上一篇文章《Tomcat在SpringBoot中是如何啓動的》從main方法啓動提及,窺探了SpringBoot是如何啓動Tomcat的,在分析Tomcat中咱們重點提到了,Tomcat主要包括2個組件,鏈接器(Connector)和容器(Container)以及他們的內部結構圖,那麼今天咱們來分析下Tomcat中的鏈接器是怎麼設計的以及它的做用是什麼。java
說明:本文tomcat版本是9.0.21,不建議零基礎讀者閱讀。apache
既然是來解析鏈接器(Connector),那麼咱們直接從源碼入手,後面全部源碼我會剔除不重要部分,因此會忽略大部分源碼細節,只關注流程。源碼以下(高能預警,大量代碼):編程
public class Connector extends LifecycleMBeanBase { public Connector() { this("org.apache.coyote.http11.Http11NioProtocol"); } public Connector(String protocol) { boolean aprConnector = AprLifecycleListener.isAprAvailable() && AprLifecycleListener.getUseAprConnector(); if ("HTTP/1.1".equals(protocol) || protocol == null) { if (aprConnector) { protocolHandlerClassName = "org.apache.coyote.http11.Http11AprProtocol"; } else { protocolHandlerClassName = "org.apache.coyote.http11.Http11NioProtocol"; } } else if ("AJP/1.3".equals(protocol)) { if (aprConnector) { protocolHandlerClassName = "org.apache.coyote.ajp.AjpAprProtocol"; } else { protocolHandlerClassName = "org.apache.coyote.ajp.AjpNioProtocol"; } } else { protocolHandlerClassName = protocol; } // Instantiate protocol handler ProtocolHandler p = null; try { Class<?> clazz = Class.forName(protocolHandlerClassName); p = (ProtocolHandler) clazz.getConstructor().newInstance(); } catch (Exception e) { log.error(sm.getString( "coyoteConnector.protocolHandlerInstantiationFailed"), e); } finally { this.protocolHandler = p; } // Default for Connector depends on this system property setThrowOnFailure(Boolean.getBoolean("org.apache.catalina.startup.EXIT_ON_INIT_FAILURE")); }
咱們來看看Connector的構造方法,其實只作了一件事情,就是根據協議設置對應的ProtocolHandler
,根據名稱咱們知道,這是協議處理類,因此鏈接器內部的一個重要子模塊就是ProtocolHandler
。tomcat
咱們看到Connector
繼承了LifecycleMBeanBase
,咱們來看看Connector
的最終繼承關係:網絡
咱們看到最終實現的是Lifecycle
接口,咱們看看這個接口是何方神聖。我把其接口的註釋拿下來解釋下app
/** * Common interface for component life cycle methods. Catalina components * may implement this interface (as well as the appropriate interface(s) for * the functionality they support) in order to provide a consistent mechanism * to start and stop the component. * start() * ----------------------------- * | | * | init() | * NEW -»-- INITIALIZING | * | | | | ------------------«----------------------- * | | |auto | | | * | | \|/ start() \|/ \|/ auto auto stop() | * | | INITIALIZED --»-- STARTING_PREP --»- STARTING --»- STARTED --»--- | * | | | | | * | |destroy()| | | * | --»-----«-- ------------------------«-------------------------------- ^ * | | | | * | | \|/ auto auto start() | * | | STOPPING_PREP ----»---- STOPPING ------»----- STOPPED -----»----- * | \|/ ^ | ^ * | | stop() | | | * | | -------------------------- | | * | | | | | * | | | destroy() destroy() | | * | | FAILED ----»------ DESTROYING ---«----------------- | * | | ^ | | * | | destroy() | |auto | * | --------»----------------- \|/ | * | DESTROYED | * | | * | stop() | * ----»-----------------------------»------------------------------ * * Any state can transition to FAILED. * * Calling start() while a component is in states STARTING_PREP, STARTING or * STARTED has no effect. * * Calling start() while a component is in state NEW will cause init() to be * called immediately after the start() method is entered. * * Calling stop() while a component is in states STOPPING_PREP, STOPPING or * STOPPED has no effect. * * Calling stop() while a component is in state NEW transitions the component * to STOPPED. This is typically encountered when a component fails to start and * does not start all its sub-components. When the component is stopped, it will * try to stop all sub-components - even those it didn't start. * * Attempting any other transition will throw {@link LifecycleException}. * * </pre> * The {@link LifecycleEvent}s fired during state changes are defined in the * methods that trigger the changed. No {@link LifecycleEvent}s are fired if the * attempted transition is not valid.
這段註釋翻譯就是,這個接口是提供給組件聲明週期管理的,而且提供了聲明週期流轉圖。這裏咱們只須要知道正常流程便可:dom
New--->Init()---->Start()---->Stop()--->Destory()異步
根據上面的生命週期說明,咱們能夠知道鏈接器(Connector
)就是按照如此的聲明週期管理的,因此咱們找到了線索,因此鏈接器確定會先初始化而後再啓動。咱們查看其initInternal()
方法能夠知道鏈接器初始化作了什麼事情,源碼以下:socket
@Override protected void initInternal() throws LifecycleException { super.initInternal(); if (protocolHandler == null) { throw new LifecycleException( sm.getString("coyoteConnector.protocolHandlerInstantiationFailed")); } // Initialize adapter adapter = new CoyoteAdapter(this); protocolHandler.setAdapter(adapter); if (service != null) { protocolHandler.setUtilityExecutor(service.getServer().getUtilityExecutor()); } // Make sure parseBodyMethodsSet has a default if (null == parseBodyMethodsSet) { setParseBodyMethods(getParseBodyMethods()); } if (protocolHandler.isAprRequired() && !AprLifecycleListener.isInstanceCreated()) { throw new LifecycleException(sm.getString("coyoteConnector.protocolHandlerNoAprListener", getProtocolHandlerClassName())); } if (protocolHandler.isAprRequired() && !AprLifecycleListener.isAprAvailable()) { throw new LifecycleException(sm.getString("coyoteConnector.protocolHandlerNoAprLibrary", getProtocolHandlerClassName())); } if (AprLifecycleListener.isAprAvailable() && AprLifecycleListener.getUseOpenSSL() && protocolHandler instanceof AbstractHttp11JsseProtocol) { AbstractHttp11JsseProtocol<?> jsseProtocolHandler = (AbstractHttp11JsseProtocol<?>) protocolHandler; if (jsseProtocolHandler.isSSLEnabled() && jsseProtocolHandler.getSslImplementationName() == null) { // OpenSSL is compatible with the JSSE configuration, so use it if APR is available jsseProtocolHandler.setSslImplementationName(OpenSSLImplementation.class.getName()); } } try { protocolHandler.init(); } catch (Exception e) { throw new LifecycleException( sm.getString("coyoteConnector.protocolHandlerInitializationFailed"), e); } } }
根據上面源碼,咱們發現主要是處理protocolHandler
並初始化它,同時咱們注意到了protocolHandler
設置了一個適配器,咱們看看這個適配器是作啥的,跟蹤源碼以下:async
/** * The adapter, used to call the connector. * * @param adapter The adapter to associate */ public void setAdapter(Adapter adapter);
這個註釋已經說的很直白了,這個適配器就是用來調用鏈接器的。咱們再繼續看看protocolHandler
的初始化方法
/** * Endpoint that provides low-level network I/O - must be matched to the * ProtocolHandler implementation (ProtocolHandler using NIO, requires NIO * Endpoint etc.). */ private final AbstractEndpoint<S,?> endpoint; public void init() throws Exception { if (getLog().isInfoEnabled()) { getLog().info(sm.getString("abstractProtocolHandler.init", getName())); logPortOffset(); } if (oname == null) { // Component not pre-registered so register it oname = createObjectName(); if (oname != null) { Registry.getRegistry(null, null).registerComponent(this, oname, null); } } if (this.domain != null) { rgOname = new ObjectName(domain + ":type=GlobalRequestProcessor,name=" + getName()); Registry.getRegistry(null, null).registerComponent( getHandler().getGlobal(), rgOname, null); } String endpointName = getName(); endpoint.setName(endpointName.substring(1, endpointName.length()-1)); endpoint.setDomain(domain); endpoint.init(); }
這裏出現了一個新的對象,endpoint
,根據註釋咱們能夠知道endpoint
是用來處理網絡IO的,並且必須匹配到指定的子類(好比Nio,就是NioEndPoint處理)。endpoint.init()
實際上就是作一些網絡的配置,而後就是初始化完畢了。根據咱們上面的週期管理,咱們知道init()
後就是start()
,因此咱們查看Connector
的start()
源碼:
protected void startInternal() throws LifecycleException { // Validate settings before starting if (getPortWithOffset() < 0) { throw new LifecycleException(sm.getString( "coyoteConnector.invalidPort", Integer.valueOf(getPortWithOffset()))); } setState(LifecycleState.STARTING); try { protocolHandler.start(); } catch (Exception e) { throw new LifecycleException( sm.getString("coyoteConnector.protocolHandlerStartFailed"), e); } }
其實就是主要調用 protocolHandler.start()
方法,繼續跟蹤,爲了方便表述,我會把接下來的代碼統一放在一塊兒說明,代碼以下:
//1.類:AbstractProtocol implements ProtocolHandler, MBeanRegistration public void start() throws Exception { // 省略部分代碼 endpoint.start(); } //2. 類:AbstractEndPoint public final void start() throws Exception { // 省略部分代碼 startInternal(); } /**3.類:NioEndPoint extends AbstractJsseEndpoint<NioChannel,SocketChannel> * Start the NIO endpoint, creating acceptor, poller threads. */ @Override public void startInternal() throws Exception { //省略部分代碼 // Start poller thread poller = new Poller(); Thread pollerThread = new Thread(poller, getName() + "-ClientPoller"); pollerThread.setPriority(threadPriority); pollerThread.setDaemon(true); pollerThread.start(); startAcceptorThread(); } }
到這裏,其實整個啓動代碼就完成了,咱們看到最後是在NioEndPoint
建立了一個Poller
,而且啓動它,這裏須要補充說明下,這裏只是以NioEndPoint爲示列,其實Tomcat 主要提供了三種實現,分別是AprEndPoint
,NioEndPoint
,Nio2EndPoint
,這裏表示了tomcat支持的I/O模型:
APR:採用 Apache 可移植運行庫實現,它根據不一樣操做系統,分別用c重寫了大部分IO和系統線程操做模塊,聽說性能要比其餘模式要好(未實測)。
NIO:非阻塞 I/O
NIO.2:異步 I/O
上述代碼主要是開啓兩個線程,一個是Poller,一個是開啓Acceptor,既然是線程,核心的代碼確定是run方法
,咱們來查看源碼,代碼以下:
//4.類:Acceptor<U> implements Runnable public void run() { //省略了部分代碼 U socket = null; socket = endpoint.serverSocketAccept(); // Configure the socket if (endpoint.isRunning() && !endpoint.isPaused()) { // setSocketOptions() will hand the socket off to // an appropriate processor if successful //核心邏輯 if (!endpoint.setSocketOptions(socket)) { endpoint.closeSocket(socket); } } else { endpoint.destroySocket(socket); } state = AcceptorState.ENDED; } //5.類:NioEndpoint protected boolean setSocketOptions(SocketChannel socket) { // Process the connection //省略部分代碼 try { // Disable blocking, polling will be used socket.configureBlocking(false); Socket sock = socket.socket(); socketProperties.setProperties(sock); NioSocketWrapper socketWrapper = new NioSocketWrapper(channel, this); channel.setSocketWrapper(socketWrapper); socketWrapper.setReadTimeout(getConnectionTimeout()); socketWrapper.setWriteTimeout(getConnectionTimeout()); socketWrapper.setKeepAliveLeft(NioEndpoint.this.getMaxKeepAliveRequests()); socketWrapper.setSecure(isSSLEnabled()); //核心邏輯 poller.register(channel, socketWrapper); return true; }
這裏能夠發現Acceptor
主要就是接受socket
,而後把它註冊到poller
中,咱們繼續看看是如何註冊的。
/**6.類NioEndpoint * Registers a newly created socket with the poller. * * @param socket The newly created socket * @param socketWrapper The socket wrapper */ public void register(final NioChannel socket, final NioSocketWrapper socketWrapper) { socketWrapper.interestOps(SelectionKey.OP_READ);//this is what OP_REGISTER turns into. PollerEvent r = null; if (eventCache != null) { r = eventCache.pop(); } if (r == null) { r = new PollerEvent(socket, OP_REGISTER); } else { r.reset(socket, OP_REGISTER); } addEvent(r); } /** 7.類:PollerEvent implements Runnable public void run() { //省略部分代碼 socket.getIOChannel().register(socket.getSocketWrapper().getPoller().getSelector(), SelectionKey.OP_READ, socket.getSocketWrapper()); }
這裏發現最終就是採用NIO模型把其註冊到通道中。(這裏涉及NIO網絡編程知識,不瞭解的同窗能夠傳送這裏)。那麼註冊完畢後,咱們看看Poller作了什麼事情。
*/ /**8.類:NioEndPoint內部類 Poller implements Runnable **/ @Override public void run() { // Loop until destroy() is called while (true) { //省略部分代碼 Iterator<SelectionKey> iterator = keyCount > 0 ? selector.selectedKeys().iterator() : null; // Walk through the collection of ready keys and dispatch // any active event. while (iterator != null && iterator.hasNext()) { SelectionKey sk = iterator.next(); NioSocketWrapper socketWrapper = (NioSocketWrapper) sk.attachment(); // Attachment may be null if another thread has called // cancelledKey() if (socketWrapper == null) { iterator.remove(); } else { iterator.remove(); //sock處理 processKey(sk, socketWrapper); } } //省略部分代碼 }
這個就是經過selector把以前註冊的事件取出來,從而完成了調用。
//9.類: NioEndPoint內部類 Poller implements Runnable protected void processKey(SelectionKey sk, NioSocketWrapper socketWrapper) { //省略大部分代碼 processSocket(socketWrapper, SocketEvent.OPEN_WRITE, true) } //10.類:AbstractEndPoint public boolean processSocket(SocketWrapperBase<S> socketWrapper, SocketEvent event, boolean dispatch) { //省略部分代碼 Executor executor = getExecutor(); if (dispatch && executor != null) { executor.execute(sc); } else { sc.run(); } return true; } //11.類:SocketProcessorBase implements Runnable public final void run() { synchronized (socketWrapper) { // It is possible that processing may be triggered for read and // write at the same time. The sync above makes sure that processing // does not occur in parallel. The test below ensures that if the // first event to be processed results in the socket being closed, // the subsequent events are not processed. if (socketWrapper.isClosed()) { return; } doRun(); } } //類:12.NioEndPoint extends AbstractJsseEndpoint<NioChannel,SocketChannel> protected void doRun() { //省略部分代碼 if (handshake == 0) { SocketState state = SocketState.OPEN; // Process the request from this socket if (event == null) { state = getHandler().process(socketWrapper, SocketEvent.OPEN_READ); } else { state = getHandler().process(socketWrapper, event); } if (state == SocketState.CLOSED) { poller.cancelledKey(key, socketWrapper); } } }
Poller
調用的run
方法或者用Executor線程池去執行run()
,最終調用都是各個子EndPoint
中的doRun()
方法,最終會取一個Handler
去處理socketWrapper
。繼續看源碼:
//類:13.AbstractProtocol內部類ConnectionHandler implements AbstractEndpoint.Handler<S> public SocketState process(SocketWrapperBase<S> wrapper, SocketEvent status) { //省略部分代碼 state = processor.process(wrapper, status); return SocketState.CLOSED; } //類:14.AbstractProcessorLight implements Processor public SocketState process(SocketWrapperBase<?> socketWrapper, SocketEvent status) throws IOException { //省略部分代碼 state = service(socketWrapper); return state; }
這部分源碼代表最終調用的process是經過一個Processor
接口的實現類來完成的,這裏最終也是會調用到各個子類中,那麼這裏的處理器其實就是處理應用協議,咱們能夠查看AbstractProcessorLight
的實現類,分別有AjpProcessor
、Http11Processor
、StreamProcessor
,分別表明tomcat支持三種應用層協議,分別是:
這裏咱們以經常使用的HTTP1.1爲例,繼續看源碼:
//類:15. Http11Processor extends AbstractProcessor public SocketState service(SocketWrapperBase<?> socketWrapper) throws IOException { //省略大部分代碼 getAdapter().service(request, response); //省略大部分代碼 } //類:16 CoyoteAdapter implements Adapter public void service(org.apache.coyote.Request req, org.apache.coyote.Response res) throws Exception { Request request = (Request) req.getNote(ADAPTER_NOTES); Response response = (Response) res.getNote(ADAPTER_NOTES); postParseSuccess = postParseRequest(req, request, res, response); if (postParseSuccess) { //check valves if we support async request.setAsyncSupported( connector.getService().getContainer().getPipeline().isAsyncSupported()); // Calling the container connector.getService().getContainer().getPipeline().getFirst().invoke( request, response); } }
這裏咱們發現協議處理器最終會調用適配器(CoyoteAdapter
),而適配器最終的工做是轉換Request
和Response
對象爲HttpServletRequest
和HttpServletResponse
,從而能夠去調用容器,到這裏整個鏈接器的流程和做用咱們就已經分析完了。
那麼咱們來回憶下整個流程,我畫了一張時序圖來講明:
這張圖包含了兩個流程,一個是組件的初始化,一個是調用的流程。鏈接器(Connector)主要初始化了兩個組件,ProtcoHandler
和EndPoint
,可是咱們從代碼結構發現,他們兩個是父子關係,也就是說ProtcoHandler
包含了EndPoint
。後面的流程就是各個子組件的調用鏈關係,總結來講就是Acceptor
負責接收請求,而後註冊到Poller
,Poller
負責處理請求,而後調用processor
處理器來處理,最後把請求轉成符合Servlet
規範的request
和response
去調用容器(Container
)。
咱們流程梳理清楚了,接下來咱們來結構化的梳理下:
回到鏈接器(Connector
)是源碼,咱們發現,上述說的模塊只有ProtocolHandler
和Adapter
兩個屬於鏈接器中,也就是說,鏈接器只包含了這兩大子模塊,那麼後續的EndPoint
、Acceptor
、Poller
、Processor
都是ProtocolHandler
的子模塊。 而Acceptor
和Poller
兩個模塊的核心功能都是在EndPoint
中完成的,因此是其子模塊,而Processor
比較獨立,因此它和EndPoint
是一個級別的子模塊。
咱們用圖來講明下上述的關係:
根據上圖咱們能夠知道,鏈接器主要負責處理鏈接請求,而後經過適配器調用容器。那麼具體流程細化能夠以下:
Acceptor
監聽網絡請求,獲取請求。Poller
獲取到監聽的請求提交線程池進行處理。Processor
根據具體的應用協議(HTTP/AJP)來生成Tomcat Request對象。Adapter
把Request對象轉換成Servlet標準的Request對象,調用容器。咱們從鏈接器的源碼,一步一步解析,分析了鏈接器主要包含了兩大模塊,ProtocolHandler
和Adapter
。ProtocolHandler
主要包含了Endpoint
模塊和Processor
模塊。Endpoint
模塊主要的做用是鏈接的處理,它委託了Acceptor
子模塊進行鏈接的監聽和註冊,委託子模塊Poller
進行鏈接的處理;而Processor
模塊主要是應用協議的處理,最後提交給Adapter
進行對象的轉換,以即可以調用容器(Container)。另外咱們也在分析源碼的過程當中補充了一些額外知識點: