深入阅读Mina源码(3) ―― Mina之IOAdapter(二)

 （接：深入阅读Mina源码(3) —— Mina之IOAdapter(一)）

 

4. AbstractPollingIoAcceptor

 

这个类在org.apache.mina.core.polling，这个包一共有四个类，这个类就可以代表整个包的内容了，细致说一下，之后就不再深入介绍这个包了。从名字可以看出包里面的类主要完成轮询。

AbstractPollingIoAcceptor主要完成了AbstractIoAcceptor定义的实际实现方法，如unbind0、bindInternal、dispose0，另外下方了一些更细致的accept、open、select、colse方法让子类实现，也清晰的可以看出它主要为Socket提供父类的一些实现，同时也有了NIO的影子。

看一下AbstractPollingIoAcceptor的属性字段，可以看到

 

 

/** A lock used to protect the selector to be waked up before it's created */
private final Semaphore lock = new Semaphore(1);

private final IoProcessor<S> processor;

private final boolean createdProcessor;

private final Queue<AcceptorOperationFuture> registerQueue = new ConcurrentLinkedQueue<AcceptorOperationFuture>();

private final Queue<AcceptorOperationFuture> cancelQueue = new ConcurrentLinkedQueue<AcceptorOperationFuture>();

private final Map<SocketAddress, H> boundHandles = Collections.synchronizedMap(new HashMap<SocketAddress, H>());

private final ServiceOperationFuture disposalFuture = new ServiceOperationFuture();

/** A flag set when the acceptor has been created and initialized */
private volatile boolean selectable;

/** The thread responsible of accepting incoming requests */
private AtomicReference<Acceptor> acceptorRef = new AtomicReference<Acceptor>();

protected boolean reuseAddress = false;

/**
 * Define the number of socket that can wait to be accepted. Default
 * to 50 (as in the SocketServer default).
 */
protected int backlog = 50;

/** A lock used to protect the selector to be waked up before it's created */
    private final Semaphore lock = new Semaphore(1);

    private final IoProcessor<S> processor;

    private final boolean createdProcessor;

    private final Queue<AcceptorOperationFuture> registerQueue = new ConcurrentLinkedQueue<AcceptorOperationFuture>();

    private final Queue<AcceptorOperationFuture> cancelQueue = new ConcurrentLinkedQueue<AcceptorOperationFuture>();

    private final Map<SocketAddress, H> boundHandles = Collections.synchronizedMap(new HashMap<SocketAddress, H>());

    private final ServiceOperationFuture disposalFuture = new ServiceOperationFuture();

    /** A flag set when the acceptor has been created and initialized */
    private volatile boolean selectable;

    /** The thread responsible of accepting incoming requests */
    private AtomicReference<Acceptor> acceptorRef = new AtomicReference<Acceptor>();

    protected boolean reuseAddress = false;

    /**
     * Define the number of socket that can wait to be accepted. Default
     * to 50 (as in the SocketServer default).
     */
    protected int backlog = 50;

这个类讲下去确实有种无从下手的感觉，这里注意一个acceptorRef是处理请求的AtomicReference类型的Acceptor，Acceptor是一个用于接收客户端请求，里面代码比较多，但是我感觉此部分是整个类比较关键的实现部分，需要仔细读一下，但是实在不太好讲，这里可以同样只关注整体，不关注细节，包括processor一些方法的内部完成。我只总结一个流程：

 

 

主要的两个方法把父类中的注释写一下，方便理解，看了注释再看代码就比较容易想明白了：

 

bindInternal：Starts the acceptor, and register the given addresses

unbind0：Implement this method to perform the actual unbind operation.