ViewVC Help

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

root/jsr166/jsr166/src/jsr166y/ForkJoinPool.java

Comparing jsr166/src/jsr166y/ForkJoinPool.java (file contents):
Revision 1.64 by dl, Tue Aug 17 18:30:32 2010 UTC vs.
Revision 1.92 by dl, Tue Feb 22 10:50:51 2011 UTC

#	Line 6 | Line 6
6
7		package jsr166y;
8
9	–	import java.util.concurrent.*;
10	–
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.Random;
15	+	import java.util.concurrent.AbstractExecutorService;
16	+	import java.util.concurrent.Callable;
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	<	import java.util.concurrent.atomic.AtomicInteger;
19	<	import java.util.concurrent.CountDownLatch;
26	>	import java.util.concurrent.locks.Condition;
27
28		/**
29		* An {@link ExecutorService} for running {@link ForkJoinTask}s.
#	Line 69 | Line 76 | import java.util.concurrent.CountDownLat
76		*    <td ALIGN=CENTER> <b>Call from within fork/join computations</b></td>
77		*  </tr>
78		*  <tr>
79	<	*    <td> <b>Arange async execution</td>
79	>	*    <td> <b>Arrange async execution</td>
80		*    <td> {@link #execute(ForkJoinTask)}</td>
81		*    <td> {@link ForkJoinTask#fork}</td>
82		*  </tr>
#	Line 124 | Line 131 | public class ForkJoinPool extends Abstra
131		* set of worker threads: Submissions from non-FJ threads enter
132		* into a submission queue. Workers take these tasks and typically
133		* split them into subtasks that may be stolen by other workers.
134	<	* The main work-stealing mechanics implemented in class
135	<	* ForkJoinWorkerThread give first priority to processing tasks
136	<	* from their own queues (LIFO or FIFO, depending on mode), then
137	<	* to randomized FIFO steals of tasks in other worker queues, and
138	<	* lastly to new submissions. These mechanics do not consider
139	<	* affinities, loads, cache localities, etc, so rarely provide the
140	<	* best possible performance on a given machine, but portably
141	<	* provide good throughput by averaging over these factors.
142	<	* (Further, even if we did try to use such information, we do not
143	<	* usually have a basis for exploiting it. For example, some sets
144	<	* of tasks profit from cache affinities, but others are harmed by
145	<	* cache pollution effects.)
146	<	*
147	<	* Beyond work-stealing support and essential bookkeeping, the
148	<	* main responsibility of this framework is to take actions when
149	<	* one worker is waiting to join a task stolen (or always held by)
150	<	* another.  Becauae we are multiplexing many tasks on to a pool
151	<	* of workers, we can't just let them block (as in Thread.join).
152	<	* We also cannot just reassign the joiner's run-time stack with
153	<	* another and replace it later, which would be a form of
154	<	* "continuation", that even if possible is not necessarily a good
155	<	* idea. Given that the creation costs of most threads on most
156	<	* systems mainly surrounds setting up runtime stacks, thread
157	<	* creation and switching is usually not much more expensive than
158	<	* stack creation and switching, and is more flexible). Instead we
134	>	* Preference rules give first priority to processing tasks from
135	>	* their own queues (LIFO or FIFO, depending on mode), then to
136	>	* randomized FIFO steals of tasks in other worker queues, and
137	>	* lastly to new submissions.
138	>	*
139	>	* The main throughput advantages of work-stealing stem from
140	>	* decentralized control -- workers mostly take tasks from
141	>	* themselves or each other. We cannot negate this in the
142	>	* implementation of other management responsibilities. The main
143	>	* tactic for avoiding bottlenecks is packing nearly all
144	>	* essentially atomic control state into a single 64bit volatile
145	>	* variable ("ctl"). This variable is read on the order of 10-100
146	>	* times as often as it is modified (always via CAS). (There is
147	>	* some additional control state, for example variable "shutdown"
148	>	* for which we can cope with uncoordinated updates.)  This
149	>	* streamlines synchronization and control at the expense of messy
150	>	* constructions needed to repack status bits upon updates.
151	>	* Updates tend not to contend with each other except during
152	>	* bursts while submitted tasks begin or end.  In some cases when
153	>	* they do contend, threads can instead do something else
154	>	* (usually, scan for tesks) until contention subsides.
155	>	*
156	>	* To enable packing, we restrict maximum parallelism to (1<<15)-1
157	>	* (which is far in excess of normal operating range) to allow
158	>	* ids, counts, and their negations (used for thresholding) to fit
159	>	* into 16bit fields.
160	>	*
161	>	* Recording Workers.  Workers are recorded in the "workers" array
162	>	* that is created upon pool construction and expanded if (rarely)
163	>	* necessary.  This is an array as opposed to some other data
164	>	* structure to support index-based random steals by workers.
165	>	* Updates to the array recording new workers and unrecording
166	>	* terminated ones are protected from each other by a seqLock
167	>	* (scanGuard) but the array is otherwise concurrently readable,
168	>	* and accessed directly by workers. To simplify index-based
169	>	* operations, the array size is always a power of two, and all
170	>	* readers must tolerate null slots. To avoid flailing during
171	>	* start-up, the array is presized to hold twice #parallelism
172	>	* workers (which is unlikely to need further resizing during
173	>	* execution). But to avoid dealing with so many null slots,
174	>	* variable scanGuard includes a mask for the nearest power of two
175	>	* that contains all current workers.  All worker thread creation
176	>	* is on-demand, triggered by task submissions, replacement of
177	>	* terminated workers, and/or compensation for blocked
178	>	* workers. However, all other support code is set up to work with
179	>	* other policies.  To ensure that we do not hold on to worker
180	>	* references that would prevent GC, ALL accesses to workers are
181	>	* via indices into the workers array (which is one source of some
182	>	* of the messy code constructions here). In essence, the workers
183	>	* array serves as a weak reference mechanism. Thus for example
184	>	* the wait queue field of ctl stores worker indices, not worker
185	>	* references.  Access to the workers in associated methods (for
186	>	* example signalWork) must both index-check and null-check the
187	>	* IDs. All such accesses ignore bad IDs by returning out early
188	>	* from what they are doing, since this can only be associated
189	>	* with termination, in which case it is OK to give up.
190	>	*
191	>	* All uses of the workers array, as well as queue arrays, check
192	>	* that the array is non-null (even if previously non-null). This
193	>	* allows nulling during termination, which is currently not
194	>	* necessary, but remains an option for resource-revocation-based
195	>	* shutdown schemes.
196	>	*
197	>	* Wait Queuing. Unlike HPC work-stealing frameworks, we cannot
198	>	* let workers spin indefinitely scanning for tasks when none are
199	>	* can be immediately found, and we cannot start/resume workers
200	>	* unless there appear to be tasks available.  On the other hand,
201	>	* we must quickly prod them into action when new tasks are
202	>	* submitted or generated.  We park/unpark workers after placing
203	>	* in an event wait queue when they cannot find work. This "queue"
204	>	* is actually a simple Treiber stack, headed by the "id" field of
205	>	* ctl, plus a 15bit counter value to both wake up waiters (by
206	>	* advancing their count) and avoid ABA effects. Successors are
207	>	* held in worker field "nextWait".  Queuing deals with several
208	>	* intrinsic races, mainly that a task-producing thread can miss
209	>	* seeing (and signalling) another thread that gave up looking for
210	>	* work but has not yet entered the wait queue. We solve this by
211	>	* requiring a full sweep of all workers both before (in scan())
212	>	* and after (in awaitWork()) a newly waiting worker is added to
213	>	* the wait queue. During a rescan, the worker might release some
214	>	* other queued worker rather than itself, which has the same net
215	>	* effect.
216	>	*
217	>	* Signalling.  We create or wake up workers only when there
218	>	* appears to be at least one task they might be able to find and
219	>	* execute.  When a submission is added or another worker adds a
220	>	* task to a queue that previously had two or fewer tasks, they
221	>	* signal waiting workers (or trigger creation of new ones if
222	>	* fewer than the given parallelism level -- see signalWork).
223	>	* These primary signals are buttressed by signals during rescans
224	>	* as well as those performed when a worker steals a task and
225	>	* notices that there are more tasks too; together these cover the
226	>	* signals needed in cases when more than two tasks are pushed
227	>	* but untaken.
228	>	*
229	>	* Trimming workers. To release resources after periods of lack of
230	>	* use, a worker starting to wait when the pool is quiescent will
231	>	* time out and terminate if the pool has remained quiescent for
232	>	* SHRINK_RATE nanosecs.
233	>	*
234	>	* Submissions. External submissions are maintained in an
235	>	* array-based queue that is structured identically to
236	>	* ForkJoinWorkerThread queues (which see) except for the use of
237	>	* submissionLock in method addSubmission. Unlike worker queues,
238	>	* multiple external threads can add new submissions.
239	>	*
240	>	* Compensation. Beyond work-stealing support and lifecycle
241	>	* control, the main responsibility of this framework is to take
242	>	* actions when one worker is waiting to join a task stolen (or
243	>	* always held by) another.  Because we are multiplexing many
244	>	* tasks on to a pool of workers, we can't just let them block (as
245	>	* in Thread.join).  We also cannot just reassign the joiner's
246	>	* run-time stack with another and replace it later, which would
247	>	* be a form of "continuation", that even if possible is not
248	>	* necessarily a good idea since we sometimes need both an
249	>	* unblocked task and its continuation to progress. Instead we
250		* combine two tactics:
251		*
252		*   Helping: Arranging for the joiner to execute some task that it
253		*      would be running if the steal had not occurred.  Method
254	<	*      ForkJoinWorkerThread.helpJoinTask tracks joining->stealing
254	>	*      ForkJoinWorkerThread.joinTask tracks joining->stealing
255		*      links to try to find such a task.
256		*
257		*   Compensating: Unless there are already enough live threads,
258	<	*      method helpMaintainParallelism() may create or or
259	<	*      re-activate a spare thread to compensate for blocked
260	<	*      joiners until they unblock.
163	<	*
164	<	* Because the determining existence of conservatively safe
165	<	* helping targets, the availability of already-created spares,
166	<	* and the apparent need to create new spares are all racy and
167	<	* require heuristic guidance, we rely on multiple retries of
168	<	* each. Further, because it is impossible to keep exactly the
169	<	* target (parallelism) number of threads running at any given
170	<	* time, we allow compensation during joins to fail, and enlist
171	<	* all other threads to help out whenever they are not otherwise
172	<	* occupied (i.e., mainly in method preStep).
258	>	*      method tryPreBlock() may create or re-activate a spare
259	>	*      thread to compensate for blocked joiners until they
260	>	*      unblock.
261		*
262		* The ManagedBlocker extension API can't use helping so relies
263		* only on compensation in method awaitBlocker.
264		*
265	<	* The main throughput advantages of work-stealing stem from
266	<	* decentralized control -- workers mostly steal tasks from each
267	<	* other. We do not want to negate this by creating bottlenecks
268	<	* implementing other management responsibilities. So we use a
269	<	* collection of techniques that avoid, reduce, or cope well with
270	<	* contention. These entail several instances of bit-packing into
271	<	* CASable fields to maintain only the minimally required
272	<	* atomicity. To enable such packing, we restrict maximum
273	<	* parallelism to (1<<15)-1 (enabling twice this (to accommodate
274	<	* unbalanced increments and decrements) to fit into a 16 bit
275	<	* field, which is far in excess of normal operating range.  Even
276	<	* though updates to some of these bookkeeping fields do sometimes
277	<	* contend with each other, they don't normally cache-contend with
278	<	* updates to others enough to warrant memory padding or
279	<	* isolation. So they are all held as fields of ForkJoinPool
280	<	* objects.  The main capabilities are as follows:
265	>	* It is impossible to keep exactly the target parallelism number
266	>	* of threads running at any given time.  Determining the
267	>	* existence of conservatively safe helping targets, the
268	>	* availability of already-created spares, and the apparent need
269	>	* to create new spares are all racy and require heuristic
270	>	* guidance, so we rely on multiple retries of each.  Currently,
271	>	* in keeping with on-demand signalling policy, we compensate only
272	>	* if blocking would leave less than one active (non-waiting,
273	>	* non-blocked) worker. Additionally, to avoid some false alarms
274	>	* due to GC, lagging counters, system activity, etc, compensated
275	>	* blocking for joins is only attempted after a number of rechecks
276	>	* proportional to the current apparent deficit (where retries are
277	>	* interspersed with Thread.yield, for good citizenship).  The
278	>	* variable blockedCount, incremented before blocking and
279	>	* decremented after, is sometimes needed to distinguish cases of
280	>	* waiting for work vs blocking on joins or other managed sync,
281	>	* but both the cases are equivalent for most pool control, so we
282	>	* can update non-atomically. (Additionally, contention on
283	>	* blockedCount alleviates some contention on ctl).
284	>	*
285	>	* Shutdown and Termination. A call to shutdownNow atomically sets
286	>	* the ctl stop bit and then (non-atomically) sets each workers
287	>	* "terminate" status, cancels all unprocessed tasks, and wakes up
288	>	* all waiting workers.  Detecting whether termination should
289	>	* commence after a non-abrupt shutdown() call requires more work
290	>	* and bookkeeping. We need consensus about quiesence (i.e., that
291	>	* there is no more work) which is reflected in active counts so
292	>	* long as there are no current blockers, as well as possible
293	>	* re-evaluations during independent changes in blocking or
294	>	* quiescing workers.
295		*
296	<	* 1. Creating and removing workers. Workers are recorded in the
195	<	* "workers" array. This is an array as opposed to some other data
196	<	* structure to support index-based random steals by workers.
197	<	* Updates to the array recording new workers and unrecording
198	<	* terminated ones are protected from each other by a lock
199	<	* (workerLock) but the array is otherwise concurrently readable,
200	<	* and accessed directly by workers. To simplify index-based
201	<	* operations, the array size is always a power of two, and all
202	<	* readers must tolerate null slots. Currently, all worker thread
203	<	* creation is on-demand, triggered by task submissions,
204	<	* replacement of terminated workers, and/or compensation for
205	<	* blocked workers. However, all other support code is set up to
206	<	* work with other policies.
207	<	*
208	<	* To ensure that we do not hold on to worker references that
209	<	* would prevent GC, ALL accesses to workers are via indices into
210	<	* the workers array (which is one source of some of the unusual
211	<	* code constructions here). In essence, the workers array serves
212	<	* as a WeakReference mechanism. Thus for example the event queue
213	<	* stores worker indices, not worker references. Access to the
214	<	* workers in associated methods (for example releaseEventWaiters)
215	<	* must both index-check and null-check the IDs. All such accesses
216	<	* ignore bad IDs by returning out early from what they are doing,
217	<	* since this can only be associated with shutdown, in which case
218	<	* it is OK to give up. On termination, we just clobber these
219	<	* data structures without trying to use them.
220	<	*
221	<	* 2. Bookkeeping for dynamically adding and removing workers. We
222	<	* aim to approximately maintain the given level of parallelism.
223	<	* When some workers are known to be blocked (on joins or via
224	<	* ManagedBlocker), we may create or resume others to take their
225	<	* place until they unblock (see below). Implementing this
226	<	* requires counts of the number of "running" threads (i.e., those
227	<	* that are neither blocked nor artifically suspended) as well as
228	<	* the total number.  These two values are packed into one field,
229	<	* "workerCounts" because we need accurate snapshots when deciding
230	<	* to create, resume or suspend.  Note however that the
231	<	* correspondance of these counts to reality is not guaranteed. In
232	<	* particular updates for unblocked threads may lag until they
233	<	* actually wake up.
234	<	*
235	<	* 3. Maintaining global run state. The run state of the pool
236	<	* consists of a runLevel (SHUTDOWN, TERMINATING, etc) similar to
237	<	* those in other Executor implementations, as well as a count of
238	<	* "active" workers -- those that are, or soon will be, or
239	<	* recently were executing tasks. The runLevel and active count
240	<	* are packed together in order to correctly trigger shutdown and
241	<	* termination. Without care, active counts can be subject to very
242	<	* high contention.  We substantially reduce this contention by
243	<	* relaxing update rules.  A worker must claim active status
244	<	* prospectively, by activating if it sees that a submitted or
245	<	* stealable task exists (it may find after activating that the
246	<	* task no longer exists). It stays active while processing this
247	<	* task (if it exists) and any other local subtasks it produces,
248	<	* until it cannot find any other tasks. It then tries
249	<	* inactivating (see method preStep), but upon update contention
250	<	* instead scans for more tasks, later retrying inactivation if it
251	<	* doesn't find any.
252	<	*
253	<	* 4. Managing idle workers waiting for tasks. We cannot let
254	<	* workers spin indefinitely scanning for tasks when none are
255	<	* available. On the other hand, we must quickly prod them into
256	<	* action when new tasks are submitted or generated.  We
257	<	* park/unpark these idle workers using an event-count scheme.
258	<	* Field eventCount is incremented upon events that may enable
259	<	* workers that previously could not find a task to now find one:
260	<	* Submission of a new task to the pool, or another worker pushing
261	<	* a task onto a previously empty queue.  (We also use this
262	<	* mechanism for configuration and termination actions that
263	<	* require wakeups of idle workers).  Each worker maintains its
264	<	* last known event count, and blocks when a scan for work did not
265	<	* find a task AND its lastEventCount matches the current
266	<	* eventCount. Waiting idle workers are recorded in a variant of
267	<	* Treiber stack headed by field eventWaiters which, when nonzero,
268	<	* encodes the thread index and count awaited for by the worker
269	<	* thread most recently calling eventSync. This thread in turn has
270	<	* a record (field nextEventWaiter) for the next waiting worker.
271	<	* In addition to allowing simpler decisions about need for
272	<	* wakeup, the event count bits in eventWaiters serve the role of
273	<	* tags to avoid ABA errors in Treiber stacks. Upon any wakeup,
274	<	* released threads also try to release others (but give up upon
275	<	* contention to reduce useless flailing).  The net effect is a
276	<	* tree-like diffusion of signals, where released threads (and
277	<	* possibly others) help with unparks.  To further reduce
278	<	* contention effects a bit, failed CASes to increment field
279	<	* eventCount are tolerated without retries in signalWork.
280	<	* Conceptually they are merged into the same event, which is OK
281	<	* when their only purpose is to enable workers to scan for work.
282	<	*
283	<	* 5. Managing suspension of extra workers. When a worker is about
284	<	* to block waiting for a join (or via ManagedBlockers), we may
285	<	* create a new thread to maintain parallelism level, or at least
286	<	* avoid starvation. Usually, extra threads are needed for only
287	<	* very short periods, yet join dependencies are such that we
288	<	* sometimes need them in bursts. Rather than create new threads
289	<	* each time this happens, we suspend no-longer-needed extra ones
290	<	* as "spares". For most purposes, we don't distinguish "extra"
291	<	* spare threads from normal "core" threads: On each call to
292	<	* preStep (the only point at which we can do this) a worker
293	<	* checks to see if there are now too many running workers, and if
294	<	* so, suspends itself.  Method helpMaintainParallelism looks for
295	<	* suspended threads to resume before considering creating a new
296	<	* replacement. The spares themselves are encoded on another
297	<	* variant of a Treiber Stack, headed at field "spareWaiters".
298	<	* Note that the use of spares is intrinsically racy.  One thread
299	<	* may become a spare at about the same time as another is
300	<	* needlessly being created. We counteract this and related slop
301	<	* in part by requiring resumed spares to immediately recheck (in
302	<	* preStep) to see whether they they should re-suspend.
303	<	*
304	<	* 6. Killing off unneeded workers. The Spare and Event queues use
305	<	* similar mechanisms to shed unused workers: The oldest (first)
306	<	* waiter uses a timed rather than hard wait. When this wait times
307	<	* out without a normal wakeup, it tries to shutdown any one (for
308	<	* convenience the newest) other waiter via tryShutdownSpare or
309	<	* tryShutdownWaiter, respectively. The wakeup rates for spares
310	<	* are much shorter than for waiters. Together, they will
311	<	* eventually reduce the number of worker threads to a minimum of
312	<	* one after a long enough period without use.
313	<	*
314	<	* 7. Deciding when to create new workers. The main dynamic
315	<	* control in this class is deciding when to create extra threads
316	<	* in method helpMaintainParallelism. We would like to keep
317	<	* exactly #parallelism threads running, which is an impossble
318	<	* task. We always need to create one when the number of running
319	<	* threads would become zero and all workers are busy. Beyond
320	<	* this, we must rely on heuristics that work well in the the
321	<	* presence of transients phenomena such as GC stalls, dynamic
322	<	* compilation, and wake-up lags. These transients are extremely
323	<	* common -- we are normally trying to fully saturate the CPUs on
324	<	* a machine, so almost any activity other than running tasks
325	<	* impedes accuracy. Our main defense is to allow some slack in
326	<	* creation thresholds, using rules that reflect the fact that the
327	<	* more threads we have running, the more likely that we are
328	<	* underestimating the number running threads. (We also include
329	<	* some heuristic use of Thread.yield when all workers appear to
330	<	* be busy, to improve likelihood of counts settling.) The rules
331	<	* also better cope with the fact that some of the methods in this
332	<	* class tend to never become compiled (but are interpreted), so
333	<	* some components of the entire set of controls might execute 100
334	<	* times faster than others. And similarly for cases where the
335	<	* apparent lack of work is just due to GC stalls and other
336	<	* transient system activity.
337	<	*
338	<	* Beware that there is a lot of representation-level coupling
296	>	* Style notes: There is a lot of representation-level coupling
297		* among classes ForkJoinPool, ForkJoinWorkerThread, and
298	<	* ForkJoinTask.  For example, direct access to "workers" array by
298	>	* ForkJoinTask.  Most fields of ForkJoinWorkerThread maintain
299	>	* data structures managed by ForkJoinPool, so are directly
300	>	* accessed.  Conversely we allow access to "workers" array by
301		* workers, and direct access to ForkJoinTask.status by both
302		* ForkJoinPool and ForkJoinWorkerThread.  There is little point
303		* trying to reduce this, since any associated future changes in
304		* representations will need to be accompanied by algorithmic
305	<	* changes anyway.
306	<	*
307	<	* Style notes: There are lots of inline assignments (of form
308	<	* "while ((local = field) != 0)") which are usually the simplest
309	<	* way to ensure the required read orderings (which are sometimes
310	<	* critical). Also several occurrences of the unusual "do {}
311	<	* while(!cas...)" which is the simplest way to force an update of
312	<	* a CAS'ed variable. There are also other coding oddities that
313	<	* help some methods perform reasonably even when interpreted (not
314	<	* compiled), at the expense of some messy constructions that
315	<	* reduce byte code counts.
316	<	*
317	<	* The order of declarations in this file is: (1) statics (2)
318	<	* fields (along with constants used when unpacking some of them)
319	<	* (3) internal control methods (4) callbacks and other support
320	<	* for ForkJoinTask and ForkJoinWorkerThread classes, (5) exported
321	<	* methods (plus a few little helpers).
305	>	* changes anyway. All together, these low-level implementation
306	>	* choices produce as much as a factor of 4 performance
307	>	* improvement compared to naive implementations, and enable the
308	>	* processing of billions of tasks per second, at the expense of
309	>	* some ugliness.
310	>	*
311	>	* Methods signalWork() and scan() are the main bottlenecks so are
312	>	* especially heavily micro-optimized/mangled.  There are lots of
313	>	* inline assignments (of form "while ((local = field) != 0)")
314	>	* which are usually the simplest way to ensure the required read
315	>	* orderings (which are sometimes critical). This leads to a
316	>	* "C"-like style of listing declarations of these locals at the
317	>	* heads of methods or blocks.  There are several occurrences of
318	>	* the unusual "do {} while (!cas...)"  which is the simplest way
319	>	* to force an update of a CAS'ed variable. There are also other
320	>	* coding oddities that help some methods perform reasonably even
321	>	* when interpreted (not compiled).
322	>	*
323	>	* The order of declarations in this file is: (1) declarations of
324	>	* statics (2) fields (along with constants used when unpacking
325	>	* some of them), listed in an order that tends to reduce
326	>	* contention among them a bit under most JVMs.  (3) internal
327	>	* control methods (4) callbacks and other support for
328	>	* ForkJoinTask and ForkJoinWorkerThread classes, (5) exported
329	>	* methods (plus a few little helpers). (6) static block
330	>	* initializing all statics in a minimally dependent order.
331		*/
332
333		/**
#	Line 393 | Line 362 | public class ForkJoinPool extends Abstra
362		* overridden in ForkJoinPool constructors.
363		*/
364		public static final ForkJoinWorkerThreadFactory
365	<	defaultForkJoinWorkerThreadFactory =
397	<	new DefaultForkJoinWorkerThreadFactory();
365	>	defaultForkJoinWorkerThreadFactory;
366
367		/**
368		* Permission required for callers of methods that may start or
369		* kill threads.
370		*/
371	<	private static final RuntimePermission modifyThreadPermission =
404	<	new RuntimePermission("modifyThread");
371	>	private static final RuntimePermission modifyThreadPermission;
372
373		/**
374		* If there is a security manager, makes sure caller has
#	Line 416 | Line 383 | public class ForkJoinPool extends Abstra
383		/**
384		* Generator for assigning sequence numbers as pool names.
385		*/
386	<	private static final AtomicInteger poolNumberGenerator =
420	<	new AtomicInteger();
386	>	private static final AtomicInteger poolNumberGenerator;
387
388		/**
389	<	* The wakeup interval (in nanoseconds) for the oldest worker
390	<	* worker waiting for an event invokes tryShutdownWaiter to shrink
391	<	* the number of workers.  The exact value does not matter too
392	<	* much, but should be long enough to slowly release resources
393	<	* during long periods without use without disrupting normal use.
389	>	* Generator for initial random seeds for worker victim
390	>	* selection. This is used only to create initial seeds. Random
391	>	* steals use a cheaper xorshift generator per steal attempt. We
392	>	* don't expect much contention on seedGenerator, so just use a
393	>	* plain Random.
394		*/
395	<	private static final long SHRINK_RATE_NANOS =
430	<	60L * 1000L * 1000L * 1000L; // one minute
395	>	static final Random workerSeedGenerator;
396
397		/**
398	<	* Absolute bound for parallelism level. Twice this number plus
399	<	* one (i.e., 0xfff) must fit into a 16bit field to enable
400	<	* word-packing for some counts and indices.
398	>	* Array holding all worker threads in the pool.  Initialized upon
399	>	* construction. Array size must be a power of two.  Updates and
400	>	* replacements are protected by scanGuard, but the array is
401	>	* always kept in a consistent enough state to be randomly
402	>	* accessed without locking by workers performing work-stealing,
403	>	* as well as other traversal-based methods in this class, so long
404	>	* as reads memory-acquire by first reading ctl. All readers must
405	>	* tolerate that some array slots may be null.
406		*/
407	<	private static final int MAX_WORKERS   = 0x7fff;
407	>	ForkJoinWorkerThread[] workers;
408
409		/**
410	<	* Array holding all worker threads in the pool.  Array size must
411	<	* be a power of two.  Updates and replacements are protected by
412	<	* workerLock, but the array is always kept in a consistent enough
443	<	* state to be randomly accessed without locking by workers
444	<	* performing work-stealing, as well as other traversal-based
445	<	* methods in this class. All readers must tolerate that some
446	<	* array slots may be null.
410	>	* Initial size for submission queue array. Must be a power of
411	>	* two.  In many applications, these always stay small so we use a
412	>	* small initial cap.
413		*/
414	<	volatile ForkJoinWorkerThread[] workers;
414	>	private static final int INITIAL_QUEUE_CAPACITY = 8;
415
416		/**
417	<	* Queue for external submissions.
417	>	* Maximum size for submission queue array. Must be a power of two
418	>	* less than or equal to 1 << (31 - width of array entry) to
419	>	* ensure lack of index wraparound, but is capped at a lower
420	>	* value to help users trap runaway computations.
421		*/
422	<	private final LinkedTransferQueue<ForkJoinTask<?>> submissionQueue;
422	>	private static final int MAXIMUM_QUEUE_CAPACITY = 1 << 24; // 16M
423
424		/**
425	<	* Lock protecting updates to workers array.
425	>	* Array serving as submission queue. Initialized upon construction.
426		*/
427	<	private final ReentrantLock workerLock;
427	>	private ForkJoinTask<?>[] submissionQueue;
428
429		/**
430	<	* Latch released upon termination.
430	>	* Lock protecting submissions array for addSubmission
431		*/
432	<	private final Phaser termination;
432	>	private final ReentrantLock submissionLock;
433
434		/**
435	<	* Creation factory for worker threads.
435	>	* Condition for awaitTermination, using submissionLock for
436	>	* convenience.
437		*/
438	<	private final ForkJoinWorkerThreadFactory factory;
438	>	private final Condition termination;
439
440		/**
441	<	* Sum of per-thread steal counts, updated only when threads are
472	<	* idle or terminating.
441	>	* Creation factory for worker threads.
442		*/
443	<	private volatile long stealCount;
443	>	private final ForkJoinWorkerThreadFactory factory;
444
445		/**
446	<	* Encoded record of top of treiber stack of threads waiting for
447	<	* events. The top 32 bits contain the count being waited for. The
479	<	* bottom 16 bits contains one plus the pool index of waiting
480	<	* worker thread. (Bits 16-31 are unused.)
446	>	* The uncaught exception handler used when any worker abruptly
447	>	* terminates.
448		*/
449	<	private volatile long eventWaiters;
483	<
484	<	private static final int  EVENT_COUNT_SHIFT = 32;
485	<	private static final long WAITER_ID_MASK    = (1L << 16) - 1L;
449	>	final Thread.UncaughtExceptionHandler ueh;
450
451		/**
452	<	* A counter for events that may wake up worker threads:
489	<	*   - Submission of a new task to the pool
490	<	*   - A worker pushing a task on an empty queue
491	<	*   - termination
452	>	* Prefix for assigning names to worker threads
453		*/
454	<	private volatile int eventCount;
454	>	private final String workerNamePrefix;
455
456		/**
457	<	* Encoded record of top of treiber stack of spare threads waiting
458	<	* for resumption. The top 16 bits contain an arbitrary count to
498	<	* avoid ABA effects. The bottom 16bits contains one plus the pool
499	<	* index of waiting worker thread.
457	>	* Sum of per-thread steal counts, updated only when threads are
458	>	* idle or terminating.
459		*/
460	<	private volatile int spareWaiters;
460	>	private volatile long stealCount;
461
462	<	private static final int SPARE_COUNT_SHIFT = 16;
463	<	private static final int SPARE_ID_MASK     = (1 << 16) - 1;
462	>	/**
463	>	* Main pool control -- a long packed with:
464	>	* AC: Number of active running workers minus target parallelism (16 bits)
465	>	* TC: Number of total workers minus target parallelism (16bits)
466	>	* ST: true if pool is terminating (1 bit)
467	>	* EC: the wait count of top waiting thread (15 bits)
468	>	* ID: ~poolIndex of top of Treiber stack of waiting threads (16 bits)
469	>	*
470	>	* When convenient, we can extract the upper 32 bits of counts and
471	>	* the lower 32 bits of queue state, u = (int)(ctl >>> 32) and e =
472	>	* (int)ctl.  The ec field is never accessed alone, but always
473	>	* together with id and st. The offsets of counts by the target
474	>	* parallelism and the positionings of fields makes it possible to
475	>	* perform the most common checks via sign tests of fields: When
476	>	* ac is negative, there are not enough active workers, when tc is
477	>	* negative, there are not enough total workers, when id is
478	>	* negative, there is at least one waiting worker, and when e is
479	>	* negative, the pool is terminating.  To deal with these possibly
480	>	* negative fields, we use casts in and out of "short" and/or
481	>	* signed shifts to maintain signedness.  Note: AC_SHIFT is
482	>	* redundantly declared in ForkJoinWorkerThread in order to
483	>	* integrate a surplus-threads check.
484	>	*/
485	>	volatile long ctl;
486	>
487	>	// bit positions/shifts for fields
488	>	private static final int  AC_SHIFT   = 48;
489	>	private static final int  TC_SHIFT   = 32;
490	>	private static final int  ST_SHIFT   = 31;
491	>	private static final int  EC_SHIFT   = 16;
492	>
493	>	// bounds
494	>	private static final int  MAX_ID     = 0x7fff;  // max poolIndex
495	>	private static final int  SMASK      = 0xffff;  // mask short bits
496	>	private static final int  SHORT_SIGN = 1 << 15;
497	>	private static final int  INT_SIGN   = 1 << 31;
498	>
499	>	// masks
500	>	private static final long STOP_BIT   = 0x0001L << ST_SHIFT;
501	>	private static final long AC_MASK    = ((long)SMASK) << AC_SHIFT;
502	>	private static final long TC_MASK    = ((long)SMASK) << TC_SHIFT;
503	>
504	>	// units for incrementing and decrementing
505	>	private static final long TC_UNIT    = 1L << TC_SHIFT;
506	>	private static final long AC_UNIT    = 1L << AC_SHIFT;
507	>
508	>	// masks and units for dealing with u = (int)(ctl >>> 32)
509	>	private static final int  UAC_SHIFT  = AC_SHIFT - 32;
510	>	private static final int  UTC_SHIFT  = TC_SHIFT - 32;
511	>	private static final int  UAC_MASK   = SMASK << UAC_SHIFT;
512	>	private static final int  UTC_MASK   = SMASK << UTC_SHIFT;
513	>	private static final int  UAC_UNIT   = 1 << UAC_SHIFT;
514	>	private static final int  UTC_UNIT   = 1 << UTC_SHIFT;
515	>
516	>	// masks and units for dealing with e = (int)ctl
517	>	private static final int  E_MASK     = 0x7fffffff; // no STOP_BIT
518	>	private static final int  EC_UNIT    = 1 << EC_SHIFT;
519
520		/**
521	<	* Lifecycle control. The low word contains the number of workers
508	<	* that are (probably) executing tasks. This value is atomically
509	<	* incremented before a worker gets a task to run, and decremented
510	<	* when worker has no tasks and cannot find any.  Bits 16-18
511	<	* contain runLevel value. When all are zero, the pool is
512	<	* running. Level transitions are monotonic (running -> shutdown
513	<	* -> terminating -> terminated) so each transition adds a bit.
514	<	* These are bundled together to ensure consistent read for
515	<	* termination checks (i.e., that runLevel is at least SHUTDOWN
516	<	* and active threads is zero).
517	<	*
518	<	* Notes: Most direct CASes are dependent on these bitfield
519	<	* positions.  Also, this field is non-private to enable direct
520	<	* performance-sensitive CASes in ForkJoinWorkerThread.
521	>	* The target parallelism level.
522		*/
523	<	volatile int runState;
523	<
524	<	// Note: The order among run level values matters.
525	<	private static final int RUNLEVEL_SHIFT     = 16;
526	<	private static final int SHUTDOWN           = 1 << RUNLEVEL_SHIFT;
527	<	private static final int TERMINATING        = 1 << (RUNLEVEL_SHIFT + 1);
528	<	private static final int TERMINATED         = 1 << (RUNLEVEL_SHIFT + 2);
529	<	private static final int ACTIVE_COUNT_MASK  = (1 << RUNLEVEL_SHIFT) - 1;
523	>	final int parallelism;
524
525		/**
526	<	* Holds number of total (i.e., created and not yet terminated)
527	<	* and running (i.e., not blocked on joins or other managed sync)
534	<	* threads, packed together to ensure consistent snapshot when
535	<	* making decisions about creating and suspending spare
536	<	* threads. Updated only by CAS. Note that adding a new worker
537	<	* requires incrementing both counts, since workers start off in
538	<	* running state.
526	>	* Index (mod submission queue length) of next element to take
527	>	* from submission queue.
528		*/
529	<	private volatile int workerCounts;
529	>	volatile int queueBase;
530
531	<	private static final int TOTAL_COUNT_SHIFT  = 16;
532	<	private static final int RUNNING_COUNT_MASK = (1 << TOTAL_COUNT_SHIFT) - 1;
533	<	private static final int ONE_RUNNING        = 1;
534	<	private static final int ONE_TOTAL          = 1 << TOTAL_COUNT_SHIFT;
531	>	/**
532	>	* Index (mod submission queue length) of next element to add
533	>	* in submission queue.
534	>	*/
535	>	int queueTop;
536
537		/**
538	<	* The target parallelism level.
549	<	* Accessed directly by ForkJoinWorkerThreads.
538	>	* True when shutdown() has been called.
539		*/
540	<	final int parallelism;
540	>	volatile boolean shutdown;
541
542		/**
543		* True if use local fifo, not default lifo, for local polling
#	Line 557 | Line 546 | public class ForkJoinPool extends Abstra
546		final boolean locallyFifo;
547
548		/**
549	<	* The uncaught exception handler used when any worker abruptly
550	<	* terminates.
549	>	* The number of threads in ForkJoinWorkerThreads.helpQuiescePool.
550	>	* When non-zero, suppresses automatic shutdown when active
551	>	* counts become zero.
552		*/
553	<	private final Thread.UncaughtExceptionHandler ueh;
553	>	volatile int quiescerCount;
554
555		/**
556	<	* Pool number, just for assigning useful names to worker threads
556	>	* The number of threads blocked in join.
557		*/
558	<	private final int poolNumber;
569	<
570	<
571	<	// Utilities for CASing fields. Note that most of these
572	<	// are usually manually inlined by callers
558	>	volatile int blockedCount;
559
560		/**
561	<	* Increments running count part of workerCounts
561	>	* Counter for worker Thread names (unrelated to their poolIndex)
562		*/
563	<	final void incrementRunningCount() {
578	<	int c;
579	<	do {} while (!UNSAFE.compareAndSwapInt(this, workerCountsOffset,
580	<	c = workerCounts,
581	<	c + ONE_RUNNING));
582	<	}
563	>	private volatile int nextWorkerNumber;
564
565		/**
566	<	* Tries to decrement running count unless already zero
566	>	* The index for the next created worker. Accessed under scanGuard.
567		*/
568	<	final boolean tryDecrementRunningCount() {
588	<	int wc = workerCounts;
589	<	if ((wc & RUNNING_COUNT_MASK) == 0)
590	<	return false;
591	<	return UNSAFE.compareAndSwapInt(this, workerCountsOffset,
592	<	wc, wc - ONE_RUNNING);
593	<	}
568	>	private int nextWorkerIndex;
569
570		/**
571	<	* Forces decrement of encoded workerCounts, awaiting nonzero if
572	<	* (rarely) necessary when other count updates lag.
573	<	*
574	<	* @param dr -- either zero or ONE_RUNNING
575	<	* @param dt == either zero or ONE_TOTAL
571	>	* SeqLock and index masking for for updates to workers array.
572	>	* Locked when SG_UNIT is set. Unlocking clears bit by adding
573	>	* SG_UNIT. Staleness of read-only operations can be checked by
574	>	* comparing scanGuard to value before the reads. The low 16 bits
575	>	* (i.e, anding with SMASK) hold (the smallest power of two
576	>	* covering all worker indices, minus one, and is used to avoid
577	>	* dealing with large numbers of null slots when the workers array
578	>	* is overallocated.
579		*/
580	<	private void decrementWorkerCounts(int dr, int dt) {
603	<	for (;;) {
604	<	int wc = workerCounts;
605	<	if ((wc & RUNNING_COUNT_MASK)  - dr < 0 ||
606	<	(wc >>> TOTAL_COUNT_SHIFT) - dt < 0) {
607	<	if ((runState & TERMINATED) != 0)
608	<	return; // lagging termination on a backout
609	<	Thread.yield();
610	<	}
611	<	if (UNSAFE.compareAndSwapInt(this, workerCountsOffset,
612	<	wc, wc - (dr + dt)))
613	<	return;
614	<	}
615	<	}
580	>	volatile int scanGuard;
581
582	<	/**
618	<	* Increments event count
619	<	*/
620	<	private void advanceEventCount() {
621	<	int c;
622	<	do {} while(!UNSAFE.compareAndSwapInt(this, eventCountOffset,
623	<	c = eventCount, c+1));
624	<	}
625	<
626	<	/**
627	<	* Tries incrementing active count; fails on contention.
628	<	* Called by workers before executing tasks.
629	<	*
630	<	* @return true on success
631	<	*/
632	<	final boolean tryIncrementActiveCount() {
633	<	int c;
634	<	return UNSAFE.compareAndSwapInt(this, runStateOffset,
635	<	c = runState, c + 1);
636	<	}
582	>	private static final int SG_UNIT = 1 << 16;
583
584		/**
585	<	* Tries decrementing active count; fails on contention.
586	<	* Called when workers cannot find tasks to run.
585	>	* The wakeup interval (in nanoseconds) for a worker waiting for a
586	>	* task when the pool is quiescent to instead try to shrink the
587	>	* number of workers.  The exact value does not matter too
588	>	* much. It must be short enough to release resources during
589	>	* sustained periods of idleness, but not so short that threads
590	>	* are continually re-created.
591		*/
592	<	final boolean tryDecrementActiveCount() {
593	<	int c;
644	<	return UNSAFE.compareAndSwapInt(this, runStateOffset,
645	<	c = runState, c - 1);
646	<	}
592	>	private static final long SHRINK_RATE =
593	>	4L * 1000L * 1000L * 1000L; // 4 seconds
594
595		/**
596	<	* Advances to at least the given level. Returns true if not
597	<	* already in at least the given level.
596	>	* Top-level loop for worker threads: On each step: if the
597	>	* previous step swept through all queues and found no tasks, or
598	>	* there are excess threads, then possibly blocks. Otherwise,
599	>	* scans for and, if found, executes a task. Returns when pool
600	>	* and/or worker terminate.
601	>	*
602	>	* @param w the worker
603		*/
604	<	private boolean advanceRunLevel(int level) {
605	<	for (;;) {
606	<	int s = runState;
607	<	if ((s & level) != 0)
608	<	return false;
609	<	if (UNSAFE.compareAndSwapInt(this, runStateOffset, s, s | level))
610	<	return true;
604	>	final void work(ForkJoinWorkerThread w) {
605	>	boolean swept = false;                // true on empty scans
606	>	long c;
607	>	while (!w.terminate && (int)(c = ctl) >= 0) {
608	>	int a;                            // active count
609	>	if (!swept && (a = (int)(c >> AC_SHIFT)) <= 0)
610	>	swept = scan(w, a);
611	>	else if (tryAwaitWork(w, c))
612	>	swept = false;
613		}
614		}
615
616	<	// workers array maintenance
616	>	// Signalling
617
618		/**
619	<	* Records and returns a workers array index for new worker.
619	>	* Wakes up or creates a worker.
620		*/
621	<	private int recordWorker(ForkJoinWorkerThread w) {
622	<	// Try using slot totalCount-1. If not available, scan and/or resize
623	<	int k = (workerCounts >>> TOTAL_COUNT_SHIFT) - 1;
624	<	final ReentrantLock lock = this.workerLock;
625	<	lock.lock();
626	<	try {
627	<	ForkJoinWorkerThread[] ws = workers;
628	<	int n = ws.length;
629	<	if (k < 0 || k >= n || ws[k] != null) {
630	<	for (k = 0; k < n && ws[k] != null; ++k)
631	<	;
632	<	if (k == n)
633	<	ws = Arrays.copyOf(ws, n << 1);
621	>	final void signalWork() {
622	>	/*
623	>	* The while condition is true if: (there is are too few total
624	>	* workers OR there is at least one waiter) AND (there are too
625	>	* few active workers OR the pool is terminating).  The value
626	>	* of e distinguishes the remaining cases: zero (no waiters)
627	>	* for create, negative if terminating (in which case do
628	>	* nothing), else release a waiter. The secondary checks for
629	>	* release (non-null array etc) can fail if the pool begins
630	>	* terminating after the test, and don't impose any added cost
631	>	* because JVMs must perform null and bounds checks anyway.
632	>	*/
633	>	long c; int e, u;
634	>	while ((((e = (int)(c = ctl)) | (u = (int)(c >>> 32))) &
635	>	(INT_SIGN|SHORT_SIGN)) == (INT_SIGN|SHORT_SIGN) && e >= 0) {
636	>	if (e > 0) {                         // release a waiting worker
637	>	int i; ForkJoinWorkerThread w; ForkJoinWorkerThread[] ws;
638	>	if ((ws = workers) == null ||
639	>	(i = ~e & SMASK) >= ws.length ||
640	>	(w = ws[i]) == null)
641	>	break;
642	>	long nc = (((long)(w.nextWait & E_MASK)) |
643	>	((long)(u + UAC_UNIT) << 32));
644	>	if (w.eventCount == e &&
645	>	UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc)) {
646	>	w.eventCount = (e + EC_UNIT) & E_MASK;
647	>	if (w.parked)
648	>	UNSAFE.unpark(w);
649	>	break;
650	>	}
651	>	}
652	>	else if (UNSAFE.compareAndSwapLong
653	>	(this, ctlOffset, c,
654	>	(long)(((u + UTC_UNIT) & UTC_MASK) |
655	>	((u + UAC_UNIT) & UAC_MASK)) << 32)) {
656	>	addWorker();
657	>	break;
658		}
681	–	ws[k] = w;
682	–	workers = ws; // volatile array write ensures slot visibility
683	–	} finally {
684	–	lock.unlock();
659		}
686	–	return k;
660		}
661
662		/**
663	<	* Nulls out record of worker in workers array
663	>	* Variant of signalWork to help release waiters on rescans.
664	>	* Tries once to release a waiter if active count < 0.
665	>	*
666	>	* @return false if failed due to contention, else true
667		*/
668	<	private void forgetWorker(ForkJoinWorkerThread w) {
669	<	int idx = w.poolIndex;
670	<	// Locking helps method recordWorker avoid unecessary expansion
671	<	final ReentrantLock lock = this.workerLock;
672	<	lock.lock();
673	<	try {
674	<	ForkJoinWorkerThread[] ws = workers;
675	<	if (idx >= 0 && idx < ws.length && ws[idx] == w) // verify
676	<	ws[idx] = null;
677	<	} finally {
678	<	lock.unlock();
668	>	private boolean tryReleaseWaiter() {
669	>	long c; int e, i; ForkJoinWorkerThread w; ForkJoinWorkerThread[] ws;
670	>	if ((e = (int)(c = ctl)) > 0 &&
671	>	(int)(c >> AC_SHIFT) < 0 &&
672	>	(ws = workers) != null &&
673	>	(i = ~e & SMASK) < ws.length &&
674	>	(w = ws[i]) != null) {
675	>	long nc = ((long)(w.nextWait & E_MASK) |
676	>	((c + AC_UNIT) & (AC_MASK|TC_MASK)));
677	>	if (w.eventCount != e ||
678	>	!UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc))
679	>	return false;
680	>	w.eventCount = (e + EC_UNIT) & E_MASK;
681	>	if (w.parked)
682	>	UNSAFE.unpark(w);
683		}
684	+	return true;
685		}
686
687	<	// adding and removing workers
687	>	// Scanning for tasks
688
689		/**
690	<	* Tries to create and add new worker. Assumes that worker counts
691	<	* are already updated to accommodate the worker, so adjusts on
692	<	* failure.
690	>	* Scans for and, if found, executes one task. Scans start at a
691	>	* random index of workers array, and randomly select the first
692	>	* (2*#workers)-1 probes, and then, if all empty, resort to 2
693	>	* circular sweeps, which is necessary to check quiescence. and
694	>	* taking a submission only if no stealable tasks were found.  The
695	>	* steal code inside the loop is a specialized form of
696	>	* ForkJoinWorkerThread.deqTask, followed bookkeeping to support
697	>	* helpJoinTask and signal propagation. The code for submission
698	>	* queues is almost identical. On each steal, the worker completes
699	>	* not only the task, but also all local tasks that this task may
700	>	* have generated. On detecting staleness or contention when
701	>	* trying to take a task, this method returns without finishing
702	>	* sweep, which allows global state rechecks before retry.
703		*
704	<	* @return the worker, or null on failure
704	>	* @param w the worker
705	>	* @param a the number of active workers
706	>	* @return true if swept all queues without finding a task
707		*/
708	<	private ForkJoinWorkerThread addWorker() {
709	<	ForkJoinWorkerThread w = null;
710	<	try {
711	<	w = factory.newThread(this);
712	<	} finally { // Adjust on either null or exceptional factory return
713	<	if (w == null) {
714	<	decrementWorkerCounts(ONE_RUNNING, ONE_TOTAL);
715	<	tryTerminate(false); // in case of failure during shutdown
708	>	private boolean scan(ForkJoinWorkerThread w, int a) {
709	>	int g = scanGuard; // mask 0 avoids useless scans if only one active
710	>	int m = parallelism == 1 - a? 0 : g & SMASK;
711	>	ForkJoinWorkerThread[] ws = workers;
712	>	if (ws == null || ws.length <= m)         // staleness check
713	>	return false;
714	>	for (int r = w.seed, k = r, j = -(m + m); j <= m + m; ++j) {
715	>	ForkJoinTask<?> t; ForkJoinTask<?>[] q; int b, i;
716	>	ForkJoinWorkerThread v = ws[k & m];
717	>	if (v != null && (b = v.queueBase) != v.queueTop &&
718	>	(q = v.queue) != null && (i = (q.length - 1) & b) >= 0) {
719	>	long u = (i << ASHIFT) + ABASE;
720	>	if ((t = q[i]) != null && v.queueBase == b &&
721	>	UNSAFE.compareAndSwapObject(q, u, t, null)) {
722	>	int d = (v.queueBase = b + 1) - v.queueTop;
723	>	v.stealHint = w.poolIndex;
724	>	if (d != 0)
725	>	signalWork();             // propagate if nonempty
726	>	w.execTask(t);
727	>	}
728	>	r ^= r << 13; r ^= r >>> 17; w.seed = r ^ (r << 5);
729	>	return false;                     // store next seed
730		}
731	+	else if (j < 0) {                     // xorshift
732	+	r ^= r << 13; r ^= r >>> 17; k = r ^= r << 5;
733	+	}
734	+	else
735	+	++k;
736		}
737	<	if (w != null) {
738	<	w.start(recordWorker(w), ueh);
739	<	advanceEventCount();
737	>	if (scanGuard != g)                       // staleness check
738	>	return false;
739	>	else {                                    // try to take submission
740	>	ForkJoinTask<?> t; ForkJoinTask<?>[] q; int b, i;
741	>	if ((b = queueBase) != queueTop &&
742	>	(q = submissionQueue) != null &&
743	>	(i = (q.length - 1) & b) >= 0) {
744	>	long u = (i << ASHIFT) + ABASE;
745	>	if ((t = q[i]) != null && queueBase == b &&
746	>	UNSAFE.compareAndSwapObject(q, u, t, null)) {
747	>	queueBase = b + 1;
748	>	w.execTask(t);
749	>	}
750	>	return false;
751	>	}
752	>	return true;                         // all queues empty
753		}
729	–	return w;
754		}
755
756		/**
757	<	* Final callback from terminating worker.  Removes record of
758	<	* worker from array, and adjusts counts. If pool is shutting
759	<	* down, tries to complete terminatation.
757	>	* Tries to enqueue worker in wait queue and await change in
758	>	* worker's eventCount.  Before blocking, rescans queues to avoid
759	>	* missed signals.  If the pool is quiescent, possibly terminates
760	>	* worker upon exit.
761		*
762	<	* @param w the worker
762	>	* @param w the calling worker
763	>	* @param c the ctl value on entry
764	>	* @return true if waited or another thread was released upon enq
765		*/
766	<	final void workerTerminated(ForkJoinWorkerThread w) {
767	<	forgetWorker(w);
768	<	decrementWorkerCounts(w.isTrimmed()? 0 : ONE_RUNNING, ONE_TOTAL);
769	<	while (w.stealCount != 0) // collect final count
770	<	tryAccumulateStealCount(w);
771	<	tryTerminate(false);
772	<	}
773	<
774	<	// Waiting for and signalling events
766	>	private boolean tryAwaitWork(ForkJoinWorkerThread w, long c) {
767	>	int v = w.eventCount;
768	>	w.nextWait = (int)c;                       // w's successor record
769	>	long nc = (long)(v & E_MASK) | ((c - AC_UNIT) & (AC_MASK|TC_MASK));
770	>	if (ctl != c || !UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc)) {
771	>	long d = ctl; // return true if lost to a deq, to force rescan
772	>	return (int)d != (int)c && ((d - c) & AC_MASK) >= 0L;
773	>	}
774	>	if (parallelism + (int)(c >> AC_SHIFT) == 1 &&
775	>	blockedCount == 0 && quiescerCount == 0)
776	>	idleAwaitWork(w, v);               // quiescent -- maybe shrink
777
778	<	/**
779	<	* Releases workers blocked on a count not equal to current count.
780	<	* Normally called after precheck that eventWaiters isn't zero to
781	<	* avoid wasted array checks. Gives up upon a change in count or
782	<	* contention, letting other workers take over.
783	<	*/
784	<	private void releaseEventWaiters() {
785	<	ForkJoinWorkerThread[] ws = workers;
786	<	int n = ws.length;
787	<	long h = eventWaiters;
788	<	int ec = eventCount;
789	<	ForkJoinWorkerThread w; int id;
790	<	while ((int)(h >>> EVENT_COUNT_SHIFT) != ec &&
791	<	(id = ((int)(h & WAITER_ID_MASK)) - 1) >= 0 &&
792	<	id < n && (w = ws[id]) != null &&
793	<	UNSAFE.compareAndSwapLong(this, eventWaitersOffset,
794	<	h,  h = w.nextWaiter)) {
795	<	LockSupport.unpark(w);
796	<	if (eventWaiters != h || eventCount != ec)
797	<	break;
778	>	boolean rescanned = false;
779	>	for (int sc;;) {
780	>	if (w.eventCount != v)
781	>	return true;
782	>	if ((sc = w.stealCount) != 0) {
783	>	long s = stealCount;               // accumulate stealCount
784	>	if (UNSAFE.compareAndSwapLong(this, stealCountOffset, s, s+sc))
785	>	w.stealCount = 0;
786	>	}
787	>	else if (!rescanned) {
788	>	int g = scanGuard, m = g & SMASK;
789	>	ForkJoinWorkerThread[] ws = workers;
790	>	if (ws != null && m < ws.length) {
791	>	rescanned = true;
792	>	for (int i = 0; i <= m; ++i) {
793	>	ForkJoinWorkerThread u = ws[i];
794	>	if (u != null) {
795	>	if (u.queueBase != u.queueTop &&
796	>	!tryReleaseWaiter())
797	>	rescanned = false; // contended
798	>	if (w.eventCount != v)
799	>	return true;
800	>	}
801	>	}
802	>	}
803	>	if (scanGuard != g ||              // stale
804	>	(queueBase != queueTop && !tryReleaseWaiter()))
805	>	rescanned = false;
806	>	if (!rescanned)
807	>	Thread.yield();                // reduce contention
808	>	else
809	>	Thread.interrupted();          // clear before park
810	>	}
811	>	else {
812	>	w.parked = true;                   // must recheck
813	>	if (w.eventCount != v) {
814	>	w.parked = false;
815	>	return true;
816	>	}
817	>	LockSupport.park(this);
818	>	rescanned = w.parked = false;
819	>	}
820		}
821		}
822
823		/**
824	<	* Tries to advance eventCount and releases waiters. Called only
825	<	* from workers.
826	<	*/
827	<	final void signalWork() {
828	<	int c; // try to increment event count -- CAS failure OK
829	<	UNSAFE.compareAndSwapInt(this, eventCountOffset, c = eventCount, c+1);
830	<	if (eventWaiters != 0L)
831	<	releaseEventWaiters();
824	>	* If pool is quiescent, checks for termination, and waits for
825	>	* event signal for up to SHRINK_RATE nanosecs. On timeout, if ctl
826	>	* has not changed, terminates the worker. Upon its termination
827	>	* (see deregisterWorker), it may wake up another worker to
828	>	* possibly repeat this process.
829	>	*
830	>	* @param w the calling worker
831	>	* @param v the eventCount w must wait until changed
832	>	*/
833	>	private void idleAwaitWork(ForkJoinWorkerThread w, int v) {
834	>	ForkJoinTask.helpExpungeStaleExceptions(); // help clean weak refs
835	>	if (shutdown)
836	>	tryTerminate(false);
837	>	long c = ctl;
838	>	long nc = (((c & (AC_MASK|TC_MASK)) + AC_UNIT) |
839	>	(long)(w.nextWait & E_MASK)); // ctl value to release w
840	>	if (w.eventCount == v &&
841	>	parallelism + (int)(c >> AC_SHIFT) == 0 &&
842	>	blockedCount == 0 && quiescerCount == 0) {
843	>	long startTime = System.nanoTime();
844	>	Thread.interrupted();
845	>	if (w.eventCount == v) {
846	>	w.parked = true;
847	>	if (w.eventCount == v)
848	>	LockSupport.parkNanos(this, SHRINK_RATE);
849	>	w.parked = false;
850	>	if (w.eventCount == v && ctl == c &&
851	>	System.nanoTime() - startTime >= SHRINK_RATE &&
852	>	UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc)) {
853	>	w.terminate = true;
854	>	w.eventCount = ((int)c + EC_UNIT) & E_MASK;
855	>	}
856	>	}
857	>	}
858		}
859
860	+	// Submissions
861	+
862		/**
863	<	* Adds the given worker to event queue and blocks until
864	<	* terminating or event count advances from the workers
865	<	* lastEventCount value
866	<	*
867	<	* @param w the calling worker thread
868	<	*/
869	<	private void eventSync(ForkJoinWorkerThread w) {
870	<	int ec = w.lastEventCount;
871	<	long nh = (((long)ec) << EVENT_COUNT_SHIFT) | ((long)(w.poolIndex+1));
872	<	long h;
873	<	while ((runState < SHUTDOWN || !tryTerminate(false)) &&
874	<	(((int)((h = eventWaiters) & WAITER_ID_MASK)) == 0 ||
875	<	(int)(h >>> EVENT_COUNT_SHIFT) == ec) &&
876	<	eventCount == ec) {
877	<	if (UNSAFE.compareAndSwapLong(this, eventWaitersOffset,
878	<	w.nextWaiter = h, nh)) {
800	<	awaitEvent(w, ec);
801	<	break;
863	>	* Enqueues the given task in the submissionQueue.  Same idea as
864	>	* ForkJoinWorkerThread.pushTask except for use of submissionLock.
865	>	*
866	>	* @param t the task
867	>	*/
868	>	private void addSubmission(ForkJoinTask<?> t) {
869	>	final ReentrantLock lock = this.submissionLock;
870	>	lock.lock();
871	>	try {
872	>	ForkJoinTask<?>[] q; int s, m;
873	>	if ((q = submissionQueue) != null) {    // ignore if queue removed
874	>	long u = (((s = queueTop) & (m = q.length-1)) << ASHIFT)+ABASE;
875	>	UNSAFE.putOrderedObject(q, u, t);
876	>	queueTop = s + 1;
877	>	if (s - queueBase == m)
878	>	growSubmissionQueue();
879		}
880	+	} finally {
881	+	lock.unlock();
882		}
883	+	signalWork();
884		}
885
886	+	//  (pollSubmission is defined below with exported methods)
887	+
888		/**
889	<	* Blocks the given worker (that has already been entered as an
890	<	* event waiter) until terminating or event count advances from
891	<	* the given value. The oldest (first) waiter uses a timed wait to
892	<	* occasionally one-by-one shrink the number of workers (to a
893	<	* minumum of one) if the pool has not been used for extended
894	<	* periods.
895	<	*
896	<	* @param w the calling worker thread
897	<	* @param ec the count
898	<	*/
899	<	private void awaitEvent(ForkJoinWorkerThread w, int ec) {
900	<	while (eventCount == ec) {
901	<	if (tryAccumulateStealCount(w)) { // transfer while idle
902	<	boolean untimed = (w.nextWaiter != 0L ||
903	<	(workerCounts & RUNNING_COUNT_MASK) <= 1);
904	<	long startTime = untimed? 0 : System.nanoTime();
905	<	Thread.interrupted();         // clear/ignore interrupt
906	<	if (eventCount != ec || !w.isRunning() ||
907	<	runState >= TERMINATING)  // recheck after clear
908	<	break;
909	<	if (untimed)
828	<	LockSupport.park(w);
829	<	else {
830	<	LockSupport.parkNanos(w, SHRINK_RATE_NANOS);
831	<	if (eventCount != ec || !w.isRunning() ||
832	<	runState >= TERMINATING)
833	<	break;
834	<	if (System.nanoTime() - startTime >= SHRINK_RATE_NANOS)
835	<	tryShutdownWaiter(ec);
836	<	}
889	>	* Creates or doubles submissionQueue array.
890	>	* Basically identical to ForkJoinWorkerThread version
891	>	*/
892	>	private void growSubmissionQueue() {
893	>	ForkJoinTask<?>[] oldQ = submissionQueue;
894	>	int size = oldQ != null ? oldQ.length << 1 : INITIAL_QUEUE_CAPACITY;
895	>	if (size > MAXIMUM_QUEUE_CAPACITY)
896	>	throw new RejectedExecutionException("Queue capacity exceeded");
897	>	if (size < INITIAL_QUEUE_CAPACITY)
898	>	size = INITIAL_QUEUE_CAPACITY;
899	>	ForkJoinTask<?>[] q = submissionQueue = new ForkJoinTask<?>[size];
900	>	int mask = size - 1;
901	>	int top = queueTop;
902	>	int oldMask;
903	>	if (oldQ != null && (oldMask = oldQ.length - 1) >= 0) {
904	>	for (int b = queueBase; b != top; ++b) {
905	>	long u = ((b & oldMask) << ASHIFT) + ABASE;
906	>	Object x = UNSAFE.getObjectVolatile(oldQ, u);
907	>	if (x != null && UNSAFE.compareAndSwapObject(oldQ, u, x, null))
908	>	UNSAFE.putObjectVolatile
909	>	(q, ((b & mask) << ASHIFT) + ABASE, x);
910		}
911		}
912		}
913
914	+	// Blocking support
915	+
916		/**
917	<	* Callback from the oldest waiter in awaitEvent waking up after a
918	<	* period of non-use. Tries (once) to shutdown an event waiter (or
919	<	* a spare, if one exists). Note that we don't need CAS or locks
920	<	* here because the method is called only from one thread
921	<	* occasionally waking (and even misfires are OK). Note that
922	<	* until the shutdown worker fully terminates, workerCounts
923	<	* will overestimate total count, which is tolerable.
924	<	*
925	<	* @param ec the event count waited on by caller (to abort
926	<	* attempt if count has since changed).
927	<	*/
928	<	private void tryShutdownWaiter(int ec) {
929	<	if (spareWaiters != 0) { // prefer killing spares
930	<	tryShutdownSpare();
931	<	return;
932	<	}
933	<	ForkJoinWorkerThread[] ws = workers;
934	<	int n = ws.length;
935	<	long h = eventWaiters;
936	<	ForkJoinWorkerThread w; int id; long nh;
937	<	if (runState == 0 &&
938	<	submissionQueue.isEmpty() &&
939	<	eventCount == ec &&
940	<	(id = ((int)(h & WAITER_ID_MASK)) - 1) >= 0 &&
941	<	id < n && (w = ws[id]) != null &&
942	<	(nh = w.nextWaiter) != 0L && // keep at least one worker
943	<	UNSAFE.compareAndSwapLong(this, eventWaitersOffset, h, nh)) {
944	<	w.shutdown();
945	<	LockSupport.unpark(w);
917	>	* Tries to increment blockedCount, decrement active count
918	>	* (sometimes implicitly) and possibly release or create a
919	>	* compensating worker in preparation for blocking. Fails
920	>	* on contention or termination.
921	>	*
922	>	* @return true if the caller can block, else should recheck and retry
923	>	*/
924	>	private boolean tryPreBlock() {
925	>	int b = blockedCount;
926	>	if (UNSAFE.compareAndSwapInt(this, blockedCountOffset, b, b + 1)) {
927	>	int pc = parallelism;
928	>	do {
929	>	ForkJoinWorkerThread[] ws; ForkJoinWorkerThread w;
930	>	int e, ac, tc, rc, i;
931	>	long c = ctl;
932	>	int u = (int)(c >>> 32);
933	>	if ((e = (int)c) < 0) {
934	>	// skip -- terminating
935	>	}
936	>	else if ((ac = (u >> UAC_SHIFT)) <= 0 && e != 0 &&
937	>	(ws = workers) != null &&
938	>	(i = ~e & SMASK) < ws.length &&
939	>	(w = ws[i]) != null) {
940	>	long nc = ((long)(w.nextWait & E_MASK) |
941	>	(c & (AC_MASK|TC_MASK)));
942	>	if (w.eventCount == e &&
943	>	UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc)) {
944	>	w.eventCount = (e + EC_UNIT) & E_MASK;
945	>	if (w.parked)
946	>	UNSAFE.unpark(w);
947	>	return true;             // release an idle worker
948	>	}
949	>	}
950	>	else if ((tc = (short)(u >>> UTC_SHIFT)) >= 0 && ac + pc > 1) {
951	>	long nc = ((c - AC_UNIT) & AC_MASK) | (c & ~AC_MASK);
952	>	if (UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc))
953	>	return true;             // no compensation needed
954	>	}
955	>	else if (tc + pc < MAX_ID) {
956	>	long nc = ((c + TC_UNIT) & TC_MASK) | (c & ~TC_MASK);
957	>	if (UNSAFE.compareAndSwapLong(this, ctlOffset, c, nc)) {
958	>	addWorker();
959	>	return true;            // create a replacement
960	>	}
961	>	}
962	>	// try to back out on any failure and let caller retry
963	>	} while (!UNSAFE.compareAndSwapInt(this, blockedCountOffset,
964	>	b = blockedCount, b - 1));
965		}
966	<	releaseEventWaiters();
966	>	return false;
967		}
968
875	–	// Maintaining spares
876	–
969		/**
970	<	* Pushes worker onto the spare stack
970	>	* Decrements blockedCount and increments active count
971		*/
972	<	final void pushSpare(ForkJoinWorkerThread w) {
973	<	int ns = (++w.spareCount << SPARE_COUNT_SHIFT) | (w.poolIndex + 1);
974	<	do {} while (!UNSAFE.compareAndSwapInt(this, spareWaitersOffset,
975	<	w.nextSpare = spareWaiters,ns));
972	>	private void postBlock() {
973	>	long c;
974	>	do {} while (!UNSAFE.compareAndSwapLong(this, ctlOffset,  // no mask
975	>	c = ctl, c + AC_UNIT));
976	>	int b;
977	>	do {} while(!UNSAFE.compareAndSwapInt(this, blockedCountOffset,
978	>	b = blockedCount, b - 1));
979		}
980
981		/**
982	<	* Callback from oldest spare occasionally waking up.  Tries
983	<	* (once) to shutdown a spare. Same idea as tryShutdownWaiter.
982	>	* Possibly blocks waiting for the given task to complete, or
983	>	* cancels the task if terminating.  Fails to wait if contended.
984	>	*
985	>	* @param joinMe the task
986		*/
987	<	final void tryShutdownSpare() {
988	<	int sw, id;
989	<	ForkJoinWorkerThread w;
990	<	ForkJoinWorkerThread[] ws;
991	<	if ((id = ((sw = spareWaiters) & SPARE_ID_MASK) - 1) >= 0 &&
992	<	id < (ws = workers).length && (w = ws[id]) != null &&
993	<	(workerCounts & RUNNING_COUNT_MASK) >= parallelism &&
994	<	UNSAFE.compareAndSwapInt(this, spareWaitersOffset,
995	<	sw, w.nextSpare)) {
996	<	w.shutdown();
900	<	LockSupport.unpark(w);
901	<	advanceEventCount();
987	>	final void tryAwaitJoin(ForkJoinTask<?> joinMe) {
988	>	int s;
989	>	Thread.interrupted(); // clear interrupts before checking termination
990	>	if (joinMe.status >= 0) {
991	>	if (tryPreBlock()) {
992	>	joinMe.tryAwaitDone(0L);
993	>	postBlock();
994	>	}
995	>	if ((ctl & STOP_BIT) != 0L)
996	>	joinMe.cancelIgnoringExceptions();
997		}
998		}
999
1000		/**
1001	<	* Tries (once) to resume a spare if worker counts match
1002	<	* the given count.
1001	>	* Possibly blocks the given worker waiting for joinMe to
1002	>	* complete or timeout
1003		*
1004	<	* @param wc workerCounts value on invocation of this method
1004	>	* @param joinMe the task
1005	>	* @param millis the wait time for underlying Object.wait
1006		*/
1007	<	private void tryResumeSpare(int wc) {
1008	<	ForkJoinWorkerThread[] ws = workers;
1009	<	int n = ws.length;
1010	<	int sw, id, rs;  ForkJoinWorkerThread w;
1011	<	if ((id = ((sw = spareWaiters) & SPARE_ID_MASK) - 1) >= 0 &&
1012	<	id < n && (w = ws[id]) != null &&
1013	<	(rs = runState) < TERMINATING &&
1014	<	eventWaiters == 0L && workerCounts == wc) {
1015	<	// In case all workers busy, heuristically back off to let settle
1016	<	Thread.yield();
1017	<	if (eventWaiters == 0L && runState == rs && // recheck
1018	<	workerCounts == wc && spareWaiters == sw &&
1019	<	UNSAFE.compareAndSwapInt(this, spareWaitersOffset,
1020	<	sw, w.nextSpare)) {
1021	<	int c;              // increment running count before resume
1022	<	do {} while(!UNSAFE.compareAndSwapInt
1023	<	(this, workerCountsOffset,
1024	<	c = workerCounts, c + ONE_RUNNING));
1025	<	if (w.tryUnsuspend())
1026	<	LockSupport.unpark(w);
1027	<	else               // back out if w was shutdown
1028	<	decrementWorkerCounts(ONE_RUNNING, 0);
1007	>	final void timedAwaitJoin(ForkJoinTask<?> joinMe, long nanos) {
1008	>	while (joinMe.status >= 0) {
1009	>	Thread.interrupted();
1010	>	if ((ctl & STOP_BIT) != 0L) {
1011	>	joinMe.cancelIgnoringExceptions();
1012	>	break;
1013	>	}
1014	>	if (tryPreBlock()) {
1015	>	long last = System.nanoTime();
1016	>	while (joinMe.status >= 0) {
1017	>	long millis = TimeUnit.NANOSECONDS.toMillis(nanos);
1018	>	if (millis <= 0)
1019	>	break;
1020	>	joinMe.tryAwaitDone(millis);
1021	>	if (joinMe.status < 0)
1022	>	break;
1023	>	if ((ctl & STOP_BIT) != 0L) {
1024	>	joinMe.cancelIgnoringExceptions();
1025	>	break;
1026	>	}
1027	>	long now = System.nanoTime();
1028	>	nanos -= now - last;
1029	>	last = now;
1030	>	}
1031	>	postBlock();
1032	>	break;
1033		}
1034		}
1035		}
1036
937	–	// adding workers on demand
938	–
1037		/**
1038	<	* Adds one or more workers if needed to establish target parallelism.
941	<	* Retries upon contention.
1038	>	* If necessary, compensates for blocker, and blocks
1039		*/
1040	<	private void addWorkerIfBelowTarget() {
1041	<	int pc = parallelism;
1042	<	int wc;
1043	<	while (((wc = workerCounts) >>> TOTAL_COUNT_SHIFT) < pc &&
1044	<	runState < TERMINATING) {
1045	<	if (UNSAFE.compareAndSwapInt(this, workerCountsOffset, wc,
1046	<	wc + (ONE_RUNNING|ONE_TOTAL))) {
1047	<	if (addWorker() == null)
1048	<	break;
1040	>	private void awaitBlocker(ManagedBlocker blocker)
1041	>	throws InterruptedException {
1042	>	while (!blocker.isReleasable()) {
1043	>	if (tryPreBlock()) {
1044	>	try {
1045	>	do {} while (!blocker.isReleasable() && !blocker.block());
1046	>	} finally {
1047	>	postBlock();
1048	>	}
1049	>	break;
1050		}
1051		}
1052		}
1053
1054	+	// Creating, registering and deregistring workers
1055	+
1056		/**
1057	<	* Tries (once) to add a new worker if all existing workers are
1058	<	* busy, and there are either no running workers or the deficit is
959	<	* at least twice the surplus.
960	<	*
961	<	* @param wc workerCounts value on invocation of this method
1057	>	* Tries to create and start a worker; minimally rolls back counts
1058	>	* on failure.
1059		*/
1060	<	private void tryAddWorkerIfBusy(int wc) {
1061	<	int tc, rc, rs;
1062	<	int pc = parallelism;
1063	<	if ((tc = wc >>> TOTAL_COUNT_SHIFT) < MAX_WORKERS &&
1064	<	((rc = wc & RUNNING_COUNT_MASK) == 0 ||
1065	<	rc < pc - ((tc - pc) << 1)) &&
1066	<	(rs = runState) < TERMINATING &&
1067	<	(rs & ACTIVE_COUNT_MASK) == tc) {
1068	<	// Since all workers busy, heuristically back off to let settle
1069	<	Thread.yield();
1070	<	if (eventWaiters == 0L && spareWaiters == 0 && // recheck
1071	<	runState == rs && workerCounts == wc &&
1072	<	UNSAFE.compareAndSwapInt(this, workerCountsOffset, wc,
1073	<	wc + (ONE_RUNNING|ONE_TOTAL)))
1074	<	addWorker();
1060	>	private void addWorker() {
1061	>	Throwable ex = null;
1062	>	ForkJoinWorkerThread t = null;
1063	>	try {
1064	>	t = factory.newThread(this);
1065	>	} catch (Throwable e) {
1066	>	ex = e;
1067	>	}
1068	>	if (t == null) {  // null or exceptional factory return
1069	>	long c;       // adjust counts
1070	>	do {} while (!UNSAFE.compareAndSwapLong
1071	>	(this, ctlOffset, c = ctl,
1072	>	(((c - AC_UNIT) & AC_MASK) |
1073	>	((c - TC_UNIT) & TC_MASK) |
1074	>	(c & ~(AC_MASK|TC_MASK)))));
1075	>	// Propagate exception if originating from an external caller
1076	>	if (!tryTerminate(false) && ex != null &&
1077	>	!(Thread.currentThread() instanceof ForkJoinWorkerThread))
1078	>	UNSAFE.throwException(ex);
1079		}
1080	+	else
1081	+	t.start();
1082		}
1083
1084		/**
1085	<	* Does at most one of:
1086	<	*
1087	<	* 1. Help wake up existing workers waiting for work via
1088	<	*    releaseEventWaiters. (If any exist, then it doesn't
1089	<	*    matter right now if under target parallelism level.)
1090	<	*
1091	<	* 2. If a spare exists, try (once) to resume it via tryResumeSpare.
1092	<	*
990	<	* 3. If there are not enough total workers, add some
991	<	*    via addWorkerIfBelowTarget;
992	<	*
993	<	* 4. Try (once) to add a new worker if all existing workers
994	<	*     are busy, via tryAddWorkerIfBusy
995	<	*/
996	<	private void helpMaintainParallelism() {
997	<	long h; int pc, wc;
998	<	if (((int)((h = eventWaiters) & WAITER_ID_MASK)) != 0) {
999	<	if ((int)(h >>> EVENT_COUNT_SHIFT) != eventCount)
1000	<	releaseEventWaiters(); // avoid useless call
1001	<	}
1002	<	else if ((pc = parallelism) >
1003	<	((wc = workerCounts) & RUNNING_COUNT_MASK)) {
1004	<	if (spareWaiters != 0)
1005	<	tryResumeSpare(wc);
1006	<	else if ((wc >>> TOTAL_COUNT_SHIFT) < pc)
1007	<	addWorkerIfBelowTarget();
1008	<	else
1009	<	tryAddWorkerIfBusy(wc);
1085	>	* Callback from ForkJoinWorkerThread constructor to assign a
1086	>	* public name
1087	>	*/
1088	>	final String nextWorkerName() {
1089	>	for (int n;;) {
1090	>	if (UNSAFE.compareAndSwapInt(this, nextWorkerNumberOffset,
1091	>	n = nextWorkerNumber, ++n))
1092	>	return workerNamePrefix + n;
1093		}
1094		}
1095
1096		/**
1097	<	* Callback from workers invoked upon each top-level action (i.e.,
1098	<	* stealing a task or taking a submission and running it).
1016	<	* Performs one or more of the following:
1017	<	*
1018	<	* 1. If the worker is active, try to set its active status to
1019	<	*    inactive and update activeCount. On contention, we may try
1020	<	*    again on this or subsequent call.
1021	<	*
1022	<	* 2. Release any existing event waiters that are now relesable
1023	<	*
1024	<	* 3. If there are too many running threads, suspend this worker
1025	<	*    (first forcing inactive if necessary).  If it is not
1026	<	*    needed, it may be killed while suspended via
1027	<	*    tryShutdownSpare. Otherwise, upon resume it rechecks to make
1028	<	*    sure that it is still needed.
1029	<	*
1030	<	* 4. If more than 1 miss, await the next task event via
1031	<	*    eventSync (first forcing inactivation if necessary), upon
1032	<	*    which worker may also be killed, via tryShutdownWaiter.
1033	<	*
1034	<	* 5. Help reactivate other workers via helpMaintainParallelism
1097	>	* Callback from ForkJoinWorkerThread constructor to
1098	>	* determine its poolIndex and record in workers array.
1099		*
1100		* @param w the worker
1101	<	* @param misses the number of scans by caller failing to find work
1038	<	* (saturating at 2 to avoid wraparound)
1101	>	* @return the worker's pool index
1102		*/
1103	<	final void preStep(ForkJoinWorkerThread w, int misses) {
1104	<	boolean active = w.active;
1105	<	int pc = parallelism;
1106	<	for (;;) {
1107	<	int rs, wc, rc, ec; long h;
1108	<	if (active && UNSAFE.compareAndSwapInt(this, runStateOffset,
1109	<	rs = runState, rs - 1))
1110	<	active = w.active = false;
1111	<	if (((int)((h = eventWaiters) & WAITER_ID_MASK)) != 0 &&
1112	<	(int)(h >>> EVENT_COUNT_SHIFT) != eventCount) {
1113	<	releaseEventWaiters();
1114	<	if (misses > 1)
1115	<	continue;                  // clear before sync below
1116	<	}
1117	<	if ((rc = ((wc = workerCounts) & RUNNING_COUNT_MASK)) > pc) {
1118	<	if (!active &&                 // must inactivate to suspend
1119	<	workerCounts == wc &&      // try to suspend as spare
1120	<	UNSAFE.compareAndSwapInt(this, workerCountsOffset,
1121	<	wc, wc - ONE_RUNNING)) {
1122	<	w.suspendAsSpare();
1123	<	if (!w.isRunning())
1124	<	break;                 // was killed while spare
1103	>	final int registerWorker(ForkJoinWorkerThread w) {
1104	>	/*
1105	>	* In the typical case, a new worker acquires the lock, uses
1106	>	* next available index and returns quickly.  Since we should
1107	>	* not block callers (ultimately from signalWork or
1108	>	* tryPreBlock) waiting for the lock needed to do this, we
1109	>	* instead help release other workers while waiting for the
1110	>	* lock.
1111	>	*/
1112	>	for (int g;;) {
1113	>	ForkJoinWorkerThread[] ws;
1114	>	if (((g = scanGuard) & SG_UNIT) == 0 &&
1115	>	UNSAFE.compareAndSwapInt(this, scanGuardOffset,
1116	>	g, g | SG_UNIT)) {
1117	>	int k = nextWorkerIndex;
1118	>	try {
1119	>	if ((ws = workers) != null) { // ignore on shutdown
1120	>	int n = ws.length;
1121	>	if (k < 0 || k >= n || ws[k] != null) {
1122	>	for (k = 0; k < n && ws[k] != null; ++k)
1123	>	;
1124	>	if (k == n)
1125	>	ws = workers = Arrays.copyOf(ws, n << 1);
1126	>	}
1127	>	ws[k] = w;
1128	>	nextWorkerIndex = k + 1;
1129	>	int m = g & SMASK;
1130	>	g = k >= m? ((m << 1) + 1) & SMASK : g + (SG_UNIT<<1);
1131	>	}
1132	>	} finally {
1133	>	scanGuard = g;
1134		}
1135	<	continue;
1135	>	return k;
1136		}
1137	<	if (misses > 0) {
1138	<	if ((ec = eventCount) == w.lastEventCount && misses > 1) {
1139	<	if (!active) {             // must inactivate to sync
1140	<	eventSync(w);
1141	<	if (w.isRunning())
1070	<	misses = 1;        // don't re-sync
1071	<	else
1072	<	break;             // was killed while waiting
1137	>	else if ((ws = workers) != null) { // help release others
1138	>	for (ForkJoinWorkerThread u : ws) {
1139	>	if (u != null && u.queueBase != u.queueTop) {
1140	>	if (tryReleaseWaiter())
1141	>	break;
1142		}
1074	–	continue;
1143		}
1076	–	w.lastEventCount = ec;
1144		}
1078	–	if (rc < pc)
1079	–	helpMaintainParallelism();
1080	–	break;
1145		}
1146		}
1147
1148		/**
1149	<	* Helps and/or blocks awaiting join of the given task.
1150	<	* Alternates between helpJoinTask() and helpMaintainParallelism()
1151	<	* as many times as there is a deficit in running count (or longer
1088	<	* if running count would become zero), then blocks if task still
1089	<	* not done.
1149	>	* Final callback from terminating worker.  Removes record of
1150	>	* worker from array, and adjusts counts. If pool is shutting
1151	>	* down, tries to complete termination.
1152		*
1153	<	* @param joinMe the task to join
1153	>	* @param w the worker
1154		*/
1155	<	final void awaitJoin(ForkJoinTask<?> joinMe, ForkJoinWorkerThread worker) {
1156	<	int threshold = parallelism;         // descend blocking thresholds
1157	<	while (joinMe.status >= 0) {
1158	<	boolean block; int wc;
1159	<	worker.helpJoinTask(joinMe);
1160	<	if (joinMe.status < 0)
1161	<	break;
1162	<	if (((wc = workerCounts) & RUNNING_COUNT_MASK) <= threshold) {
1163	<	if (threshold > 0)
1164	<	--threshold;
1165	<	else
1166	<	advanceEventCount(); // force release
1167	<	block = false;
1155	>	final void deregisterWorker(ForkJoinWorkerThread w, Throwable ex) {
1156	>	int idx = w.poolIndex;
1157	>	int sc = w.stealCount;
1158	>	int steps = 0;
1159	>	// Remove from array, adjust worker counts and collect steal count.
1160	>	// We can intermix failed removes or adjusts with steal updates
1161	>	do {
1162	>	long s, c;
1163	>	int g;
1164	>	if (steps == 0 && ((g = scanGuard) & SG_UNIT) == 0 &&
1165	>	UNSAFE.compareAndSwapInt(this, scanGuardOffset,
1166	>	g, g |= SG_UNIT)) {
1167	>	ForkJoinWorkerThread[] ws = workers;
1168	>	if (ws != null && idx >= 0 &&
1169	>	idx < ws.length && ws[idx] == w)
1170	>	ws[idx] = null;    // verify
1171	>	nextWorkerIndex = idx;
1172	>	scanGuard = g + SG_UNIT;
1173	>	steps = 1;
1174		}
1175	+	if (steps == 1 &&
1176	+	UNSAFE.compareAndSwapLong(this, ctlOffset, c = ctl,
1177	+	(((c - AC_UNIT) & AC_MASK) |
1178	+	((c - TC_UNIT) & TC_MASK) |
1179	+	(c & ~(AC_MASK|TC_MASK)))))
1180	+	steps = 2;
1181	+	if (sc != 0 &&
1182	+	UNSAFE.compareAndSwapLong(this, stealCountOffset,
1183	+	s = stealCount, s + sc))
1184	+	sc = 0;
1185	+	} while (steps != 2 || sc != 0);
1186	+	if (!tryTerminate(false)) {
1187	+	if (ex != null)   // possibly replace if died abnormally
1188	+	signalWork();
1189		else
1190	<	block = UNSAFE.compareAndSwapInt(this, workerCountsOffset,
1109	<	wc, wc - ONE_RUNNING);
1110	<	helpMaintainParallelism();
1111	<	if (block) {
1112	<	int c;
1113	<	joinMe.internalAwaitDone();
1114	<	do {} while (!UNSAFE.compareAndSwapInt
1115	<	(this, workerCountsOffset,
1116	<	c = workerCounts, c + ONE_RUNNING));
1117	<	break;
1118	<	}
1190	>	tryReleaseWaiter();
1191		}
1192		}
1193
1194	<	/**
1123	<	* Same idea as awaitJoin, but no helping
1124	<	*/
1125	<	final void awaitBlocker(ManagedBlocker blocker)
1126	<	throws InterruptedException {
1127	<	int threshold = parallelism;
1128	<	while (!blocker.isReleasable()) {
1129	<	boolean block; int wc;
1130	<	if (((wc = workerCounts) & RUNNING_COUNT_MASK) <= threshold) {
1131	<	if (threshold > 0)
1132	<	--threshold;
1133	<	else
1134	<	advanceEventCount();
1135	<	block = false;
1136	<	}
1137	<	else
1138	<	block = UNSAFE.compareAndSwapInt(this, workerCountsOffset,
1139	<	wc, wc - ONE_RUNNING);
1140	<	helpMaintainParallelism();
1141	<	if (block) {
1142	<	try {
1143	<	do {} while (!blocker.isReleasable() && !blocker.block());
1144	<	} finally {
1145	<	int c;
1146	<	do {} while (!UNSAFE.compareAndSwapInt
1147	<	(this, workerCountsOffset,
1148	<	c = workerCounts, c + ONE_RUNNING));
1149	<	}
1150	<	break;
1151	<	}
1152	<	}
1153	<	}
1194	>	// Shutdown and termination
1195
1196		/**
1197		* Possibly initiates and/or completes termination.
#	Line 1160 | Line 1201 | public class ForkJoinPool extends Abstra
1201		* @return true if now terminating or terminated
1202		*/
1203		private boolean tryTerminate(boolean now) {
1204	<	if (now)
1205	<	advanceRunLevel(SHUTDOWN); // ensure at least SHUTDOWN
1206	<	else if (runState < SHUTDOWN ||
1207	<	!submissionQueue.isEmpty() ||
1208	<	(runState & ACTIVE_COUNT_MASK) != 0)
1209	<	return false;
1210	<
1211	<	if (advanceRunLevel(TERMINATING))
1212	<	startTerminating();
1213	<
1214	<	// Finish now if all threads terminated; else in some subsequent call
1215	<	if ((workerCounts >>> TOTAL_COUNT_SHIFT) == 0) {
1216	<	advanceRunLevel(TERMINATED);
1217	<	termination.arrive();
1204	>	long c;
1205	>	while (((c = ctl) & STOP_BIT) == 0) {
1206	>	if (!now) {
1207	>	if ((int)(c >> AC_SHIFT) != -parallelism)
1208	>	return false;
1209	>	if (!shutdown || blockedCount != 0 || quiescerCount != 0 ||
1210	>	queueTop - queueBase > 0) {
1211	>	if (ctl == c) // staleness check
1212	>	return false;
1213	>	continue;
1214	>	}
1215	>	}
1216	>	if (UNSAFE.compareAndSwapLong(this, ctlOffset, c, c | STOP_BIT))
1217	>	startTerminating();
1218	>	}
1219	>	if ((short)(c >>> TC_SHIFT) == -parallelism) {
1220	>	submissionLock.lock();
1221	>	termination.signalAll();
1222	>	submissionLock.unlock();
1223		}
1224		return true;
1225		}
1226
1227		/**
1228	<	* Actions on transition to TERMINATING
1229	<	*
1230	<	* Runs up to four passes through workers: (0) shutting down each
1231	<	* (without waking up if parked) to quickly spread notifications
1232	<	* without unnecessary bouncing around event queues etc (1) wake
1233	<	* up and help cancel tasks (2) interrupt (3) mop up races with
1188	<	* interrupted workers
1228	>	* Runs up to three passes through workers: (0) Setting
1229	>	* termination status for each worker, followed by wakeups up
1230	>	* queued workers (1) helping cancel tasks (2) interrupting
1231	>	* lagging threads (likely in external tasks, but possibly also
1232	>	* blocked in joins).  Each pass repeats previous steps because of
1233	>	* potential lagging thread creation.
1234		*/
1235		private void startTerminating() {
1236		cancelSubmissions();
1237	<	for (int passes = 0; passes < 4 && workerCounts != 0; ++passes) {
1193	<	advanceEventCount();
1194	<	eventWaiters = 0L; // clobber lists
1195	<	spareWaiters = 0;
1237	>	for (int pass = 0; pass < 3; ++pass) {
1238		ForkJoinWorkerThread[] ws = workers;
1239	<	int n = ws.length;
1240	<	for (int i = 0; i < n; ++i) {
1241	<	ForkJoinWorkerThread w = ws[i];
1242	<	if (w != null) {
1243	<	w.shutdown();
1244	<	if (passes > 0 && !w.isTerminated()) {
1245	<	w.cancelTasks();
1246	<	LockSupport.unpark(w);
1247	<	if (passes > 1) {
1248	<	try {
1249	<	w.interrupt();
1208	<	} catch (SecurityException ignore) {
1239	>	if (ws != null) {
1240	>	for (ForkJoinWorkerThread w : ws) {
1241	>	if (w != null) {
1242	>	w.terminate = true;
1243	>	if (pass > 0) {
1244	>	w.cancelTasks();
1245	>	if (pass > 1 && !w.isInterrupted()) {
1246	>	try {
1247	>	w.interrupt();
1248	>	} catch (SecurityException ignore) {
1249	>	}
1250		}
1251		}
1252		}
1253		}
1254	+	terminateWaiters();
1255		}
1256		}
1257		}
1258
1259		/**
1260	<	* Clear out and cancel submissions, ignoring exceptions
1260	>	* Polls and cancels all submissions. Called only during termination.
1261		*/
1262		private void cancelSubmissions() {
1263	<	ForkJoinTask<?> task;
1264	<	while ((task = submissionQueue.poll()) != null) {
1265	<	try {
1266	<	task.cancel(false);
1267	<	} catch (Throwable ignore) {
1263	>	while (queueBase != queueTop) {
1264	>	ForkJoinTask<?> task = pollSubmission();
1265	>	if (task != null) {
1266	>	try {
1267	>	task.cancel(false);
1268	>	} catch (Throwable ignore) {
1269	>	}
1270		}
1271		}
1272		}
1273
1274	<	// misc support for ForkJoinWorkerThread
1274	>	/**
1275	>	* Tries to set the termination status of waiting workers, and
1276	>	* then wake them up (after which they will terminate).
1277	>	*/
1278	>	private void terminateWaiters() {
1279	>	ForkJoinWorkerThread[] ws = workers;
1280	>	if (ws != null) {
1281	>	ForkJoinWorkerThread w; long c; int i, e;
1282	>	int n = ws.length;
1283	>	while ((i = ~(e = (int)(c = ctl)) & SMASK) < n &&
1284	>	(w = ws[i]) != null && w.eventCount == (e & E_MASK)) {
1285	>	if (UNSAFE.compareAndSwapLong(this, ctlOffset, c,
1286	>	(long)(w.nextWait & E_MASK) |
1287	>	((c + AC_UNIT) & AC_MASK) |
1288	>	(c & (TC_MASK|STOP_BIT)))) {
1289	>	w.terminate = true;
1290	>	w.eventCount = e + EC_UNIT;
1291	>	if (w.parked)
1292	>	UNSAFE.unpark(w);
1293	>	}
1294	>	}
1295	>	}
1296	>	}
1297	>
1298	>	// misc ForkJoinWorkerThread support
1299
1300		/**
1301	<	* Returns pool number
1301	>	* Increment or decrement quiescerCount. Needed only to prevent
1302	>	* triggering shutdown if a worker is transiently inactive while
1303	>	* checking quiescence.
1304	>	*
1305	>	* @param delta 1 for increment, -1 for decrement
1306		*/
1307	<	final int getPoolNumber() {
1308	<	return poolNumber;
1307	>	final void addQuiescerCount(int delta) {
1308	>	int c;
1309	>	do {} while(!UNSAFE.compareAndSwapInt(this, quiescerCountOffset,
1310	>	c = quiescerCount, c + delta));
1311		}
1312
1313		/**
1314	<	* Tries to accumulates steal count from a worker, clearing
1315	<	* the worker's value.
1314	>	* Directly increment or decrement active count without
1315	>	* queuing. This method is used to transiently assert inactivation
1316	>	* while checking quiescence.
1317		*
1318	<	* @return true if worker steal count now zero
1318	>	* @param delta 1 for increment, -1 for decrement
1319		*/
1320	<	final boolean tryAccumulateStealCount(ForkJoinWorkerThread w) {
1321	<	int sc = w.stealCount;
1322	<	long c = stealCount;
1323	<	// CAS even if zero, for fence effects
1324	<	if (UNSAFE.compareAndSwapLong(this, stealCountOffset, c, c + sc)) {
1325	<	if (sc != 0)
1251	<	w.stealCount = 0;
1252	<	return true;
1253	<	}
1254	<	return sc == 0;
1320	>	final void addActiveCount(int delta) {
1321	>	long d = delta < 0 ? -AC_UNIT : AC_UNIT;
1322	>	long c;
1323	>	do {} while (!UNSAFE.compareAndSwapLong(this, ctlOffset, c = ctl,
1324	>	((c + d) & AC_MASK) |
1325	>	(c & ~AC_MASK)));
1326		}
1327
1328		/**
#	Line 1259 | Line 1330 | public class ForkJoinPool extends Abstra
1330		* active thread.
1331		*/
1332		final int idlePerActive() {
1333	<	int pc = parallelism; // use parallelism, not rc
1334	<	int ac = runState;    // no mask -- artifically boosts during shutdown
1335	<	// Use exact results for small values, saturate past 4
1336	<	return pc <= ac? 0 : pc >>> 1 <= ac? 1 : pc >>> 2 <= ac? 3 : pc >>> 3;
1333	>	// Approximate at powers of two for small values, saturate past 4
1334	>	int p = parallelism;
1335	>	int a = p + (int)(ctl >> AC_SHIFT);
1336	>	return (a > (p >>>= 1) ? 0 :
1337	>	a > (p >>>= 1) ? 1 :
1338	>	a > (p >>>= 1) ? 2 :
1339	>	a > (p >>>= 1) ? 4 :
1340	>	8);
1341		}
1342
1343	<	// Public and protected methods
1343	>	// Exported methods
1344
1345		// Constructors
1346
#	Line 1312 | Line 1387 | public class ForkJoinPool extends Abstra
1387		* use {@link #defaultForkJoinWorkerThreadFactory}.
1388		* @param handler the handler for internal worker threads that
1389		* terminate due to unrecoverable errors encountered while executing
1390	<	* tasks. For default value, use <code>null</code>.
1390	>	* tasks. For default value, use {@code null}.
1391		* @param asyncMode if true,
1392		* establishes local first-in-first-out scheduling mode for forked
1393		* tasks that are never joined. This mode may be more appropriate
1394		* than default locally stack-based mode in applications in which
1395		* worker threads only process event-style asynchronous tasks.
1396	<	* For default value, use <code>false</code>.
1396	>	* For default value, use {@code false}.
1397		* @throws IllegalArgumentException if parallelism less than or
1398		*         equal to zero, or greater than implementation limit
1399		* @throws NullPointerException if the factory is null
#	Line 1334 | Line 1409 | public class ForkJoinPool extends Abstra
1409		checkPermission();
1410		if (factory == null)
1411		throw new NullPointerException();
1412	<	if (parallelism <= 0 || parallelism > MAX_WORKERS)
1412	>	if (parallelism <= 0 || parallelism > MAX_ID)
1413		throw new IllegalArgumentException();
1414		this.parallelism = parallelism;
1415		this.factory = factory;
1416		this.ueh = handler;
1417		this.locallyFifo = asyncMode;
1418	<	int arraySize = initialArraySizeFor(parallelism);
1419	<	this.workers = new ForkJoinWorkerThread[arraySize];
1420	<	this.submissionQueue = new LinkedTransferQueue<ForkJoinTask<?>>();
1421	<	this.workerLock = new ReentrantLock();
1422	<	this.termination = new Phaser(1);
1423	<	this.poolNumber = poolNumberGenerator.incrementAndGet();
1424	<	}
1425	<
1426	<	/**
1427	<	* Returns initial power of two size for workers array.
1428	<	* @param pc the initial parallelism level
1429	<	*/
1430	<	private static int initialArraySizeFor(int pc) {
1431	<	// See Hackers Delight, sec 3.2. We know MAX_WORKERS < (1 >>> 16)
1432	<	int size = pc < MAX_WORKERS ? pc + 1 : MAX_WORKERS;
1433	<	size |= size >>> 1;
1434	<	size |= size >>> 2;
1360	<	size |= size >>> 4;
1361	<	size |= size >>> 8;
1362	<	return size + 1;
1418	>	long np = (long)(-parallelism); // offset ctl counts
1419	>	this.ctl = ((np << AC_SHIFT) & AC_MASK) | ((np << TC_SHIFT) & TC_MASK);
1420	>	this.submissionQueue = new ForkJoinTask<?>[INITIAL_QUEUE_CAPACITY];
1421	>	// initialize workers array with room for 2*parallelism if possible
1422	>	int n = parallelism << 1;
1423	>	if (n >= MAX_ID)
1424	>	n = MAX_ID;
1425	>	else { // See Hackers Delight, sec 3.2, where n < (1 << 16)
1426	>	n |= n >>> 1; n |= n >>> 2; n |= n >>> 4; n |= n >>> 8;
1427	>	}
1428	>	workers = new ForkJoinWorkerThread[n + 1];
1429	>	this.submissionLock = new ReentrantLock();
1430	>	this.termination = submissionLock.newCondition();
1431	>	StringBuilder sb = new StringBuilder("ForkJoinPool-");
1432	>	sb.append(poolNumberGenerator.incrementAndGet());
1433	>	sb.append("-worker-");
1434	>	this.workerNamePrefix = sb.toString();
1435		}
1436
1437		// Execution methods
1438
1439		/**
1368	–	* Common code for execute, invoke and submit
1369	–	*/
1370	–	private <T> void doSubmit(ForkJoinTask<T> task) {
1371	–	if (task == null)
1372	–	throw new NullPointerException();
1373	–	if (runState >= SHUTDOWN)
1374	–	throw new RejectedExecutionException();
1375	–	submissionQueue.offer(task);
1376	–	advanceEventCount();
1377	–	if (eventWaiters != 0L)
1378	–	releaseEventWaiters();
1379	–	if ((workerCounts >>> TOTAL_COUNT_SHIFT) < parallelism)
1380	–	addWorkerIfBelowTarget();
1381	–	}
1382	–
1383	–	/**
1440		* Performs the given task, returning its result upon completion.
1441	+	* If the computation encounters an unchecked Exception or Error,
1442	+	* it is rethrown as the outcome of this invocation.  Rethrown
1443	+	* exceptions behave in the same way as regular exceptions, but,
1444	+	* when possible, contain stack traces (as displayed for example
1445	+	* using {@code ex.printStackTrace()}) of both the current thread
1446	+	* as well as the thread actually encountering the exception;
1447	+	* minimally only the latter.
1448		*
1449		* @param task the task
1450		* @return the task's result
#	Line 1390 | Line 1453 | public class ForkJoinPool extends Abstra
1453		*         scheduled for execution
1454		*/
1455		public <T> T invoke(ForkJoinTask<T> task) {
1456	<	doSubmit(task);
1457	<	return task.join();
1456	>	Thread t = Thread.currentThread();
1457	>	if (task == null)
1458	>	throw new NullPointerException();
1459	>	if (shutdown)
1460	>	throw new RejectedExecutionException();
1461	>	if ((t instanceof ForkJoinWorkerThread) &&
1462	>	((ForkJoinWorkerThread)t).pool == this)
1463	>	return task.invoke();  // bypass submit if in same pool
1464	>	else {
1465	>	addSubmission(task);
1466	>	return task.join();
1467	>	}
1468	>	}
1469	>
1470	>	/**
1471	>	* Unless terminating, forks task if within an ongoing FJ
1472	>	* computation in the current pool, else submits as external task.
1473	>	*/
1474	>	private <T> void forkOrSubmit(ForkJoinTask<T> task) {
1475	>	ForkJoinWorkerThread w;
1476	>	Thread t = Thread.currentThread();
1477	>	if (shutdown)
1478	>	throw new RejectedExecutionException();
1479	>	if ((t instanceof ForkJoinWorkerThread) &&
1480	>	(w = (ForkJoinWorkerThread)t).pool == this)
1481	>	w.pushTask(task);
1482	>	else
1483	>	addSubmission(task);
1484		}
1485
1486		/**
#	Line 1403 | Line 1492 | public class ForkJoinPool extends Abstra
1492		*         scheduled for execution
1493		*/
1494		public void execute(ForkJoinTask<?> task) {
1495	<	doSubmit(task);
1495	>	if (task == null)
1496	>	throw new NullPointerException();
1497	>	forkOrSubmit(task);
1498		}
1499
1500		// AbstractExecutorService methods
#	Line 1414 | Line 1505 | public class ForkJoinPool extends Abstra
1505		*         scheduled for execution
1506		*/
1507		public void execute(Runnable task) {
1508	+	if (task == null)
1509	+	throw new NullPointerException();
1510		ForkJoinTask<?> job;
1511		if (task instanceof ForkJoinTask<?>) // avoid re-wrap
1512		job = (ForkJoinTask<?>) task;
1513		else
1514		job = ForkJoinTask.adapt(task, null);
1515	<	doSubmit(job);
1515	>	forkOrSubmit(job);
1516		}
1517
1518		/**
#	Line 1432 | Line 1525 | public class ForkJoinPool extends Abstra
1525		*         scheduled for execution
1526		*/
1527		public <T> ForkJoinTask<T> submit(ForkJoinTask<T> task) {
1528	<	doSubmit(task);
1528	>	if (task == null)
1529	>	throw new NullPointerException();
1530	>	forkOrSubmit(task);
1531		return task;
1532		}
1533
#	Line 1442 | Line 1537 | public class ForkJoinPool extends Abstra
1537		*         scheduled for execution
1538		*/
1539		public <T> ForkJoinTask<T> submit(Callable<T> task) {
1540	+	if (task == null)
1541	+	throw new NullPointerException();
1542		ForkJoinTask<T> job = ForkJoinTask.adapt(task);
1543	<	doSubmit(job);
1543	>	forkOrSubmit(job);
1544		return job;
1545		}
1546
#	Line 1453 | Line 1550 | public class ForkJoinPool extends Abstra
1550		*         scheduled for execution
1551		*/
1552		public <T> ForkJoinTask<T> submit(Runnable task, T result) {
1553	+	if (task == null)
1554	+	throw new NullPointerException();
1555		ForkJoinTask<T> job = ForkJoinTask.adapt(task, result);
1556	<	doSubmit(job);
1556	>	forkOrSubmit(job);
1557		return job;
1558		}
1559
#	Line 1464 | Line 1563 | public class ForkJoinPool extends Abstra
1563		*         scheduled for execution
1564		*/
1565		public ForkJoinTask<?> submit(Runnable task) {
1566	+	if (task == null)
1567	+	throw new NullPointerException();
1568		ForkJoinTask<?> job;
1569		if (task instanceof ForkJoinTask<?>) // avoid re-wrap
1570		job = (ForkJoinTask<?>) task;
1571		else
1572		job = ForkJoinTask.adapt(task, null);
1573	<	doSubmit(job);
1573	>	forkOrSubmit(job);
1574		return job;
1575		}
1576
#	Line 1529 | Line 1630 | public class ForkJoinPool extends Abstra
1630
1631		/**
1632		* Returns the number of worker threads that have started but not
1633	<	* yet terminated.  This result returned by this method may differ
1633	>	* yet terminated.  The result returned by this method may differ
1634		* from {@link #getParallelism} when threads are created to
1635		* maintain parallelism when others are cooperatively blocked.
1636		*
1637		* @return the number of worker threads
1638		*/
1639		public int getPoolSize() {
1640	<	return workerCounts >>> TOTAL_COUNT_SHIFT;
1640	>	return parallelism + (short)(ctl >>> TC_SHIFT);
1641		}
1642
1643		/**
#	Line 1558 | Line 1659 | public class ForkJoinPool extends Abstra
1659		* @return the number of worker threads
1660		*/
1661		public int getRunningThreadCount() {
1662	<	return workerCounts & RUNNING_COUNT_MASK;
1662	>	int r = parallelism + (int)(ctl >> AC_SHIFT);
1663	>	return r <= 0? 0 : r; // suppress momentarily negative values
1664		}
1665
1666		/**
#	Line 1569 | Line 1671 | public class ForkJoinPool extends Abstra
1671		* @return the number of active threads
1672		*/
1673		public int getActiveThreadCount() {
1674	<	return runState & ACTIVE_COUNT_MASK;
1674	>	int r = parallelism + (int)(ctl >> AC_SHIFT) + blockedCount;
1675	>	return r <= 0? 0 : r; // suppress momentarily negative values
1676		}
1677
1678		/**
#	Line 1584 | Line 1687 | public class ForkJoinPool extends Abstra
1687		* @return {@code true} if all threads are currently idle
1688		*/
1689		public boolean isQuiescent() {
1690	<	return (runState & ACTIVE_COUNT_MASK) == 0;
1690	>	return parallelism + (int)(ctl >> AC_SHIFT) + blockedCount == 0;
1691		}
1692
1693		/**
#	Line 1614 | Line 1717 | public class ForkJoinPool extends Abstra
1717		*/
1718		public long getQueuedTaskCount() {
1719		long count = 0;
1720	<	ForkJoinWorkerThread[] ws = workers;
1721	<	int n = ws.length;
1722	<	for (int i = 0; i < n; ++i) {
1723	<	ForkJoinWorkerThread w = ws[i];
1724	<	if (w != null)
1725	<	count += w.getQueueSize();
1720	>	ForkJoinWorkerThread[] ws;
1721	>	if ((short)(ctl >>> TC_SHIFT) > -parallelism &&
1722	>	(ws = workers) != null) {
1723	>	for (ForkJoinWorkerThread w : ws)
1724	>	if (w != null)
1725	>	count -= w.queueBase - w.queueTop; // must read base first
1726		}
1727		return count;
1728		}
1729
1730		/**
1731		* Returns an estimate of the number of tasks submitted to this
1732	<	* pool that have not yet begun executing.  This method takes time
1733	<	* proportional to the number of submissions.
1732	>	* pool that have not yet begun executing.  This meThod may take
1733	>	* time proportional to the number of submissions.
1734		*
1735		* @return the number of queued submissions
1736		*/
1737		public int getQueuedSubmissionCount() {
1738	<	return submissionQueue.size();
1738	>	return -queueBase + queueTop;
1739		}
1740
1741		/**
#	Line 1642 | Line 1745 | public class ForkJoinPool extends Abstra
1745		* @return {@code true} if there are any queued submissions
1746		*/
1747		public boolean hasQueuedSubmissions() {
1748	<	return !submissionQueue.isEmpty();
1748	>	return queueBase != queueTop;
1749		}
1750
1751		/**
#	Line 1653 | Line 1756 | public class ForkJoinPool extends Abstra
1756		* @return the next submission, or {@code null} if none
1757		*/
1758		protected ForkJoinTask<?> pollSubmission() {
1759	<	return submissionQueue.poll();
1759	>	ForkJoinTask<?> t; ForkJoinTask<?>[] q; int b, i;
1760	>	while ((b = queueBase) != queueTop &&
1761	>	(q = submissionQueue) != null &&
1762	>	(i = (q.length - 1) & b) >= 0) {
1763	>	long u = (i << ASHIFT) + ABASE;
1764	>	if ((t = q[i]) != null &&
1765	>	queueBase == b &&
1766	>	UNSAFE.compareAndSwapObject(q, u, t, null)) {
1767	>	queueBase = b + 1;
1768	>	return t;
1769	>	}
1770	>	}
1771	>	return null;
1772		}
1773
1774		/**
#	Line 1674 | Line 1789 | public class ForkJoinPool extends Abstra
1789		* @return the number of elements transferred
1790		*/
1791		protected int drainTasksTo(Collection<? super ForkJoinTask<?>> c) {
1792	<	int count = submissionQueue.drainTo(c);
1793	<	ForkJoinWorkerThread[] ws = workers;
1794	<	int n = ws.length;
1795	<	for (int i = 0; i < n; ++i) {
1796	<	ForkJoinWorkerThread w = ws[i];
1797	<	if (w != null)
1798	<	count += w.drainTasksTo(c);
1792	>	int count = 0;
1793	>	while (queueBase != queueTop) {
1794	>	ForkJoinTask<?> t = pollSubmission();
1795	>	if (t != null) {
1796	>	c.add(t);
1797	>	++count;
1798	>	}
1799	>	}
1800	>	ForkJoinWorkerThread[] ws;
1801	>	if ((short)(ctl >>> TC_SHIFT) > -parallelism &&
1802	>	(ws = workers) != null) {
1803	>	for (ForkJoinWorkerThread w : ws)
1804	>	if (w != null)
1805	>	count += w.drainTasksTo(c);
1806		}
1807		return count;
1808		}
#	Line 1696 | Line 1818 | public class ForkJoinPool extends Abstra
1818		long st = getStealCount();
1819		long qt = getQueuedTaskCount();
1820		long qs = getQueuedSubmissionCount();
1699	–	int wc = workerCounts;
1700	–	int tc = wc >>> TOTAL_COUNT_SHIFT;
1701	–	int rc = wc & RUNNING_COUNT_MASK;
1821		int pc = parallelism;
1822	<	int rs = runState;
1823	<	int ac = rs & ACTIVE_COUNT_MASK;
1822	>	long c = ctl;
1823	>	int tc = pc + (short)(c >>> TC_SHIFT);
1824	>	int rc = pc + (int)(c >> AC_SHIFT);
1825	>	if (rc < 0) // ignore transient negative
1826	>	rc = 0;
1827	>	int ac = rc + blockedCount;
1828	>	String level;
1829	>	if ((c & STOP_BIT) != 0)
1830	>	level = (tc == 0)? "Terminated" : "Terminating";
1831	>	else
1832	>	level = shutdown? "Shutting down" : "Running";
1833		return super.toString() +
1834	<	"[" + runLevelToString(rs) +
1834	>	"[" + level +
1835		", parallelism = " + pc +
1836		", size = " + tc +
1837		", active = " + ac +
#	Line 1714 | Line 1842 | public class ForkJoinPool extends Abstra
1842		"]";
1843		}
1844
1717	–	private static String runLevelToString(int s) {
1718	–	return ((s & TERMINATED) != 0 ? "Terminated" :
1719	–	((s & TERMINATING) != 0 ? "Terminating" :
1720	–	((s & SHUTDOWN) != 0 ? "Shutting down" :
1721	–	"Running")));
1722	–	}
1723	–
1845		/**
1846		* Initiates an orderly shutdown in which previously submitted
1847		* tasks are executed, but no new tasks will be accepted.
#	Line 1735 | Line 1856 | public class ForkJoinPool extends Abstra
1856		*/
1857		public void shutdown() {
1858		checkPermission();
1859	<	advanceRunLevel(SHUTDOWN);
1859	>	shutdown = true;
1860		tryTerminate(false);
1861		}
1862
#	Line 1757 | Line 1878 | public class ForkJoinPool extends Abstra
1878		*/
1879		public List<Runnable> shutdownNow() {
1880		checkPermission();
1881	+	shutdown = true;
1882		tryTerminate(true);
1883		return Collections.emptyList();
1884		}
#	Line 1767 | Line 1889 | public class ForkJoinPool extends Abstra
1889		* @return {@code true} if all tasks have completed following shut down
1890		*/
1891		public boolean isTerminated() {
1892	<	return runState >= TERMINATED;
1892	>	long c = ctl;
1893	>	return ((c & STOP_BIT) != 0L &&
1894	>	(short)(c >>> TC_SHIFT) == -parallelism);
1895		}
1896
1897		/**
#	Line 1775 | Line 1899 | public class ForkJoinPool extends Abstra
1899		* commenced but not yet completed.  This method may be useful for
1900		* debugging. A return of {@code true} reported a sufficient
1901		* period after shutdown may indicate that submitted tasks have
1902	<	* ignored or suppressed interruption, causing this executor not
1903	<	* to properly terminate.
1902	>	* ignored or suppressed interruption, or are waiting for IO,
1903	>	* causing this executor not to properly terminate. (See the
1904	>	* advisory notes for class {@link ForkJoinTask} stating that
1905	>	* tasks should not normally entail blocking operations.  But if
1906	>	* they do, they must abort them on interrupt.)
1907		*
1908		* @return {@code true} if terminating but not yet terminated
1909		*/
1910		public boolean isTerminating() {
1911	<	return (runState & (TERMINATING|TERMINATED)) == TERMINATING;
1911	>	long c = ctl;
1912	>	return ((c & STOP_BIT) != 0L &&
1913	>	(short)(c >>> TC_SHIFT) != -parallelism);
1914	>	}
1915	>
1916	>	/**
1917	>	* Returns true if terminating or terminated. Used by ForkJoinWorkerThread.
1918	>	*/
1919	>	final boolean isAtLeastTerminating() {
1920	>	return (ctl & STOP_BIT) != 0L;
1921		}
1922
1923		/**
#	Line 1790 | Line 1926 | public class ForkJoinPool extends Abstra
1926		* @return {@code true} if this pool has been shut down
1927		*/
1928		public boolean isShutdown() {
1929	<	return runState >= SHUTDOWN;
1929	>	return shutdown;
1930		}
1931
1932		/**
#	Line 1806 | Line 1942 | public class ForkJoinPool extends Abstra
1942		*/
1943		public boolean awaitTermination(long timeout, TimeUnit unit)
1944		throws InterruptedException {
1945	+	long nanos = unit.toNanos(timeout);
1946	+	final ReentrantLock lock = this.submissionLock;
1947	+	lock.lock();
1948		try {
1949	<	return termination.awaitAdvanceInterruptibly(0, timeout, unit) > 0;
1950	<	} catch(TimeoutException ex) {
1951	<	return false;
1949	>	for (;;) {
1950	>	if (isTerminated())
1951	>	return true;
1952	>	if (nanos <= 0)
1953	>	return false;
1954	>	nanos = termination.awaitNanos(nanos);
1955	>	}
1956	>	} finally {
1957	>	lock.unlock();
1958		}
1959		}
1960
#	Line 1855 | Line 2000 | public class ForkJoinPool extends Abstra
2000		*   QueueTaker(BlockingQueue<E> q) { this.queue = q; }
2001		*   public boolean block() throws InterruptedException {
2002		*     if (item == null)
2003	<	*       item = queue.take
2003	>	*       item = queue.take();
2004		*     return true;
2005		*   }
2006		*   public boolean isReleasable() {
2007	<	*     return item != null || (item = queue.poll) != null;
2007	>	*     return item != null || (item = queue.poll()) != null;
2008		*   }
2009		*   public E getItem() { // call after pool.managedBlock completes
2010		*     return item;
#	Line 1929 | Line 2074 | public class ForkJoinPool extends Abstra
2074		}
2075
2076		// Unsafe mechanics
2077	<
2078	<	private static final sun.misc.Unsafe UNSAFE = getUnsafe();
2079	<	private static final long workerCountsOffset =
2080	<	objectFieldOffset("workerCounts", ForkJoinPool.class);
2081	<	private static final long runStateOffset =
2082	<	objectFieldOffset("runState", ForkJoinPool.class);
2083	<	private static final long eventCountOffset =
2084	<	objectFieldOffset("eventCount", ForkJoinPool.class);
2085	<	private static final long eventWaitersOffset =
2086	<	objectFieldOffset("eventWaiters",ForkJoinPool.class);
2087	<	private static final long stealCountOffset =
2088	<	objectFieldOffset("stealCount",ForkJoinPool.class);
2089	<	private static final long spareWaitersOffset =
2090	<	objectFieldOffset("spareWaiters",ForkJoinPool.class);
2091	<
2092	<	private static long objectFieldOffset(String field, Class<?> klazz) {
2077	>	private static final sun.misc.Unsafe UNSAFE;
2078	>	private static final long ctlOffset;
2079	>	private static final long stealCountOffset;
2080	>	private static final long blockedCountOffset;
2081	>	private static final long quiescerCountOffset;
2082	>	private static final long scanGuardOffset;
2083	>	private static final long nextWorkerNumberOffset;
2084	>	private static final long ABASE;
2085	>	private static final int ASHIFT;
2086	>
2087	>	static {
2088	>	poolNumberGenerator = new AtomicInteger();
2089	>	workerSeedGenerator = new Random();
2090	>	modifyThreadPermission = new RuntimePermission("modifyThread");
2091	>	defaultForkJoinWorkerThreadFactory =
2092	>	new DefaultForkJoinWorkerThreadFactory();
2093	>	int s;
2094		try {
2095	<	return UNSAFE.objectFieldOffset(klazz.getDeclaredField(field));
2096	<	} catch (NoSuchFieldException e) {
2097	<	// Convert Exception to corresponding Error
2098	<	NoSuchFieldError error = new NoSuchFieldError(field);
2099	<	error.initCause(e);
2100	<	throw error;
2101	<	}
2095	>	UNSAFE = getUnsafe();
2096	>	Class k = ForkJoinPool.class;
2097	>	ctlOffset = UNSAFE.objectFieldOffset
2098	>	(k.getDeclaredField("ctl"));
2099	>	stealCountOffset = UNSAFE.objectFieldOffset
2100	>	(k.getDeclaredField("stealCount"));
2101	>	blockedCountOffset = UNSAFE.objectFieldOffset
2102	>	(k.getDeclaredField("blockedCount"));
2103	>	quiescerCountOffset = UNSAFE.objectFieldOffset
2104	>	(k.getDeclaredField("quiescerCount"));
2105	>	scanGuardOffset = UNSAFE.objectFieldOffset
2106	>	(k.getDeclaredField("scanGuard"));
2107	>	nextWorkerNumberOffset = UNSAFE.objectFieldOffset
2108	>	(k.getDeclaredField("nextWorkerNumber"));
2109	>	Class a = ForkJoinTask[].class;
2110	>	ABASE = UNSAFE.arrayBaseOffset(a);
2111	>	s = UNSAFE.arrayIndexScale(a);
2112	>	} catch (Exception e) {
2113	>	throw new Error(e);
2114	>	}
2115	>	if ((s & (s-1)) != 0)
2116	>	throw new Error("data type scale not a power of two");
2117	>	ASHIFT = 31 - Integer.numberOfLeadingZeros(s);
2118		}
2119
2120		/**

Diff Legend

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