Spring Bean 循环依赖详解

• 同步Git Pages 博客： Spring Bean 循环依赖详解

什么是循环依赖？

循环依赖就是循环引用，就是两个或多个bean相互之间的持有对方。比如CircleA引用CircleB，CircleB引用CircleC，CircleC引用CircleA。

graph LR; A-->B B-->C C-->A

Spring如何解决循环依赖？

Spring容器循环依赖包括如下两种：

1. 构造器循环依赖
2. setter循环依赖

构造器循环依赖，表示通过构造器注入的循环依赖，此依赖是无法解决的，只能抛出BeanCurrentlyInCreationException异常。

原型模式下的依赖检查，只有单例模式下才会尝试解决循环依赖，原型模式下直接抛出异常BeanCurrentlyInCreationException异常。

AbstractBeanFactory抽象类：

// Fail if we're already creating this bean instance:
// We're assumably within a circular reference.
if (isPrototypeCurrentlyInCreation(beanName)) {
    throw new BeanCurrentlyInCreationException(beanName);
}

单例模式下，setter循环依赖可以解决(使用@Async的单例除外，特殊情况下会报错）

单例模式下，构造器循环依赖无法解决

原型模式下，循环依赖无法解决

Spring 循环依赖场景分析

不管是单例 Bean还是prototype Bean，注入Bean的逻辑起点都是在AbstractBeanFactory.java中的 getBean方法。

以a，b 两个bean互相依赖为例，假设Spring容器先加载 a 后加载 b，时序图如下：

sequenceDiagram AbstractBeanFactory.java->>DefaultSingletonBeanRegistry.java:getBean("a") AbstractBeanFactory.java->>DefaultSingletonBeanRegistry.java:getSingleton("a", true) DefaultSingletonBeanRegistry.java ->> DefaultSingletonBeanRegistry.java:this.singletonFactories.get("a")为空 DefaultSingletonBeanRegistry.java-->>AbstractBeanFactory.java: 返回 sharedInstance "a" 为空（缓存里没有） AbstractBeanFactory.java->>DefaultSingletonBeanRegistry.java: getSingleton("a", ObjectFactory singletonFactory) DefaultSingletonBeanRegistry.java -->> DefaultSingletonBeanRegistry.java: singletonFactory.getObject() DefaultSingletonBeanRegistry.java->>AbstractAutowireCapableBeanFactory: createBean("a", mbd, args) AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory: doCreateBean("a", mbdToUse, args) AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:createBeanInstance("a", mbd, args) 实例化 AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:addSingletonFactory("a", singletonFactory) 添加三级缓存工厂方法 AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:populateBean，寻找注入所依赖的"b" AbstractAutowireCapableBeanFactory-->> AbstractBeanFactory.java: getBean("b") AbstractBeanFactory.java->>DefaultSingletonBeanRegistry.java:getSingleton("b", true) DefaultSingletonBeanRegistry.java-->>AbstractBeanFactory.java: 返回 sharedInstance "b" 为空（缓存里没有） AbstractBeanFactory.java->>DefaultSingletonBeanRegistry.java: getSingleton("b", ObjectFactory singletonFactory) DefaultSingletonBeanRegistry.java -->> DefaultSingletonBeanRegistry.java: singletonFactory.getObject() DefaultSingletonBeanRegistry.java->>AbstractAutowireCapableBeanFactory: createBean("b", mbd, args) AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory: doCreateBean("b", mbdToUse, args) AbstractAutowireCapableBeanFactory-->> AbstractAutowireCapableBeanFactory:createBeanInstance("b", mbd, args) 实例化 AbstractAutowireCapableBeanFactory-->> AbstractAutowireCapableBeanFactory:addSingletonFactory("b", singletonFactory) 添加三级缓存工厂方法 AbstractAutowireCapableBeanFactory-->> AbstractAutowireCapableBeanFactory:populateBean，寻找注入所依赖的"a" AbstractAutowireCapableBeanFactory -->> AbstractBeanFactory.java: getBean("a") AbstractBeanFactory.java->> DefaultSingletonBeanRegistry.java:getBean("a") AbstractBeanFactory.java->>DefaultSingletonBeanRegistry.java:getSingleton("a", true) DefaultSingletonBeanRegistry.java -->>DefaultSingletonBeanRegistry.java:this.singletonFactories.get("a")（三级缓存有a值） DefaultSingletonBeanRegistry.java -->> DefaultSingletonBeanRegistry.java:singletonFactory.getObject() DefaultSingletonBeanRegistry.java->> AbstractAutowireCapableBeanFactory: getEarlyBeanReference 获取早期的Bean "a" AbstractAutowireCapableBeanFactory -->> AbstractAutowireCapableBeanFactory: 若有AOP，会提前处理AOP相关逻辑 AbstractAutowireCapableBeanFactory -->> DefaultSingletonBeanRegistry.java: exposedObject DefaultSingletonBeanRegistry.java -->>DefaultSingletonBeanRegistry.java: 把 Bean "a" 放到二级缓存，移除三级缓存 DefaultSingletonBeanRegistry.java -->> AbstractBeanFactory.java:返回提早暴露的Bean "a" AbstractBeanFactory.java ->> AbstractAutowireCapableBeanFactory: 返回Bean "a",将"a"注入到 "b"当中(b终于找到a了) AbstractAutowireCapableBeanFactory -->> AbstractAutowireCapableBeanFactory:initializeBean("b", exposedObject, mbd)，有AOP的话，会在这里处理 AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:Object earlySingletonReference = getSingleton("b", false) 为null，因为二级缓存中为空 AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:exposedObject != earlySingletonReference 返回 b AbstractAutowireCapableBeanFactory -->> DefaultSingletonBeanRegistry.java:返回 DefaultSingletonBeanRegistry.java -->> DefaultSingletonBeanRegistry.java:addSingleton("b", singletonObject)，放到一级缓存中 DefaultSingletonBeanRegistry.java -->> AbstractBeanFactory.java: 返回b (getSingleton("b", ObjectFactory singletonFactory)) AbstractBeanFactory.java ->> AbstractAutowireCapableBeanFactory: a终于找到了依赖的b AbstractAutowireCapableBeanFactory -->> AbstractAutowireCapableBeanFactory:initializeBean("a", exposedObject, mbd)，有AOP的话会在此处理 AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:Object earlySingletonReference = getSingleton("a", false)，这里返回 a 不是null，因为二级缓存里有a AbstractAutowireCapableBeanFactory-->>AbstractAutowireCapableBeanFactory:exposedObject = earlySingletonReference 返回 a AbstractAutowireCapableBeanFactory -->> AbstractBeanFactory.java:返回a

从上面的时序图可以看出主要的流程，涉及到 AbstractBeanFactory、DefaultSingletonBeanRegistry、AbstractAutowireCapableBeanFactory 三个类。

org.springframework.beans.factory.support.DefaultSingletonBeanRegistry.java

/**
	 * Return the (raw) singleton object registered under the given name.
	 * <p>Checks already instantiated singletons and also allows for an early
	 * reference to a currently created singleton (resolving a circular reference).
	 * @param beanName the name of the bean to look for
	 * @param allowEarlyReference whether early references should be created or not
	 * @return the registered singleton object, or {@code null} if none found
	 */
	protected Object getSingleton(String beanName, boolean allowEarlyReference) {
		Object singletonObject = this.singletonObjects.get(beanName);
		if (singletonObject == null && isSingletonCurrentlyInCreation(beanName)) {
			synchronized (this.singletonObjects) {
				singletonObject = this.earlySingletonObjects.get(beanName);
				if (singletonObject == null && allowEarlyReference) {
					ObjectFactory<?> singletonFactory = this.singletonFactories.get(beanName);
					if (singletonFactory != null) {
						singletonObject = singletonFactory.getObject();
						this.earlySingletonObjects.put(beanName, singletonObject);
						this.singletonFactories.remove(beanName);
					}
				}
			}
		}
		return (singletonObject != NULL_OBJECT ? singletonObject : null);
	}

bean加载时，从一开始的 调用 getSingleton() 方法从单例缓存中获取说起。

这个方法中涉及到三个缓存：

	/** Cache of singleton objects: bean name --> bean instance */
	private final Map<String, Object> singletonObjects = new ConcurrentHashMap<String, Object>(64);

	/** Cache of singleton factories: bean name --> ObjectFactory */
	private final Map<String, ObjectFactory<?>> singletonFactories = new HashMap<String, ObjectFactory<?>>(16);

	/** Cache of early singleton objects: bean name --> bean instance */
	private final Map<String, Object> earlySingletonObjects = new HashMap<String, Object>(16);

• singletonObjects 一级缓存，存放最终版Bean
• earlySingletonObjects 二级缓存，存放单例工厂创建的对象
• singletonFactories 三级缓存，存放单例工厂（是一个函数）

这就是我们常说的三级缓存，第一级是singletonObjects ，第二级是earlySingletonObjects ,第三级是 singletonFactories

getSingleton()方法整个过程如下：

首先从单例对象缓存singletonObjects 中获取，如果没有，且通过isSingletonCurrentlyInCreation() 方法判断当前要获取的bean正在处于创建中，那么就从earlySingletonObjects 二级缓存中获取，如果还没有

则从singletonFactories 三级缓存中获取。获取到的话，就通过singletonFactory.getObject()获取对象，并将其加入到二级缓存当中，从三级缓存当中删除。

这时候肯定有疑问，缓存中的数据是从哪里来的呢？

一直往下跟代码我们会发现doCreateBean() 方法

sharedInstance = getSingleton(beanName, new ObjectFactory<Object>() {
						@Override
						public Object getObject() throws BeansException {
							try {
								return createBean(beanName, mbd, args);
							}
							catch (BeansException ex) {
								// Explicitly remove instance from singleton cache: It might have been put there
								// eagerly by the creation process, to allow for circular reference resolution.
								// Also remove any beans that received a temporary reference to the bean.
								destroySingleton(beanName);
								throw ex;
							}
						}
					});

AbstractAutowireCapableBeanFactory 中：

/**
	 * Central method of this class: creates a bean instance,
	 * populates the bean instance, applies post-processors, etc.
	 * @see #doCreateBean
	 */
	@Override
	protected Object createBean(final String beanName, final RootBeanDefinition mbd, final Object[] args)
			throws BeanCreationException {

		if (logger.isDebugEnabled()) {
			logger.debug("Creating instance of bean '" + beanName + "'");
		}
		// Make sure bean class is actually resolved at this point.
		resolveBeanClass(mbd, beanName);

		// Prepare method overrides.
		try {
			mbd.prepareMethodOverrides();
		}
		catch (BeanDefinitionValidationException ex) {
			throw new BeanDefinitionStoreException(mbd.getResourceDescription(),
					beanName, "Validation of method overrides failed", ex);
		}

		try {
			// Give BeanPostProcessors a chance to return a proxy instead of the target bean instance.
			Object bean = resolveBeforeInstantiation(beanName, mbd);
			if (bean != null) {
				return bean;
			}
		}
		catch (Throwable ex) {
			throw new BeanCreationException(mbd.getResourceDescription(), beanName,
					"BeanPostProcessor before instantiation of bean failed", ex);
		}

		Object beanInstance = doCreateBean(beanName, mbd, args);
		if (logger.isDebugEnabled()) {
			logger.debug("Finished creating instance of bean '" + beanName + "'");
		}
		return beanInstance;
	}

doCreateBean()：

// Eagerly cache singletons to be able to resolve circular references
		// even when triggered by lifecycle interfaces like BeanFactoryAware.
		boolean earlySingletonExposure = (mbd.isSingleton() && this.allowCircularReferences &&
				isSingletonCurrentlyInCreation(beanName));
		if (earlySingletonExposure) {
			if (logger.isDebugEnabled()) {
				logger.debug("Eagerly caching bean '" + beanName +
						"' to allow for resolving potential circular references");
			}
			addSingletonFactory(beanName, new ObjectFactory<Object>() {
				@Override
				public Object getObject() throws BeansException {
					return getEarlyBeanReference(beanName, mbd, bean);
				}
			});
		}

我们可以看到 earlySingletonExposure是true 的时候，才会调用addSingletonFactory()将bean对象加入到三级缓存当中。

earlySingletonExposure 为true 的前提是：

• bean是单例的
• bean 允许提前引用
• 当前bean正在创建当中

addSingletonFactory() 代码如下:

	protected void addSingletonFactory(String beanName, ObjectFactory<?> singletonFactory) {
		Assert.notNull(singletonFactory, "Singleton factory must not be null");
		synchronized (this.singletonObjects) {
			if (!this.singletonObjects.containsKey(beanName)) {
				this.singletonFactories.put(beanName, singletonFactory);
				this.earlySingletonObjects.remove(beanName);
				this.registeredSingletons.add(beanName);
			}
		}
	}	

将bean添加到三级缓存，从二级缓存中删掉

那么一级缓存是在哪里添加的呢？

前边我们看到的 doCreateBean() 方法其实来自匿名内部类 new ObjectFactory<Object>() {} 重写的getObject() 方法

// Create bean instance.
				if (mbd.isSingleton()) {
					sharedInstance = getSingleton(beanName, new ObjectFactory<Object>() {
						@Override
						public Object getObject() throws BeansException {
							try {
								return createBean(beanName, mbd, args);
							}
							catch (BeansException ex) {
								// Explicitly remove instance from singleton cache: It might have been put there
								// eagerly by the creation process, to allow for circular reference resolution.
								// Also remove any beans that received a temporary reference to the bean.
								destroySingleton(beanName);
								throw ex;
							}
						}
					});
					bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd);
				}

在调用 getSingleton() 方法时， getSingleton() 方法中( DefaultSingletonBeanRegistry )，调用了上边重写的getObject() 方法后，后边调用了addSingleton(beanName, singletonObject);

addSingleton(beanName, singletonObject) 方法如下：

	/**
	 * Add the given singleton object to the singleton cache of this factory.
	 * <p>To be called for eager registration of singletons.
	 * @param beanName the name of the bean
	 * @param singletonObject the singleton object
	 */
	protected void addSingleton(String beanName, Object singletonObject) {
		synchronized (this.singletonObjects) {
			this.singletonObjects.put(beanName, (singletonObject != null ? singletonObject : NULL_OBJECT));
			this.singletonFactories.remove(beanName);
			this.earlySingletonObjects.remove(beanName);
			this.registeredSingletons.add(beanName);
		}
	}

添加到一级缓存，从二级缓存中删掉，从三级缓存中删掉，将已经创建成功的bean添加registeredSingletons

集合中。

总结

​ 至此，Spring 关于 singleton bean 循环依赖已经分析完毕了。所以我们基本上可以确定 Spring 解决循环依赖的方案了：Spring 在创建 bean 的时候并不是等它完全完成，而是在创建过程中将创建中的 bean 的 ObjectFactory 提前曝光（即加入到 singletonFactories 缓存中，代码在上文的**"addSingletonFactory() 代码"**），这样一旦下一个 bean 创建的时候需要依赖 bean ，则直接使用 ObjectFactory 的 getObject() 获取了，也就是 getSingleton() 中的代码片段了。

​ 以A，B两个Bean互相依赖为例，在获取A Bean的时候，发现三级缓存中都是空的，则走后续实例化Bean的逻辑，将自己的ObjectFactory放到三级缓存当中，再接着进行加载自己依赖的其他属性，这时候去加载B Bean，同样的流程，B Bean也将自己的ObjectFactory放入到三级缓存中，再接着加载自己依赖的A Bean。这时候从三级缓存中获取到 A Bean ，同时将获取到的Bean放到二级缓存中。继续走后面的初始化逻辑。这时B就加载到了早期的A Bean。A Bean 继续走后续逻辑就成了一个完整的Bean。

​ 到这里，关于 Spring 解决 bean 循环依赖就已经分析完毕了。最后来描述下就上面那个循环依赖 Spring 解决的过程：首先 A 完成初始化第一步并将自己提前曝光出来（通过 ObjectFactory 将自己提前曝光），在初始化的时候，发现自己依赖对象 B，此时就会去尝试 get(B)，这个时候发现 B 还没有被创建出来，然后 B 就走创建流程，在 B 初始化的时候，同样发现自己依赖 C，C 也没有被创建出来，这个时候 C 又开始初始化进程，但是在初始化的过程中发现自己依赖 A，于是尝试 get(A)，这个时候由于 A 已经添加至缓存中（一般都是添加至三级缓存 singletonFactories ），通过 ObjectFactory 提前曝光，所以可以通过 ObjectFactory.getObject() 拿到 A 对象，C 拿到 A 对象后顺利完成初始化，然后将自己添加到一级缓存中，回到 B ，B 也可以拿到 C 对象，完成初始化，A 可以顺利拿到 B 完成初始化。到这里整个链路就已经完成了初始化过程了。

参考

• 从源码层面深度剖析Spring循环依赖 | 京东云技术团队