Netty学习之旅------Netty Channel 概述

1、Channel官方概述

先附上官方关于Channel的介绍。

/**
 * A nexus to a network socket or a component which is capable of I/O
 * operations such as read, write, connect, and bind.
 * <p>
 * A channel provides a user:
 * <ul>
 * <li>the current state of the channel (e.g. is it open? is it connected?),</li>
 * <li>the {@linkplain ChannelConfig configuration parameters} of the channel (e.g. receive buffer size),</li>
 * <li>the I/O operations that the channel supports (e.g. read, write, connect, and bind), and</li>
 * <li>the {@link ChannelPipeline} which handles all I/O events and requests
 *     associated with the channel.</li>
 * </ul>
 *
 * <h3>All I/O operations are asynchronous.</h3>
 * <p>
 * All I/O operations in Netty are asynchronous.  It means any I/O calls will
 * return immediately with no guarantee that the requested I/O operation has
 * been completed at the end of the call.  Instead, you will be returned with
 * a {@link ChannelFuture} instance which will notify you when the requested I/O
 * operation has succeeded, failed, or canceled.
 *
 * <h3>Channels are hierarchical</h3>
 * <p>
 * A {@link Channel} can have a {@linkplain #parent() parent} depending on
 * how it was created.  For instance, a {@link SocketChannel}, that was accepted
 * by {@link ServerSocketChannel}, will return the {@link ServerSocketChannel}
 * as its parent on {@link #parent()}.
 * <p>
 * The semantics of the hierarchical structure depends on the transport
 * implementation where the {@link Channel} belongs to.  For example, you could
 * write a new {@link Channel} implementation that creates the sub-channels that
 * share one socket connection, as <a href="http://beepcore.org/">BEEP</a> and
 * <a href="http://en.wikipedia.org/wiki/Secure_Shell">SSH</a> do.
 *
 * <h3>Downcast to access transport-specific operations</h3>
 * <p>
 * Some transports exposes additional operations that is specific to the
 * transport.  Down-cast the {@link Channel} to sub-type to invoke such
 * operations.  For example, with the old I/O datagram transport, multicast
 * join / leave operations are provided by {@link DatagramChannel}.
 *
 * <h3>Release resources</h3>
 * <p>
 * It is important to call {@link #close()} or {@link #close(ChannelPromise)} to release all
 * resources once you are done with the {@link Channel}. This ensures all resources are
 * released in a proper way, i.e. filehandles.
 */

• 通道状态主要包括：打开、关闭、连接。
• 通道主要的IO操作，读(read)、写(write)、连接(connect)、绑定(bind)。
• 所有的IO操作都是异步的，调用诸如read,write方法后，并不保证IO操作完成，但会返回一个凭证，在IO操作成功，取消或失败后会记录在该凭证中。
• channel有父子关系，SocketChannel是通过ServerSocketChannel接受创建的，故SocketChannel的parent()方法返回的就是ServerSocketChannel。
• 在Channel使用完毕后，请调用close方法，释放通道占用的资源。

2、Netty Channel API

接下来重点介绍一下Channel的API。

/**
     * Returns the globally unique identifier of this {@link Channel}.
     * 返回全局唯一的channel id
     */
    ChannelId id();

    /**
     * Return the {@link EventLoop} this {@link Channel} was registered too.
     * 返回该Channel注册的线程模型，先理解为Ractor模型的Ractor线程。
     */
    EventLoop eventLoop();

    /**
     * Returns the parent of this channel.
     *
     * @return the parent channel.
     *         {@code null} if this channel does not have a parent channel.
     *    返回该Channel由谁创建的，ServerSocketChannel返回null,SocketChannel返回创建它的ServerSocketChannel
     */
    Channel parent();

    /**
     * Returns the configuration of this channel.
     *  返回通道的配置信息
     */
    ChannelConfig config();

    /**
     * Returns {@code true} if the {@link Channel} is open an may get active later
     * 通道是否打开
     */
    boolean isOpen();

    /**
     * Returns {@code true} if the {@link Channel} is registered with an {@link EventLoop}.
     * 该通道是否已经注册在事件模型中，此处先参考Nio编程模型，一个通过需要注册在Register上
     */
    boolean isRegistered();

    /**
     * Return {@code true} if the {@link Channel} is active and so connected.
     *  通道是否激活
     */
    boolean isActive();

    /**
     * Return the {@link ChannelMetadata} of the {@link Channel} which describe the nature of the {@link Channel}.
     *  通道是否支持 调用disconnect方法后，调用connect方法
     */
    ChannelMetadata metadata();

    /**
     * Returns the local address where this channel is bound to.  The returned
     * {@link SocketAddress} is supposed to be down-cast into more concrete
     * type such as {@link InetSocketAddress} to retrieve the detailed
     * information.
     *
     * @return the local address of this channel.
     *         {@code null} if this channel is not bound.
      *      返回绑定的地址，服务端的Channel返回监听的地址，而客户端的Channel返回连接到服务端的本地套接字。
     */
    SocketAddress localAddress();

    /**
     * Returns the remote address where this channel is connected to.  The
     * returned {@link SocketAddress} is supposed to be down-cast into more
     * concrete type such as {@link InetSocketAddress} to retrieve the detailed
     * information.
     *
     * @return the remote address of this channel.
     *         {@code null} if this channel is not connected.
     *         If this channel is not connected but it can receive messages
     *         from arbitrary remote addresses (e.g. {@link DatagramChannel},
     *         use {@link DatagramPacket#recipient()} to determine
     *         the origination of the received message as this method will
     *         return {@code null}.
     *         返回channel的远程套接字。
     */
    SocketAddress remoteAddress();

    /**
     * Returns the {@link ChannelFuture} which will be notified when this
     * channel is closed.  This method always returns the same future instance.
     *  通道的关闭凭证（许可），这里是多线程编程一种典型的设计模式，一个channle返回一个固定的
     */
    ChannelFuture closeFuture();

    /**
     * Returns {@code true} if and only if the I/O thread will perform the
     * requested write operation immediately.  Any write requests made when
     * this method returns {@code false} are queued until the I/O thread is
     * ready to process the queued write requests.
     *  是否可写，如果通道的写缓冲区未满，即返回true，表示写操作可以立即     
     *   操作缓冲区，然后返回。

     */
    boolean isWritable();

    /**
     * Returns an <em>internal-use-only</em> object that provides unsafe operations.
     */
    Unsafe unsafe();

    /**
     * Return the assigned {@link ChannelPipeline}
     *  返回管道
     */
    ChannelPipeline pipeline();

    /**
     * Return the assigned {@link ByteBufAllocator} which will be used to allocate {@link ByteBuf}s.
     * 返回ByteBuf内存分配器
     */
    ByteBufAllocator alloc();

    /**
     * Return a new {@link ChannelPromise}.
     * 诸如newPromise,newSuccessedFuture()方法，就是返回一个凭证，用来保存通知结果的，是多线程编程一         * 中典型的设计模式
     */
    ChannelPromise newPromise();

    /**
     * Return an new {@link ChannelProgressivePromise}
     */
    ChannelProgressivePromise newProgressivePromise();

    /**
     * Create a new {@link ChannelFuture} which is marked as succeeded already. So {@link ChannelFuture#isSuccess()}
     * will return {@code true}. All {@link FutureListener} added to it will be notified directly. Also
     * every call of blocking methods will just return without blocking.
     */
    ChannelFuture newSucceededFuture();

    /**
     * Create a new {@link ChannelFuture} which is marked as failed already. So {@link ChannelFuture#isSuccess()}
     * will return {@code false}. All {@link FutureListener} added to it will be notified directly. Also
     * every call of blocking methods will just return without blocking.
     */
    ChannelFuture newFailedFuture(Throwable cause);

    /**
     * Return a special ChannelPromise which can be reused for different operations.
     * <p>
     * It's only supported to use
     * it for {@link Channel#write(Object, ChannelPromise)}.
     * </p>
     * <p>
     * Be aware that the returned {@link ChannelPromise} will not support most operations and should only be used
     * if you want to save an object allocation for every write operation. You will not be able to detect if the
     * operation  was complete, only if it failed as the implementation will call
     * {@link ChannelPipeline#fireExceptionCaught(Throwable)} in this case.
     * </p>
     * <strong>Be aware this is an expert feature and should be used with care!</strong>
     */
    ChannelPromise voidPromise();

    /**
     * Request to bind to the given {@link SocketAddress} and notify the {@link ChannelFuture} once the operation
     * completes, either because the operation was successful or because of an error.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#bind(ChannelHandlerContext, SocketAddress, ChannelPromise)} method
     * called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     *  绑定
     */
    ChannelFuture bind(SocketAddress localAddress);

    /**
     * Request to connect to the given {@link SocketAddress} and notify the {@link ChannelFuture} once the operation
     * completes, either because the operation was successful or because of an error.
     * <p>
     * If the connection fails because of a connection timeout, the {@link ChannelFuture} will get failed with
     * a {@link ConnectTimeoutException}. If it fails because of connection refused a {@link ConnectException}
     * will be used.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#connect(ChannelHandlerContext, SocketAddress, SocketAddress, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     * 连接
     */
    ChannelFuture connect(SocketAddress remoteAddress);

    /**
     * Request to connect to the given {@link SocketAddress} while bind to the localAddress and notify the
     * {@link ChannelFuture} once the operation completes, either because the operation was successful or because of
     * an error.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#connect(ChannelHandlerContext, SocketAddress, SocketAddress, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    ChannelFuture connect(SocketAddress remoteAddress, SocketAddress localAddress);

    /**
     * Request to disconnect from the remote peer and notify the {@link ChannelFuture} once the operation completes,
     * either because the operation was successful or because of an error.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#disconnect(ChannelHandlerContext, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     * 断开连接
     */
    ChannelFuture disconnect();

    /**
     * Request to close this {@link Channel} and notify the {@link ChannelFuture} once the operation completes,
     * either because the operation was successful or because of
     * an error.
     *
     * After it is closed it is not possible to reuse it again.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#close(ChannelHandlerContext, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     *  关闭，释放通道资源
     */
    ChannelFuture close();

    /**
     * Request to deregister this {@link Channel} from its assigned {@link EventLoop} and notify the
     * {@link ChannelFuture} once the operation completes, either because the operation was successful or because of
     * an error.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#deregister(ChannelHandlerContext, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     * <p>
     * After this method completes (not the {@link ChannelFuture}!) one can not submit new tasks to the
     * {@link Channel}'s {@link EventLoop} until the {@link Channel} is again registered with an {@link EventLoop}.
     * Any attempt to do so will result in a {@link RejectedExecutionException} being thrown.
     * Any tasks that were submitted before the call to {@link #deregister()} will finish before the
     * {@link ChannelFuture} completes. Furthermore, periodic and delayed tasks will not be executed until the
     * {@link Channel} is registered with an {@link EventLoop} again. Theses are tasks submitted
     * to the {@link EventLoop} via one of the methods declared by {@link ScheduledExecutorService}.
     * Please note that all of the above only applies to tasks created from within the deregistered {@link Channel}'s
     * {@link ChannelHandler}s.
     * <p>
     * It's only safe to {@linkplain EventLoop#register(Channel)} the {@link Channel} with another (or the same)
     * {@link EventLoop} after the {@link ChannelFuture} has completed.
     */
    ChannelFuture deregister();

    /**
     * Request to bind to the given {@link SocketAddress} and notify the {@link ChannelFuture} once the operation
     * completes, either because the operation was successful or because of an error.
     *
     * The given {@link ChannelPromise} will be notified.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#bind(ChannelHandlerContext, SocketAddress, ChannelPromise)} method
     * called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    ChannelFuture bind(SocketAddress localAddress, ChannelPromise promise);

    /**
     * Request to connect to the given {@link SocketAddress} and notify the {@link ChannelFuture} once the operation
     * completes, either because the operation was successful or because of an error.
     *
     * The given {@link ChannelFuture} will be notified.
     *
     * <p>
     * If the connection fails because of a connection timeout, the {@link ChannelFuture} will get failed with
     * a {@link ConnectTimeoutException}. If it fails because of connection refused a {@link ConnectException}
     * will be used.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#connect(ChannelHandlerContext, SocketAddress, SocketAddress, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    ChannelFuture connect(SocketAddress remoteAddress, ChannelPromise promise);

    /**
     * Request to connect to the given {@link SocketAddress} while bind to the localAddress and notify the
     * {@link ChannelFuture} once the operation completes, either because the operation was successful or because of
     * an error.
     *
     * The given {@link ChannelPromise} will be notified and also returned.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#connect(ChannelHandlerContext, SocketAddress, SocketAddress, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    ChannelFuture connect(SocketAddress remoteAddress, SocketAddress localAddress, ChannelPromise promise);

    /**
     * Request to disconnect from the remote peer and notify the {@link ChannelFuture} once the operation completes,
     * either because the operation was successful or because of an error.
     *
     * The given {@link ChannelPromise} will be notified.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#disconnect(ChannelHandlerContext, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    ChannelFuture disconnect(ChannelPromise promise);

    /**
     * Request to close this {@link Channel} and notify the {@link ChannelFuture} once the operation completes,
     * either because the operation was successful or because of
     * an error.
     *
     * After it is closed it is not possible to reuse it again.
     * The given {@link ChannelPromise} will be notified.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#close(ChannelHandlerContext, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    ChannelFuture close(ChannelPromise promise);

    /**
     * Request to deregister this {@link Channel} from its assigned {@link EventLoop} and notify the
     * {@link ChannelPromise} once the operation completes, either because the operation was successful or because of
     * an error.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#deregister(ChannelHandlerContext, ChannelPromise)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     * <p>
     * After this method completes (not the {@link ChannelPromise}!) one can not submit new tasks to the
     * {@link Channel}'s {@link EventLoop} until the {@link Channel} is again registered with an {@link EventLoop}.
     * Any attempt to do so will result in a {@link RejectedExecutionException} being thrown.
     * Any tasks that were submitted before the call to {@link #deregister()} will finish before the
     * {@link ChannelPromise} completes. Furthermore, periodic and delayed tasks will not be executed until the
     * {@link Channel} is registered with an {@link EventLoop} again. Theses are tasks submitted
     * to the {@link EventLoop} via one of the methods declared by {@link ScheduledExecutorService}.
     * Please note that all of the above only applies to tasks created from within the deregistered {@link Channel}'s
     * {@link ChannelHandler}s.
     * <p>
     * It's only safe to {@linkplain EventLoop#register(Channel)} the {@link Channel} with another (or the same)
     * {@link EventLoop} after the {@link ChannelPromise} has completed.
     */
    ChannelFuture deregister(ChannelPromise promise);

    /**
     * Request to Read data from the {@link Channel} into the first inbound buffer, triggers an
     * {@link ChannelHandler#channelRead(ChannelHandlerContext, Object)} event if data was
     * read, and triggers a
     * {@link ChannelHandler#channelReadComplete(ChannelHandlerContext) channelReadComplete} event so the
     * handler can decide to continue reading.  If there's a pending read operation already, this method does nothing.
     * <p>
     * This will result in having the
     * {@link ChannelHandler#read(ChannelHandlerContext)}
     * method called of the next {@link ChannelHandler} contained in the  {@link ChannelPipeline} of the
     * {@link Channel}.
     */
    Channel read();

    /**
     * Request to write a message via this {@link Channel} through the {@link ChannelPipeline}.
     * This method will not request to actual flush, so be sure to call {@link #flush()}
     * once you want to request to flush all pending data to the actual transport.
     */
    ChannelFuture write(Object msg);

    /**
     * Request to write a message via this {@link Channel} through the {@link ChannelPipeline}.
     * This method will not request to actual flush, so be sure to call {@link #flush()}
     * once you want to request to flush all pending data to the actual transport.
     */
    ChannelFuture write(Object msg, ChannelPromise promise);

    /**
     * Request to flush all pending messages.
     */
    Channel flush();

    /**
     * Shortcut for call {@link #write(Object, ChannelPromise)} and {@link #flush()}.
     */
    ChannelFuture writeAndFlush(Object msg, ChannelPromise promise);

    /**
     * Shortcut for call {@link #write(Object)} and {@link #flush()}.
     */
    ChannelFuture writeAndFlush(Object msg);

    /**
     * <em>Unsafe</em> operations that should <em>never</em> be called from user-code. These methods
     * are only provided to implement the actual transport, and must be invoked from an I/O thread except for the
     * following methods:
     * <ul>
     *   <li>{@link #invoker()}</li>
     *   <li>{@link #localAddress()}</li>
     *   <li>{@link #remoteAddress()}</li>
     *   <li>{@link #closeForcibly()}</li>
     *   <li>{@link #register(EventLoop, ChannelPromise)}</li>
     *   <li>{@link #deregister(ChannelPromise)}</li>
     *   <li>{@link #voidPromise()}</li>
     * </ul>
     */
    interface Unsafe {

        /**
         * Return the assigned {@link RecvByteBufAllocator.Handle} which will be used to allocate {@link ByteBuf}'s when
         * receiving data.
         */
        RecvByteBufAllocator.Handle recvBufAllocHandle();

        /**
         * Returns the {@link ChannelHandlerInvoker} which is used by default unless specified by a user.
         */
        ChannelHandlerInvoker invoker();

        /**
         * Return the {@link SocketAddress} to which is bound local or
         * {@code null} if none.
         */
        SocketAddress localAddress();

        /**
         * Return the {@link SocketAddress} to which is bound remote or
         * {@code null} if none is bound yet.
         */
        SocketAddress remoteAddress();

        /**
         * Register the {@link Channel} of the {@link ChannelPromise} and notify
         * the {@link ChannelFuture} once the registration was complete.
         * <p>
         * It's only safe to submit a new task to the {@link EventLoop} from within a
         * {@link ChannelHandler} once the {@link ChannelPromise} succeeded. Otherwise
         * the task may or may not be rejected.
         * </p>
         */
        void register(EventLoop eventLoop, ChannelPromise promise);

        /**
         * Bind the {@link SocketAddress} to the {@link Channel} of the {@link ChannelPromise} and notify
         * it once its done.
         */
        void bind(SocketAddress localAddress, ChannelPromise promise);

        /**
         * Connect the {@link Channel} of the given {@link ChannelFuture} with the given remote {@link SocketAddress}.
         * If a specific local {@link SocketAddress} should be used it need to be given as argument. Otherwise just
         * pass {@code null} to it.
         *
         * The {@link ChannelPromise} will get notified once the connect operation was complete.
         */
        void connect(SocketAddress remoteAddress, SocketAddress localAddress, ChannelPromise promise);

        /**
         * Disconnect the {@link Channel} of the {@link ChannelFuture} and notify the {@link ChannelPromise} once the
         * operation was complete.
         */
        void disconnect(ChannelPromise promise);

        /**
         * Close the {@link Channel} of the {@link ChannelPromise} and notify the {@link ChannelPromise} once the
         * operation was complete.
         */
        void close(ChannelPromise promise);

        /**
         * Closes the {@link Channel} immediately without firing any events.  Probably only useful
         * when registration attempt failed.
         */
        void closeForcibly();

        /**
         * Deregister the {@link Channel} of the {@link ChannelPromise} from {@link EventLoop} and notify the
         * {@link ChannelPromise} once the operation was complete.
         */
        void deregister(ChannelPromise promise);

        /**
         * Schedules a read operation that fills the inbound buffer of the first {@link ChannelHandler} in the
         * {@link ChannelPipeline}.  If there's already a pending read operation, this method does nothing.
         */
        void beginRead();

        /**
         * Schedules a write operation.
         */
        void write(Object msg, ChannelPromise promise);

        /**
         * Flush out all write operations scheduled via {@link #write(Object, ChannelPromise)}.
         */
        void flush();

        /**
         * Return a special ChannelPromise which can be reused and passed to the operations in {@link Unsafe}.
         * It will never be notified of a success or error and so is only a placeholder for operations
         * that take a {@link ChannelPromise} as argument but for which you not want to get notified.
         */
        ChannelPromise voidPromise();

        /**
         * Returns the {@link ChannelOutboundBuffer} of the {@link Channel} where the pending write requests are stored.
         * 返回通道的环形缓存区
         */
        ChannelOutboundBuffer outboundBuffer();
    }

3、Netty Channel主要类图

本文关于Netty Channel接口就描述到这里，主要从Channel概述、Channel接口API,Channel类图三个方面简单介绍一下Channel,为进一步解剖Netty Channel，ChanelPiple,ChannelHandler等打下基础。

4、Netty Channel 研究方向（不成熟想法）

2.1 Channel接口概述

2.2、AbstractChannel源码学习

2.2.1 环形缓存区

2.2.2 Netty线程模型

2.2.3 Chanel read,write方法实现探究

2.3、ServerChannel

2.4、SOcketChannel

 

欢迎加笔者微信号(dingwpmz)，加群探讨，笔者优质专栏目录：

1、源码分析RocketMQ专栏(40篇+)
2、源码分析Sentinel专栏(12篇+)
3、源码分析Dubbo专栏(28篇+)
4、源码分析Mybatis专栏
5、源码分析Netty专栏(18篇+)
6、源码分析JUC专栏
7、源码分析Elasticjob专栏
8、Elasticsearch专栏
9、源码分析Mycat专栏