简介
在前面的章节中,我们介绍了在Netty中可以使用kequeue或者epoll来实现更为高效的native传输方式。那么kequeue和epoll和NIO传输协议有什么不同呢?
本章将会以kequeue为例进行深入探讨。
在上面我们介绍的native的例子中,关于kqueue的类有这样几个,分别是KQueueEventLoopGroup,KQueueServerSocketChannel和KQueueSocketChannel,通过简单的替换和添加对应的依赖包,我们可以轻松的将普通的NIO netty服务替换成为native的Kqueue服务。
是时候揭开Kqueue的秘密了。
KQueueEventLoopGroup
eventLoop和eventLoopGroup是用来接受event和事件处理的。先来看下KQueueEventLoopGroup的定义:
public final class KQueueEventLoopGroup extends MultithreadEventLoopGroup
作为一个MultithreadEventLoopGroup,必须实现一个newChild方法,用来创建child EventLoop。在KQueueEventLoopGroup中,除了构造函数之外,额外需要实现的方法就是newChild:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }
newChild中的所有参数都是从KQueueEventLoopGroup的构造函数中传入的。除了maxEvents,selectStrategyFactory和rejectedExecutionHandler之外,还可以接收taskQueueFactory和tailTaskQueueFactory两个参数,最后把这些参数都传到KQueueEventLoop的构造函数中去,最终返回一个KQueueEventLoop对象。
另外在使用KQueueEventLoopGroup之前我们还需要确保Kqueue在系统中是可用的,这个判断是通过调用KQueue.ensureAvailability();
来实现的。
KQueue.ensureAvailability首先判断是否定义了系统属性io.netty.transport.noNative,如果定了,说明native transport被禁用了,后续也就没有必要再进行判断了。
如果io.netty.transport.noNative没有被定义,那么会调用Native.newKQueue()
来尝试从native中获取一个kqueue的FileDescriptor,如果上述的获取过程中没有任何异常,则说明kqueue在native方法中存在,我们可以继续使用了。
以下是判断kqueue是否可用的代码:
static { Throwable cause = null; if (SystemPropertyUtil.getBoolean("io.netty.transport.noNative", false)) { cause = new UnsupportedOperationException( "Native transport was explicit disabled with -Dio.netty.transport.noNative=true"); } else { FileDescriptor kqueueFd = null; try { kqueueFd = Native.newKQueue(); } catch (Throwable t) { cause = t; } finally { if (kqueueFd != null) { try { kqueueFd.close(); } catch (Exception ignore) { // ignore } } } } UNAVAILABILITY_CAUSE = cause; }
KQueueEventLoop
KQueueEventLoop是从KQueueEventLoopGroup中创建出来的,用来执行具体的IO任务。
先来看一下KQueueEventLoop的定义:
final class KQueueEventLoop extends SingleThreadEventLoop
不管是NIO还是KQueue或者是Epoll,因为使用了更加高级的IO技术,所以他们使用的EventLoop都是SingleThreadEventLoop,也就是说使用单线程就够了。
和KQueueEventLoopGroup一样,KQueueEventLoop也需要判断当前的系统环境是否支持kqueue:
static { KQueue.ensureAvailability(); }
上一节讲到了,KQueueEventLoopGroup会调用KQueueEventLoop的构造函数来返回一个eventLoop对象, 我们先来看下KQueueEventLoop的构造函数:
KQueueEventLoop(EventLoopGroup parent, Executor executor, int maxEvents, SelectStrategy strategy, RejectedExecutionHandler rejectedExecutionHandler, EventLoopTaskQueueFactory taskQueueFactory, EventLoopTaskQueueFactory tailTaskQueueFactory) { super(parent, executor, false, newTaskQueue(taskQueueFactory), newTaskQueue(tailTaskQueueFactory), rejectedExecutionHandler); this.selectStrategy = ObjectUtil.checkNotNull(strategy, "strategy"); this.kqueueFd = Native.newKQueue(); if (maxEvents == 0) { allowGrowing = true; maxEvents = 4096; } else { allowGrowing = false; } this.changeList = new KQueueEventArray(maxEvents); this.eventList = new KQueueEventArray(maxEvents); int result = Native.keventAddUserEvent(kqueueFd.intValue(), KQUEUE_WAKE_UP_IDENT); if (result < 0) { cleanup(); throw new IllegalStateException("kevent failed to add user event with errno: " + (-result)); } }
传入的maxEvents表示的是这个KQueueEventLoop能够接受的最大的event个数。如果maxEvents=0,则表示KQueueEventLoop的event容量可以动态扩展,并且最大值是4096。否则的话,KQueueEventLoop的event容量不能扩展。
maxEvents是作为数组的大小用来构建changeList和eventList。
KQueueEventLoop中还定义了一个map叫做channels,用来保存注册的channels:
private final IntObjectMap<AbstractKQueueChannel> channels = new IntObjectHashMap<AbstractKQueueChannel>(4096);
来看一下channel的add和remote方法:
void add(AbstractKQueueChannel ch) { assert inEventLoop(); AbstractKQueueChannel old = channels.put(ch.fd().intValue(), ch); assert old == null || !old.isOpen(); } void remove(AbstractKQueueChannel ch) throws Exception { assert inEventLoop(); int fd = ch.fd().intValue(); AbstractKQueueChannel old = channels.remove(fd); if (old != null && old != ch) { channels.put(fd, old); assert !ch.isOpen(); } else if (ch.isOpen()) { ch.unregisterFilters(); } }
可以看到添加和删除的都是AbstractKQueueChannel,后面的章节中我们会详细讲解KQueueChannel,这里我们只需要知道channel map中的key是kequeue中特有的FileDescriptor的int值。
再来看一下EventLoop中最重要的run方法:
protected void run() { for (;;) { try { int strategy = selectStrategy.calculateStrategy(selectNowSupplier, hasTasks()); switch (strategy) { case SelectStrategy.CONTINUE: continue; case SelectStrategy.BUSY_WAIT: case SelectStrategy.SELECT: strategy = kqueueWait(WAKEN_UP_UPDATER.getAndSet(this, 0) == 1); if (wakenUp == 1) { wakeup(); } default: } final int ioRatio = this.ioRatio; if (ioRatio == 100) { try { if (strategy > 0) { processReady(strategy); } } finally { runAllTasks(); } } else { final long ioStartTime = System.nanoTime(); try { if (strategy > 0) { processReady(strategy); } } finally { final long ioTime = System.nanoTime() - ioStartTime; runAllTasks(ioTime * (100 - ioRatio) / ioRatio); }
它的逻辑是先使用selectStrategy.calculateStrategy获取当前的select strategy,然后根据strategy的值来判断是否需要执行processReady方法,最后执行runAllTasks,从task queue中拿到要执行的任务去执行。
selectStrategy.calculateStrategy用来判断当前的select状态,默认情况下有三个状态,分别是:SELECT,CONTINUE,BUSY_WAIT。 这三个状态都是负数:
int SELECT = -1; int CONTINUE = -2; int BUSY_WAIT = -3;
分别表示当前的IO在slect的block状态,或者跳过当前IO的状态,和正在IO loop pull的状态。BUSY_WAIT是一个非阻塞的IO PULL,kqueue并不支持,所以会fallback到SELECT。
除了这三个状态之外,calculateStrategy还会返回一个正值,表示当前要执行的任务的个数。
在run方法中,如果strategy的结果是SELECT,那么最终会调用Native.keventWait方法返回当前ready的events个数,并且将ready的event放到KQueueEventArray的eventList中去。
如果ready的event个数大于零,则会调用processReady方法对这些event进行状态回调处理。
怎么处理的呢?下面是处理的核心逻辑:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }0
这里的fd是从eventList中读取到的:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }1
根据eventList的fd,我们可以从channels中拿到对应的KQueueChannel,然后根据event的filter状态来决定KQueueChannel的具体操作,是writeReady,readReady或者readEOF。
最后就是执行runAllTasks方法了,runAllTasks的逻辑很简单,就是从taskQueue中读取任务然后执行。
KQueueServerSocketChannel和KQueueSocketChannel
KQueueServerSocketChannel是用在server端的channel:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }2
KQueueServerSocketChannel继承自AbstractKQueueServerChannel,除了构造函数之外,最重要的一个方法就是newChildChannel:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }3
这个方法用来创建一个新的child channel。从上面的代码中,我们可以看到生成的child channel是一个KQueueSocketChannel的实例。
它的构造函数接受三个参数,分别是parent channel,BsdSocket和InetSocketAddress。
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }4
这里的fd是socket accept acceptedAddress的结果:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }5
下面是KQueueSocketChannel的定义:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }6
KQueueSocketChannel和KQueueServerSocketChannel的关系是父子的关系,在KQueueSocketChannel中有一个parent方法,用来返回ServerSocketChannel对象,这也是前面提到的newChildChannel方法中传入KQueueSocketChannel构造函数中的serverChannel:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }7
KQueueSocketChannel还有一个特性就是支持tcp fastopen,它的本质是调用BsdSocket的connectx方法,在建立连接的同时传递数据:
protected EventLoop newChild(Executor executor, Object... args) throws Exception { Integer maxEvents = (Integer) args[0]; SelectStrategyFactory selectStrategyFactory = (SelectStrategyFactory) args[1]; RejectedExecutionHandler rejectedExecutionHandler = (RejectedExecutionHandler) args[2]; EventLoopTaskQueueFactory taskQueueFactory = null; EventLoopTaskQueueFactory tailTaskQueueFactory = null; int argsLength = args.length; if (argsLength > 3) { taskQueueFactory = (EventLoopTaskQueueFactory) args[3]; } if (argsLength > 4) { tailTaskQueueFactory = (EventLoopTaskQueueFactory) args[4]; } return new KQueueEventLoop(this, executor, maxEvents, selectStrategyFactory.newSelectStrategy(), rejectedExecutionHandler, taskQueueFactory, tailTaskQueueFactory); }8
总结
以上就是KqueueEventLoop和KqueueSocketChannel的详细介绍,基本上和NIO没有太大的区别,只不过性能根据优秀。
更多内容请参考 http://www.flydean.com/53-1-netty-kqueue-transport/
最通俗的解读,最深刻的干货,最简洁的教程,众多你不知道的小技巧等你来发现!
欢迎关注我的公众号:「程序那些事」,懂技术,更懂你!
原文:https://juejin.cn/post/7103037223775240205