         int r = Math.min(i + step, hi);
151	<	*         build(actions, i, r, new Phaser(b));
152	<	*         i = r;
153	<	*       }
154	<	*    }
155	<	*    else {
156	<	*      for (int i = lo; i &lt; hi; ++i)
157	<	*        actions[i] = new Task(b);
158	<	*        // assumes new Task(b) performs b.register()
159	<	*    }
160	<	*  }
161	<	*  // .. initially called, for n tasks via
163	<	*  build(new Task[n], 0, n, new Phaser());
164	<	* </pre>
144	>	*  <pre> {@code
145	>	* void build(Task[] actions, int lo, int hi, Phaser b) {
146	>	*   int step = (hi - lo) / TASKS_PER_PHASER;
147	>	*   if (step > 1) {
148	>	*     int i = lo;
149	>	*     while (i < hi) {
150	>	*       int r = Math.min(i + step, hi);
151	>	*       build(actions, i, r, new Phaser(b));
152	>	*       i = r;
153	>	*     }
154	>	*   } else {
155	>	*     for (int i = lo; i < hi; ++i)
156	>	*       actions[i] = new Task(b);
157	>	*       // assumes new Task(b) performs b.register()
158	>	*   }
159	>	* }
160	>	* // .. initially called, for n tasks via
161	>	* build(new Task[n], 0, n, new Phaser());}</pre>
162		*
163	<	* The best value of <tt>TASKS_PER_PHASER</tt> depends mainly on
163	>	* The best value of {@code TASKS_PER_PHASER} depends mainly on
164		* expected barrier synchronization rates. A value as low as four may
165		* be appropriate for extremely small per-barrier task bodies (thus
166		* high rates), or up to hundreds for extremely large ones.
#	Line 175 | Line 172 | import java.lang.reflect.*;
172		* parties result in IllegalStateExceptions. However, you can and
173		* should create tiered phasers to accommodate arbitrarily large sets
174		* of participants.
175	+	*
176	+	* @since 1.7
177	+	* @author Doug Lea
178		*/
179		public class Phaser {
180		/*
#	Line 195 | 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);
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) {
219	<	return (int)(s >>> 32);
219	>	return (int) (s >>> 32);
220		}
221
222		private static int arrivedOf(long s) {
#	Line 224 | 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 243 | Line 247 | public class Phaser {
247		private final Phaser parent;
248
249		/**
250	<	* The root of Phaser tree. Equals this if not in a tree.  Used to
250	>	* The root of phaser tree. Equals this if not in a tree.  Used to
251		* support faster state push-down.
252		*/
253		private final Phaser root;
#	Line 251 | 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 259 | Line 263 | public class Phaser {
263		private final AtomicReference<QNode> oddQ  = new AtomicReference<QNode>();
264
265		private AtomicReference<QNode> queueFor(int phase) {
266	<	return (phase & 1) == 0? evenQ : oddQ;
266	>	return ((phase & 1) == 0) ? evenQ : oddQ;
267		}
268
269		/**
#	Line 267 | Line 271 | public class Phaser {
271		* root if necessary.
272		*/
273		private long getReconciledState() {
274	<	return parent == null? state : reconcileState();
274	>	return (parent == null) ? state : reconcileState();
275		}
276
277		/**
#	Line 294 | Line 298 | public class Phaser {
298		}
299
300		/**
301	<	* Creates a new Phaser without any initially registered parties,
302	<	* initial phase number 0, and no parent.
301	>	* Creates a new phaser without any initially registered parties,
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);
307		}
308
309		/**
310	<	* Creates a new Phaser with the given numbers of registered
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);
319		}
320
321		/**
322	<	* Creates a new Phaser with the given parent, without any
322	>	* Creates a new phaser with the given parent, without any
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 332 | Line 339 | public class Phaser {
339		}
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
342	>	* Creates a new phaser with the given parent and numbers of
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 357 | 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 367 | 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 393 | 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 406 | 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 426 | Line 438 | public class Phaser {
438		if (par == null) {      // directly trip
439		if (casState
440		(s,
441	<	trippedStateFor(onAdvance(phase, parties)? -1 :
441	>	trippedStateFor(onAdvance(phase, parties) ? -1 :
442		((phase + 1) & phaseMask), parties))) {
443		releaseWaiters(phase);
444		break;
#	Line 440 | Line 452 | public class Phaser {
452		}
453		}
454		}
443	–	else if (phase < 0) // Don't throw exception if terminated
444	–	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		}
462
463		/**
464	<	* Arrives at the barrier, and deregisters from it, without
465	<	* waiting for others. Deregistration reduces number of parties
464	>	* Arrives at the barrier and deregisters from it without waiting
465	>	* for others. Deregistration reduces the number of parties
466		* required to trip the barrier in future phases.  If this phaser
467		* has a parent, and deregistration causes this phaser to have
468	<	* zero parties, this phaser is also deregistered from its parent.
468	>	* zero parties, this phaser also arrives at and is deregistered
469	>	* from its parent.
470		*
471		* @return the current barrier phase number upon entry to
472	<	* this method, or a negative value if terminated;
472	>	* this method, or a negative value if terminated
473		* @throws IllegalStateException if not terminated and the number
474	<	* of registered or unarrived parties would become negative.
474	>	* of registered or unarrived parties would become negative
475		*/
476		public int arriveAndDeregister() {
477		// similar code to arrive, but too different to merge
#	Line 469 | Line 480 | public class Phaser {
480		for (;;) {
481		long s = state;
482		phase = phaseOf(s);
483	+	if (phase < 0)
484	+	break;
485		int parties = partiesOf(s) - 1;
486		int unarrived = unarrivedOf(s) - 1;
487		if (parties >= 0) {
#	Line 487 | Line 500 | public class Phaser {
500		if (unarrived == 0) {
501		if (casState
502		(s,
503	<	trippedStateFor(onAdvance(phase, parties)? -1 :
503	>	trippedStateFor(onAdvance(phase, parties) ? -1 :
504		((phase + 1) & phaseMask), parties))) {
505		releaseWaiters(phase);
506		break;
507		}
508		continue;
509		}
497	–	if (phase < 0)
498	–	break;
510		if (par != null && phase != phaseOf(root.state)) {
511		reconcileState();
512		continue;
513		}
514		}
515	<	throw badBounds(parties, unarrived);
515	>	throw new IllegalStateException(badBounds(parties, unarrived));
516		}
517		return phase;
518		}
519
520		/**
521		* Arrives at the barrier and awaits others. Equivalent in effect
522	<	* to <tt>awaitAdvance(arrive())</tt>.  If you instead need to
523	<	* await with interruption of timeout, and/or deregister upon
524	<	* arrival, you can arrange them using analogous constructions.
522	>	* to {@code awaitAdvance(arrive())}.  If you need to await with
523	>	* interruption or timeout, you can arrange this with an analogous
524	>	* construction using one of the other forms of the awaitAdvance
525	>	* method.  If instead you need to deregister upon arrival use
526	>	* {@code arriveAndDeregister}.
527	>	*
528		* @return the phase on entry to this method
529		* @throws IllegalStateException if not terminated and the number
530	<	* of unarrived parties would become negative.
530	>	* of unarrived parties would become negative
531		*/
532		public int arriveAndAwaitAdvance() {
533		return awaitAdvance(arrive());
534		}
535
536		/**
537	<	* Awaits the phase of the barrier to advance from the given
538	<	* value, or returns immediately if argument is negative or this
539	<	* barrier is terminated.
537	>	* Awaits the phase of the barrier to advance from the given phase
538	>	* value, or returns immediately if the current phase of the barrier
539	>	* is not equal to the given phase value or this barrier is
540	>	* terminated.
541	>	*
542		* @param phase the phase on entry to this method
543		* @return the phase on exit from this method
544		*/
#	Line 533 | Line 549 | public class Phaser {
549		int p = phaseOf(s);
550		if (p != phase)
551		return p;
552	<	if (unarrivedOf(s) == 0)
552	>	if (unarrivedOf(s) == 0 && parent != null)
553		parent.awaitAdvance(phase);
554		// Fall here even if parent waited, to reconcile and help release
555		return untimedWait(phase);
#	Line 541 | Line 557 | public class Phaser {
557
558		/**
559		* Awaits the phase of the barrier to advance from the given
560	<	* value, or returns immediately if argumet is negative or this
560	>	* value, or returns immediately if argument is negative or this
561		* barrier is terminated, or throws InterruptedException if
562		* interrupted while waiting.
563	+	*
564		* @param phase the phase on entry to this method
565		* @return the phase on exit from this method
566		* @throws InterruptedException if thread interrupted while waiting
567		*/
568	<	public int awaitAdvanceInterruptibly(int phase) throws InterruptedException {
568	>	public int awaitAdvanceInterruptibly(int phase)
569	>	throws InterruptedException {
570		if (phase < 0)
571		return phase;
572		long s = getReconciledState();
573		int p = phaseOf(s);
574		if (p != phase)
575		return p;
576	<	if (unarrivedOf(s) != 0)
576	>	if (unarrivedOf(s) == 0 && parent != null)
577		parent.awaitAdvanceInterruptibly(phase);
578		return interruptibleWait(phase);
579		}
#	Line 564 | Line 582 | public class Phaser {
582		* Awaits the phase of the barrier to advance from the given value
583		* or the given timeout elapses, or returns immediately if
584		* argument is negative or this barrier is terminated.
585	+	*
586		* @param phase the phase on entry to this method
587		* @return the phase on exit from this method
588		* @throws InterruptedException if thread interrupted while waiting
589		* @throws TimeoutException if timed out while waiting
590		*/
591	<	public int awaitAdvanceInterruptibly(int phase, long timeout, TimeUnit unit)
591	>	public int awaitAdvanceInterruptibly(int phase,
592	>	long timeout, TimeUnit unit)
593		throws InterruptedException, TimeoutException {
594		if (phase < 0)
595		return phase;
#	Line 577 | Line 597 | public class Phaser {
597		int p = phaseOf(s);
598		if (p != phase)
599		return p;
600	<	if (unarrivedOf(s) == 0)
600	>	if (unarrivedOf(s) == 0 && parent != null)
601		parent.awaitAdvanceInterruptibly(phase, timeout, unit);
602		return timedWait(phase, unit.toNanos(timeout));
603		}
#	Line 608 | Line 628 | public class Phaser {
628
629		/**
630		* Returns the current phase number. The maximum phase number is
631	<	* <tt>Integer.MAX_VALUE</tt>, after which it restarts at
631	>	* {@code Integer.MAX_VALUE}, after which it restarts at
632		* zero. Upon termination, the phase number is negative.
633	+	*
634		* @return the phase number, or a negative value if terminated
635		*/
636		public final int getPhase() {
#	Line 617 | Line 638 | public class Phaser {
638		}
639
640		/**
620	–	* Returns true if the current phase number equals the given phase.
621	–	* @param phase the phase
622	–	* @return true if the current phase number equals the given phase.
623	–	*/
624	–	public final boolean hasPhase(int phase) {
625	–	return phaseOf(getReconciledState()) == phase;
626	–	}
627	–
628	–	/**
641		* Returns the number of parties registered at this barrier.
642	+	*
643		* @return the number of parties
644		*/
645		public int getRegisteredParties() {
#	Line 636 | Line 649 | public class Phaser {
649		/**
650		* Returns the number of parties that have arrived at the current
651		* phase of this barrier.
652	+	*
653		* @return the number of arrived parties
654		*/
655		public int getArrivedParties() {
#	Line 645 | Line 659 | public class Phaser {
659		/**
660		* Returns the number of registered parties that have not yet
661		* arrived at the current phase of this barrier.
662	+	*
663		* @return the number of unarrived parties
664		*/
665		public int getUnarrivedParties() {
#	Line 652 | Line 667 | public class Phaser {
667		}
668
669		/**
670	<	* Returns the parent of this phaser, or null if none.
671	<	* @return the parent of this phaser, or null if none.
670	>	* Returns the parent of this phaser, or {@code null} if none.
671	>	*
672	>	* @return the parent of this phaser, or {@code null} if none
673		*/
674		public Phaser getParent() {
675		return parent;
#	Line 662 | Line 678 | public class Phaser {
678		/**
679		* Returns the root ancestor of this phaser, which is the same as
680		* this phaser if it has no parent.
681	<	* @return the root ancestor of this phaser.
681	>	*
682	>	* @return the root ancestor of this phaser
683		*/
684		public Phaser getRoot() {
685		return root;
686		}
687
688		/**
689	<	* Returns true if this barrier has been terminated.
690	<	* @return true if this barrier has been terminated
689	>	* Returns {@code true} if this barrier has been terminated.
690	>	*
691	>	* @return {@code true} if this barrier has been terminated
692		*/
693		public boolean isTerminated() {
694		return getPhase() < 0;
#	Line 680 | Line 698 | public class Phaser {
698		* Overridable method to perform an action upon phase advance, and
699		* to control termination. This method is invoked whenever the
700		* barrier is tripped (and thus all other waiting parties are
701	<	* dormant). If it returns true, then, rather than advance the
702	<	* phase number, this barrier will be set to a final termination
703	<	* state, and subsequent calls to <tt>isTerminated</tt> will
704	<	* return true.
701	>	* dormant). If it returns {@code true}, then, rather than advance
702	>	* the phase number, this barrier will be set to a final
703	>	* termination state, and subsequent calls to {@link #isTerminated}
704	>	* will return true.
705		*
706	<	* <p> The default version returns true when the number of
706	>	* <p>The default version returns {@code true} when the number of
707		* registered parties is zero. Normally, overrides that arrange
708		* termination for other reasons should also preserve this
709		* property.
710		*
711	<	* <p> You may override this method to perform an action with side
711	>	* <p>You may override this method to perform an action with side
712		* effects visible to participating tasks, but it is in general
713		* only sensible to do so in designs where all parties register
714	<	* before any arrive, and all <tt>awaitAdvance</tt> at each phase.
715	<	* Otherwise, you cannot ensure lack of interference. In
716	<	* particular, this method may be invoked more than once per
699	<	* transition if other parties successfully register while the
700	<	* invocation of this method is in progress, thus postponing the
701	<	* transition until those parties also arrive, re-triggering this
702	<	* method.
714	>	* before any arrive, and all {@link #awaitAdvance} at each phase.
715	>	* Otherwise, you cannot ensure lack of interference from other
716	>	* parties during the invocation of this method.
717		*
718		* @param phase the phase number on entering the barrier
719	<	* @param registeredParties the current number of registered
720	<	* parties.
707	<	* @return true if this barrier should terminate
719	>	* @param registeredParties the current number of registered parties
720	>	* @return {@code true} if this barrier should terminate
721		*/
722		protected boolean onAdvance(int phase, int registeredParties) {
723		return registeredParties <= 0;
#	Line 713 | Line 726 | public class Phaser {
726		/**
727		* Returns a string identifying this phaser, as well as its
728		* state.  The state, in brackets, includes the String {@code
729	<	* "phase ="} followed by the phase number, {@code "parties ="}
729	>	* "phase = "} followed by the phase number, {@code "parties = "}
730		* followed by the number of registered parties, and {@code
731	<	* "arrived ="} followed by the number of arrived parties
731	>	* "arrived = "} followed by the number of arrived parties.
732		*
733		* @return a string identifying this barrier, as well as its state
734		*/
735		public String toString() {
736		long s = getReconciledState();
737	<	return super.toString() + "[phase = " + phaseOf(s) + " parties = " + partiesOf(s) + " arrived = " + arrivedOf(s) + "]";
737	>	return super.toString() +
738	>	"[phase = " + phaseOf(s) +
739	>	" parties = " + partiesOf(s) +
740	>	" arrived = " + arrivedOf(s) + "]";
741		}
742
743		// methods for waiting
744
729	–	/** The number of CPUs, for spin control */
730	–	static final int NCPUS = Runtime.getRuntime().availableProcessors();
731	–
732	–	/**
733	–	* The number of times to spin before blocking in timed waits.
734	–	* The value is empirically derived.
735	–	*/
736	–	static final int maxTimedSpins = (NCPUS < 2)? 0 : 32;
737	–
738	–	/**
739	–	* The number of times to spin before blocking in untimed waits.
740	–	* This is greater than timed value because untimed waits spin
741	–	* faster since they don't need to check times on each spin.
742	–	*/
743	–	static final int maxUntimedSpins = maxTimedSpins * 32;
744	–
745	–	/**
746	–	* The number of nanoseconds for which it is faster to spin
747	–	* rather than to use timed park. A rough estimate suffices.
748	–	*/
749	–	static final long spinForTimeoutThreshold = 1000L;
750	–
745		/**
746	<	* Wait nodes for Treiber stack representing wait queue for non-FJ
753	<	* tasks.
746	>	* Wait nodes for Treiber stack representing wait queue
747		*/
748	<	static final class QNode {
749	<	QNode next;
748	>	static final class QNode implements ForkJoinPool.ManagedBlocker {
749	>	final Phaser phaser;
750	>	final int phase;
751	>	final long startTime;
752	>	final long nanos;
753	>	final boolean timed;
754	>	final boolean interruptible;
755	>	volatile boolean wasInterrupted = false;
756		volatile Thread thread; // nulled to cancel wait
757	<	QNode() {
757	>	QNode next;
758	>	QNode(Phaser phaser, int phase, boolean interruptible,
759	>	boolean timed, long startTime, long nanos) {
760	>	this.phaser = phaser;
761	>	this.phase = phase;
762	>	this.timed = timed;
763	>	this.interruptible = interruptible;
764	>	this.startTime = startTime;
765	>	this.nanos = nanos;
766		thread = Thread.currentThread();
767		}
768	+	public boolean isReleasable() {
769	+	return (thread == null ||
770	+	phaser.getPhase() != phase ||
771	+	(interruptible && wasInterrupted) ||
772	+	(timed && (nanos - (System.nanoTime() - startTime)) <= 0));
773	+	}
774	+	public boolean block() {
775	+	if (Thread.interrupted()) {
776	+	wasInterrupted = true;
777	+	if (interruptible)
778	+	return true;
779	+	}
780	+	if (!timed)
781	+	LockSupport.park(this);
782	+	else {
783	+	long waitTime = nanos - (System.nanoTime() - startTime);
784	+	if (waitTime <= 0)
785	+	return true;
786	+	LockSupport.parkNanos(this, waitTime);
787	+	}
788	+	return isReleasable();
789	+	}
790		void signal() {
791		Thread t = thread;
792		if (t != null) {
#	Line 765 | Line 794 | public class Phaser {
794		LockSupport.unpark(t);
795		}
796		}
797	+	boolean doWait() {
798	+	if (thread != null) {
799	+	try {
800	+	ForkJoinPool.managedBlock(this, false);
801	+	} catch (InterruptedException ie) {
802	+	}
803	+	}
804	+	return wasInterrupted;
805	+	}
806	+
807		}
808
809		/**
810	<	* Removes and signals waiting threads from wait queue
810	>	* Removes and signals waiting threads from wait queue.
811		*/
812		private void releaseWaiters(int phase) {
813		AtomicReference<QNode> head = queueFor(phase);
#	Line 780 | Line 819 | public class Phaser {
819		}
820
821		/**
822	+	* Tries to enqueue given node in the appropriate wait queue.
823	+	*
824	+	* @return true if successful
825	+	*/
826	+	private boolean tryEnqueue(QNode node) {
827	+	AtomicReference<QNode> head = queueFor(node.phase);
828	+	return head.compareAndSet(node.next = head.get(), node);
829	+	}
830	+
831	+	/**
832		* Enqueues node and waits unless aborted or signalled.
833	+	*
834	+	* @return current phase
835		*/
836		private int untimedWait(int phase) {
786	–	int spins = maxUntimedSpins;
837		QNode node = null;
788	–	boolean interrupted = false;
838		boolean queued = false;
839	+	boolean interrupted = false;
840		int p;
841		while ((p = getPhase()) == phase) {
842	<	interrupted = Thread.interrupted();
843	<	if (node != null) {
844	<	if (!queued) {
845	<	AtomicReference<QNode> head = queueFor(phase);
846	<	queued = head.compareAndSet(node.next = head.get(), node);
847	<	}
798	<	else if (node.thread != null)
799	<	LockSupport.park(this);
800	<	}
801	<	else if (spins <= 0)
802	<	node = new QNode();
842	>	if (Thread.interrupted())
843	>	interrupted = true;
844	>	else if (node == null)
845	>	node = new QNode(this, phase, false, false, 0, 0);
846	>	else if (!queued)
847	>	queued = tryEnqueue(node);
848		else
849	<	--spins;
849	>	interrupted = node.doWait();
850		}
851		if (node != null)
852		node.thread = null;
853	+	releaseWaiters(phase);
854		if (interrupted)
855		Thread.currentThread().interrupt();
810	–	releaseWaiters(phase);
856		return p;
857		}
858
859		/**
860	<	* Messier interruptible version
860	>	* Interruptible version
861	>	* @return current phase
862		*/
863		private int interruptibleWait(int phase) throws InterruptedException {
818	–	int spins = maxUntimedSpins;
864		QNode node = null;
865		boolean queued = false;
866		boolean interrupted = false;
867		int p;
868	<	while ((p = getPhase()) == phase) {
869	<	if (interrupted = Thread.interrupted())
870	<	break;
871	<	if (node != null) {
872	<	if (!queued) {
873	<	AtomicReference<QNode> head = queueFor(phase);
874	<	queued = head.compareAndSet(node.next = head.get(), node);
830	<	}
831	<	else if (node.thread != null)
832	<	LockSupport.park(this);
833	<	}
834	<	else if (spins <= 0)
835	<	node = new QNode();
868	>	while ((p = getPhase()) == phase && !interrupted) {
869	>	if (Thread.interrupted())
870	>	interrupted = true;
871	>	else if (node == null)
872	>	node = new QNode(this, phase, true, false, 0, 0);
873	>	else if (!queued)
874	>	queued = tryEnqueue(node);
875		else
876	<	--spins;
876	>	interrupted = node.doWait();
877		}
878		if (node != null)
879		node.thread = null;
880	+	if (p != phase || (p = getPhase()) != phase)
881	+	releaseWaiters(phase);
882		if (interrupted)
883		throw new InterruptedException();
843	–	releaseWaiters(phase);
884		return p;
885		}
886
887		/**
888	<	* Even messier timeout version.
888	>	* Timeout version.
889	>	* @return current phase
890		*/
891		private int timedWait(int phase, long nanos)
892		throws InterruptedException, TimeoutException {
893	+	long startTime = System.nanoTime();
894	+	QNode node = null;
895	+	boolean queued = false;
896	+	boolean interrupted = false;
897		int p;
898	<	if ((p = getPhase()) == phase) {
899	<	long lastTime = System.nanoTime();
900	<	int spins = maxTimedSpins;
901	<	QNode node = null;
902	<	boolean queued = false;
903	<	boolean interrupted = false;
904	<	while ((p = getPhase()) == phase) {
905	<	if (interrupted = Thread.interrupted())
906	<	break;
907	<	long now = System.nanoTime();
908	<	if ((nanos -= now - lastTime) <= 0)
864	<	break;
865	<	lastTime = now;
866	<	if (node != null) {
867	<	if (!queued) {
868	<	AtomicReference<QNode> head = queueFor(phase);
869	<	queued = head.compareAndSet(node.next = head.get(), node);
870	<	}
871	<	else if (node.thread != null &&
872	<	nanos > spinForTimeoutThreshold) {
873	<	LockSupport.parkNanos(this, nanos);
874	<	}
875	<	}
876	<	else if (spins <= 0)
877	<	node = new QNode();
878	<	else
879	<	--spins;
880	<	}
881	<	if (node != null)
882	<	node.thread = null;
883	<	if (interrupted)
884	<	throw new InterruptedException();
885	<	if (p == phase && (p = getPhase()) == phase)
886	<	throw new TimeoutException();
898	>	while ((p = getPhase()) == phase && !interrupted) {
899	>	if (Thread.interrupted())
900	>	interrupted = true;
901	>	else if (nanos - (System.nanoTime() - startTime) <= 0)
902	>	break;
903	>	else if (node == null)
904	>	node = new QNode(this, phase, true, true, startTime, nanos);
905	>	else if (!queued)
906	>	queued = tryEnqueue(node);
907	>	else
908	>	interrupted = node.doWait();
909		}
910	<	releaseWaiters(phase);
910	>	if (node != null)
911	>	node.thread = null;
912	>	if (p != phase || (p = getPhase()) != phase)
913	>	releaseWaiters(phase);
914	>	if (interrupted)
915	>	throw new InterruptedException();
916	>	if (p == phase)
917	>	throw new TimeoutException();
918		return p;
919		}
920
921	<	// Temporary Unsafe mechanics for preliminary release
921	>	// Unsafe mechanics
922
923	<	static final Unsafe _unsafe;
924	<	static final long stateOffset;
923	>	private static final sun.misc.Unsafe UNSAFE = getUnsafe();
924	>	private static final long stateOffset =
925	>	objectFieldOffset("state", Phaser.class);
926
927	<	static {
927	>	private final boolean casState(long cmp, long val) {
928	>	return UNSAFE.compareAndSwapLong(this, stateOffset, cmp, val);
929	>	}
930	>
931	>	private static long objectFieldOffset(String field, Class<?> klazz) {
932		try {
933	<	if (Phaser.class.getClassLoader() != null) {
934	<	Field f = Unsafe.class.getDeclaredField("theUnsafe");
935	<	f.setAccessible(true);
936	<	_unsafe = (Unsafe)f.get(null);
937	<	}
938	<	else
905	<	_unsafe = Unsafe.getUnsafe();
906	<	stateOffset = _unsafe.objectFieldOffset
907	<	(Phaser.class.getDeclaredField("state"));
908	<	} catch (Exception e) {
909	<	throw new RuntimeException("Could not initialize intrinsics", e);
933	>	return UNSAFE.objectFieldOffset(klazz.getDeclaredField(field));
934	>	} catch (NoSuchFieldException e) {
935	>	// Convert Exception to corresponding Error
936	>	NoSuchFieldError error = new NoSuchFieldError(field);
937	>	error.initCause(e);
938	>	throw error;
939		}
940		}
941
942	<	final boolean casState(long cmp, long val) {
943	<	return _unsafe.compareAndSwapLong(this, stateOffset, cmp, val);
942	>	/**
943	>	* Returns a sun.misc.Unsafe.  Suitable for use in a 3rd party package.
944	>	* Replace with a simple call to Unsafe.getUnsafe when integrating
945	>	* into a jdk.
946	>	*
947	>	* @return a sun.misc.Unsafe
948	>	*/
949	>	private static sun.misc.Unsafe getUnsafe() {
950	>	try {
951	>	return sun.misc.Unsafe.getUnsafe();
952	>	} catch (SecurityException se) {
953	>	try {
954	>	return java.security.AccessController.doPrivileged
955	>	(new java.security
956	>	.PrivilegedExceptionAction<sun.misc.Unsafe>() {
957	>	public sun.misc.Unsafe run() throws Exception {
958	>	java.lang.reflect.Field f = sun.misc
959	>	.Unsafe.class.getDeclaredField("theUnsafe");
960	>	f.setAccessible(true);
961	>	return (sun.misc.Unsafe) f.get(null);
962	>	}});
963	>	} catch (java.security.PrivilegedActionException e) {
964	>	throw new RuntimeException("Could not initialize intrinsics",
965	>	e.getCause());
966	>	}
967	>	}
968		}
969		}

Diff Legend

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