NioEventLoop启动触发条件：
1.服务端绑定本地端口
2.新连接接入通过chooser绑定一个NioEventLoop
服务端绑定本地端口
绑定本地端口，使用下面方法;
ChannelFuture future = bootstrap.bind(host, port).sync();
最终会调用doBind0()方法：
private static void doBind0(final ChannelFuture regFuture, final Channel channel, final SocketAddress localAddress, final ChannelPromise promise) { ???channel.eventLoop().execute(new Runnable() { ???????public void run() { ???????????if(regFuture.isSuccess()) { ???????????????channel.bind(localAddress, promise).addListener(ChannelFutureListener.CLOSE_ON_FAILURE); ???????????} else { ???????????????promise.setFailure(regFuture.cause()); ???????????} ???????} ???});}这个时候就会调用channel对应NioEventLoop的execute方法，会判断是否在当前的eventloop对应的thread中，如果在，直接向任务队列中添加绑定端口的任务，如果不在，首先要start当前eventLoop对应的thread,再将任务放到任务队列中。这里的excute(task)方法，并不是让线程直接执行它，而是将它放到线程的任务队列中，等待线程去执行它。
public void execute(Runnable task) { ???if(task == null) { ???????throw new NullPointerException("task"); ???} else { ???????boolean inEventLoop = this.inEventLoop(); ???????if(inEventLoop) { ???????????this.addTask(task); ???????} else { ???????????this.startThread(); ???????????this.addTask(task); ???????????if(this.isShutdown() && this.removeTask(task)) { ???????????????reject(); ???????????} ???????} ???????if(!this.addTaskWakesUp && this.wakesUpForTask(task)) { ???????????this.wakeup(inEventLoop); ???????} ???}}NioEventLoop线程执行逻辑
NioEventLoop对应线程的run方法，run()方法里面是一个死循环，主要的逻辑是首先采用select检查是否有IO事件，如果有IO事件，就采用processSelectedKey()对IO事件进行处理，最后调用runAllTasks()处理任务队列中的任务。
protected void run() { ???while(true) { ???????boolean oldWakenUp = this.wakenUp.getAndSet(false); ???????try { ???????????if(this.hasTasks()) { ???????????????this.selectNow(); ???????????} else { ???????????????this.select(oldWakenUp); ???????????????if(this.wakenUp.get()) { ???????????????????this.selector.wakeup(); ???????????????} ???????????} ???????????this.cancelledKeys = 0; ???????????this.needsToSelectAgain = false; ???????????int t = this.ioRatio; ???????????if(t == 100) { ???????????????this.processSelectedKeys(); ???????????????this.runAllTasks(); ???????????} else { ???????????????long e = System.nanoTime(); ???????????????this.processSelectedKeys(); ???????????????long ioTime = System.nanoTime() - e; ???????????????this.runAllTasks(ioTime * (long)(100 - t) / (long)t); ???????????} ???????????if(this.isShuttingDown()) { ???????????????this.closeAll(); ???????????????if(this.confirmShutdown()) { ???????????????????return; ???????????????} ???????????} ???????} catch (Throwable var8) { ???????????logger.warn("Unexpected exception in the selector loop.", var8); ???????????try { ???????????????Thread.sleep(1000L); ???????????} catch (InterruptedException var7) { ???????????????; ???????????} ???????} ???}}这段代码中的ioRadio是控制执行IO事件和执行任务队列中的任务的一个事件比，默认是50，代表执行IO事件处理和执行任务队列的任务事件比是1:1。
1)使用select检测IO事件
通过Selector的select()方法可以选择已经准备就绪的通道 （这些通道包含你感兴趣的的事件）。比如你对读就绪的通道感兴趣，那么select()方法就会返回读事件已经就绪的那些通道。Java中的Selector几个重载的select()方法：
- int select()：阻塞到至少有一个通道在你注册的事件上就绪了。
- int select(long timeout)：和select()一样，但最长阻塞时间为timeout毫秒。
- int selectNow()：非阻塞，只要有通道就绪就立刻返回。
select()方法返回的int值表示有多少通道已经就绪,是自上次调用select()方法后有多少通道变成就绪状态。之前在select()调用时进入就绪的通道不会在本次调用中被记入，而在前一次select()调用进入就绪但现在已经不在处于就绪的通道也不会被记入。例如：首次调用select()方法，如果有一个通道变成就绪状态，返回了1，若再次调用select()方法，如果另一个通道就绪了，它会再次返回1。如果对第一个就绪的channel没有做任何操作，现在就有两个就绪的通道，但在每次select()方法调用之间，只有一个通道就绪了。
一旦调用select()方法，并且返回值不为0时，则可以通过调用Selector的selectedKeys()方法来访问已选择键集合 。如下：
Set selectedKeys=selector.selectedKeys();
Netty中首先判断任务队列是否为空，如果为空的话，就采用select(ltimeout)有超时设置的阻塞方法，如果不为空的话，就调用非阻塞的selectNow()方法，因为即使没有IO事件处理，也可以对任务队列中的任务进行处理。Netty中NioEventLoop的select和selectNow方法其实底层还是依靠selector的select方法。
void selectNow() throws IOException { ???try { ???????this.selector.selectNow(); ???} finally { ???????if(this.wakenUp.get()) { ???????????this.selector.wakeup(); ???????} ???}}private void select(boolean oldWakenUp) throws IOException { ???Selector selector = this.selector; ???try { ???????int e = 0; ???????long currentTimeNanos = System.nanoTime(); ???????long selectDeadLineNanos = currentTimeNanos + this.delayNanos(currentTimeNanos); ???????while(true) { ???????????long timeoutMillis = (selectDeadLineNanos - currentTimeNanos + 500000L) / 1000000L; ???????????if(timeoutMillis <= 0L) { ???????????????if(e == 0) { ???????????????????selector.selectNow(); ???????????????????e = 1; ???????????????} ???????????????break; ???????????} ???????????int selectedKeys = selector.select(timeoutMillis); ???????????++e; ???????????if(selectedKeys != 0 || oldWakenUp || this.wakenUp.get() || this.hasTasks() || this.hasScheduledTasks()) { ???????????????break; ???????????} ???????????if(Thread.interrupted()) { ???????????????if(logger.isDebugEnabled()) { ???????????????????logger.debug("Selector.select() returned prematurely because Thread.currentThread().interrupt() was called. Use NioEventLoop.shutdownGracefully() to shutdown the NioEventLoop."); ???????????????} ???????????????e = 1; ???????????????break; ???????????} ???????????long time = System.nanoTime(); ???????????if(time - TimeUnit.MILLISECONDS.toNanos(timeoutMillis) >= currentTimeNanos) { ???????????????e = 1; ???????????} else if(SELECTOR_AUTO_REBUILD_THRESHOLD > 0 && e >= SELECTOR_AUTO_REBUILD_THRESHOLD) { ???????????????logger.warn("Selector.select() returned prematurely {} times in a row; rebuilding selector.", Integer.valueOf(e)); ???????????????this.rebuildSelector(); ???????????????selector = this.selector; ???????????????selector.selectNow(); ???????????????e = 1; ???????????????break; ???????????} ???????????currentTimeNanos = time; ???????} ???????if(e > 3 && logger.isDebugEnabled()) { ???????????logger.debug("Selector.select() returned prematurely {} times in a row.", Integer.valueOf(e - 1)); ???????} ???} catch (CancelledKeyException var13) { ???????if(logger.isDebugEnabled()) { ???????????logger.debug(CancelledKeyException.class.getSimpleName() + " raised by a Selector - JDK bug?", var13); ???????} ???}}可以看到调用selectNow方法是直接调用java nio的select.selectNow方法，而Netty的select方法中有一个参数oldWakeUp记录当前操作是否是唤醒状态（不太清楚这个唤醒状态的作用），每次进行select操作之前，会将其标志位false，表示要进行select操作，而且是未唤醒状态。
Netty中的select方法首先是根据当前时间时间去计算截止时间，这里使用到了超时队列（超时队列的作用也不太清楚），然后根据截止时间去计算超时时间，如果超时时间小于0，就执行selectNow操作，并退出此次select操作，否则执行带有超时时间的select方法，如果返回的selectKey不等于0，也就是有channel在select上注册了，或者该select操作被唤醒了（？），或者任务队列中有了任务，定时任务队列中有了任务，都会break出来。
接下来的代码逻辑是避免JDK空轮询的，当JDK发生了空轮训，select会直接返回，这时并没有IO事件到达，也没有超过超时时间，这样会导致线程进入死循环，CPU利用率飙升至100%，JDK到现在也并没有解决这个问题。
而Netty是通过记录空轮询的次数，如果这个次数达到了一个上限，上限默认是512，那么就新建一个selector，将注册在老selector上的channel注册到新的selector上，并且关闭老的selector，将新的selector替代老的selector。Netty通过rebuildSelector方法重建selector。
public void rebuildSelector() { ???if(!this.inEventLoop()) { ???????this.execute(new Runnable() { ???????????public void run() { ???????????????NioEventLoop.this.rebuildSelector(); ???????????} ???????}); ???} else { ???????Selector oldSelector = this.selector; ???????if(oldSelector != null) { ???????????Selector newSelector; ???????????try { ???????????????newSelector = this.openSelector(); ???????????} catch (Exception var9) { ???????????????logger.warn("Failed to create a new Selector.", var9); ???????????????return; ???????????} ???????????int nChannels = 0; ???????????label69: ???????????while(true) { ???????????????try { ???????????????????Iterator t = oldSelector.keys().iterator(); ???????????????????while(true) { ???????????????????????if(!t.hasNext()) { ???????????????????????????break label69; ???????????????????????} ???????????????????????SelectionKey key = (SelectionKey)t.next(); ???????????????????????Object a = key.attachment(); ???????????????????????try { ???????????????????????????if(key.isValid() && key.channel().keyFor(newSelector) == null) { ???????????????????????????????int e = key.interestOps(); ???????????????????????????????key.cancel(); ???????????????????????????????SelectionKey var14 = key.channel().register(newSelector, e, a); ???????????????????????????????if(a instanceof AbstractNioChannel) { ???????????????????????????????????((AbstractNioChannel)a).selectionKey = var14; ???????????????????????????????} ???????????????????????????????++nChannels; ???????????????????????????} ???????????????????????} catch (Exception var11) { ???????????????????????????logger.warn("Failed to re-register a Channel to the new Selector.", var11); ???????????????????????????if(a instanceof AbstractNioChannel) { ???????????????????????????????AbstractNioChannel var13 = (AbstractNioChannel)a; ???????????????????????????????var13.unsafe().close(var13.unsafe().voidPromise()); ???????????????????????????} else { ???????????????????????????????NioTask task = (NioTask)a; ???????????????????????????????invokeChannelUnregistered(task, key, var11); ???????????????????????????} ???????????????????????} ???????????????????} ???????????????} catch (ConcurrentModificationException var12) { ???????????????????; ???????????????} ???????????} ???????????this.selector = newSelector; ???????????try { ???????????????oldSelector.close(); ???????????} catch (Throwable var10) { ???????????????if(logger.isWarnEnabled()) { ???????????????????logger.warn("Failed to close the old Selector.", var10); ???????????????} ???????????} ???????????logger.info("Migrated " + nChannels + " channel(s) to the new Selector."); ???????} ???}}2)processSelectedKey()
netty中selectedKey的优化
通过调用Selector的selectedKeys()方法来访问已选择键集合，此时返回的是HashSet。但是netty是通过反射的方式，将HashSet替换成数组pssSelectedKeysOptimized去处理IO事件。
private Selector openSelector() { ???AbstractSelector selector; ???try { ???????selector = this.provider.openSelector(); ???} catch (IOException var7) { ???????throw new ChannelException("failed to open a new selector", var7); ???} ???if(DISABLE_KEYSET_OPTIMIZATION) { ???????return selector; ???} else { ???????try { ???????????SelectedSelectionKeySet t = new SelectedSelectionKeySet(); ???????????Class selectorImplClass = Class.forName("sun.nio.ch.SelectorImpl", false, PlatformDependent.getSystemClassLoader()); ???????????if(!selectorImplClass.isAssignableFrom(selector.getClass())) { ???????????????return selector; ???????????} ???????????Field selectedKeysField = selectorImplClass.getDeclaredField("selectedKeys"); ???????????Field publicSelectedKeysField = selectorImplClass.getDeclaredField("publicSelectedKeys"); ???????????selectedKeysField.setAccessible(true); ???????????publicSelectedKeysField.setAccessible(true); ???????????selectedKeysField.set(selector, t); ???????????publicSelectedKeysField.set(selector, t); ???????????this.selectedKeys = t; ???????????logger.trace("Instrumented an optimized java.util.Set into: {}", selector); ???????} catch (Throwable var6) { ???????????this.selectedKeys = null; ???????????logger.trace("Failed to instrument an optimized java.util.Set into: {}", selector, var6); ???????} ???????return selector; ???}}首先会调用JDK的openSelector方法返回创建的selector，然后会判断是否要对keySet进行优化，通过判断DISABLE_KEYSET_OPTIMIZATION，是否要对keyset进行优化，默认是要对keyset进行优化的。这里的SelectedSelectionKeySet是优化过后的keyset，底层是通过两个数组加上两个数组的大小进行实现的，这样可以使得add操作达到o(1)的时间复杂度（但是是HashSet的add操作时间复杂度不也是o（1））嘛，
processSelectedKey调用processSelectedKeysOptimized
该方法的流程就是遍历数组中所有的selectedKey，一旦遍历完，就将该引用指向为空。获取每一个selectorKey对应的channel，然后通过调用processSelectedKey去处理该channel上感兴趣的事件。
private void processSelectedKeysOptimized(SelectionKey[] selectedKeys) { ???int i = 0;
//遍历SelectedKsys ???while(true) { ???????SelectionKey k = selectedKeys[i]; ???????if(k == null) { ???????????return; ???????} ???????selectedKeys[i] = null;
//获取selectKey对应的channel ???????Object a = k.attachment(); ???????if(a instanceof AbstractNioChannel) { ???????????processSelectedKey(k, (AbstractNioChannel)((AbstractNioChannel)a)); ???????} else { ???????????NioTask task = (NioTask)a; ???????????processSelectedKey(k, (NioTask)task); ???????} ???????if(this.needsToSelectAgain) { ???????????while(selectedKeys[i] != null) { ???????????????selectedKeys[i] = null; ???????????????++i; ???????????} ???????????this.selectAgain(); ???????????selectedKeys = this.selectedKeys.flip(); ???????????i = -1; ???????} ???????++i; ???}}这里处理selector上面的IO事件，底层其实都是通过channel的unsafe类进行操作的，这里read和accept事件对应的都是channel的read方法。
private static void processSelectedKey(SelectionKey k, AbstractNioChannel ch) { ???AbstractNioChannel.NioUnsafe unsafe = ch.unsafe(); ???if(!k.isValid()) { ???????unsafe.close(unsafe.voidPromise()); ???} else { ???????try { ???????????int ignored = k.readyOps();
//如果是read或者accept事件就对channel进行读操作 ???????????if((ignored & 17) != 0 || ignored == 0) { ???????????????unsafe.read(); ???????????????if(!ch.isOpen()) { ???????????????????return; ???????????????} ???????????}
//write事件 ???????????if((ignored & 4) != 0) { ???????????????ch.unsafe().forceFlush(); ???????????}
//connect事件 ???????????if((ignored & 8) != 0) { ???????????????int ops = k.interestOps(); ???????????????ops &= -9; ???????????????k.interestOps(ops); ???????????????unsafe.finishConnect(); ???????????} ???????} catch (CancelledKeyException var5) { ???????????unsafe.close(unsafe.voidPromise()); ???????} ???}}3)使用runAllTasks()执行任务队列中的事件
PriorityQueue（优先级队列），也是非线程安全的队列。
private <V> ScheduledFuture<V> schedule(final ScheduledFutureTask<V> task) { ???if(task == null) { ???????throw new NullPointerException("task"); ???} else { ???????if(this.inEventLoop()) { ???????????this.delayedTaskQueue.add(task); ???????} else { ???????????this.execute(new Runnable() { ???????????????public void run() { ???????????????????SingleThreadEventExecutor.this.delayedTaskQueue.add(task); ???????????????} ???????????}); ???????} ???????return task; ???}}runAllTask首先从定时任务队列中拉取定时任务，将需要执行的定时任务加入到普通任务队列中，并计算截止时间，然后循环的从普通任务队列中拉取任务，并执行任务，这里判断是否到达超时时间，是每相隔64个任务，就判断是否到达最大任务执行时间。为啥要每隔64个任务判断是否超时呢？因为nanoTime也是比较费时的。
protected boolean runAllTasks(long timeoutNanos) { ???this.fetchFromDelayedQueue(); ???Runnable task = this.pollTask(); ???if(task == null) { ???????return false; ???} else { ???????long deadline = ScheduledFutureTask.nanoTime() + timeoutNanos; ???????long runTasks = 0L; ???????long lastExecutionTime; ???????while(true) { ???????????try { ???????????????task.run(); ???????????} catch (Throwable var11) { ???????????????logger.warn("A task raised an exception.", var11); ???????????} ???????????++runTasks; ???????????if((runTasks & 63L) == 0L) { ???????????????lastExecutionTime = ScheduledFutureTask.nanoTime(); ???????????????if(lastExecutionTime >= deadline) { ???????????????????break; ???????????????} ???????????} ???????????task = this.pollTask(); ???????????if(task == null) { ???????????????lastExecutionTime = ScheduledFutureTask.nanoTime(); ???????????????break; ???????????} ???????} ???????this.lastExecutionTime = lastExecutionTime; ???????return true; ???}}从定时队列中拉取任务，这里拉取的任务是拉取截止时间不超过nanoTime的任务，将任务从定时任务队列中删除，将任务加入到普通任务队列中。这个while循环执行完成之后，所有需要执行的定时任务全部都加入到普通任务队列中。
private void fetchFromDelayedQueue() { ???long nanoTime = 0L; ???while (true) { ???????ScheduledFutureTask delayedTask = (ScheduledFutureTask) this.delayedTaskQueue.peek(); ???????if (delayedTask == null) { ???????????break; ???????} ???????if (nanoTime == 0L) { ???????????nanoTime = ScheduledFutureTask.nanoTime(); ???????} ???????if (delayedTask.deadlineNanos() > nanoTime) { ???????????break; ???????} ???????this.delayedTaskQueue.remove(); ???????this.taskQueue.add(delayedTask); ???}}定时任务队列是一个优先级队列，队列按照优先级进行排序，这里的优先级是每个任务的截止时间，队列是按照截止时间的早晚对任务进行排序的。
public int compareTo(Delayed o) { ???if(this == o) { ???????return 0; ???} else { ???????ScheduledFutureTask that = (ScheduledFutureTask)o; ???????long d = this.deadlineNanos() - that.deadlineNanos(); ???????if(d < 0L) { ???????????return -1; ???????} else if(d > 0L) { ???????????return 1; ???????} else if(this.id < that.id) { ???????????return -1; ???????} else if(this.id == that.id) { ???????????throw new Error(); ???????} else { ???????????return 1; ???????} ???}}总结：
1.默认情况下，NioEventLoopGroup会创建2*cpu个数的线程池，在调用NioEventLoop.execute(task)的时候，如果当前的NioEventLoop没有创建自己的线程，就会创建线程。
2.Netty如何解决JDK空轮训bug?通过计算空轮训操作的个数，这里的空轮训的判断是既没有IO事件的到达，也没有达到超时时间，如果空轮训的个数超过阈值（512），就会新建一个selector，将旧selector的selectorKey注册到新的selector上，将旧的selector关闭，用新的selector替代旧的selector。
3.Netty在所有外部线程调用NioEventLoop的操作时，如果通过InEventLoop判断是否在NioEventLoop所属的线程，如果不在通过startThread启动NioEventLoop的线程，并且将任务添加到NioEventLoop的任务队列中，所有NioEventLoop对应一个线程，其中的操作只会被一个线程所执行，实现了异步串行无锁化。
Netty源码分析之NioEventLoop执行流程
原文地址：https://www.cnblogs.com/xiaobaituyun/p/10801336.html