Package com.google.common.util.concurrent

Source Code of com.google.common.util.concurrent.Monitor$Guard

/*
* Copyright (C) 2010 The Guava Authors
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/

package com.google.common.util.concurrent;

import static com.google.common.base.Preconditions.checkNotNull;

import com.google.common.annotations.Beta;
import com.google.common.base.Throwables;

import java.util.concurrent.TimeUnit;
import java.util.concurrent.locks.Condition;
import java.util.concurrent.locks.ReentrantLock;

import javax.annotation.concurrent.GuardedBy;

/**
* A synchronization abstraction supporting waiting on arbitrary boolean conditions.
*
* <p>This class is intended as a replacement for {@link ReentrantLock}. Code using {@code Monitor}
* is less error-prone and more readable than code using {@code ReentrantLock}, without significant
* performance loss. {@code Monitor} even has the potential for performance gain by optimizing the
* evaluation and signaling of conditions.  Signaling is entirely
* <a href="http://en.wikipedia.org/wiki/Monitor_(synchronization)#Implicit_signaling">
* implicit</a>.
* By eliminating explicit signaling, this class can guarantee that only one thread is awakened
* when a condition becomes true (no "signaling storms" due to use of {@link
* java.util.concurrent.locks.Condition#signalAll Condition.signalAll}) and that no signals are lost
* (no "hangs" due to incorrect use of {@link java.util.concurrent.locks.Condition#signal
* Condition.signal}).
*
* <p>A thread is said to <i>occupy</i> a monitor if it has <i>entered</i> the monitor but not yet
* <i>left</i>. Only one thread may occupy a given monitor at any moment. A monitor is also
* reentrant, so a thread may enter a monitor any number of times, and then must leave the same
* number of times. The <i>enter</i> and <i>leave</i> operations have the same synchronization
* semantics as the built-in Java language synchronization primitives.
*
* <p>A call to any of the <i>enter</i> methods with <b>void</b> return type should always be
* followed immediately by a <i>try/finally</i> block to ensure that the current thread leaves the
* monitor cleanly: <pre>   {@code
*
*   monitor.enter();
*   try {
*     // do things while occupying the monitor
*   } finally {
*     monitor.leave();
*   }}</pre>
*
* <p>A call to any of the <i>enter</i> methods with <b>boolean</b> return type should always
* appear as the condition of an <i>if</i> statement containing a <i>try/finally</i> block to
* ensure that the current thread leaves the monitor cleanly: <pre>   {@code
*
*   if (monitor.tryEnter()) {
*     try {
*       // do things while occupying the monitor
*     } finally {
*       monitor.leave();
*     }
*   } else {
*     // do other things since the monitor was not available
*   }}</pre>
*
* <h2>Comparison with {@code synchronized} and {@code ReentrantLock}</h2>
*
* <p>The following examples show a simple threadsafe holder expressed using {@code synchronized},
* {@link ReentrantLock}, and {@code Monitor}.
*
* <h3>{@code synchronized}</h3>
*
* <p>This version is the fewest lines of code, largely because the synchronization mechanism used
* is built into the language and runtime. But the programmer has to remember to avoid a couple of
* common bugs: The {@code wait()} must be inside a {@code while} instead of an {@code if}, and
* {@code notifyAll()} must be used instead of {@code notify()} because there are two different
* logical conditions being awaited. <pre>   {@code
*
*   public class SafeBox<V> {
*     private V value;
*
*     public synchronized V get() throws InterruptedException {
*       while (value == null) {
*         wait();
*       }
*       V result = value;
*       value = null;
*       notifyAll();
*       return result;
*     }
*
*     public synchronized void set(V newValue) throws InterruptedException {
*       while (value != null) {
*         wait();
*       }
*       value = newValue;
*       notifyAll();
*     }
*   }}</pre>
*
* <h3>{@code ReentrantLock}</h3>
*
* <p>This version is much more verbose than the {@code synchronized} version, and still suffers
* from the need for the programmer to remember to use {@code while} instead of {@code if}.
* However, one advantage is that we can introduce two separate {@code Condition} objects, which
* allows us to use {@code signal()} instead of {@code signalAll()}, which may be a performance
* benefit. <pre>   {@code
*
*   public class SafeBox<V> {
*     private final ReentrantLock lock = new ReentrantLock();
*     private final Condition valuePresent = lock.newCondition();
*     private final Condition valueAbsent = lock.newCondition();
*     private V value;
*
*     public V get() throws InterruptedException {
*       lock.lock();
*       try {
*         while (value == null) {
*           valuePresent.await();
*         }
*         V result = value;
*         value = null;
*         valueAbsent.signal();
*         return result;
*       } finally {
*         lock.unlock();
*       }
*     }
*
*     public void set(V newValue) throws InterruptedException {
*       lock.lock();
*       try {
*         while (value != null) {
*           valueAbsent.await();
*         }
*         value = newValue;
*         valuePresent.signal();
*       } finally {
*         lock.unlock();
*       }
*     }
*   }}</pre>
*
* <h3>{@code Monitor}</h3>
*
* <p>This version adds some verbosity around the {@code Guard} objects, but removes that same
* verbosity, and more, from the {@code get} and {@code set} methods. {@code Monitor} implements the
* same efficient signaling as we had to hand-code in the {@code ReentrantLock} version above.
* Finally, the programmer no longer has to hand-code the wait loop, and therefore doesn't have to
* remember to use {@code while} instead of {@code if}. <pre>   {@code
*
*   public class SafeBox<V> {
*     private final Monitor monitor = new Monitor();
*     private final Monitor.Guard valuePresent = new Monitor.Guard(monitor) {
*       public boolean isSatisfied() {
*         return value != null;
*       }
*     };
*     private final Monitor.Guard valueAbsent = new Monitor.Guard(monitor) {
*       public boolean isSatisfied() {
*         return value == null;
*       }
*     };
*     private V value;
*
*     public V get() throws InterruptedException {
*       monitor.enterWhen(valuePresent);
*       try {
*         V result = value;
*         value = null;
*         return result;
*       } finally {
*         monitor.leave();
*       }
*     }
*
*     public void set(V newValue) throws InterruptedException {
*       monitor.enterWhen(valueAbsent);
*       try {
*         value = newValue;
*       } finally {
*         monitor.leave();
*       }
*     }
*   }}</pre>
*
* @author Justin T. Sampson
* @author Martin Buchholz
* @since 10.0
*/
@Beta
public final class Monitor {
  // TODO(user): Use raw LockSupport or AbstractQueuedSynchronizer instead of ReentrantLock.
  // TODO(user): "Port" jsr166 tests for ReentrantLock.
  //
  // TODO(user): Change API to make it impossible to use a Guard with the "wrong" monitor,
  //    by making the monitor implicit, and to eliminate other sources of IMSE.
  //    Imagine:
  //    guard.lock();
  //    try { /* monitor locked and guard satisfied here */ }
  //    finally { guard.unlock(); }
  // Here are Justin's design notes about this:
  //
  // This idea has come up from time to time, and I think one of my
  // earlier versions of Monitor even did something like this. I ended
  // up strongly favoring the current interface.
  //
  // I probably can't remember all the reasons (it's possible you
  // could find them in the code review archives), but here are a few:
  //
  // 1. What about leaving/unlocking? Are you going to do
  //    guard.enter() paired with monitor.leave()? That might get
  //    confusing. It's nice for the finally block to look as close as
  //    possible to the thing right before the try. You could have
  //    guard.leave(), but that's a little odd as well because the
  //    guard doesn't have anything to do with leaving. You can't
  //    really enforce that the guard you're leaving is the same one
  //    you entered with, and it doesn't actually matter.
  //
  // 2. Since you can enter the monitor without a guard at all, some
  //    places you'll have monitor.enter()/monitor.leave() and other
  //    places you'll have guard.enter()/guard.leave() even though
  //    it's the same lock being acquired underneath. Always using
  //    monitor.enterXXX()/monitor.leave() will make it really clear
  //    which lock is held at any point in the code.
  //
  // 3. I think "enterWhen(notEmpty)" reads better than "notEmpty.enter()".
  //
  // TODO(user): Implement ReentrantLock features:
  //    - toString() method
  //    - getOwner() method
  //    - getQueuedThreads() method
  //    - getWaitingThreads(Guard) method
  //    - implement Serializable
  //    - redo the API to be as close to identical to ReentrantLock as possible,
  //      since, after all, this class is also a reentrant mutual exclusion lock!?

  /*
   * One of the key challenges of this class is to prevent lost signals, while trying hard to
   * minimize unnecessary signals.  One simple and correct algorithm is to signal some other
   * waiter with a satisfied guard (if one exists) whenever any thread occupying the monitor
   * exits the monitor, either by unlocking all of its held locks, or by starting to wait for a
   * guard.  This includes exceptional exits, so all control paths involving signalling must be
   * protected by a finally block.
   *
   * Further optimizations of this algorithm become increasingly subtle.  A wait that terminates
   * without the guard being satisfied (due to timeout, but not interrupt) can then immediately
   * exit the monitor without signalling.  If it timed out without being signalled, it does not
   * need to "pass on" the signal to another thread.  If it *was* signalled, then its guard must
   * have been satisfied at the time of signal, and has since been modified by some other thread
   * to be non-satisfied before reacquiring the lock, and that other thread takes over the
   * responsibility of signaling the next waiter.
   *
   * Unlike the underlying Condition, if we are not careful, an interrupt *can* cause a signal to
   * be lost, because the signal may be sent to a condition whose sole waiter has just been
   * interrupted.
   *
   * Imagine a monitor with multiple guards.  A thread enters the monitor, satisfies all the
   * guards, and leaves, calling signalNextWaiter.  With traditional locks and conditions, all
   * the conditions need to be signalled because it is not known which if any of them have
   * waiters (and hasWaiters can't be used reliably because of a check-then-act race).  With our
   * Monitor guards, we only signal the first active guard that is satisfied.  But the
   * corresponding thread may have already been interrupted and is waiting to reacquire the lock
   * while still registered in activeGuards, in which case the signal is a no-op, and the
   * bigger-picture signal is lost unless interrupted threads take special action by
   * participating in the signal-passing game.
   */

  /*
   * Timeout handling is intricate, especially given our ambitious goals:
   * - Avoid underflow and overflow of timeout values when specified timeouts are close to
   *   Long.MIN_VALUE or Long.MAX_VALUE.
   * - Favor responding to interrupts over timeouts.
   * - System.nanoTime() is expensive enough that we want to call it the minimum required number of
   *   times, typically once before invoking a blocking method.  This often requires keeping track
   *   of the first time in a method that nanoTime() has been invoked, for which the special value
   *   0L is reserved to mean "uninitialized".  If timeout is non-positive, then nanoTime need
   *   never be called.
   * - Keep behavior of fair and non-fair instances consistent.
   */

  /**
   * A boolean condition for which a thread may wait. A {@code Guard} is associated with a single
   * {@code Monitor}. The monitor may check the guard at arbitrary times from any thread occupying
   * the monitor, so code should not be written to rely on how often a guard might or might not be
   * checked.
   *
   * <p>If a {@code Guard} is passed into any method of a {@code Monitor} other than the one it is
   * associated with, an {@link IllegalMonitorStateException} is thrown.
   *
   * @since 10.0
   */
  @Beta
  public abstract static class Guard {

    final Monitor monitor;
    final Condition condition;

    @GuardedBy("monitor.lock")
    int waiterCount = 0;

    /** The next active guard */
    @GuardedBy("monitor.lock")
    Guard next;

    protected Guard(Monitor monitor) {
      this.monitor = checkNotNull(monitor, "monitor");
      this.condition = monitor.lock.newCondition();
    }

    /**
     * Evaluates this guard's boolean condition. This method is always called with the associated
     * monitor already occupied. Implementations of this method must depend only on state protected
     * by the associated monitor, and must not modify that state.
     */
    public abstract boolean isSatisfied();

  }

  /**
   * Whether this monitor is fair.
   */
  private final boolean fair;

  /**
   * The lock underlying this monitor.
   */
  private final ReentrantLock lock;

  /**
   * The guards associated with this monitor that currently have waiters ({@code waiterCount > 0}).
   * A linked list threaded through the Guard.next field.
   */
  @GuardedBy("lock")
  private Guard activeGuards = null;

  /**
   * Creates a monitor with a non-fair (but fast) ordering policy. Equivalent to {@code
   * Monitor(false)}.
   */
  public Monitor() {
    this(false);
  }

  /**
   * Creates a monitor with the given ordering policy.
   *
   * @param fair whether this monitor should use a fair ordering policy rather than a non-fair (but
   *        fast) one
   */
  public Monitor(boolean fair) {
    this.fair = fair;
    this.lock = new ReentrantLock(fair);
  }

  /**
   * Enters this monitor. Blocks indefinitely.
   */
  public void enter() {
    lock.lock();
  }

  /**
   * Enters this monitor. Blocks indefinitely, but may be interrupted.
   *
   * @throws InterruptedException if interrupted while waiting
   */
  public void enterInterruptibly() throws InterruptedException {
    lock.lockInterruptibly();
  }

  /**
   * Enters this monitor. Blocks at most the given time.
   *
   * @return whether the monitor was entered
   */
  public boolean enter(long time, TimeUnit unit) {
    final long timeoutNanos = toSafeNanos(time, unit);
    final ReentrantLock lock = this.lock;
    if (!fair && lock.tryLock()) {
      return true;
    }
    boolean interrupted = Thread.interrupted();
    try {
      final long startTime = System.nanoTime();
      for (long remainingNanos = timeoutNanos;;) {
        try {
          return lock.tryLock(remainingNanos, TimeUnit.NANOSECONDS);
        } catch (InterruptedException interrupt) {
          interrupted = true;
          remainingNanos = remainingNanos(startTime, timeoutNanos);
        }
      }
    } finally {
      if (interrupted) {
        Thread.currentThread().interrupt();
      }
    }
  }

  /**
   * Enters this monitor. Blocks at most the given time, and may be interrupted.
   *
   * @return whether the monitor was entered
   * @throws InterruptedException if interrupted while waiting
   */
  public boolean enterInterruptibly(long time, TimeUnit unit) throws InterruptedException {
    return lock.tryLock(time, unit);
  }

  /**
   * Enters this monitor if it is possible to do so immediately. Does not block.
   *
   * <p><b>Note:</b> This method disregards the fairness setting of this monitor.
   *
   * @return whether the monitor was entered
   */
  public boolean tryEnter() {
    return lock.tryLock();
  }

  /**
   * Enters this monitor when the guard is satisfied. Blocks indefinitely, but may be interrupted.
   *
   * @throws InterruptedException if interrupted while waiting
   */
  public void enterWhen(Guard guard) throws InterruptedException {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    boolean signalBeforeWaiting = lock.isHeldByCurrentThread();
    lock.lockInterruptibly();

    boolean satisfied = false;
    try {
      if (!guard.isSatisfied()) {
        await(guard, signalBeforeWaiting);
      }
      satisfied = true;
    } finally {
      if (!satisfied) {
        leave();
      }
    }
  }

  /**
   * Enters this monitor when the guard is satisfied. Blocks indefinitely.
   */
  public void enterWhenUninterruptibly(Guard guard) {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    boolean signalBeforeWaiting = lock.isHeldByCurrentThread();
    lock.lock();

    boolean satisfied = false;
    try {
      if (!guard.isSatisfied()) {
        awaitUninterruptibly(guard, signalBeforeWaiting);
      }
      satisfied = true;
    } finally {
      if (!satisfied) {
        leave();
      }
    }
  }

  /**
   * Enters this monitor when the guard is satisfied. Blocks at most the given time, including both
   * the time to acquire the lock and the time to wait for the guard to be satisfied, and may be
   * interrupted.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   * @throws InterruptedException if interrupted while waiting
   */
  public boolean enterWhen(Guard guard, long time, TimeUnit unit) throws InterruptedException {
    final long timeoutNanos = toSafeNanos(time, unit);
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    boolean reentrant = lock.isHeldByCurrentThread();
    long startTime = 0L;

locked: {
      if (!fair) {
        // Check interrupt status to get behavior consistent with fair case.
        if (Thread.interrupted()) {
          throw new InterruptedException();
        }
        if (lock.tryLock()) {
          break locked;
        }
      }
      startTime = initNanoTime(timeoutNanos);
      if (!lock.tryLock(time, unit)) {
        return false;
      }
    }

    boolean satisfied = false;
    boolean threw = true;
    try {
      satisfied = guard.isSatisfied()
          || awaitNanos(guard,
                        (startTime == 0L)
                            ? timeoutNanos
                            : remainingNanos(startTime, timeoutNanos),
                        reentrant);
      threw = false;
      return satisfied;
    } finally {
      if (!satisfied) {
        try {
          // Don't need to signal if timed out, but do if interrupted
          if (threw && !reentrant) {
            signalNextWaiter();
          }
        } finally {
          lock.unlock();
        }
      }
    }
  }

  /**
   * Enters this monitor when the guard is satisfied. Blocks at most the given time, including
   * both the time to acquire the lock and the time to wait for the guard to be satisfied.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   */
  public boolean enterWhenUninterruptibly(Guard guard, long time, TimeUnit unit) {
    final long timeoutNanos = toSafeNanos(time, unit);
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    long startTime = 0L;
    boolean signalBeforeWaiting = lock.isHeldByCurrentThread();
    boolean interrupted = Thread.interrupted();
    try {
      if (fair || !lock.tryLock()) {
        startTime = initNanoTime(timeoutNanos);
        for (long remainingNanos = timeoutNanos;;) {
          try {
            if (lock.tryLock(remainingNanos, TimeUnit.NANOSECONDS)) {
              break;
            } else {
              return false;
            }
          } catch (InterruptedException interrupt) {
            interrupted = true;
            remainingNanos = remainingNanos(startTime, timeoutNanos);
          }
        }
      }

      boolean satisfied = false;
      try {
        while (true) {
          try {
            if (guard.isSatisfied()) {
              satisfied = true;
            } else {
              final long remainingNanos;
              if (startTime == 0L) {
                startTime = initNanoTime(timeoutNanos);
                remainingNanos = timeoutNanos;
              } else {
                remainingNanos = remainingNanos(startTime, timeoutNanos);
              }
              satisfied = awaitNanos(guard, remainingNanos, signalBeforeWaiting);
            }
            return satisfied;
          } catch (InterruptedException interrupt) {
            interrupted = true;
            signalBeforeWaiting = false;
          }
        }
      } finally {
        if (!satisfied) {
          lock.unlock()// No need to signal if timed out
        }
      }
    } finally {
      if (interrupted) {
        Thread.currentThread().interrupt();
      }
    }
  }

  /**
   * Enters this monitor if the guard is satisfied. Blocks indefinitely acquiring the lock, but
   * does not wait for the guard to be satisfied.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   */
  public boolean enterIf(Guard guard) {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    lock.lock();

    boolean satisfied = false;
    try {
      return satisfied = guard.isSatisfied();
    } finally {
      if (!satisfied) {
        lock.unlock();
      }
    }
  }

  /**
   * Enters this monitor if the guard is satisfied. Blocks indefinitely acquiring the lock, but does
   * not wait for the guard to be satisfied, and may be interrupted.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   * @throws InterruptedException if interrupted while waiting
   */
  public boolean enterIfInterruptibly(Guard guard) throws InterruptedException {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    lock.lockInterruptibly();

    boolean satisfied = false;
    try {
      return satisfied = guard.isSatisfied();
    } finally {
      if (!satisfied) {
        lock.unlock();
      }
    }
  }

  /**
   * Enters this monitor if the guard is satisfied. Blocks at most the given time acquiring the
   * lock, but does not wait for the guard to be satisfied.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   */
  public boolean enterIf(Guard guard, long time, TimeUnit unit) {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    if (!enter(time, unit)) {
      return false;
    }

    boolean satisfied = false;
    try {
      return satisfied = guard.isSatisfied();
    } finally {
      if (!satisfied) {
        lock.unlock();
      }
    }
  }

  /**
   * Enters this monitor if the guard is satisfied. Blocks at most the given time acquiring the
   * lock, but does not wait for the guard to be satisfied, and may be interrupted.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   */
  public boolean enterIfInterruptibly(Guard guard, long time, TimeUnit unit)
      throws InterruptedException {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    if (!lock.tryLock(time, unit)) {
      return false;
    }

    boolean satisfied = false;
    try {
      return satisfied = guard.isSatisfied();
    } finally {
      if (!satisfied) {
        lock.unlock();
      }
    }
  }

  /**
   * Enters this monitor if it is possible to do so immediately and the guard is satisfied. Does not
   * block acquiring the lock and does not wait for the guard to be satisfied.
   *
   * <p><b>Note:</b> This method disregards the fairness setting of this monitor.
   *
   * @return whether the monitor was entered, which guarantees that the guard is now satisfied
   */
  public boolean tryEnterIf(Guard guard) {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    final ReentrantLock lock = this.lock;
    if (!lock.tryLock()) {
      return false;
    }

    boolean satisfied = false;
    try {
      return satisfied = guard.isSatisfied();
    } finally {
      if (!satisfied) {
        lock.unlock();
      }
    }
  }

  /**
   * Waits for the guard to be satisfied. Waits indefinitely, but may be interrupted. May be
   * called only by a thread currently occupying this monitor.
   *
   * @throws InterruptedException if interrupted while waiting
   */
  public void waitFor(Guard guard) throws InterruptedException {
    if (!((guard.monitor == this) & lock.isHeldByCurrentThread())) {
      throw new IllegalMonitorStateException();
    }
    if (!guard.isSatisfied()) {
      await(guard, true);
    }
  }

  /**
   * Waits for the guard to be satisfied. Waits indefinitely. May be called only by a thread
   * currently occupying this monitor.
   */
  public void waitForUninterruptibly(Guard guard) {
    if (!((guard.monitor == this) & lock.isHeldByCurrentThread())) {
      throw new IllegalMonitorStateException();
    }
    if (!guard.isSatisfied()) {
      awaitUninterruptibly(guard, true);
    }
  }

  /**
   * Waits for the guard to be satisfied. Waits at most the given time, and may be interrupted.
   * May be called only by a thread currently occupying this monitor.
   *
   * @return whether the guard is now satisfied
   * @throws InterruptedException if interrupted while waiting
   */
  public boolean waitFor(Guard guard, long time, TimeUnit unit) throws InterruptedException {
    final long timeoutNanos = toSafeNanos(time, unit);
    if (!((guard.monitor == this) & lock.isHeldByCurrentThread())) {
      throw new IllegalMonitorStateException();
    }
    if (guard.isSatisfied()) {
      return true;
    }
    if (Thread.interrupted()) {
      throw new InterruptedException();
    }
    return awaitNanos(guard, timeoutNanos, true);
  }

  /**
   * Waits for the guard to be satisfied. Waits at most the given time. May be called only by a
   * thread currently occupying this monitor.
   *
   * @return whether the guard is now satisfied
   */
  public boolean waitForUninterruptibly(Guard guard, long time, TimeUnit unit) {
    final long timeoutNanos = toSafeNanos(time, unit);
    if (!((guard.monitor == this) & lock.isHeldByCurrentThread())) {
      throw new IllegalMonitorStateException();
    }
    if (guard.isSatisfied()) {
      return true;
    }
    boolean signalBeforeWaiting = true;
    final long startTime = initNanoTime(timeoutNanos);
    boolean interrupted = Thread.interrupted();
    try {
      for (long remainingNanos = timeoutNanos;;) {
        try {
          return awaitNanos(guard, remainingNanos, signalBeforeWaiting);
        } catch (InterruptedException interrupt) {
          interrupted = true;
          if (guard.isSatisfied()) {
            return true;
          }
          signalBeforeWaiting = false;
          remainingNanos = remainingNanos(startTime, timeoutNanos);
        }
      }
    } finally {
      if (interrupted) {
        Thread.currentThread().interrupt();
      }
    }
  }

  /**
   * Leaves this monitor. May be called only by a thread currently occupying this monitor.
   */
  public void leave() {
    final ReentrantLock lock = this.lock;
    try {
      // No need to signal if we will still be holding the lock when we return
      if (lock.getHoldCount() == 1) {
        signalNextWaiter();
      }
    } finally {
      lock.unlock()// Will throw IllegalMonitorStateException if not held
    }
  }

  /**
   * Returns whether this monitor is using a fair ordering policy.
   */
  public boolean isFair() {
    return fair;
  }

  /**
   * Returns whether this monitor is occupied by any thread. This method is designed for use in
   * monitoring of the system state, not for synchronization control.
   */
  public boolean isOccupied() {
    return lock.isLocked();
  }

  /**
   * Returns whether the current thread is occupying this monitor (has entered more times than it
   * has left).
   */
  public boolean isOccupiedByCurrentThread() {
    return lock.isHeldByCurrentThread();
  }

  /**
   * Returns the number of times the current thread has entered this monitor in excess of the number
   * of times it has left. Returns 0 if the current thread is not occupying this monitor.
   */
  public int getOccupiedDepth() {
    return lock.getHoldCount();
  }

  /**
   * Returns an estimate of the number of threads waiting to enter this monitor. The value is only
   * an estimate because the number of threads may change dynamically while this method traverses
   * internal data structures. This method is designed for use in monitoring of the system state,
   * not for synchronization control.
   */
  public int getQueueLength() {
    return lock.getQueueLength();
  }

  /**
   * Returns whether any threads are waiting to enter this monitor. Note that because cancellations
   * may occur at any time, a {@code true} return does not guarantee that any other thread will ever
   * enter this monitor. This method is designed primarily for use in monitoring of the system
   * state.
   */
  public boolean hasQueuedThreads() {
    return lock.hasQueuedThreads();
  }

  /**
   * Queries whether the given thread is waiting to enter this monitor. Note that because
   * cancellations may occur at any time, a {@code true} return does not guarantee that this thread
   * will ever enter this monitor. This method is designed primarily for use in monitoring of the
   * system state.
   */
  public boolean hasQueuedThread(Thread thread) {
    return lock.hasQueuedThread(thread);
  }

  /**
   * Queries whether any threads are waiting for the given guard to become satisfied. Note that
   * because timeouts and interrupts may occur at any time, a {@code true} return does not guarantee
   * that the guard becoming satisfied in the future will awaken any threads. This method is
   * designed primarily for use in monitoring of the system state.
   */
  public boolean hasWaiters(Guard guard) {
    return getWaitQueueLength(guard) > 0;
  }

  /**
   * Returns an estimate of the number of threads waiting for the given guard to become satisfied.
   * Note that because timeouts and interrupts may occur at any time, the estimate serves only as an
   * upper bound on the actual number of waiters. This method is designed for use in monitoring of
   * the system state, not for synchronization control.
   */
  public int getWaitQueueLength(Guard guard) {
    if (guard.monitor != this) {
      throw new IllegalMonitorStateException();
    }
    lock.lock();
    try {
      return guard.waiterCount;
    } finally {
      lock.unlock();
    }
  }

  /**
   * Returns unit.toNanos(time), additionally ensuring the returned value is not at risk of
   * overflowing or underflowing, by bounding the value between 0 and (Long.MAX_VALUE / 4) * 3.
   * Actually waiting for more than 219 years is not supported!
   */
  private static long toSafeNanos(long time, TimeUnit unit) {
    long timeoutNanos = unit.toNanos(time);
    return (timeoutNanos <= 0L) ? 0L
        : (timeoutNanos > (Long.MAX_VALUE / 4) * 3) ? (Long.MAX_VALUE / 4) * 3
        : timeoutNanos;
  }

  /**
   * Returns System.nanoTime() unless the timeout has already elapsed.
   * Returns 0L if and only if the timeout has already elapsed.
   */
  private static long initNanoTime(long timeoutNanos) {
    if (timeoutNanos <= 0L) {
      return 0L;
    } else {
      long startTime = System.nanoTime();
      return (startTime == 0L) ? 1L : startTime;
    }
  }

  /**
   * Returns the remaining nanos until the given timeout, or 0L if the timeout has already elapsed.
   * Caller must have previously sanitized timeoutNanos using toSafeNanos.
   */
  private static long remainingNanos(long startTime, long timeoutNanos) {
    // assert timeoutNanos == 0L || startTime != 0L;

    // TODO : NOT CORRECT, BUT TESTS PASS ANYWAYS!
    // if (true) return timeoutNanos;
    // ONLY 2 TESTS FAIL IF WE DO:
    // if (true) return 0;

    return (timeoutNanos <= 0L) ? 0L : timeoutNanos - (System.nanoTime() - startTime);
  }

  /**
   * Signals some other thread waiting on a satisfied guard, if one exists.
   *
   * We manage calls to this method carefully, to signal only when necessary, but never losing a
   * signal, which is the classic problem of this kind of concurrency construct.  We must signal if
   * the current thread is about to relinquish the lock and may have changed the state protected by
   * the monitor, thereby causing some guard to be satisfied.
   *
   * In addition, any thread that has been signalled when its guard was satisfied acquires the
   * responsibility of signalling the next thread when it again relinquishes the lock.  Unlike a
   * normal Condition, there is no guarantee that an interrupted thread has not been signalled,
   * since the concurrency control must manage multiple Conditions.  So this method must generally
   * be called when waits are interrupted.
   *
   * On the other hand, if a signalled thread wakes up to discover that its guard is still not
   * satisfied, it does *not* need to call this method before returning to wait.  This can only
   * happen due to spurious wakeup (ignorable) or another thread acquiring the lock before the
   * current thread can and returning the guard to the unsatisfied state.  In the latter case the
   * other thread (last thread modifying the state protected by the monitor) takes over the
   * responsibility of signalling the next waiter.
   *
   * This method must not be called from within a beginWaitingFor/endWaitingFor block, or else the
   * current thread's guard might be mistakenly signalled, leading to a lost signal.
   */
  @GuardedBy("lock")
  private void signalNextWaiter() {
    for (Guard guard = activeGuards; guard != null; guard = guard.next) {
      if (isSatisfied(guard)) {
        guard.condition.signal();
        break;
      }
    }
  }

  /**
   * Exactly like signalNextWaiter, but caller guarantees that guardToSkip need not be considered,
   * because caller has previously checked that guardToSkip.isSatisfied() returned false.
   * An optimization for the case that guardToSkip.isSatisfied() may be expensive.
   *
   * We decided against using this method, since in practice, isSatisfied() is likely to be very
   * cheap (typically one field read).  Resurrect this method if you find that not to be true.
   */
//   @GuardedBy("lock")
//   private void signalNextWaiterSkipping(Guard guardToSkip) {
//     for (Guard guard = activeGuards; guard != null; guard = guard.next) {
//       if (guard != guardToSkip && isSatisfied(guard)) {
//         guard.condition.signal();
//         break;
//       }
//     }
//   }

  /**
   * Exactly like guard.isSatisfied(), but in addition signals all waiting threads in the
   * (hopefully unlikely) event that isSatisfied() throws.
   */
  @GuardedBy("lock")
  private boolean isSatisfied(Guard guard) {
    try {
      return guard.isSatisfied();
    } catch (Throwable throwable) {
      signalAllWaiters();
      throw Throwables.propagate(throwable);
    }
  }

  /**
   * Signals all threads waiting on guards.
   */
  @GuardedBy("lock")
  private void signalAllWaiters() {
    for (Guard guard = activeGuards; guard != null; guard = guard.next) {
      guard.condition.signalAll();
    }
  }

  /**
   * Records that the current thread is about to wait on the specified guard.
   */
  @GuardedBy("lock")
  private void beginWaitingFor(Guard guard) {
    int waiters = guard.waiterCount++;
    if (waiters == 0) {
      // push guard onto activeGuards
      guard.next = activeGuards;
      activeGuards = guard;
    }
  }

  /**
   * Records that the current thread is no longer waiting on the specified guard.
   */
  @GuardedBy("lock")
  private void endWaitingFor(Guard guard) {
    int waiters = --guard.waiterCount;
    if (waiters == 0) {
      // unlink guard from activeGuards
      for (Guard p = activeGuards, pred = null;; pred = p, p = p.next) {
        if (p == guard) {
          if (pred == null) {
            activeGuards = p.next;
          } else {
            pred.next = p.next;
          }
          p.next = null// help GC
          break;
        }
      }
    }
  }

  /*
   * Methods that loop waiting on a guard's condition until the guard is satisfied, while
   * recording this fact so that other threads know to check our guard and signal us.
   * It's caller's responsibility to ensure that the guard is *not* currently satisfied.
   */

  @GuardedBy("lock")
  private void await(Guard guard, boolean signalBeforeWaiting)
      throws InterruptedException {
    if (signalBeforeWaiting) {
      signalNextWaiter();
    }
    beginWaitingFor(guard);
    try {
      do {
        guard.condition.await();
      } while (!guard.isSatisfied());
    } finally {
      endWaitingFor(guard);
    }
  }

  @GuardedBy("lock")
  private void awaitUninterruptibly(Guard guard, boolean signalBeforeWaiting) {
    if (signalBeforeWaiting) {
      signalNextWaiter();
    }
    beginWaitingFor(guard);
    try {
      do {
        guard.condition.awaitUninterruptibly();
      } while (!guard.isSatisfied());
    } finally {
      endWaitingFor(guard);
    }
  }

  /**
   * Caller should check before calling that guard is not satisfied.
   */
  @GuardedBy("lock")
  private boolean awaitNanos(Guard guard, long nanos, boolean signalBeforeWaiting)
      throws InterruptedException {
    boolean firstTime = true;
    try {
      do {
        if (nanos <= 0L) {
          return false;
        }
        if (firstTime) {
          if (signalBeforeWaiting) {
            signalNextWaiter();
          }
          beginWaitingFor(guard);
          firstTime = false;
        }
        nanos = guard.condition.awaitNanos(nanos);
      } while (!guard.isSatisfied());
      return true;
    } finally {
      if (!firstTime) {
        endWaitingFor(guard);
      }
    }
  }

}
TOP

Related Classes of com.google.common.util.concurrent.Monitor$Guard

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.