Semaphore

http://tutorials.jenkov.com/java-concurrency/semaphores.html

线程之间的同步通信，就需要信号，最简单的就是所有对象自带的notify 和wait， 但是其缺点在于可能会丢掉信号，导致死锁，并且能表示的状态一般只有一个。

semaphore可以看做是在wait和notify上封装的，更高级，可以避免丢失信号

condition是由ReentrantLock生成的，功能更为强大，能表示多种状态的同步。

http://blog.csdn.net/sunmenggmail/article/details/9851025

semaphore可以看做是在wait和notify的基础上的包装。此类同步用法可以用以协调线程之间的运行。

A Semaphore is a thread synchronization construct that can be used either to send signals between threads to avoidmissed signals, or to guard a critical section like you would with a lock. Java 5 comes with semaphore implementations in the java.util.concurrent package so you don't have to implement your own semaphores. Still, it can be useful to know the theory behind their implementation and use.

Java 5 comes with a built-in Semaphore so you don't have to implement your own. You can read more about it in thejava.util.concurrent.Semaphore text, in my java.util.concurrent tutorial.

Here is a list of the topics covered in this text:

1. Simple Semaphore
2. Using Semaphores for Signaling
3. Counting Semaphore
4. Bounded Semaphore
5. Using Semaphores as Locks

Simple Semaphore

Here is a simple Semaphore implementation:

public class Semaphore {
  private boolean signal = false;

  public synchronized void take() {
    this.signal = true;
    this.notify();
  }

  public synchronized void release() throws InterruptedException{
    while(!this.signal) wait();
    this.signal = false;
  }

}

The take() method sends a signal which is stored internally in the Semaphore. The release() method waits for a signal. When received the signal flag is cleared again, and the release() method exited.

Using a semaphore like this you can avoid missed signals. You will call take() instead of notify() and release()instead of wait(). If the call to take() happens before the call to release() the thread calling release() will still know that take() was called, because the signal is stored internally in the signal variable. This is not the case with wait() and notify().

The names take() and release() may seem a bit odd when using a semaphore for signaling. The names origin from the use of semaphores as locks, as explained later in this text. In that case the names make more sense.

Using Semaphores for Signaling

Here is a simplified example of two threads signaling each other using a Semaphore:

Semaphore semaphore = new Semaphore();

SendingThread sender = new SendingThread(semaphore);

ReceivingThread receiver = new ReceivingThread(semaphore);

receiver.start();
sender.start();

public class SendingThread {
  Semaphore semaphore = null;

  public SendingThread(Semaphore semaphore){
    this.semaphore = semaphore;
  }

  public void run(){
    while(true){
      //do something, then signal
      this.semaphore.take();

    }
  }
}

public class RecevingThread {
  Semaphore semaphore = null;

  public ReceivingThread(Semaphore semaphore){
    this.semaphore = semaphore;
  }

  public void run(){
    while(true){
      this.semaphore.release();
      //receive signal, then do something...
    }
  }
}

Counting Semaphore

The Semaphore implementation in the previous section does not count the number of signals sent to it by take()method calls. We can change the Semaphore to do so. This is called a counting semaphore. Here is a simple implementation of a counting semaphore:

public class CountingSemaphore {
  private int signals = 0;

  public synchronized void take() {
    this.signals++;
    this.notify();
  }

  public synchronized void release() throws InterruptedException{
    while(this.signals == 0) wait();
    this.signals--;
  }

}

Bounded Semaphore

The CoutingSemaphore has no upper bound on how many signals it can store. We can change the semaphore implementation to have an upper bound, like this:

public class BoundedSemaphore {
  private int signals = 0;
  private int bound   = 0;

  public BoundedSemaphore(int upperBound){
    this.bound = upperBound;
  }

  public synchronized void take() throws InterruptedException{
    while(this.signals == bound) wait();
    this.signals++;
    this.notify();
  }

  public synchronized void release() throws InterruptedException{
    while(this.signals == 0) wait();
    this.signals--;
    this.notify();
  }
}

Notice how the take() method now blocks if the number of signals is equal to the upper bound. Not until a thread has called receive will the thread calling take() be allowed to deliver its signal, if the BoundedSemaphore has reached its upper signal limit.

Using Semaphores as Locks

It is possible to use a bounded semaphore as a lock. To do so, set the upper bound to 1, and have the call to take()and release() guard the critical section. Here is an example:

BoundedSemaphore semaphore = new BoundedSemaphore(1);

...

semaphore.take();

try{
  //critical section
} finally {
  semaphore.release();
}

In contrast to the signaling use case the methods take() and release() are now called by the same thread. Since only one thread is allowed to take the semaphore, all other threads calling take() will be blocked until release() is called. The call to release() will never block since there has always been a call to take() first.

You can also use a bounded semaphore to limit the number of threads allowed into a section of code. For instance, in the example above, what would happen if you set the limit of the BoundedSemaphore to 5? 5 threads would be allowed to enter the critical section at a time. You would have to make sure though, that the thread operations do not conflict for these 5 threads, or you application will fail.

The relase() method is called from inside a finally-block to make sure it is called even if an exception is thrown from the critical section.