SpringCache框架加载/拦截原理详解

官网文档

背景

项目A中需要多数据源的实现，比如UserDao.getAllUserList() 需要从readonly库中读取，但是UserDao.insert() 需要插入主(写)库

就需要在dao层的方法调用上面添加注解！

了解后知道-接口通过jdk代理(mybatis的mapper接口就是通过jdk代理动态生成的-> MapperFactoryBean.class )的，没办法被aop的拦截(注解配置的拦截)

//dao
 @Pointcut("@annotation(com.kaola.cs.data.common.aspect.DataSourceSelect)")
 public void dao() {
 }

然后碰巧接触了项目B，使用了SpringCache模块，但是Spring的Cache模块居然能够拦截(spring-cache也是通过注解拦截！！！)

引起了我的兴趣，就把源码翻了一遍

SpringCache的用途

与 mybatis 对比

1.  spring-cache 是基于spring的方法级别的，也就是说你方法做了啥不关心，它只负责缓存方法结果

mybatis 的缓存(CachingExecutor / BaseExecutor) 是基于数据库查询结果的缓存

 2.  spring-cache 可以配置各种类型的缓存介质(redis , ehcache , hashmap, 甚至db等等) -> 它仅仅是提供接口和默认实现，可以自己拓展

mybatis 的缓存是hashmap，单一！！lowb

SpringCache 的配置

1.注解(spring-boot) 2.xml配置

这里只讲注解，但是初始化的类都是一样的！！！

定义 CacheConfigure.java 就能直接使用

@EnableCaching
@Configuration
public class CacheConfigure extends CachingConfigurerSupport {
 @Override
 @Bean
 public CacheManager cacheManager() {
   SimpleCacheManager result = new SimpleCacheManager();
   List<Cache> caches = new ArrayList<>();
   caches.add(new ConcurrentMapCache("testCache"));
   result.setCaches(caches);
   return result;
 }

@Override
 @Bean
 public CacheErrorHandler errorHandler() {
   return new SimpleCacheErrorHandler();
 }

}

通过 @EnableCaching 注解可以找到 Spring-Cache 初始化的核心类

ProxyCachingConfiguration.java

@Configuration
@Role(BeanDefinition.ROLE_INFRASTRUCTURE)
public class ProxyCachingConfiguration extends AbstractCachingConfiguration {

@Bean(name = CacheManagementConfigUtils.CACHE_ADVISOR_BEAN_NAME)
@Role(BeanDefinition.ROLE_INFRASTRUCTURE)
public BeanFactoryCacheOperationSourceAdvisor cacheAdvisor() {
BeanFactoryCacheOperationSourceAdvisor advisor = new BeanFactoryCacheOperationSourceAdvisor();
advisor.setCacheOperationSource(cacheOperationSource());
advisor.setAdvice(cacheInterceptor());
if (this.enableCaching != null) {
 advisor.setOrder(this.enableCaching.<Integer>getNumber("order"));
}
return advisor;
}

@Bean
@Role(BeanDefinition.ROLE_INFRASTRUCTURE)
public CacheOperationSource cacheOperationSource() {
return new AnnotationCacheOperationSource();
}

@Bean
@Role(BeanDefinition.ROLE_INFRASTRUCTURE)
public CacheInterceptor cacheInterceptor() {
CacheInterceptor interceptor = new CacheInterceptor();
interceptor.configure(this.errorHandler, this.keyGenerator, this.cacheResolver, this.cacheManager);
interceptor.setCacheOperationSource(cacheOperationSource());
return interceptor;
}

}

通过注解，把3个类的bean 实例化: BeanFactoryCacheOperationSourceAdvisor 、CacheOperationSource 、CacheInterceptor
说一下这3个类的作用

BeanFactoryCacheOperationSourceAdvisor.java

/*
BeanFactoryCacheOperationSourceAdvisor 继承了 AbstractBeanFactoryPointcutAdvisor
在spring 中的效果就是，在每个bean的初始化时 (每个bean都会被加载成 advised 对象 -> 有 targetSource 和 Advisor[] 数组)
每个bean被调用方法的时候都是先遍历advisor的方法，然后在调用原生bean(也就是targetSource)的方法，实现了aop的效果

bean 加载的时候 BeanFactoryCacheOperationSourceAdvisor 的 getPointcut()-> 也就是 CacheOperationSourcePointcut 就会被获取，然后调用
CacheOperationSourcePointcut.matches()方法, 用来匹配对应的bean
假设bean 在 BeanFactoryCacheOperationSourceAdvisor 的扫描中 matchs() 方法返回了true
结果就是
 在每个bean的方法被调用的时候 CacheInterceptor 中的 invoke() 方法就会被调用

总结：
 spring-cache 也完成了aop一样的实现(spring-aop也是这样做的)

重点就是在 CacheOperationSourcePointcut.matchs() 方法中，怎么匹配接口的了 这里先不说后面具体介绍！！！！

*/
public class BeanFactoryCacheOperationSourceAdvisor extends AbstractBeanFactoryPointcutAdvisor {

@Nullable
private CacheOperationSource cacheOperationSource;

private final CacheOperationSourcePointcut pointcut = new CacheOperationSourcePointcut() {
@Override
@Nullable
protected CacheOperationSource getCacheOperationSource() {
 return cacheOperationSource;
}
};

/**
* Set the cache operation attribute source which is used to find cache
* attributes. This should usually be identical to the source reference
* set on the cache interceptor itself.
*/
public void setCacheOperationSource(CacheOperationSource cacheOperationSource) {
this.cacheOperationSource = cacheOperationSource;
}

/**
* Set the {@link ClassFilter} to use for this pointcut.
* Default is {@link ClassFilter#TRUE}.
*/
public void setClassFilter(ClassFilter classFilter) {
this.pointcut.setClassFilter(classFilter);
}

@Override
public Pointcut getPointcut() {
return this.pointcut;
}

}

CacheOperationSource.java 是个接口

实现类是 -> AnnotationCacheOperationSource.java 重点是父类 -> AbstractFallbackCacheOperationSource.java

讲解一下：

代码量很少，主要是 attributeCache 的封装使用，通过把 method - CacheOperation

然后在 CacheInterceptor.invoke() 的时候通过invocation 获取到 method-class 然后调用CacheOperationSource.getCacheOperations() 获取到 CacheOperation
CacheOperation 其实就是触发对应spring-cache 注解的操作-获取缓存的实现了

public abstract class AbstractFallbackCacheOperationSource implements CacheOperationSource {

/**
* Canonical value held in cache to indicate no caching attribute was
* found for this method and we don't need to look again.
*/
private static final Collection<CacheOperation> NULL_CACHING_ATTRIBUTE = Collections.emptyList();

/**
* Logger available to subclasses.
* <p>As this base class is not marked Serializable, the logger will be recreated
* after serialization - provided that the concrete subclass is Serializable.
*/
protected final Log logger = LogFactory.getLog(getClass());

/**
* Cache of CacheOperations, keyed by method on a specific target class.
* <p>As this base class is not marked Serializable, the cache will be recreated
* after serialization - provided that the concrete subclass is Serializable.
*/
private final Map<Object, Collection<CacheOperation>> attributeCache = new ConcurrentHashMap<>(1024);

/**
* Determine the caching attribute for this method invocation.
* <p>Defaults to the class's caching attribute if no method attribute is found.
* @param method the method for the current invocation (never {@code null})
* @param targetClass the target class for this invocation (may be {@code null})
* @return {@link CacheOperation} for this method, or {@code null} if the method
* is not cacheable
*/
@Override
@Nullable
public Collection<CacheOperation> getCacheOperations(Method method, @Nullable Class<?> targetClass) {
if (method.getDeclaringClass() == Object.class) {
 return null;
}

Object cacheKey = getCacheKey(method, targetClass);
Collection<CacheOperation> cached = this.attributeCache.get(cacheKey);

if (cached != null) {
 return (cached != NULL_CACHING_ATTRIBUTE ? cached : null);
}
else {
 Collection<CacheOperation> cacheOps = computeCacheOperations(method, targetClass);
 if (cacheOps != null) {
 if (logger.isTraceEnabled()) {
  logger.trace("Adding cacheable method '" + method.getName() + "' with attribute: " + cacheOps);
 }
 this.attributeCache.put(cacheKey, cacheOps);
 }
 else {
 this.attributeCache.put(cacheKey, NULL_CACHING_ATTRIBUTE);
 }
 return cacheOps;
}
}

/**
* Determine a cache key for the given method and target class.
* <p>Must not produce same key for overloaded methods.
* Must produce same key for different instances of the same method.
* @param method the method (never {@code null})
* @param targetClass the target class (may be {@code null})
* @return the cache key (never {@code null})
*/
protected Object getCacheKey(Method method, @Nullable Class<?> targetClass) {
return new MethodClassKey(method, targetClass);
}

@Nullable
private Collection<CacheOperation> computeCacheOperations(Method method, @Nullable Class<?> targetClass) {
// Don't allow no-public methods as required.
if (allowPublicMethodsOnly() && !Modifier.isPublic(method.getModifiers())) {
 return null;
}

// The method may be on an interface, but we need attributes from the target class.
// If the target class is null, the method will be unchanged.
Method specificMethod = AopUtils.getMostSpecificMethod(method, targetClass);

// First try is the method in the target class.
Collection<CacheOperation> opDef = findCacheOperations(specificMethod);
if (opDef != null) {
 return opDef;
}

// Second try is the caching operation on the target class.
opDef = findCacheOperations(specificMethod.getDeclaringClass());
if (opDef != null && ClassUtils.isUserLevelMethod(method)) {
 return opDef;
}

if (specificMethod != method) {
 // Fallback is to look at the original method.
 opDef = findCacheOperations(method);
 if (opDef != null) {
 return opDef;
 }
 // Last fallback is the class of the original method.
 opDef = findCacheOperations(method.getDeclaringClass());
 if (opDef != null && ClassUtils.isUserLevelMethod(method)) {
 return opDef;
 }
}

return null;
}

/**
* Subclasses need to implement this to return the caching attribute for the
* given class, if any.
* @param clazz the class to retrieve the attribute for
* @return all caching attribute associated with this class, or {@code null} if none
*/
@Nullable
protected abstract Collection<CacheOperation> findCacheOperations(Class<?> clazz);

/**
* Subclasses need to implement this to return the caching attribute for the
* given method, if any.
* @param method the method to retrieve the attribute for
* @return all caching attribute associated with this method, or {@code null} if none
*/
@Nullable
protected abstract Collection<CacheOperation> findCacheOperations(Method method);

/**
* Should only public methods be allowed to have caching semantics?
* <p>The default implementation returns {@code false}.
*/
protected boolean allowPublicMethodsOnly() {
return false;
}

}

！！！！  CacheOperationSourcePointcut.java 的 matchs() 方法

用来判断类是不是符合spring-cache 拦截条件 也就是 @Cachable @CachePut 等等的注解怎么识别的地方

经过跟踪代码发现是 AnnotationCacheOperationSource.findCacheOperations() 调用的

省略部分代码....

public class AnnotationCacheOperationSource extends AbstractFallbackCacheOperationSource implements Serializable {

private final Set<CacheAnnotationParser> annotationParsers;

@Override
@Nullable
protected Collection<CacheOperation> findCacheOperations(Class<?> clazz) {
return determineCacheOperations(parser -> parser.parseCacheAnnotations(clazz));
}

@Override
@Nullable
protected Collection<CacheOperation> findCacheOperations(Method method) {
return determineCacheOperations(parser -> parser.parseCacheAnnotations(method));
}

/**
* Determine the cache operation(s) for the given {@link CacheOperationProvider}.
* <p>This implementation delegates to configured
* {@link CacheAnnotationParser CacheAnnotationParsers}
* for parsing known annotations into Spring's metadata attribute class.
* <p>Can be overridden to support custom annotations that carry caching metadata.
* @param provider the cache operation provider to use
* @return the configured caching operations, or {@code null} if none found
*/
@Nullable
protected Collection<CacheOperation> determineCacheOperations(CacheOperationProvider provider) {
Collection<CacheOperation> ops = null;
for (CacheAnnotationParser annotationParser : this.annotationParsers) {
 Collection<CacheOperation> annOps = provider.getCacheOperations(annotationParser);
 if (annOps != null) {
 if (ops == null) {
  ops = annOps;
 }
 else {
  Collection<CacheOperation> combined = new ArrayList<>(ops.size() + annOps.size());
  combined.addAll(ops);
  combined.addAll(annOps);
  ops = combined;
 }
 }
}
return ops;
}
}

然后就是注解的解析方法 SpringCacheAnnotationParser.java

代码很简单-就不多说了

@Nullable
private Collection<CacheOperation> parseCacheAnnotations(
 DefaultCacheConfig cachingConfig, AnnotatedElement ae, boolean localOnly) {

Collection<? extends Annotation> anns = (localOnly ?
 AnnotatedElementUtils.getAllMergedAnnotations(ae, CACHE_OPERATION_ANNOTATIONS) :
 AnnotatedElementUtils.findAllMergedAnnotations(ae, CACHE_OPERATION_ANNOTATIONS));
if (anns.isEmpty()) {
 return null;
}

final Collection<CacheOperation> ops = new ArrayList<>(1);
anns.stream().filter(ann -> ann instanceof Cacheable).forEach(
 ann -> ops.add(parseCacheableAnnotation(ae, cachingConfig, (Cacheable) ann)));
anns.stream().filter(ann -> ann instanceof CacheEvict).forEach(
 ann -> ops.add(parseEvictAnnotation(ae, cachingConfig, (CacheEvict) ann)));
anns.stream().filter(ann -> ann instanceof CachePut).forEach(
 ann -> ops.add(parsePutAnnotation(ae, cachingConfig, (CachePut) ann)));
anns.stream().filter(ann -> ann instanceof Caching).forEach(
 ann -> parseCachingAnnotation(ae, cachingConfig, (Caching) ann, ops));
return ops;
}

总结

1.spring-cache 实现了 AbstractBeanFactoryPointcutAdvisor 提供 CacheOperationSourcePointcut (PointCut) 作切点判断，提供 CacheInterceptor (MethodInterceptor) 作方法拦截

2.spring-cache 提供 CacheOperationSource 作为 method 对应 CacheOperation(缓存操作) 的查询和加载

3.spring-cache 通过 SpringCacheAnnotationParser 来解析自己定义的 @Cacheable @CacheEvict @Caching 等注解类
所以 spring-cache 不使用 aspectj 的方式，通过 CacheOperationSource.getCacheOperations() 方式可以使jdk代理的类也能匹配到

jdk代理的类的匹配

代码类在 CacheOperationSource.getCacheOperations()

重点在于 targetClass 和 method ，如果是对应的 dao.xxx() 就能matchs() 并且拦截

CacheInterceptor -> CacheAspectSupport.execute() 方法

// 代码自己看吧。也很简单 -> 结果就是spring-cache 也可以拦截到mybatis的dao层接口，进行缓存

@Nullable
protected Object execute(CacheOperationInvoker invoker, Object target, Method method, Object[] args) {
// Check whether aspect is enabled (to cope with cases where the AJ is pulled in automatically)
if (this.initialized) {
 Class<?> targetClass = getTargetClass(target);
 CacheOperationSource cacheOperationSource = getCacheOperationSource();
 if (cacheOperationSource != null) {
 Collection<CacheOperation> operations = cacheOperationSource.getCacheOperations(method, targetClass);
 if (!CollectionUtils.isEmpty(operations)) {
  return execute(invoker, method,
   new CacheOperationContexts(operations, method, args, target, targetClass));
 }
 }
}

return invoker.invoke();
}

来源：https://juejin.im/post/5cc4445fe51d455c6049b7cd