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

Comparing jsr166/src/jsr166y/ForkJoinPool.java (file contents):
Revision 1.123 by dl, Mon Feb 20 18:20:06 2012 UTC vs.
Revision 1.136 by dl, Mon Oct 29 17:23:34 2012 UTC

# Line 5 | Line 5
5   */
6  
7   package jsr166y;
8 +
9   import java.util.ArrayList;
10   import java.util.Arrays;
11   import java.util.Collection;
# Line 41 | Line 42 | import java.util.concurrent.locks.Condit
42   * ForkJoinPool}s may also be appropriate for use with event-style
43   * tasks that are never joined.
44   *
45 < * <p>A {@code ForkJoinPool} is constructed with a given target
46 < * parallelism level; by default, equal to the number of available
47 < * processors. The pool attempts to maintain enough active (or
48 < * available) threads by dynamically adding, suspending, or resuming
49 < * internal worker threads, even if some tasks are stalled waiting to
50 < * join others. However, no such adjustments are guaranteed in the
51 < * face of blocked IO or other unmanaged synchronization. The nested
52 < * {@link ManagedBlocker} interface enables extension of the kinds of
45 > * <p>A static {@link #commonPool} is available and appropriate for
46 > * most applications. The common pool is used by any ForkJoinTask that
47 > * is not explicitly submitted to a specified pool. Using the common
48 > * pool normally reduces resource usage (its threads are slowly
49 > * reclaimed during periods of non-use, and reinstated upon subsequent
50 > * use).  The common pool is by default constructed with default
51 > * parameters, but these may be controlled by setting any or all of
52 > * the three properties {@code
53 > * java.util.concurrent.ForkJoinPool.common.{parallelism,
54 > * threadFactory, exceptionHandler}}.
55 > *
56 > * <p>For applications that require separate or custom pools, a {@code
57 > * ForkJoinPool} may be constructed with a given target parallelism
58 > * level; by default, equal to the number of available processors. The
59 > * pool attempts to maintain enough active (or available) threads by
60 > * dynamically adding, suspending, or resuming internal worker
61 > * threads, even if some tasks are stalled waiting to join
62 > * others. However, no such adjustments are guaranteed in the face of
63 > * blocked IO or other unmanaged synchronization. The nested {@link
64 > * ManagedBlocker} interface enables extension of the kinds of
65   * synchronization accommodated.
66   *
67   * <p>In addition to execution and lifecycle control methods, this
# Line 93 | Line 106 | import java.util.concurrent.locks.Condit
106   *  </tr>
107   * </table>
108   *
96 * <p><b>Sample Usage.</b> Normally a single {@code ForkJoinPool} is
97 * used for all parallel task execution in a program or subsystem.
98 * Otherwise, use would not usually outweigh the construction and
99 * bookkeeping overhead of creating a large set of threads. For
100 * example, a common pool could be used for the {@code SortTasks}
101 * illustrated in {@link RecursiveAction}. Because {@code
102 * ForkJoinPool} uses threads in {@linkplain java.lang.Thread#isDaemon
103 * daemon} mode, there is typically no need to explicitly {@link
104 * #shutdown} such a pool upon program exit.
105 *
106 *  <pre> {@code
107 * static final ForkJoinPool mainPool = new ForkJoinPool();
108 * ...
109 * public void sort(long[] array) {
110 *   mainPool.invoke(new SortTask(array, 0, array.length));
111 * }}</pre>
112 *
109   * <p><b>Implementation notes</b>: This implementation restricts the
110   * maximum number of running threads to 32767. Attempts to create
111   * pools with greater than the maximum number result in
# Line 239 | Line 235 | public class ForkJoinPool extends Abstra
235       * when locked remains available to check consistency.
236       *
237       * Recording WorkQueues.  WorkQueues are recorded in the
238 <     * "workQueues" array that is created upon pool construction and
239 <     * expanded if necessary.  Updates to the array while recording
240 <     * new workers and unrecording terminated ones are protected from
241 <     * each other by a lock but the array is otherwise concurrently
242 <     * readable, and accessed directly.  To simplify index-based
243 <     * operations, the array size is always a power of two, and all
244 <     * readers must tolerate null slots. Shared (submission) queues
245 <     * are at even indices, worker queues at odd indices. Grouping
246 <     * them together in this way simplifies and speeds up task
251 <     * scanning.
238 >     * "workQueues" array that is created upon first use and expanded
239 >     * if necessary.  Updates to the array while recording new workers
240 >     * and unrecording terminated ones are protected from each other
241 >     * by a lock but the array is otherwise concurrently readable, and
242 >     * accessed directly.  To simplify index-based operations, the
243 >     * array size is always a power of two, and all readers must
244 >     * tolerate null slots. Shared (submission) queues are at even
245 >     * indices, worker queues at odd indices. Grouping them together
246 >     * in this way simplifies and speeds up task scanning.
247       *
248       * All worker thread creation is on-demand, triggered by task
249       * submissions, replacement of terminated workers, and/or
# Line 320 | Line 315 | public class ForkJoinPool extends Abstra
315       *
316       * Trimming workers. To release resources after periods of lack of
317       * use, a worker starting to wait when the pool is quiescent will
318 <     * time out and terminate if the pool has remained quiescent for
319 <     * SHRINK_RATE nanosecs. This will slowly propagate, eventually
320 <     * terminating all workers after long periods of non-use.
318 >     * time out and terminate if the pool has remained quiescent for a
319 >     * given period -- a short period if there are more threads than
320 >     * parallelism, longer as the number of threads decreases. This
321 >     * will slowly propagate, eventually terminating all workers after
322 >     * periods of non-use.
323       *
324       * Shutdown and Termination. A call to shutdownNow atomically sets
325       * a runState bit and then (non-atomically) sets each worker's
# Line 413 | Line 410 | public class ForkJoinPool extends Abstra
410       * unblocked threads to the point that we know they are available)
411       * leading to more situations requiring more threads, and so
412       * on. This aspect of control can be seen as an (analytically
413 <     * intractible) game with an opponent that may choose the worst
413 >     * intractable) game with an opponent that may choose the worst
414       * (for us) active thread to stall at any time.  We take several
415       * precautions to bound losses (and thus bound gains), mainly in
416       * methods tryCompensate and awaitJoin: (1) We only try
# Line 504 | Line 501 | public class ForkJoinPool extends Abstra
501      }
502  
503      /**
507     * A simple non-reentrant lock used for exclusion when managing
508     * queues and workers. We use a custom lock so that we can readily
509     * probe lock state in constructions that check among alternative
510     * actions. The lock is normally only very briefly held, and
511     * sometimes treated as a spinlock, but other usages block to
512     * reduce overall contention in those cases where locked code
513     * bodies perform allocation/resizing.
514     */
515    static final class Mutex extends AbstractQueuedSynchronizer {
516        public final boolean tryAcquire(int ignore) {
517            return compareAndSetState(0, 1);
518        }
519        public final boolean tryRelease(int ignore) {
520            setState(0);
521            return true;
522        }
523        public final void lock() { acquire(0); }
524        public final void unlock() { release(0); }
525        public final boolean isHeldExclusively() { return getState() == 1; }
526        public final Condition newCondition() { return new ConditionObject(); }
527    }
528
529    /**
504       * Class for artificial tasks that are used to replace the target
505       * of local joins if they are removed from an interior queue slot
506       * in WorkQueue.tryRemoveAndExec. We don't need the proxy to
# Line 629 | Line 603 | public class ForkJoinPool extends Abstra
603          final ForkJoinPool pool;   // the containing pool (may be null)
604          final ForkJoinWorkerThread owner; // owning thread or null if shared
605          volatile Thread parker;    // == owner during call to park; else null
606 <        ForkJoinTask<?> currentJoin;  // task being joined in awaitJoin
606 >        volatile ForkJoinTask<?> currentJoin;  // task being joined in awaitJoin
607          ForkJoinTask<?> currentSteal; // current non-local task being executed
608          // Heuristic padding to ameliorate unfortunate memory placements
609          Object p00, p01, p02, p03, p04, p05, p06, p07;
# Line 721 | Line 695 | public class ForkJoinPool extends Abstra
695           * version of this method because it is never needed.)
696           */
697          final ForkJoinTask<?> pop() {
698 <            ForkJoinTask<?> t; int m;
699 <            ForkJoinTask<?>[] a = array;
726 <            if (a != null && (m = a.length - 1) >= 0) {
698 >            ForkJoinTask<?>[] a; ForkJoinTask<?> t; int m;
699 >            if ((a = array) != null && (m = a.length - 1) >= 0) {
700                  for (int s; (s = top - 1) - base >= 0;) {
701 <                    int j = ((m & s) << ASHIFT) + ABASE;
702 <                    if ((t = (ForkJoinTask<?>)U.getObjectVolatile(a, j)) == null)
701 >                    long j = ((m & s) << ASHIFT) + ABASE;
702 >                    if ((t = (ForkJoinTask<?>)U.getObject(a, j)) == null)
703                          break;
704                      if (U.compareAndSwapObject(a, j, t, null)) {
705                          top = s;
# Line 814 | Line 787 | public class ForkJoinPool extends Abstra
787          }
788  
789          /**
790 +         * Version of tryUnpush for shared queues; called by non-FJ
791 +         * submitters after prechecking that task probably exists.
792 +         */
793 +        final boolean trySharedUnpush(ForkJoinTask<?> t) {
794 +            boolean success = false;
795 +            if (runState == 0 && U.compareAndSwapInt(this, RUNSTATE, 0, 1)) {
796 +                try {
797 +                    ForkJoinTask<?>[] a; int s;
798 +                    if ((a = array) != null && (s = top) != base &&
799 +                        U.compareAndSwapObject
800 +                        (a, (((a.length - 1) & --s) << ASHIFT) + ABASE, t, null)) {
801 +                        top = s;
802 +                        success = true;
803 +                    }
804 +                } finally {
805 +                    runState = 0;                         // unlock
806 +                }
807 +            }
808 +            return success;
809 +        }
810 +
811 +        /**
812           * Polls the given task only if it is at the current base.
813           */
814          final boolean pollFor(ForkJoinTask<?> task) {
# Line 830 | Line 825 | public class ForkJoinPool extends Abstra
825          }
826  
827          /**
833         * If present, removes from queue and executes the given task, or
834         * any other cancelled task. Returns (true) immediately on any CAS
835         * or consistency check failure so caller can retry.
836         *
837         * @return false if no progress can be made
838         */
839        final boolean tryRemoveAndExec(ForkJoinTask<?> task) {
840            boolean removed = false, empty = true, progress = true;
841            ForkJoinTask<?>[] a; int m, s, b, n;
842            if ((a = array) != null && (m = a.length - 1) >= 0 &&
843                (n = (s = top) - (b = base)) > 0) {
844                for (ForkJoinTask<?> t;;) {           // traverse from s to b
845                    int j = ((--s & m) << ASHIFT) + ABASE;
846                    t = (ForkJoinTask<?>)U.getObjectVolatile(a, j);
847                    if (t == null)                    // inconsistent length
848                        break;
849                    else if (t == task) {
850                        if (s + 1 == top) {           // pop
851                            if (!U.compareAndSwapObject(a, j, task, null))
852                                break;
853                            top = s;
854                            removed = true;
855                        }
856                        else if (base == b)           // replace with proxy
857                            removed = U.compareAndSwapObject(a, j, task,
858                                                             new EmptyTask());
859                        break;
860                    }
861                    else if (t.status >= 0)
862                        empty = false;
863                    else if (s + 1 == top) {          // pop and throw away
864                        if (U.compareAndSwapObject(a, j, t, null))
865                            top = s;
866                        break;
867                    }
868                    if (--n == 0) {
869                        if (!empty && base == b)
870                            progress = false;
871                        break;
872                    }
873                }
874            }
875            if (removed)
876                task.doExec();
877            return progress;
878        }
879
880        /**
828           * Initializes or doubles the capacity of array. Call either
829           * by owner or with lock held -- it is OK for base, but not
830           * top, to move while resizings are in progress.
# Line 939 | Line 886 | public class ForkJoinPool extends Abstra
886          // Execution methods
887  
888          /**
889 <         * Removes and runs tasks until empty, using local mode
890 <         * ordering. Normally called only after checking for apparent
891 <         * non-emptiness.
892 <         */
893 <        final void runLocalTasks() {
894 <            // hoist checks from repeated pop/poll
895 <            ForkJoinTask<?>[] a; int m;
896 <            if ((a = array) != null && (m = a.length - 1) >= 0) {
897 <                if (mode == 0) {
898 <                    for (int s; (s = top - 1) - base >= 0;) {
899 <                        int j = ((m & s) << ASHIFT) + ABASE;
900 <                        ForkJoinTask<?> t =
901 <                            (ForkJoinTask<?>)U.getObjectVolatile(a, j);
955 <                        if (t != null) {
956 <                            if (U.compareAndSwapObject(a, j, t, null)) {
957 <                                top = s;
958 <                                t.doExec();
959 <                            }
960 <                        }
961 <                        else
962 <                            break;
963 <                    }
889 >         * Pops and runs tasks until empty.
890 >         */
891 >        private void popAndExecAll() {
892 >            // A bit faster than repeated pop calls
893 >            ForkJoinTask<?>[] a; int m, s; long j; ForkJoinTask<?> t;
894 >            while ((a = array) != null && (m = a.length - 1) >= 0 &&
895 >                   (s = top - 1) - base >= 0 &&
896 >                   (t = ((ForkJoinTask<?>)
897 >                         U.getObject(a, j = ((m & s) << ASHIFT) + ABASE)))
898 >                   != null) {
899 >                if (U.compareAndSwapObject(a, j, t, null)) {
900 >                    top = s;
901 >                    t.doExec();
902                  }
903 <                else {
904 <                    for (int b; (b = base) - top < 0;) {
905 <                        int j = ((m & b) << ASHIFT) + ABASE;
906 <                        ForkJoinTask<?> t =
907 <                            (ForkJoinTask<?>)U.getObjectVolatile(a, j);
908 <                        if (t != null) {
909 <                            if (base == b &&
910 <                                U.compareAndSwapObject(a, j, t, null)) {
911 <                                base = b + 1;
912 <                                t.doExec();
913 <                            }
914 <                        } else if (base == b) {
915 <                            if (b + 1 == top)
903 >            }
904 >        }
905 >
906 >        /**
907 >         * Polls and runs tasks until empty.
908 >         */
909 >        private void pollAndExecAll() {
910 >            for (ForkJoinTask<?> t; (t = poll()) != null;)
911 >                t.doExec();
912 >        }
913 >
914 >        /**
915 >         * If present, removes from queue and executes the given task, or
916 >         * any other cancelled task. Returns (true) immediately on any CAS
917 >         * or consistency check failure so caller can retry.
918 >         *
919 >         * @return 0 if no progress can be made, else positive
920 >         * (this unusual convention simplifies use with tryHelpStealer.)
921 >         */
922 >        final int tryRemoveAndExec(ForkJoinTask<?> task) {
923 >            int stat = 1;
924 >            boolean removed = false, empty = true;
925 >            ForkJoinTask<?>[] a; int m, s, b, n;
926 >            if ((a = array) != null && (m = a.length - 1) >= 0 &&
927 >                (n = (s = top) - (b = base)) > 0) {
928 >                for (ForkJoinTask<?> t;;) {           // traverse from s to b
929 >                    int j = ((--s & m) << ASHIFT) + ABASE;
930 >                    t = (ForkJoinTask<?>)U.getObjectVolatile(a, j);
931 >                    if (t == null)                    // inconsistent length
932 >                        break;
933 >                    else if (t == task) {
934 >                        if (s + 1 == top) {           // pop
935 >                            if (!U.compareAndSwapObject(a, j, task, null))
936                                  break;
937 <                            Thread.yield(); // wait for lagging update
937 >                            top = s;
938 >                            removed = true;
939                          }
940 +                        else if (base == b)           // replace with proxy
941 +                            removed = U.compareAndSwapObject(a, j, task,
942 +                                                             new EmptyTask());
943 +                        break;
944 +                    }
945 +                    else if (t.status >= 0)
946 +                        empty = false;
947 +                    else if (s + 1 == top) {          // pop and throw away
948 +                        if (U.compareAndSwapObject(a, j, t, null))
949 +                            top = s;
950 +                        break;
951 +                    }
952 +                    if (--n == 0) {
953 +                        if (!empty && base == b)
954 +                            stat = 0;
955 +                        break;
956                      }
957                  }
958              }
959 +            if (removed)
960 +                task.doExec();
961 +            return stat;
962          }
963  
964          /**
965           * Executes a top-level task and any local tasks remaining
966           * after execution.
989         *
990         * @return true unless terminating
967           */
968 <        final boolean runTask(ForkJoinTask<?> t) {
993 <            boolean alive = true;
968 >        final void runTask(ForkJoinTask<?> t) {
969              if (t != null) {
970                  currentSteal = t;
971                  t.doExec();
972 <                if (top != base)        // conservative guard
973 <                    runLocalTasks();
972 >                if (top != base) {       // process remaining local tasks
973 >                    if (mode == 0)
974 >                        popAndExecAll();
975 >                    else
976 >                        pollAndExecAll();
977 >                }
978                  ++nsteals;
979                  currentSteal = null;
980              }
1002            else if (runState < 0)      // terminating
1003                alive = false;
1004            return alive;
981          }
982  
983          /**
# Line 1106 | Line 1082 | public class ForkJoinPool extends Abstra
1082      public static final ForkJoinWorkerThreadFactory
1083          defaultForkJoinWorkerThreadFactory;
1084  
1085 +
1086 +    /** Property prefix for constructing common pool */
1087 +    private static final String propPrefix =
1088 +        "java.util.concurrent.ForkJoinPool.common.";
1089 +
1090 +    /**
1091 +     * Common (static) pool. Non-null for public use unless a static
1092 +     * construction exception, but internal usages must null-check on
1093 +     * use.
1094 +     */
1095 +    static final ForkJoinPool commonPool;
1096 +
1097 +    /**
1098 +     * Common pool parallelism. Must equal commonPool.parallelism.
1099 +     */
1100 +    static final int commonPoolParallelism;
1101 +
1102      /**
1103       * Generator for assigning sequence numbers as pool names.
1104       */
# Line 1124 | Line 1117 | public class ForkJoinPool extends Abstra
1117      private static final RuntimePermission modifyThreadPermission;
1118  
1119      /**
1120 <     * Per-thread submission bookeeping. Shared across all pools
1120 >     * Per-thread submission bookkeeping. Shared across all pools
1121       * to reduce ThreadLocal pollution and because random motion
1122       * to avoid contention in one pool is likely to hold for others.
1123       */
# Line 1133 | Line 1126 | public class ForkJoinPool extends Abstra
1126      // static constants
1127  
1128      /**
1129 <     * The wakeup interval (in nanoseconds) for a worker waiting for a
1130 <     * task when the pool is quiescent to instead try to shrink the
1131 <     * number of workers.  The exact value does not matter too
1139 <     * much. It must be short enough to release resources during
1140 <     * sustained periods of idleness, but not so short that threads
1141 <     * are continually re-created.
1129 >     * Initial timeout value (in nanoseconds) for the thread triggering
1130 >     * quiescence to park waiting for new work. On timeout, the thread
1131 >     * will instead try to shrink the number of workers.
1132       */
1133 <    private static final long SHRINK_RATE =
1144 <        4L * 1000L * 1000L * 1000L; // 4 seconds
1133 >    private static final long IDLE_TIMEOUT      = 1000L * 1000L * 1000L; // 1sec
1134  
1135      /**
1136 <     * The timeout value for attempted shrinkage, includes
1148 <     * some slop to cope with system timer imprecision.
1136 >     * Timeout value when there are more threads than parallelism level
1137       */
1138 <    private static final long SHRINK_TIMEOUT = SHRINK_RATE - (SHRINK_RATE / 10);
1138 >    private static final long FAST_IDLE_TIMEOUT =  100L * 1000L * 1000L;
1139  
1140      /**
1141       * The maximum stolen->joining link depth allowed in method
# Line 1160 | Line 1148 | public class ForkJoinPool extends Abstra
1148       * traversal parameters at the expense of sometimes blocking when
1149       * we could be helping.
1150       */
1151 <    private static final int MAX_HELP = 32;
1151 >    private static final int MAX_HELP = 64;
1152  
1153      /**
1154       * Secondary time-based bound (in nanosecs) for helping attempts
# Line 1170 | Line 1158 | public class ForkJoinPool extends Abstra
1158       * value should roughly approximate the time required to create
1159       * and/or activate a worker thread.
1160       */
1161 <    private static final long COMPENSATION_DELAY = 100L * 1000L; // 0.1 millisec
1161 >    private static final long COMPENSATION_DELAY = 1L << 18; // ~0.25 millisec
1162  
1163      /**
1164       * Increment for seed generators. See class ThreadLocal for
# Line 1267 | Line 1255 | public class ForkJoinPool extends Abstra
1255       * empirically works OK on current JVMs.
1256       */
1257  
1258 +    volatile long stealCount;                  // collects worker counts
1259      volatile long ctl;                         // main pool control
1260      final int parallelism;                     // parallelism level
1261      final int localMode;                       // per-worker scheduling mode
1262 +    volatile int nextWorkerNumber;             // to create worker name string
1263      final int submitMask;                      // submit queue index bound
1264      int nextSeed;                              // for initializing worker seeds
1265 +    volatile int mainLock;                     // spinlock for array updates
1266      volatile int runState;                     // shutdown status and seq
1267      WorkQueue[] workQueues;                    // main registry
1277    final Mutex lock;                          // for registration
1278    final Condition termination;               // for awaitTermination
1268      final ForkJoinWorkerThreadFactory factory; // factory for new workers
1269      final Thread.UncaughtExceptionHandler ueh; // per-worker UEH
1281    final AtomicLong stealCount;               // collect counts when terminated
1282    final AtomicInteger nextWorkerNumber;      // to create worker name string
1270      final String workerNamePrefix;             // to create worker name string
1271  
1272 +    /*
1273 +     * Mechanics for main lock protecting worker array updates.  Uses
1274 +     * the same strategy as ConcurrentHashMap bins -- a spinLock for
1275 +     * normal cases, but falling back to builtin lock when (rarely)
1276 +     * needed.  See internal ConcurrentHashMap documentation for
1277 +     * explanation.
1278 +     */
1279 +
1280 +    static final int LOCK_WAITING = 2; // bit to indicate need for signal
1281 +    static final int MAX_LOCK_SPINS = 1 << 8;
1282 +
1283 +    private void tryAwaitMainLock() {
1284 +        int spins = MAX_LOCK_SPINS, r = 0, h;
1285 +        while (((h = mainLock) & 1) != 0) {
1286 +            if (r == 0)
1287 +                r = ThreadLocalRandom.current().nextInt(); // randomize spins
1288 +            else if (spins >= 0) {
1289 +                r ^= r << 1; r ^= r >>> 3; r ^= r << 10; // xorshift
1290 +                if (r >= 0)
1291 +                    --spins;
1292 +            }
1293 +            else if (U.compareAndSwapInt(this, MAINLOCK, h, h | LOCK_WAITING)) {
1294 +                synchronized (this) {
1295 +                    if ((mainLock & LOCK_WAITING) != 0) {
1296 +                        try {
1297 +                            wait();
1298 +                        } catch (InterruptedException ie) {
1299 +                            Thread.currentThread().interrupt();
1300 +                        }
1301 +                    }
1302 +                    else
1303 +                        notifyAll(); // possibly won race vs signaller
1304 +                }
1305 +                break;
1306 +            }
1307 +        }
1308 +    }
1309 +
1310      //  Creating, registering, and deregistering workers
1311  
1312      /**
# Line 1308 | Line 1333 | public class ForkJoinPool extends Abstra
1333       * ForkJoinWorkerThread.
1334       */
1335      final String nextWorkerName() {
1336 <        return workerNamePrefix.concat
1337 <            (Integer.toString(nextWorkerNumber.addAndGet(1)));
1336 >        int n;
1337 >        do {} while(!U.compareAndSwapInt(this, NEXTWORKERNUMBER,
1338 >                                         n = nextWorkerNumber, ++n));
1339 >        return workerNamePrefix.concat(Integer.toString(n));
1340      }
1341  
1342      /**
# Line 1322 | Line 1349 | public class ForkJoinPool extends Abstra
1349       * @param w the worker's queue
1350       */
1351      final void registerWorker(WorkQueue w) {
1352 <        Mutex lock = this.lock;
1353 <        lock.lock();
1352 >        while (!U.compareAndSwapInt(this, MAINLOCK, 0, 1))
1353 >            tryAwaitMainLock();
1354          try {
1355 <            WorkQueue[] ws = workQueues;
1356 <            if (w != null && ws != null) {          // skip on shutdown/failure
1357 <                int rs, n;
1358 <                while ((n = ws.length) <            // ensure can hold total
1359 <                       (parallelism + (short)(ctl >>> TC_SHIFT) << 1))
1333 <                    workQueues = ws = Arrays.copyOf(ws, n << 1);
1334 <                int m = n - 1;
1355 >            WorkQueue[] ws;
1356 >            if ((ws = workQueues) == null)
1357 >                ws = workQueues = new WorkQueue[submitMask + 1];
1358 >            if (w != null) {
1359 >                int rs, n =  ws.length, m = n - 1;
1360                  int s = nextSeed += SEED_INCREMENT; // rarely-colliding sequence
1361                  w.seed = (s == 0) ? 1 : s;          // ensure non-zero seed
1362                  int r = (s << 1) | 1;               // use odd-numbered indices
1363 <                while (ws[r &= m] != null)          // step by approx half size
1364 <                    r += ((n >>> 1) & SQMASK) + 2;
1363 >                if (ws[r &= m] != null) {           // collision
1364 >                    int probes = 0;                 // step by approx half size
1365 >                    int step = (n <= 4) ? 2 : ((n >>> 1) & SQMASK) + 2;
1366 >                    while (ws[r = (r + step) & m] != null) {
1367 >                        if (++probes >= n) {
1368 >                            workQueues = ws = Arrays.copyOf(ws, n <<= 1);
1369 >                            m = n - 1;
1370 >                            probes = 0;
1371 >                        }
1372 >                    }
1373 >                }
1374                  w.eventCount = w.poolIndex = r;     // establish before recording
1375                  ws[r] = w;                          // also update seq
1376                  runState = ((rs = runState) & SHUTDOWN) | ((rs + 2) & ~SHUTDOWN);
1377              }
1378          } finally {
1379 <            lock.unlock();
1379 >            if (!U.compareAndSwapInt(this, MAINLOCK, 1, 0)) {
1380 >                mainLock = 0;
1381 >                synchronized (this) { notifyAll(); };
1382 >            }
1383          }
1384 +
1385      }
1386  
1387      /**
# Line 1356 | Line 1394 | public class ForkJoinPool extends Abstra
1394       * @param ex the exception causing failure, or null if none
1395       */
1396      final void deregisterWorker(ForkJoinWorkerThread wt, Throwable ex) {
1359        Mutex lock = this.lock;
1397          WorkQueue w = null;
1398          if (wt != null && (w = wt.workQueue) != null) {
1399              w.runState = -1;                // ensure runState is set
1400 <            stealCount.getAndAdd(w.totalSteals + w.nsteals);
1400 >            long steals = w.totalSteals + w.nsteals, sc;
1401 >            do {} while(!U.compareAndSwapLong(this, STEALCOUNT,
1402 >                                              sc = stealCount, sc + steals));
1403              int idx = w.poolIndex;
1404 <            lock.lock();
1405 <            try {                           // remove record from array
1404 >            while (!U.compareAndSwapInt(this, MAINLOCK, 0, 1))
1405 >                tryAwaitMainLock();
1406 >            try {
1407                  WorkQueue[] ws = workQueues;
1408                  if (ws != null && idx >= 0 && idx < ws.length && ws[idx] == w)
1409                      ws[idx] = null;
1410              } finally {
1411 <                lock.unlock();
1411 >                if (!U.compareAndSwapInt(this, MAINLOCK, 1, 0)) {
1412 >                    mainLock = 0;
1413 >                    synchronized (this) { notifyAll(); };
1414 >                }
1415              }
1416          }
1417  
# Line 1390 | Line 1433 | public class ForkJoinPool extends Abstra
1433              U.throwException(ex);
1434      }
1435  
1393
1436      // Submissions
1437  
1438      /**
# Line 1399 | Line 1441 | public class ForkJoinPool extends Abstra
1441       * range). If no queue exists at the index, one is created.  If
1442       * the queue is busy, another index is randomly chosen. The
1443       * submitMask bounds the effective number of queues to the
1444 <     * (nearest poswer of two for) parallelism level.
1444 >     * (nearest power of two for) parallelism level.
1445       *
1446       * @param task the task. Caller must ensure non-null.
1447       */
# Line 1408 | Line 1450 | public class ForkJoinPool extends Abstra
1450          for (int r = s.seed, m = submitMask;;) {
1451              WorkQueue[] ws; WorkQueue q;
1452              int k = r & m & SQMASK;          // use only even indices
1453 <            if (runState < 0 || (ws = workQueues) == null || ws.length <= k)
1453 >            if (runState < 0)
1454                  throw new RejectedExecutionException(); // shutting down
1455 +            else if ((ws = workQueues) == null || ws.length <= k) {
1456 +                while (!U.compareAndSwapInt(this, MAINLOCK, 0, 1))
1457 +                    tryAwaitMainLock();
1458 +                try {
1459 +                    if (workQueues == null)
1460 +                        workQueues = new WorkQueue[submitMask + 1];
1461 +                } finally {
1462 +                    if (!U.compareAndSwapInt(this, MAINLOCK, 1, 0)) {
1463 +                        mainLock = 0;
1464 +                        synchronized (this) { notifyAll(); };
1465 +                    }
1466 +                }
1467 +            }
1468              else if ((q = ws[k]) == null) {  // create new queue
1469                  WorkQueue nq = new WorkQueue(this, null, SHARED_QUEUE);
1470 <                Mutex lock = this.lock;      // construct outside lock
1471 <                lock.lock();
1472 <                try {                        // recheck under lock
1470 >                while (!U.compareAndSwapInt(this, MAINLOCK, 0, 1))
1471 >                    tryAwaitMainLock();
1472 >                try {
1473                      int rs = runState;       // to update seq
1474                      if (ws == workQueues && ws[k] == null) {
1475                          ws[k] = nq;
1476                          runState = ((rs & SHUTDOWN) | ((rs + 2) & ~SHUTDOWN));
1477                      }
1478                  } finally {
1479 <                    lock.unlock();
1479 >                    if (!U.compareAndSwapInt(this, MAINLOCK, 1, 0)) {
1480 >                        mainLock = 0;
1481 >                        synchronized (this) { notifyAll(); };
1482 >                    }
1483                  }
1484              }
1485              else if (q.trySharedPush(task)) {
# Line 1438 | Line 1496 | public class ForkJoinPool extends Abstra
1496          }
1497      }
1498  
1499 +    /**
1500 +     * Submits the given (non-null) task to the common pool, if possible.
1501 +     */
1502 +    static void submitToCommonPool(ForkJoinTask<?> task) {
1503 +        ForkJoinPool p;
1504 +        if ((p = commonPool) == null)
1505 +            throw new RejectedExecutionException("Common Pool Unavailable");
1506 +        p.doSubmit(task);
1507 +    }
1508 +
1509 +    /**
1510 +     * Returns true if the given task was submitted to common pool
1511 +     * and has not yet commenced execution, and is available for
1512 +     * removal according to execution policies; if so removing the
1513 +     * submission from the pool.
1514 +     *
1515 +     * @param task the task
1516 +     * @return true if successful
1517 +     */
1518 +    static boolean tryUnsubmitFromCommonPool(ForkJoinTask<?> task) {
1519 +        // Peek, looking for task and eligibility before
1520 +        // using trySharedUnpush to actually take it under lock
1521 +        ForkJoinPool p; WorkQueue[] ws; WorkQueue q;
1522 +        ForkJoinTask<?>[] a; int t, s, n;
1523 +        int k = submitters.get().seed & SQMASK;
1524 +        return ((p = commonPool) != null &&
1525 +                (ws = p.workQueues) != null &&
1526 +                ws.length > (k &= p.submitMask) &&
1527 +                (q = ws[k]) != null &&
1528 +                (a = q.array) != null &&
1529 +                (n = (t = q.top) - q.base) > 0 &&
1530 +                (n > 1 || (int)(p.ctl >> AC_SHIFT) < 0) &&
1531 +                (s = t - 1) >= 0 && s < a.length && a[s] == task &&
1532 +                q.trySharedUnpush(task));
1533 +    }
1534 +
1535      // Maintaining ctl counts
1536  
1537      /**
# Line 1449 | Line 1543 | public class ForkJoinPool extends Abstra
1543      }
1544  
1545      /**
1546 <     * Tries to activate or create a worker if too few are active.
1546 >     * Tries to create one or activate one or more workers if too few are active.
1547       */
1548      final void signalWork() {
1549          long c; int u;
# Line 1483 | Line 1577 | public class ForkJoinPool extends Abstra
1577          }
1578      }
1579  
1486
1580      // Scanning for tasks
1581  
1582      /**
# Line 1491 | Line 1584 | public class ForkJoinPool extends Abstra
1584       */
1585      final void runWorker(WorkQueue w) {
1586          w.growArray(false);         // initialize queue array in this thread
1587 <        do {} while (w.runTask(scan(w)));
1587 >        do { w.runTask(scan(w)); } while (w.runState >= 0);
1588      }
1589  
1590      /**
# Line 1534 | Line 1627 | public class ForkJoinPool extends Abstra
1627       * awaiting signal,
1628       *
1629       * @param w the worker (via its WorkQueue)
1630 <     * @return a task or null of none found
1630 >     * @return a task or null if none found
1631       */
1632      private final ForkJoinTask<?> scan(WorkQueue w) {
1633          WorkQueue[] ws;                       // first update random seed
# Line 1551 | Line 1644 | public class ForkJoinPool extends Abstra
1644                      t = (ForkJoinTask<?>)U.getObjectVolatile(a, i);
1645                      if (q.base == b && ec >= 0 && t != null &&
1646                          U.compareAndSwapObject(a, i, t, null)) {
1647 <                        q.base = b + 1;       // specialization of pollAt
1647 >                        if (q.top - (q.base = b + 1) > 0)
1648 >                            signalWork();    // help pushes signal
1649                          return t;
1650                      }
1651 <                    else if ((t != null || b + 1 != q.top) &&
1558 <                             (ec < 0 || j <= m)) {
1651 >                    else if (ec < 0 || j <= m) {
1652                          rs = 0;               // mark scan as imcomplete
1653                          break;                // caller can retry after release
1654                      }
# Line 1563 | Line 1656 | public class ForkJoinPool extends Abstra
1656                  if (--j < 0)
1657                      break;
1658              }
1659 +
1660              long c = ctl; int e = (int)c, a = (int)(c >> AC_SHIFT), nr, ns;
1661              if (e < 0)                        // decode ctl on empty scan
1662                  w.runState = -1;              // pool is terminating
# Line 1588 | Line 1682 | public class ForkJoinPool extends Abstra
1682                  else {
1683                      if ((ns = w.nsteals) != 0) {
1684                          w.nsteals = 0;        // set rescans if ran task
1685 <                        w.rescans = (a > 0)? 0 : a + parallelism;
1685 >                        w.rescans = (a > 0) ? 0 : a + parallelism;
1686                          w.totalSteals += ns;
1687                      }
1688                      if (a == 1 - parallelism) // quiescent
# Line 1596 | Line 1690 | public class ForkJoinPool extends Abstra
1690                  }
1691              }
1692              else if (w.eventCount < 0) {      // already queued
1693 <                if ((nr = w.rescans) > 0) {   // continue rescanning
1694 <                    int ac = a + parallelism;
1695 <                    if (((w.rescans = (ac < nr) ? ac : nr - 1) & 3) == 0)
1696 <                        Thread.yield();       // yield before block
1603 <                }
1604 <                else {
1693 >                int ac = a + parallelism;
1694 >                if ((nr = w.rescans) > 0)     // continue rescanning
1695 >                    w.rescans = (ac < nr) ? ac : nr - 1;
1696 >                else if (((w.seed >>> 16) & ac) == 0) { // randomize park
1697                      Thread.interrupted();     // clear status
1698                      Thread wt = Thread.currentThread();
1699                      U.putObject(wt, PARKBLOCKER, this);
# Line 1619 | Line 1711 | public class ForkJoinPool extends Abstra
1711      /**
1712       * If inactivating worker w has caused the pool to become
1713       * quiescent, checks for pool termination, and, so long as this is
1714 <     * not the only worker, waits for event for up to SHRINK_RATE
1715 <     * nanosecs.  On timeout, if ctl has not changed, terminates the
1714 >     * not the only worker, waits for event for up to a given
1715 >     * duration.  On timeout, if ctl has not changed, terminates the
1716       * worker, which will in turn wake up another worker to possibly
1717       * repeat this process.
1718       *
# Line 1630 | Line 1722 | public class ForkJoinPool extends Abstra
1722       */
1723      private void idleAwaitWork(WorkQueue w, long currentCtl, long prevCtl) {
1724          if (w.eventCount < 0 && !tryTerminate(false, false) &&
1725 <            (int)prevCtl != 0 && ctl == currentCtl) {
1725 >            (int)prevCtl != 0 && !hasQueuedSubmissions() && ctl == currentCtl) {
1726 >            int dc = -(short)(currentCtl >>> TC_SHIFT);
1727 >            long parkTime = dc < 0 ? FAST_IDLE_TIMEOUT: (dc + 1) * IDLE_TIMEOUT;
1728 >            long deadline = System.nanoTime() + parkTime - 100000L; // 1ms slop
1729              Thread wt = Thread.currentThread();
1635            Thread.yield();            // yield before block
1730              while (ctl == currentCtl) {
1637                long startTime = System.nanoTime();
1731                  Thread.interrupted();  // timed variant of version in scan()
1732                  U.putObject(wt, PARKBLOCKER, this);
1733                  w.parker = wt;
1734                  if (ctl == currentCtl)
1735 <                    U.park(false, SHRINK_RATE);
1735 >                    U.park(false, parkTime);
1736                  w.parker = null;
1737                  U.putObject(wt, PARKBLOCKER, null);
1738                  if (ctl != currentCtl)
1739                      break;
1740 <                if (System.nanoTime() - startTime >= SHRINK_TIMEOUT &&
1740 >                if (deadline - System.nanoTime() <= 0L &&
1741                      U.compareAndSwapLong(this, CTL, currentCtl, prevCtl)) {
1742                      w.eventCount = (w.eventCount + E_SEQ) | E_MASK;
1743                      w.runState = -1;   // shrink
# Line 1665 | Line 1758 | public class ForkJoinPool extends Abstra
1758       * leaves hints in workers to speed up subsequent calls. The
1759       * implementation is very branchy to cope with potential
1760       * inconsistencies or loops encountering chains that are stale,
1761 <     * unknown, or so long that they are likely cyclic.  All of these
1669 <     * cases are dealt with by just retrying by caller.
1761 >     * unknown, or so long that they are likely cyclic.
1762       *
1763       * @param joiner the joining worker
1764       * @param task the task to join
1765 <     * @return true if found or ran a task (and so is immediately retryable)
1765 >     * @return 0 if no progress can be made, negative if task
1766 >     * known complete, else positive
1767       */
1768 <    private boolean tryHelpStealer(WorkQueue joiner, ForkJoinTask<?> task) {
1769 <        WorkQueue[] ws;
1770 <        int m, depth = MAX_HELP;                // remaining chain depth
1771 <        boolean progress = false;
1772 <        if ((ws = workQueues) != null && (m = ws.length - 1) > 0 &&
1773 <            task.status >= 0) {
1774 <            ForkJoinTask<?> subtask = task;     // current target
1775 <            outer: for (WorkQueue j = joiner;;) {
1776 <                WorkQueue stealer = null;       // find stealer of subtask
1777 <                WorkQueue v = ws[j.stealHint & m]; // try hint
1778 <                if (v != null && v.currentSteal == subtask)
1779 <                    stealer = v;
1780 <                else {                          // scan
1781 <                    for (int i = 1; i <= m; i += 2) {
1782 <                        if ((v = ws[i]) != null && v.currentSteal == subtask &&
1783 <                            v != joiner) {
1784 <                            stealer = v;
1785 <                            j.stealHint = i;    // save hint
1786 <                            break;
1768 >    private int tryHelpStealer(WorkQueue joiner, ForkJoinTask<?> task) {
1769 >        int stat = 0, steps = 0;                    // bound to avoid cycles
1770 >        if (joiner != null && task != null) {       // hoist null checks
1771 >            restart: for (;;) {
1772 >                ForkJoinTask<?> subtask = task;     // current target
1773 >                for (WorkQueue j = joiner, v;;) {   // v is stealer of subtask
1774 >                    WorkQueue[] ws; int m, s, h;
1775 >                    if ((s = task.status) < 0) {
1776 >                        stat = s;
1777 >                        break restart;
1778 >                    }
1779 >                    if ((ws = workQueues) == null || (m = ws.length - 1) <= 0)
1780 >                        break restart;              // shutting down
1781 >                    if ((v = ws[h = (j.stealHint | 1) & m]) == null ||
1782 >                        v.currentSteal != subtask) {
1783 >                        for (int origin = h;;) {    // find stealer
1784 >                            if (((h = (h + 2) & m) & 15) == 1 &&
1785 >                                (subtask.status < 0 || j.currentJoin != subtask))
1786 >                                continue restart;   // occasional staleness check
1787 >                            if ((v = ws[h]) != null &&
1788 >                                v.currentSteal == subtask) {
1789 >                                j.stealHint = h;    // save hint
1790 >                                break;
1791 >                            }
1792 >                            if (h == origin)
1793 >                                break restart;      // cannot find stealer
1794                          }
1795                      }
1796 <                    if (stealer == null)
1797 <                        break;
1798 <                }
1799 <
1800 <                for (WorkQueue q = stealer;;) { // try to help stealer
1801 <                    ForkJoinTask[] a; ForkJoinTask<?> t; int b;
1802 <                    if (task.status < 0)
1803 <                        break outer;
1804 <                    if ((b = q.base) - q.top < 0 && (a = q.array) != null) {
1805 <                        progress = true;
1806 <                        int i = (((a.length - 1) & b) << ASHIFT) + ABASE;
1807 <                        t = (ForkJoinTask<?>)U.getObjectVolatile(a, i);
1808 <                        if (subtask.status < 0) // must recheck before taking
1809 <                            break outer;
1810 <                        if (t != null &&
1811 <                            q.base == b &&
1812 <                            U.compareAndSwapObject(a, i, t, null)) {
1813 <                            q.base = b + 1;
1814 <                            joiner.runSubtask(t);
1796 >                    for (;;) { // help stealer or descend to its stealer
1797 >                        ForkJoinTask[] a;  int b;
1798 >                        if (subtask.status < 0)     // surround probes with
1799 >                            continue restart;       //   consistency checks
1800 >                        if ((b = v.base) - v.top < 0 && (a = v.array) != null) {
1801 >                            int i = (((a.length - 1) & b) << ASHIFT) + ABASE;
1802 >                            ForkJoinTask<?> t =
1803 >                                (ForkJoinTask<?>)U.getObjectVolatile(a, i);
1804 >                            if (subtask.status < 0 || j.currentJoin != subtask ||
1805 >                                v.currentSteal != subtask)
1806 >                                continue restart;   // stale
1807 >                            stat = 1;               // apparent progress
1808 >                            if (t != null && v.base == b &&
1809 >                                U.compareAndSwapObject(a, i, t, null)) {
1810 >                                v.base = b + 1;     // help stealer
1811 >                                joiner.runSubtask(t);
1812 >                            }
1813 >                            else if (v.base == b && ++steps == MAX_HELP)
1814 >                                break restart;      // v apparently stalled
1815 >                        }
1816 >                        else {                      // empty -- try to descend
1817 >                            ForkJoinTask<?> next = v.currentJoin;
1818 >                            if (subtask.status < 0 || j.currentJoin != subtask ||
1819 >                                v.currentSteal != subtask)
1820 >                                continue restart;   // stale
1821 >                            else if (next == null || ++steps == MAX_HELP)
1822 >                                break restart;      // dead-end or maybe cyclic
1823 >                            else {
1824 >                                subtask = next;
1825 >                                j = v;
1826 >                                break;
1827 >                            }
1828                          }
1716                        else if (q.base == b)
1717                            break outer;        // possibly stalled
1718                    }
1719                    else {                      // descend
1720                        ForkJoinTask<?> next = stealer.currentJoin;
1721                        if (--depth <= 0 || subtask.status < 0 ||
1722                            next == null || next == subtask)
1723                            break outer;        // stale, dead-end, or cyclic
1724                        subtask = next;
1725                        j = stealer;
1726                        break;
1829                      }
1830                  }
1831              }
1832          }
1833 <        return progress;
1833 >        return stat;
1834      }
1835  
1836      /**
# Line 1757 | Line 1859 | public class ForkJoinPool extends Abstra
1859       * adds a new thread if no idle workers are available and either
1860       * pool would become completely starved or: (at least half
1861       * starved, and fewer than 50% spares exist, and there is at least
1862 <     * one task apparently available). Even though the availablity
1862 >     * one task apparently available). Even though the availability
1863       * check requires a full scan, it is worthwhile in reducing false
1864       * alarms.
1865       *
1866 <     * @param task if nonnull, a task being waited for
1867 <     * @param blocker if nonnull, a blocker being waited for
1866 >     * @param task if non-null, a task being waited for
1867 >     * @param blocker if non-null, a blocker being waited for
1868       * @return true if the caller can block, else should recheck and retry
1869       */
1870      final boolean tryCompensate(ForkJoinTask<?> task, ManagedBlocker blocker) {
# Line 1821 | Line 1923 | public class ForkJoinPool extends Abstra
1923      }
1924  
1925      /**
1926 <     * Helps and/or blocks until the given task is done
1926 >     * Helps and/or blocks until the given task is done.
1927       *
1928       * @param joiner the joining worker
1929       * @param task the task
1930       * @return task status on exit
1931       */
1932      final int awaitJoin(WorkQueue joiner, ForkJoinTask<?> task) {
1933 <        ForkJoinTask<?> prevJoin = joiner.currentJoin;
1934 <        joiner.currentJoin = task;
1935 <        long startTime = 0L;
1936 <        for (int k = 0, s; ; ++k) {
1937 <            if ((joiner.isEmpty() ?                  // try to help
1938 <                 !tryHelpStealer(joiner, task) :
1939 <                 !joiner.tryRemoveAndExec(task))) {
1940 <                if (k == 0) {
1941 <                    startTime = System.nanoTime();
1942 <                    tryPollForAndExec(joiner, task); // check uncommon case
1943 <                }
1944 <                else if ((k & (MAX_HELP - 1)) == 0 &&
1945 <                         System.nanoTime() - startTime >= COMPENSATION_DELAY &&
1946 <                         tryCompensate(task, null)) {
1947 <                    if (task.trySetSignal() && task.status >= 0) {
1948 <                        synchronized (task) {
1949 <                            if (task.status >= 0) {
1950 <                                try {                // see ForkJoinTask
1951 <                                    task.wait();     //  for explanation
1952 <                                } catch (InterruptedException ie) {
1933 >        int s;
1934 >        if ((s = task.status) >= 0) {
1935 >            ForkJoinTask<?> prevJoin = joiner.currentJoin;
1936 >            joiner.currentJoin = task;
1937 >            long startTime = 0L;
1938 >            for (int k = 0;;) {
1939 >                if ((s = (joiner.isEmpty() ?           // try to help
1940 >                          tryHelpStealer(joiner, task) :
1941 >                          joiner.tryRemoveAndExec(task))) == 0 &&
1942 >                    (s = task.status) >= 0) {
1943 >                    if (k == 0) {
1944 >                        startTime = System.nanoTime();
1945 >                        tryPollForAndExec(joiner, task); // check uncommon case
1946 >                    }
1947 >                    else if ((k & (MAX_HELP - 1)) == 0 &&
1948 >                             System.nanoTime() - startTime >=
1949 >                             COMPENSATION_DELAY &&
1950 >                             tryCompensate(task, null)) {
1951 >                        if (task.trySetSignal()) {
1952 >                            synchronized (task) {
1953 >                                if (task.status >= 0) {
1954 >                                    try {                // see ForkJoinTask
1955 >                                        task.wait();     //  for explanation
1956 >                                    } catch (InterruptedException ie) {
1957 >                                    }
1958                                  }
1959 +                                else
1960 +                                    task.notifyAll();
1961                              }
1853                            else
1854                                task.notifyAll();
1962                          }
1963 +                        long c;                          // re-activate
1964 +                        do {} while (!U.compareAndSwapLong
1965 +                                     (this, CTL, c = ctl, c + AC_UNIT));
1966                      }
1857                    long c;                          // re-activate
1858                    do {} while (!U.compareAndSwapLong
1859                                 (this, CTL, c = ctl, c + AC_UNIT));
1967                  }
1968 +                if (s < 0 || (s = task.status) < 0) {
1969 +                    joiner.currentJoin = prevJoin;
1970 +                    break;
1971 +                }
1972 +                else if ((k++ & (MAX_HELP - 1)) == MAX_HELP >>> 1)
1973 +                    Thread.yield();                     // for politeness
1974              }
1862            if ((s = task.status) < 0) {
1863                joiner.currentJoin = prevJoin;
1864                return s;
1865            }
1866            else if ((k & (MAX_HELP - 1)) == MAX_HELP >>> 1)
1867                Thread.yield();                     // for politeness
1975          }
1976 +        return s;
1977      }
1978  
1979      /**
# Line 1882 | Line 1990 | public class ForkJoinPool extends Abstra
1990          while ((s = task.status) >= 0 &&
1991                 (joiner.isEmpty() ?
1992                  tryHelpStealer(joiner, task) :
1993 <                joiner.tryRemoveAndExec(task)))
1993 >                joiner.tryRemoveAndExec(task)) != 0)
1994              ;
1995          return s;
1996      }
# Line 1922 | Line 2030 | public class ForkJoinPool extends Abstra
2030       */
2031      final void helpQuiescePool(WorkQueue w) {
2032          for (boolean active = true;;) {
2033 <            if (w.base - w.top < 0)
2034 <                w.runLocalTasks();  // exhaust local queue
2033 >            ForkJoinTask<?> localTask; // exhaust local queue
2034 >            while ((localTask = w.nextLocalTask()) != null)
2035 >                localTask.doExec();
2036              WorkQueue q = findNonEmptyStealQueue(w);
2037              if (q != null) {
2038                  ForkJoinTask<?> t; int b;
# Line 1955 | Line 2064 | public class ForkJoinPool extends Abstra
2064      }
2065  
2066      /**
2067 +     * Restricted version of helpQuiescePool for non-FJ callers
2068 +     */
2069 +    static void externalHelpQuiescePool() {
2070 +        ForkJoinPool p; WorkQueue[] ws; WorkQueue w, q;
2071 +        ForkJoinTask<?> t; int b;
2072 +        int k = submitters.get().seed & SQMASK;
2073 +        if ((p = commonPool) != null &&
2074 +            (ws = p.workQueues) != null &&
2075 +            ws.length > (k &= p.submitMask) &&
2076 +            (w = ws[k]) != null &&
2077 +            (q = p.findNonEmptyStealQueue(w)) != null &&
2078 +            (b = q.base) - q.top < 0 &&
2079 +            (t = q.pollAt(b)) != null)
2080 +            t.doExec();
2081 +    }
2082 +
2083 +    /**
2084       * Gets and removes a local or stolen task for the given worker.
2085       *
2086       * @return a task, if available
# Line 1987 | Line 2113 | public class ForkJoinPool extends Abstra
2113                  8);
2114      }
2115  
2116 +    /**
2117 +     * Returns approximate submission queue length for the given caller
2118 +     */
2119 +    static int getEstimatedSubmitterQueueLength() {
2120 +        ForkJoinPool p; WorkQueue[] ws; WorkQueue q;
2121 +        int k = submitters.get().seed & SQMASK;
2122 +        return ((p = commonPool) != null &&
2123 +                p.runState >= 0 &&
2124 +                (ws = p.workQueues) != null &&
2125 +                ws.length > (k &= p.submitMask) &&
2126 +                (q = ws[k]) != null) ?
2127 +            q.queueSize() : 0;
2128 +    }
2129 +
2130      //  Termination
2131  
2132      /**
# Line 2004 | Line 2144 | public class ForkJoinPool extends Abstra
2144       * @return true if now terminating or terminated
2145       */
2146      private boolean tryTerminate(boolean now, boolean enable) {
2007        Mutex lock = this.lock;
2147          for (long c;;) {
2148              if (((c = ctl) & STOP_BIT) != 0) {      // already terminating
2149                  if ((short)(c >>> TC_SHIFT) == -parallelism) {
2150 <                    lock.lock();                    // don't need try/finally
2151 <                    termination.signalAll();        // signal when 0 workers
2152 <                    lock.unlock();
2150 >                    synchronized(this) {
2151 >                        notifyAll();                // signal when 0 workers
2152 >                    }
2153                  }
2154                  return true;
2155              }
2156              if (runState >= 0) {                    // not yet enabled
2157                  if (!enable)
2158                      return false;
2159 <                lock.lock();
2160 <                runState |= SHUTDOWN;
2161 <                lock.unlock();
2159 >                while (!U.compareAndSwapInt(this, MAINLOCK, 0, 1))
2160 >                    tryAwaitMainLock();
2161 >                try {
2162 >                    runState |= SHUTDOWN;
2163 >                } finally {
2164 >                    if (!U.compareAndSwapInt(this, MAINLOCK, 1, 0)) {
2165 >                        mainLock = 0;
2166 >                        synchronized (this) { notifyAll(); };
2167 >                    }
2168 >                }
2169              }
2170              if (!now) {                             // check if idle & no tasks
2171                  if ((int)(c >> AC_SHIFT) != -parallelism ||
# Line 2152 | Line 2298 | public class ForkJoinPool extends Abstra
2298          // Use nearest power 2 for workQueues size. See Hackers Delight sec 3.2.
2299          int n = parallelism - 1;
2300          n |= n >>> 1; n |= n >>> 2; n |= n >>> 4; n |= n >>> 8; n |= n >>> 16;
2301 <        int size = (n + 1) << 1;        // #slots = 2*#workers
2156 <        this.submitMask = size - 1;     // room for max # of submit queues
2157 <        this.workQueues = new WorkQueue[size];
2158 <        this.termination = (this.lock = new Mutex()).newCondition();
2159 <        this.stealCount = new AtomicLong();
2160 <        this.nextWorkerNumber = new AtomicInteger();
2301 >        this.submitMask = ((n + 1) << 1) - 1;
2302          int pn = poolNumberGenerator.incrementAndGet();
2303          StringBuilder sb = new StringBuilder("ForkJoinPool-");
2304          sb.append(Integer.toString(pn));
2305          sb.append("-worker-");
2306          this.workerNamePrefix = sb.toString();
2166        lock.lock();
2307          this.runState = 1;              // set init flag
2308 <        lock.unlock();
2308 >    }
2309 >
2310 >    /**
2311 >     * Constructor for common pool, suitable only for static initialization.
2312 >     * Basically the same as above, but uses smallest possible initial footprint.
2313 >     */
2314 >    ForkJoinPool(int parallelism, int submitMask,
2315 >                 ForkJoinWorkerThreadFactory factory,
2316 >                 Thread.UncaughtExceptionHandler handler) {
2317 >        this.factory = factory;
2318 >        this.ueh = handler;
2319 >        this.submitMask = submitMask;
2320 >        this.parallelism = parallelism;
2321 >        long np = (long)(-parallelism);
2322 >        this.ctl = ((np << AC_SHIFT) & AC_MASK) | ((np << TC_SHIFT) & TC_MASK);
2323 >        this.localMode = LIFO_QUEUE;
2324 >        this.workerNamePrefix = "ForkJoinPool.commonPool-worker-";
2325 >        this.runState = 1;
2326 >    }
2327 >
2328 >    /**
2329 >     * Returns the common pool instance.
2330 >     *
2331 >     * @return the common pool instance
2332 >     */
2333 >    public static ForkJoinPool commonPool() {
2334 >        ForkJoinPool p;
2335 >        if ((p = commonPool) == null)
2336 >            throw new Error("Common Pool Unavailable");
2337 >        return p;
2338      }
2339  
2340      // Execution methods
# Line 2341 | Line 2510 | public class ForkJoinPool extends Abstra
2510      }
2511  
2512      /**
2513 +     * Returns the targeted parallelism level of the common pool.
2514 +     *
2515 +     * @return the targeted parallelism level of the common pool
2516 +     */
2517 +    public static int getCommonPoolParallelism() {
2518 +        return commonPoolParallelism;
2519 +    }
2520 +
2521 +    /**
2522       * Returns the number of worker threads that have started but not
2523       * yet terminated.  The result returned by this method may differ
2524       * from {@link #getParallelism} when threads are created to
# Line 2421 | Line 2599 | public class ForkJoinPool extends Abstra
2599       * @return the number of steals
2600       */
2601      public long getStealCount() {
2602 <        long count = stealCount.get();
2602 >        long count = stealCount;
2603          WorkQueue[] ws; WorkQueue w;
2604          if ((ws = workQueues) != null) {
2605              for (int i = 1; i < ws.length; i += 2) {
# Line 2551 | Line 2729 | public class ForkJoinPool extends Abstra
2729      public String toString() {
2730          // Use a single pass through workQueues to collect counts
2731          long qt = 0L, qs = 0L; int rc = 0;
2732 <        long st = stealCount.get();
2732 >        long st = stealCount;
2733          long c = ctl;
2734          WorkQueue[] ws; WorkQueue w;
2735          if ((ws = workQueues) != null) {
# Line 2592 | Line 2770 | public class ForkJoinPool extends Abstra
2770      }
2771  
2772      /**
2773 <     * Initiates an orderly shutdown in which previously submitted
2774 <     * tasks are executed, but no new tasks will be accepted.
2775 <     * Invocation has no additional effect if already shut down.
2776 <     * Tasks that are in the process of being submitted concurrently
2777 <     * during the course of this method may or may not be rejected.
2773 >     * Possibly initiates an orderly shutdown in which previously
2774 >     * submitted tasks are executed, but no new tasks will be
2775 >     * accepted. Invocation has no effect on execution state if this
2776 >     * is the {@link #commonPool}, and no additional effect if
2777 >     * already shut down.  Tasks that are in the process of being
2778 >     * submitted concurrently during the course of this method may or
2779 >     * may not be rejected.
2780       *
2781       * @throws SecurityException if a security manager exists and
2782       *         the caller is not permitted to modify threads
# Line 2605 | Line 2785 | public class ForkJoinPool extends Abstra
2785       */
2786      public void shutdown() {
2787          checkPermission();
2788 <        tryTerminate(false, true);
2788 >        if (this != commonPool)
2789 >            tryTerminate(false, true);
2790      }
2791  
2792      /**
2793 <     * Attempts to cancel and/or stop all tasks, and reject all
2794 <     * subsequently submitted tasks.  Tasks that are in the process of
2795 <     * being submitted or executed concurrently during the course of
2796 <     * this method may or may not be rejected. This method cancels
2797 <     * both existing and unexecuted tasks, in order to permit
2798 <     * termination in the presence of task dependencies. So the method
2799 <     * always returns an empty list (unlike the case for some other
2800 <     * Executors).
2793 >     * Possibly attempts to cancel and/or stop all tasks, and reject
2794 >     * all subsequently submitted tasks.  Invocation has no effect on
2795 >     * execution state if this is the {@link #commonPool}, and no
2796 >     * additional effect if already shut down. Otherwise, tasks that
2797 >     * are in the process of being submitted or executed concurrently
2798 >     * during the course of this method may or may not be
2799 >     * rejected. This method cancels both existing and unexecuted
2800 >     * tasks, in order to permit termination in the presence of task
2801 >     * dependencies. So the method always returns an empty list
2802 >     * (unlike the case for some other Executors).
2803       *
2804       * @return an empty list
2805       * @throws SecurityException if a security manager exists and
# Line 2626 | Line 2809 | public class ForkJoinPool extends Abstra
2809       */
2810      public List<Runnable> shutdownNow() {
2811          checkPermission();
2812 <        tryTerminate(true, true);
2812 >        if (this != commonPool)
2813 >            tryTerminate(true, true);
2814          return Collections.emptyList();
2815      }
2816  
# Line 2683 | Line 2867 | public class ForkJoinPool extends Abstra
2867      public boolean awaitTermination(long timeout, TimeUnit unit)
2868          throws InterruptedException {
2869          long nanos = unit.toNanos(timeout);
2870 <        final Mutex lock = this.lock;
2871 <        lock.lock();
2872 <        try {
2873 <            for (;;) {
2874 <                if (isTerminated())
2875 <                    return true;
2876 <                if (nanos <= 0)
2877 <                    return false;
2878 <                nanos = termination.awaitNanos(nanos);
2870 >        if (isTerminated())
2871 >            return true;
2872 >        long startTime = System.nanoTime();
2873 >        boolean terminated = false;
2874 >        synchronized(this) {
2875 >            for (long waitTime = nanos, millis = 0L;;) {
2876 >                if (terminated = isTerminated() ||
2877 >                    waitTime <= 0L ||
2878 >                    (millis = unit.toMillis(waitTime)) <= 0L)
2879 >                    break;
2880 >                wait(millis);
2881 >                waitTime = nanos - (System.nanoTime() - startTime);
2882              }
2696        } finally {
2697            lock.unlock();
2883          }
2884 +        return terminated;
2885      }
2886  
2887      /**
# Line 2827 | Line 3013 | public class ForkJoinPool extends Abstra
3013      private static final long PARKBLOCKER;
3014      private static final int ABASE;
3015      private static final int ASHIFT;
3016 +    private static final long NEXTWORKERNUMBER;
3017 +    private static final long STEALCOUNT;
3018 +    private static final long MAINLOCK;
3019  
3020      static {
3021          poolNumberGenerator = new AtomicInteger();
# Line 2842 | Line 3031 | public class ForkJoinPool extends Abstra
3031              Class<?> ak = ForkJoinTask[].class;
3032              CTL = U.objectFieldOffset
3033                  (k.getDeclaredField("ctl"));
3034 +            NEXTWORKERNUMBER = U.objectFieldOffset
3035 +                (k.getDeclaredField("nextWorkerNumber"));
3036 +            STEALCOUNT = U.objectFieldOffset
3037 +                (k.getDeclaredField("stealCount"));
3038 +            MAINLOCK = U.objectFieldOffset
3039 +                (k.getDeclaredField("mainLock"));
3040              Class<?> tk = Thread.class;
3041              PARKBLOCKER = U.objectFieldOffset
3042                  (tk.getDeclaredField("parkBlocker"));
3043              ABASE = U.arrayBaseOffset(ak);
3044              s = U.arrayIndexScale(ak);
3045 +            ASHIFT = 31 - Integer.numberOfLeadingZeros(s);
3046          } catch (Exception e) {
3047              throw new Error(e);
3048          }
3049          if ((s & (s-1)) != 0)
3050              throw new Error("data type scale not a power of two");
3051 <        ASHIFT = 31 - Integer.numberOfLeadingZeros(s);
3051 >        try { // Establish common pool
3052 >            String pp = System.getProperty(propPrefix + "parallelism");
3053 >            String fp = System.getProperty(propPrefix + "threadFactory");
3054 >            String up = System.getProperty(propPrefix + "exceptionHandler");
3055 >            ForkJoinWorkerThreadFactory fac = (fp == null) ?
3056 >                defaultForkJoinWorkerThreadFactory :
3057 >                ((ForkJoinWorkerThreadFactory)ClassLoader.
3058 >                 getSystemClassLoader().loadClass(fp).newInstance());
3059 >            Thread.UncaughtExceptionHandler ueh = (up == null)? null :
3060 >                ((Thread.UncaughtExceptionHandler)ClassLoader.
3061 >                 getSystemClassLoader().loadClass(up).newInstance());
3062 >            int par;
3063 >            if ((pp == null || (par = Integer.parseInt(pp)) <= 0))
3064 >                par = Runtime.getRuntime().availableProcessors();
3065 >            if (par > MAX_CAP)
3066 >                par = MAX_CAP;
3067 >            commonPoolParallelism = par;
3068 >            int n = par - 1; // precompute submit mask
3069 >            n |= n >>> 1; n |= n >>> 2; n |= n >>> 4;
3070 >            n |= n >>> 8; n |= n >>> 16;
3071 >            int mask = ((n + 1) << 1) - 1;
3072 >            commonPool = new ForkJoinPool(par, mask, fac, ueh);
3073 >        } catch (Exception e) {
3074 >            throw new Error(e);
3075 >        }
3076      }
3077  
3078      /**

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines