21 |
|
/** |
22 |
|
* An {@link ExecutorService} for running {@link ForkJoinTask}s. |
23 |
|
* A {@code ForkJoinPool} provides the entry point for submissions |
24 |
< |
* from non-{@code ForkJoinTask}s, as well as management and |
24 |
> |
* from non-{@code ForkJoinTask} clients, as well as management and |
25 |
|
* monitoring operations. |
26 |
|
* |
27 |
|
* <p>A {@code ForkJoinPool} differs from other kinds of {@link |
30 |
|
* execute subtasks created by other active tasks (eventually blocking |
31 |
|
* waiting for work if none exist). This enables efficient processing |
32 |
|
* when most tasks spawn other subtasks (as do most {@code |
33 |
< |
* ForkJoinTask}s). A {@code ForkJoinPool} may also be used for mixed |
34 |
< |
* execution of some plain {@code Runnable}- or {@code Callable}- |
35 |
< |
* based activities along with {@code ForkJoinTask}s. When setting |
36 |
< |
* {@linkplain #setAsyncMode async mode}, a {@code ForkJoinPool} may |
37 |
< |
* also be appropriate for use with fine-grained tasks of any form |
38 |
< |
* that are never joined. Otherwise, other {@code ExecutorService} |
39 |
< |
* implementations are typically more appropriate choices. |
33 |
> |
* ForkJoinTask}s). When setting <em>asyncMode</em> to true in |
34 |
> |
* constructors, {@code ForkJoinPool}s may also be appropriate for use |
35 |
> |
* with event-style tasks that are never joined. |
36 |
|
* |
37 |
|
* <p>A {@code ForkJoinPool} is constructed with a given target |
38 |
|
* parallelism level; by default, equal to the number of available |
39 |
< |
* processors. Unless configured otherwise via {@link |
40 |
< |
* #setMaintainsParallelism}, the pool attempts to maintain this |
41 |
< |
* number of active (or available) threads by dynamically adding, |
42 |
< |
* suspending, or resuming internal worker threads, even if some tasks |
43 |
< |
* are stalled waiting to join others. However, no such adjustments |
44 |
< |
* are performed in the face of blocked IO or other unmanaged |
45 |
< |
* synchronization. The nested {@link ManagedBlocker} interface |
50 |
< |
* enables extension of the kinds of synchronization accommodated. |
51 |
< |
* The target parallelism level may also be changed dynamically |
52 |
< |
* ({@link #setParallelism}). The total number of threads may be |
53 |
< |
* limited using method {@link #setMaximumPoolSize}, in which case it |
54 |
< |
* may become possible for the activities of a pool to stall due to |
55 |
< |
* the lack of available threads to process new tasks. When the pool |
56 |
< |
* is executing tasks, these and other configuration setting methods |
57 |
< |
* may only gradually affect actual pool sizes. It is normally best |
58 |
< |
* practice to invoke these methods only when the pool is known to be |
59 |
< |
* quiescent. |
39 |
> |
* processors. The pool attempts to maintain enough active (or |
40 |
> |
* available) threads by dynamically adding, suspending, or resuming |
41 |
> |
* internal worker threads, even if some tasks are stalled waiting to |
42 |
> |
* join others. However, no such adjustments are guaranteed in the |
43 |
> |
* face of blocked IO or other unmanaged synchronization. The nested |
44 |
> |
* {@link ManagedBlocker} interface enables extension of the kinds of |
45 |
> |
* synchronization accommodated. |
46 |
|
* |
47 |
|
* <p>In addition to execution and lifecycle control methods, this |
48 |
|
* class provides status check methods (for example |
51 |
|
* {@link #toString} returns indications of pool state in a |
52 |
|
* convenient form for informal monitoring. |
53 |
|
* |
54 |
+ |
* <p> As is the case with other ExecutorServices, there are three |
55 |
+ |
* main task execution methods summarized in the follwoing |
56 |
+ |
* table. These are designed to be used by clients not already engaged |
57 |
+ |
* in fork/join computations in the current pool. The main forms of |
58 |
+ |
* these methods accept instances of {@code ForkJoinTask}, but |
59 |
+ |
* overloaded forms also allow mixed execution of plain {@code |
60 |
+ |
* Runnable}- or {@code Callable}- based activities as well. However, |
61 |
+ |
* tasks that are already executing in a pool should normally |
62 |
+ |
* <em>NOT</em> use these pool execution methods, but instead use the |
63 |
+ |
* within-computation forms listed in the table. To avoid inadvertant |
64 |
+ |
* cyclic task dependencies and to improve performance, task |
65 |
+ |
* submissions to the current pool by an ongoing fork/join |
66 |
+ |
* computations may be implicitly translated to the corresponding |
67 |
+ |
* ForkJoinTask forms. |
68 |
+ |
* |
69 |
+ |
* <table BORDER CELLPADDING=3 CELLSPACING=1> |
70 |
+ |
* <tr> |
71 |
+ |
* <td></td> |
72 |
+ |
* <td ALIGN=CENTER> <b>Call from non-fork/join clients</b></td> |
73 |
+ |
* <td ALIGN=CENTER> <b>Call from within fork/join computations</b></td> |
74 |
+ |
* </tr> |
75 |
+ |
* <tr> |
76 |
+ |
* <td> <b>Arange async execution</td> |
77 |
+ |
* <td> {@link #execute(ForkJoinTask)}</td> |
78 |
+ |
* <td> {@link ForkJoinTask#fork}</td> |
79 |
+ |
* </tr> |
80 |
+ |
* <tr> |
81 |
+ |
* <td> <b>Await and obtain result</td> |
82 |
+ |
* <td> {@link #invoke(ForkJoinTask)}</td> |
83 |
+ |
* <td> {@link ForkJoinTask#invoke}</td> |
84 |
+ |
* </tr> |
85 |
+ |
* <tr> |
86 |
+ |
* <td> <b>Arrange exec and obtain Future</td> |
87 |
+ |
* <td> {@link #submit(ForkJoinTask)}</td> |
88 |
+ |
* <td> {@link ForkJoinTask#fork} (ForkJoinTasks <em>are</em> Futures)</td> |
89 |
+ |
* </tr> |
90 |
+ |
* </table> |
91 |
+ |
* |
92 |
|
* <p><b>Sample Usage.</b> Normally a single {@code ForkJoinPool} is |
93 |
|
* used for all parallel task execution in a program or subsystem. |
94 |
|
* Otherwise, use would not usually outweigh the construction and |
164 |
|
* (workerLock) but the array is otherwise concurrently readable, |
165 |
|
* and accessed directly by workers. To simplify index-based |
166 |
|
* operations, the array size is always a power of two, and all |
167 |
< |
* readers must tolerate null slots. Currently, all but the first |
168 |
< |
* worker thread creation is on-demand, triggered by task |
169 |
< |
* submissions, replacement of terminated workers, and/or |
170 |
< |
* compensation for blocked workers. However, all other support |
171 |
< |
* code is set up to work with other policies. |
167 |
> |
* readers must tolerate null slots. Currently, all worker thread |
168 |
> |
* creation is on-demand, triggered by task submissions, |
169 |
> |
* replacement of terminated workers, and/or compensation for |
170 |
> |
* blocked workers. However, all other support code is set up to |
171 |
> |
* work with other policies. |
172 |
|
* |
173 |
|
* 2. Bookkeeping for dynamically adding and removing workers. We |
174 |
< |
* maintain a given level of parallelism (or, if |
175 |
< |
* maintainsParallelism is false, at least avoid starvation). When |
152 |
< |
* some workers are known to be blocked (on joins or via |
174 |
> |
* aim to approximately maintain the given level of parallelism. |
175 |
> |
* When some workers are known to be blocked (on joins or via |
176 |
|
* ManagedBlocker), we may create or resume others to take their |
177 |
|
* place until they unblock (see below). Implementing this |
178 |
|
* requires counts of the number of "running" threads (i.e., those |
180 |
|
* the total number. These two values are packed into one field, |
181 |
|
* "workerCounts" because we need accurate snapshots when deciding |
182 |
|
* to create, resume or suspend. To support these decisions, |
183 |
< |
* updates must be prospective (not retrospective). For example, |
184 |
< |
* the running count is decremented before blocking by a thread |
185 |
< |
* about to block, but incremented by the thread about to unblock |
186 |
< |
* it. (In a few cases, these prospective updates may need to be |
187 |
< |
* rolled back, for example when deciding to create a new worker |
188 |
< |
* but the thread factory fails or returns null. In these cases, |
189 |
< |
* we are no worse off wrt other decisions than we would be |
190 |
< |
* otherwise.) Updates to the workerCounts field sometimes |
191 |
< |
* transiently encounter a fair amount of contention when join |
192 |
< |
* dependencies are such that many threads block or unblock at |
193 |
< |
* about the same time. We alleviate this by sometimes bundling |
194 |
< |
* updates (for example blocking one thread on join and resuming a |
172 |
< |
* spare cancel each other out), and in most other cases |
173 |
< |
* performing an alternative action (like releasing waiters and |
174 |
< |
* finding spares; see below) as a more productive form of |
175 |
< |
* backoff. |
183 |
> |
* updates to spare counts must be prospective (not |
184 |
> |
* retrospective). For example, the running count is decremented |
185 |
> |
* before blocking by a thread about to block as a spare, but |
186 |
> |
* incremented by the thread about to unblock it. Updates upon |
187 |
> |
* resumption ofr threads blocking in awaitJoin or awaitBlocker |
188 |
> |
* cannot usually be prospective, so the running count is in |
189 |
> |
* general an upper bound of the number of productively running |
190 |
> |
* threads Updates to the workerCounts field sometimes transiently |
191 |
> |
* encounter a fair amount of contention when join dependencies |
192 |
> |
* are such that many threads block or unblock at about the same |
193 |
> |
* time. We alleviate this by sometimes performing an alternative |
194 |
> |
* action on contention like releasing waiters or locating spares. |
195 |
|
* |
196 |
|
* 3. Maintaining global run state. The run state of the pool |
197 |
|
* consists of a runLevel (SHUTDOWN, TERMINATING, etc) similar to |
240 |
|
* that only releases idle workers until it detects interference |
241 |
|
* by other threads trying to release, and lets them take |
242 |
|
* over. The net effect is a tree-like diffusion of signals, where |
243 |
< |
* released threads and possibly others) help with unparks. To |
243 |
> |
* released threads (and possibly others) help with unparks. To |
244 |
|
* further reduce contention effects a bit, failed CASes to |
245 |
|
* increment field eventCount are tolerated without retries. |
246 |
|
* Conceptually they are merged into the same event, which is OK |
257 |
|
* "extra" spare threads from normal "core" threads: On each call |
258 |
|
* to preStep (the only point at which we can do this) a worker |
259 |
|
* checks to see if there are now too many running workers, and if |
260 |
< |
* so, suspends itself. Methods preJoin and doBlock look for |
261 |
< |
* suspended threads to resume before considering creating a new |
262 |
< |
* replacement. We don't need a special data structure to maintain |
263 |
< |
* spares; simply scanning the workers array looking for |
260 |
> |
* so, suspends itself. Methods awaitJoin and awaitBlocker look |
261 |
> |
* for suspended threads to resume before considering creating a |
262 |
> |
* new replacement. We don't need a special data structure to |
263 |
> |
* maintain spares; simply scanning the workers array looking for |
264 |
|
* worker.isSuspended() is fine because the calling thread is |
265 |
|
* otherwise not doing anything useful anyway; we are at least as |
266 |
|
* happy if after locating a spare, the caller doesn't actually |
279 |
|
* |
280 |
|
* 6. Deciding when to create new workers. The main dynamic |
281 |
|
* control in this class is deciding when to create extra threads, |
282 |
< |
* in methods preJoin and doBlock. We always need to create one |
283 |
< |
* when the number of running threads becomes zero. But because |
284 |
< |
* blocked joins are typically dependent, we don't necessarily |
285 |
< |
* need or want one-to-one replacement. Using a one-to-one |
286 |
< |
* compensation rule often leads to enough useless overhead |
287 |
< |
* creating, suspending, resuming, and/or killing threads to |
288 |
< |
* signficantly degrade throughput. We use a rule reflecting the |
289 |
< |
* idea that, the more spare threads you already have, the more |
290 |
< |
* evidence you need to create another one; where "evidence" is |
291 |
< |
* expressed as the current deficit -- target minus running |
292 |
< |
* threads. To reduce flickering and drift around target values, |
293 |
< |
* the relation is quadratic: adding a spare if (dc*dc)>=(sc*pc) |
294 |
< |
* (where dc is deficit, sc is number of spare threads and pc is |
295 |
< |
* target parallelism.) This effectively reduces churn at the |
296 |
< |
* price of systematically undershooting target parallelism when |
297 |
< |
* many threads are blocked. However, biasing toward undeshooting |
298 |
< |
* partially compensates for the above mechanics to suspend extra |
280 |
< |
* threads, that normally lead to overshoot because we can only |
281 |
< |
* suspend workers in-between top-level actions. It also better |
282 |
< |
* copes with the fact that some of the methods in this class tend |
283 |
< |
* to never become compiled (but are interpreted), so some |
284 |
< |
* components of the entire set of controls might execute many |
285 |
< |
* times faster than others. And similarly for cases where the |
286 |
< |
* apparent lack of work is just due to GC stalls and other |
287 |
< |
* transient system activity. |
288 |
< |
* |
289 |
< |
* 7. Maintaining other configuration parameters and monitoring |
290 |
< |
* statistics. Updates to fields controlling parallelism level, |
291 |
< |
* max size, etc can only meaningfully take effect for individual |
292 |
< |
* threads upon their next top-level actions; i.e., between |
293 |
< |
* stealing/running tasks/submission, which are separated by calls |
294 |
< |
* to preStep. Memory ordering for these (assumed infrequent) |
295 |
< |
* reconfiguration calls is ensured by using reads and writes to |
296 |
< |
* volatile field workerCounts (that must be read in preStep anyway) |
297 |
< |
* as "fences" -- user-level reads are preceded by reads of |
298 |
< |
* workCounts, and writes are followed by no-op CAS to |
299 |
< |
* workerCounts. The values reported by other management and |
300 |
< |
* monitoring methods are either computed on demand, or are kept |
301 |
< |
* in fields that are only updated when threads are otherwise |
302 |
< |
* idle. |
282 |
> |
* in methods awaitJoin and awaitBlocker. We always need to create |
283 |
> |
* one when the number of running threads becomes zero. But |
284 |
> |
* because blocked joins are typically dependent, we don't |
285 |
> |
* necessarily need or want one-to-one replacement. Instead, we |
286 |
> |
* use a combination of heuristics that adds threads only when the |
287 |
> |
* pool appears to be approaching starvation. These effectively |
288 |
> |
* reduce churn at the price of systematically undershooting |
289 |
> |
* target parallelism when many threads are blocked. However, |
290 |
> |
* biasing toward undeshooting partially compensates for the above |
291 |
> |
* mechanics to suspend extra threads, that normally lead to |
292 |
> |
* overshoot because we can only suspend workers in-between |
293 |
> |
* top-level actions. It also better copes with the fact that some |
294 |
> |
* of the methods in this class tend to never become compiled (but |
295 |
> |
* are interpreted), so some components of the entire set of |
296 |
> |
* controls might execute many times faster than others. And |
297 |
> |
* similarly for cases where the apparent lack of work is just due |
298 |
> |
* to GC stalls and other transient system activity. |
299 |
|
* |
300 |
|
* Beware that there is a lot of representation-level coupling |
301 |
|
* among classes ForkJoinPool, ForkJoinWorkerThread, and |
341 |
|
* Default ForkJoinWorkerThreadFactory implementation; creates a |
342 |
|
* new ForkJoinWorkerThread. |
343 |
|
*/ |
344 |
< |
static class DefaultForkJoinWorkerThreadFactory |
344 |
> |
static class DefaultForkJoinWorkerThreadFactory |
345 |
|
implements ForkJoinWorkerThreadFactory { |
346 |
|
public ForkJoinWorkerThread newThread(ForkJoinPool pool) { |
347 |
|
return new ForkJoinWorkerThread(pool); |
409 |
|
/** |
410 |
|
* Latch released upon termination. |
411 |
|
*/ |
412 |
< |
private final CountDownLatch terminationLatch; |
412 |
> |
private final Phaser termination; |
413 |
|
|
414 |
|
/** |
415 |
|
* Creation factory for worker threads. |
480 |
|
private static final int ONE_RUNNING = 1; |
481 |
|
private static final int ONE_TOTAL = 1 << TOTAL_COUNT_SHIFT; |
482 |
|
|
487 |
– |
/* |
488 |
– |
* Fields parallelism. maxPoolSize, locallyFifo, |
489 |
– |
* maintainsParallelism, and ueh are non-volatile, but external |
490 |
– |
* reads/writes use workerCount fences to ensure visability. |
491 |
– |
*/ |
492 |
– |
|
483 |
|
/** |
484 |
|
* The target parallelism level. |
485 |
+ |
* Accessed directly by ForkJoinWorkerThreads. |
486 |
|
*/ |
487 |
< |
private int parallelism; |
497 |
< |
|
498 |
< |
/** |
499 |
< |
* The maximum allowed pool size. |
500 |
< |
*/ |
501 |
< |
private int maxPoolSize; |
487 |
> |
final int parallelism; |
488 |
|
|
489 |
|
/** |
490 |
|
* True if use local fifo, not default lifo, for local polling |
491 |
< |
* Replicated by ForkJoinWorkerThreads |
491 |
> |
* Read by, and replicated by ForkJoinWorkerThreads |
492 |
|
*/ |
493 |
< |
private boolean locallyFifo; |
493 |
> |
final boolean locallyFifo; |
494 |
|
|
495 |
|
/** |
496 |
< |
* Controls whether to add spares to maintain parallelism |
496 |
> |
* The uncaught exception handler used when any worker abruptly |
497 |
> |
* terminates. |
498 |
|
*/ |
499 |
< |
private boolean maintainsParallelism; |
513 |
< |
|
514 |
< |
/** |
515 |
< |
* The uncaught exception handler used when any worker |
516 |
< |
* abruptly terminates |
517 |
< |
*/ |
518 |
< |
private Thread.UncaughtExceptionHandler ueh; |
499 |
> |
private final Thread.UncaughtExceptionHandler ueh; |
500 |
|
|
501 |
|
/** |
502 |
|
* Pool number, just for assigning useful names to worker threads |
506 |
|
// utilities for updating fields |
507 |
|
|
508 |
|
/** |
509 |
< |
* Adds delta to running count. Used mainly by ForkJoinTask. |
529 |
< |
* |
530 |
< |
* @param delta the number to add |
509 |
> |
* Increments running count. Also used by ForkJoinTask. |
510 |
|
*/ |
511 |
< |
final void updateRunningCount(int delta) { |
512 |
< |
int wc; |
511 |
> |
final void incrementRunningCount() { |
512 |
> |
int c; |
513 |
|
do {} while (!UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
514 |
< |
wc = workerCounts, |
515 |
< |
wc + delta)); |
537 |
< |
} |
538 |
< |
|
539 |
< |
/** |
540 |
< |
* Write fence for user modifications of pool parameters |
541 |
< |
* (parallelism. etc). Note that it doesn't matter if CAS fails. |
542 |
< |
*/ |
543 |
< |
private void workerCountWriteFence() { |
544 |
< |
int wc; |
545 |
< |
UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
546 |
< |
wc = workerCounts, wc); |
514 |
> |
c = workerCounts, |
515 |
> |
c + ONE_RUNNING)); |
516 |
|
} |
517 |
< |
|
517 |
> |
|
518 |
|
/** |
519 |
< |
* Read fence for external reads of pool parameters |
551 |
< |
* (parallelism. maxPoolSize, etc). |
519 |
> |
* Tries to decrement running count unless already zero |
520 |
|
*/ |
521 |
< |
private void workerCountReadFence() { |
522 |
< |
int ignore = workerCounts; |
521 |
> |
final boolean tryDecrementRunningCount() { |
522 |
> |
int wc = workerCounts; |
523 |
> |
if ((wc & RUNNING_COUNT_MASK) == 0) |
524 |
> |
return false; |
525 |
> |
return UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
526 |
> |
wc, wc - ONE_RUNNING); |
527 |
|
} |
528 |
|
|
529 |
|
/** |
574 |
|
lock.lock(); |
575 |
|
try { |
576 |
|
ForkJoinWorkerThread[] ws = workers; |
577 |
< |
int len = ws.length; |
578 |
< |
if (k < 0 || k >= len || ws[k] != null) { |
579 |
< |
for (k = 0; k < len && ws[k] != null; ++k) |
577 |
> |
int nws = ws.length; |
578 |
> |
if (k < 0 || k >= nws || ws[k] != null) { |
579 |
> |
for (k = 0; k < nws && ws[k] != null; ++k) |
580 |
|
; |
581 |
< |
if (k == len) |
582 |
< |
ws = Arrays.copyOf(ws, len << 1); |
581 |
> |
if (k == nws) |
582 |
> |
ws = Arrays.copyOf(ws, nws << 1); |
583 |
|
} |
584 |
|
ws[k] = w; |
585 |
|
workers = ws; // volatile array write ensures slot visibility |
625 |
|
return null; |
626 |
|
} |
627 |
|
} |
628 |
< |
w.start(recordWorker(w), locallyFifo, ueh); |
628 |
> |
w.start(recordWorker(w), ueh); |
629 |
|
return w; |
630 |
|
} |
631 |
|
|
633 |
|
* Adjusts counts upon failure to create worker |
634 |
|
*/ |
635 |
|
private void onWorkerCreationFailure() { |
636 |
< |
int c; |
637 |
< |
do {} while (!UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
638 |
< |
c = workerCounts, |
639 |
< |
c - (ONE_RUNNING|ONE_TOTAL))); |
636 |
> |
for (;;) { |
637 |
> |
int wc = workerCounts; |
638 |
> |
if ((wc >>> TOTAL_COUNT_SHIFT) > 0 && |
639 |
> |
UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
640 |
> |
wc, wc - (ONE_RUNNING|ONE_TOTAL))) |
641 |
> |
break; |
642 |
> |
} |
643 |
|
tryTerminate(false); // in case of failure during shutdown |
644 |
|
} |
645 |
|
|
649 |
|
*/ |
650 |
|
private void ensureEnoughTotalWorkers() { |
651 |
|
int wc; |
652 |
< |
while (runState < TERMINATING && |
653 |
< |
((wc = workerCounts) >>> TOTAL_COUNT_SHIFT) < parallelism) { |
652 |
> |
while (((wc = workerCounts) >>> TOTAL_COUNT_SHIFT) < parallelism && |
653 |
> |
runState < TERMINATING) { |
654 |
|
if ((UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
655 |
|
wc, wc + (ONE_RUNNING|ONE_TOTAL)) && |
656 |
|
addWorker() == null)) |
673 |
|
} |
674 |
|
forgetWorker(w); |
675 |
|
|
676 |
< |
// decrement total count, and if was running, running count |
677 |
< |
int unit = w.isTrimmed()? ONE_TOTAL : (ONE_RUNNING|ONE_TOTAL); |
678 |
< |
int wc; |
679 |
< |
do {} while (!UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
680 |
< |
wc = workerCounts, wc - unit)); |
676 |
> |
// Decrement total count, and if was running, running count |
677 |
> |
// Spin (waiting for other updates) if either would be negative |
678 |
> |
int nr = w.isTrimmed() ? 0 : ONE_RUNNING; |
679 |
> |
int unit = ONE_TOTAL + nr; |
680 |
> |
for (;;) { |
681 |
> |
int wc = workerCounts; |
682 |
> |
int rc = wc & RUNNING_COUNT_MASK; |
683 |
> |
if (rc - nr < 0 || (wc >>> TOTAL_COUNT_SHIFT) == 0) |
684 |
> |
Thread.yield(); // back off if waiting for other updates |
685 |
> |
else if (UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
686 |
> |
wc, wc - unit)) |
687 |
> |
break; |
688 |
> |
} |
689 |
|
|
690 |
|
accumulateStealCount(w); // collect final count |
691 |
|
if (!tryTerminate(false)) |
695 |
|
// Waiting for and signalling events |
696 |
|
|
697 |
|
/** |
715 |
– |
* Ensures eventCount on exit is different (mod 2^32) than on |
716 |
– |
* entry. CAS failures are OK -- any change in count suffices. |
717 |
– |
*/ |
718 |
– |
private void advanceEventCount() { |
719 |
– |
int c; |
720 |
– |
UNSAFE.compareAndSwapInt(this, eventCountOffset, c = eventCount, c+1); |
721 |
– |
} |
722 |
– |
|
723 |
– |
/** |
698 |
|
* Releases workers blocked on a count not equal to current count. |
699 |
|
*/ |
700 |
< |
final void releaseWaiters() { |
700 |
> |
private void releaseWaiters() { |
701 |
|
long top; |
702 |
|
int id; |
703 |
|
while ((id = (int)((top = eventWaiters) & WAITER_INDEX_MASK)) > 0 && |
712 |
|
} |
713 |
|
|
714 |
|
/** |
715 |
+ |
* Ensures eventCount on exit is different (mod 2^32) than on |
716 |
+ |
* entry and wakes up all waiters |
717 |
+ |
*/ |
718 |
+ |
private void signalEvent() { |
719 |
+ |
int c; |
720 |
+ |
do {} while (!UNSAFE.compareAndSwapInt(this, eventCountOffset, |
721 |
+ |
c = eventCount, c+1)); |
722 |
+ |
releaseWaiters(); |
723 |
+ |
} |
724 |
+ |
|
725 |
+ |
/** |
726 |
|
* Advances eventCount and releases waiters until interference by |
727 |
|
* other releasing threads is detected. |
728 |
|
*/ |
729 |
|
final void signalWork() { |
730 |
+ |
// EventCount CAS failures are OK -- any change in count suffices. |
731 |
|
int ec; |
732 |
|
UNSAFE.compareAndSwapInt(this, eventCountOffset, ec=eventCount, ec+1); |
733 |
|
outer:for (;;) { |
806 |
|
boolean inactivate = !worked & active; |
807 |
|
for (;;) { |
808 |
|
if (inactivate) { |
809 |
< |
int c = runState; |
809 |
> |
int rs = runState; |
810 |
|
if (UNSAFE.compareAndSwapInt(this, runStateOffset, |
811 |
< |
c, c - ONE_ACTIVE)) |
811 |
> |
rs, rs - ONE_ACTIVE)) |
812 |
|
inactivate = active = w.active = false; |
813 |
|
} |
814 |
|
int wc = workerCounts; |
826 |
|
} |
827 |
|
|
828 |
|
/** |
829 |
< |
* Adjusts counts and creates or resumes compensating threads for |
830 |
< |
* a worker about to block on task joinMe, returning early if |
831 |
< |
* joinMe becomes ready. First tries resuming an existing spare |
832 |
< |
* (which usually also avoids any count adjustment), but must then |
833 |
< |
* decrement running count to determine whether a new thread is |
834 |
< |
* needed. See above for fuller explanation. |
835 |
< |
*/ |
836 |
< |
final void preJoin(ForkJoinTask<?> joinMe) { |
837 |
< |
boolean dec = false; // true when running count decremented |
838 |
< |
for (;;) { |
839 |
< |
releaseWaiters(); // help other threads progress |
840 |
< |
|
841 |
< |
if (joinMe.status < 0) // surround spare search with done checks |
842 |
< |
return; |
843 |
< |
ForkJoinWorkerThread spare = null; |
844 |
< |
for (ForkJoinWorkerThread w : workers) { |
845 |
< |
if (w != null && w.isSuspended()) { |
846 |
< |
spare = w; |
829 |
> |
* Tries to decrement running count, and if so, possibly creates |
830 |
> |
* or resumes compensating threads before blocking on task joinMe. |
831 |
> |
* This code is sprawled out with manual inlining to evade some |
832 |
> |
* JIT oddities. |
833 |
> |
* |
834 |
> |
* @param joinMe the task to join |
835 |
> |
* @return task status on exit |
836 |
> |
*/ |
837 |
> |
final int tryAwaitJoin(ForkJoinTask<?> joinMe) { |
838 |
> |
int cw = workerCounts; // read now to spoil CAS if counts change as ... |
839 |
> |
releaseWaiters(); // ... a byproduct of releaseWaiters |
840 |
> |
int stat = joinMe.status; |
841 |
> |
if (stat >= 0 && // inline variant of tryDecrementRunningCount |
842 |
> |
(cw & RUNNING_COUNT_MASK) > 0 && |
843 |
> |
UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
844 |
> |
cw, cw - ONE_RUNNING)) { |
845 |
> |
int pc = parallelism; |
846 |
> |
int scans = 0; // to require confirming passes to add threads |
847 |
> |
outer: while ((workerCounts & RUNNING_COUNT_MASK) < pc) { |
848 |
> |
if ((stat = joinMe.status) < 0) |
849 |
|
break; |
850 |
+ |
ForkJoinWorkerThread spare = null; |
851 |
+ |
ForkJoinWorkerThread[] ws = workers; |
852 |
+ |
int nws = ws.length; |
853 |
+ |
for (int i = 0; i < nws; ++i) { |
854 |
+ |
ForkJoinWorkerThread w = ws[i]; |
855 |
+ |
if (w != null && w.isSuspended()) { |
856 |
+ |
spare = w; |
857 |
+ |
break; |
858 |
+ |
} |
859 |
+ |
} |
860 |
+ |
if ((stat = joinMe.status) < 0) // recheck to narrow race |
861 |
+ |
break; |
862 |
+ |
int wc = workerCounts; |
863 |
+ |
int rc = wc & RUNNING_COUNT_MASK; |
864 |
+ |
if (rc >= pc) |
865 |
+ |
break; |
866 |
+ |
if (spare != null) { |
867 |
+ |
if (spare.tryUnsuspend()) { |
868 |
+ |
int c; // inline incrementRunningCount |
869 |
+ |
do {} while (!UNSAFE.compareAndSwapInt |
870 |
+ |
(this, workerCountsOffset, |
871 |
+ |
c = workerCounts, c + ONE_RUNNING)); |
872 |
+ |
LockSupport.unpark(spare); |
873 |
+ |
break; |
874 |
+ |
} |
875 |
+ |
continue; |
876 |
|
} |
863 |
– |
} |
864 |
– |
if (joinMe.status < 0) |
865 |
– |
return; |
866 |
– |
|
867 |
– |
if (spare != null && spare.tryUnsuspend()) { |
868 |
– |
if (dec || joinMe.requestSignal() < 0) { |
869 |
– |
int c; |
870 |
– |
do {} while (!UNSAFE.compareAndSwapInt(this, |
871 |
– |
workerCountsOffset, |
872 |
– |
c = workerCounts, |
873 |
– |
c + ONE_RUNNING)); |
874 |
– |
} // else no net count change |
875 |
– |
LockSupport.unpark(spare); |
876 |
– |
return; |
877 |
– |
} |
878 |
– |
|
879 |
– |
int wc = workerCounts; // decrement running count |
880 |
– |
if (!dec && (wc & RUNNING_COUNT_MASK) != 0 && |
881 |
– |
(dec = UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
882 |
– |
wc, wc -= ONE_RUNNING)) && |
883 |
– |
joinMe.requestSignal() < 0) { // cannot block |
884 |
– |
int c; // back out |
885 |
– |
do {} while (!UNSAFE.compareAndSwapInt(this, |
886 |
– |
workerCountsOffset, |
887 |
– |
c = workerCounts, |
888 |
– |
c + ONE_RUNNING)); |
889 |
– |
return; |
890 |
– |
} |
891 |
– |
|
892 |
– |
if (dec) { |
877 |
|
int tc = wc >>> TOTAL_COUNT_SHIFT; |
878 |
< |
int pc = parallelism; |
879 |
< |
int dc = pc - (wc & RUNNING_COUNT_MASK); // deficit count |
880 |
< |
if ((dc < pc && (dc <= 0 || (dc * dc < (tc - pc) * pc) || |
881 |
< |
!maintainsParallelism)) || |
882 |
< |
tc >= maxPoolSize) // cannot add |
883 |
< |
return; |
884 |
< |
if (spare == null && |
878 |
> |
int sc = tc - pc; |
879 |
> |
if (rc > 0) { |
880 |
> |
int p = pc; |
881 |
> |
int s = sc; |
882 |
> |
while (s-- >= 0) { // try keeping 3/4 live |
883 |
> |
if (rc > (p -= (p >>> 2) + 1)) |
884 |
> |
break outer; |
885 |
> |
} |
886 |
> |
} |
887 |
> |
if (scans++ > sc && tc < MAX_THREADS && |
888 |
|
UNSAFE.compareAndSwapInt(this, workerCountsOffset, wc, |
889 |
|
wc + (ONE_RUNNING|ONE_TOTAL))) { |
890 |
< |
addWorker(); |
891 |
< |
return; |
890 |
> |
addWorker(); |
891 |
> |
break; |
892 |
|
} |
893 |
|
} |
894 |
+ |
if (stat >= 0) |
895 |
+ |
stat = joinMe.internalAwaitDone(); |
896 |
+ |
int c; // inline incrementRunningCount |
897 |
+ |
do {} while (!UNSAFE.compareAndSwapInt |
898 |
+ |
(this, workerCountsOffset, |
899 |
+ |
c = workerCounts, c + ONE_RUNNING)); |
900 |
|
} |
901 |
+ |
return stat; |
902 |
|
} |
903 |
|
|
904 |
|
/** |
905 |
< |
* Same idea as preJoin but with too many differing details to |
906 |
< |
* integrate: There are no task-based signal counts, and only one |
913 |
< |
* way to do the actual blocking. So for simplicity it is directly |
914 |
< |
* incorporated into this method. |
905 |
> |
* Same idea as (and mostly pasted from) tryAwaitJoin, but |
906 |
> |
* self-contained |
907 |
|
*/ |
908 |
< |
final void doBlock(ManagedBlocker blocker, boolean maintainPar) |
908 |
> |
final void awaitBlocker(ManagedBlocker blocker) |
909 |
|
throws InterruptedException { |
918 |
– |
maintainPar &= maintainsParallelism; // override |
919 |
– |
boolean dec = false; |
920 |
– |
boolean done = false; |
910 |
|
for (;;) { |
911 |
+ |
if (blocker.isReleasable()) |
912 |
+ |
return; |
913 |
+ |
int cw = workerCounts; |
914 |
|
releaseWaiters(); |
915 |
+ |
if ((cw & RUNNING_COUNT_MASK) > 0 && |
916 |
+ |
UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
917 |
+ |
cw, cw - ONE_RUNNING)) |
918 |
+ |
break; |
919 |
+ |
} |
920 |
+ |
boolean done = false; |
921 |
+ |
int pc = parallelism; |
922 |
+ |
int scans = 0; |
923 |
+ |
outer: while ((workerCounts & RUNNING_COUNT_MASK) < pc) { |
924 |
|
if (done = blocker.isReleasable()) |
925 |
|
break; |
926 |
|
ForkJoinWorkerThread spare = null; |
927 |
< |
for (ForkJoinWorkerThread w : workers) { |
927 |
> |
ForkJoinWorkerThread[] ws = workers; |
928 |
> |
int nws = ws.length; |
929 |
> |
for (int i = 0; i < nws; ++i) { |
930 |
> |
ForkJoinWorkerThread w = ws[i]; |
931 |
|
if (w != null && w.isSuspended()) { |
932 |
|
spare = w; |
933 |
|
break; |
935 |
|
} |
936 |
|
if (done = blocker.isReleasable()) |
937 |
|
break; |
938 |
< |
if (spare != null && spare.tryUnsuspend()) { |
939 |
< |
if (dec) { |
938 |
> |
int wc = workerCounts; |
939 |
> |
int rc = wc & RUNNING_COUNT_MASK; |
940 |
> |
if (rc >= pc) |
941 |
> |
break; |
942 |
> |
if (spare != null) { |
943 |
> |
if (spare.tryUnsuspend()) { |
944 |
|
int c; |
945 |
< |
do {} while (!UNSAFE.compareAndSwapInt(this, |
946 |
< |
workerCountsOffset, |
947 |
< |
c = workerCounts, |
948 |
< |
c + ONE_RUNNING)); |
945 |
> |
do {} while (!UNSAFE.compareAndSwapInt |
946 |
> |
(this, workerCountsOffset, |
947 |
> |
c = workerCounts, c + ONE_RUNNING)); |
948 |
> |
LockSupport.unpark(spare); |
949 |
> |
break; |
950 |
|
} |
951 |
< |
LockSupport.unpark(spare); |
943 |
< |
break; |
951 |
> |
continue; |
952 |
|
} |
953 |
< |
int wc = workerCounts; |
954 |
< |
if (!dec && (wc & RUNNING_COUNT_MASK) != 0) |
955 |
< |
dec = UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
956 |
< |
wc, wc -= ONE_RUNNING); |
957 |
< |
if (dec) { |
958 |
< |
int tc = wc >>> TOTAL_COUNT_SHIFT; |
959 |
< |
int pc = parallelism; |
960 |
< |
int dc = pc - (wc & RUNNING_COUNT_MASK); |
953 |
< |
if ((dc < pc && (dc <= 0 || (dc * dc < (tc - pc) * pc) || |
954 |
< |
!maintainPar)) || |
955 |
< |
tc >= maxPoolSize) |
956 |
< |
break; |
957 |
< |
if (spare == null && |
958 |
< |
UNSAFE.compareAndSwapInt(this, workerCountsOffset, wc, |
959 |
< |
wc + (ONE_RUNNING|ONE_TOTAL))){ |
960 |
< |
addWorker(); |
961 |
< |
break; |
953 |
> |
int tc = wc >>> TOTAL_COUNT_SHIFT; |
954 |
> |
int sc = tc - pc; |
955 |
> |
if (rc > 0) { |
956 |
> |
int p = pc; |
957 |
> |
int s = sc; |
958 |
> |
while (s-- >= 0) { |
959 |
> |
if (rc > (p -= (p >>> 2) + 1)) |
960 |
> |
break outer; |
961 |
|
} |
962 |
|
} |
963 |
+ |
if (scans++ > sc && tc < MAX_THREADS && |
964 |
+ |
UNSAFE.compareAndSwapInt(this, workerCountsOffset, wc, |
965 |
+ |
wc + (ONE_RUNNING|ONE_TOTAL))) { |
966 |
+ |
addWorker(); |
967 |
+ |
break; |
968 |
+ |
} |
969 |
|
} |
965 |
– |
|
970 |
|
try { |
971 |
|
if (!done) |
972 |
< |
do {} while (!blocker.isReleasable() && !blocker.block()); |
972 |
> |
do {} while (!blocker.isReleasable() && |
973 |
> |
!blocker.block()); |
974 |
|
} finally { |
975 |
< |
if (dec) { |
976 |
< |
int c; |
977 |
< |
do {} while (!UNSAFE.compareAndSwapInt(this, |
978 |
< |
workerCountsOffset, |
974 |
< |
c = workerCounts, |
975 |
< |
c + ONE_RUNNING)); |
976 |
< |
} |
977 |
< |
} |
978 |
< |
} |
979 |
< |
|
980 |
< |
/** |
981 |
< |
* Unless there are not enough other running threads, adjusts |
982 |
< |
* counts for a a worker in performing helpJoin that cannot find |
983 |
< |
* any work, so that this worker can now block. |
984 |
< |
* |
985 |
< |
* @return true if worker may block |
986 |
< |
*/ |
987 |
< |
final boolean preBlockHelpingJoin(ForkJoinTask<?> joinMe) { |
988 |
< |
while (joinMe.status >= 0) { |
989 |
< |
releaseWaiters(); // help other threads progress |
990 |
< |
|
991 |
< |
// if a spare exists, resume it to maintain parallelism level |
992 |
< |
if ((workerCounts & RUNNING_COUNT_MASK) <= parallelism) { |
993 |
< |
ForkJoinWorkerThread spare = null; |
994 |
< |
for (ForkJoinWorkerThread w : workers) { |
995 |
< |
if (w != null && w.isSuspended()) { |
996 |
< |
spare = w; |
997 |
< |
break; |
998 |
< |
} |
999 |
< |
} |
1000 |
< |
if (joinMe.status < 0) |
1001 |
< |
break; |
1002 |
< |
if (spare != null) { |
1003 |
< |
if (spare.tryUnsuspend()) { |
1004 |
< |
boolean canBlock = true; |
1005 |
< |
if (joinMe.requestSignal() < 0) { |
1006 |
< |
canBlock = false; // already done |
1007 |
< |
int c; |
1008 |
< |
do {} while (!UNSAFE.compareAndSwapInt |
1009 |
< |
(this, workerCountsOffset, |
1010 |
< |
c = workerCounts, c + ONE_RUNNING)); |
1011 |
< |
} |
1012 |
< |
LockSupport.unpark(spare); |
1013 |
< |
return canBlock; |
1014 |
< |
} |
1015 |
< |
continue; // recheck -- another spare may exist |
1016 |
< |
} |
1017 |
< |
} |
1018 |
< |
|
1019 |
< |
int wc = workerCounts; // reread to shorten CAS window |
1020 |
< |
int rc = wc & RUNNING_COUNT_MASK; |
1021 |
< |
if (rc <= 2) // keep this and at most one other thread alive |
1022 |
< |
break; |
1023 |
< |
|
1024 |
< |
if (UNSAFE.compareAndSwapInt(this, workerCountsOffset, |
1025 |
< |
wc, wc - ONE_RUNNING)) { |
1026 |
< |
if (joinMe.requestSignal() >= 0) |
1027 |
< |
return true; |
1028 |
< |
int c; // back out |
1029 |
< |
do {} while (!UNSAFE.compareAndSwapInt |
1030 |
< |
(this, workerCountsOffset, |
1031 |
< |
c = workerCounts, c + ONE_RUNNING)); |
1032 |
< |
break; |
1033 |
< |
} |
975 |
> |
int c; |
976 |
> |
do {} while (!UNSAFE.compareAndSwapInt |
977 |
> |
(this, workerCountsOffset, |
978 |
> |
c = workerCounts, c + ONE_RUNNING)); |
979 |
|
} |
980 |
< |
return false; |
1036 |
< |
} |
980 |
> |
} |
981 |
|
|
982 |
|
/** |
983 |
|
* Possibly initiates and/or completes termination. |
1000 |
|
// Finish now if all threads terminated; else in some subsequent call |
1001 |
|
if ((workerCounts >>> TOTAL_COUNT_SHIFT) == 0) { |
1002 |
|
advanceRunLevel(TERMINATED); |
1003 |
< |
terminationLatch.countDown(); |
1003 |
> |
termination.arrive(); |
1004 |
|
} |
1005 |
|
return true; |
1006 |
|
} |
1009 |
|
* Actions on transition to TERMINATING |
1010 |
|
*/ |
1011 |
|
private void startTerminating() { |
1012 |
< |
// Clear out and cancel submissions, ignoring exceptions |
1012 |
> |
for (int i = 0; i < 2; ++i) { // twice to mop up newly created workers |
1013 |
> |
cancelSubmissions(); |
1014 |
> |
shutdownWorkers(); |
1015 |
> |
cancelWorkerTasks(); |
1016 |
> |
signalEvent(); |
1017 |
> |
interruptWorkers(); |
1018 |
> |
} |
1019 |
> |
} |
1020 |
> |
|
1021 |
> |
/** |
1022 |
> |
* Clear out and cancel submissions, ignoring exceptions |
1023 |
> |
*/ |
1024 |
> |
private void cancelSubmissions() { |
1025 |
|
ForkJoinTask<?> task; |
1026 |
|
while ((task = submissionQueue.poll()) != null) { |
1027 |
|
try { |
1029 |
|
} catch (Throwable ignore) { |
1030 |
|
} |
1031 |
|
} |
1032 |
< |
// Propagate run level |
1033 |
< |
for (ForkJoinWorkerThread w : workers) { |
1032 |
> |
} |
1033 |
> |
|
1034 |
> |
/** |
1035 |
> |
* Sets all worker run states to at least shutdown, |
1036 |
> |
* also resuming suspended workers |
1037 |
> |
*/ |
1038 |
> |
private void shutdownWorkers() { |
1039 |
> |
ForkJoinWorkerThread[] ws = workers; |
1040 |
> |
int nws = ws.length; |
1041 |
> |
for (int i = 0; i < nws; ++i) { |
1042 |
> |
ForkJoinWorkerThread w = ws[i]; |
1043 |
|
if (w != null) |
1044 |
< |
w.shutdown(); // also resumes suspended workers |
1044 |
> |
w.shutdown(); |
1045 |
|
} |
1046 |
< |
// Ensure no straggling local tasks |
1047 |
< |
for (ForkJoinWorkerThread w : workers) { |
1046 |
> |
} |
1047 |
> |
|
1048 |
> |
/** |
1049 |
> |
* Clears out and cancels all locally queued tasks |
1050 |
> |
*/ |
1051 |
> |
private void cancelWorkerTasks() { |
1052 |
> |
ForkJoinWorkerThread[] ws = workers; |
1053 |
> |
int nws = ws.length; |
1054 |
> |
for (int i = 0; i < nws; ++i) { |
1055 |
> |
ForkJoinWorkerThread w = ws[i]; |
1056 |
|
if (w != null) |
1057 |
|
w.cancelTasks(); |
1058 |
|
} |
1059 |
< |
// Wake up idle workers |
1060 |
< |
advanceEventCount(); |
1061 |
< |
releaseWaiters(); |
1062 |
< |
// Unstick pending joins |
1063 |
< |
for (ForkJoinWorkerThread w : workers) { |
1059 |
> |
} |
1060 |
> |
|
1061 |
> |
/** |
1062 |
> |
* Unsticks all workers blocked on joins etc |
1063 |
> |
*/ |
1064 |
> |
private void interruptWorkers() { |
1065 |
> |
ForkJoinWorkerThread[] ws = workers; |
1066 |
> |
int nws = ws.length; |
1067 |
> |
for (int i = 0; i < nws; ++i) { |
1068 |
> |
ForkJoinWorkerThread w = ws[i]; |
1069 |
|
if (w != null && !w.isTerminated()) { |
1070 |
|
try { |
1071 |
|
w.interrupt(); |
1103 |
|
* active thread. |
1104 |
|
*/ |
1105 |
|
final int idlePerActive() { |
1128 |
– |
int ac = runState; // no mask -- artifically boosts during shutdown |
1106 |
|
int pc = parallelism; // use targeted parallelism, not rc |
1107 |
+ |
int ac = runState; // no mask -- artifically boosts during shutdown |
1108 |
|
// Use exact results for small values, saturate past 4 |
1109 |
|
return pc <= ac? 0 : pc >>> 1 <= ac? 1 : pc >>> 2 <= ac? 3 : pc >>> 3; |
1110 |
|
} |
1115 |
|
|
1116 |
|
/** |
1117 |
|
* Creates a {@code ForkJoinPool} with parallelism equal to {@link |
1118 |
< |
* java.lang.Runtime#availableProcessors}, and using the {@linkplain |
1119 |
< |
* #defaultForkJoinWorkerThreadFactory default thread factory}. |
1118 |
> |
* java.lang.Runtime#availableProcessors}, using the {@linkplain |
1119 |
> |
* #defaultForkJoinWorkerThreadFactory default thread factory}, |
1120 |
> |
* no UncaughtExceptionHandler, and non-async LIFO processing mode. |
1121 |
|
* |
1122 |
|
* @throws SecurityException if a security manager exists and |
1123 |
|
* the caller is not permitted to modify threads |
1126 |
|
*/ |
1127 |
|
public ForkJoinPool() { |
1128 |
|
this(Runtime.getRuntime().availableProcessors(), |
1129 |
< |
defaultForkJoinWorkerThreadFactory); |
1129 |
> |
defaultForkJoinWorkerThreadFactory, null, false); |
1130 |
|
} |
1131 |
|
|
1132 |
|
/** |
1133 |
|
* Creates a {@code ForkJoinPool} with the indicated parallelism |
1134 |
< |
* level and using the {@linkplain |
1135 |
< |
* #defaultForkJoinWorkerThreadFactory default thread factory}. |
1134 |
> |
* level, the {@linkplain |
1135 |
> |
* #defaultForkJoinWorkerThreadFactory default thread factory}, |
1136 |
> |
* no UncaughtExceptionHandler, and non-async LIFO processing mode. |
1137 |
|
* |
1138 |
|
* @param parallelism the parallelism level |
1139 |
|
* @throws IllegalArgumentException if parallelism less than or |
1144 |
|
* java.lang.RuntimePermission}{@code ("modifyThread")} |
1145 |
|
*/ |
1146 |
|
public ForkJoinPool(int parallelism) { |
1147 |
< |
this(parallelism, defaultForkJoinWorkerThreadFactory); |
1168 |
< |
} |
1169 |
< |
|
1170 |
< |
/** |
1171 |
< |
* Creates a {@code ForkJoinPool} with parallelism equal to {@link |
1172 |
< |
* java.lang.Runtime#availableProcessors}, and using the given |
1173 |
< |
* thread factory. |
1174 |
< |
* |
1175 |
< |
* @param factory the factory for creating new threads |
1176 |
< |
* @throws NullPointerException if the factory is null |
1177 |
< |
* @throws SecurityException if a security manager exists and |
1178 |
< |
* the caller is not permitted to modify threads |
1179 |
< |
* because it does not hold {@link |
1180 |
< |
* java.lang.RuntimePermission}{@code ("modifyThread")} |
1181 |
< |
*/ |
1182 |
< |
public ForkJoinPool(ForkJoinWorkerThreadFactory factory) { |
1183 |
< |
this(Runtime.getRuntime().availableProcessors(), factory); |
1147 |
> |
this(parallelism, defaultForkJoinWorkerThreadFactory, null, false); |
1148 |
|
} |
1149 |
|
|
1150 |
|
/** |
1151 |
< |
* Creates a {@code ForkJoinPool} with the given parallelism and |
1188 |
< |
* thread factory. |
1151 |
> |
* Creates a {@code ForkJoinPool} with the given parameters. |
1152 |
|
* |
1153 |
< |
* @param parallelism the parallelism level |
1154 |
< |
* @param factory the factory for creating new threads |
1153 |
> |
* @param parallelism the parallelism level. For default value, |
1154 |
> |
* use {@link java.lang.Runtime#availableProcessors}. |
1155 |
> |
* @param factory the factory for creating new threads. For default value, |
1156 |
> |
* use {@link #defaultForkJoinWorkerThreadFactory}. |
1157 |
> |
* @param handler the handler for internal worker threads that |
1158 |
> |
* terminate due to unrecoverable errors encountered while executing |
1159 |
> |
* tasks. For default value, use <code>null</code>. |
1160 |
> |
* @param asyncMode if true, |
1161 |
> |
* establishes local first-in-first-out scheduling mode for forked |
1162 |
> |
* tasks that are never joined. This mode may be more appropriate |
1163 |
> |
* than default locally stack-based mode in applications in which |
1164 |
> |
* worker threads only process event-style asynchronous tasks. |
1165 |
> |
* For default value, use <code>false</code>. |
1166 |
|
* @throws IllegalArgumentException if parallelism less than or |
1167 |
|
* equal to zero, or greater than implementation limit |
1168 |
|
* @throws NullPointerException if the factory is null |
1171 |
|
* because it does not hold {@link |
1172 |
|
* java.lang.RuntimePermission}{@code ("modifyThread")} |
1173 |
|
*/ |
1174 |
< |
public ForkJoinPool(int parallelism, ForkJoinWorkerThreadFactory factory) { |
1174 |
> |
public ForkJoinPool(int parallelism, |
1175 |
> |
ForkJoinWorkerThreadFactory factory, |
1176 |
> |
Thread.UncaughtExceptionHandler handler, |
1177 |
> |
boolean asyncMode) { |
1178 |
|
checkPermission(); |
1179 |
|
if (factory == null) |
1180 |
|
throw new NullPointerException(); |
1181 |
|
if (parallelism <= 0 || parallelism > MAX_THREADS) |
1182 |
|
throw new IllegalArgumentException(); |
1206 |
– |
this.poolNumber = poolNumberGenerator.incrementAndGet(); |
1207 |
– |
int arraySize = initialArraySizeFor(parallelism); |
1183 |
|
this.parallelism = parallelism; |
1184 |
|
this.factory = factory; |
1185 |
< |
this.maxPoolSize = MAX_THREADS; |
1186 |
< |
this.maintainsParallelism = true; |
1185 |
> |
this.ueh = handler; |
1186 |
> |
this.locallyFifo = asyncMode; |
1187 |
> |
int arraySize = initialArraySizeFor(parallelism); |
1188 |
|
this.workers = new ForkJoinWorkerThread[arraySize]; |
1189 |
|
this.submissionQueue = new LinkedTransferQueue<ForkJoinTask<?>>(); |
1190 |
|
this.workerLock = new ReentrantLock(); |
1191 |
< |
this.terminationLatch = new CountDownLatch(1); |
1192 |
< |
// Start first worker; remaining workers added upon first submission |
1217 |
< |
workerCounts = ONE_RUNNING | ONE_TOTAL; |
1218 |
< |
addWorker(); |
1191 |
> |
this.termination = new Phaser(1); |
1192 |
> |
this.poolNumber = poolNumberGenerator.incrementAndGet(); |
1193 |
|
} |
1194 |
|
|
1195 |
|
/** |
1216 |
|
throw new NullPointerException(); |
1217 |
|
if (runState >= SHUTDOWN) |
1218 |
|
throw new RejectedExecutionException(); |
1219 |
< |
submissionQueue.offer(task); |
1220 |
< |
advanceEventCount(); |
1221 |
< |
releaseWaiters(); |
1222 |
< |
if ((workerCounts >>> TOTAL_COUNT_SHIFT) < parallelism) |
1219 |
> |
// Convert submissions to current pool into forks |
1220 |
> |
Thread t = Thread.currentThread(); |
1221 |
> |
ForkJoinWorkerThread w; |
1222 |
> |
if ((t instanceof ForkJoinWorkerThread) && |
1223 |
> |
(w = (ForkJoinWorkerThread) t).pool == this) |
1224 |
> |
w.pushTask(task); |
1225 |
> |
else { |
1226 |
> |
submissionQueue.offer(task); |
1227 |
> |
signalEvent(); |
1228 |
|
ensureEnoughTotalWorkers(); |
1229 |
+ |
} |
1230 |
|
} |
1231 |
|
|
1232 |
|
/** |
1233 |
|
* Performs the given task, returning its result upon completion. |
1234 |
+ |
* If the caller is already engaged in a fork/join computation in |
1235 |
+ |
* the current pool, this method is equivalent in effect to |
1236 |
+ |
* {@link ForkJoinTask#invoke}. |
1237 |
|
* |
1238 |
|
* @param task the task |
1239 |
|
* @return the task's result |
1248 |
|
|
1249 |
|
/** |
1250 |
|
* Arranges for (asynchronous) execution of the given task. |
1251 |
+ |
* If the caller is already engaged in a fork/join computation in |
1252 |
+ |
* the current pool, this method is equivalent in effect to |
1253 |
+ |
* {@link ForkJoinTask#fork}. |
1254 |
|
* |
1255 |
|
* @param task the task |
1256 |
|
* @throws NullPointerException if the task is null |
1278 |
|
} |
1279 |
|
|
1280 |
|
/** |
1281 |
+ |
* Submits a ForkJoinTask for execution. |
1282 |
+ |
* If the caller is already engaged in a fork/join computation in |
1283 |
+ |
* the current pool, this method is equivalent in effect to |
1284 |
+ |
* {@link ForkJoinTask#fork}. |
1285 |
+ |
* |
1286 |
+ |
* @param task the task to submit |
1287 |
+ |
* @return the task |
1288 |
+ |
* @throws NullPointerException if the task is null |
1289 |
+ |
* @throws RejectedExecutionException if the task cannot be |
1290 |
+ |
* scheduled for execution |
1291 |
+ |
*/ |
1292 |
+ |
public <T> ForkJoinTask<T> submit(ForkJoinTask<T> task) { |
1293 |
+ |
doSubmit(task); |
1294 |
+ |
return task; |
1295 |
+ |
} |
1296 |
+ |
|
1297 |
+ |
/** |
1298 |
|
* @throws NullPointerException if the task is null |
1299 |
|
* @throws RejectedExecutionException if the task cannot be |
1300 |
|
* scheduled for execution |
1332 |
|
} |
1333 |
|
|
1334 |
|
/** |
1332 |
– |
* Submits a ForkJoinTask for execution. |
1333 |
– |
* |
1334 |
– |
* @param task the task to submit |
1335 |
– |
* @return the task |
1336 |
– |
* @throws NullPointerException if the task is null |
1337 |
– |
* @throws RejectedExecutionException if the task cannot be |
1338 |
– |
* scheduled for execution |
1339 |
– |
*/ |
1340 |
– |
public <T> ForkJoinTask<T> submit(ForkJoinTask<T> task) { |
1341 |
– |
doSubmit(task); |
1342 |
– |
return task; |
1343 |
– |
} |
1344 |
– |
|
1345 |
– |
/** |
1335 |
|
* @throws NullPointerException {@inheritDoc} |
1336 |
|
* @throws RejectedExecutionException {@inheritDoc} |
1337 |
|
*/ |
1373 |
|
* @return the handler, or {@code null} if none |
1374 |
|
*/ |
1375 |
|
public Thread.UncaughtExceptionHandler getUncaughtExceptionHandler() { |
1387 |
– |
workerCountReadFence(); |
1376 |
|
return ueh; |
1377 |
|
} |
1378 |
|
|
1379 |
|
/** |
1392 |
– |
* Sets the handler for internal worker threads that terminate due |
1393 |
– |
* to unrecoverable errors encountered while executing tasks. |
1394 |
– |
* Unless set, the current default or ThreadGroup handler is used |
1395 |
– |
* as handler. |
1396 |
– |
* |
1397 |
– |
* @param h the new handler |
1398 |
– |
* @return the old handler, or {@code null} if none |
1399 |
– |
* @throws SecurityException if a security manager exists and |
1400 |
– |
* the caller is not permitted to modify threads |
1401 |
– |
* because it does not hold {@link |
1402 |
– |
* java.lang.RuntimePermission}{@code ("modifyThread")} |
1403 |
– |
*/ |
1404 |
– |
public Thread.UncaughtExceptionHandler |
1405 |
– |
setUncaughtExceptionHandler(Thread.UncaughtExceptionHandler h) { |
1406 |
– |
checkPermission(); |
1407 |
– |
workerCountReadFence(); |
1408 |
– |
Thread.UncaughtExceptionHandler old = ueh; |
1409 |
– |
if (h != old) { |
1410 |
– |
ueh = h; |
1411 |
– |
workerCountWriteFence(); |
1412 |
– |
for (ForkJoinWorkerThread w : workers) { |
1413 |
– |
if (w != null) |
1414 |
– |
w.setUncaughtExceptionHandler(h); |
1415 |
– |
} |
1416 |
– |
} |
1417 |
– |
return old; |
1418 |
– |
} |
1419 |
– |
|
1420 |
– |
/** |
1421 |
– |
* Sets the target parallelism level of this pool. |
1422 |
– |
* |
1423 |
– |
* @param parallelism the target parallelism |
1424 |
– |
* @throws IllegalArgumentException if parallelism less than or |
1425 |
– |
* equal to zero or greater than maximum size bounds |
1426 |
– |
* @throws SecurityException if a security manager exists and |
1427 |
– |
* the caller is not permitted to modify threads |
1428 |
– |
* because it does not hold {@link |
1429 |
– |
* java.lang.RuntimePermission}{@code ("modifyThread")} |
1430 |
– |
*/ |
1431 |
– |
public void setParallelism(int parallelism) { |
1432 |
– |
checkPermission(); |
1433 |
– |
if (parallelism <= 0 || parallelism > maxPoolSize) |
1434 |
– |
throw new IllegalArgumentException(); |
1435 |
– |
workerCountReadFence(); |
1436 |
– |
int pc = this.parallelism; |
1437 |
– |
if (pc != parallelism) { |
1438 |
– |
this.parallelism = parallelism; |
1439 |
– |
workerCountWriteFence(); |
1440 |
– |
// Release spares. If too many, some will die after re-suspend |
1441 |
– |
for (ForkJoinWorkerThread w : workers) { |
1442 |
– |
if (w != null && w.tryUnsuspend()) { |
1443 |
– |
updateRunningCount(1); |
1444 |
– |
LockSupport.unpark(w); |
1445 |
– |
} |
1446 |
– |
} |
1447 |
– |
ensureEnoughTotalWorkers(); |
1448 |
– |
advanceEventCount(); |
1449 |
– |
releaseWaiters(); // force config recheck by existing workers |
1450 |
– |
} |
1451 |
– |
} |
1452 |
– |
|
1453 |
– |
/** |
1380 |
|
* Returns the targeted parallelism level of this pool. |
1381 |
|
* |
1382 |
|
* @return the targeted parallelism level of this pool |
1383 |
|
*/ |
1384 |
|
public int getParallelism() { |
1459 |
– |
// workerCountReadFence(); // inlined below |
1460 |
– |
int ignore = workerCounts; |
1385 |
|
return parallelism; |
1386 |
|
} |
1387 |
|
|
1398 |
|
} |
1399 |
|
|
1400 |
|
/** |
1477 |
– |
* Returns the maximum number of threads allowed to exist in the |
1478 |
– |
* pool. Unless set using {@link #setMaximumPoolSize}, the |
1479 |
– |
* maximum is an implementation-defined value designed only to |
1480 |
– |
* prevent runaway growth. |
1481 |
– |
* |
1482 |
– |
* @return the maximum |
1483 |
– |
*/ |
1484 |
– |
public int getMaximumPoolSize() { |
1485 |
– |
workerCountReadFence(); |
1486 |
– |
return maxPoolSize; |
1487 |
– |
} |
1488 |
– |
|
1489 |
– |
/** |
1490 |
– |
* Sets the maximum number of threads allowed to exist in the |
1491 |
– |
* pool. The given value should normally be greater than or equal |
1492 |
– |
* to the {@link #getParallelism parallelism} level. Setting this |
1493 |
– |
* value has no effect on current pool size. It controls |
1494 |
– |
* construction of new threads. The use of this method may cause |
1495 |
– |
* tasks that intrinsically require extra threads for dependent |
1496 |
– |
* computations to indefinitely stall. If you are instead trying |
1497 |
– |
* to minimize internal thread creation, consider setting {@link |
1498 |
– |
* #setMaintainsParallelism} as false. |
1499 |
– |
* |
1500 |
– |
* @throws IllegalArgumentException if negative or greater than |
1501 |
– |
* internal implementation limit |
1502 |
– |
*/ |
1503 |
– |
public void setMaximumPoolSize(int newMax) { |
1504 |
– |
if (newMax < 0 || newMax > MAX_THREADS) |
1505 |
– |
throw new IllegalArgumentException(); |
1506 |
– |
maxPoolSize = newMax; |
1507 |
– |
workerCountWriteFence(); |
1508 |
– |
} |
1509 |
– |
|
1510 |
– |
/** |
1511 |
– |
* Returns {@code true} if this pool dynamically maintains its |
1512 |
– |
* target parallelism level. If false, new threads are added only |
1513 |
– |
* to avoid possible starvation. This setting is by default true. |
1514 |
– |
* |
1515 |
– |
* @return {@code true} if maintains parallelism |
1516 |
– |
*/ |
1517 |
– |
public boolean getMaintainsParallelism() { |
1518 |
– |
workerCountReadFence(); |
1519 |
– |
return maintainsParallelism; |
1520 |
– |
} |
1521 |
– |
|
1522 |
– |
/** |
1523 |
– |
* Sets whether this pool dynamically maintains its target |
1524 |
– |
* parallelism level. If false, new threads are added only to |
1525 |
– |
* avoid possible starvation. |
1526 |
– |
* |
1527 |
– |
* @param enable {@code true} to maintain parallelism |
1528 |
– |
*/ |
1529 |
– |
public void setMaintainsParallelism(boolean enable) { |
1530 |
– |
maintainsParallelism = enable; |
1531 |
– |
workerCountWriteFence(); |
1532 |
– |
} |
1533 |
– |
|
1534 |
– |
/** |
1535 |
– |
* Establishes local first-in-first-out scheduling mode for forked |
1536 |
– |
* tasks that are never joined. This mode may be more appropriate |
1537 |
– |
* than default locally stack-based mode in applications in which |
1538 |
– |
* worker threads only process asynchronous tasks. This method is |
1539 |
– |
* designed to be invoked only when the pool is quiescent, and |
1540 |
– |
* typically only before any tasks are submitted. The effects of |
1541 |
– |
* invocations at other times may be unpredictable. |
1542 |
– |
* |
1543 |
– |
* @param async if {@code true}, use locally FIFO scheduling |
1544 |
– |
* @return the previous mode |
1545 |
– |
* @see #getAsyncMode |
1546 |
– |
*/ |
1547 |
– |
public boolean setAsyncMode(boolean async) { |
1548 |
– |
workerCountReadFence(); |
1549 |
– |
boolean oldMode = locallyFifo; |
1550 |
– |
if (oldMode != async) { |
1551 |
– |
locallyFifo = async; |
1552 |
– |
workerCountWriteFence(); |
1553 |
– |
for (ForkJoinWorkerThread w : workers) { |
1554 |
– |
if (w != null) |
1555 |
– |
w.setAsyncMode(async); |
1556 |
– |
} |
1557 |
– |
} |
1558 |
– |
return oldMode; |
1559 |
– |
} |
1560 |
– |
|
1561 |
– |
/** |
1401 |
|
* Returns {@code true} if this pool uses local first-in-first-out |
1402 |
|
* scheduling mode for forked tasks that are never joined. |
1403 |
|
* |
1404 |
|
* @return {@code true} if this pool uses async mode |
1566 |
– |
* @see #setAsyncMode |
1405 |
|
*/ |
1406 |
|
public boolean getAsyncMode() { |
1569 |
– |
workerCountReadFence(); |
1407 |
|
return locallyFifo; |
1408 |
|
} |
1409 |
|
|
1472 |
|
*/ |
1473 |
|
public long getQueuedTaskCount() { |
1474 |
|
long count = 0; |
1475 |
< |
for (ForkJoinWorkerThread w : workers) { |
1475 |
> |
ForkJoinWorkerThread[] ws = workers; |
1476 |
> |
int nws = ws.length; |
1477 |
> |
for (int i = 0; i < nws; ++i) { |
1478 |
> |
ForkJoinWorkerThread w = ws[i]; |
1479 |
|
if (w != null) |
1480 |
|
count += w.getQueueSize(); |
1481 |
|
} |
1533 |
|
*/ |
1534 |
|
protected int drainTasksTo(Collection<? super ForkJoinTask<?>> c) { |
1535 |
|
int n = submissionQueue.drainTo(c); |
1536 |
< |
for (ForkJoinWorkerThread w : workers) { |
1536 |
> |
ForkJoinWorkerThread[] ws = workers; |
1537 |
> |
int nws = ws.length; |
1538 |
> |
for (int i = 0; i < nws; ++i) { |
1539 |
> |
ForkJoinWorkerThread w = ws[i]; |
1540 |
|
if (w != null) |
1541 |
|
n += w.drainTasksTo(c); |
1542 |
|
} |
1544 |
|
} |
1545 |
|
|
1546 |
|
/** |
1547 |
+ |
* Returns count of total parks by existing workers. |
1548 |
+ |
* Used during development only since not meaningful to users. |
1549 |
+ |
*/ |
1550 |
+ |
private int collectParkCount() { |
1551 |
+ |
int count = 0; |
1552 |
+ |
ForkJoinWorkerThread[] ws = workers; |
1553 |
+ |
int nws = ws.length; |
1554 |
+ |
for (int i = 0; i < nws; ++i) { |
1555 |
+ |
ForkJoinWorkerThread w = ws[i]; |
1556 |
+ |
if (w != null) |
1557 |
+ |
count += w.parkCount; |
1558 |
+ |
} |
1559 |
+ |
return count; |
1560 |
+ |
} |
1561 |
+ |
|
1562 |
+ |
/** |
1563 |
|
* Returns a string identifying this pool, as well as its state, |
1564 |
|
* including indications of run state, parallelism level, and |
1565 |
|
* worker and task counts. |
1576 |
|
int pc = parallelism; |
1577 |
|
int rs = runState; |
1578 |
|
int ac = rs & ACTIVE_COUNT_MASK; |
1579 |
+ |
// int pk = collectParkCount(); |
1580 |
|
return super.toString() + |
1581 |
|
"[" + runLevelToString(rs) + |
1582 |
|
", parallelism = " + pc + |
1586 |
|
", steals = " + st + |
1587 |
|
", tasks = " + qt + |
1588 |
|
", submissions = " + qs + |
1589 |
+ |
// ", parks = " + pk + |
1590 |
|
"]"; |
1591 |
|
} |
1592 |
|
|
1682 |
|
*/ |
1683 |
|
public boolean awaitTermination(long timeout, TimeUnit unit) |
1684 |
|
throws InterruptedException { |
1685 |
< |
return terminationLatch.await(timeout, unit); |
1685 |
> |
try { |
1686 |
> |
return termination.awaitAdvanceInterruptibly(0, timeout, unit) > 0; |
1687 |
> |
} catch(TimeoutException ex) { |
1688 |
> |
return false; |
1689 |
> |
} |
1690 |
|
} |
1691 |
|
|
1692 |
|
/** |
1738 |
|
* Blocks in accord with the given blocker. If the current thread |
1739 |
|
* is a {@link ForkJoinWorkerThread}, this method possibly |
1740 |
|
* arranges for a spare thread to be activated if necessary to |
1741 |
< |
* ensure parallelism while the current thread is blocked. |
1877 |
< |
* |
1878 |
< |
* <p>If {@code maintainParallelism} is {@code true} and the pool |
1879 |
< |
* supports it ({@link #getMaintainsParallelism}), this method |
1880 |
< |
* attempts to maintain the pool's nominal parallelism. Otherwise |
1881 |
< |
* it activates a thread only if necessary to avoid complete |
1882 |
< |
* starvation. This option may be preferable when blockages use |
1883 |
< |
* timeouts, or are almost always brief. |
1741 |
> |
* ensure sufficient parallelism while the current thread is blocked. |
1742 |
|
* |
1743 |
|
* <p>If the caller is not a {@link ForkJoinTask}, this method is |
1744 |
|
* behaviorally equivalent to |
1752 |
|
* first be expanded to ensure parallelism, and later adjusted. |
1753 |
|
* |
1754 |
|
* @param blocker the blocker |
1897 |
– |
* @param maintainParallelism if {@code true} and supported by |
1898 |
– |
* this pool, attempt to maintain the pool's nominal parallelism; |
1899 |
– |
* otherwise activate a thread only if necessary to avoid |
1900 |
– |
* complete starvation. |
1755 |
|
* @throws InterruptedException if blocker.block did so |
1756 |
|
*/ |
1757 |
< |
public static void managedBlock(ManagedBlocker blocker, |
1904 |
< |
boolean maintainParallelism) |
1757 |
> |
public static void managedBlock(ManagedBlocker blocker) |
1758 |
|
throws InterruptedException { |
1759 |
|
Thread t = Thread.currentThread(); |
1760 |
|
if (t instanceof ForkJoinWorkerThread) |
1761 |
< |
((ForkJoinWorkerThread) t).pool. |
1762 |
< |
doBlock(blocker, maintainParallelism); |
1763 |
< |
else |
1764 |
< |
awaitBlocker(blocker); |
1912 |
< |
} |
1913 |
< |
|
1914 |
< |
/** |
1915 |
< |
* Performs Non-FJ blocking |
1916 |
< |
*/ |
1917 |
< |
private static void awaitBlocker(ManagedBlocker blocker) |
1918 |
< |
throws InterruptedException { |
1919 |
< |
do {} while (!blocker.isReleasable() && !blocker.block()); |
1761 |
> |
((ForkJoinWorkerThread) t).pool.awaitBlocker(blocker); |
1762 |
> |
else { |
1763 |
> |
do {} while (!blocker.isReleasable() && !blocker.block()); |
1764 |
> |
} |
1765 |
|
} |
1766 |
|
|
1767 |
|
// AbstractExecutorService overrides. These rely on undocumented |
1790 |
|
private static final long stealCountOffset = |
1791 |
|
objectFieldOffset("stealCount",ForkJoinPool.class); |
1792 |
|
|
1948 |
– |
|
1793 |
|
private static long objectFieldOffset(String field, Class<?> klazz) { |
1794 |
|
try { |
1795 |
|
return UNSAFE.objectFieldOffset(klazz.getDeclaredField(field)); |