BIO到NIO源码的一些事儿之NIO 下 之 Selector

3,025 阅读28分钟

前言

此系列文章会详细解读NIO的功能逐步丰满的路程,为Reactor-Netty 库的讲解铺平道路。

关于Java编程方法论-Reactor与Webflux的视频分享,已经完成了Rxjava 与 Reactor,b站地址如下:

Rxjava源码解读与分享:www.bilibili.com/video/av345…

Reactor源码解读与分享:www.bilibili.com/video/av353…

本系列源码解读基于JDK11 api细节可能与其他版本有所差别,请自行解决jdk版本问题。

本系列前几篇:

BIO到NIO源码的一些事儿之BIO

BIO到NIO源码的一些事儿之NIO 上

BIO到NIO源码的一些事儿之NIO 中

SelectionKey的引入

如我们在前面内容所讲,在学生确定之后,我们就要对其状态进行设定,然后再交由Selector进行管理,其状态的设定我们就通过SelectionKey来进行。

那这里我们先通过之前在Channel中并未仔细讲解的SelectableChannel下的register方法。我们前面有提到过, SelectableChannelchannel打造成可以通过Selector来进行多路复用。作为管理者,channel想要实现复用,就必须在管理者这里进行注册登记。所以,SelectableChannel下的register方法也就是我们值得二次关注的核心了,也是对接我们接下来内容的切入点,对于register方法的解读,请看我们之前的文章BIO到NIO源码的一些事儿之NIO 上赋予Channel可被多路复用的能力这一节的内容。

这里要记住的是SelectableChannel是对接channel特征(即SelectionKey)的关键所在,这有点类似于表设计,原本可以将特征什么的设定在一张表内,但为了操作更加具有针对性,即为了让代码功能更易于管理,就进行抽取并设计了第二张表,这个就有点像人体器官,整体上大家共同协作完成一件事,但器官内部自己专注于自己的主要特定功能,偶尔也具备其他器官的一些小功能。

由此,我们也就可以知道,SelectionKey表示一个SelectableChannelSelector关联的标记,可以简单理解为一个token。就好比是我们做权限管理系统用户登录后前台会从后台拿到的一个token一样,用户可以凭借此token来访问操作相应的资源信息。

//java.nio.channels.spi.AbstractSelectableChannel#register
public final SelectionKey register(Selector sel, int ops, Object att)
    throws ClosedChannelException
{       ...
    synchronized (regLock) {
       ...
        synchronized (keyLock) {
           ...
            SelectionKey k = findKey(sel);
            if (k != null) {
                k.attach(att);
                k.interestOps(ops);
            } else {
                // New registration
                k = ((AbstractSelector)sel).register(this, ops, att);
                addKey(k);
            }
            return k;
        }
    }
}

结合上下两段源码,在每次Selector使用register方法注册channel时,都会创建并返回一个SelectionKey

//sun.nio.ch.SelectorImpl#register
@Override
protected final SelectionKey register(AbstractSelectableChannel ch,
                                        int ops,
                                        Object attachment)
{
    if (!(ch instanceof SelChImpl))
        throw new IllegalSelectorException();
    SelectionKeyImpl k = new SelectionKeyImpl((SelChImpl)ch, this);
    k.attach(attachment);

    // register (if needed) before adding to key set
    implRegister(k);

    // add to the selector's key set, removing it immediately if the selector
    // is closed. The key is not in the channel's key set at this point but
    // it may be observed by a thread iterating over the selector's key set.
    keys.add(k);
    try {
        k.interestOps(ops);
    } catch (ClosedSelectorException e) {
        assert ch.keyFor(this) == null;
        keys.remove(k);
        k.cancel();
        throw e;
    }
    return k;
}

我们在BIO到NIO源码的一些事儿之NIO 上赋予Channel可被多路复用的能力这一节的内容知道,一旦注册到Selector上,Channel将一直保持注册直到其被解除注册。在解除注册的时候会解除Selector分配给Channel的所有资源。 也就是SelectionKey在其调用SelectionKey#channel方法,或这个key所代表的channel 关闭,抑或此key所关联的Selector关闭之前,都是有效。我们在前面的文章分析中也知道,取消一个SelectionKey,不会立刻从Selector移除,它将被添加到SelectorcancelledKeys这个Set集合中,以便在下一次选择操作期间删除,我们可以通过java.nio.channels.SelectionKey#isValid判断一个SelectionKey是否有效。

SelectionKey包含四个操作集,每个操作集用一个Int来表示,int值中的低四位的bit 用于表示channel支持的可选操作种类。


   /**
    * Operation-set bit for read operations.
    */
   public static final int OP_READ = 1 << 0;

   /**
    * Operation-set bit for write operations.
    */
   public static final int OP_WRITE = 1 << 2;

   /**
    * Operation-set bit for socket-connect operations.
    */
   public static final int OP_CONNECT = 1 << 3;

   /**
    * Operation-set bit for socket-accept operations.
    */
   public static final int OP_ACCEPT = 1 << 4;

interestOps

通过interestOps来确定了selector在下一个选择操作的过程中将测试哪些操作类别的准备情况,操作事件是否是channel关注的。interestOpsSelectionKey创建时,初始化为注册Selector时的ops值,这个值可通过sun.nio.ch.SelectionKeyImpl#interestOps(int)来改变,这点我们在SelectorImpl#register可以清楚的看到。

//sun.nio.ch.SelectionKeyImpl
public final class SelectionKeyImpl
   extends AbstractSelectionKey
{
   private static final VarHandle INTERESTOPS =
           ConstantBootstraps.fieldVarHandle(
                   MethodHandles.lookup(),
                   "interestOps",
                   VarHandle.class,
                   SelectionKeyImpl.class, int.class);

   private final SelChImpl channel;
   private final SelectorImpl selector;

   private volatile int interestOps;
   private volatile int readyOps;

   // registered events in kernel, used by some Selector implementations
   private int registeredEvents;

   // index of key in pollfd array, used by some Selector implementations
   private int index;

   SelectionKeyImpl(SelChImpl ch, SelectorImpl sel) {
       channel = ch;
       selector = sel;
   }
  ...
}

readyOps

readyOps表示通过Selector检测到channel已经准备就绪的操作事件。在SelectionKey创建时(即上面源码所示),readyOps值为0,在Selectorselect操作中可能会更新,但是需要注意的是我们不能直接调用来更新。

SelectionKeyreadyOps表示一个channel已经为某些操作准备就绪,但不能保证在针对这个就绪事件类型的操作过程中不会发生阻塞,即该操作所在线程有可能会发生阻塞。在完成select操作后,大部分情况下会立即对readyOps更新,此时readyOps值最准确,如果外部的事件或在该channel有IO操作,readyOps可能不准确。所以,我们有看到其是volatile类型。

SelectionKey定义了所有的操作事件,但是具体channel支持的操作事件依赖于具体的channel,即具体问题具体分析。 所有可选择的channel(即SelectableChannel的子类)都可以通过SelectableChannel#validOps方法,判断一个操作事件是否被channel所支持,即每个子类都会有对validOps的实现,返回一个数字,仅标识channel支持的哪些操作。尝试设置或测试一个不被channel所支持的操作设定,将会抛出相关的运行时异常。 不同应用场景下,其所支持的Ops是不同的,摘取部分如下所示:

//java.nio.channels.SocketChannel#validOps
public final int validOps() {
    //即1|4|8  1101
    return (SelectionKey.OP_READ
            | SelectionKey.OP_WRITE
            | SelectionKey.OP_CONNECT);
}
//java.nio.channels.ServerSocketChannel#validOps
public final int validOps() {
    // 16
    return SelectionKey.OP_ACCEPT;
}
//java.nio.channels.DatagramChannel#validOps
public final int validOps() {
    // 1|4
    return (SelectionKey.OP_READ
            | SelectionKey.OP_WRITE);
}

如果需要经常关联一些我们程序中指定数据到SelectionKey,比如一个我们使用一个object表示上层的一种高级协议的状态,object用于通知实现协议处理器。所以,SelectionKey支持通过attach方法将一个对象附加到SelectionKeyattachment上。attachment可以通过java.nio.channels.SelectionKey#attachment方法进行访问。如果要取消该对象,则可以通过该种方式:selectionKey.attach(null)

需要注意的是如果附加的对象不再使用,一定要人为清除,如果没有,假如此SelectionKey一直存在,由于此处属于强引用,那么垃圾回收器不会回收该对象,若不清除的话会成内存泄漏。

SelectionKey在由多线程并发使用时,是线程安全的。我们只需要知道,Selectorselect操作会一直使用在调用该操作开始时当前的interestOps所设定的值。

Selector探究

到现在为止,我们已经多多少少接触了Selector,其是一个什么样的角色,想必都很清楚了,那我们就在我们已经接触到的来进一步深入探究Selector的设计运行机制。

Selector的open方法

从命名上就可以知道 SelectableChannel对象是依靠Selector来实现多路复用的。 我们可以通过调用java.nio.channels.Selector#open来创建一个selector对象:

//java.nio.channels.Selector#open
public static Selector open() throws IOException {
    return SelectorProvider.provider().openSelector();
}

关于这个SelectorProvider.provider(),其使用了根据所在系统的默认实现,我这里是windows系统,那么其默认实现为sun.nio.ch.WindowsSelectorProvider,这样,就可以调用基于相应系统的具体实现了。

//java.nio.channels.spi.SelectorProvider#provider
public static SelectorProvider provider() {
    synchronized (lock) {
        if (provider != null)
            return provider;
        return AccessController.doPrivileged(
            new PrivilegedAction<>() {
                public SelectorProvider run() {
                        if (loadProviderFromProperty())
                            return provider;
                        if (loadProviderAsService())
                            return provider;
                        provider = sun.nio.ch.DefaultSelectorProvider.create();
                        return provider;
                    }
                });
    }
}
//sun.nio.ch.DefaultSelectorProvider
public class DefaultSelectorProvider {

/**
 * Prevent instantiation.
 */
private DefaultSelectorProvider() { }

/**
 * Returns the default SelectorProvider.
 */
public static SelectorProvider create() {
    return new sun.nio.ch.WindowsSelectorProvider();
}

}

基于windows来讲,selector这里最终会使用sun.nio.ch.WindowsSelectorImpl来做一些核心的逻辑。

public class WindowsSelectorProvider extends SelectorProviderImpl {

    public AbstractSelector openSelector() throws IOException {
        return new WindowsSelectorImpl(this);
    }
}

这里,我们需要来看一下WindowsSelectorImpl的构造函数:

//sun.nio.ch.WindowsSelectorImpl#WindowsSelectorImpl
WindowsSelectorImpl(SelectorProvider sp) throws IOException {
    super(sp);
    pollWrapper = new PollArrayWrapper(INIT_CAP);
    wakeupPipe = Pipe.open();
    wakeupSourceFd = ((SelChImpl)wakeupPipe.source()).getFDVal();

    // Disable the Nagle algorithm so that the wakeup is more immediate
    SinkChannelImpl sink = (SinkChannelImpl)wakeupPipe.sink();
    (sink.sc).socket().setTcpNoDelay(true);
    wakeupSinkFd = ((SelChImpl)sink).getFDVal();

    pollWrapper.addWakeupSocket(wakeupSourceFd, 0);
}

我们由Pipe.open()就可知道selector会保持打开的状态,直到其调用它的close方法:

//java.nio.channels.spi.AbstractSelector#close
public final void close() throws IOException {
    boolean open = selectorOpen.getAndSet(false);
    if (!open)
        return;
    implCloseSelector();
}
//sun.nio.ch.SelectorImpl#implCloseSelector
@Override
public final void implCloseSelector() throws IOException {
    wakeup();
    synchronized (this) {
        implClose();
        synchronized (publicSelectedKeys) {
            // Deregister channels
            Iterator<SelectionKey> i = keys.iterator();
            while (i.hasNext()) {
                SelectionKeyImpl ski = (SelectionKeyImpl)i.next();
                deregister(ski);
                SelectableChannel selch = ski.channel();
                if (!selch.isOpen() && !selch.isRegistered())
                    ((SelChImpl)selch).kill();
                selectedKeys.remove(ski);
                i.remove();
            }
            assert selectedKeys.isEmpty() && keys.isEmpty();
        }
    }
}
//sun.nio.ch.WindowsSelectorImpl#implClose
@Override
protected void implClose() throws IOException {
    assert !isOpen();
    assert Thread.holdsLock(this);

    // prevent further wakeup
    synchronized (interruptLock) {
        interruptTriggered = true;
    }

    wakeupPipe.sink().close();
    wakeupPipe.source().close();
    pollWrapper.free();

    // Make all remaining helper threads exit
    for (SelectThread t: threads)
            t.makeZombie();
    startLock.startThreads();
}

可以看到,前面的wakeupPipe在close方法中关闭掉了。这里的close方法中又涉及了wakeupPipe.sink()wakeupPipe.source()的关闭与pollWrapper.free()的释放,此处也是我们本篇的难点所在,这里,我们来看看它们到底是什么样的存在。 首先,我们对WindowsSelectorImpl(SelectorProvider sp)这个构造函数做下梳理:

  • 创建一个PollArrayWrapper对象(pollWrapper);
  • Pipe.open()打开一个管道;
  • 拿到wakeupSourceFdwakeupSinkFd两个文件描述符;
  • 把pipe内Source端的文件描述符(wakeupSourceFd)放到pollWrapper里;

Pipe.open()的解惑

这里我们会有疑惑,为什么要创建一个管道,它是用来做什么的。

我们来看Pipe.open()源码实现:

//java.nio.channels.Pipe#open
public static Pipe open() throws IOException {
    return SelectorProvider.provider().openPipe();
}
//sun.nio.ch.SelectorProviderImpl#openPipe
public Pipe openPipe() throws IOException {
    return new PipeImpl(this);
}
//sun.nio.ch.PipeImpl#PipeImpl
PipeImpl(final SelectorProvider sp) throws IOException {
    try {
        AccessController.doPrivileged(new Initializer(sp));
    } catch (PrivilegedActionException x) {
        throw (IOException)x.getCause();
    }
}
private class Initializer
implements PrivilegedExceptionAction<Void>
{

private final SelectorProvider sp;

private IOException ioe = null;

private Initializer(SelectorProvider sp) {
    this.sp = sp;
}

@Override
public Void run() throws IOException {
    LoopbackConnector connector = new LoopbackConnector();
    connector.run();
    if (ioe instanceof ClosedByInterruptException) {
        ioe = null;
        Thread connThread = new Thread(connector) {
            @Override
            public void interrupt() {}
        };
        connThread.start();
        for (;;) {
            try {
                connThread.join();
                break;
            } catch (InterruptedException ex) {}
        }
        Thread.currentThread().interrupt();
    }

    if (ioe != null)
        throw new IOException("Unable to establish loopback connection", ioe);

    return null;
}

从上述源码我们可以知道,创建了一个PipeImpl对象, 在PipeImpl的构造函数里会执行AccessController.doPrivileged,在它调用后紧接着会执行Initializerrun方法:

//sun.nio.ch.PipeImpl.Initializer.LoopbackConnector
private class LoopbackConnector implements Runnable {

    @Override
    public void run() {
        ServerSocketChannel ssc = null;
        SocketChannel sc1 = null;
        SocketChannel sc2 = null;

        try {
            // Create secret with a backing array.
            ByteBuffer secret = ByteBuffer.allocate(NUM_SECRET_BYTES);
            ByteBuffer bb = ByteBuffer.allocate(NUM_SECRET_BYTES);

            // Loopback address
            InetAddress lb = InetAddress.getLoopbackAddress();
            assert(lb.isLoopbackAddress());
            InetSocketAddress sa = null;
            for(;;) {
                // Bind ServerSocketChannel to a port on the loopback
                // address
                if (ssc == null || !ssc.isOpen()) {
                    ssc = ServerSocketChannel.open();
                    ssc.socket().bind(new InetSocketAddress(lb, 0));
                    sa = new InetSocketAddress(lb, ssc.socket().getLocalPort());
                }

                // Establish connection (assume connections are eagerly
                // accepted)
                sc1 = SocketChannel.open(sa);
                RANDOM_NUMBER_GENERATOR.nextBytes(secret.array());
                do {
                    sc1.write(secret);
                } while (secret.hasRemaining());
                secret.rewind();

                // Get a connection and verify it is legitimate
                sc2 = ssc.accept();
                do {
                    sc2.read(bb);
                } while (bb.hasRemaining());
                bb.rewind();

                if (bb.equals(secret))
                    break;

                sc2.close();
                sc1.close();
            }

            // Create source and sink channels
            source = new SourceChannelImpl(sp, sc1);
            sink = new SinkChannelImpl(sp, sc2);
        } catch (IOException e) {
            try {
                if (sc1 != null)
                    sc1.close();
                if (sc2 != null)
                    sc2.close();
            } catch (IOException e2) {}
            ioe = e;
        } finally {
            try {
                if (ssc != null)
                    ssc.close();
            } catch (IOException e2) {}
        }
    }
}
}

这里即为创建pipe的过程,windows下的实现是创建两个本地的socketChannel,然后连接(连接的过程通过写一个随机数据做两个socket的连接校验),两个socketChannel分别实现了管道pipesourcesink端。 而我们依然不清楚这个pipe到底干什么用的, 假如大家熟悉系统调用的C/C++的话,就可以知道,一个阻塞在select上的线程有以下三种方式可以被唤醒:

  1. 有数据可读/写,或出现异常。
  2. 阻塞时间到,即time out
  3. 收到一个non-block的信号。可由killpthread_kill发出。

所以,Selector.wakeup()要唤醒阻塞的select,那么也只能通过这三种方法,其中:

  • 第二种方法可以排除,因为select一旦阻塞,无法修改其time out时间。
  • 而第三种看来只能在Linux上实现,Windows上没有这种信号通知的机制。

看来只有第一种方法了。假如我们多次调用Selector.open(),那么在Windows上会每调用一次,就会建立一对自己和自己的loopbackTCP连接;在Linux上的话,每调用一次,会开一对pipe(pipe在Linux下一般都成对打开),到这里,估计我们能够猜得出来——那就是如果想要唤醒select,只需要朝着自己的这个loopback连接发点数据过去,于是,就可以唤醒阻塞在select上的线程了。

我们对上面所述做下总结:在Windows下,Java虚拟机在Selector.open()时会自己和自己建立loopbackTCP连接;在Linux下,Selector会创建pipe。这主要是为了Selector.wakeup()可以方便唤醒阻塞在select()系统调用上的线程(通过向自己所建立的TCP链接和管道上随便写点什么就可以唤醒阻塞线程)。

PollArrayWrapper解读

WindowsSelectorImpl构造器最后,我们看到这一句代码:pollWrapper.addWakeupSocket(wakeupSourceFd, 0);,即把pipe内Source端的文件描述符(wakeupSourceFd)放到pollWrapper里。pollWrapper作为PollArrayWrapper的实例,它到底是什么,这一节,我们就来对其探索一番。

class PollArrayWrapper {

    private AllocatedNativeObject pollArray; // The fd array

    long pollArrayAddress; // pollArrayAddress

    @Native private static final short FD_OFFSET     = 0; // fd offset in pollfd
    @Native private static final short EVENT_OFFSET  = 4; // events offset in pollfd

    static short SIZE_POLLFD = 8; // sizeof pollfd struct

    private int size; // Size of the pollArray

    PollArrayWrapper(int newSize) {
        int allocationSize = newSize * SIZE_POLLFD;
        pollArray = new AllocatedNativeObject(allocationSize, true);
        pollArrayAddress = pollArray.address();
        this.size = newSize;
    }

    ...

    // Access methods for fd structures
    void putDescriptor(int i, int fd) {
        pollArray.putInt(SIZE_POLLFD * i + FD_OFFSET, fd);
    }

    void putEventOps(int i, int event) {
        pollArray.putShort(SIZE_POLLFD * i + EVENT_OFFSET, (short)event);
    }
    ...
   // Adds Windows wakeup socket at a given index.
    void addWakeupSocket(int fdVal, int index) {
        putDescriptor(index, fdVal);
        putEventOps(index, Net.POLLIN);
    }
}

这里将wakeupSourceFdPOLLIN事件标识为pollArrayEventOps的对应的值,这里使用的是unsafe直接操作的内存,也就是相对于这个pollArray所在内存地址的偏移量SIZE_POLLFD * i + EVENT_OFFSET这个位置上写入Net.POLLIN所代表的值,即参考下面本地方法相关源码所展示的值。putDescriptor同样是这种类似操作。当sink端有数据写入时,source对应的文件描述符wakeupSourceFd就会处于就绪状态。

//java.base/windows/native/libnio/ch/nio_util.h
    /* WSAPoll()/WSAPOLLFD and the corresponding constants are only defined   */
    /* in Windows Vista / Windows Server 2008 and later. If we are on an      */
    /* older release we just use the Solaris constants as this was previously */
    /* done in PollArrayWrapper.java.                                         */
    #define POLLIN       0x0001
    #define POLLOUT      0x0004
    #define POLLERR      0x0008
    #define POLLHUP      0x0010
    #define POLLNVAL     0x0020
    #define POLLCONN     0x0002

AllocatedNativeObject这个类的父类有大量的unsafe类的操作,这些都是直接基于内存级别的操作。从其父类的构造器中,我们能也清楚的看到pollArray是通过unsafe.allocateMemory(size + ps)分配的一块系统内存。

class AllocatedNativeObject                             // package-private
    extends NativeObject
{
    /**
     * Allocates a memory area of at least {@code size} bytes outside of the
     * Java heap and creates a native object for that area.
     */
    AllocatedNativeObject(int size, boolean pageAligned) {
        super(size, pageAligned);
    }

    /**
     * Frees the native memory area associated with this object.
     */
    synchronized void free() {
        if (allocationAddress != 0) {
            unsafe.freeMemory(allocationAddress);
            allocationAddress = 0;
        }
    }

}
//sun.nio.ch.NativeObject#NativeObject(int, boolean)
protected NativeObject(int size, boolean pageAligned) {
        if (!pageAligned) {
            this.allocationAddress = unsafe.allocateMemory(size);
            this.address = this.allocationAddress;
        } else {
            int ps = pageSize();
            long a = unsafe.allocateMemory(size + ps);
            this.allocationAddress = a;
            this.address = a + ps - (a & (ps - 1));
        }
    }

至此,我们算是完成了对Selector.open()的解读,其主要任务就是完成建立Pipe,并把pipe source端的wakeupSourceFd放入pollArray中,这个pollArraySelector完成其角色任务的枢纽。本篇主要围绕Windows的实现来进行分析,即在windows下通过两个连接的socketChannel实现了Pipelinux下则直接使用系统的pipe即可。

SelectionKey在selector中的管理

SelectionKey在selector中注册

所谓的注册,其实就是将一个对象放到注册地对象内的一个容器字段上,这个字段可以是数组,队列,也可以是一个set集合,也可以是一个list。这里,同样是这样,只不过,其需要有个返回值,那么把这个要放入集合的对象返回即可。

//sun.nio.ch.SelectorImpl#register
@Override
protected final SelectionKey register(AbstractSelectableChannel ch,
                                        int ops,
                                        Object attachment)
{
    if (!(ch instanceof SelChImpl))
        throw new IllegalSelectorException();
    SelectionKeyImpl k = new SelectionKeyImpl((SelChImpl)ch, this);
    k.attach(attachment);

    // register (if needed) before adding to key set
    implRegister(k);

    // add to the selector's key set, removing it immediately if the selector
    // is closed. The key is not in the channel's key set at this point but
    // it may be observed by a thread iterating over the selector's key set.
    keys.add(k);
    try {
        k.interestOps(ops);
    } catch (ClosedSelectorException e) {
        assert ch.keyFor(this) == null;
        keys.remove(k);
        k.cancel();
        throw e;
    }
    return k;
}
//sun.nio.ch.WindowsSelectorImpl#implRegister
@Override
protected void implRegister(SelectionKeyImpl ski) {
    ensureOpen();
    synchronized (updateLock) {
        newKeys.addLast(ski);
    }
}

这段代码我们之前已经有看过,这里我们再次温习下。 首先会新建一个SelectionKeyImpl对象,这个对象就是对Channel的包装,不仅如此,还顺带把当前这个Selector对象给收了进去,这样,我们也可以通过SelectionKey的对象来拿到其对应的Selector对象。

接着,基于windows平台实现的implRegister,先通过ensureOpen()来确保该Selector是打开的。接着将这个SelectionKeyImpl加入到WindowsSelectorImpl内针对于新注册SelectionKey进行管理的newKeys之中,newKeys是一个ArrayDeque对象。对于ArrayDeque有不懂的,可以参考Java 容器源码分析之 Deque 与 ArrayDeque这篇文章。

然后再将此这个SelectionKeyImpl加入到sun.nio.ch.SelectorImpl#keys中去,这个Set<SelectionKey>集合代表那些已经注册到当前这个Selector对象上的SelectionKey集合。我们来看sun.nio.ch.SelectorImpl的构造函数:

//sun.nio.ch.SelectorImpl#SelectorImpl
protected SelectorImpl(SelectorProvider sp) {
    super(sp);
    keys = ConcurrentHashMap.newKeySet();
    selectedKeys = new HashSet<>();
    publicKeys = Collections.unmodifiableSet(keys);
    publicSelectedKeys = Util.ungrowableSet(selectedKeys);
}

也就是说,这里的publicKeys就来源于keys,只是publicKeys属于只读的,我们想要知道当前Selector对象上所注册的keys,就可以调用sun.nio.ch.SelectorImpl#keys来得到:

//sun.nio.ch.SelectorImpl#keys
@Override
public final Set<SelectionKey> keys() {
    ensureOpen();
    return publicKeys;
}

再回到这个构造函数中,selectedKeys,顾名思义,其属于已选择Keys,即前一次操作期间,已经准备就绪的Channel所对应的SelectionKey。此集合为keys的子集。通过selector.selectedKeys()获取。

//sun.nio.ch.SelectorImpl#selectedKeys
@Override
public final Set<SelectionKey> selectedKeys() {
    ensureOpen();
    return publicSelectedKeys;
}

我们看到其返回的是publicSelectedKeys,针对这个字段里的元素操作可以做删除,但不能做增加。 在前面的内容中,我们有涉及到SelectionKey的取消,所以,我们在java.nio.channels.spi.AbstractSelector方法内,是有定义cancelledKeys的,也是一个HashSet对象。其代表已经被取消但尚未取消注册(deregister)的SelectionKey。此Set集合无法直接访问,同样,它也是keys()的子集。

对于新的Selector实例,上面几个集合均为空。由上面展示的源码可知,通过channel.registerSelectionKey添加keys中,此为key的来源。 如果某个selectionKey.cancel()被调用,那么此key将会被添加到cancelledKeys这个集合中,然后在下一次调用selector select方法期间,此时canceldKeys不为空,将会触发此SelectionKeyderegister操作(释放资源,并从keys中移除)。无论通过channel.close()还是通过selectionKey.cancel(),都会导致SelectionKey被加入到cannceldKey中.

每次选择操作(select)期间,都可以将key添加到selectedKeys中或者将从cancelledKeys中移除。

Selector的select方法的解读

了解了上面的这些,我们来进入到select方法中,观察下它的细节。由Selector的api可知,select操作有两种形式,一种为 select(),selectNow(),select(long timeout);另一种为select(Consumer<SelectionKey> action, long timeout)select(Consumer<SelectionKey> action)selectNow(Consumer<SelectionKey> action)。后者为JDK11新加入的api,主要针对那些准备好进行I/O操作的channels在select过程中对相应的key进行的一个自定义操作。

需要注意的是,有Consumer<SelectionKey> action参数的select操作是阻塞的,只有在选择了至少一个Channel的情况下,才会调用此Selector实例的wakeup方法来唤醒,同样,其所在线程被打断也可以。

//sun.nio.ch.SelectorImpl
@Override
public final int select(long timeout) throws IOException {
    if (timeout < 0)
        throw new IllegalArgumentException("Negative timeout");
    return lockAndDoSelect(null, (timeout == 0) ? -1 : timeout);
}

//sun.nio.ch.SelectorImpl
@Override
public final int select(Consumer<SelectionKey> action, long timeout)
    throws IOException
{
    Objects.requireNonNull(action);
    if (timeout < 0)
        throw new IllegalArgumentException("Negative timeout");
    return lockAndDoSelect(action, (timeout == 0) ? -1 : timeout);
}
//sun.nio.ch.SelectorImpl#lockAndDoSelect
private int lockAndDoSelect(Consumer<SelectionKey> action, long timeout)
        throws IOException
    {
        synchronized (this) {
            ensureOpen();
            if (inSelect)
                throw new IllegalStateException("select in progress");
            inSelect = true;
            try {
                synchronized (publicSelectedKeys) {
                    return doSelect(action, timeout);
                }
            } finally {
                inSelect = false;
            }
        }
    }

我们可以观察,无论哪种,它们最后都落在了lockAndDoSelect这个方法上,最终会执行特定系统上的doSelect(action, timeout)实现。 这里我们以sun.nio.ch.WindowsSelectorImpl#doSelect为例来讲述其操作执行的步骤:

// sun.nio.ch.WindowsSelectorImpl#doSelect
@Override
protected int doSelect(Consumer<SelectionKey> action, long timeout)
    throws IOException
    {
        assert Thread.holdsLock(this);
        this.timeout = timeout; // set selector timeout
        processUpdateQueue();  // <1>
        processDeregisterQueue(); // <2>
        if (interruptTriggered) {
            resetWakeupSocket();
            return 0;
        }
        // Calculate number of helper threads needed for poll. If necessary
        // threads are created here and start waiting on startLock
        adjustThreadsCount();
        finishLock.reset(); // reset finishLock
        // Wakeup helper threads, waiting on startLock, so they start polling.
        // Redundant threads will exit here after wakeup.
        startLock.startThreads();
        // do polling in the main thread. Main thread is responsible for
        // first MAX_SELECTABLE_FDS entries in pollArray.
        try {
            begin();
            try {
                subSelector.poll();  // <3>
            } catch (IOException e) {
                finishLock.setException(e); // Save this exception
            }
            // Main thread is out of poll(). Wakeup others and wait for them
            if (threads.size() > 0)
                finishLock.waitForHelperThreads();
          } finally {
              end();
          }
        // Done with poll(). Set wakeupSocket to nonsignaled  for the next run.
        finishLock.checkForException();
        processDeregisterQueue();  // <4>
        int updated = updateSelectedKeys(action); // <5>
        // Done with poll(). Set wakeupSocket to nonsignaled  for the next run.
        resetWakeupSocket(); // <6>
        return updated;
    }

processUpdateQueue解读

  1. 首先通过相应操作系统实现类(此处是WindowsSelectorImpl)的具体实现我们可以知道,通过<1> 处的 processUpdateQueue()获得关于每个剩余Channel(有些Channel取消了)的在此刻的interestOps,这里包括新注册的和updateKeys,并对其进行pollWrapper的管理操作。

    • 即对于新注册的SelectionKeyImpl,我们在相对于这个pollArray所在内存地址的偏移量SIZE_POLLFD * totalChannels + FD_OFFSETSIZE_POLLFD * totalChannels + EVENT_OFFSET分别存入SelectionKeyImpl的文件描述符fd与其对应的EventOps(初始为0)。

    • updateKeys,因为是其之前已经在pollArray的某个相对位置上存储过,这里我们还需要对拿到的key的有效性进行判断,如果有效,只需要将正在操作的这个SelectionKeyImpl对象的interestOps写入到在pollWrapper中的存放它的EventOps位置上。

    注意: 在对newKeys进行key的有效性判断之后,如果有效,会调用growIfNeeded()方法,这里首先会判断channelArray.length == totalChannels,此为一个SelectionKeyImpl的数组,初始容量大小为8。channelArray其实就是方便Selector管理在册SelectionKeyImpl数量的一个数组而已,通过判断它的数组长度大小,如果和totalChannels(初始值为1)相等,不仅仅是为了channelArray扩容,更重要的是为了辅助pollWrapper,让pollWrapper扩容才是目的所在。

    而当totalChannels % MAX_SELECTABLE_FDS == 0时,则多开一个线程处理selectorwindowsselect系统调用有最大文件描述符限制,一次只能轮询1024个文件描述符,如果多于1024个,需要多线程进行轮询。同时调用pollWrapper.addWakeupSocket(wakeupSourceFd, totalChannels)在相对于这个pollArray所在内存地址的偏移量SIZE_POLLFD * totalChannels + FD_OFFSET这个位置上写入wakeupSourceFd所代表的fdVal值。这样在新起的线程就可以通过MAX_SELECTABLE_FDS来确定这个用来监控的wakeupSourceFd,方便唤醒selector。通过ski.setIndex(totalChannels)记录下SelectionKeyImpl在数组中的索引位置,以待后续使用。

   /**
    * sun.nio.ch.WindowsSelectorImpl#processUpdateQueue
    * Process new registrations and changes to the interest ops.
    */
private void processUpdateQueue() {
    assert Thread.holdsLock(this);

    synchronized (updateLock) {
        SelectionKeyImpl ski;

        // new registrations
        while ((ski = newKeys.pollFirst()) != null) {
            if (ski.isValid()) {
                growIfNeeded();
                channelArray[totalChannels] = ski;
                ski.setIndex(totalChannels);
                pollWrapper.putEntry(totalChannels, ski);
                totalChannels++;
                MapEntry previous = fdMap.put(ski);
                assert previous == null;
            }
        }

        // changes to interest ops
        while ((ski = updateKeys.pollFirst()) != null) {
            int events = ski.translateInterestOps();
            int fd = ski.getFDVal();
            if (ski.isValid() && fdMap.containsKey(fd)) {
                int index = ski.getIndex();
                assert index >= 0 && index < totalChannels;
                pollWrapper.putEventOps(index, events);
            }
        }
    }
}

//sun.nio.ch.PollArrayWrapper#putEntry
// Prepare another pollfd struct for use.
void putEntry(int index, SelectionKeyImpl ski) {
    putDescriptor(index, ski.getFDVal());
    putEventOps(index, 0);
}
//sun.nio.ch.WindowsSelectorImpl#growIfNeeded
private void growIfNeeded() {
    if (channelArray.length == totalChannels) {
        int newSize = totalChannels * 2; // Make a larger array
        SelectionKeyImpl temp[] = new SelectionKeyImpl[newSize];
        System.arraycopy(channelArray, 1, temp, 1, totalChannels - 1);
        channelArray = temp;
        pollWrapper.grow(newSize);
    }
    if (totalChannels % MAX_SELECTABLE_FDS == 0) { // more threads needed
        pollWrapper.addWakeupSocket(wakeupSourceFd, totalChannels);
        totalChannels++;
        threadsCount++;
    }
}
// Initial capacity of the poll array
private final int INIT_CAP = 8;
// Maximum number of sockets for select().
// Should be INIT_CAP times a power of 2
private static final int MAX_SELECTABLE_FDS = 1024;

// The list of SelectableChannels serviced by this Selector. Every mod
// MAX_SELECTABLE_FDS entry is bogus, to align this array with the poll
// array,  where the corresponding entry is occupied by the wakeupSocket
private SelectionKeyImpl[] channelArray = new SelectionKeyImpl[INIT_CAP];
// The number of valid entries in  poll array, including entries occupied
// by wakeup socket handle.
private int totalChannels = 1;

//sun.nio.ch.PollArrayWrapper#grow
// Grows the pollfd array to new size
void grow(int newSize) {
    PollArrayWrapper temp = new PollArrayWrapper(newSize);
    for (int i = 0; i < size; i++)
        replaceEntry(this, i, temp, i);
    pollArray.free();
    pollArray = temp.pollArray;
    this.size = temp.size;
    pollArrayAddress = pollArray.address();
}

// Maps file descriptors to their indices in  pollArray
private static final class FdMap extends HashMap<Integer, MapEntry> {
    static final long serialVersionUID = 0L;
    private MapEntry get(int desc) {
        return get(Integer.valueOf(desc));
    }
    private MapEntry put(SelectionKeyImpl ski) {
        return put(Integer.valueOf(ski.getFDVal()), new MapEntry(ski));
    }
    private MapEntry remove(SelectionKeyImpl ski) {
        Integer fd = Integer.valueOf(ski.getFDVal());
        MapEntry x = get(fd);
        if ((x != null) && (x.ski.channel() == ski.channel()))
            return remove(fd);
        return null;
    }
}

// class for fdMap entries
private static final class MapEntry {
    final SelectionKeyImpl ski;
    long updateCount = 0;
    MapEntry(SelectionKeyImpl ski) {
        this.ski = ski;
    }
}
private final FdMap fdMap = new FdMap();
processDeregisterQueue解读
  1. 接着通过上面WindowsSelectorImpl#doSelect展示源码中<2> 处的 processDeregisterQueue()
    • cancelledKeys进行清除,遍历cancelledKeys,并对每个key进行deregister操作,然后从cancelledKeys集合中删除,从keys集合与selectedKeys中删除,以此来释放引用,方便gc回收,
    • 其内调用implDereg方法,将会从channelArray中移除对应的Channel代表的SelectionKeyImpl,调整totalChannels和线程数,从mapkeys中移除SelectionKeyImpl,移除Channel上的SelectionKeyImpl并关闭Channel
    • 同时还发现该processDeregisterQueue()方法在调用poll方法前后都进行调用,这是确保能够正确处理在调用poll方法阻塞的这一段时间之内取消的键能被及时清理。
    • 最后,还会判断这个cancelledKey所代表的channel是否打开和解除注册,如果关闭并解除注册,则应该将相应的文件描述符对应占用的资源给关闭掉。
   /**
    * sun.nio.ch.SelectorImpl#processDeregisterQueue
    * Invoked by selection operations to process the cancelled-key set
    */
protected final void processDeregisterQueue() throws IOException {
    assert Thread.holdsLock(this);
    assert Thread.holdsLock(publicSelectedKeys);

    Set<SelectionKey> cks = cancelledKeys();
    synchronized (cks) {
        if (!cks.isEmpty()) {
            Iterator<SelectionKey> i = cks.iterator();
            while (i.hasNext()) {
                SelectionKeyImpl ski = (SelectionKeyImpl)i.next();
                i.remove();

                // remove the key from the selector
                implDereg(ski);

                selectedKeys.remove(ski);
                keys.remove(ski);

                // remove from channel's key set
                deregister(ski);

                SelectableChannel ch = ski.channel();
                if (!ch.isOpen() && !ch.isRegistered())
                    ((SelChImpl)ch).kill();
            }
        }
    }
}
//sun.nio.ch.WindowsSelectorImpl#implDereg
@Override
protected void implDereg(SelectionKeyImpl ski) {
    assert !ski.isValid();
    assert Thread.holdsLock(this);

    if (fdMap.remove(ski) != null) {
        int i = ski.getIndex();
        assert (i >= 0);

        if (i != totalChannels - 1) {
            // Copy end one over it
            SelectionKeyImpl endChannel = channelArray[totalChannels-1];
            channelArray[i] = endChannel;
            endChannel.setIndex(i);
            pollWrapper.replaceEntry(pollWrapper, totalChannels-1, pollWrapper, i);
        }
        ski.setIndex(-1);

        channelArray[totalChannels - 1] = null;
        totalChannels--;
        if (totalChannels != 1 && totalChannels % MAX_SELECTABLE_FDS == 1) {
            totalChannels--;
            threadsCount--; // The last thread has become redundant.
        }
    }
}

//sun.nio.ch.SocketChannelImpl#kill
@Override
public void kill() throws IOException {
    synchronized (stateLock) {
        if (state == ST_KILLPENDING) {
            state = ST_KILLED;
            nd.close(fd);
        }
    }
}
//C:/Program Files/Java/jdk-11.0.1/lib/src.zip!/java.base/sun/nio/ch/SocketChannelImpl.java:1126
static {
    IOUtil.load();
    nd = new SocketDispatcher();
}
//sun.nio.ch.SocketDispatcher#close
void close(FileDescriptor fd) throws IOException {
    close0(fd);
}
adjustThreadsCount解读
  1. 接着我们来看到上面WindowsSelectorImpl#doSelect展示源码中adjustThreadsCount()方法的调用。
    • 前面有提到如果totalChannels % MAX_SELECTABLE_FDS == 0,则多开一个线程处理selector。这里就是根据分配的线程数量值来增加或减少线程,其实就是针对操作系统的最大select操作的文件描述符限制对线程个数进行调整。
    • 我们来观察所建线程做了什么事情,即观察SelectThreadrun方法实现。通过观察其源码可以看到它首先是while (true),通过startLock.waitForStart(this)来控制该线程是否运行还是等待,运行状态的话,会进而调用subSelector.poll(index)(这个我们后面内容详细解读),
    • 当此线程poll结束,而且相对于当前主线程假如有多条SelectThread子线程的话,当前这条SelectThread线程第一个结束poll的话,就调用finishLock.threadFinished()来通知主线程。在刚新建这个线程并调用其run方法的时候,此时lastRun = 0,在第一次启动的时候sun.nio.ch.WindowsSelectorImpl.StartLock#runsCounter同样为0,所以会调用startLock.wait()进而进入等待状态。

注意:

  • sun.nio.ch.WindowsSelectorImpl.StartLock同样会判断当前其所检测的线程是否废弃,废弃的话就返回true,这样被检测线程也就能跳出其内run方法的while循环从而结束线程运行。
  • 在调整线程的时候(调用adjustThreadsCount方法)与Selector调用close方法会间接调用到sun.nio.ch.WindowsSelectorImpl#implClose,这两个方法都会涉及到Selector线程的释放,即调用sun.nio.ch.WindowsSelectorImpl.SelectThread#makeZombie
  • finishLock.threadFinished()会调用wakeup()方法来通知主线程,这里,我们可以学到一个细节,如果线程正阻塞在select方法上,就可以调用wakeup方法会使阻塞的选择操作立即返回,通过Windows的相关实现,原理其实是向pipesink端写入了一个字节,source文件描述符就会处于就绪状态,poll方法会返回,从而导致select方法返回。而在其他solaris或者linux系统上其实采用系统调用pipe来完成管道的创建,相当于直接用了系统的管道。通过wakeup()相关实现还可以看出,调用wakeup会设置interruptTriggered的标志位,所以连续多次调用wakeup的效果等同于一次调用,不会引起无所谓的bug出现。
//sun.nio.ch.WindowsSelectorImpl#adjustThreadsCount
// After some channels registered/deregistered, the number of required
// helper threads may have changed. Adjust this number.
private void adjustThreadsCount() {
    if (threadsCount > threads.size()) {
        // More threads needed. Start more threads.
        for (int i = threads.size(); i < threadsCount; i++) {
            SelectThread newThread = new SelectThread(i);
            threads.add(newThread);
            newThread.setDaemon(true);
            newThread.start();
        }
    } else if (threadsCount < threads.size()) {
        // Some threads become redundant. Remove them from the threads List.
        for (int i = threads.size() - 1 ; i >= threadsCount; i--)
            threads.remove(i).makeZombie();
    }
}

//sun.nio.ch.WindowsSelectorImpl.SelectThread
// Represents a helper thread used for select.
private final class SelectThread extends Thread {
    private final int index; // index of this thread
    final SubSelector subSelector;
    private long lastRun = 0; // last run number
    private volatile boolean zombie;
    // Creates a new thread
    private SelectThread(int i) {
        super(null, null, "SelectorHelper", 0, false);
        this.index = i;
        this.subSelector = new SubSelector(i);
        //make sure we wait for next round of poll
        this.lastRun = startLock.runsCounter;
    }
    void makeZombie() {
        zombie = true;
    }
    boolean isZombie() {
        return zombie;
    }
    public void run() {
        while (true) { // poll loop
            // wait for the start of poll. If this thread has become
            // redundant, then exit.
            if (startLock.waitForStart(this))
                return;
            // call poll()
            try {
                subSelector.poll(index);
            } catch (IOException e) {
                // Save this exception and let other threads finish.
                finishLock.setException(e);
            }
            // notify main thread, that this thread has finished, and
            // wakeup others, if this thread is the first to finish.
            finishLock.threadFinished();
        }
    }
}

// sun.nio.ch.WindowsSelectorImpl.FinishLock#threadFinished
// Each helper thread invokes this function on finishLock, when
// the thread is done with poll().
private synchronized void threadFinished() {
    if (threadsToFinish == threads.size()) { // finished poll() first
        // if finished first, wakeup others
        wakeup();
    }
    threadsToFinish--;
    if (threadsToFinish == 0) // all helper threads finished poll().
        notify();             // notify the main thread
}

//sun.nio.ch.WindowsSelectorImpl#wakeup
@Override
public Selector wakeup() {
    synchronized (interruptLock) {
        if (!interruptTriggered) {
            setWakeupSocket();
            interruptTriggered = true;
        }
    }
    return this;
}
//sun.nio.ch.WindowsSelectorImpl#setWakeupSocket
// Sets Windows wakeup socket to a signaled state.
private void setWakeupSocket() {
    setWakeupSocket0(wakeupSinkFd);
}
private native void setWakeupSocket0(int wakeupSinkFd);

JNIEXPORT void JNICALL
Java_sun_nio_ch_WindowsSelectorImpl_setWakeupSocket0(JNIEnv *env, jclass this,
                                                jint scoutFd)
{
    /* Write one byte into the pipe */
    const char byte = 1;
    send(scoutFd, &byte, 1, 0);
}
subSelector的poll方法解读
  1. subSelector.poll() 是select的核心,由native函数poll0实现,并把pollWrapper.pollArrayAddress作为参数传给poll0readFdswriteFdsexceptFds数组用来保存底层select的结果,数组的第一个位置都是存放发生事件的socket的总数,其余位置存放发生事件的socket句柄fd。 我们通过下面的代码可知: 这个poll0()会监听pollWrapper中的FD有没有数据进出,这里会造成IO阻塞,直到有数据读写事件发生。由于pollWrapper中保存的也有ServerSocketChannelFD,所以只要ClientSocket发一份数据到ServerSocket,那么poll0()就会返回;又由于pollWrapper中保存的也有pipewrite端的FD,所以只要pipewrite端向FD发一份数据,也会造成poll0()返回;如果这两种情况都没有发生,那么poll0()就一直阻塞,也就是selector.select()会一直阻塞;如果有任何一种情况发生,那么selector.select()就会返回,所有在SelectThreadrun()里要用while (true) {},这样就可以保证在selector接收到数据并处理完后继续监听poll();

可以看出,NIO依然是阻塞式的IO,那么它和BIO的区别究竟在哪呢。 其实它的区别在于阻塞的位置不同,BIO是阻塞在read方法(recvfrom),而NIO阻塞在select方法。那么这样做有什么好处呢。如果单纯的改变阻塞的位置,自然是没有什么变化的,但epoll等的实现的巧妙之处就在于,它利用回调机制,让监听能够只需要知晓哪些socket上的数据已经准备好了,只需要处理这些线程上面的数据就行了。采用BIO,假设有1000个连接,需要开1000个线程,然后有1000read的位置在阻塞(我们在讲解BIO部分已经通过Demo体现),采用NIO编程,只需要1个线程,它利用select的轮询策略配合epoll的事件机制及红黑树数据结构,降低了其内部轮询的开销,同时极大的减小了线程上下文切换的开销。

//sun.nio.ch.WindowsSelectorImpl.SubSelector
private final class SubSelector {
        private final int pollArrayIndex; // starting index in pollArray to poll
        // These arrays will hold result of native select().
        // The first element of each array is the number of selected sockets.
        // Other elements are file descriptors of selected sockets.
        // 保存发生read的FD
        private final int[] readFds = new int [MAX_SELECTABLE_FDS + 1];
        // 保存发生write的FD
        private final int[] writeFds = new int [MAX_SELECTABLE_FDS + 1];
        //保存发生except的FD
        private final int[] exceptFds = new int [MAX_SELECTABLE_FDS + 1];

        private SubSelector() {
            this.pollArrayIndex = 0; // main thread
        }

        private SubSelector(int threadIndex) { // helper threads
            this.pollArrayIndex = (threadIndex + 1) * MAX_SELECTABLE_FDS;
        }

        private int poll() throws IOException{ // poll for the main thread
            return poll0(pollWrapper.pollArrayAddress,
                         Math.min(totalChannels, MAX_SELECTABLE_FDS),
                         readFds, writeFds, exceptFds, timeout);
        }

        private int poll(int index) throws IOException {
            // poll for helper threads
            return  poll0(pollWrapper.pollArrayAddress +
                     (pollArrayIndex * PollArrayWrapper.SIZE_POLLFD),
                     Math.min(MAX_SELECTABLE_FDS,
                             totalChannels - (index + 1) * MAX_SELECTABLE_FDS),
                     readFds, writeFds, exceptFds, timeout);
        }

        private native int poll0(long pollAddress, int numfds,
             int[] readFds, int[] writeFds, int[] exceptFds, long timeout);
             ...
}
updateSelectedKeys解读
  1. 接下来将通过上面WindowsSelectorImpl#doSelect展示源码中<5> 处的 updateSelectedKeys(action)来处理每个channel准备就绪的信息。
  • 如果该通道的key尚未在selectedKeys中存在,则将其添加到该集合中。
  • 如果该通道的key已经存在selectedKeys中,即这个channel存在所支持的ReadyOps就绪操作中必须包含一个这种操作(由(ski.nioReadyOps() & ski.nioInterestOps()) != 0来确定),此时修改其ReadyOps为当前所要进行的操作。而我们之前看到的Consumer<SelectionKey>这个动作也是在此处进行。而由下面源码可知,先前记录在ReadyOps中的任何就绪信息在调用此action之前被丢弃掉,直接进行设定。
//sun.nio.ch.WindowsSelectorImpl#updateSelectedKeys
private int updateSelectedKeys(Consumer<SelectionKey> action) {
    updateCount++;
    int numKeysUpdated = 0;
    numKeysUpdated += subSelector.processSelectedKeys(updateCount, action);
    for (SelectThread t: threads) {
        numKeysUpdated += t.subSelector.processSelectedKeys(updateCount, action);
    }
    return numKeysUpdated;
}
//sun.nio.ch.SelectorImpl#processReadyEvents
protected final int processReadyEvents(int rOps,
                                        SelectionKeyImpl ski,
                                        Consumer<SelectionKey> action) {
    if (action != null) {
        ski.translateAndSetReadyOps(rOps);
        if ((ski.nioReadyOps() & ski.nioInterestOps()) != 0) {
            action.accept(ski);
            ensureOpen();
            return 1;
        }
    } else {
        assert Thread.holdsLock(publicSelectedKeys);
        if (selectedKeys.contains(ski)) {
            if (ski.translateAndUpdateReadyOps(rOps)) {
                return 1;
            }
        } else {
            ski.translateAndSetReadyOps(rOps);
            if ((ski.nioReadyOps() & ski.nioInterestOps()) != 0) {
                selectedKeys.add(ski);
                return 1;
            }
        }
    }
    return 0;
}
//sun.nio.ch.WindowsSelectorImpl.SubSelector#processSelectedKeys
private int processSelectedKeys(long updateCount, Consumer<SelectionKey> action) {
    int numKeysUpdated = 0;
    numKeysUpdated += processFDSet(updateCount, action, readFds,
                                    Net.POLLIN,
                                    false);
    numKeysUpdated += processFDSet(updateCount, action, writeFds,
                                    Net.POLLCONN |
                                    Net.POLLOUT,
                                    false);
    numKeysUpdated += processFDSet(updateCount, action, exceptFds,
                                    Net.POLLIN |
                                    Net.POLLCONN |
                                    Net.POLLOUT,
                                    true);
    return numKeysUpdated;
}

    /**
    * sun.nio.ch.WindowsSelectorImpl.SubSelector#processFDSet
    * updateCount is used to tell if a key has been counted as updated
    * in this select operation.
    *
    * me.updateCount <= updateCount
    */
private int processFDSet(long updateCount,
                            Consumer<SelectionKey> action,
                            int[] fds, int rOps,
                            boolean isExceptFds)
{
    int numKeysUpdated = 0;
    for (int i = 1; i <= fds[0]; i++) {
        int desc = fds[i];
        if (desc == wakeupSourceFd) {
            synchronized (interruptLock) {
                interruptTriggered = true;
            }
            continue;
        }
        MapEntry me = fdMap.get(desc);
        // If me is null, the key was deregistered in the previous
        // processDeregisterQueue.
        if (me == null)
            continue;
        SelectionKeyImpl sk = me.ski;

        // The descriptor may be in the exceptfds set because there is
        // OOB data queued to the socket. If there is OOB data then it
        // is discarded and the key is not added to the selected set.
        if (isExceptFds &&
            (sk.channel() instanceof SocketChannelImpl) &&
            discardUrgentData(desc))
        {
            continue;
        }
        //我们应该关注的
        int updated = processReadyEvents(rOps, sk, action);
        if (updated > 0 && me.updateCount != updateCount) {
            me.updateCount = updateCount;
            numKeysUpdated++;
        }
    }
    return numKeysUpdated;
}

至此,关于Selector的内容就暂时告一段落,在下一篇中,我会针对Java NIO Buffer进行相关解读。