ViewVC Help
View File | Revision Log | Show Annotations | Download File | Root Listing
root/jsr166/jsr166/src/jsr166y/ForkJoinPool.java
(Generate patch)

Comparing jsr166/src/jsr166y/ForkJoinPool.java (file contents):
Revision 1.10 by jsr166, Mon Jul 20 23:07:43 2009 UTC vs.
Revision 1.78 by dl, Tue Sep 7 14:52:24 2010 UTC

# Line 5 | Line 5
5   */
6  
7   package jsr166y;
8 < import java.util.*;
9 < import java.util.concurrent.*;
10 < import java.util.concurrent.locks.*;
11 < import java.util.concurrent.atomic.*;
12 < import sun.misc.Unsafe;
13 < import java.lang.reflect.*;
8 >
9 > import java.util.ArrayList;
10 > import java.util.Arrays;
11 > import java.util.Collection;
12 > import java.util.Collections;
13 > import java.util.List;
14 > import java.util.concurrent.AbstractExecutorService;
15 > import java.util.concurrent.Callable;
16 > import java.util.concurrent.CountDownLatch;
17 > import java.util.concurrent.ExecutorService;
18 > import java.util.concurrent.Future;
19 > import java.util.concurrent.RejectedExecutionException;
20 > import java.util.concurrent.RunnableFuture;
21 > import java.util.concurrent.TimeUnit;
22 > import java.util.concurrent.TimeoutException;
23 > import java.util.concurrent.atomic.AtomicInteger;
24 > import java.util.concurrent.locks.LockSupport;
25 > import java.util.concurrent.locks.ReentrantLock;
26  
27   /**
28 < * An {@link ExecutorService} for running {@link ForkJoinTask}s.  A
29 < * ForkJoinPool provides the entry point for submissions from
30 < * non-ForkJoinTasks, as well as management and monitoring operations.
31 < * Normally a single ForkJoinPool is used for a large number of
20 < * submitted tasks. Otherwise, use would not usually outweigh the
21 < * construction and bookkeeping overhead of creating a large set of
22 < * threads.
28 > * An {@link ExecutorService} for running {@link ForkJoinTask}s.
29 > * A {@code ForkJoinPool} provides the entry point for submissions
30 > * from non-{@code ForkJoinTask} clients, as well as management and
31 > * monitoring operations.
32   *
33 < * <p>ForkJoinPools differ from other kinds of Executors mainly in
34 < * that they provide <em>work-stealing</em>: all threads in the pool
35 < * attempt to find and execute subtasks created by other active tasks
36 < * (eventually blocking if none exist). This makes them efficient when
37 < * most tasks spawn other subtasks (as do most ForkJoinTasks), as well
38 < * as the mixed execution of some plain Runnable- or Callable- based
39 < * activities along with ForkJoinTasks. When setting
40 < * {@code setAsyncMode}, a ForkJoinPools may also be appropriate for
41 < * use with fine-grained tasks that are never joined. Otherwise, other
33 < * ExecutorService implementations are typically more appropriate
34 < * choices.
33 > * <p>A {@code ForkJoinPool} differs from other kinds of {@link
34 > * ExecutorService} mainly by virtue of employing
35 > * <em>work-stealing</em>: all threads in the pool attempt to find and
36 > * execute subtasks created by other active tasks (eventually blocking
37 > * waiting for work if none exist). This enables efficient processing
38 > * when most tasks spawn other subtasks (as do most {@code
39 > * ForkJoinTask}s). When setting <em>asyncMode</em> to true in
40 > * constructors, {@code ForkJoinPool}s may also be appropriate for use
41 > * with event-style tasks that are never joined.
42   *
43 < * <p>A ForkJoinPool may be constructed with a given parallelism level
44 < * (target pool size), which it attempts to maintain by dynamically
45 < * adding, suspending, or resuming threads, even if some tasks are
46 < * waiting to join others. However, no such adjustments are performed
47 < * in the face of blocked IO or other unmanaged synchronization. The
48 < * nested {@code ManagedBlocker} interface enables extension of
49 < * the kinds of synchronization accommodated.  The target parallelism
50 < * level may also be changed dynamically ({@code setParallelism})
51 < * and thread construction can be limited using methods
45 < * {@code setMaximumPoolSize} and/or
46 < * {@code setMaintainsParallelism}.
43 > * <p>A {@code ForkJoinPool} is constructed with a given target
44 > * parallelism level; by default, equal to the number of available
45 > * processors. The pool attempts to maintain enough active (or
46 > * available) threads by dynamically adding, suspending, or resuming
47 > * internal worker threads, even if some tasks are stalled waiting to
48 > * join others. However, no such adjustments are guaranteed in the
49 > * face of blocked IO or other unmanaged synchronization. The nested
50 > * {@link ManagedBlocker} interface enables extension of the kinds of
51 > * synchronization accommodated.
52   *
53   * <p>In addition to execution and lifecycle control methods, this
54   * class provides status check methods (for example
55 < * {@code getStealCount}) that are intended to aid in developing,
55 > * {@link #getStealCount}) that are intended to aid in developing,
56   * tuning, and monitoring fork/join applications. Also, method
57 < * {@code toString} returns indications of pool state in a
57 > * {@link #toString} returns indications of pool state in a
58   * convenient form for informal monitoring.
59   *
60 + * <p> As is the case with other ExecutorServices, there are three
61 + * main task execution methods summarized in the following
62 + * table. These are designed to be used by clients not already engaged
63 + * in fork/join computations in the current pool.  The main forms of
64 + * these methods accept instances of {@code ForkJoinTask}, but
65 + * overloaded forms also allow mixed execution of plain {@code
66 + * Runnable}- or {@code Callable}- based activities as well.  However,
67 + * tasks that are already executing in a pool should normally
68 + * <em>NOT</em> use these pool execution methods, but instead use the
69 + * within-computation forms listed in the table.
70 + *
71 + * <table BORDER CELLPADDING=3 CELLSPACING=1>
72 + *  <tr>
73 + *    <td></td>
74 + *    <td ALIGN=CENTER> <b>Call from non-fork/join clients</b></td>
75 + *    <td ALIGN=CENTER> <b>Call from within fork/join computations</b></td>
76 + *  </tr>
77 + *  <tr>
78 + *    <td> <b>Arrange async execution</td>
79 + *    <td> {@link #execute(ForkJoinTask)}</td>
80 + *    <td> {@link ForkJoinTask#fork}</td>
81 + *  </tr>
82 + *  <tr>
83 + *    <td> <b>Await and obtain result</td>
84 + *    <td> {@link #invoke(ForkJoinTask)}</td>
85 + *    <td> {@link ForkJoinTask#invoke}</td>
86 + *  </tr>
87 + *  <tr>
88 + *    <td> <b>Arrange exec and obtain Future</td>
89 + *    <td> {@link #submit(ForkJoinTask)}</td>
90 + *    <td> {@link ForkJoinTask#fork} (ForkJoinTasks <em>are</em> Futures)</td>
91 + *  </tr>
92 + * </table>
93 + *
94 + * <p><b>Sample Usage.</b> Normally a single {@code ForkJoinPool} is
95 + * used for all parallel task execution in a program or subsystem.
96 + * Otherwise, use would not usually outweigh the construction and
97 + * bookkeeping overhead of creating a large set of threads. For
98 + * example, a common pool could be used for the {@code SortTasks}
99 + * illustrated in {@link RecursiveAction}. Because {@code
100 + * ForkJoinPool} uses threads in {@linkplain java.lang.Thread#isDaemon
101 + * daemon} mode, there is typically no need to explicitly {@link
102 + * #shutdown} such a pool upon program exit.
103 + *
104 + * <pre>
105 + * static final ForkJoinPool mainPool = new ForkJoinPool();
106 + * ...
107 + * public void sort(long[] array) {
108 + *   mainPool.invoke(new SortTask(array, 0, array.length));
109 + * }
110 + * </pre>
111 + *
112   * <p><b>Implementation notes</b>: This implementation restricts the
113   * maximum number of running threads to 32767. Attempts to create
114 < * pools with greater than the maximum result in
115 < * IllegalArgumentExceptions.
114 > * pools with greater than the maximum number result in
115 > * {@code IllegalArgumentException}.
116 > *
117 > * <p>This implementation rejects submitted tasks (that is, by throwing
118 > * {@link RejectedExecutionException}) only when the pool is shut down
119 > * or internal resources have been exhausted.
120 > *
121 > * @since 1.7
122 > * @author Doug Lea
123   */
124   public class ForkJoinPool extends AbstractExecutorService {
125  
126      /*
127 <     * See the extended comments interspersed below for design,
128 <     * rationale, and walkthroughs.
129 <     */
130 <
131 <    /** Mask for packing and unpacking shorts */
132 <    private static final int  shortMask = 0xffff;
133 <
134 <    /** Max pool size -- must be a power of two minus 1 */
135 <    private static final int MAX_THREADS =  0x7FFF;
136 <
137 <    /**
138 <     * Factory for creating new ForkJoinWorkerThreads.  A
139 <     * ForkJoinWorkerThreadFactory must be defined and used for
140 <     * ForkJoinWorkerThread subclasses that extend base functionality
141 <     * or initialize threads with different contexts.
127 >     * Implementation Overview
128 >     *
129 >     * This class provides the central bookkeeping and control for a
130 >     * set of worker threads: Submissions from non-FJ threads enter
131 >     * into a submission queue. Workers take these tasks and typically
132 >     * split them into subtasks that may be stolen by other workers.
133 >     * The main work-stealing mechanics implemented in class
134 >     * ForkJoinWorkerThread give first priority to processing tasks
135 >     * from their own queues (LIFO or FIFO, depending on mode), then
136 >     * to randomized FIFO steals of tasks in other worker queues, and
137 >     * lastly to new submissions. These mechanics do not consider
138 >     * affinities, loads, cache localities, etc, so rarely provide the
139 >     * best possible performance on a given machine, but portably
140 >     * provide good throughput by averaging over these factors.
141 >     * (Further, even if we did try to use such information, we do not
142 >     * usually have a basis for exploiting it. For example, some sets
143 >     * of tasks profit from cache affinities, but others are harmed by
144 >     * cache pollution effects.)
145 >     *
146 >     * Beyond work-stealing support and essential bookkeeping, the
147 >     * main responsibility of this framework is to take actions when
148 >     * one worker is waiting to join a task stolen (or always held by)
149 >     * another.  Because we are multiplexing many tasks on to a pool
150 >     * of workers, we can't just let them block (as in Thread.join).
151 >     * We also cannot just reassign the joiner's run-time stack with
152 >     * another and replace it later, which would be a form of
153 >     * "continuation", that even if possible is not necessarily a good
154 >     * idea. Given that the creation costs of most threads on most
155 >     * systems mainly surrounds setting up runtime stacks, thread
156 >     * creation and switching is usually not much more expensive than
157 >     * stack creation and switching, and is more flexible). Instead we
158 >     * combine two tactics:
159 >     *
160 >     *   Helping: Arranging for the joiner to execute some task that it
161 >     *      would be running if the steal had not occurred.  Method
162 >     *      ForkJoinWorkerThread.helpJoinTask tracks joining->stealing
163 >     *      links to try to find such a task.
164 >     *
165 >     *   Compensating: Unless there are already enough live threads,
166 >     *      method helpMaintainParallelism() may create or
167 >     *      re-activate a spare thread to compensate for blocked
168 >     *      joiners until they unblock.
169 >     *
170 >     * It is impossible to keep exactly the target (parallelism)
171 >     * number of threads running at any given time.  Determining
172 >     * existence of conservatively safe helping targets, the
173 >     * availability of already-created spares, and the apparent need
174 >     * to create new spares are all racy and require heuristic
175 >     * guidance, so we rely on multiple retries of each.  Compensation
176 >     * occurs in slow-motion. It is triggered only upon timeouts of
177 >     * Object.wait used for joins. This reduces poor decisions that
178 >     * would otherwise be made when threads are waiting for others
179 >     * that are stalled because of unrelated activities such as
180 >     * garbage collection.
181 >     *
182 >     * The ManagedBlocker extension API can't use helping so relies
183 >     * only on compensation in method awaitBlocker.
184 >     *
185 >     * The main throughput advantages of work-stealing stem from
186 >     * decentralized control -- workers mostly steal tasks from each
187 >     * other. We do not want to negate this by creating bottlenecks
188 >     * implementing other management responsibilities. So we use a
189 >     * collection of techniques that avoid, reduce, or cope well with
190 >     * contention. These entail several instances of bit-packing into
191 >     * CASable fields to maintain only the minimally required
192 >     * atomicity. To enable such packing, we restrict maximum
193 >     * parallelism to (1<<15)-1 (enabling twice this (to accommodate
194 >     * unbalanced increments and decrements) to fit into a 16 bit
195 >     * field, which is far in excess of normal operating range.  Even
196 >     * though updates to some of these bookkeeping fields do sometimes
197 >     * contend with each other, they don't normally cache-contend with
198 >     * updates to others enough to warrant memory padding or
199 >     * isolation. So they are all held as fields of ForkJoinPool
200 >     * objects.  The main capabilities are as follows:
201 >     *
202 >     * 1. Creating and removing workers. Workers are recorded in the
203 >     * "workers" array. This is an array as opposed to some other data
204 >     * structure to support index-based random steals by workers.
205 >     * Updates to the array recording new workers and unrecording
206 >     * terminated ones are protected from each other by a lock
207 >     * (workerLock) but the array is otherwise concurrently readable,
208 >     * and accessed directly by workers. To simplify index-based
209 >     * operations, the array size is always a power of two, and all
210 >     * readers must tolerate null slots. Currently, all worker thread
211 >     * creation is on-demand, triggered by task submissions,
212 >     * replacement of terminated workers, and/or compensation for
213 >     * blocked workers. However, all other support code is set up to
214 >     * work with other policies.
215 >     *
216 >     * To ensure that we do not hold on to worker references that
217 >     * would prevent GC, ALL accesses to workers are via indices into
218 >     * the workers array (which is one source of some of the unusual
219 >     * code constructions here). In essence, the workers array serves
220 >     * as a WeakReference mechanism. Thus for example the event queue
221 >     * stores worker indices, not worker references. Access to the
222 >     * workers in associated methods (for example releaseEventWaiters)
223 >     * must both index-check and null-check the IDs. All such accesses
224 >     * ignore bad IDs by returning out early from what they are doing,
225 >     * since this can only be associated with shutdown, in which case
226 >     * it is OK to give up. On termination, we just clobber these
227 >     * data structures without trying to use them.
228 >     *
229 >     * 2. Bookkeeping for dynamically adding and removing workers. We
230 >     * aim to approximately maintain the given level of parallelism.
231 >     * When some workers are known to be blocked (on joins or via
232 >     * ManagedBlocker), we may create or resume others to take their
233 >     * place until they unblock (see below). Implementing this
234 >     * requires counts of the number of "running" threads (i.e., those
235 >     * that are neither blocked nor artificially suspended) as well as
236 >     * the total number.  These two values are packed into one field,
237 >     * "workerCounts" because we need accurate snapshots when deciding
238 >     * to create, resume or suspend.  Note however that the
239 >     * correspondence of these counts to reality is not guaranteed. In
240 >     * particular updates for unblocked threads may lag until they
241 >     * actually wake up.
242 >     *
243 >     * 3. Maintaining global run state. The run state of the pool
244 >     * consists of a runLevel (SHUTDOWN, TERMINATING, etc) similar to
245 >     * those in other Executor implementations, as well as a count of
246 >     * "active" workers -- those that are, or soon will be, or
247 >     * recently were executing tasks. The runLevel and active count
248 >     * are packed together in order to correctly trigger shutdown and
249 >     * termination. Without care, active counts can be subject to very
250 >     * high contention.  We substantially reduce this contention by
251 >     * relaxing update rules.  A worker must claim active status
252 >     * prospectively, by activating if it sees that a submitted or
253 >     * stealable task exists (it may find after activating that the
254 >     * task no longer exists). It stays active while processing this
255 >     * task (if it exists) and any other local subtasks it produces,
256 >     * until it cannot find any other tasks. It then tries
257 >     * inactivating (see method preStep), but upon update contention
258 >     * instead scans for more tasks, later retrying inactivation if it
259 >     * doesn't find any.
260 >     *
261 >     * 4. Managing idle workers waiting for tasks. We cannot let
262 >     * workers spin indefinitely scanning for tasks when none are
263 >     * available. On the other hand, we must quickly prod them into
264 >     * action when new tasks are submitted or generated.  We
265 >     * park/unpark these idle workers using an event-count scheme.
266 >     * Field eventCount is incremented upon events that may enable
267 >     * workers that previously could not find a task to now find one:
268 >     * Submission of a new task to the pool, or another worker pushing
269 >     * a task onto a previously empty queue.  (We also use this
270 >     * mechanism for configuration and termination actions that
271 >     * require wakeups of idle workers).  Each worker maintains its
272 >     * last known event count, and blocks when a scan for work did not
273 >     * find a task AND its lastEventCount matches the current
274 >     * eventCount. Waiting idle workers are recorded in a variant of
275 >     * Treiber stack headed by field eventWaiters which, when nonzero,
276 >     * encodes the thread index and count awaited for by the worker
277 >     * thread most recently calling eventSync. This thread in turn has
278 >     * a record (field nextEventWaiter) for the next waiting worker.
279 >     * In addition to allowing simpler decisions about need for
280 >     * wakeup, the event count bits in eventWaiters serve the role of
281 >     * tags to avoid ABA errors in Treiber stacks. Upon any wakeup,
282 >     * released threads also try to release at most two others.  The
283 >     * net effect is a tree-like diffusion of signals, where released
284 >     * threads (and possibly others) help with unparks.  To further
285 >     * reduce contention effects a bit, failed CASes to increment
286 >     * field eventCount are tolerated without retries in signalWork.
287 >     * Conceptually they are merged into the same event, which is OK
288 >     * when their only purpose is to enable workers to scan for work.
289 >     *
290 >     * 5. Managing suspension of extra workers. When a worker notices
291 >     * (usually upon timeout of a wait()) that there are too few
292 >     * running threads, we may create a new thread to maintain
293 >     * parallelism level, or at least avoid starvation. Usually, extra
294 >     * threads are needed for only very short periods, yet join
295 >     * dependencies are such that we sometimes need them in
296 >     * bursts. Rather than create new threads each time this happens,
297 >     * we suspend no-longer-needed extra ones as "spares". For most
298 >     * purposes, we don't distinguish "extra" spare threads from
299 >     * normal "core" threads: On each call to preStep (the only point
300 >     * at which we can do this) a worker checks to see if there are
301 >     * now too many running workers, and if so, suspends itself.
302 >     * Method helpMaintainParallelism looks for suspended threads to
303 >     * resume before considering creating a new replacement. The
304 >     * spares themselves are encoded on another variant of a Treiber
305 >     * Stack, headed at field "spareWaiters".  Note that the use of
306 >     * spares is intrinsically racy.  One thread may become a spare at
307 >     * about the same time as another is needlessly being created. We
308 >     * counteract this and related slop in part by requiring resumed
309 >     * spares to immediately recheck (in preStep) to see whether they
310 >     * should re-suspend.
311 >     *
312 >     * 6. Killing off unneeded workers. A timeout mechanism is used to
313 >     * shed unused workers: The oldest (first) event queue waiter uses
314 >     * a timed rather than hard wait. When this wait times out without
315 >     * a normal wakeup, it tries to shutdown any one (for convenience
316 >     * the newest) other spare or event waiter via
317 >     * tryShutdownUnusedWorker. This eventually reduces the number of
318 >     * worker threads to a minimum of one after a long enough period
319 >     * without use.
320 >     *
321 >     * 7. Deciding when to create new workers. The main dynamic
322 >     * control in this class is deciding when to create extra threads
323 >     * in method helpMaintainParallelism. We would like to keep
324 >     * exactly #parallelism threads running, which is an impossible
325 >     * task. We always need to create one when the number of running
326 >     * threads would become zero and all workers are busy. Beyond
327 >     * this, we must rely on heuristics that work well in the
328 >     * presence of transient phenomena such as GC stalls, dynamic
329 >     * compilation, and wake-up lags. These transients are extremely
330 >     * common -- we are normally trying to fully saturate the CPUs on
331 >     * a machine, so almost any activity other than running tasks
332 >     * impedes accuracy. Our main defense is to allow parallelism to
333 >     * lapse for a while during joins, and use a timeout to see if,
334 >     * after the resulting settling, there is still a need for
335 >     * additional workers.  This also better copes with the fact that
336 >     * some of the methods in this class tend to never become compiled
337 >     * (but are interpreted), so some components of the entire set of
338 >     * controls might execute 100 times faster than others. And
339 >     * similarly for cases where the apparent lack of work is just due
340 >     * to GC stalls and other transient system activity.
341 >     *
342 >     * Beware that there is a lot of representation-level coupling
343 >     * among classes ForkJoinPool, ForkJoinWorkerThread, and
344 >     * ForkJoinTask.  For example, direct access to "workers" array by
345 >     * workers, and direct access to ForkJoinTask.status by both
346 >     * ForkJoinPool and ForkJoinWorkerThread.  There is little point
347 >     * trying to reduce this, since any associated future changes in
348 >     * representations will need to be accompanied by algorithmic
349 >     * changes anyway.
350 >     *
351 >     * Style notes: There are lots of inline assignments (of form
352 >     * "while ((local = field) != 0)") which are usually the simplest
353 >     * way to ensure the required read orderings (which are sometimes
354 >     * critical). Also several occurrences of the unusual "do {}
355 >     * while (!cas...)" which is the simplest way to force an update of
356 >     * a CAS'ed variable. There are also other coding oddities that
357 >     * help some methods perform reasonably even when interpreted (not
358 >     * compiled), at the expense of some messy constructions that
359 >     * reduce byte code counts.
360 >     *
361 >     * The order of declarations in this file is: (1) statics (2)
362 >     * fields (along with constants used when unpacking some of them)
363 >     * (3) internal control methods (4) callbacks and other support
364 >     * for ForkJoinTask and ForkJoinWorkerThread classes, (5) exported
365 >     * methods (plus a few little helpers).
366 >     */
367 >
368 >    /**
369 >     * Factory for creating new {@link ForkJoinWorkerThread}s.
370 >     * A {@code ForkJoinWorkerThreadFactory} must be defined and used
371 >     * for {@code ForkJoinWorkerThread} subclasses that extend base
372 >     * functionality or initialize threads with different contexts.
373       */
374      public static interface ForkJoinWorkerThreadFactory {
375          /**
376           * Returns a new worker thread operating in the given pool.
377           *
378           * @param pool the pool this thread works in
379 <         * @throws NullPointerException if pool is null;
379 >         * @throws NullPointerException if the pool is null
380           */
381          public ForkJoinWorkerThread newThread(ForkJoinPool pool);
382      }
383  
384      /**
385 <     * Default ForkJoinWorkerThreadFactory implementation, creates a
385 >     * Default ForkJoinWorkerThreadFactory implementation; creates a
386       * new ForkJoinWorkerThread.
387       */
388 <    static class  DefaultForkJoinWorkerThreadFactory
388 >    static class DefaultForkJoinWorkerThreadFactory
389          implements ForkJoinWorkerThreadFactory {
390          public ForkJoinWorkerThread newThread(ForkJoinPool pool) {
391 <            try {
97 <                return new ForkJoinWorkerThread(pool);
98 <            } catch (OutOfMemoryError oom)  {
99 <                return null;
100 <            }
391 >            return new ForkJoinWorkerThread(pool);
392          }
393      }
394  
# Line 133 | Line 424 | public class ForkJoinPool extends Abstra
424          new AtomicInteger();
425  
426      /**
427 <     * Array holding all worker threads in the pool. Initialized upon
428 <     * first use. Array size must be a power of two.  Updates and
429 <     * replacements are protected by workerLock, but it is always kept
430 <     * in a consistent enough state to be randomly accessed without
431 <     * locking by workers performing work-stealing.
427 >     * The time to block in a join (see awaitJoin) before checking if
428 >     * a new worker should be (re)started to maintain parallelism
429 >     * level. The value should be short enough to maintain global
430 >     * responsiveness and progress but long enough to avoid
431 >     * counterproductive firings during GC stalls or unrelated system
432 >     * activity, and to not bog down systems with continual re-firings
433 >     * on GCs or legitimately long waits.
434 >     */
435 >    private static final long JOIN_TIMEOUT_MILLIS = 250L; // 4 per second
436 >
437 >    /**
438 >     * The wakeup interval (in nanoseconds) for the oldest worker
439 >     * waiting for an event to invoke tryShutdownUnusedWorker to
440 >     * shrink the number of workers.  The exact value does not matter
441 >     * too much. It must be short enough to release resources during
442 >     * sustained periods of idleness, but not so short that threads
443 >     * are continually re-created.
444 >     */
445 >    private static final long SHRINK_RATE_NANOS =
446 >        30L * 1000L * 1000L * 1000L; // 2 per minute
447 >
448 >    /**
449 >     * Absolute bound for parallelism level. Twice this number plus
450 >     * one (i.e., 0xfff) must fit into a 16bit field to enable
451 >     * word-packing for some counts and indices.
452 >     */
453 >    private static final int MAX_WORKERS   = 0x7fff;
454 >
455 >    /**
456 >     * Array holding all worker threads in the pool.  Array size must
457 >     * be a power of two.  Updates and replacements are protected by
458 >     * workerLock, but the array is always kept in a consistent enough
459 >     * state to be randomly accessed without locking by workers
460 >     * performing work-stealing, as well as other traversal-based
461 >     * methods in this class. All readers must tolerate that some
462 >     * array slots may be null.
463       */
464      volatile ForkJoinWorkerThread[] workers;
465  
466      /**
467 <     * Lock protecting access to workers.
467 >     * Queue for external submissions.
468       */
469 <    private final ReentrantLock workerLock;
469 >    private final LinkedTransferQueue<ForkJoinTask<?>> submissionQueue;
470  
471      /**
472 <     * Condition for awaitTermination.
472 >     * Lock protecting updates to workers array.
473       */
474 <    private final Condition termination;
474 >    private final ReentrantLock workerLock;
475  
476      /**
477 <     * The uncaught exception handler used when any worker
156 <     * abruptly terminates
477 >     * Latch released upon termination.
478       */
479 <    private Thread.UncaughtExceptionHandler ueh;
479 >    private final Phaser termination;
480  
481      /**
482       * Creation factory for worker threads.
# Line 163 | Line 484 | public class ForkJoinPool extends Abstra
484      private final ForkJoinWorkerThreadFactory factory;
485  
486      /**
166     * Head of stack of threads that were created to maintain
167     * parallelism when other threads blocked, but have since
168     * suspended when the parallelism level rose.
169     */
170    private volatile WaitQueueNode spareStack;
171
172    /**
487       * Sum of per-thread steal counts, updated only when threads are
488       * idle or terminating.
489       */
490 <    private final AtomicLong stealCount;
490 >    private volatile long stealCount;
491  
492      /**
493 <     * Queue for external submissions.
493 >     * Encoded record of top of Treiber stack of threads waiting for
494 >     * events. The top 32 bits contain the count being waited for. The
495 >     * bottom 16 bits contains one plus the pool index of waiting
496 >     * worker thread. (Bits 16-31 are unused.)
497       */
498 <    private final LinkedTransferQueue<ForkJoinTask<?>> submissionQueue;
498 >    private volatile long eventWaiters;
499  
500 <    /**
501 <     * Head of Treiber stack for barrier sync. See below for explanation
185 <     */
186 <    private volatile WaitQueueNode syncStack;
500 >    private static final int  EVENT_COUNT_SHIFT = 32;
501 >    private static final long WAITER_ID_MASK    = (1L << 16) - 1L;
502  
503      /**
504 <     * The count for event barrier
504 >     * A counter for events that may wake up worker threads:
505 >     *   - Submission of a new task to the pool
506 >     *   - A worker pushing a task on an empty queue
507 >     *   - termination
508       */
509 <    private volatile long eventCount;
509 >    private volatile int eventCount;
510  
511      /**
512 <     * Pool number, just for assigning useful names to worker threads
512 >     * Encoded record of top of Treiber stack of spare threads waiting
513 >     * for resumption. The top 16 bits contain an arbitrary count to
514 >     * avoid ABA effects. The bottom 16bits contains one plus the pool
515 >     * index of waiting worker thread.
516       */
517 <    private final int poolNumber;
517 >    private volatile int spareWaiters;
518  
519 <    /**
520 <     * The maximum allowed pool size
200 <     */
201 <    private volatile int maxPoolSize;
519 >    private static final int SPARE_COUNT_SHIFT = 16;
520 >    private static final int SPARE_ID_MASK     = (1 << 16) - 1;
521  
522      /**
523 <     * The desired parallelism level, updated only under workerLock.
523 >     * Lifecycle control. The low word contains the number of workers
524 >     * that are (probably) executing tasks. This value is atomically
525 >     * incremented before a worker gets a task to run, and decremented
526 >     * when worker has no tasks and cannot find any.  Bits 16-18
527 >     * contain runLevel value. When all are zero, the pool is
528 >     * running. Level transitions are monotonic (running -> shutdown
529 >     * -> terminating -> terminated) so each transition adds a bit.
530 >     * These are bundled together to ensure consistent read for
531 >     * termination checks (i.e., that runLevel is at least SHUTDOWN
532 >     * and active threads is zero).
533 >     *
534 >     * Notes: Most direct CASes are dependent on these bitfield
535 >     * positions.  Also, this field is non-private to enable direct
536 >     * performance-sensitive CASes in ForkJoinWorkerThread.
537       */
538 <    private volatile int parallelism;
538 >    volatile int runState;
539  
540 <    /**
541 <     * True if use local fifo, not default lifo, for local polling
542 <     */
543 <    private volatile boolean locallyFifo;
540 >    // Note: The order among run level values matters.
541 >    private static final int RUNLEVEL_SHIFT     = 16;
542 >    private static final int SHUTDOWN           = 1 << RUNLEVEL_SHIFT;
543 >    private static final int TERMINATING        = 1 << (RUNLEVEL_SHIFT + 1);
544 >    private static final int TERMINATED         = 1 << (RUNLEVEL_SHIFT + 2);
545 >    private static final int ACTIVE_COUNT_MASK  = (1 << RUNLEVEL_SHIFT) - 1;
546  
547      /**
548       * Holds number of total (i.e., created and not yet terminated)
549       * and running (i.e., not blocked on joins or other managed sync)
550 <     * threads, packed into one int to ensure consistent snapshot when
550 >     * threads, packed together to ensure consistent snapshot when
551       * making decisions about creating and suspending spare
552 <     * threads. Updated only by CAS.  Note: CASes in
553 <     * updateRunningCount and preJoin running active count is in low
554 <     * word, so need to be modified if this changes
552 >     * threads. Updated only by CAS. Note that adding a new worker
553 >     * requires incrementing both counts, since workers start off in
554 >     * running state.
555       */
556      private volatile int workerCounts;
557  
558 <    private static int totalCountOf(int s)           { return s >>> 16;  }
559 <    private static int runningCountOf(int s)         { return s & shortMask; }
560 <    private static int workerCountsFor(int t, int r) { return (t << 16) + r; }
558 >    private static final int TOTAL_COUNT_SHIFT  = 16;
559 >    private static final int RUNNING_COUNT_MASK = (1 << TOTAL_COUNT_SHIFT) - 1;
560 >    private static final int ONE_RUNNING        = 1;
561 >    private static final int ONE_TOTAL          = 1 << TOTAL_COUNT_SHIFT;
562  
563      /**
564 <     * Adds delta (which may be negative) to running count.  This must
565 <     * be called before (with negative arg) and after (with positive)
231 <     * any managed synchronization (i.e., mainly, joins).
232 <     * @param delta the number to add
564 >     * The target parallelism level.
565 >     * Accessed directly by ForkJoinWorkerThreads.
566       */
567 <    final void updateRunningCount(int delta) {
235 <        int s;
236 <        do;while (!casWorkerCounts(s = workerCounts, s + delta));
237 <    }
567 >    final int parallelism;
568  
569      /**
570 <     * Adds delta (which may be negative) to both total and running
571 <     * count.  This must be called upon creation and termination of
242 <     * worker threads.
243 <     * @param delta the number to add
570 >     * True if use local fifo, not default lifo, for local polling
571 >     * Read by, and replicated by ForkJoinWorkerThreads
572       */
573 <    private void updateWorkerCount(int delta) {
246 <        int d = delta + (delta << 16); // add to both lo and hi parts
247 <        int s;
248 <        do;while (!casWorkerCounts(s = workerCounts, s + d));
249 <    }
573 >    final boolean locallyFifo;
574  
575      /**
576 <     * Lifecycle control. High word contains runState, low word
577 <     * contains the number of workers that are (probably) executing
254 <     * tasks. This value is atomically incremented before a worker
255 <     * gets a task to run, and decremented when worker has no tasks
256 <     * and cannot find any. These two fields are bundled together to
257 <     * support correct termination triggering.  Note: activeCount
258 <     * CAS'es cheat by assuming active count is in low word, so need
259 <     * to be modified if this changes
576 >     * The uncaught exception handler used when any worker abruptly
577 >     * terminates.
578       */
579 <    private volatile int runControl;
579 >    private final Thread.UncaughtExceptionHandler ueh;
580  
581 <    // RunState values. Order among values matters
582 <    private static final int RUNNING     = 0;
583 <    private static final int SHUTDOWN    = 1;
584 <    private static final int TERMINATING = 2;
267 <    private static final int TERMINATED  = 3;
581 >    /**
582 >     * Pool number, just for assigning useful names to worker threads
583 >     */
584 >    private final int poolNumber;
585  
586 <    private static int runStateOf(int c)             { return c >>> 16; }
587 <    private static int activeCountOf(int c)          { return c & shortMask; }
271 <    private static int runControlFor(int r, int a)   { return (r << 16) + a; }
586 >    // Utilities for CASing fields. Note that most of these
587 >    // are usually manually inlined by callers
588  
589      /**
590 <     * Try incrementing active count; fail on contention. Called by
275 <     * workers before/during executing tasks.
276 <     * @return true on success;
590 >     * Increments running count part of workerCounts
591       */
592 <    final boolean tryIncrementActiveCount() {
593 <        int c = runControl;
594 <        return casRunControl(c, c+1);
592 >    final void incrementRunningCount() {
593 >        int c;
594 >        do {} while (!UNSAFE.compareAndSwapInt(this, workerCountsOffset,
595 >                                               c = workerCounts,
596 >                                               c + ONE_RUNNING));
597      }
598  
599      /**
600 <     * Tries decrementing active count; fails on contention.
285 <     * Possibly triggers termination on success.
286 <     * Called by workers when they can't find tasks.
287 <     * @return true on success
600 >     * Tries to decrement running count unless already zero
601       */
602 <    final boolean tryDecrementActiveCount() {
603 <        int c = runControl;
604 <        int nextc = c - 1;
292 <        if (!casRunControl(c, nextc))
602 >    final boolean tryDecrementRunningCount() {
603 >        int wc = workerCounts;
604 >        if ((wc & RUNNING_COUNT_MASK) == 0)
605              return false;
606 <        if (canTerminateOnShutdown(nextc))
607 <            terminateOnShutdown();
608 <        return true;
606 >        return UNSAFE.compareAndSwapInt(this, workerCountsOffset,
607 >                                        wc, wc - ONE_RUNNING);
608 >    }
609 >
610 >    /**
611 >     * Forces decrement of encoded workerCounts, awaiting nonzero if
612 >     * (rarely) necessary when other count updates lag.
613 >     *
614 >     * @param dr -- either zero or ONE_RUNNING
615 >     * @param dt -- either zero or ONE_TOTAL
616 >     */
617 >    private void decrementWorkerCounts(int dr, int dt) {
618 >        for (;;) {
619 >            int wc = workerCounts;
620 >            if ((wc & RUNNING_COUNT_MASK)  - dr < 0 ||
621 >                (wc >>> TOTAL_COUNT_SHIFT) - dt < 0) {
622 >                if ((runState & TERMINATED) != 0)
623 >                    return; // lagging termination on a backout
624 >                Thread.yield();
625 >            }
626 >            if (UNSAFE.compareAndSwapInt(this, workerCountsOffset,
627 >                                         wc, wc - (dr + dt)))
628 >                return;
629 >        }
630      }
631  
632      /**
633 <     * Returns true if argument represents zero active count and
634 <     * nonzero runstate, which is the triggering condition for
302 <     * terminating on shutdown.
633 >     * Tries decrementing active count; fails on contention.
634 >     * Called when workers cannot find tasks to run.
635       */
636 <    private static boolean canTerminateOnShutdown(int c) {
637 <        return ((c & -c) >>> 16) != 0; // i.e. least bit is nonzero runState bit
636 >    final boolean tryDecrementActiveCount() {
637 >        int c;
638 >        return UNSAFE.compareAndSwapInt(this, runStateOffset,
639 >                                        c = runState, c - 1);
640      }
641  
642      /**
643 <     * Transition run state to at least the given state. Return true
644 <     * if not already at least given state.
643 >     * Advances to at least the given level. Returns true if not
644 >     * already in at least the given level.
645       */
646 <    private boolean transitionRunStateTo(int state) {
646 >    private boolean advanceRunLevel(int level) {
647          for (;;) {
648 <            int c = runControl;
649 <            if (runStateOf(c) >= state)
648 >            int s = runState;
649 >            if ((s & level) != 0)
650                  return false;
651 <            if (casRunControl(c, runControlFor(state, activeCountOf(c))))
651 >            if (UNSAFE.compareAndSwapInt(this, runStateOffset, s, s | level))
652                  return true;
653          }
654      }
655  
656 +    // workers array maintenance
657 +
658      /**
659 <     * Controls whether to add spares to maintain parallelism
659 >     * Records and returns a workers array index for new worker.
660       */
661 <    private volatile boolean maintainsParallelism;
661 >    private int recordWorker(ForkJoinWorkerThread w) {
662 >        // Try using slot totalCount-1. If not available, scan and/or resize
663 >        int k = (workerCounts >>> TOTAL_COUNT_SHIFT) - 1;
664 >        final ReentrantLock lock = this.workerLock;
665 >        lock.lock();
666 >        try {
667 >            ForkJoinWorkerThread[] ws = workers;
668 >            int n = ws.length;
669 >            if (k < 0 || k >= n || ws[k] != null) {
670 >                for (k = 0; k < n && ws[k] != null; ++k)
671 >                    ;
672 >                if (k == n)
673 >                    ws = Arrays.copyOf(ws, n << 1);
674 >            }
675 >            ws[k] = w;
676 >            workers = ws; // volatile array write ensures slot visibility
677 >        } finally {
678 >            lock.unlock();
679 >        }
680 >        return k;
681 >    }
682  
683 <    // Constructors
683 >    /**
684 >     * Nulls out record of worker in workers array.
685 >     */
686 >    private void forgetWorker(ForkJoinWorkerThread w) {
687 >        int idx = w.poolIndex;
688 >        // Locking helps method recordWorker avoid unnecessary expansion
689 >        final ReentrantLock lock = this.workerLock;
690 >        lock.lock();
691 >        try {
692 >            ForkJoinWorkerThread[] ws = workers;
693 >            if (idx >= 0 && idx < ws.length && ws[idx] == w) // verify
694 >                ws[idx] = null;
695 >        } finally {
696 >            lock.unlock();
697 >        }
698 >    }
699  
700      /**
701 <     * Creates a ForkJoinPool with a pool size equal to the number of
702 <     * processors available on the system and using the default
703 <     * ForkJoinWorkerThreadFactory,
704 <     * @throws SecurityException if a security manager exists and
705 <     *         the caller is not permitted to modify threads
335 <     *         because it does not hold {@link
336 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
701 >     * Final callback from terminating worker.  Removes record of
702 >     * worker from array, and adjusts counts. If pool is shutting
703 >     * down, tries to complete termination.
704 >     *
705 >     * @param w the worker
706       */
707 <    public ForkJoinPool() {
708 <        this(Runtime.getRuntime().availableProcessors(),
709 <             defaultForkJoinWorkerThreadFactory);
707 >    final void workerTerminated(ForkJoinWorkerThread w) {
708 >        forgetWorker(w);
709 >        decrementWorkerCounts(w.isTrimmed()? 0 : ONE_RUNNING, ONE_TOTAL);
710 >        while (w.stealCount != 0) // collect final count
711 >            tryAccumulateStealCount(w);
712 >        tryTerminate(false);
713      }
714  
715 +    // Waiting for and signalling events
716 +
717      /**
718 <     * Creates a ForkJoinPool with the indicated parallelism level
719 <     * threads, and using the default ForkJoinWorkerThreadFactory,
720 <     * @param parallelism the number of worker threads
721 <     * @throws IllegalArgumentException if parallelism less than or
348 <     * equal to zero
349 <     * @throws SecurityException if a security manager exists and
350 <     *         the caller is not permitted to modify threads
351 <     *         because it does not hold {@link
352 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
718 >     * Releases workers blocked on a count not equal to current count.
719 >     * Normally called after precheck that eventWaiters isn't zero to
720 >     * avoid wasted array checks. Gives up upon a change in count or
721 >     * upon releasing two workers, letting others take over.
722       */
723 <    public ForkJoinPool(int parallelism) {
724 <        this(parallelism, defaultForkJoinWorkerThreadFactory);
723 >    private void releaseEventWaiters() {
724 >        ForkJoinWorkerThread[] ws = workers;
725 >        int n = ws.length;
726 >        long h = eventWaiters;
727 >        int ec = eventCount;
728 >        boolean releasedOne = false;
729 >        ForkJoinWorkerThread w; int id;
730 >        while ((id = ((int)(h & WAITER_ID_MASK)) - 1) >= 0 &&
731 >               (int)(h >>> EVENT_COUNT_SHIFT) != ec &&
732 >               id < n && (w = ws[id]) != null) {
733 >            if (UNSAFE.compareAndSwapLong(this, eventWaitersOffset,
734 >                                          h,  w.nextWaiter)) {
735 >                LockSupport.unpark(w);
736 >                if (releasedOne) // exit on second release
737 >                    break;
738 >                releasedOne = true;
739 >            }
740 >            if (eventCount != ec)
741 >                break;
742 >            h = eventWaiters;
743 >        }
744      }
745  
746      /**
747 <     * Creates a ForkJoinPool with parallelism equal to the number of
748 <     * processors available on the system and using the given
361 <     * ForkJoinWorkerThreadFactory,
362 <     * @param factory the factory for creating new threads
363 <     * @throws NullPointerException if factory is null
364 <     * @throws SecurityException if a security manager exists and
365 <     *         the caller is not permitted to modify threads
366 <     *         because it does not hold {@link
367 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
747 >     * Tries to advance eventCount and releases waiters. Called only
748 >     * from workers.
749       */
750 <    public ForkJoinPool(ForkJoinWorkerThreadFactory factory) {
751 <        this(Runtime.getRuntime().availableProcessors(), factory);
750 >    final void signalWork() {
751 >        int c; // try to increment event count -- CAS failure OK
752 >        UNSAFE.compareAndSwapInt(this, eventCountOffset, c = eventCount, c+1);
753 >        if (eventWaiters != 0L)
754 >            releaseEventWaiters();
755      }
756  
757      /**
758 <     * Creates a ForkJoinPool with the given parallelism and factory.
758 >     * Adds the given worker to event queue and blocks until
759 >     * terminating or event count advances from the given value
760       *
761 <     * @param parallelism the targeted number of worker threads
762 <     * @param factory the factory for creating new threads
378 <     * @throws IllegalArgumentException if parallelism less than or
379 <     * equal to zero, or greater than implementation limit.
380 <     * @throws NullPointerException if factory is null
381 <     * @throws SecurityException if a security manager exists and
382 <     *         the caller is not permitted to modify threads
383 <     *         because it does not hold {@link
384 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
761 >     * @param w the calling worker thread
762 >     * @param ec the count
763       */
764 <    public ForkJoinPool(int parallelism, ForkJoinWorkerThreadFactory factory) {
765 <        if (parallelism <= 0 || parallelism > MAX_THREADS)
766 <            throw new IllegalArgumentException();
767 <        if (factory == null)
768 <            throw new NullPointerException();
769 <        checkPermission();
770 <        this.factory = factory;
771 <        this.parallelism = parallelism;
772 <        this.maxPoolSize = MAX_THREADS;
773 <        this.maintainsParallelism = true;
774 <        this.poolNumber = poolNumberGenerator.incrementAndGet();
775 <        this.workerLock = new ReentrantLock();
776 <        this.termination = workerLock.newCondition();
399 <        this.stealCount = new AtomicLong();
400 <        this.submissionQueue = new LinkedTransferQueue<ForkJoinTask<?>>();
401 <        // worker array and workers are lazily constructed
764 >    private void eventSync(ForkJoinWorkerThread w, int ec) {
765 >        long nh = (((long)ec) << EVENT_COUNT_SHIFT) | ((long)(w.poolIndex+1));
766 >        long h;
767 >        while ((runState < SHUTDOWN || !tryTerminate(false)) &&
768 >               (((int)((h = eventWaiters) & WAITER_ID_MASK)) == 0 ||
769 >                (int)(h >>> EVENT_COUNT_SHIFT) == ec) &&
770 >               eventCount == ec) {
771 >            if (UNSAFE.compareAndSwapLong(this, eventWaitersOffset,
772 >                                          w.nextWaiter = h, nh)) {
773 >                awaitEvent(w, ec);
774 >                break;
775 >            }
776 >        }
777      }
778  
779      /**
780 <     * Create new worker using factory.
781 <     * @param index the index to assign worker
782 <     * @return new worker, or null of factory failed
780 >     * Blocks the given worker (that has already been entered as an
781 >     * event waiter) until terminating or event count advances from
782 >     * the given value. The oldest (first) waiter uses a timed wait to
783 >     * occasionally one-by-one shrink the number of workers (to a
784 >     * minimum of one) if the pool has not been used for extended
785 >     * periods.
786 >     *
787 >     * @param w the calling worker thread
788 >     * @param ec the count
789       */
790 <    private ForkJoinWorkerThread createWorker(int index) {
791 <        Thread.UncaughtExceptionHandler h = ueh;
792 <        ForkJoinWorkerThread w = factory.newThread(this);
793 <        if (w != null) {
794 <            w.poolIndex = index;
795 <            w.setDaemon(true);
796 <            w.setAsyncMode(locallyFifo);
797 <            w.setName("ForkJoinPool-" + poolNumber + "-worker-" + index);
798 <            if (h != null)
799 <                w.setUncaughtExceptionHandler(h);
790 >    private void awaitEvent(ForkJoinWorkerThread w, int ec) {
791 >        while (eventCount == ec) {
792 >            if (tryAccumulateStealCount(w)) { // transfer while idle
793 >                boolean untimed = (w.nextWaiter != 0L ||
794 >                                   (workerCounts & RUNNING_COUNT_MASK) <= 1);
795 >                long startTime = untimed? 0 : System.nanoTime();
796 >                Thread.interrupted();         // clear/ignore interrupt
797 >                if (eventCount != ec || w.runState != 0 ||
798 >                    runState >= TERMINATING)  // recheck after clear
799 >                    break;
800 >                if (untimed)
801 >                    LockSupport.park(w);
802 >                else {
803 >                    LockSupport.parkNanos(w, SHRINK_RATE_NANOS);
804 >                    if (eventCount != ec || w.runState != 0 ||
805 >                        runState >= TERMINATING)
806 >                        break;
807 >                    if (System.nanoTime() - startTime >= SHRINK_RATE_NANOS)
808 >                        tryShutdownUnusedWorker(ec);
809 >                }
810 >            }
811          }
420        return w;
812      }
813  
814 +    // Maintaining parallelism
815 +
816      /**
817 <     * Returns a good size for worker array given pool size.
425 <     * Currently requires size to be a power of two.
817 >     * Pushes worker onto the spare stack.
818       */
819 <    private static int arraySizeFor(int ps) {
820 <        return ps <= 1? 1 : (1 << (32 - Integer.numberOfLeadingZeros(ps-1)));
819 >    final void pushSpare(ForkJoinWorkerThread w) {
820 >        int ns = (++w.spareCount << SPARE_COUNT_SHIFT) | (w.poolIndex + 1);
821 >        do {} while (!UNSAFE.compareAndSwapInt(this, spareWaitersOffset,
822 >                                               w.nextSpare = spareWaiters,ns));
823      }
824  
825      /**
826 <     * Creates or resizes array if necessary to hold newLength.
827 <     * Call only under exclusion or lock.
434 <     * @return the array
826 >     * Tries (once) to resume a spare if the number of running
827 >     * threads is less than target.
828       */
829 <    private ForkJoinWorkerThread[] ensureWorkerArrayCapacity(int newLength) {
829 >    private void tryResumeSpare() {
830 >        int sw, id;
831          ForkJoinWorkerThread[] ws = workers;
832 <        if (ws == null)
833 <            return workers = new ForkJoinWorkerThread[arraySizeFor(newLength)];
834 <        else if (newLength > ws.length)
835 <            return workers = Arrays.copyOf(ws, arraySizeFor(newLength));
836 <        else
837 <            return ws;
832 >        int n = ws.length;
833 >        ForkJoinWorkerThread w;
834 >        if ((sw = spareWaiters) != 0 &&
835 >            (id = (sw & SPARE_ID_MASK) - 1) >= 0 &&
836 >            id < n && (w = ws[id]) != null &&
837 >            (workerCounts & RUNNING_COUNT_MASK) < parallelism &&
838 >            spareWaiters == sw &&
839 >            UNSAFE.compareAndSwapInt(this, spareWaitersOffset,
840 >                                     sw, w.nextSpare)) {
841 >            int c; // increment running count before resume
842 >            do {} while (!UNSAFE.compareAndSwapInt
843 >                         (this, workerCountsOffset,
844 >                          c = workerCounts, c + ONE_RUNNING));
845 >            if (w.tryUnsuspend())
846 >                LockSupport.unpark(w);
847 >            else   // back out if w was shutdown
848 >                decrementWorkerCounts(ONE_RUNNING, 0);
849 >        }
850      }
851  
852      /**
853 <     * Try to shrink workers into smaller array after one or more terminate
854 <     */
855 <    private void tryShrinkWorkerArray() {
856 <        ForkJoinWorkerThread[] ws = workers;
857 <        if (ws != null) {
858 <            int len = ws.length;
859 <            int last = len - 1;
860 <            while (last >= 0 && ws[last] == null)
861 <                --last;
862 <            int newLength = arraySizeFor(last+1);
863 <            if (newLength < len)
864 <                workers = Arrays.copyOf(ws, newLength);
853 >     * Tries to increase the number of running workers if below target
854 >     * parallelism: If a spare exists tries to resume it via
855 >     * tryResumeSpare.  Otherwise, if not enough total workers or all
856 >     * existing workers are busy, adds a new worker. In all cases also
857 >     * helps wake up releasable workers waiting for work.
858 >     */
859 >    private void helpMaintainParallelism() {
860 >        int pc = parallelism;
861 >        int wc, rs, tc;
862 >        while (((wc = workerCounts) & RUNNING_COUNT_MASK) < pc &&
863 >               (rs = runState) < TERMINATING) {
864 >            if (spareWaiters != 0)
865 >                tryResumeSpare();
866 >            else if ((tc = wc >>> TOTAL_COUNT_SHIFT) >= MAX_WORKERS ||
867 >                     (tc >= pc && (rs & ACTIVE_COUNT_MASK) != tc))
868 >                break;   // enough total
869 >            else if (runState == rs && workerCounts == wc &&
870 >                     UNSAFE.compareAndSwapInt(this, workerCountsOffset, wc,
871 >                                              wc + (ONE_RUNNING|ONE_TOTAL))) {
872 >                ForkJoinWorkerThread w = null;
873 >                try {
874 >                    w = factory.newThread(this);
875 >                } finally { // adjust on null or exceptional factory return
876 >                    if (w == null) {
877 >                        decrementWorkerCounts(ONE_RUNNING, ONE_TOTAL);
878 >                        tryTerminate(false); // handle failure during shutdown
879 >                    }
880 >                }
881 >                if (w == null)
882 >                    break;
883 >                w.start(recordWorker(w), ueh);
884 >                if ((workerCounts >>> TOTAL_COUNT_SHIFT) >= pc) {
885 >                    int c; // advance event count
886 >                    UNSAFE.compareAndSwapInt(this, eventCountOffset,
887 >                                             c = eventCount, c+1);
888 >                    break; // add at most one unless total below target
889 >                }
890 >            }
891          }
892 +        if (eventWaiters != 0L)
893 +            releaseEventWaiters();
894      }
895  
896      /**
897 <     * Initialize workers if necessary
897 >     * Callback from the oldest waiter in awaitEvent waking up after a
898 >     * period of non-use. If all workers are idle, tries (once) to
899 >     * shutdown an event waiter or a spare, if one exists. Note that
900 >     * we don't need CAS or locks here because the method is called
901 >     * only from one thread occasionally waking (and even misfires are
902 >     * OK). Note that until the shutdown worker fully terminates,
903 >     * workerCounts will overestimate total count, which is tolerable.
904 >     *
905 >     * @param ec the event count waited on by caller (to abort
906 >     * attempt if count has since changed).
907       */
908 <    final void ensureWorkerInitialization() {
909 <        ForkJoinWorkerThread[] ws = workers;
910 <        if (ws == null) {
911 <            final ReentrantLock lock = this.workerLock;
912 <            lock.lock();
913 <            try {
914 <                ws = workers;
915 <                if (ws == null) {
916 <                    int ps = parallelism;
917 <                    ws = ensureWorkerArrayCapacity(ps);
918 <                    for (int i = 0; i < ps; ++i) {
919 <                        ForkJoinWorkerThread w = createWorker(i);
920 <                        if (w != null) {
921 <                            ws[i] = w;
922 <                            w.start();
923 <                            updateWorkerCount(1);
924 <                        }
925 <                    }
908 >    private void tryShutdownUnusedWorker(int ec) {
909 >        if (runState == 0 && eventCount == ec) { // only trigger if all idle
910 >            ForkJoinWorkerThread[] ws = workers;
911 >            int n = ws.length;
912 >            ForkJoinWorkerThread w = null;
913 >            boolean shutdown = false;
914 >            int sw;
915 >            long h;
916 >            if ((sw = spareWaiters) != 0) { // prefer killing spares
917 >                int id = (sw & SPARE_ID_MASK) - 1;
918 >                if (id >= 0 && id < n && (w = ws[id]) != null &&
919 >                    UNSAFE.compareAndSwapInt(this, spareWaitersOffset,
920 >                                             sw, w.nextSpare))
921 >                    shutdown = true;
922 >            }
923 >            else if ((h = eventWaiters) != 0L) {
924 >                long nh;
925 >                int id = ((int)(h & WAITER_ID_MASK)) - 1;
926 >                if (id >= 0 && id < n && (w = ws[id]) != null &&
927 >                    (nh = w.nextWaiter) != 0L && // keep at least one worker
928 >                    UNSAFE.compareAndSwapLong(this, eventWaitersOffset, h, nh))
929 >                    shutdown = true;
930 >            }
931 >            if (w != null && shutdown) {
932 >                w.shutdown();
933 >                LockSupport.unpark(w);
934 >            }
935 >        }
936 >        releaseEventWaiters(); // in case of interference
937 >    }
938 >
939 >    /**
940 >     * Callback from workers invoked upon each top-level action (i.e.,
941 >     * stealing a task or taking a submission and running it).
942 >     * Performs one or more of the following:
943 >     *
944 >     * 1. If the worker is active and either did not run a task
945 >     *    or there are too many workers, try to set its active status
946 >     *    to inactive and update activeCount. On contention, we may
947 >     *    try again in this or a subsequent call.
948 >     *
949 >     * 2. If not enough total workers, help create some.
950 >     *
951 >     * 3. If there are too many running workers, suspend this worker
952 >     *    (first forcing inactive if necessary).  If it is not needed,
953 >     *    it may be shutdown while suspended (via
954 >     *    tryShutdownUnusedWorker).  Otherwise, upon resume it
955 >     *    rechecks running thread count and need for event sync.
956 >     *
957 >     * 4. If worker did not run a task, await the next task event via
958 >     *    eventSync if necessary (first forcing inactivation), upon
959 >     *    which the worker may be shutdown via
960 >     *    tryShutdownUnusedWorker.  Otherwise, help release any
961 >     *    existing event waiters that are now releasable,
962 >     *
963 >     * @param w the worker
964 >     * @param ran true if worker ran a task since last call to this method
965 >     */
966 >    final void preStep(ForkJoinWorkerThread w, boolean ran) {
967 >        int wec = w.lastEventCount;
968 >        boolean active = w.active;
969 >        boolean inactivate = false;
970 >        int pc = parallelism;
971 >        int rs;
972 >        while (w.runState == 0 && (rs = runState) < TERMINATING) {
973 >            if ((inactivate || (active && (rs & ACTIVE_COUNT_MASK) >= pc)) &&
974 >                UNSAFE.compareAndSwapInt(this, runStateOffset, rs, rs - 1))
975 >                inactivate = active = w.active = false;
976 >            int wc = workerCounts;
977 >            if ((wc & RUNNING_COUNT_MASK) > pc) {
978 >                if (!(inactivate |= active) && // must inactivate to suspend
979 >                    workerCounts == wc &&      // try to suspend as spare
980 >                    UNSAFE.compareAndSwapInt(this, workerCountsOffset,
981 >                                             wc, wc - ONE_RUNNING))
982 >                    w.suspendAsSpare();
983 >            }
984 >            else if ((wc >>> TOTAL_COUNT_SHIFT) < pc)
985 >                helpMaintainParallelism();     // not enough workers
986 >            else if (!ran) {
987 >                long h = eventWaiters;
988 >                int ec = eventCount;
989 >                if (h != 0L && (int)(h >>> EVENT_COUNT_SHIFT) != ec)
990 >                    releaseEventWaiters();     // release others before waiting
991 >                else if (ec != wec) {
992 >                    w.lastEventCount = ec;     // no need to wait
993 >                    break;
994                  }
995 <            } finally {
996 <                lock.unlock();
995 >                else if (!(inactivate |= active))
996 >                    eventSync(w, wec);         // must inactivate before sync
997              }
998 +            else
999 +                break;
1000          }
1001      }
1002  
1003      /**
1004 <     * Worker creation and startup for threads added via setParallelism.
1004 >     * Helps and/or blocks awaiting join of the given task.
1005 >     * See above for explanation.
1006 >     *
1007 >     * @param joinMe the task to join
1008 >     * @param worker the current worker thread
1009       */
1010 <    private void createAndStartAddedWorkers() {
1011 <        resumeAllSpares();  // Allow spares to convert to nonspare
1012 <        int ps = parallelism;
1013 <        ForkJoinWorkerThread[] ws = ensureWorkerArrayCapacity(ps);
1014 <        int len = ws.length;
1015 <        // Sweep through slots, to keep lowest indices most populated
1016 <        int k = 0;
1017 <        while (k < len) {
1018 <            if (ws[k] != null) {
1019 <                ++k;
1020 <                continue;
1010 >    final void awaitJoin(ForkJoinTask<?> joinMe, ForkJoinWorkerThread worker) {
1011 >        int retries = 2 + (parallelism >> 2); // #helpJoins before blocking
1012 >        while (joinMe.status >= 0) {
1013 >            int wc;
1014 >            worker.helpJoinTask(joinMe);
1015 >            if (joinMe.status < 0)
1016 >                break;
1017 >            else if (retries > 0)
1018 >                --retries;
1019 >            else if (((wc = workerCounts) & RUNNING_COUNT_MASK) != 0 &&
1020 >                     UNSAFE.compareAndSwapInt(this, workerCountsOffset,
1021 >                                              wc, wc - ONE_RUNNING)) {
1022 >                int stat, c; long h;
1023 >                while ((stat = joinMe.status) >= 0 &&
1024 >                       (h = eventWaiters) != 0L && // help release others
1025 >                       (int)(h >>> EVENT_COUNT_SHIFT) != eventCount)
1026 >                    releaseEventWaiters();
1027 >                if (stat >= 0 &&
1028 >                    ((workerCounts & RUNNING_COUNT_MASK) == 0 ||
1029 >                     (stat =
1030 >                      joinMe.internalAwaitDone(JOIN_TIMEOUT_MILLIS)) >= 0))
1031 >                    helpMaintainParallelism(); // timeout or no running workers
1032 >                do {} while (!UNSAFE.compareAndSwapInt
1033 >                             (this, workerCountsOffset,
1034 >                              c = workerCounts, c + ONE_RUNNING));
1035 >                if (stat < 0)
1036 >                    break;   // else restart
1037              }
1038 <            int s = workerCounts;
1039 <            int tc = totalCountOf(s);
1040 <            int rc = runningCountOf(s);
1041 <            if (rc >= ps || tc >= ps)
1038 >        }
1039 >    }
1040 >
1041 >    /**
1042 >     * Same idea as awaitJoin, but no helping, retries, or timeouts.
1043 >     */
1044 >    final void awaitBlocker(ManagedBlocker blocker)
1045 >        throws InterruptedException {
1046 >        while (!blocker.isReleasable()) {
1047 >            int wc = workerCounts;
1048 >            if ((wc & RUNNING_COUNT_MASK) != 0 &&
1049 >                UNSAFE.compareAndSwapInt(this, workerCountsOffset,
1050 >                                         wc, wc - ONE_RUNNING)) {
1051 >                try {
1052 >                    while (!blocker.isReleasable()) {
1053 >                        long h = eventWaiters;
1054 >                        if (h != 0L &&
1055 >                            (int)(h >>> EVENT_COUNT_SHIFT) != eventCount)
1056 >                            releaseEventWaiters();
1057 >                        else if ((workerCounts & RUNNING_COUNT_MASK) == 0 &&
1058 >                                 runState < TERMINATING)
1059 >                            helpMaintainParallelism();
1060 >                        else if (blocker.block())
1061 >                            break;
1062 >                    }
1063 >                } finally {
1064 >                    int c;
1065 >                    do {} while (!UNSAFE.compareAndSwapInt
1066 >                                 (this, workerCountsOffset,
1067 >                                  c = workerCounts, c + ONE_RUNNING));
1068 >                }
1069                  break;
1070 <            if (casWorkerCounts (s, workerCountsFor(tc+1, rc+1))) {
1071 <                ForkJoinWorkerThread w = createWorker(k);
1070 >            }
1071 >        }
1072 >    }
1073 >
1074 >    /**
1075 >     * Possibly initiates and/or completes termination.
1076 >     *
1077 >     * @param now if true, unconditionally terminate, else only
1078 >     * if shutdown and empty queue and no active workers
1079 >     * @return true if now terminating or terminated
1080 >     */
1081 >    private boolean tryTerminate(boolean now) {
1082 >        if (now)
1083 >            advanceRunLevel(SHUTDOWN); // ensure at least SHUTDOWN
1084 >        else if (runState < SHUTDOWN ||
1085 >                 !submissionQueue.isEmpty() ||
1086 >                 (runState & ACTIVE_COUNT_MASK) != 0)
1087 >            return false;
1088 >
1089 >        if (advanceRunLevel(TERMINATING))
1090 >            startTerminating();
1091 >
1092 >        // Finish now if all threads terminated; else in some subsequent call
1093 >        if ((workerCounts >>> TOTAL_COUNT_SHIFT) == 0) {
1094 >            advanceRunLevel(TERMINATED);
1095 >            termination.arrive();
1096 >        }
1097 >        return true;
1098 >    }
1099 >
1100 >    /**
1101 >     * Actions on transition to TERMINATING
1102 >     *
1103 >     * Runs up to four passes through workers: (0) shutting down each
1104 >     * (without waking up if parked) to quickly spread notifications
1105 >     * without unnecessary bouncing around event queues etc (1) wake
1106 >     * up and help cancel tasks (2) interrupt (3) mop up races with
1107 >     * interrupted workers
1108 >     */
1109 >    private void startTerminating() {
1110 >        cancelSubmissions();
1111 >        for (int passes = 0; passes < 4 && workerCounts != 0; ++passes) {
1112 >            int c; // advance event count
1113 >            UNSAFE.compareAndSwapInt(this, eventCountOffset,
1114 >                                     c = eventCount, c+1);
1115 >            eventWaiters = 0L; // clobber lists
1116 >            spareWaiters = 0;
1117 >            for (ForkJoinWorkerThread w : workers) {
1118                  if (w != null) {
1119 <                    ws[k++] = w;
1120 <                    w.start();
1121 <                }
1122 <                else {
1123 <                    updateWorkerCount(-1); // back out on failed creation
1124 <                    break;
1119 >                    w.shutdown();
1120 >                    if (passes > 0 && !w.isTerminated()) {
1121 >                        w.cancelTasks();
1122 >                        LockSupport.unpark(w);
1123 >                        if (passes > 1) {
1124 >                            try {
1125 >                                w.interrupt();
1126 >                            } catch (SecurityException ignore) {
1127 >                            }
1128 >                        }
1129 >                    }
1130                  }
1131              }
1132          }
1133      }
1134  
1135 +    /**
1136 +     * Clears out and cancels submissions, ignoring exceptions.
1137 +     */
1138 +    private void cancelSubmissions() {
1139 +        ForkJoinTask<?> task;
1140 +        while ((task = submissionQueue.poll()) != null) {
1141 +            try {
1142 +                task.cancel(false);
1143 +            } catch (Throwable ignore) {
1144 +            }
1145 +        }
1146 +    }
1147 +
1148 +    // misc support for ForkJoinWorkerThread
1149 +
1150 +    /**
1151 +     * Returns pool number.
1152 +     */
1153 +    final int getPoolNumber() {
1154 +        return poolNumber;
1155 +    }
1156 +
1157 +    /**
1158 +     * Tries to accumulate steal count from a worker, clearing
1159 +     * the worker's value if successful.
1160 +     *
1161 +     * @return true if worker steal count now zero
1162 +     */
1163 +    final boolean tryAccumulateStealCount(ForkJoinWorkerThread w) {
1164 +        int sc = w.stealCount;
1165 +        long c = stealCount;
1166 +        // CAS even if zero, for fence effects
1167 +        if (UNSAFE.compareAndSwapLong(this, stealCountOffset, c, c + sc)) {
1168 +            if (sc != 0)
1169 +                w.stealCount = 0;
1170 +            return true;
1171 +        }
1172 +        return sc == 0;
1173 +    }
1174 +
1175 +    /**
1176 +     * Returns the approximate (non-atomic) number of idle threads per
1177 +     * active thread.
1178 +     */
1179 +    final int idlePerActive() {
1180 +        int pc = parallelism; // use parallelism, not rc
1181 +        int ac = runState;    // no mask -- artificially boosts during shutdown
1182 +        // Use exact results for small values, saturate past 4
1183 +        return ((pc <= ac) ? 0 :
1184 +                (pc >>> 1 <= ac) ? 1 :
1185 +                (pc >>> 2 <= ac) ? 3 :
1186 +                pc >>> 3);
1187 +    }
1188 +
1189 +    // Public and protected methods
1190 +
1191 +    // Constructors
1192 +
1193 +    /**
1194 +     * Creates a {@code ForkJoinPool} with parallelism equal to {@link
1195 +     * java.lang.Runtime#availableProcessors}, using the {@linkplain
1196 +     * #defaultForkJoinWorkerThreadFactory default thread factory},
1197 +     * no UncaughtExceptionHandler, and non-async LIFO processing mode.
1198 +     *
1199 +     * @throws SecurityException if a security manager exists and
1200 +     *         the caller is not permitted to modify threads
1201 +     *         because it does not hold {@link
1202 +     *         java.lang.RuntimePermission}{@code ("modifyThread")}
1203 +     */
1204 +    public ForkJoinPool() {
1205 +        this(Runtime.getRuntime().availableProcessors(),
1206 +             defaultForkJoinWorkerThreadFactory, null, false);
1207 +    }
1208 +
1209 +    /**
1210 +     * Creates a {@code ForkJoinPool} with the indicated parallelism
1211 +     * level, the {@linkplain
1212 +     * #defaultForkJoinWorkerThreadFactory default thread factory},
1213 +     * no UncaughtExceptionHandler, and non-async LIFO processing mode.
1214 +     *
1215 +     * @param parallelism the parallelism level
1216 +     * @throws IllegalArgumentException if parallelism less than or
1217 +     *         equal to zero, or greater than implementation limit
1218 +     * @throws SecurityException if a security manager exists and
1219 +     *         the caller is not permitted to modify threads
1220 +     *         because it does not hold {@link
1221 +     *         java.lang.RuntimePermission}{@code ("modifyThread")}
1222 +     */
1223 +    public ForkJoinPool(int parallelism) {
1224 +        this(parallelism, defaultForkJoinWorkerThreadFactory, null, false);
1225 +    }
1226 +
1227 +    /**
1228 +     * Creates a {@code ForkJoinPool} with the given parameters.
1229 +     *
1230 +     * @param parallelism the parallelism level. For default value,
1231 +     * use {@link java.lang.Runtime#availableProcessors}.
1232 +     * @param factory the factory for creating new threads. For default value,
1233 +     * use {@link #defaultForkJoinWorkerThreadFactory}.
1234 +     * @param handler the handler for internal worker threads that
1235 +     * terminate due to unrecoverable errors encountered while executing
1236 +     * tasks. For default value, use {@code null}.
1237 +     * @param asyncMode if true,
1238 +     * establishes local first-in-first-out scheduling mode for forked
1239 +     * tasks that are never joined. This mode may be more appropriate
1240 +     * than default locally stack-based mode in applications in which
1241 +     * worker threads only process event-style asynchronous tasks.
1242 +     * For default value, use {@code false}.
1243 +     * @throws IllegalArgumentException if parallelism less than or
1244 +     *         equal to zero, or greater than implementation limit
1245 +     * @throws NullPointerException if the factory is null
1246 +     * @throws SecurityException if a security manager exists and
1247 +     *         the caller is not permitted to modify threads
1248 +     *         because it does not hold {@link
1249 +     *         java.lang.RuntimePermission}{@code ("modifyThread")}
1250 +     */
1251 +    public ForkJoinPool(int parallelism,
1252 +                        ForkJoinWorkerThreadFactory factory,
1253 +                        Thread.UncaughtExceptionHandler handler,
1254 +                        boolean asyncMode) {
1255 +        checkPermission();
1256 +        if (factory == null)
1257 +            throw new NullPointerException();
1258 +        if (parallelism <= 0 || parallelism > MAX_WORKERS)
1259 +            throw new IllegalArgumentException();
1260 +        this.parallelism = parallelism;
1261 +        this.factory = factory;
1262 +        this.ueh = handler;
1263 +        this.locallyFifo = asyncMode;
1264 +        int arraySize = initialArraySizeFor(parallelism);
1265 +        this.workers = new ForkJoinWorkerThread[arraySize];
1266 +        this.submissionQueue = new LinkedTransferQueue<ForkJoinTask<?>>();
1267 +        this.workerLock = new ReentrantLock();
1268 +        this.termination = new Phaser(1);
1269 +        this.poolNumber = poolNumberGenerator.incrementAndGet();
1270 +    }
1271 +
1272 +    /**
1273 +     * Returns initial power of two size for workers array.
1274 +     * @param pc the initial parallelism level
1275 +     */
1276 +    private static int initialArraySizeFor(int pc) {
1277 +        // If possible, initially allocate enough space for one spare
1278 +        int size = pc < MAX_WORKERS ? pc + 1 : MAX_WORKERS;
1279 +        // See Hackers Delight, sec 3.2. We know MAX_WORKERS < (1 >>> 16)
1280 +        size |= size >>> 1;
1281 +        size |= size >>> 2;
1282 +        size |= size >>> 4;
1283 +        size |= size >>> 8;
1284 +        return size + 1;
1285 +    }
1286 +
1287      // Execution methods
1288  
1289      /**
1290       * Common code for execute, invoke and submit
1291       */
1292      private <T> void doSubmit(ForkJoinTask<T> task) {
1293 <        if (isShutdown())
1293 >        if (task == null)
1294 >            throw new NullPointerException();
1295 >        if (runState >= SHUTDOWN)
1296              throw new RejectedExecutionException();
532        if (workers == null)
533            ensureWorkerInitialization();
1297          submissionQueue.offer(task);
1298 <        signalIdleWorkers();
1298 >        int c; // try to increment event count -- CAS failure OK
1299 >        UNSAFE.compareAndSwapInt(this, eventCountOffset, c = eventCount, c+1);
1300 >        helpMaintainParallelism(); // create, start, or resume some workers
1301      }
1302  
1303      /**
1304 <     * Performs the given task; returning its result upon completion
1304 >     * Performs the given task, returning its result upon completion.
1305 >     *
1306       * @param task the task
1307       * @return the task's result
1308 <     * @throws NullPointerException if task is null
1309 <     * @throws RejectedExecutionException if pool is shut down
1308 >     * @throws NullPointerException if the task is null
1309 >     * @throws RejectedExecutionException if the task cannot be
1310 >     *         scheduled for execution
1311       */
1312      public <T> T invoke(ForkJoinTask<T> task) {
1313          doSubmit(task);
# Line 549 | Line 1316 | public class ForkJoinPool extends Abstra
1316  
1317      /**
1318       * Arranges for (asynchronous) execution of the given task.
1319 +     *
1320       * @param task the task
1321 <     * @throws NullPointerException if task is null
1322 <     * @throws RejectedExecutionException if pool is shut down
1321 >     * @throws NullPointerException if the task is null
1322 >     * @throws RejectedExecutionException if the task cannot be
1323 >     *         scheduled for execution
1324       */
1325 <    public <T> void execute(ForkJoinTask<T> task) {
1325 >    public void execute(ForkJoinTask<?> task) {
1326          doSubmit(task);
1327      }
1328  
1329      // AbstractExecutorService methods
1330  
1331 +    /**
1332 +     * @throws NullPointerException if the task is null
1333 +     * @throws RejectedExecutionException if the task cannot be
1334 +     *         scheduled for execution
1335 +     */
1336      public void execute(Runnable task) {
1337 <        doSubmit(new AdaptedRunnable<Void>(task, null));
1337 >        ForkJoinTask<?> job;
1338 >        if (task instanceof ForkJoinTask<?>) // avoid re-wrap
1339 >            job = (ForkJoinTask<?>) task;
1340 >        else
1341 >            job = ForkJoinTask.adapt(task, null);
1342 >        doSubmit(job);
1343 >    }
1344 >
1345 >    /**
1346 >     * Submits a ForkJoinTask for execution.
1347 >     *
1348 >     * @param task the task to submit
1349 >     * @return the task
1350 >     * @throws NullPointerException if the task is null
1351 >     * @throws RejectedExecutionException if the task cannot be
1352 >     *         scheduled for execution
1353 >     */
1354 >    public <T> ForkJoinTask<T> submit(ForkJoinTask<T> task) {
1355 >        doSubmit(task);
1356 >        return task;
1357      }
1358  
1359 +    /**
1360 +     * @throws NullPointerException if the task is null
1361 +     * @throws RejectedExecutionException if the task cannot be
1362 +     *         scheduled for execution
1363 +     */
1364      public <T> ForkJoinTask<T> submit(Callable<T> task) {
1365 <        ForkJoinTask<T> job = new AdaptedCallable<T>(task);
1365 >        ForkJoinTask<T> job = ForkJoinTask.adapt(task);
1366          doSubmit(job);
1367          return job;
1368      }
1369  
1370 +    /**
1371 +     * @throws NullPointerException if the task is null
1372 +     * @throws RejectedExecutionException if the task cannot be
1373 +     *         scheduled for execution
1374 +     */
1375      public <T> ForkJoinTask<T> submit(Runnable task, T result) {
1376 <        ForkJoinTask<T> job = new AdaptedRunnable<T>(task, result);
1376 >        ForkJoinTask<T> job = ForkJoinTask.adapt(task, result);
1377          doSubmit(job);
1378          return job;
1379      }
1380  
1381 +    /**
1382 +     * @throws NullPointerException if the task is null
1383 +     * @throws RejectedExecutionException if the task cannot be
1384 +     *         scheduled for execution
1385 +     */
1386      public ForkJoinTask<?> submit(Runnable task) {
1387 <        ForkJoinTask<Void> job = new AdaptedRunnable<Void>(task, null);
1387 >        ForkJoinTask<?> job;
1388 >        if (task instanceof ForkJoinTask<?>) // avoid re-wrap
1389 >            job = (ForkJoinTask<?>) task;
1390 >        else
1391 >            job = ForkJoinTask.adapt(task, null);
1392          doSubmit(job);
1393          return job;
1394      }
1395  
1396      /**
1397 <     * Adaptor for Runnables. This implements RunnableFuture
1398 <     * to be compliant with AbstractExecutorService constraints
1397 >     * @throws NullPointerException       {@inheritDoc}
1398 >     * @throws RejectedExecutionException {@inheritDoc}
1399       */
588    static final class AdaptedRunnable<T> extends ForkJoinTask<T>
589        implements RunnableFuture<T> {
590        final Runnable runnable;
591        final T resultOnCompletion;
592        T result;
593        AdaptedRunnable(Runnable runnable, T result) {
594            if (runnable == null) throw new NullPointerException();
595            this.runnable = runnable;
596            this.resultOnCompletion = result;
597        }
598        public T getRawResult() { return result; }
599        public void setRawResult(T v) { result = v; }
600        public boolean exec() {
601            runnable.run();
602            result = resultOnCompletion;
603            return true;
604        }
605        public void run() { invoke(); }
606    }
607
608    /**
609     * Adaptor for Callables
610     */
611    static final class AdaptedCallable<T> extends ForkJoinTask<T>
612        implements RunnableFuture<T> {
613        final Callable<T> callable;
614        T result;
615        AdaptedCallable(Callable<T> callable) {
616            if (callable == null) throw new NullPointerException();
617            this.callable = callable;
618        }
619        public T getRawResult() { return result; }
620        public void setRawResult(T v) { result = v; }
621        public boolean exec() {
622            try {
623                result = callable.call();
624                return true;
625            } catch (Error err) {
626                throw err;
627            } catch (RuntimeException rex) {
628                throw rex;
629            } catch (Exception ex) {
630                throw new RuntimeException(ex);
631            }
632        }
633        public void run() { invoke(); }
634    }
635
1400      public <T> List<Future<T>> invokeAll(Collection<? extends Callable<T>> tasks) {
1401 <        ArrayList<ForkJoinTask<T>> ts =
1401 >        ArrayList<ForkJoinTask<T>> forkJoinTasks =
1402              new ArrayList<ForkJoinTask<T>>(tasks.size());
1403 <        for (Callable<T> c : tasks)
1404 <            ts.add(new AdaptedCallable<T>(c));
1405 <        invoke(new InvokeAll<T>(ts));
1406 <        return (List<Future<T>>)(List)ts;
1403 >        for (Callable<T> task : tasks)
1404 >            forkJoinTasks.add(ForkJoinTask.adapt(task));
1405 >        invoke(new InvokeAll<T>(forkJoinTasks));
1406 >
1407 >        @SuppressWarnings({"unchecked", "rawtypes"})
1408 >            List<Future<T>> futures = (List<Future<T>>) (List) forkJoinTasks;
1409 >        return futures;
1410      }
1411  
1412      static final class InvokeAll<T> extends RecursiveAction {
1413          final ArrayList<ForkJoinTask<T>> tasks;
1414          InvokeAll(ArrayList<ForkJoinTask<T>> tasks) { this.tasks = tasks; }
1415          public void compute() {
1416 <            try { invokeAll(tasks); } catch(Exception ignore) {}
1416 >            try { invokeAll(tasks); }
1417 >            catch (Exception ignore) {}
1418          }
1419 +        private static final long serialVersionUID = -7914297376763021607L;
1420      }
1421  
653    // Configuration and status settings and queries
654
1422      /**
1423 <     * Returns the factory used for constructing new workers
1423 >     * Returns the factory used for constructing new workers.
1424       *
1425       * @return the factory used for constructing new workers
1426       */
# Line 664 | Line 1431 | public class ForkJoinPool extends Abstra
1431      /**
1432       * Returns the handler for internal worker threads that terminate
1433       * due to unrecoverable errors encountered while executing tasks.
667     * @return the handler, or null if none
668     */
669    public Thread.UncaughtExceptionHandler getUncaughtExceptionHandler() {
670        Thread.UncaughtExceptionHandler h;
671        final ReentrantLock lock = this.workerLock;
672        lock.lock();
673        try {
674            h = ueh;
675        } finally {
676            lock.unlock();
677        }
678        return h;
679    }
680
681    /**
682     * Sets the handler for internal worker threads that terminate due
683     * to unrecoverable errors encountered while executing tasks.
684     * Unless set, the current default or ThreadGroup handler is used
685     * as handler.
1434       *
1435 <     * @param h the new handler
688 <     * @return the old handler, or null if none
689 <     * @throws SecurityException if a security manager exists and
690 <     *         the caller is not permitted to modify threads
691 <     *         because it does not hold {@link
692 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
693 <     */
694 <    public Thread.UncaughtExceptionHandler
695 <        setUncaughtExceptionHandler(Thread.UncaughtExceptionHandler h) {
696 <        checkPermission();
697 <        Thread.UncaughtExceptionHandler old = null;
698 <        final ReentrantLock lock = this.workerLock;
699 <        lock.lock();
700 <        try {
701 <            old = ueh;
702 <            ueh = h;
703 <            ForkJoinWorkerThread[] ws = workers;
704 <            if (ws != null) {
705 <                for (int i = 0; i < ws.length; ++i) {
706 <                    ForkJoinWorkerThread w = ws[i];
707 <                    if (w != null)
708 <                        w.setUncaughtExceptionHandler(h);
709 <                }
710 <            }
711 <        } finally {
712 <            lock.unlock();
713 <        }
714 <        return old;
715 <    }
716 <
717 <
718 <    /**
719 <     * Sets the target parallelism level of this pool.
720 <     * @param parallelism the target parallelism
721 <     * @throws IllegalArgumentException if parallelism less than or
722 <     * equal to zero or greater than maximum size bounds.
723 <     * @throws SecurityException if a security manager exists and
724 <     *         the caller is not permitted to modify threads
725 <     *         because it does not hold {@link
726 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
1435 >     * @return the handler, or {@code null} if none
1436       */
1437 <    public void setParallelism(int parallelism) {
1438 <        checkPermission();
730 <        if (parallelism <= 0 || parallelism > maxPoolSize)
731 <            throw new IllegalArgumentException();
732 <        final ReentrantLock lock = this.workerLock;
733 <        lock.lock();
734 <        try {
735 <            if (!isTerminating()) {
736 <                int p = this.parallelism;
737 <                this.parallelism = parallelism;
738 <                if (parallelism > p)
739 <                    createAndStartAddedWorkers();
740 <                else
741 <                    trimSpares();
742 <            }
743 <        } finally {
744 <            lock.unlock();
745 <        }
746 <        signalIdleWorkers();
1437 >    public Thread.UncaughtExceptionHandler getUncaughtExceptionHandler() {
1438 >        return ueh;
1439      }
1440  
1441      /**
1442 <     * Returns the targeted number of worker threads in this pool.
1442 >     * Returns the targeted parallelism level of this pool.
1443       *
1444 <     * @return the targeted number of worker threads in this pool
1444 >     * @return the targeted parallelism level of this pool
1445       */
1446      public int getParallelism() {
1447          return parallelism;
# Line 757 | Line 1449 | public class ForkJoinPool extends Abstra
1449  
1450      /**
1451       * Returns the number of worker threads that have started but not
1452 <     * yet terminated.  This result returned by this method may differ
1453 <     * from {@code getParallelism} when threads are created to
1452 >     * yet terminated.  The result returned by this method may differ
1453 >     * from {@link #getParallelism} when threads are created to
1454       * maintain parallelism when others are cooperatively blocked.
1455       *
1456       * @return the number of worker threads
1457       */
1458      public int getPoolSize() {
1459 <        return totalCountOf(workerCounts);
768 <    }
769 <
770 <    /**
771 <     * Returns the maximum number of threads allowed to exist in the
772 <     * pool, even if there are insufficient unblocked running threads.
773 <     * @return the maximum
774 <     */
775 <    public int getMaximumPoolSize() {
776 <        return maxPoolSize;
777 <    }
778 <
779 <    /**
780 <     * Sets the maximum number of threads allowed to exist in the
781 <     * pool, even if there are insufficient unblocked running threads.
782 <     * Setting this value has no effect on current pool size. It
783 <     * controls construction of new threads.
784 <     * @throws IllegalArgumentException if negative or greater then
785 <     * internal implementation limit.
786 <     */
787 <    public void setMaximumPoolSize(int newMax) {
788 <        if (newMax < 0 || newMax > MAX_THREADS)
789 <            throw new IllegalArgumentException();
790 <        maxPoolSize = newMax;
791 <    }
792 <
793 <
794 <    /**
795 <     * Returns true if this pool dynamically maintains its target
796 <     * parallelism level. If false, new threads are added only to
797 <     * avoid possible starvation.
798 <     * This setting is by default true;
799 <     * @return true if maintains parallelism
800 <     */
801 <    public boolean getMaintainsParallelism() {
802 <        return maintainsParallelism;
803 <    }
804 <
805 <    /**
806 <     * Sets whether this pool dynamically maintains its target
807 <     * parallelism level. If false, new threads are added only to
808 <     * avoid possible starvation.
809 <     * @param enable true to maintains parallelism
810 <     */
811 <    public void setMaintainsParallelism(boolean enable) {
812 <        maintainsParallelism = enable;
813 <    }
814 <
815 <    /**
816 <     * Establishes local first-in-first-out scheduling mode for forked
817 <     * tasks that are never joined. This mode may be more appropriate
818 <     * than default locally stack-based mode in applications in which
819 <     * worker threads only process asynchronous tasks.  This method is
820 <     * designed to be invoked only when pool is quiescent, and
821 <     * typically only before any tasks are submitted. The effects of
822 <     * invocations at other times may be unpredictable.
823 <     *
824 <     * @param async if true, use locally FIFO scheduling
825 <     * @return the previous mode.
826 <     */
827 <    public boolean setAsyncMode(boolean async) {
828 <        boolean oldMode = locallyFifo;
829 <        locallyFifo = async;
830 <        ForkJoinWorkerThread[] ws = workers;
831 <        if (ws != null) {
832 <            for (int i = 0; i < ws.length; ++i) {
833 <                ForkJoinWorkerThread t = ws[i];
834 <                if (t != null)
835 <                    t.setAsyncMode(async);
836 <            }
837 <        }
838 <        return oldMode;
1459 >        return workerCounts >>> TOTAL_COUNT_SHIFT;
1460      }
1461  
1462      /**
1463 <     * Returns true if this pool uses local first-in-first-out
1463 >     * Returns {@code true} if this pool uses local first-in-first-out
1464       * scheduling mode for forked tasks that are never joined.
1465       *
1466 <     * @return true if this pool uses async mode.
1466 >     * @return {@code true} if this pool uses async mode
1467       */
1468      public boolean getAsyncMode() {
1469          return locallyFifo;
# Line 851 | Line 1472 | public class ForkJoinPool extends Abstra
1472      /**
1473       * Returns an estimate of the number of worker threads that are
1474       * not blocked waiting to join tasks or for other managed
1475 <     * synchronization.
1475 >     * synchronization. This method may overestimate the
1476 >     * number of running threads.
1477       *
1478       * @return the number of worker threads
1479       */
1480      public int getRunningThreadCount() {
1481 <        return runningCountOf(workerCounts);
1481 >        return workerCounts & RUNNING_COUNT_MASK;
1482      }
1483  
1484      /**
1485       * Returns an estimate of the number of threads that are currently
1486       * stealing or executing tasks. This method may overestimate the
1487       * number of active threads.
1488 <     * @return the number of active threads.
1488 >     *
1489 >     * @return the number of active threads
1490       */
1491      public int getActiveThreadCount() {
1492 <        return activeCountOf(runControl);
870 <    }
871 <
872 <    /**
873 <     * Returns an estimate of the number of threads that are currently
874 <     * idle waiting for tasks. This method may underestimate the
875 <     * number of idle threads.
876 <     * @return the number of idle threads.
877 <     */
878 <    final int getIdleThreadCount() {
879 <        int c = runningCountOf(workerCounts) - activeCountOf(runControl);
880 <        return (c <= 0)? 0 : c;
1492 >        return runState & ACTIVE_COUNT_MASK;
1493      }
1494  
1495      /**
1496 <     * Returns true if all worker threads are currently idle. An idle
1497 <     * worker is one that cannot obtain a task to execute because none
1498 <     * are available to steal from other threads, and there are no
1499 <     * pending submissions to the pool. This method is conservative:
1500 <     * It might not return true immediately upon idleness of all
1501 <     * threads, but will eventually become true if threads remain
1502 <     * inactive.
1503 <     * @return true if all threads are currently idle
1496 >     * Returns {@code true} if all worker threads are currently idle.
1497 >     * An idle worker is one that cannot obtain a task to execute
1498 >     * because none are available to steal from other threads, and
1499 >     * there are no pending submissions to the pool. This method is
1500 >     * conservative; it might not return {@code true} immediately upon
1501 >     * idleness of all threads, but will eventually become true if
1502 >     * threads remain inactive.
1503 >     *
1504 >     * @return {@code true} if all threads are currently idle
1505       */
1506      public boolean isQuiescent() {
1507 <        return activeCountOf(runControl) == 0;
1507 >        return (runState & ACTIVE_COUNT_MASK) == 0;
1508      }
1509  
1510      /**
# Line 899 | Line 1512 | public class ForkJoinPool extends Abstra
1512       * one thread's work queue by another. The reported value
1513       * underestimates the actual total number of steals when the pool
1514       * is not quiescent. This value may be useful for monitoring and
1515 <     * tuning fork/join programs: In general, steal counts should be
1515 >     * tuning fork/join programs: in general, steal counts should be
1516       * high enough to keep threads busy, but low enough to avoid
1517       * overhead and contention across threads.
1518 <     * @return the number of steals.
1518 >     *
1519 >     * @return the number of steals
1520       */
1521      public long getStealCount() {
1522 <        return stealCount.get();
909 <    }
910 <
911 <    /**
912 <     * Accumulate steal count from a worker. Call only
913 <     * when worker known to be idle.
914 <     */
915 <    private void updateStealCount(ForkJoinWorkerThread w) {
916 <        int sc = w.getAndClearStealCount();
917 <        if (sc != 0)
918 <            stealCount.addAndGet(sc);
1522 >        return stealCount;
1523      }
1524  
1525      /**
# Line 925 | Line 1529 | public class ForkJoinPool extends Abstra
1529       * an approximation, obtained by iterating across all threads in
1530       * the pool. This method may be useful for tuning task
1531       * granularities.
1532 <     * @return the number of queued tasks.
1532 >     *
1533 >     * @return the number of queued tasks
1534       */
1535      public long getQueuedTaskCount() {
1536          long count = 0;
1537 <        ForkJoinWorkerThread[] ws = workers;
1538 <        if (ws != null) {
1539 <            for (int i = 0; i < ws.length; ++i) {
935 <                ForkJoinWorkerThread t = ws[i];
936 <                if (t != null)
937 <                    count += t.getQueueSize();
938 <            }
939 <        }
1537 >        for (ForkJoinWorkerThread w : workers)
1538 >            if (w != null)
1539 >                count += w.getQueueSize();
1540          return count;
1541      }
1542  
1543      /**
1544 <     * Returns an estimate of the number tasks submitted to this pool
1545 <     * that have not yet begun executing. This method takes time
1544 >     * Returns an estimate of the number of tasks submitted to this
1545 >     * pool that have not yet begun executing.  This method takes time
1546       * proportional to the number of submissions.
1547 <     * @return the number of queued submissions.
1547 >     *
1548 >     * @return the number of queued submissions
1549       */
1550      public int getQueuedSubmissionCount() {
1551          return submissionQueue.size();
1552      }
1553  
1554      /**
1555 <     * Returns true if there are any tasks submitted to this pool
1556 <     * that have not yet begun executing.
1557 <     * @return {@code true} if there are any queued submissions.
1555 >     * Returns {@code true} if there are any tasks submitted to this
1556 >     * pool that have not yet begun executing.
1557 >     *
1558 >     * @return {@code true} if there are any queued submissions
1559       */
1560      public boolean hasQueuedSubmissions() {
1561          return !submissionQueue.isEmpty();
# Line 963 | Line 1565 | public class ForkJoinPool extends Abstra
1565       * Removes and returns the next unexecuted submission if one is
1566       * available.  This method may be useful in extensions to this
1567       * class that re-assign work in systems with multiple pools.
1568 <     * @return the next submission, or null if none
1568 >     *
1569 >     * @return the next submission, or {@code null} if none
1570       */
1571      protected ForkJoinTask<?> pollSubmission() {
1572          return submissionQueue.poll();
# Line 973 | Line 1576 | public class ForkJoinPool extends Abstra
1576       * Removes all available unexecuted submitted and forked tasks
1577       * from scheduling queues and adds them to the given collection,
1578       * without altering their execution status. These may include
1579 <     * artificially generated or wrapped tasks. This method is designed
1580 <     * to be invoked only when the pool is known to be
1579 >     * artificially generated or wrapped tasks. This method is
1580 >     * designed to be invoked only when the pool is known to be
1581       * quiescent. Invocations at other times may not remove all
1582       * tasks. A failure encountered while attempting to add elements
1583       * to collection {@code c} may result in elements being in
# Line 982 | Line 1585 | public class ForkJoinPool extends Abstra
1585       * exception is thrown.  The behavior of this operation is
1586       * undefined if the specified collection is modified while the
1587       * operation is in progress.
1588 +     *
1589       * @param c the collection to transfer elements into
1590       * @return the number of elements transferred
1591       */
1592 <    protected int drainTasksTo(Collection<ForkJoinTask<?>> c) {
1593 <        int n = submissionQueue.drainTo(c);
1594 <        ForkJoinWorkerThread[] ws = workers;
1595 <        if (ws != null) {
1596 <            for (int i = 0; i < ws.length; ++i) {
1597 <                ForkJoinWorkerThread w = ws[i];
994 <                if (w != null)
995 <                    n += w.drainTasksTo(c);
996 <            }
997 <        }
998 <        return n;
1592 >    protected int drainTasksTo(Collection<? super ForkJoinTask<?>> c) {
1593 >        int count = submissionQueue.drainTo(c);
1594 >        for (ForkJoinWorkerThread w : workers)
1595 >            if (w != null)
1596 >                count += w.drainTasksTo(c);
1597 >        return count;
1598      }
1599  
1600      /**
# Line 1006 | Line 1605 | public class ForkJoinPool extends Abstra
1605       * @return a string identifying this pool, as well as its state
1606       */
1607      public String toString() {
1009        int ps = parallelism;
1010        int wc = workerCounts;
1011        int rc = runControl;
1608          long st = getStealCount();
1609          long qt = getQueuedTaskCount();
1610          long qs = getQueuedSubmissionCount();
1611 +        int wc = workerCounts;
1612 +        int tc = wc >>> TOTAL_COUNT_SHIFT;
1613 +        int rc = wc & RUNNING_COUNT_MASK;
1614 +        int pc = parallelism;
1615 +        int rs = runState;
1616 +        int ac = rs & ACTIVE_COUNT_MASK;
1617          return super.toString() +
1618 <            "[" + runStateToString(runStateOf(rc)) +
1619 <            ", parallelism = " + ps +
1620 <            ", size = " + totalCountOf(wc) +
1621 <            ", active = " + activeCountOf(rc) +
1622 <            ", running = " + runningCountOf(wc) +
1618 >            "[" + runLevelToString(rs) +
1619 >            ", parallelism = " + pc +
1620 >            ", size = " + tc +
1621 >            ", active = " + ac +
1622 >            ", running = " + rc +
1623              ", steals = " + st +
1624              ", tasks = " + qt +
1625              ", submissions = " + qs +
1626              "]";
1627      }
1628  
1629 <    private static String runStateToString(int rs) {
1630 <        switch(rs) {
1631 <        case RUNNING: return "Running";
1632 <        case SHUTDOWN: return "Shutting down";
1633 <        case TERMINATING: return "Terminating";
1032 <        case TERMINATED: return "Terminated";
1033 <        default: throw new Error("Unknown run state");
1034 <        }
1629 >    private static String runLevelToString(int s) {
1630 >        return ((s & TERMINATED) != 0 ? "Terminated" :
1631 >                ((s & TERMINATING) != 0 ? "Terminating" :
1632 >                 ((s & SHUTDOWN) != 0 ? "Shutting down" :
1633 >                  "Running")));
1634      }
1635  
1037    // lifecycle control
1038
1636      /**
1637       * Initiates an orderly shutdown in which previously submitted
1638       * tasks are executed, but no new tasks will be accepted.
1639       * Invocation has no additional effect if already shut down.
1640       * Tasks that are in the process of being submitted concurrently
1641       * during the course of this method may or may not be rejected.
1642 +     *
1643       * @throws SecurityException if a security manager exists and
1644       *         the caller is not permitted to modify threads
1645       *         because it does not hold {@link
1646 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
1646 >     *         java.lang.RuntimePermission}{@code ("modifyThread")}
1647       */
1648      public void shutdown() {
1649          checkPermission();
1650 <        transitionRunStateTo(SHUTDOWN);
1651 <        if (canTerminateOnShutdown(runControl))
1054 <            terminateOnShutdown();
1650 >        advanceRunLevel(SHUTDOWN);
1651 >        tryTerminate(false);
1652      }
1653  
1654      /**
1655 <     * Attempts to stop all actively executing tasks, and cancels all
1656 <     * waiting tasks.  Tasks that are in the process of being
1657 <     * submitted or executed concurrently during the course of this
1658 <     * method may or may not be rejected. Unlike some other executors,
1659 <     * this method cancels rather than collects non-executed tasks
1660 <     * upon termination, so always returns an empty list. However, you
1661 <     * can use method {@code drainTasksTo} before invoking this
1662 <     * method to transfer unexecuted tasks to another collection.
1655 >     * Attempts to cancel and/or stop all tasks, and reject all
1656 >     * subsequently submitted tasks.  Tasks that are in the process of
1657 >     * being submitted or executed concurrently during the course of
1658 >     * this method may or may not be rejected. This method cancels
1659 >     * both existing and unexecuted tasks, in order to permit
1660 >     * termination in the presence of task dependencies. So the method
1661 >     * always returns an empty list (unlike the case for some other
1662 >     * Executors).
1663 >     *
1664       * @return an empty list
1665       * @throws SecurityException if a security manager exists and
1666       *         the caller is not permitted to modify threads
1667       *         because it does not hold {@link
1668 <     *         java.lang.RuntimePermission}{@code ("modifyThread")},
1668 >     *         java.lang.RuntimePermission}{@code ("modifyThread")}
1669       */
1670      public List<Runnable> shutdownNow() {
1671          checkPermission();
1672 <        terminate();
1672 >        tryTerminate(true);
1673          return Collections.emptyList();
1674      }
1675  
# Line 1081 | Line 1679 | public class ForkJoinPool extends Abstra
1679       * @return {@code true} if all tasks have completed following shut down
1680       */
1681      public boolean isTerminated() {
1682 <        return runStateOf(runControl) == TERMINATED;
1682 >        return runState >= TERMINATED;
1683      }
1684  
1685      /**
1686       * Returns {@code true} if the process of termination has
1687 <     * commenced but possibly not yet completed.
1687 >     * commenced but not yet completed.  This method may be useful for
1688 >     * debugging. A return of {@code true} reported a sufficient
1689 >     * period after shutdown may indicate that submitted tasks have
1690 >     * ignored or suppressed interruption, causing this executor not
1691 >     * to properly terminate.
1692       *
1693 <     * @return {@code true} if terminating
1693 >     * @return {@code true} if terminating but not yet terminated
1694       */
1695      public boolean isTerminating() {
1696 <        return runStateOf(runControl) >= TERMINATING;
1696 >        return (runState & (TERMINATING|TERMINATED)) == TERMINATING;
1697      }
1698  
1699      /**
# Line 1100 | Line 1702 | public class ForkJoinPool extends Abstra
1702       * @return {@code true} if this pool has been shut down
1703       */
1704      public boolean isShutdown() {
1705 <        return runStateOf(runControl) >= SHUTDOWN;
1705 >        return runState >= SHUTDOWN;
1706      }
1707  
1708      /**
# Line 1116 | Line 1718 | public class ForkJoinPool extends Abstra
1718       */
1719      public boolean awaitTermination(long timeout, TimeUnit unit)
1720          throws InterruptedException {
1119        long nanos = unit.toNanos(timeout);
1120        final ReentrantLock lock = this.workerLock;
1121        lock.lock();
1122        try {
1123            for (;;) {
1124                if (isTerminated())
1125                    return true;
1126                if (nanos <= 0)
1127                    return false;
1128                nanos = termination.awaitNanos(nanos);
1129            }
1130        } finally {
1131            lock.unlock();
1132        }
1133    }
1134
1135    // Shutdown and termination support
1136
1137    /**
1138     * Callback from terminating worker. Null out the corresponding
1139     * workers slot, and if terminating, try to terminate, else try to
1140     * shrink workers array.
1141     * @param w the worker
1142     */
1143    final void workerTerminated(ForkJoinWorkerThread w) {
1144        updateStealCount(w);
1145        updateWorkerCount(-1);
1146        final ReentrantLock lock = this.workerLock;
1147        lock.lock();
1148        try {
1149            ForkJoinWorkerThread[] ws = workers;
1150            if (ws != null) {
1151                int idx = w.poolIndex;
1152                if (idx >= 0 && idx < ws.length && ws[idx] == w)
1153                    ws[idx] = null;
1154                if (totalCountOf(workerCounts) == 0) {
1155                    terminate(); // no-op if already terminating
1156                    transitionRunStateTo(TERMINATED);
1157                    termination.signalAll();
1158                }
1159                else if (!isTerminating()) {
1160                    tryShrinkWorkerArray();
1161                    tryResumeSpare(true); // allow replacement
1162                }
1163            }
1164        } finally {
1165            lock.unlock();
1166        }
1167        signalIdleWorkers();
1168    }
1169
1170    /**
1171     * Initiate termination.
1172     */
1173    private void terminate() {
1174        if (transitionRunStateTo(TERMINATING)) {
1175            stopAllWorkers();
1176            resumeAllSpares();
1177            signalIdleWorkers();
1178            cancelQueuedSubmissions();
1179            cancelQueuedWorkerTasks();
1180            interruptUnterminatedWorkers();
1181            signalIdleWorkers(); // resignal after interrupt
1182        }
1183    }
1184
1185    /**
1186     * Possibly terminates when on shutdown state.
1187     */
1188    private void terminateOnShutdown() {
1189        if (!hasQueuedSubmissions() && canTerminateOnShutdown(runControl))
1190            terminate();
1191    }
1192
1193    /**
1194     * Clears out and cancels submissions.
1195     */
1196    private void cancelQueuedSubmissions() {
1197        ForkJoinTask<?> task;
1198        while ((task = pollSubmission()) != null)
1199            task.cancel(false);
1200    }
1201
1202    /**
1203     * Cleans out worker queues.
1204     */
1205    private void cancelQueuedWorkerTasks() {
1206        final ReentrantLock lock = this.workerLock;
1207        lock.lock();
1208        try {
1209            ForkJoinWorkerThread[] ws = workers;
1210            if (ws != null) {
1211                for (int i = 0; i < ws.length; ++i) {
1212                    ForkJoinWorkerThread t = ws[i];
1213                    if (t != null)
1214                        t.cancelTasks();
1215                }
1216            }
1217        } finally {
1218            lock.unlock();
1219        }
1220    }
1221
1222    /**
1223     * Sets each worker's status to terminating. Requires lock to avoid
1224     * conflicts with add/remove.
1225     */
1226    private void stopAllWorkers() {
1227        final ReentrantLock lock = this.workerLock;
1228        lock.lock();
1721          try {
1722 <            ForkJoinWorkerThread[] ws = workers;
1723 <            if (ws != null) {
1232 <                for (int i = 0; i < ws.length; ++i) {
1233 <                    ForkJoinWorkerThread t = ws[i];
1234 <                    if (t != null)
1235 <                        t.shutdownNow();
1236 <                }
1237 <            }
1238 <        } finally {
1239 <            lock.unlock();
1240 <        }
1241 <    }
1242 <
1243 <    /**
1244 <     * Interrupts all unterminated workers.  This is not required for
1245 <     * sake of internal control, but may help unstick user code during
1246 <     * shutdown.
1247 <     */
1248 <    private void interruptUnterminatedWorkers() {
1249 <        final ReentrantLock lock = this.workerLock;
1250 <        lock.lock();
1251 <        try {
1252 <            ForkJoinWorkerThread[] ws = workers;
1253 <            if (ws != null) {
1254 <                for (int i = 0; i < ws.length; ++i) {
1255 <                    ForkJoinWorkerThread t = ws[i];
1256 <                    if (t != null && !t.isTerminated()) {
1257 <                        try {
1258 <                            t.interrupt();
1259 <                        } catch (SecurityException ignore) {
1260 <                        }
1261 <                    }
1262 <                }
1263 <            }
1264 <        } finally {
1265 <            lock.unlock();
1266 <        }
1267 <    }
1268 <
1269 <
1270 <    /*
1271 <     * Nodes for event barrier to manage idle threads.  Queue nodes
1272 <     * are basic Treiber stack nodes, also used for spare stack.
1273 <     *
1274 <     * The event barrier has an event count and a wait queue (actually
1275 <     * a Treiber stack).  Workers are enabled to look for work when
1276 <     * the eventCount is incremented. If they fail to find work, they
1277 <     * may wait for next count. Upon release, threads help others wake
1278 <     * up.
1279 <     *
1280 <     * Synchronization events occur only in enough contexts to
1281 <     * maintain overall liveness:
1282 <     *
1283 <     *   - Submission of a new task to the pool
1284 <     *   - Resizes or other changes to the workers array
1285 <     *   - pool termination
1286 <     *   - A worker pushing a task on an empty queue
1287 <     *
1288 <     * The case of pushing a task occurs often enough, and is heavy
1289 <     * enough compared to simple stack pushes, to require special
1290 <     * handling: Method signalWork returns without advancing count if
1291 <     * the queue appears to be empty.  This would ordinarily result in
1292 <     * races causing some queued waiters not to be woken up. To avoid
1293 <     * this, the first worker enqueued in method sync (see
1294 <     * syncIsReleasable) rescans for tasks after being enqueued, and
1295 <     * helps signal if any are found. This works well because the
1296 <     * worker has nothing better to do, and so might as well help
1297 <     * alleviate the overhead and contention on the threads actually
1298 <     * doing work.  Also, since event counts increments on task
1299 <     * availability exist to maintain liveness (rather than to force
1300 <     * refreshes etc), it is OK for callers to exit early if
1301 <     * contending with another signaller.
1302 <     */
1303 <    static final class WaitQueueNode {
1304 <        WaitQueueNode next; // only written before enqueued
1305 <        volatile ForkJoinWorkerThread thread; // nulled to cancel wait
1306 <        final long count; // unused for spare stack
1307 <
1308 <        WaitQueueNode(long c, ForkJoinWorkerThread w) {
1309 <            count = c;
1310 <            thread = w;
1311 <        }
1312 <
1313 <        /**
1314 <         * Wakes up waiter, returning false if known to already
1315 <         */
1316 <        boolean signal() {
1317 <            ForkJoinWorkerThread t = thread;
1318 <            if (t == null)
1319 <                return false;
1320 <            thread = null;
1321 <            LockSupport.unpark(t);
1322 <            return true;
1323 <        }
1324 <
1325 <        /**
1326 <         * Awaits release on sync.
1327 <         */
1328 <        void awaitSyncRelease(ForkJoinPool p) {
1329 <            while (thread != null && !p.syncIsReleasable(this))
1330 <                LockSupport.park(this);
1331 <        }
1332 <
1333 <        /**
1334 <         * Awaits resumption as spare.
1335 <         */
1336 <        void awaitSpareRelease() {
1337 <            while (thread != null) {
1338 <                if (!Thread.interrupted())
1339 <                    LockSupport.park(this);
1340 <            }
1341 <        }
1342 <    }
1343 <
1344 <    /**
1345 <     * Ensures that no thread is waiting for count to advance from the
1346 <     * current value of eventCount read on entry to this method, by
1347 <     * releasing waiting threads if necessary.
1348 <     * @return the count
1349 <     */
1350 <    final long ensureSync() {
1351 <        long c = eventCount;
1352 <        WaitQueueNode q;
1353 <        while ((q = syncStack) != null && q.count < c) {
1354 <            if (casBarrierStack(q, null)) {
1355 <                do {
1356 <                    q.signal();
1357 <                } while ((q = q.next) != null);
1358 <                break;
1359 <            }
1360 <        }
1361 <        return c;
1362 <    }
1363 <
1364 <    /**
1365 <     * Increments event count and releases waiting threads.
1366 <     */
1367 <    private void signalIdleWorkers() {
1368 <        long c;
1369 <        do;while (!casEventCount(c = eventCount, c+1));
1370 <        ensureSync();
1371 <    }
1372 <
1373 <    /**
1374 <     * Signals threads waiting to poll a task. Because method sync
1375 <     * rechecks availability, it is OK to only proceed if queue
1376 <     * appears to be non-empty, and OK to skip under contention to
1377 <     * increment count (since some other thread succeeded).
1378 <     */
1379 <    final void signalWork() {
1380 <        long c;
1381 <        WaitQueueNode q;
1382 <        if (syncStack != null &&
1383 <            casEventCount(c = eventCount, c+1) &&
1384 <            (((q = syncStack) != null && q.count <= c) &&
1385 <             (!casBarrierStack(q, q.next) || !q.signal())))
1386 <            ensureSync();
1387 <    }
1388 <
1389 <    /**
1390 <     * Waits until event count advances from last value held by
1391 <     * caller, or if excess threads, caller is resumed as spare, or
1392 <     * caller or pool is terminating. Updates caller's event on exit.
1393 <     * @param w the calling worker thread
1394 <     */
1395 <    final void sync(ForkJoinWorkerThread w) {
1396 <        updateStealCount(w); // Transfer w's count while it is idle
1397 <
1398 <        while (!w.isShutdown() && !isTerminating() && !suspendIfSpare(w)) {
1399 <            long prev = w.lastEventCount;
1400 <            WaitQueueNode node = null;
1401 <            WaitQueueNode h;
1402 <            while (eventCount == prev &&
1403 <                   ((h = syncStack) == null || h.count == prev)) {
1404 <                if (node == null)
1405 <                    node = new WaitQueueNode(prev, w);
1406 <                if (casBarrierStack(node.next = h, node)) {
1407 <                    node.awaitSyncRelease(this);
1408 <                    break;
1409 <                }
1410 <            }
1411 <            long ec = ensureSync();
1412 <            if (ec != prev) {
1413 <                w.lastEventCount = ec;
1414 <                break;
1415 <            }
1416 <        }
1417 <    }
1418 <
1419 <    /**
1420 <     * Returns true if worker waiting on sync can proceed:
1421 <     *  - on signal (thread == null)
1422 <     *  - on event count advance (winning race to notify vs signaller)
1423 <     *  - on Interrupt
1424 <     *  - if the first queued node, we find work available
1425 <     * If node was not signalled and event count not advanced on exit,
1426 <     * then we also help advance event count.
1427 <     * @return true if node can be released
1428 <     */
1429 <    final boolean syncIsReleasable(WaitQueueNode node) {
1430 <        long prev = node.count;
1431 <        if (!Thread.interrupted() && node.thread != null &&
1432 <            (node.next != null ||
1433 <             !ForkJoinWorkerThread.hasQueuedTasks(workers)) &&
1434 <            eventCount == prev)
1722 >            return termination.awaitAdvanceInterruptibly(0, timeout, unit) > 0;
1723 >        } catch (TimeoutException ex) {
1724              return false;
1436        if (node.thread != null) {
1437            node.thread = null;
1438            long ec = eventCount;
1439            if (prev <= ec) // help signal
1440                casEventCount(ec, ec+1);
1441        }
1442        return true;
1443    }
1444
1445    /**
1446     * Returns true if a new sync event occurred since last call to
1447     * sync or this method, if so, updating caller's count.
1448     */
1449    final boolean hasNewSyncEvent(ForkJoinWorkerThread w) {
1450        long lc = w.lastEventCount;
1451        long ec = ensureSync();
1452        if (ec == lc)
1453            return false;
1454        w.lastEventCount = ec;
1455        return true;
1456    }
1457
1458    //  Parallelism maintenance
1459
1460    /**
1461     * Decrements running count; if too low, adds spare.
1462     *
1463     * Conceptually, all we need to do here is add or resume a
1464     * spare thread when one is about to block (and remove or
1465     * suspend it later when unblocked -- see suspendIfSpare).
1466     * However, implementing this idea requires coping with
1467     * several problems: We have imperfect information about the
1468     * states of threads. Some count updates can and usually do
1469     * lag run state changes, despite arrangements to keep them
1470     * accurate (for example, when possible, updating counts
1471     * before signalling or resuming), especially when running on
1472     * dynamic JVMs that don't optimize the infrequent paths that
1473     * update counts. Generating too many threads can make these
1474     * problems become worse, because excess threads are more
1475     * likely to be context-switched with others, slowing them all
1476     * down, especially if there is no work available, so all are
1477     * busy scanning or idling.  Also, excess spare threads can
1478     * only be suspended or removed when they are idle, not
1479     * immediately when they aren't needed. So adding threads will
1480     * raise parallelism level for longer than necessary.  Also,
1481     * FJ applications often encounter highly transient peaks when
1482     * many threads are blocked joining, but for less time than it
1483     * takes to create or resume spares.
1484     *
1485     * @param joinMe if non-null, return early if done
1486     * @param maintainParallelism if true, try to stay within
1487     * target counts, else create only to avoid starvation
1488     * @return true if joinMe known to be done
1489     */
1490    final boolean preJoin(ForkJoinTask<?> joinMe, boolean maintainParallelism) {
1491        maintainParallelism &= maintainsParallelism; // overrride
1492        boolean dec = false;  // true when running count decremented
1493        while (spareStack == null || !tryResumeSpare(dec)) {
1494            int counts = workerCounts;
1495            if (dec || (dec = casWorkerCounts(counts, --counts))) { // CAS cheat
1496                if (!needSpare(counts, maintainParallelism))
1497                    break;
1498                if (joinMe.status < 0)
1499                    return true;
1500                if (tryAddSpare(counts))
1501                    break;
1502            }
1503        }
1504        return false;
1505    }
1506
1507    /**
1508     * Same idea as preJoin
1509     */
1510    final boolean preBlock(ManagedBlocker blocker, boolean maintainParallelism){
1511        maintainParallelism &= maintainsParallelism;
1512        boolean dec = false;
1513        while (spareStack == null || !tryResumeSpare(dec)) {
1514            int counts = workerCounts;
1515            if (dec || (dec = casWorkerCounts(counts, --counts))) {
1516                if (!needSpare(counts, maintainParallelism))
1517                    break;
1518                if (blocker.isReleasable())
1519                    return true;
1520                if (tryAddSpare(counts))
1521                    break;
1522            }
1523        }
1524        return false;
1525    }
1526
1527    /**
1528     * Returns true if a spare thread appears to be needed.  If
1529     * maintaining parallelism, returns true when the deficit in
1530     * running threads is more than the surplus of total threads, and
1531     * there is apparently some work to do.  This self-limiting rule
1532     * means that the more threads that have already been added, the
1533     * less parallelism we will tolerate before adding another.
1534     * @param counts current worker counts
1535     * @param maintainParallelism try to maintain parallelism
1536     */
1537    private boolean needSpare(int counts, boolean maintainParallelism) {
1538        int ps = parallelism;
1539        int rc = runningCountOf(counts);
1540        int tc = totalCountOf(counts);
1541        int runningDeficit = ps - rc;
1542        int totalSurplus = tc - ps;
1543        return (tc < maxPoolSize &&
1544                (rc == 0 || totalSurplus < 0 ||
1545                 (maintainParallelism &&
1546                  runningDeficit > totalSurplus &&
1547                  ForkJoinWorkerThread.hasQueuedTasks(workers))));
1548    }
1549
1550    /**
1551     * Adds a spare worker if lock available and no more than the
1552     * expected numbers of threads exist.
1553     * @return true if successful
1554     */
1555    private boolean tryAddSpare(int expectedCounts) {
1556        final ReentrantLock lock = this.workerLock;
1557        int expectedRunning = runningCountOf(expectedCounts);
1558        int expectedTotal = totalCountOf(expectedCounts);
1559        boolean success = false;
1560        boolean locked = false;
1561        // confirm counts while locking; CAS after obtaining lock
1562        try {
1563            for (;;) {
1564                int s = workerCounts;
1565                int tc = totalCountOf(s);
1566                int rc = runningCountOf(s);
1567                if (rc > expectedRunning || tc > expectedTotal)
1568                    break;
1569                if (!locked && !(locked = lock.tryLock()))
1570                    break;
1571                if (casWorkerCounts(s, workerCountsFor(tc+1, rc+1))) {
1572                    createAndStartSpare(tc);
1573                    success = true;
1574                    break;
1575                }
1576            }
1577        } finally {
1578            if (locked)
1579                lock.unlock();
1580        }
1581        return success;
1582    }
1583
1584    /**
1585     * Adds the kth spare worker. On entry, pool counts are already
1586     * adjusted to reflect addition.
1587     */
1588    private void createAndStartSpare(int k) {
1589        ForkJoinWorkerThread w = null;
1590        ForkJoinWorkerThread[] ws = ensureWorkerArrayCapacity(k + 1);
1591        int len = ws.length;
1592        // Probably, we can place at slot k. If not, find empty slot
1593        if (k < len && ws[k] != null) {
1594            for (k = 0; k < len && ws[k] != null; ++k)
1595                ;
1596        }
1597        if (k < len && !isTerminating() && (w = createWorker(k)) != null) {
1598            ws[k] = w;
1599            w.start();
1600        }
1601        else
1602            updateWorkerCount(-1); // adjust on failure
1603        signalIdleWorkers();
1604    }
1605
1606    /**
1607     * Suspends calling thread w if there are excess threads.  Called
1608     * only from sync.  Spares are enqueued in a Treiber stack using
1609     * the same WaitQueueNodes as barriers.  They are resumed mainly
1610     * in preJoin, but are also woken on pool events that require all
1611     * threads to check run state.
1612     * @param w the caller
1613     */
1614    private boolean suspendIfSpare(ForkJoinWorkerThread w) {
1615        WaitQueueNode node = null;
1616        int s;
1617        while (parallelism < runningCountOf(s = workerCounts)) {
1618            if (node == null)
1619                node = new WaitQueueNode(0, w);
1620            if (casWorkerCounts(s, s-1)) { // representation-dependent
1621                // push onto stack
1622                do;while (!casSpareStack(node.next = spareStack, node));
1623                // block until released by resumeSpare
1624                node.awaitSpareRelease();
1625                return true;
1626            }
1627        }
1628        return false;
1629    }
1630
1631    /**
1632     * Tries to pop and resume a spare thread.
1633     * @param updateCount if true, increment running count on success
1634     * @return true if successful
1635     */
1636    private boolean tryResumeSpare(boolean updateCount) {
1637        WaitQueueNode q;
1638        while ((q = spareStack) != null) {
1639            if (casSpareStack(q, q.next)) {
1640                if (updateCount)
1641                    updateRunningCount(1);
1642                q.signal();
1643                return true;
1644            }
1645        }
1646        return false;
1647    }
1648
1649    /**
1650     * Pops and resumes all spare threads. Same idea as ensureSync.
1651     * @return true if any spares released
1652     */
1653    private boolean resumeAllSpares() {
1654        WaitQueueNode q;
1655        while ( (q = spareStack) != null) {
1656            if (casSpareStack(q, null)) {
1657                do {
1658                    updateRunningCount(1);
1659                    q.signal();
1660                } while ((q = q.next) != null);
1661                return true;
1662            }
1663        }
1664        return false;
1665    }
1666
1667    /**
1668     * Pops and shuts down excessive spare threads. Call only while
1669     * holding lock. This is not guaranteed to eliminate all excess
1670     * threads, only those suspended as spares, which are the ones
1671     * unlikely to be needed in the future.
1672     */
1673    private void trimSpares() {
1674        int surplus = totalCountOf(workerCounts) - parallelism;
1675        WaitQueueNode q;
1676        while (surplus > 0 && (q = spareStack) != null) {
1677            if (casSpareStack(q, null)) {
1678                do {
1679                    updateRunningCount(1);
1680                    ForkJoinWorkerThread w = q.thread;
1681                    if (w != null && surplus > 0 &&
1682                        runningCountOf(workerCounts) > 0 && w.shutdown())
1683                        --surplus;
1684                    q.signal();
1685                } while ((q = q.next) != null);
1686            }
1725          }
1726      }
1727  
1728      /**
1729       * Interface for extending managed parallelism for tasks running
1730 <     * in ForkJoinPools. A ManagedBlocker provides two methods.
1731 <     * Method {@code isReleasable} must return true if blocking is not
1732 <     * necessary. Method {@code block} blocks the current thread
1733 <     * if necessary (perhaps internally invoking isReleasable before
1734 <     * actually blocking.).
1730 >     * in {@link ForkJoinPool}s.
1731 >     *
1732 >     * <p>A {@code ManagedBlocker} provides two methods.  Method
1733 >     * {@code isReleasable} must return {@code true} if blocking is
1734 >     * not necessary. Method {@code block} blocks the current thread
1735 >     * if necessary (perhaps internally invoking {@code isReleasable}
1736 >     * before actually blocking). The unusual methods in this API
1737 >     * accommodate synchronizers that may, but don't usually, block
1738 >     * for long periods. Similarly, they allow more efficient internal
1739 >     * handling of cases in which additional workers may be, but
1740 >     * usually are not, needed to ensure sufficient parallelism.
1741 >     * Toward this end, implementations of method {@code isReleasable}
1742 >     * must be amenable to repeated invocation.
1743 >     *
1744       * <p>For example, here is a ManagedBlocker based on a
1745       * ReentrantLock:
1746 <     * <pre>
1747 <     *   class ManagedLocker implements ManagedBlocker {
1748 <     *     final ReentrantLock lock;
1749 <     *     boolean hasLock = false;
1750 <     *     ManagedLocker(ReentrantLock lock) { this.lock = lock; }
1751 <     *     public boolean block() {
1752 <     *        if (!hasLock)
1753 <     *           lock.lock();
1754 <     *        return true;
1755 <     *     }
1756 <     *     public boolean isReleasable() {
1757 <     *        return hasLock || (hasLock = lock.tryLock());
1758 <     *     }
1746 >     *  <pre> {@code
1747 >     * class ManagedLocker implements ManagedBlocker {
1748 >     *   final ReentrantLock lock;
1749 >     *   boolean hasLock = false;
1750 >     *   ManagedLocker(ReentrantLock lock) { this.lock = lock; }
1751 >     *   public boolean block() {
1752 >     *     if (!hasLock)
1753 >     *       lock.lock();
1754 >     *     return true;
1755 >     *   }
1756 >     *   public boolean isReleasable() {
1757 >     *     return hasLock || (hasLock = lock.tryLock());
1758 >     *   }
1759 >     * }}</pre>
1760 >     *
1761 >     * <p>Here is a class that possibly blocks waiting for an
1762 >     * item on a given queue:
1763 >     *  <pre> {@code
1764 >     * class QueueTaker<E> implements ManagedBlocker {
1765 >     *   final BlockingQueue<E> queue;
1766 >     *   volatile E item = null;
1767 >     *   QueueTaker(BlockingQueue<E> q) { this.queue = q; }
1768 >     *   public boolean block() throws InterruptedException {
1769 >     *     if (item == null)
1770 >     *       item = queue.take();
1771 >     *     return true;
1772 >     *   }
1773 >     *   public boolean isReleasable() {
1774 >     *     return item != null || (item = queue.poll()) != null;
1775       *   }
1776 <     * </pre>
1776 >     *   public E getItem() { // call after pool.managedBlock completes
1777 >     *     return item;
1778 >     *   }
1779 >     * }}</pre>
1780       */
1781      public static interface ManagedBlocker {
1782          /**
1783           * Possibly blocks the current thread, for example waiting for
1784           * a lock or condition.
1785 <         * @return true if no additional blocking is necessary (i.e.,
1786 <         * if isReleasable would return true).
1785 >         *
1786 >         * @return {@code true} if no additional blocking is necessary
1787 >         * (i.e., if isReleasable would return true)
1788           * @throws InterruptedException if interrupted while waiting
1789 <         * (the method is not required to do so, but is allowed to).
1789 >         * (the method is not required to do so, but is allowed to)
1790           */
1791          boolean block() throws InterruptedException;
1792  
1793          /**
1794 <         * Returns true if blocking is unnecessary.
1794 >         * Returns {@code true} if blocking is unnecessary.
1795           */
1796          boolean isReleasable();
1797      }
1798  
1799      /**
1800       * Blocks in accord with the given blocker.  If the current thread
1801 <     * is a ForkJoinWorkerThread, this method possibly arranges for a
1802 <     * spare thread to be activated if necessary to ensure parallelism
1803 <     * while the current thread is blocked.  If
1804 <     * {@code maintainParallelism} is true and the pool supports
1805 <     * it ({@link #getMaintainsParallelism}), this method attempts to
1806 <     * maintain the pool's nominal parallelism. Otherwise if activates
1807 <     * a thread only if necessary to avoid complete starvation. This
1808 <     * option may be preferable when blockages use timeouts, or are
1809 <     * almost always brief.
1810 <     *
1811 <     * <p> If the caller is not a ForkJoinTask, this method is behaviorally
1812 <     * equivalent to
1813 <     * <pre>
1814 <     *   while (!blocker.isReleasable())
1748 <     *      if (blocker.block())
1749 <     *         return;
1750 <     * </pre>
1751 <     * If the caller is a ForkJoinTask, then the pool may first
1752 <     * be expanded to ensure parallelism, and later adjusted.
1801 >     * is a {@link ForkJoinWorkerThread}, this method possibly
1802 >     * arranges for a spare thread to be activated if necessary to
1803 >     * ensure sufficient parallelism while the current thread is blocked.
1804 >     *
1805 >     * <p>If the caller is not a {@link ForkJoinTask}, this method is
1806 >     * behaviorally equivalent to
1807 >     *  <pre> {@code
1808 >     * while (!blocker.isReleasable())
1809 >     *   if (blocker.block())
1810 >     *     return;
1811 >     * }</pre>
1812 >     *
1813 >     * If the caller is a {@code ForkJoinTask}, then the pool may
1814 >     * first be expanded to ensure parallelism, and later adjusted.
1815       *
1816       * @param blocker the blocker
1817 <     * @param maintainParallelism if true and supported by this pool,
1756 <     * attempt to maintain the pool's nominal parallelism; otherwise
1757 <     * activate a thread only if necessary to avoid complete
1758 <     * starvation.
1759 <     * @throws InterruptedException if blocker.block did so.
1817 >     * @throws InterruptedException if blocker.block did so
1818       */
1819 <    public static void managedBlock(ManagedBlocker blocker,
1762 <                                    boolean maintainParallelism)
1819 >    public static void managedBlock(ManagedBlocker blocker)
1820          throws InterruptedException {
1821          Thread t = Thread.currentThread();
1822 <        ForkJoinPool pool = (t instanceof ForkJoinWorkerThread?
1823 <                             ((ForkJoinWorkerThread)t).pool : null);
1824 <        if (!blocker.isReleasable()) {
1825 <            try {
1826 <                if (pool == null ||
1827 <                    !pool.preBlock(blocker, maintainParallelism))
1771 <                    awaitBlocker(blocker);
1772 <            } finally {
1773 <                if (pool != null)
1774 <                    pool.updateRunningCount(1);
1775 <            }
1822 >        if (t instanceof ForkJoinWorkerThread) {
1823 >            ForkJoinWorkerThread w = (ForkJoinWorkerThread) t;
1824 >            w.pool.awaitBlocker(blocker);
1825 >        }
1826 >        else {
1827 >            do {} while (!blocker.isReleasable() && !blocker.block());
1828          }
1829      }
1830  
1831 <    private static void awaitBlocker(ManagedBlocker blocker)
1832 <        throws InterruptedException {
1833 <        do;while (!blocker.isReleasable() && !blocker.block());
1782 <    }
1783 <
1784 <    // AbstractExecutorService overrides
1831 >    // AbstractExecutorService overrides.  These rely on undocumented
1832 >    // fact that ForkJoinTask.adapt returns ForkJoinTasks that also
1833 >    // implement RunnableFuture.
1834  
1835      protected <T> RunnableFuture<T> newTaskFor(Runnable runnable, T value) {
1836 <        return new AdaptedRunnable(runnable, value);
1836 >        return (RunnableFuture<T>) ForkJoinTask.adapt(runnable, value);
1837      }
1838  
1839      protected <T> RunnableFuture<T> newTaskFor(Callable<T> callable) {
1840 <        return new AdaptedCallable(callable);
1840 >        return (RunnableFuture<T>) ForkJoinTask.adapt(callable);
1841      }
1842  
1843 +    // Unsafe mechanics
1844  
1845 <    // Temporary Unsafe mechanics for preliminary release
1846 <    private static Unsafe getUnsafe() throws Throwable {
1845 >    private static final sun.misc.Unsafe UNSAFE = getUnsafe();
1846 >    private static final long workerCountsOffset =
1847 >        objectFieldOffset("workerCounts", ForkJoinPool.class);
1848 >    private static final long runStateOffset =
1849 >        objectFieldOffset("runState", ForkJoinPool.class);
1850 >    private static final long eventCountOffset =
1851 >        objectFieldOffset("eventCount", ForkJoinPool.class);
1852 >    private static final long eventWaitersOffset =
1853 >        objectFieldOffset("eventWaiters", ForkJoinPool.class);
1854 >    private static final long stealCountOffset =
1855 >        objectFieldOffset("stealCount", ForkJoinPool.class);
1856 >    private static final long spareWaitersOffset =
1857 >        objectFieldOffset("spareWaiters", ForkJoinPool.class);
1858 >
1859 >    private static long objectFieldOffset(String field, Class<?> klazz) {
1860 >        try {
1861 >            return UNSAFE.objectFieldOffset(klazz.getDeclaredField(field));
1862 >        } catch (NoSuchFieldException e) {
1863 >            // Convert Exception to corresponding Error
1864 >            NoSuchFieldError error = new NoSuchFieldError(field);
1865 >            error.initCause(e);
1866 >            throw error;
1867 >        }
1868 >    }
1869 >
1870 >    /**
1871 >     * Returns a sun.misc.Unsafe.  Suitable for use in a 3rd party package.
1872 >     * Replace with a simple call to Unsafe.getUnsafe when integrating
1873 >     * into a jdk.
1874 >     *
1875 >     * @return a sun.misc.Unsafe
1876 >     */
1877 >    private static sun.misc.Unsafe getUnsafe() {
1878          try {
1879 <            return Unsafe.getUnsafe();
1879 >            return sun.misc.Unsafe.getUnsafe();
1880          } catch (SecurityException se) {
1881              try {
1882                  return java.security.AccessController.doPrivileged
1883 <                    (new java.security.PrivilegedExceptionAction<Unsafe>() {
1884 <                        public Unsafe run() throws Exception {
1885 <                            return getUnsafePrivileged();
1883 >                    (new java.security
1884 >                     .PrivilegedExceptionAction<sun.misc.Unsafe>() {
1885 >                        public sun.misc.Unsafe run() throws Exception {
1886 >                            java.lang.reflect.Field f = sun.misc
1887 >                                .Unsafe.class.getDeclaredField("theUnsafe");
1888 >                            f.setAccessible(true);
1889 >                            return (sun.misc.Unsafe) f.get(null);
1890                          }});
1891              } catch (java.security.PrivilegedActionException e) {
1892 <                throw e.getCause();
1892 >                throw new RuntimeException("Could not initialize intrinsics",
1893 >                                           e.getCause());
1894              }
1895          }
1896      }
1811
1812    private static Unsafe getUnsafePrivileged()
1813            throws NoSuchFieldException, IllegalAccessException {
1814        Field f = Unsafe.class.getDeclaredField("theUnsafe");
1815        f.setAccessible(true);
1816        return (Unsafe) f.get(null);
1817    }
1818
1819    private static long fieldOffset(String fieldName)
1820            throws NoSuchFieldException {
1821        return _unsafe.objectFieldOffset
1822            (ForkJoinPool.class.getDeclaredField(fieldName));
1823    }
1824
1825    static final Unsafe _unsafe;
1826    static final long eventCountOffset;
1827    static final long workerCountsOffset;
1828    static final long runControlOffset;
1829    static final long syncStackOffset;
1830    static final long spareStackOffset;
1831
1832    static {
1833        try {
1834            _unsafe = getUnsafe();
1835            eventCountOffset = fieldOffset("eventCount");
1836            workerCountsOffset = fieldOffset("workerCounts");
1837            runControlOffset = fieldOffset("runControl");
1838            syncStackOffset = fieldOffset("syncStack");
1839            spareStackOffset = fieldOffset("spareStack");
1840        } catch (Throwable e) {
1841            throw new RuntimeException("Could not initialize intrinsics", e);
1842        }
1843    }
1844
1845    private boolean casEventCount(long cmp, long val) {
1846        return _unsafe.compareAndSwapLong(this, eventCountOffset, cmp, val);
1847    }
1848    private boolean casWorkerCounts(int cmp, int val) {
1849        return _unsafe.compareAndSwapInt(this, workerCountsOffset, cmp, val);
1850    }
1851    private boolean casRunControl(int cmp, int val) {
1852        return _unsafe.compareAndSwapInt(this, runControlOffset, cmp, val);
1853    }
1854    private boolean casSpareStack(WaitQueueNode cmp, WaitQueueNode val) {
1855        return _unsafe.compareAndSwapObject(this, spareStackOffset, cmp, val);
1856    }
1857    private boolean casBarrierStack(WaitQueueNode cmp, WaitQueueNode val) {
1858        return _unsafe.compareAndSwapObject(this, syncStackOffset, cmp, val);
1859    }
1897   }

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines