一说到缓存，有没有脑海里立马想到的是guava cache、ehcache、redis、memerycache等等这些和缓存相关的技术实现，计算机专业出生可能还会想到cpu的一级缓存、二级缓存、三级缓存等等。以说到缓存，层出不穷的技术可以列举非常多，那么到底什么是缓存？guava cache是缓存么？是也不是；redis是缓存的，是也不是，那么下面就说下我对缓存的看法

我对缓存的认识就是：当多个系统(组件、甚至是软件分层)之间，处理一件事情的时候，有速度差异的时候，为了提高速度，将上一层的处理结果记录下来，后续的访问就直接从本层拿到记录下来的结果，而不需要去上一层去获取结果，从而提高整体的响应速度。

从这个角度来看，理解下cpu的高速缓存：

高速缓存产生的背景就是：cpu需要的数据都是从内存中读取的，但是内存的速度和cpu的速度相差好几个数量级，为了更高效的利用cpu的计算能力，平衡内存和cpu之间的效率差距，引入高速缓存。cpu不是从内存中直接获取数据，而只是和高速缓存交互，如果高速缓存中有，就直接用高速缓存中的数据；如果没有就去内存转给你读取。这个地方缓存是站在cpu的角度来说，高速缓存记录的就是从内存中拿到的结果一次提高效率。

应用缓存：

一个应用所需要的数据很多，有的是来自磁盘，比如配置文件，有的是来自db，比如业务数据。应用数据要获取这些数据都需要经过IO去获取，为了提高效率，应用中可以将拿到的数据放到自己本地内存中，这样后续需要这些数据的时候就从内存中取就好了。

所以说，凡是使用了一个更高效的方式来冗余数据，而为了获取这些数据是比较耗时的，那这种都可以认为是缓存。缓存不一定是内存，那些镜像站其实也是缓存，比如github，maven仓库等，直接去访问美国的仓库是比较慢的，所以在国内有很多镜像站，本质也是缓存。

下面从内存缓存的角度看，可以有哪些通用的实现以及他们的实现原理及细节。

HashMap

利用HashMap就可以实现一个简单的缓存，将一些获取比较耗时的结果放到HashMap中，后续就读取就好了。一个比较典型的应用场景就是反射缓存，在几乎所有的框架中都会将反射给缓存起来的，还有就是一个小的优化就是枚举的反序列化缓存，总之像这种固定不变的东西，而总量是有限的，还是比较适合用HashMap来做缓存的。

但是内存毕竟是有限的，所以当缓存的数据越来越多，最终可能导致OOM，所以缓存一定要有度，也就是说当缓存数据达到一定的阈值时，应该有策略淘汰掉一部分，不能让缓存无限增长，最终导致进程崩溃，或者说不能让缓存占用过多的内存以影响程序的正常运行，。

HashMap肯定是没有数据淘汰机制的，所以一般直接用HashMap来作为缓存的基本都是确定数据有限的场景，比如上述的反射缓存，配置文件缓存等。

一种简单的缓存淘汰方式就是FIFO，在数据写入的时候，将数据放到队头。那也就是先被更新的被先淘汰。这个其实是不合理的，缓存优化的是查询效率，那这样会让缓存命中率比较低。试想，更新频率不大，但是查询不大的数据会被淘汰出缓存，这肯定不合理，缓存优化的就是更新不大，查询大的场景，让查询更高效。所以FIFO不适合做为缓存淘汰策略，一种更加好的方式就是LRU算法那。

LRU缓存

LRU缓存就是使用LRU淘汰算法的缓存结构。所谓的LRU直接翻译就Latest Recent Used，它的理论基础就是：如果一个数据被访问了，那么它在接下来还会被访问的可能性非常大，所以当内存不够用的时候，那么先淘汰哪些最久没有被访问过的。

基于这个理论，实现LRU的常见方式就是基于链表：每次访问一个数据的时候都将这个数据移动到链表头部，然后内存不够需要淘汰的时候，直接删除链表尾部元素，从而实现LRU。

基于单个链表的实现

问题：

1. 访问速度是O(n)，即每次根据key查询数据的时候都需要遍历链表，这是不能接受的。

链表+HashMap实现

链表的方式比较方便的实现了LRU，只是需要移动一下指正就能够完成LRU算法，这个过程是O(1)的，但是查询是个O(n)的操作。HashMap的查询是O(1)的，所以链表结合HashMap可以实现O(1)的查询。

基本结构入下：

• 链表是用来存储key-value换出数据的。
• 每次访问的时候，将访问的节点都放到链表头部，然后当链表元素个数超过容量的时候，从尾部删除。这样依靠链表的头尾来实现LRU算法。
• HashMap的作用其实就是索引，实现O(1)的数据访问
• ps：这里区分下jdk中HashMap中用链表法解决hash冲突的那个链表，这里的链表不是一回事，如果将HashMap解决冲突的量表也画进来，结构应该是这样的

插入缓存put()方法逻辑：

缓存查询get()方法逻辑：

这个方式是是实现了最基本最朴素的LRU，这里的HashMap+链表的结构，jdk中已经有一个专门的结构进行了封装LinkedHashMap，也就是说LinkedHashMap已经实现了LRU了，只是需要在构建LinkedHashMap的时候，传入入参accessOrder=true，在put()/get()的时候，就会将访问的数据放到链表头部，这正是LRU淘汰的基础。但是，LinkedHashMap默认没有实现当数据量到达初始化容量的时候，是扩容了，而不是淘汰数据，但是给了扩展接口，只需要需要自己重写实现一下removeEldestEntry()这个接口，那就有LRU的效果啦。

class LinkedHashMapBaseLRUCache<K, V> extends LinkedHashMap {private int capacity;// 这是缓存的大小public LinkedHashMapBaseLRUCache(int capacity) {// 这是HashMap的大小，可以不一致，也可以一致。super(16, 0.75f, true);this.capacity = capacity;}@Overrideprotected boolean removeEldestEntry(Map.Entry eldest) {return super.size() > capacity;}
}

这仅仅是一个实现了基本的LRU算法的一个缓存，但是实际上要成为一个好用的缓存，还需要跟多额外的功能，比如：

1. 按照访问时间的自动过期机制。

2. 缓存内存控制，比如当jvm内存告警，淘汰缓存，即缓存使用WeakReference，而不是强引用。

3. 并发控制

4. 。。。

但是，很多缓存的实现的基础都是双链表+HashMap的思路，然后在这基础之上做了一些优化以及功能加强。下面会详细分析的guava cache基本结构也是如此。

堆

其实LRU就是一个按照访问时间排序的优先队列，而堆天生就是个优先队列，所以说堆天然就有LRU的特点。但是堆的操作中实际上没有查找任意元素，由于堆的特点，它不是一个搜索树，所以查找任意元素的效率O(n)，但是插入/删除一个元素都是O(lgn)的，总体相比于hashMap+链表的实现，效率上逊色不少。

guava cache

guava cache是guava实现的一个通用的进程内的内存缓存，其提供了丰富api来满足不同的使用场景。上面也提到了，它的实现基础也是HashMap+链表的方式。

理解guava cache主要看它如何实现如下4个功能

1. 缓存数据结构

2. 如何实现定时清除的

3.如何实现LRU

4. 线程安全

guava cache的实现虽然也是hashMap+链表的方式，但他没有直接使用jdk线程的HashMap，而是自己实现了一个HasnMap。考虑到线程安全问题，guava 在实现这个HashMap的时候采用了jdk7中
ConcurrentHashMap一样的思路：分段。

如下就是guava cache的整体结构：上半个虚线框是HashMap部分，下半部分就是链表部分。

其中用于保存实际数据的ReferenceEntry结构，对于guava cache至关重要，guava cache实现除了基本的缓存数据k-v存储外，一些LRU内存淘汰策略、按照最近查询时间/最近修改时间自动过期、key和value的weak引用类型等这些都依赖于ReferenceEntry结构，其结构如下：

自动过期的实现：这个是指缓存数据具有自动过期机制，而不是坐等缓存满了才淘汰。

• StrongEntry:不具备自动过期
• StrongAccessEntry：记录查询时刻，具备按照查询时间自动过期，比如查询后5s后没有查询，这条记录就自动删除
• StrongWriteEntry：记录修改时刻，具备按照修改时间自动过期，比如插入/更新后5s后没有更新，这条记录就自动删除
• StrongAccessWriteEntry：记录查询+时刻，具备按照查询时间/修改时间过期

这样只需要获得当前时间，然后和ReferenceEntry中保存的上次访问时间比较一下，如果大于指定时间，就删除对应的数据就好了。

guava cache对key和value还是先了若引用：key可以上强引用、WeakReference；value可以是强引用、SoftReference、WeakReference。key的强弱就是ReferenceEntry的子类：StrongxxxReference，key就是强引用；是WeakxxxxReference，那key就是WeakReference，所以上面的列表中还有一个对应的WeakxxxEntry。而对于value的引用类型，ValueReference来封装的

LRU实现：实现LRU主要是靠accessQueue和recencyQueue两个队列来实现LRU
按照前述使用链表实现LRU的原理，只需要每次读/写的时候，将读/写的元素都放到链表头部，然后当发现缓存中元素个数大于最大容量的时候，就直接删除链表尾部就好了，按照这个思路，一个链表就搞定了。但是guava cache中很明显有三个，这是为啥？

• accessQueue：它的地位就是我们自己实现的朴素LRU算法中那个双向链表。当发生读/写的时候，将读写涉及到的ReferenceEntry放到链表头(队头)；然后在put()的时候，如果发现缓存中已有元素个数超过了总的容量，那就从accessQueue队头出队一个元素，然后将这个元素删除(Hash表中删除+accessQueue中删除)。但是有个问题：accessQueue是LocalCache中自己实现的内部类：AccessQueue，这个类是线程不安全的，所以每次读/写的时候，要将元素移动到队头，需要先获得对应Segement的锁，然后才能操作。对于缓存来说，应用场景本身就是写少读多的，那么在写的时候，需要先获得Segement的锁才能写无可厚非，但是如果是读也要竞争这个锁就有点不好了，如果发生了竞争，就会阻塞读，这会大大影响读的QPS的，所以说读的时候不能加Segement锁，guava的解决办法就是recencyQueue。
• recencyQueue：在每次get()的时候，不会去改变accessQueue(即不会将读取到的ReferenceEntry立马放到AccessQueue的队尾)，而是直接将读取到的元素插入到recencyQueue中。recencyQueue使用的是jdk的ConcurrentLinkedQueue，它实现变成安全的，所以所有线程在get()的时候可直接使用recencyQueue#add()将读取的元素添加到队尾。在get()结束的时候，会去tryLock()一下Segement锁，如果拿到锁，就会遍历recencyQueue，然后依次出队，然后将其中的ReferenceEntry在accessQueue中移动到队尾，这样recencyQueue就变成空的了，而历史读取的元素以依次在accessQueue中都放到了队尾(ps：这里可以理解成在get()读取的时候，将读取的ReferenceEntry放到accessQueu队尾是个延迟的批量操作，recencyQueue就是这个批操作的一个缓存，这个批大小LocalCache写死的为64)。这这个逻辑就是在get()的finally块中调用的postReadCleanup()实现。

ps：

1. AccessQueue#add()这个方法很特殊：如果ReferenceEntry在队列中已存在，就是将其移动到队尾；如果不存在才是添加到队尾，先了解这个逻辑，再去看代码，看到相关代码就少些疑惑。

2. 处理recencyQueue不光是在get()的最后会尝试处理，在put()的时候也会去处理(put()的时候先拿到了Segement锁，由于使用的是ReentrantLock，由于其可重入性，tryLock()一定会成功的)

问题：

1. 因为recencyQueue的原因，guava cache的读性能，是受ConcurrentLinkedQueue的写入性能约束的。而且如果写入也比较大，持续竞争的话，可能导致recencyQueue非常大。

2. 不管是LRU淘汰，还是最近访问时间自动淘汰机制、还是WeakReference情况下具体的缓存数据被gc掉了去清理引用，这三种情况的清理都是在put()/get()方法执行过程中，同步去执行的，这势必会影响读写QPS的。

3. 所有的key的有效期都是一样的，没办法为不同key设置不同的有效期。ps：Caffine好像也不行。

4. LRU算法本身的问题：对于偶发的批量数据读取，会污染缓存中的数据，导致将真正的热点数据挤出缓存。为了解决这个问题，有很多对LRU的优化，比如Innodb中对LRU的优化LRU算法及其优化策略——算法篇 - 掘金。另外redis也是使用了LRU算法，但是实现上也做了一些优化：

redis的淘汰策略：Redis优化--LRU和LFU区别 - it610.com

guava cache的使用例子

  Cache<String, String> cache =CacheBuilder.newBuilder().maximumSize(4)// cache的最大容量，超过后将使用LRU淘汰 这个是基于缓存总容量的淘汰：即缓存的数据量达到了maximumSize，就开始按照LRU淘汰//.maximumWeight(12)// maximumWeight+weigher，maximumWeight指定的是最大权重，weigher是一个权重比例。//.weigher(null)// 如果不指定，默认实现的weigher就是OneWeigher，返回的权重就是1，这个时候跟总量maximumSize的淘汰是一样的，maximumWeight就是最大容量。.concurrencyLevel(1)// 这个命名其实还是蛮讲究的，guava cache其内部的HashMap是采用分段来实现的，一个分段独享一把锁，按有多少个分段，那么统一时刻就允许有多少线程操作HashMap// ，所以说白了这里的ConcurrentLevel就是分段个数。.initialCapacity(2)// HashMap的初始容量，也就是说初始化的时候并不会直接初始化一个maximumSize大小的hashMap，而是初始化一个initialcapacity// 大小，后续需要的时候会扩容。真实存储数据的是在Segement中，// 所以初始化的时候根据concurrencyLevel决定创建多少了Segment,然后根据initialCapacity/concurrencyLevel决定Segement中数组的大小.weakKeys()//默认情况，Cache中的key和value都是强引用。weakKeys()开启后，key将使用WeakReference引用//.weakValues() // 开启后value的引用就是WeakReference//.softValues() // 开启后value的引用就是SoftReference//.expireAfterWrite(3, TimeUnit.SECONDS)// 写入过后3s后没有访问，自动过期删除//.expireAfterAccess(3,TimeUnit.SECONDS)// 读取过后3s没有访问，自动删除过期。.refreshAfterWrite(3, TimeUnit.SECONDS)// 自从上上次写入后，没隔3s调用cacheLoader去自动更新一遍数据。// 这里需要注意，guava的实现并不是真的起了个异步任务在做这个事情，而是在查询的时候去检查，如果上次写入时间举当前超过了3s，就去重新load一遍，去看实现上虽然表面上看是异步执行的呀，// 但默认实现上是sameThreadExecutor，其实就是同步执行的，所以如果生产环境中要设置这个参数就要注意了，这个相当于没隔3s的get()查询都会去调用一遍CacheLoad。.recordStats()//打开统计信息：如缓存命中率等，这个是有一些开销的.removalListener(new MyRemovalListener<String, String>())// 当一个entry删除的时候(主动删除or缓存淘汰删除)// ，会调用指定的Listener。具体哪些场景会调用这个removeListener参考：RemovalCause//.ticker(null)// 实际就是一个时间计时器，用于计算entry是否过期。默认就是System.nanoTime()，这个跟guava StopWatch是一样的.build(new MyCacheLoader());

CacheLoader的实现例子：

 // 当设置了refreshAfterWrite(3, TimeUnit.SECONDS)参数，那么当数据距离上次写入3s后，再次调用get()的时候，会自动去调用reload()方法重新加载数据，刷新一遍。// 但是默认实现的reload()方法是同步的，即直接调用了CacheLoader#load()，这里我们可以重新实现reload()变成变成异步的。@Overridepublic ListenableFuture reload(String key, String oldValue) throws Exception {Future<String> future = executorService.submit(new Callable<String>() {@Overridepublic String call() throws Exception {System.out.println("reload thread:" + Thread.currentThread().getId());return key+"-reload";}});// 这里这么写是掩耳盗铃的，需要自己实现ListableFuture，这个地方即使是实现了异步，其实也只是和一些内存处理并行。// 如果在这些内存操作处理结束，这里的load还没有加载成功，那么就会丢掉这里reload的结果，所以要自己实现异步的时候一定要注意一下return Futures.immediateFuture(future.get());}

RemovalListener的实现

class MyRemovalListener<K, V> implements RemovalListener {@Overridepublic void onRemoval(RemovalNotification notification) {System.out.println("删除了," + notification.getKey() + "=" + notification.getValue() + "  cause:" + notification.getCause());}
}

参考：

1. LRU算法及其优化策略——算法篇 - 掘金  这个文章比较详细介绍了从最朴素的LRU以及一些优化的LRU算法

2. 你应该知道的缓存进化史 - 咖啡拿铁的技术分享的个人空间 - OSCHINA - 中文开源技术交流社区  这个文章比较详细介绍了集中缓存的实现，并结合guava cache的源码详细介绍了guava cache的使用及对应原理

附件：LocalCache的注释，以供参考：

/** Copyright (C) 2009 The Guava Authors** Licensed under the Apache License, Version 2.0 (the "License");* you may not use this file except in compliance with the License.* You may obtain a copy of the License at** http://www.apache.org/licenses/LICENSE-2.0** Unless required by applicable law or agreed to in writing, software* distributed under the License is distributed on an "AS IS" BASIS,* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.* See the License for the specific language governing permissions and* limitations under the License.*/package com.zj.cache.guavacache;import CacheLoader.InvalidCacheLoadException;
import com.google.common.annotations.GwtCompatible;
import com.google.common.annotations.GwtIncompatible;
import com.google.common.annotations.VisibleForTesting;
import com.google.common.base.Equivalence;
import com.google.common.base.Function;
import com.google.common.base.Stopwatch;
import com.google.common.base.Ticker;
import com.google.common.cache.AbstractCache.SimpleStatsCounter;
import com.google.common.cache.AbstractCache.StatsCounter;
import com.google.common.cache.*;
import com.google.common.collect.*;
import com.google.common.primitives.Ints;
import com.google.common.util.concurrent.*;import javax.annotation.Nullable;
import javax.annotation.concurrent.GuardedBy;
import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.Serializable;
import java.lang.ref.Reference;
import java.lang.ref.ReferenceQueue;
import java.lang.ref.SoftReference;
import java.lang.ref.WeakReference;
import java.util.*;
import java.util.concurrent.*;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.concurrent.atomic.AtomicReferenceArray;
import java.util.concurrent.locks.ReentrantLock;
import java.util.logging.Level;
import java.util.logging.Logger;import static com.google.common.base.Preconditions.checkNotNull;
import static com.google.common.base.Preconditions.checkState;
import static com.google.common.util.concurrent.Uninterruptibles.getUninterruptibly;
import static com.zj.cache.guavacache.CacheBuilder.NULL_TICKER;
import static com.zj.cache.guavacache.CacheBuilder.UNSET_INT;
import static java.util.concurrent.TimeUnit.NANOSECONDS;/*** The concurrent hash map implementation built by {@link com.zj.cache.guavacache.CacheBuilder}.** <p>This implementation is heavily derived from revision 1.96 of <a* href="http://tinyurl.com/ConcurrentHashMap">ConcurrentHashMap.java</a>.** @author Charles Fry* @author Bob Lee ({@code com.google.common.collect.MapMaker})* @author Doug Lea ({@code ConcurrentHashMap})*/
@GwtCompatible(emulated = true)
class LocalCache<K, V> extends AbstractMap<K, V> implements ConcurrentMap<K, V> {/** The basic strategy is to subdivide the table among Segments, each of which itself is a* concurrently readable hash table. The map supports non-blocking reads and concurrent writes* across different segments.** If a maximum size is specified, a best-effort bounding is performed per segment, using a* page-replacement algorithm to determine which entries to evict when the capacity has been* exceeded.** The page replacement algorithm's data structures are kept casually consistent with the map. The* ordering of writes to a segment is sequentially consistent. An update to the map and recording* of reads may not be immediately reflected on the algorithm's data structures. These structures* are guarded by a lock and operations are applied in batches to avoid lock contention. The* penalty of applying the batches is spread across threads so that the amortized cost is slightly* higher than performing just the operation without enforcing the capacity constraint.** This implementation uses a per-segment queue to record a memento of the additions, removals,* and accesses that were performed on the map. The queue is drained on writes and when it exceeds* its capacity threshold.** The Least Recently Used page replacement algorithm was chosen due to its simplicity, high hit* rate, and ability to be implemented with O(1) time complexity. The initial LRU implementation* operates per-segment rather than globally for increased implementation simplicity. We expect* the cache hit rate to be similar to that of a global LRU algorithm.*/// Constants/*** The maximum capacity, used if a higher value is implicitly specified by either of the constructors with arguments. MUST be a power of* two <= 1<<30 to ensure that entries are indexable using ints.*/static final int MAXIMUM_CAPACITY = 1 << 30;/*** The maximum number of segments to allow; used to bound constructor arguments.*/static final int MAX_SEGMENTS = 1 << 16; // slightly conservative/*** Number of (unsynchronized) retries in the containsValue method.*/static final int CONTAINS_VALUE_RETRIES = 3;/*** Number of cache access operations that can be buffered per segment before the cache's recency ordering information is updated. This* is used to avoid lock contention by recording a memento of reads and delaying a lock acquisition until the threshold is crossed or a* mutation occurs.** <p>This must be a (2^n)-1 as it is used as a mask.*/static final int DRAIN_THRESHOLD = 0x3F;//十进制63/*** Maximum number of entries to be drained in a single cleanup run. This applies independently to the cleanup queue and both reference* queues.*/// TODO(fry): empirically optimize thisstatic final int DRAIN_MAX = 16;// Fieldsstatic final Logger logger = Logger.getLogger(LocalCache.class.getName());static final ListeningExecutorService sameThreadExecutor = MoreExecutors.sameThreadExecutor();/*** Mask value for indexing into segments. The upper bits of a key's hash code are used to choose the segment.*/final int segmentMask;/*** Shift value for indexing within segments. Helps prevent entries that end up in the same segment from also ending up in the same* bucket.*/final int segmentShift;/*** The segments, each of which is a specialized hash table.*/final Segment<K, V>[] segments;/*** The concurrency level.*/final int concurrencyLevel;/*** Strategy for comparing keys.*/final Equivalence<Object> keyEquivalence;/*** Strategy for comparing values.*/final Equivalence<Object> valueEquivalence;/*** Strategy for referencing keys.*/final Strength keyStrength;/*** Strategy for referencing values.*/final Strength valueStrength;/*** The maximum weight of this map. UNSET_INT if there is no maximum.*/final long maxWeight;/*** Weigher to weigh cache entries.*/final Weigher<K, V> weigher;/*** How long after the last access to an entry the map will retain that entry.*/final long expireAfterAccessNanos;/*** How long after the last write to an entry the map will retain that entry.*/final long expireAfterWriteNanos;/*** How long after the last write an entry becomes a candidate for refresh.*/final long refreshNanos;/*** Entries waiting to be consumed by the removal listener.*/// TODO(fry): define a new type which creates event objects and automates the clear logicfinal Queue<RemovalNotification<K, V>> removalNotificationQueue;/*** A listener that is invoked when an entry is removed due to expiration or garbage collection of soft/weak entries.*/final RemovalListener<K, V> removalListener;/*** Measures time in a testable way.*/final Ticker ticker;/*** Factory used to create new entries.*/final EntryFactory entryFactory;/*** Accumulates global cache statistics. Note that there are also per-segments stats counters which must be aggregated to obtain a global* stats view.*/final StatsCounter globalStatsCounter;/*** The default cache loader to use on loading operations.*/@Nullablefinal CacheLoader<? super K, V> defaultLoader;/*** Creates a new, empty map with the specified strategy, initial capacity and concurrency level.*/LocalCache(CacheBuilder<? super K, ? super V> builder, @Nullable CacheLoader<? super K, V> loader) {concurrencyLevel = Math.min(builder.getConcurrencyLevel(), MAX_SEGMENTS);keyStrength = builder.getKeyStrength();valueStrength = builder.getValueStrength();keyEquivalence = builder.getKeyEquivalence();valueEquivalence = builder.getValueEquivalence();maxWeight = builder.getMaximumWeight();weigher = builder.getWeigher();expireAfterAccessNanos = builder.getExpireAfterAccessNanos();expireAfterWriteNanos = builder.getExpireAfterWriteNanos();refreshNanos = builder.getRefreshNanos();removalListener = builder.getRemovalListener();removalNotificationQueue = (removalListener == CacheBuilder.NullListener.INSTANCE)? LocalCache.<RemovalNotification<K, V>>discardingQueue(): new ConcurrentLinkedQueue<RemovalNotification<K, V>>();ticker = builder.getTicker(recordsTime());entryFactory = EntryFactory.getFactory(keyStrength, usesAccessEntries(), usesWriteEntries());globalStatsCounter = builder.getStatsCounterSupplier().get();defaultLoader = loader;int initialCapacity = Math.min(builder.getInitialCapacity(), MAXIMUM_CAPACITY);if (evictsBySize() && !customWeigher()) {initialCapacity = Math.min(initialCapacity, (int) maxWeight);}// Find the lowest power-of-two segmentCount that exceeds concurrencyLevel, unless// maximumSize/Weight is specified in which case ensure that each segment gets at least 10// entries. The special casing for size-based eviction is only necessary because that eviction// happens per segment instead of globally, so too many segments compared to the maximum size// will result in random eviction behavior.int segmentShift = 0;int segmentCount = 1;while (segmentCount < concurrencyLevel&& (!evictsBySize() || segmentCount * 20 <= maxWeight)) {++segmentShift;segmentCount <<= 1;}this.segmentShift = 32 - segmentShift;segmentMask = segmentCount - 1;this.segments = newSegmentArray(segmentCount);int segmentCapacity = initialCapacity / segmentCount;if (segmentCapacity * segmentCount < initialCapacity) {++segmentCapacity;}int segmentSize = 1;while (segmentSize < segmentCapacity) {segmentSize <<= 1;}if (evictsBySize()) {// Ensure sum of segment max weights = overall max weightslong maxSegmentWeight = maxWeight / segmentCount + 1;long remainder = maxWeight % segmentCount;for (int i = 0; i < this.segments.length; ++i) {if (i == remainder) {maxSegmentWeight--;}this.segments[i] =createSegment(segmentSize, maxSegmentWeight, builder.getStatsCounterSupplier().get());}} else {for (int i = 0; i < this.segments.length; ++i) {this.segments[i] =createSegment(segmentSize, UNSET_INT, builder.getStatsCounterSupplier().get());}}}boolean evictsBySize() {return maxWeight >= 0;}boolean customWeigher() {return weigher != OneWeigher.INSTANCE;}boolean expires() {return expiresAfterWrite() || expiresAfterAccess();}boolean expiresAfterWrite() {return expireAfterWriteNanos > 0;}boolean expiresAfterAccess() {return expireAfterAccessNanos > 0;}boolean refreshes() {return refreshNanos > 0;}boolean usesAccessQueue() {return expiresAfterAccess() || evictsBySize();}boolean usesWriteQueue() {return expiresAfterWrite();}boolean recordsWrite() {return expiresAfterWrite() || refreshes();}boolean recordsAccess() {return expiresAfterAccess();}boolean recordsTime() {return recordsWrite() || recordsAccess();}boolean usesWriteEntries() {return usesWriteQueue() || recordsWrite();}boolean usesAccessEntries() {return usesAccessQueue() || recordsAccess();}boolean usesKeyReferences() {return keyStrength != Strength.STRONG;}boolean usesValueReferences() {return valueStrength != Strength.STRONG;}enum Strength {/** TODO(kevinb): If we strongly reference the value and aren't loading, we needn't wrap the* value. This could save ~8 bytes per entry.*/STRONG {@Override<K, V> ValueReference<K, V> referenceValue(Segment<K, V> segment, ReferenceEntry<K, V> entry, V value, int weight) {return (weight == 1)? new StrongValueReference<K, V>(value): new WeightedStrongValueReference<K, V>(value, weight);}@OverrideEquivalence<Object> defaultEquivalence() {return Equivalence.equals();}},SOFT {@Override<K, V> ValueReference<K, V> referenceValue(Segment<K, V> segment, ReferenceEntry<K, V> entry, V value, int weight) {return (weight == 1)? new SoftValueReference<K, V>(segment.valueReferenceQueue, value, entry): new WeightedSoftValueReference<K, V>(segment.valueReferenceQueue, value, entry, weight);}@OverrideEquivalence<Object> defaultEquivalence() {return Equivalence.identity();}},WEAK {@Override<K, V> ValueReference<K, V> referenceValue(Segment<K, V> segment, ReferenceEntry<K, V> entry, V value, int weight) {return (weight == 1)? new WeakValueReference<K, V>(segment.valueReferenceQueue, value, entry): new WeightedWeakValueReference<K, V>(segment.valueReferenceQueue, value, entry, weight);}@OverrideEquivalence<Object> defaultEquivalence() {return Equivalence.identity();}};/*** Creates a reference for the given value according to this value strength.*/abstract <K, V> ValueReference<K, V> referenceValue(Segment<K, V> segment, ReferenceEntry<K, V> entry, V value, int weight);/*** Returns the default equivalence strategy used to compare and hash keys or values referenced at this strength. This strategy will* be used unless the user explicitly specifies an alternate strategy.*/abstract Equivalence<Object> defaultEquivalence();}/*** Creates new entries.*/enum EntryFactory {STRONG {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new StrongEntry<K, V>(key, hash, next);}},STRONG_ACCESS {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new StrongAccessEntry<K, V>(key, hash, next);}@Override<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {ReferenceEntry<K, V> newEntry = super.copyEntry(segment, original, newNext);copyAccessEntry(original, newEntry);return newEntry;}},STRONG_WRITE {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new StrongWriteEntry<K, V>(key, hash, next);}@Override<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {ReferenceEntry<K, V> newEntry = super.copyEntry(segment, original, newNext);copyWriteEntry(original, newEntry);return newEntry;}},STRONG_ACCESS_WRITE {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new StrongAccessWriteEntry<K, V>(key, hash, next);}@Override<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {ReferenceEntry<K, V> newEntry = super.copyEntry(segment, original, newNext);copyAccessEntry(original, newEntry);copyWriteEntry(original, newEntry);return newEntry;}},WEAK {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new WeakEntry<K, V>(segment.keyReferenceQueue, key, hash, next);}},WEAK_ACCESS {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new WeakAccessEntry<K, V>(segment.keyReferenceQueue, key, hash, next);}@Override<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {ReferenceEntry<K, V> newEntry = super.copyEntry(segment, original, newNext);copyAccessEntry(original, newEntry);return newEntry;}},WEAK_WRITE {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new WeakWriteEntry<K, V>(segment.keyReferenceQueue, key, hash, next);}@Override<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {ReferenceEntry<K, V> newEntry = super.copyEntry(segment, original, newNext);copyWriteEntry(original, newEntry);return newEntry;}},WEAK_ACCESS_WRITE {@Override<K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next) {return new WeakAccessWriteEntry<K, V>(segment.keyReferenceQueue, key, hash, next);}@Override<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {ReferenceEntry<K, V> newEntry = super.copyEntry(segment, original, newNext);copyAccessEntry(original, newEntry);copyWriteEntry(original, newEntry);return newEntry;}};/*** Masks used to compute indices in the following table.*/static final int ACCESS_MASK = 1;static final int WRITE_MASK  = 2;static final int WEAK_MASK   = 4;/*** Look-up table for factories.*/static final EntryFactory[] factories = {STRONG, STRONG_ACCESS, STRONG_WRITE, STRONG_ACCESS_WRITE,WEAK, WEAK_ACCESS, WEAK_WRITE, WEAK_ACCESS_WRITE,};static EntryFactory getFactory(Strength keyStrength, boolean usesAccessQueue,boolean usesWriteQueue) {int flags = ((keyStrength == Strength.WEAK) ? WEAK_MASK : 0)| (usesAccessQueue ? ACCESS_MASK : 0)| (usesWriteQueue ? WRITE_MASK : 0);return factories[flags];}/*** Creates a new entry.** @param segment to create the entry for* @param key     of the entry* @param hash    of the key* @param next    entry in the same bucket*/abstract <K, V> ReferenceEntry<K, V> newEntry(Segment<K, V> segment, K key, int hash, @Nullable ReferenceEntry<K, V> next);/*** Copies an entry, assigning it a new {@code next} entry.** @param original the entry to copy* @param newNext  entry in the same bucket*/@GuardedBy("Segment.this")<K, V> ReferenceEntry<K, V> copyEntry(Segment<K, V> segment, ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {return newEntry(segment, original.getKey(), original.getHash(), newNext);}@GuardedBy("Segment.this")<K, V> void copyAccessEntry(ReferenceEntry<K, V> original, ReferenceEntry<K, V> newEntry) {// TODO(fry): when we link values instead of entries this method can go// away, as can connectAccessOrder, nullifyAccessOrder.newEntry.setAccessTime(original.getAccessTime());connectAccessOrder(original.getPreviousInAccessQueue(), newEntry);connectAccessOrder(newEntry, original.getNextInAccessQueue());nullifyAccessOrder(original);}@GuardedBy("Segment.this")<K, V> void copyWriteEntry(ReferenceEntry<K, V> original, ReferenceEntry<K, V> newEntry) {// TODO(fry): when we link values instead of entries this method can go// away, as can connectWriteOrder, nullifyWriteOrder.newEntry.setWriteTime(original.getWriteTime());connectWriteOrder(original.getPreviousInWriteQueue(), newEntry);connectWriteOrder(newEntry, original.getNextInWriteQueue());nullifyWriteOrder(original);}}/*** A reference to a value.*/interface ValueReference<K, V> {/*** Returns the value. Does not block or throw exceptions.*/@NullableV get();/*** Waits for a value that may still be loading. Unlike get(), this method can block (in the case of FutureValueReference).** @throws ExecutionException if the loading thread throws an exception* @throws if                 the loading thread throws an error*/V waitForValue() throws ExecutionException;/*** Returns the weight of this entry. This is assumed to be static between calls to setValue.*/int getWeight();/*** Returns the entry associated with this value reference, or {@code null} if this value reference is independent of any entry.*/@NullableReferenceEntry<K, V> getEntry();/*** Creates a copy of this reference for the given entry.** <p>{@code value} may be null only for a loading reference.*/ValueReference<K, V> copyFor(ReferenceQueue<V> queue, @Nullable V value, ReferenceEntry<K, V> entry);/*** Notifify pending loads that a new value was set. This is only relevant to loading value references.*/void notifyNewValue(@Nullable V newValue);/*** Returns true if a new value is currently loading, regardless of whether or not there is an existing value. It is assumed that the* return value of this method is constant for any given ValueReference instance.*/boolean isLoading();/*** Returns true if this reference contains an active value, meaning one that is still considered present in the cache. Active values* consist of live values, which are returned by cache lookups, and dead values, which have been evicted but awaiting removal.* Non-active values consist strictly of loading values, though during refresh a value may be both active and loading.*/boolean isActive();}/*** Placeholder. Indicates that the value hasn't been set yet.*/static final ValueReference<Object, Object> UNSET = new ValueReference<Object, Object>() {@Overridepublic Object get() {return null;}@Overridepublic int getWeight() {return 0;}@Overridepublic ReferenceEntry<Object, Object> getEntry() {return null;}@Overridepublic ValueReference<Object, Object> copyFor(ReferenceQueue<Object> queue,@Nullable Object value, ReferenceEntry<Object, Object> entry) {return this;}@Overridepublic boolean isLoading() {return false;}@Overridepublic boolean isActive() {return false;}@Overridepublic Object waitForValue() {return null;}@Overridepublic void notifyNewValue(Object newValue) {}};/*** Singleton placeholder that indicates a value is being loaded.*/@SuppressWarnings("unchecked") // impl never uses a parameter or returns any non-null valuestatic <K, V> ValueReference<K, V> unset() {return (ValueReference<K, V>) UNSET;}/*** An entry in a reference map.* <p>* Entries in the map can be in the following states:* <p>* Valid: - Live: valid key/value are set - Loading: loading is pending* <p>* Invalid: - Expired: time expired (key/value may still be set) - Collected: key/value was partially collected, but not yet cleaned up* - Unset: marked as unset, awaiting cleanup or reuse*/interface ReferenceEntry<K, V> {/*** Returns the value reference from this entry.*/ValueReference<K, V> getValueReference();/*** Sets the value reference for this entry.*/void setValueReference(ValueReference<K, V> valueReference);/*** Returns the next entry in the chain.*/@NullableReferenceEntry<K, V> getNext();/*** Returns the entry's hash.*/int getHash();/*** Returns the key for this entry.*/@NullableK getKey();/** Used by entries that use access order. Access entries are maintained in a doubly-linked list.* New entries are added at the tail of the list at write time; stale entries are expired from* the head of the list.*//*** Returns the time that this entry was last accessed, in ns.*/long getAccessTime();/*** Sets the entry access time in ns.*/void setAccessTime(long time);/*** Returns the next entry in the access queue.*/ReferenceEntry<K, V> getNextInAccessQueue();/*** Sets the next entry in the access queue.*/void setNextInAccessQueue(ReferenceEntry<K, V> next);/*** Returns the previous entry in the access queue.*/ReferenceEntry<K, V> getPreviousInAccessQueue();/*** Sets the previous entry in the access queue.*/void setPreviousInAccessQueue(ReferenceEntry<K, V> previous);/** Implemented by entries that use write order. Write entries are maintained in a* doubly-linked list. New entries are added at the tail of the list at write time and stale* entries are expired from the head of the list.*//*** Returns the time that this entry was last written, in ns.*/long getWriteTime();/*** Sets the entry write time in ns.*/void setWriteTime(long time);/*** Returns the next entry in the write queue.*/ReferenceEntry<K, V> getNextInWriteQueue();/*** Sets the next entry in the write queue.*/void setNextInWriteQueue(ReferenceEntry<K, V> next);/*** Returns the previous entry in the write queue.*/ReferenceEntry<K, V> getPreviousInWriteQueue();/*** Sets the previous entry in the write queue.*/void setPreviousInWriteQueue(ReferenceEntry<K, V> previous);}private enum NullEntry implements ReferenceEntry<Object, Object> {INSTANCE;@Overridepublic ValueReference<Object, Object> getValueReference() {return null;}@Overridepublic void setValueReference(ValueReference<Object, Object> valueReference) {}@Overridepublic ReferenceEntry<Object, Object> getNext() {return null;}@Overridepublic int getHash() {return 0;}@Overridepublic Object getKey() {return null;}@Overridepublic long getAccessTime() {return 0;}@Overridepublic void setAccessTime(long time) {}@Overridepublic ReferenceEntry<Object, Object> getNextInAccessQueue() {return this;}@Overridepublic void setNextInAccessQueue(ReferenceEntry<Object, Object> next) {}@Overridepublic ReferenceEntry<Object, Object> getPreviousInAccessQueue() {return this;}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<Object, Object> previous) {}@Overridepublic long getWriteTime() {return 0;}@Overridepublic void setWriteTime(long time) {}@Overridepublic ReferenceEntry<Object, Object> getNextInWriteQueue() {return this;}@Overridepublic void setNextInWriteQueue(ReferenceEntry<Object, Object> next) {}@Overridepublic ReferenceEntry<Object, Object> getPreviousInWriteQueue() {return this;}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<Object, Object> previous) {}}static abstract class AbstractReferenceEntry<K, V> implements ReferenceEntry<K, V> {@Overridepublic ValueReference<K, V> getValueReference() {throw new UnsupportedOperationException();}@Overridepublic void setValueReference(ValueReference<K, V> valueReference) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getNext() {throw new UnsupportedOperationException();}@Overridepublic int getHash() {throw new UnsupportedOperationException();}@Overridepublic K getKey() {throw new UnsupportedOperationException();}@Overridepublic long getAccessTime() {throw new UnsupportedOperationException();}@Overridepublic void setAccessTime(long time) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {throw new UnsupportedOperationException();}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {throw new UnsupportedOperationException();}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {throw new UnsupportedOperationException();}@Overridepublic long getWriteTime() {throw new UnsupportedOperationException();}@Overridepublic void setWriteTime(long time) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {throw new UnsupportedOperationException();}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {throw new UnsupportedOperationException();}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {throw new UnsupportedOperationException();}}@SuppressWarnings("unchecked") // impl never uses a parameter or returns any non-null valuestatic <K, V> ReferenceEntry<K, V> nullEntry() {return (ReferenceEntry<K, V>) NullEntry.INSTANCE;}static final Queue<? extends Object> DISCARDING_QUEUE = new AbstractQueue<Object>() {@Overridepublic boolean offer(Object o) {return true;}@Overridepublic Object peek() {return null;}@Overridepublic Object poll() {return null;}@Overridepublic int size() {return 0;}@Overridepublic Iterator<Object> iterator() {return Iterators.emptyIterator();}};/*** Queue that discards all elements.*/@SuppressWarnings("unchecked") // impl never uses a parameter or returns any non-null valuestatic <E> Queue<E> discardingQueue() {return (Queue) DISCARDING_QUEUE;}/** Note: All of this duplicate code sucks, but it saves a lot of memory. If only Java had mixins!* To maintain this code, make a change for the strong reference type. Then, cut and paste, and* replace "Strong" with "Soft" or "Weak" within the pasted text. The primary difference is that* strong entries store the key reference directly while soft and weak entries delegate to their* respective superclasses.*//*** Used for strongly-referenced keys.*/static class StrongEntry<K, V> extends AbstractReferenceEntry<K, V> {final K key;StrongEntry(K key, int hash, @Nullable ReferenceEntry<K, V> next) {this.key = key;this.hash = hash;this.next = next;}@Overridepublic K getKey() {return this.key;}// The code below is exactly the same for each entry type.final    int                  hash;final    ReferenceEntry<K, V> next;volatile ValueReference<K, V> valueReference = unset();@Overridepublic ValueReference<K, V> getValueReference() {return valueReference;}@Overridepublic void setValueReference(ValueReference<K, V> valueReference) {this.valueReference = valueReference;}@Overridepublic int getHash() {return hash;}@Overridepublic ReferenceEntry<K, V> getNext() {return next;}}static final class StrongAccessEntry<K, V> extends StrongEntry<K, V> {StrongAccessEntry(K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(key, hash, next);}// The code below is exactly the same for each access entry type.volatile long accessTime = Long.MAX_VALUE;@Overridepublic long getAccessTime() {return accessTime;}@Overridepublic void setAccessTime(long time) {this.accessTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {return nextAccess;}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {this.nextAccess = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {return previousAccess;}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {this.previousAccess = previous;}}static final class StrongWriteEntry<K, V> extends StrongEntry<K, V> {StrongWriteEntry(K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(key, hash, next);}// The code below is exactly the same for each write entry type.volatile long writeTime = Long.MAX_VALUE;@Overridepublic long getWriteTime() {return writeTime;}@Overridepublic void setWriteTime(long time) {this.writeTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {return nextWrite;}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {this.nextWrite = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {return previousWrite;}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {this.previousWrite = previous;}}static final class StrongAccessWriteEntry<K, V> extends StrongEntry<K, V> {StrongAccessWriteEntry(K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(key, hash, next);}// The code below is exactly the same for each access entry type.volatile long accessTime = Long.MAX_VALUE;@Overridepublic long getAccessTime() {return accessTime;}@Overridepublic void setAccessTime(long time) {this.accessTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {return nextAccess;}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {this.nextAccess = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {return previousAccess;}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {this.previousAccess = previous;}// The code below is exactly the same for each write entry type.volatile long writeTime = Long.MAX_VALUE;@Overridepublic long getWriteTime() {return writeTime;}@Overridepublic void setWriteTime(long time) {this.writeTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {return nextWrite;}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {this.nextWrite = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {return previousWrite;}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {this.previousWrite = previous;}}/*** Used for weakly-referenced keys.*/static class WeakEntry<K, V> extends WeakReference<K> implements ReferenceEntry<K, V> {WeakEntry(ReferenceQueue<K> queue, K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(key, queue);this.hash = hash;this.next = next;}@Overridepublic K getKey() {return get();}/** It'd be nice to get these for free from AbstractReferenceEntry, but we're already extending* WeakReference<K>.*/// null access@Overridepublic long getAccessTime() {throw new UnsupportedOperationException();}@Overridepublic void setAccessTime(long time) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {throw new UnsupportedOperationException();}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {throw new UnsupportedOperationException();}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {throw new UnsupportedOperationException();}// null write@Overridepublic long getWriteTime() {throw new UnsupportedOperationException();}@Overridepublic void setWriteTime(long time) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {throw new UnsupportedOperationException();}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {throw new UnsupportedOperationException();}@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {throw new UnsupportedOperationException();}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {throw new UnsupportedOperationException();}// The code below is exactly the same for each entry type.final    int                  hash;final    ReferenceEntry<K, V> next;volatile ValueReference<K, V> valueReference = unset();@Overridepublic ValueReference<K, V> getValueReference() {return valueReference;}@Overridepublic void setValueReference(ValueReference<K, V> valueReference) {this.valueReference = valueReference;}@Overridepublic int getHash() {return hash;}@Overridepublic ReferenceEntry<K, V> getNext() {return next;}}static final class WeakAccessEntry<K, V> extends WeakEntry<K, V> {WeakAccessEntry(ReferenceQueue<K> queue, K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(queue, key, hash, next);}// The code below is exactly the same for each access entry type.volatile long accessTime = Long.MAX_VALUE;@Overridepublic long getAccessTime() {return accessTime;}@Overridepublic void setAccessTime(long time) {this.accessTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {return nextAccess;}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {this.nextAccess = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {return previousAccess;}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {this.previousAccess = previous;}}static final class WeakWriteEntry<K, V> extends WeakEntry<K, V> {WeakWriteEntry(ReferenceQueue<K> queue, K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(queue, key, hash, next);}// The code below is exactly the same for each write entry type.volatile long writeTime = Long.MAX_VALUE;@Overridepublic long getWriteTime() {return writeTime;}@Overridepublic void setWriteTime(long time) {this.writeTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {return nextWrite;}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {this.nextWrite = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {return previousWrite;}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {this.previousWrite = previous;}}static final class WeakAccessWriteEntry<K, V> extends WeakEntry<K, V> {WeakAccessWriteEntry(ReferenceQueue<K> queue, K key, int hash, @Nullable ReferenceEntry<K, V> next) {super(queue, key, hash, next);}// The code below is exactly the same for each access entry type.volatile long accessTime = Long.MAX_VALUE;@Overridepublic long getAccessTime() {return accessTime;}@Overridepublic void setAccessTime(long time) {this.accessTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {return nextAccess;}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {this.nextAccess = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousAccess = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {return previousAccess;}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {this.previousAccess = previous;}// The code below is exactly the same for each write entry type.volatile long writeTime = Long.MAX_VALUE;@Overridepublic long getWriteTime() {return writeTime;}@Overridepublic void setWriteTime(long time) {this.writeTime = time;}@GuardedBy("Segment.this")ReferenceEntry<K, V> nextWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {return nextWrite;}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {this.nextWrite = next;}@GuardedBy("Segment.this")ReferenceEntry<K, V> previousWrite = nullEntry();@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {return previousWrite;}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {this.previousWrite = previous;}}/*** References a weak value.*/static class WeakValueReference<K, V>extends WeakReference<V> implements ValueReference<K, V> {final ReferenceEntry<K, V> entry;WeakValueReference(ReferenceQueue<V> queue, V referent, ReferenceEntry<K, V> entry) {super(referent, queue);this.entry = entry;}@Overridepublic int getWeight() {return 1;}@Overridepublic ReferenceEntry<K, V> getEntry() {return entry;}@Overridepublic void notifyNewValue(V newValue) {}@Overridepublic ValueReference<K, V> copyFor(ReferenceQueue<V> queue, V value, ReferenceEntry<K, V> entry) {return new WeakValueReference<K, V>(queue, value, entry);}@Overridepublic boolean isLoading() {return false;}@Overridepublic boolean isActive() {return true;}@Overridepublic V waitForValue() {return get();}}/*** References a soft value.*/static class SoftValueReference<K, V>extends SoftReference<V> implements ValueReference<K, V> {final ReferenceEntry<K, V> entry;SoftValueReference(ReferenceQueue<V> queue, V referent, ReferenceEntry<K, V> entry) {super(referent, queue);this.entry = entry;}@Overridepublic int getWeight() {return 1;}@Overridepublic ReferenceEntry<K, V> getEntry() {return entry;}@Overridepublic void notifyNewValue(V newValue) {}@Overridepublic ValueReference<K, V> copyFor(ReferenceQueue<V> queue, V value, ReferenceEntry<K, V> entry) {return new SoftValueReference<K, V>(queue, value, entry);}@Overridepublic boolean isLoading() {return false;}@Overridepublic boolean isActive() {return true;}@Overridepublic V waitForValue() {return get();}}/*** References a strong value.*/static class StrongValueReference<K, V> implements ValueReference<K, V> {final V referent;StrongValueReference(V referent) {this.referent = referent;}@Overridepublic V get() {return referent;}@Overridepublic int getWeight() {return 1;}@Overridepublic ReferenceEntry<K, V> getEntry() {return null;}@Overridepublic ValueReference<K, V> copyFor(ReferenceQueue<V> queue, V value, ReferenceEntry<K, V> entry) {return this;}@Overridepublic boolean isLoading() {return false;}@Overridepublic boolean isActive() {return true;}@Overridepublic V waitForValue() {return get();}@Overridepublic void notifyNewValue(V newValue) {}}/*** References a weak value.*/static final class WeightedWeakValueReference<K, V> extends WeakValueReference<K, V> {final int weight;WeightedWeakValueReference(ReferenceQueue<V> queue, V referent, ReferenceEntry<K, V> entry,int weight) {super(queue, referent, entry);this.weight = weight;}@Overridepublic int getWeight() {return weight;}@Overridepublic ValueReference<K, V> copyFor(ReferenceQueue<V> queue, V value, ReferenceEntry<K, V> entry) {return new WeightedWeakValueReference<K, V>(queue, value, entry, weight);}}/*** References a soft value.*/static final class WeightedSoftValueReference<K, V> extends SoftValueReference<K, V> {final int weight;WeightedSoftValueReference(ReferenceQueue<V> queue, V referent, ReferenceEntry<K, V> entry,int weight) {super(queue, referent, entry);this.weight = weight;}@Overridepublic int getWeight() {return weight;}@Overridepublic ValueReference<K, V> copyFor(ReferenceQueue<V> queue, V value, ReferenceEntry<K, V> entry) {return new WeightedSoftValueReference<K, V>(queue, value, entry, weight);}}/*** References a strong value.*/static final class WeightedStrongValueReference<K, V> extends StrongValueReference<K, V> {final int weight;WeightedStrongValueReference(V referent, int weight) {super(referent);this.weight = weight;}@Overridepublic int getWeight() {return weight;}}/*** Applies a supplemental hash function to a given hash code, which defends against poor quality hash functions. This is critical when* the concurrent hash map uses power-of-two length hash tables, that otherwise encounter collisions for hash codes that do not differ* in lower or upper bits.** @param h hash code*/static int rehash(int h) {// Spread bits to regularize both segment and index locations,// using variant of single-word Wang/Jenkins hash.// TODO(kevinb): use Hashing/move this to Hashing?h += (h << 15) ^ 0xffffcd7d;h ^= (h >>> 10);h += (h << 3);h ^= (h >>> 6);h += (h << 2) + (h << 14);return h ^ (h >>> 16);}/*** This method is a convenience for testing. Code should call {@link Segment#newEntry} directly.*/@GuardedBy("Segment.this")@VisibleForTestingReferenceEntry<K, V> newEntry(K key, int hash, @Nullable ReferenceEntry<K, V> next) {return segmentFor(hash).newEntry(key, hash, next);}/*** This method is a convenience for testing. Code should call {@link Segment#copyEntry} directly.*/@GuardedBy("Segment.this")@VisibleForTestingReferenceEntry<K, V> copyEntry(ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {int hash = original.getHash();return segmentFor(hash).copyEntry(original, newNext);}/*** This method is a convenience for testing. Code should call {@link Segment#setValue} instead.*/@GuardedBy("Segment.this")@VisibleForTestingValueReference<K, V> newValueReference(ReferenceEntry<K, V> entry, V value, int weight) {int hash = entry.getHash();return valueStrength.referenceValue(segmentFor(hash), entry, checkNotNull(value), weight);}int hash(@Nullable Object key) {int h = keyEquivalence.hash(key);return rehash(h);}void reclaimValue(ValueReference<K, V> valueReference) {ReferenceEntry<K, V> entry = valueReference.getEntry();int hash = entry.getHash();segmentFor(hash).reclaimValue(entry.getKey(), hash, valueReference);}void reclaimKey(ReferenceEntry<K, V> entry) {int hash = entry.getHash();// 这里是存放到ReferenceEntry中的key的hash值segmentFor(hash).reclaimKey(entry, hash);}/*** This method is a convenience for testing. Code should call {@link Segment#getLiveValue} instead.*/@VisibleForTestingboolean isLive(ReferenceEntry<K, V> entry, long now) {return segmentFor(entry.getHash()).getLiveValue(entry, now) != null;}/*** Returns the segment that should be used for a key with the given hash.** @param hash the hash code for the key* @return the segment*/Segment<K, V> segmentFor(int hash) {// TODO(fry): Lazily create segments?return segments[(hash >>> segmentShift) & segmentMask];}Segment<K, V> createSegment(int initialCapacity, long maxSegmentWeight, StatsCounter statsCounter) {return new Segment<K, V>(this, initialCapacity, maxSegmentWeight, statsCounter);}/*** Gets the value from an entry. Returns null if the entry is invalid, partially-collected, loading, or expired. Unlike {@link* Segment#getLiveValue} this method does not attempt to cleanup stale entries. As such it should only be called outside of a segment* context, such as during iteration.*/@NullableV getLiveValue(ReferenceEntry<K, V> entry, long now) {if (entry.getKey() == null) {return null;}V value = entry.getValueReference().get();if (value == null) {return null;}if (isExpired(entry, now)) {return null;}return value;}// expiration/*** Returns true if the entry has expired.*/boolean isExpired(ReferenceEntry<K, V> entry, long now) {checkNotNull(entry);if (expiresAfterAccess()&& (now - entry.getAccessTime() >= expireAfterAccessNanos)) {return true;}if (expiresAfterWrite()&& (now - entry.getWriteTime() >= expireAfterWriteNanos)) {return true;}return false;}// queues@GuardedBy("Segment.this")static <K, V> void connectAccessOrder(ReferenceEntry<K, V> previous, ReferenceEntry<K, V> next) {previous.setNextInAccessQueue(next);next.setPreviousInAccessQueue(previous);}@GuardedBy("Segment.this")static <K, V> void nullifyAccessOrder(ReferenceEntry<K, V> nulled) {ReferenceEntry<K, V> nullEntry = nullEntry();nulled.setNextInAccessQueue(nullEntry);nulled.setPreviousInAccessQueue(nullEntry);}@GuardedBy("Segment.this")static <K, V> void connectWriteOrder(ReferenceEntry<K, V> previous, ReferenceEntry<K, V> next) {previous.setNextInWriteQueue(next);next.setPreviousInWriteQueue(previous);}@GuardedBy("Segment.this")static <K, V> void nullifyWriteOrder(ReferenceEntry<K, V> nulled) {ReferenceEntry<K, V> nullEntry = nullEntry();nulled.setNextInWriteQueue(nullEntry);nulled.setPreviousInWriteQueue(nullEntry);}/*** Notifies listeners that an entry has been automatically removed due to expiration, eviction, or eligibility for garbage collection.* This should be called every time expireEntries or evictEntry is called (once the lock is released).*/void processPendingNotifications() {RemovalNotification<K, V> notification;while ((notification = removalNotificationQueue.poll()) != null) {try {removalListener.onRemoval(notification);} catch (Throwable e) {logger.log(Level.WARNING, "Exception thrown by removal listener", e);}}}@SuppressWarnings("unchecked")final Segment<K, V>[] newSegmentArray(int ssize) {return new Segment[ssize];}// Inner Classes/*** Segments are specialized versions of hash tables. This subclass inherits from ReentrantLock opportunistically, just to simplify some* locking and avoid separate construction.*/@SuppressWarnings("serial") // This class is never serialized.// 这里为啥是集成ReentrantLock，而不是像我们平常用的那样，在Segement中：Lock lock= new ReentrantLock()，然后在需要加锁/释放锁的时候使用lock.lock()/lock.unlock()，区别是啥？// jdk中的实现很多都是采用了这种集成的方式，而不是组合方式，有啥区别？/*** 1.首先在效果上是没有任何区别的，继承的方式不需要显示的去创建ReentrantLock对象，而组合需要。但使用方式都一样，需要在finally块中释放锁。* 2.这种继承的方式，是不是有点像synchronize锁的方式了：synchronize的锁标志是放在对象头中的；集成方式是放在父类AQS的state字段中的，所以继承ReentrantLock，使用起来有没有一点像synchronize(this){* //业务逻辑的方式}*/static class Segment<K, V> extends ReentrantLock {/** TODO(fry): Consider copying variables (like evictsBySize) from outer class into this class.* It will require more memory but will reduce indirection.*//** Segments maintain a table of entry lists that are ALWAYS kept in a consistent state, so can* be read without locking. Next fields of nodes are immutable (final). All list additions are* performed at the front of each bin. This makes it easy to check changes, and also fast to* traverse. When nodes would otherwise be changed, new nodes are created to replace them. This* works well for hash tables since the bin lists tend to be short. (The average length is less* than two.)** Read operations can thus proceed without locking, but rely on selected uses of volatiles to* ensure that completed write operations performed by other threads are noticed. For most* purposes, the "count" field, tracking the number of elements, serves as that volatile* variable ensuring visibility. This is convenient because this field needs to be read in many* read operations anyway:** - All (unsynchronized) read operations must first read the "count" field, and should not* look at table entries if it is 0.** - All (synchronized) write operations should write to the "count" field after structurally* changing any bin. The operations must not take any action that could even momentarily* cause a concurrent read operation to see inconsistent data. This is made easier by the* nature of the read operations in Map. For example, no operation can reveal that the table* has grown but the threshold has not yet been updated, so there are no atomicity requirements* for this with respect to reads.** As a guide, all critical volatile reads and writes to the count field are marked in code* comments.*/final LocalCache<K, V> map;/*** The number of live elements in this segment's region.*/volatile int count;/*** The weight of the live elements in this segment's region.*/@GuardedBy("Segment.this")int totalWeight;/*** Number of updates that alter the size of the table. This is used during bulk-read methods to make sure they see a consistent* snapshot: If modCounts change during a traversal of segments loading size or checking containsValue, then we might have an* inconsistent view of state so (usually) must retry.*/int modCount;/*** The table is expanded when its size exceeds this threshold. (The value of this field is always {@code (int) (capacity * 0.75)}.)*/int threshold;/*** The per-segment table.*/volatile AtomicReferenceArray<ReferenceEntry<K, V>> table;/*** The maximum weight of this segment. UNSET_INT if there is no maximum.*/final long maxSegmentWeight;/*** The key reference queue contains entries whose keys have been garbage collected, and which need to be cleaned up internally.*/final ReferenceQueue<K> keyReferenceQueue;/*** The value reference queue contains value references whose values have been garbage collected, and which need to be cleaned up* internally.*/final ReferenceQueue<V> valueReferenceQueue;/*** The recency queue is used to record which entries were accessed for updating the access list's ordering. It is drained as a batch* operation when either the DRAIN_THRESHOLD is crossed or a write occurs on the segment.*/final Queue<ReferenceEntry<K, V>> recencyQueue;/*** A counter of the number of reads since the last write, used to drain queues on a small fraction of read operations.*/final AtomicInteger readCount = new AtomicInteger();/*** A queue of elements currently in the map, ordered by write time. Elements are added to the tail of the queue on write.*/@GuardedBy("Segment.this")final Queue<ReferenceEntry<K, V>> writeQueue;/*** A queue of elements currently in the map, ordered by access time. Elements are added to the tail of the queue on access (note* that writes count as accesses).*/@GuardedBy("Segment.this")final Queue<ReferenceEntry<K, V>> accessQueue;/*** Accumulates cache statistics.*/final StatsCounter statsCounter;Segment(LocalCache<K, V> map, int initialCapacity, long maxSegmentWeight,StatsCounter statsCounter) {this.map = map;this.maxSegmentWeight = maxSegmentWeight;this.statsCounter = checkNotNull(statsCounter);initTable(newEntryArray(initialCapacity));keyReferenceQueue = map.usesKeyReferences()? new ReferenceQueue<K>() : null;valueReferenceQueue = map.usesValueReferences()? new ReferenceQueue<V>() : null;recencyQueue = map.usesAccessQueue()? new ConcurrentLinkedQueue<ReferenceEntry<K, V>>(): LocalCache.<ReferenceEntry<K, V>>discardingQueue();writeQueue = map.usesWriteQueue()? new WriteQueue<K, V>(): LocalCache.<ReferenceEntry<K, V>>discardingQueue();accessQueue = map.usesAccessQueue()? new AccessQueue<K, V>(): LocalCache.<ReferenceEntry<K, V>>discardingQueue();}AtomicReferenceArray<ReferenceEntry<K, V>> newEntryArray(int size) {return new AtomicReferenceArray<ReferenceEntry<K, V>>(size);}void initTable(AtomicReferenceArray<ReferenceEntry<K, V>> newTable) {this.threshold = newTable.length() * 3 / 4; // 0.75if (!map.customWeigher() && this.threshold == maxSegmentWeight) {// prevent spurious expansion before evictionthis.threshold++;}this.table = newTable;}@GuardedBy("Segment.this")ReferenceEntry<K, V> newEntry(K key, int hash, @Nullable ReferenceEntry<K, V> next) {return map.entryFactory.newEntry(this, checkNotNull(key), hash, next);}/*** Copies {@code original} into a new entry chained to {@code newNext}. Returns the new entry, or {@code null} if {@code original}* was already garbage collected.*/@GuardedBy("Segment.this")ReferenceEntry<K, V> copyEntry(ReferenceEntry<K, V> original, ReferenceEntry<K, V> newNext) {if (original.getKey() == null) {// key collectedreturn null;}ValueReference<K, V> valueReference = original.getValueReference();V value = valueReference.get();if ((value == null) && valueReference.isActive()) {// value collectedreturn null;}ReferenceEntry<K, V> newEntry = map.entryFactory.copyEntry(this, original, newNext);newEntry.setValueReference(valueReference.copyFor(this.valueReferenceQueue, value, newEntry));return newEntry;}/*** Sets a new value of an entry. Adds newly created entries at the end of the access queue.*/@GuardedBy("Segment.this")void setValue(ReferenceEntry<K, V> entry, K key, V value, long now) {ValueReference<K, V> previous = entry.getValueReference();int weight = map.weigher.weigh(key, value);checkState(weight >= 0, "Weights must be non-negative");ValueReference<K, V> valueReference =map.valueStrength.referenceValue(this, entry, value, weight);entry.setValueReference(valueReference);recordWrite(entry, weight, now);previous.notifyNewValue(value);}// loadingV get(K key, int hash, CacheLoader<? super K, V> loader) throws ExecutionException {checkNotNull(key);checkNotNull(loader);try {if (count != 0) { // read-volatile// don't call getLiveEntry, which would ignore loading valuesReferenceEntry<K, V> e = getEntry(key, hash);if (e != null) {long now = map.ticker.read();V value = getLiveValue(e, now);if (value != null) {recordRead(e, now);statsCounter.recordHits(1);// scheduleRefresh()这个地方首先，如果需要reload()就返回reload的值；否则返回入参value//这个方法的作用就是：如果设置了refreshAfterWrite(3, TimeUnit.SECONDS)，那么这里就会去检查，上次写入时间举例现在是否超过了3s// 如果超过了，那就去调用CacheLoader去重新加载更新一遍数据。注意：虽然这里些的是scheduleRefresh()，但是guava27版本的实现中其实都还是同步的// 并没有一部实现。return scheduleRefresh(e, key, hash, value, now, loader);}ValueReference<K, V> valueReference = e.getValueReference();if (valueReference.isLoading()) {return waitForLoadingValue(e, key, valueReference);}}}// 缓存中不存在的时候，调用CacheLoader去重新加载进缓存return lockedGetOrLoad(key, hash, loader);} catch (ExecutionException ee) {Throwable cause = ee.getCause();if (cause instanceof Error) {throw new ExecutionError((Error) cause);} else if (cause instanceof RuntimeException) {throw new UncheckedExecutionException(cause);}throw ee;} finally {postReadCleanup();}}V lockedGetOrLoad(K key, int hash, CacheLoader<? super K, V> loader)throws ExecutionException {ReferenceEntry<K, V> e;ValueReference<K, V> valueReference = null;LoadingValueReference<K, V> loadingValueReference = null;boolean createNewEntry = true;lock();// 加锁，即要获得Segement上的锁。try {// re-read ticker once inside the locklong now = map.ticker.read();preWriteCleanup(now);int newCount = this.count - 1;AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null && map.keyEquivalence.equivalent(key, entryKey)) {valueReference = e.getValueReference();if (valueReference.isLoading()) {createNewEntry = false;} else {V value = valueReference.get();if (value == null) {enqueueNotification(entryKey, hash, valueReference, RemovalCause.COLLECTED);} else if (map.isExpired(e, now)) {// This is a duplicate check, as preWriteCleanup already purged expired// entries, but let's accomodate an incorrect expiration queue.enqueueNotification(entryKey, hash, valueReference, RemovalCause.EXPIRED);} else {recordLockedRead(e, now);statsCounter.recordHits(1);// we were concurrent with loading; don't consider refreshreturn value;}// immediately reuse invalid entrieswriteQueue.remove(e);accessQueue.remove(e);this.count = newCount; // write-volatile}break;}}if (createNewEntry) {loadingValueReference = new LoadingValueReference<K, V>();if (e == null) {e = newEntry(key, hash, first);e.setValueReference(loadingValueReference);table.set(index, e);} else {e.setValueReference(loadingValueReference);}}} finally {unlock();postWriteCleanup();}if (createNewEntry) {try {// Synchronizes on the entry to allow failing fast when a recursive load is// detected. This may be circumvented when an entry is copied, but will fail fast most// of the time.synchronized (e) {return loadSync(key, hash, loadingValueReference, loader);}} finally {statsCounter.recordMisses(1);}} else {// The entry already exists. Wait for loading.return waitForLoadingValue(e, key, valueReference);}}V waitForLoadingValue(ReferenceEntry<K, V> e, K key, ValueReference<K, V> valueReference)throws ExecutionException {if (!valueReference.isLoading()) {throw new AssertionError();}checkState(!Thread.holdsLock(e), "Recursive load");// don't consider expiration as we're concurrent with loadingtry {V value = valueReference.waitForValue();if (value == null) {throw new InvalidCacheLoadException("CacheLoader returned null for key " + key + ".");}// re-read ticker now that loading has completedlong now = map.ticker.read();recordRead(e, now);return value;} finally {statsCounter.recordMisses(1);}}// at most one of loadSync/loadAsync may be called for any given LoadingValueReferenceV loadSync(K key, int hash, LoadingValueReference<K, V> loadingValueReference,CacheLoader<? super K, V> loader) throws ExecutionException {ListenableFuture<V> loadingFuture = loadingValueReference.loadFuture(key, loader);return getAndRecordStats(key, hash, loadingValueReference, loadingFuture);}ListenableFuture<V> loadAsync(final K key, final int hash, final LoadingValueReference<K, V> loadingValueReference, CacheLoader<? super K, V> loader) {// 实际去调用loader重新加载数据的。final ListenableFuture<V> loadingFuture = loadingValueReference.loadFuture(key, loader);loadingFuture.addListener(new Runnable() {@Overridepublic void run() {try {V newValue = getAndRecordStats(key, hash, loadingValueReference, loadingFuture);} catch (Throwable t) {logger.log(Level.WARNING, "Exception thrown during refresh", t);loadingValueReference.setException(t);}}}, sameThreadExecutor);return loadingFuture;}/*** Waits uninterruptibly for {@code newValue} to be loaded, and then records loading stats.*/V getAndRecordStats(K key, int hash, LoadingValueReference<K, V> loadingValueReference,ListenableFuture<V> newValue) throws ExecutionException {V value = null;try {value = getUninterruptibly(newValue);if (value == null) {throw new InvalidCacheLoadException("CacheLoader returned null for key " + key + ".");}statsCounter.recordLoadSuccess(loadingValueReference.elapsedNanos());storeLoadedValue(key, hash, loadingValueReference, value);return value;} finally {if (value == null) {statsCounter.recordLoadException(loadingValueReference.elapsedNanos());removeLoadingValue(key, hash, loadingValueReference);}}}V scheduleRefresh(ReferenceEntry<K, V> entry, K key, int hash, V oldValue, long now, CacheLoader<? super K, V> loader) {// 设置了refreshAfterWrite() 并且上次写入时间离当前时间大于了指定时间  并且value没有处于loading中// (除了LoadingValueReference返回true外，其他都返回false，因为Loading过程中，会用LoadingValueReference来封装一个ValueReference)if (map.refreshes() && (now - entry.getWriteTime() > map.refreshNanos) && !entry.getValueReference().isLoading()) {V newValue = refresh(key, hash, loader, true);if (newValue != null) {return newValue;}}return oldValue;}/*** Refreshes the value associated with {@code key}, unless another thread is already doing so. Returns the newly refreshed value* associated with {@code key} if it was refreshed inline, or {@code null} if another thread is performing the refresh or if an* error occurs during refresh.*/@NullableV refresh(K key, int hash, CacheLoader<? super K, V> loader, boolean checkTime) {// 根据key找到要重新加载的ReferenceEntry，从而就知道了这个key对应的value引用：ValueReferfence,只是这里用LoadingValueReference封装了一下。final LoadingValueReference<K, V> loadingValueReference = insertLoadingValueReference(key, hash, checkTime);if (loadingValueReference == null) {return null;}// 开始重新加载对应的key，这里虽然写的是异步，但默认实现是同步的。ps：异步线程和当前线程一样，那不就是同步么ListenableFuture<V> result = loadAsync(key, hash, loadingValueReference, loader);// 即使实现了异步，这里要注意，如果到这里异步reload()还没执行完，这里相当于就把reload()的结果给丢掉了，返回null。在调用放看来就是缓存中没有。if (result.isDone()) {try {return Uninterruptibles.getUninterruptibly(result);} catch (Throwable t) {// don't let refresh exceptions propagate; error was already logged}}return null;}/*** Returns a newly inserted {@code LoadingValueReference}, or null if the live value reference is already loading.*/@NullableLoadingValueReference<K, V> insertLoadingValueReference(final K key, final int hash,boolean checkTime) {ReferenceEntry<K, V> e = null;lock();try {long now = map.ticker.read();preWriteCleanup(now);AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);// Look for an existing entry.for (e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {// We found an existing entry.ValueReference<K, V> valueReference = e.getValueReference();if (valueReference.isLoading()|| (checkTime && (now - e.getWriteTime() < map.refreshNanos))) {// refresh is a no-op if loading is pending// if checkTime, we want to check *after* acquiring the lock if refresh still needs// to be scheduledreturn null;}// continue returning old value while loading++modCount;LoadingValueReference<K, V> loadingValueReference = new LoadingValueReference<K, V>(valueReference);e.setValueReference(loadingValueReference);return loadingValueReference;}}++modCount;LoadingValueReference<K, V> loadingValueReference = new LoadingValueReference<K, V>();e = newEntry(key, hash, first);e.setValueReference(loadingValueReference);table.set(index, e);return loadingValueReference;} finally {unlock();postWriteCleanup();}}// reference queues, for garbage collection cleanup/*** Cleanup collected entries when the lock is available.*/void tryDrainReferenceQueues() {if (tryLock()) {try {drainReferenceQueues();} finally {unlock();}}}/*** Drain the key and value reference queues, cleaning up internal entries containing garbage collected keys or values.*/@GuardedBy("Segment.this")void drainReferenceQueues() {// 如果key不是强引用，则回收那些key实际内容被gc掉了的缓存项目if (map.usesKeyReferences()) {//keyStrength != Strength.STRONG;drainKeyReferenceQueue();}// 如果value不是强引用，则回收那些value实际内容被gc掉了的缓存项目if (map.usesValueReferences()) {//valueStrength != Strength.STRONG;drainValueReferenceQueue();}}@GuardedBy("Segment.this")void drainKeyReferenceQueue() {Reference<? extends K> ref;int i = 0;// 如果keyStrength!=Strong，即key的引用你是WeakReference或者SoftReference，在gc的时候，可能会把引用中对应的值给回收掉，// 只剩下WeakReference对象本身。在WeakReference实现的时候，会传入一个ReferenceQueue，gc的时候，jvm会将被回收了实际值的WeakReference放到这个队列中// 以便通知用户。在Segement的实现中，保留了传个WeakReference的这个ReferenceQueue，这个地方就是检查这个Queue，如果Queue不为空，说明被回收掉了// 那就将对应的缓存ReferenceEntry从缓存中删除。while ((ref = keyReferenceQueue.poll()) != null) {@SuppressWarnings("unchecked")ReferenceEntry<K, V> entry = (ReferenceEntry<K, V>) ref;// 这里就是找到entry所在的链表，然后将entry从链表中删除。map.reclaimKey(entry);if (++i == DRAIN_MAX) {break;}}}@GuardedBy("Segment.this")void drainValueReferenceQueue() {Reference<? extends V> ref;int i = 0;while ((ref = valueReferenceQueue.poll()) != null) {@SuppressWarnings("unchecked")ValueReference<K, V> valueReference = (ValueReference<K, V>) ref;map.reclaimValue(valueReference);if (++i == DRAIN_MAX) {break;}}}/*** Clears all entries from the key and value reference queues.*/void clearReferenceQueues() {if (map.usesKeyReferences()) {clearKeyReferenceQueue();}if (map.usesValueReferences()) {clearValueReferenceQueue();}}void clearKeyReferenceQueue() {while (keyReferenceQueue.poll() != null) {}}void clearValueReferenceQueue() {while (valueReferenceQueue.poll() != null) {}}// recency queue, shared by expiration and eviction/*** Records the relative order in which this read was performed by adding {@code entry} to the recency queue. At write-time, or when* the queue is full past the threshold, the queue will be drained and the entries therein processed.** <p>Note: locked reads should use {@link #recordLockedRead}.*/void recordRead(ReferenceEntry<K, V> entry, long now) {if (map.recordsAccess()) {entry.setAccessTime(now);}recencyQueue.add(entry);}/*** Updates the eviction metadata that {@code entry} was just read. This currently amounts to adding {@code entry} to relevant* eviction lists.** <p>Note: this method should only be called under lock, as it directly manipulates the* eviction queues. Unlocked reads should use {@link #recordRead}.*/@GuardedBy("Segment.this")void recordLockedRead(ReferenceEntry<K, V> entry, long now) {if (map.recordsAccess()) {entry.setAccessTime(now);}accessQueue.add(entry);}/*** Updates eviction metadata that {@code entry} was just written. This currently amounts to adding {@code entry} to relevant* eviction lists.*/@GuardedBy("Segment.this")void recordWrite(ReferenceEntry<K, V> entry, int weight, long now) {// we are already under lock, so drain the recency queue immediatelydrainRecencyQueue();totalWeight += weight;if (map.recordsAccess()) {entry.setAccessTime(now);}if (map.recordsWrite()) {entry.setWriteTime(now);}// accessQueue是用来实现LRU的，每次put()/get()都将对应的ReferenceEntry放到了accessQueeu的队尾，// 然后在淘汰的时候，如果缓存个数大于规定容量，就从accessQueue中拿出来一个，从缓存中删除accessQueue.add(entry);writeQueue.add(entry);}/*** Drains the recency queue, updating eviction metadata that the entries therein were read in the specified relative order. This* currently amounts to adding them to relevant eviction lists (accounting for the fact that they could have been removed from the* map since being added to the recency queue).*/@GuardedBy("Segment.this")void drainRecencyQueue() {ReferenceEntry<K, V> e;// recencyQueue 啥时候往里放？while ((e = recencyQueue.poll()) != null) {// An entry may be in the recency queue despite it being removed from// the map . This can occur when the entry was concurrently read while a// writer is removing it from the segment or after a clear has removed// all of the segment's entries./*** guava cache实现LRU是依靠accessQueue，相当于每次访问都将访问的Entry放到这个队列的尾部，这样在put()的时候就会判断，如果缓存中元素超过了最大容量* ，就会从accessQueue队列中直接出队元素，然后从缓存中删除这个元素，以实现LRU算法。* 但是有个问题：accessQueue是线程不安全的，所以将元素移动到队头需要获取Segment锁，然后才能移动元素到队尾。那如果每次get()都要去获得一次锁，如果有竞争，就会阻塞查询。* 所以，guava就有了recencyQueue，它是一个jdk中线程安全的Queue(ConcurrentLinkedQueue)，每次get()的时候，都将查询到的Entry往这个队列中放，* 然后在查询的最后tryLock()，如果没有获得锁，那么就会调用这个方法，然后将recencyQueue清空，然后将recencyQueue的元素都放到accessQueue的队尾。*/if (accessQueue.contains(e)) {// 这里这么判断一下的原因，就是有可能recencyQueue中的元素已经删除了，因为元素删除/过期等都不会去处理recencyQueue。// 这里accessQueue是LocalCache中自己实现的，add()方法：如果e已经在队列中存在，就是把对应的e移动到队尾；如果不存在就加到队尾accessQueue.add(e);}}}// expiration/*** Cleanup expired entries when the lock is available.*/void tryExpireEntries(long now) {if (tryLock()) {try {expireEntries(now);} finally {unlock();// don't call postWriteCleanup as we're in a read}}}@GuardedBy("Segment.this")void expireEntries(long now) {// recencyQueue暂下的一批读取过的ReferenceEntry，依次将他们在accessQueue中放到队尾去drainRecencyQueue();ReferenceEntry<K, V> e;// 清除掉那些最近写时间过期的缓存项。while ((e = writeQueue.peek()) != null && map.isExpired(e, now)) {if (!removeEntry(e, e.getHash(), RemovalCause.EXPIRED)) {throw new AssertionError();}}// 清除掉那些最近访问时间过期的缓存项。while ((e = accessQueue.peek()) != null && map.isExpired(e, now)) {if (!removeEntry(e, e.getHash(), RemovalCause.EXPIRED)) {throw new AssertionError();}}}// eviction@GuardedBy("Segment.this")void enqueueNotification(ReferenceEntry<K, V> entry, RemovalCause cause) {enqueueNotification(entry.getKey(), entry.getHash(), entry.getValueReference(), cause);}@GuardedBy("Segment.this")void enqueueNotification(@Nullable K key, int hash, ValueReference<K, V> valueReference,RemovalCause cause) {totalWeight -= valueReference.getWeight();if (cause.wasEvicted()) {statsCounter.recordEviction();}if (map.removalNotificationQueue != DISCARDING_QUEUE) {V value = valueReference.get();RemovalNotification<K, V> notification = new RemovalNotification<K, V>(key, value, cause);map.removalNotificationQueue.offer(notification);}}/*** Performs eviction if the segment is full. This should only be called prior to adding a new entry and increasing {@code count}.*/@GuardedBy("Segment.this")void evictEntries() {if (!map.evictsBySize()) {return;}drainRecencyQueue();// totalWeight是当前Segement中的ReferenceEntry的个数，maxSegmentWeight是当前Segement最多才能出的Segement个数。// maxSegmentWeight计算逻辑？while (totalWeight > maxSegmentWeight) {// getNextEvictable()这个就是在从accessQueue中出队ReferenceEntry<K, V> e = getNextEvictable();if (!removeEntry(e, e.getHash(), RemovalCause.SIZE)) {throw new AssertionError();}}}// TODO(fry): instead implement this with an eviction headReferenceEntry<K, V> getNextEvictable() {for (ReferenceEntry<K, V> e : accessQueue) {int weight = e.getValueReference().getWeight();if (weight > 0) {return e;}}throw new AssertionError();}/*** Returns first entry of bin for given hash.*/ReferenceEntry<K, V> getFirst(int hash) {// read this volatile field only onceAtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;return table.get(hash & (table.length() - 1));}// Specialized implementations of map methods@NullableReferenceEntry<K, V> getEntry(Object key, int hash) {for (ReferenceEntry<K, V> e = getFirst(hash); e != null; e = e.getNext()) {if (e.getHash() != hash) {continue;}K entryKey = e.getKey();if (entryKey == null) {tryDrainReferenceQueues();continue;}if (map.keyEquivalence.equivalent(key, entryKey)) {return e;}}return null;}@NullableReferenceEntry<K, V> getLiveEntry(Object key, int hash, long now) {ReferenceEntry<K, V> e = getEntry(key, hash);if (e == null) {return null;} else if (map.isExpired(e, now)) {tryExpireEntries(now);return null;}return e;}/*** Gets the value from an entry. Returns null if the entry is invalid, partially-collected, loading, or expired.*/V getLiveValue(ReferenceEntry<K, V> entry, long now) {if (entry.getKey() == null) {tryDrainReferenceQueues();return null;}V value = entry.getValueReference().get();if (value == null) {tryDrainReferenceQueues();return null;}if (map.isExpired(entry, now)) {tryExpireEntries(now);return null;}return value;}@NullableV get(Object key, int hash) {try {if (count != 0) { // read-volatilelong now = map.ticker.read();ReferenceEntry<K, V> e = getLiveEntry(key, hash, now);if (e == null) {return null;}V value = e.getValueReference().get();if (value != null) {recordRead(e, now);// scheduleRefresh()这个地方首先，如果需要reload()就返回reload的值；否则返回入参value//这个方法的作用就是：如果设置了refreshAfterWrite(3, TimeUnit.SECONDS)，那么这里就会去检查，上次写入时间举例现在是否超过了3s// 如果超过了，那就去调用CacheLoader去重新加载更新一遍数据。注意：虽然这里些的是scheduleRefresh()，但是guava27版本的实现中其实都还是同步的// 并没有一部实现。return scheduleRefresh(e, e.getKey(), hash, value, now, map.defaultLoader);}tryDrainReferenceQueues();}return null;} finally {// 这个就是尝试加锁，如果加锁成功将recencyQueue清空，然后将对应的ReferenceEntry一次放到了accessQueue的队头postReadCleanup();}}boolean containsKey(Object key, int hash) {try {if (count != 0) { // read-volatilelong now = map.ticker.read();ReferenceEntry<K, V> e = getLiveEntry(key, hash, now);if (e == null) {return false;}return e.getValueReference().get() != null;}return false;} finally {postReadCleanup();}}/*** This method is a convenience for testing. Code should call {@link LocalCache#containsValue} directly.*/@VisibleForTestingboolean containsValue(Object value) {try {if (count != 0) { // read-volatilelong now = map.ticker.read();AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int length = table.length();for (int i = 0; i < length; ++i) {for (ReferenceEntry<K, V> e = table.get(i); e != null; e = e.getNext()) {V entryValue = getLiveValue(e, now);if (entryValue == null) {continue;}if (map.valueEquivalence.equivalent(value, entryValue)) {return true;}}}}return false;} finally {postReadCleanup();}}@NullableV put(K key, int hash, V value, boolean onlyIfAbsent) {lock();try {long now = map.ticker.read();// 这个地方会清理另种情况的缓存项：// 1. 如果key或者value的引用类型不是强引用，那么key或者value任意一个的具体内容被清除了，这个方法就会将对应的entry从缓存中删除。// 2. 过期时间自动删除：比如设置了expireAfterWrite()/expireAfterAccess()，那么超过这个时间没有访问，对应的缓存项就会被删除。这个会在这里作一次清理。preWriteCleanup(now);int newCount = this.count + 1;if (newCount > this.threshold) {expand();// 对table数组扩容。newCount = this.count + 1;}// first就是一个table数组中链表的头结点AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);// Look for an existing entry.for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();// 这个if中条件=true，说明put(key，value)对应的缓存项已经存在。if (e.getHash() == hash && entryKey != null && map.keyEquivalence.equivalent(key, entryKey)) {// We found an existing entry.ValueReference<K, V> valueReference = e.getValueReference();V entryValue = valueReference.get();// 如果设置了softValue()/weakValue()，那么缓存的key-value中的value就不是强引用，那就可能由于内存不够发生gc的时候，将实际的数据回收掉了// 只是剩下引用对象本身，所以这个地方entryValue==null就会成立。if (entryValue == null) {++modCount;if (valueReference.isActive()) {enqueueNotification(key, hash, valueReference, RemovalCause.COLLECTED);setValue(e, key, value, now);newCount = this.count; // count remains unchanged} else {setValue(e, key, value, now);newCount = this.count + 1;}this.count = newCount; // write-volatile// LRU淘汰，超过缓存大小，淘汰。evictEntries();return null;} else if (onlyIfAbsent) {// Mimic// "if (!map.containsKey(key)) ...// else return map.get(key);recordLockedRead(e, now);return entryValue;} else {// clobber existing entry, count remains unchanged++modCount;enqueueNotification(key, hash, valueReference, RemovalCause.REPLACED);setValue(e, key, value, now);evictEntries();return entryValue;}}}// Create a new entry.++modCount;ReferenceEntry<K, V> newEntry = newEntry(key, hash, first);setValue(newEntry, key, value, now);table.set(index, newEntry);newCount = this.count + 1;this.count = newCount; // write-volatile// LRU缓存淘汰：guava cache的LRU策略是依赖Segement#accessQueue和recencyQueue来实现的：// 1. 每次put()的时候，如果对应的key存在，则将这个key对应的ReferenceEntry移动到accessQueue的队尾；如果不存在，新增一个ReferenceEntry，也会放到accessQueue的队尾// 2. 没测get()的时候，都会将查询到的这个ReferenceEntry放到recencyQueue的队尾。evictEntries();return null;} finally {unlock();postWriteCleanup();}}/*** Expands the table if possible.*/@GuardedBy("Segment.this")void expand() {AtomicReferenceArray<ReferenceEntry<K, V>> oldTable = table;int oldCapacity = oldTable.length();if (oldCapacity >= MAXIMUM_CAPACITY) {return;}/** Reclassify nodes in each list to new Map. Because we are using power-of-two expansion, the* elements from each bin must either stay at same index, or move with a power of two offset.* We eliminate unnecessary node creation by catching cases where old nodes can be reused* because their next fields won't change. Statistically, at the default threshold, only* about one-sixth of them need cloning when a table doubles. The nodes they replace will be* garbage collectable as soon as they are no longer referenced by any reader thread that may* be in the midst of traversing table right now.*/int newCount = count;AtomicReferenceArray<ReferenceEntry<K, V>> newTable = newEntryArray(oldCapacity << 1);threshold = newTable.length() * 3 / 4;int newMask = newTable.length() - 1;for (int oldIndex = 0; oldIndex < oldCapacity; ++oldIndex) {// We need to guarantee that any existing reads of old Map can// proceed. So we cannot yet null out each bin.ReferenceEntry<K, V> head = oldTable.get(oldIndex);if (head != null) {ReferenceEntry<K, V> next = head.getNext();int headIndex = head.getHash() & newMask;// Single node on listif (next == null) {newTable.set(headIndex, head);} else {// Reuse the consecutive sequence of nodes with the same target// index from the end of the list. tail points to the first// entry in the reusable list.ReferenceEntry<K, V> tail = head;int tailIndex = headIndex;for (ReferenceEntry<K, V> e = next; e != null; e = e.getNext()) {int newIndex = e.getHash() & newMask;if (newIndex != tailIndex) {// The index changed. We'll need to copy the previous entry.tailIndex = newIndex;tail = e;}}newTable.set(tailIndex, tail);// Clone nodes leading up to the tail.for (ReferenceEntry<K, V> e = head; e != tail; e = e.getNext()) {int newIndex = e.getHash() & newMask;ReferenceEntry<K, V> newNext = newTable.get(newIndex);ReferenceEntry<K, V> newFirst = copyEntry(e, newNext);if (newFirst != null) {newTable.set(newIndex, newFirst);} else {removeCollectedEntry(e);newCount--;}}}}}table = newTable;this.count = newCount;}boolean replace(K key, int hash, V oldValue, V newValue) {lock();try {long now = map.ticker.read();preWriteCleanup(now);AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> valueReference = e.getValueReference();V entryValue = valueReference.get();if (entryValue == null) {if (valueReference.isActive()) {// If the value disappeared, this entry is partially collected.int newCount = this.count - 1;++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, entryKey, hash, valueReference, RemovalCause.COLLECTED);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatile}return false;}if (map.valueEquivalence.equivalent(oldValue, entryValue)) {++modCount;enqueueNotification(key, hash, valueReference, RemovalCause.REPLACED);setValue(e, key, newValue, now);evictEntries();return true;} else {// Mimic// "if (map.containsKey(key) && map.get(key).equals(oldValue))..."recordLockedRead(e, now);return false;}}}return false;} finally {unlock();postWriteCleanup();}}@NullableV replace(K key, int hash, V newValue) {lock();try {long now = map.ticker.read();preWriteCleanup(now);AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> valueReference = e.getValueReference();V entryValue = valueReference.get();if (entryValue == null) {if (valueReference.isActive()) {// If the value disappeared, this entry is partially collected.int newCount = this.count - 1;++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, entryKey, hash, valueReference, RemovalCause.COLLECTED);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatile}return null;}++modCount;enqueueNotification(key, hash, valueReference, RemovalCause.REPLACED);setValue(e, key, newValue, now);evictEntries();return entryValue;}}return null;} finally {unlock();postWriteCleanup();}}@NullableV remove(Object key, int hash) {lock();try {long now = map.ticker.read();preWriteCleanup(now);int newCount = this.count - 1;AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> valueReference = e.getValueReference();V entryValue = valueReference.get();RemovalCause cause;if (entryValue != null) {cause = RemovalCause.EXPLICIT;} else if (valueReference.isActive()) {cause = RemovalCause.COLLECTED;} else {// currently loadingreturn null;}++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, entryKey, hash, valueReference, cause);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatilereturn entryValue;}}return null;} finally {unlock();postWriteCleanup();}}boolean storeLoadedValue(K key, int hash, LoadingValueReference<K, V> oldValueReference,V newValue) {lock();try {long now = map.ticker.read();preWriteCleanup(now);int newCount = this.count + 1;if (newCount > this.threshold) { // ensure capacityexpand();newCount = this.count + 1;}AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> valueReference = e.getValueReference();V entryValue = valueReference.get();// replace the old LoadingValueReference if it's live, otherwise// perform a putIfAbsentif (oldValueReference == valueReference|| (entryValue == null && valueReference != UNSET)) {++modCount;if (oldValueReference.isActive()) {RemovalCause cause =(entryValue == null) ? RemovalCause.COLLECTED : RemovalCause.REPLACED;enqueueNotification(key, hash, oldValueReference, cause);newCount--;}setValue(e, key, newValue, now);this.count = newCount; // write-volatileevictEntries();return true;}// the loaded value was already clobberedvalueReference = new WeightedStrongValueReference<K, V>(newValue, 0);enqueueNotification(key, hash, valueReference, RemovalCause.REPLACED);return false;}}++modCount;ReferenceEntry<K, V> newEntry = newEntry(key, hash, first);setValue(newEntry, key, newValue, now);table.set(index, newEntry);this.count = newCount; // write-volatileevictEntries();return true;} finally {unlock();postWriteCleanup();}}boolean remove(Object key, int hash, Object value) {lock();try {long now = map.ticker.read();preWriteCleanup(now);int newCount = this.count - 1;AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> valueReference = e.getValueReference();V entryValue = valueReference.get();RemovalCause cause;if (map.valueEquivalence.equivalent(value, entryValue)) {cause = RemovalCause.EXPLICIT;} else if (entryValue == null && valueReference.isActive()) {cause = RemovalCause.COLLECTED;} else {// currently loadingreturn false;}++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, entryKey, hash, valueReference, cause);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatilereturn (cause == RemovalCause.EXPLICIT);}}return false;} finally {unlock();postWriteCleanup();}}void clear() {if (count != 0) { // read-volatilelock();try {AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;for (int i = 0; i < table.length(); ++i) {for (ReferenceEntry<K, V> e = table.get(i); e != null; e = e.getNext()) {// Loading references aren't actually in the map yet.if (e.getValueReference().isActive()) {enqueueNotification(e, RemovalCause.EXPLICIT);}}}for (int i = 0; i < table.length(); ++i) {table.set(i, null);}clearReferenceQueues();writeQueue.clear();accessQueue.clear();readCount.set(0);++modCount;count = 0; // write-volatile} finally {unlock();postWriteCleanup();}}}// 删除指定entry，first是这个entry所在链表，key/hash/valueReference就是这个entry中的内容，其实这里不传也是ok的。@GuardedBy("Segment.this")@NullableReferenceEntry<K, V> removeValueFromChain(ReferenceEntry<K, V> first,ReferenceEntry<K, V> entry, @Nullable K key, int hash,ValueReference<K, V> valueReference,RemovalCause cause) {// 发布删除时间，RemovalListener会监听这个事件。enqueueNotification(key, hash, valueReference, cause);writeQueue.remove(entry);accessQueue.remove(entry);if (valueReference.isLoading()) {valueReference.notifyNewValue(null);return first;} else {return removeEntryFromChain(first, entry);}}// 从first为头节点的链表中删除指定的entry，返回的是删除entry后的头结点@GuardedBy("Segment.this")@NullableReferenceEntry<K, V> removeEntryFromChain(ReferenceEntry<K, V> first, ReferenceEntry<K, V> entry) {int newCount = count;ReferenceEntry<K, V> newFirst = entry.getNext();for (ReferenceEntry<K, V> e = first; e != entry; e = e.getNext()) {ReferenceEntry<K, V> next = copyEntry(e, newFirst);if (next != null) {newFirst = next;} else {removeCollectedEntry(e);newCount--;}}this.count = newCount;return newFirst;}@GuardedBy("Segment.this")void removeCollectedEntry(ReferenceEntry<K, V> entry) {enqueueNotification(entry, RemovalCause.COLLECTED);writeQueue.remove(entry);accessQueue.remove(entry);}/*** Removes an entry whose key has been garbage collected.*/boolean reclaimKey(ReferenceEntry<K, V> entry, int hash) {lock();try {int newCount = count - 1;// first是entry所在的table的坑位上的链表的头结点。AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);// 遍历链表，找到入参的entry。for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {if (e == entry) {++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, e.getKey(), hash, e.getValueReference(),RemovalCause.COLLECTED);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatilereturn true;}}return false;} finally {unlock();postWriteCleanup();}}/*** Removes an entry whose value has been garbage collected.*/boolean reclaimValue(K key, int hash, ValueReference<K, V> valueReference) {lock();try {int newCount = this.count - 1;AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> v = e.getValueReference();if (v == valueReference) {++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, entryKey, hash, valueReference, RemovalCause.COLLECTED);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatilereturn true;}return false;}}return false;} finally {unlock();if (!isHeldByCurrentThread()) { // don't cleanup inside of putpostWriteCleanup();}}}boolean removeLoadingValue(K key, int hash, LoadingValueReference<K, V> valueReference) {lock();try {AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {K entryKey = e.getKey();if (e.getHash() == hash && entryKey != null&& map.keyEquivalence.equivalent(key, entryKey)) {ValueReference<K, V> v = e.getValueReference();if (v == valueReference) {if (valueReference.isActive()) {e.setValueReference(valueReference.getOldValue());} else {ReferenceEntry<K, V> newFirst = removeEntryFromChain(first, e);table.set(index, newFirst);}return true;}return false;}}return false;} finally {unlock();postWriteCleanup();}}// 删除指定entry，并发布一个删除事件。@GuardedBy("Segment.this")boolean removeEntry(ReferenceEntry<K, V> entry, int hash, RemovalCause cause) {int newCount = this.count - 1;AtomicReferenceArray<ReferenceEntry<K, V>> table = this.table;int index = hash & (table.length() - 1);ReferenceEntry<K, V> first = table.get(index);for (ReferenceEntry<K, V> e = first; e != null; e = e.getNext()) {if (e == entry) {++modCount;ReferenceEntry<K, V> newFirst = removeValueFromChain(first, e, e.getKey(), hash, e.getValueReference(), cause);newCount = this.count - 1;table.set(index, newFirst);this.count = newCount; // write-volatilereturn true;}}return false;}/*** Performs routine cleanup following a read. Normally cleanup happens during writes. If cleanup is not observed after a sufficient* number of reads, try cleaning up from the read thread.*/void postReadCleanup() {// DRAIN_THRESHOLD=63，这个if就是看读取次数有没有到了63。这也是在debug的时候会发现recencyQueue命名是在不断往里add()，但是这个队列的总个数"始终"没有超过63的原因// 这里不是说recencyQueue真的不会超过63，是因为debug的时候没有什么竞争，一到了63个的时候，tryLock()能够获得锁，所以就给清除了，如果有竞争的情况下，这里tryLock()不成功，是完全有可能超过63的if ((readCount.incrementAndGet() & DRAIN_THRESHOLD) == 0) {cleanUp();}}/*** Performs routine cleanup prior to executing a write. This should be called every time a write thread acquires the segment lock,* immediately after acquiring the lock.** <p>Post-condition: expireEntries has been run.*/@GuardedBy("Segment.this")void preWriteCleanup(long now) {runLockedCleanup(now);}/*** Performs routine cleanup following a write.*/void postWriteCleanup() {runUnlockedCleanup();}void cleanUp() {long now = map.ticker.read();runLockedCleanup(now);runUnlockedCleanup();}void runLockedCleanup(long now) {// 这里是tryLock()，而不是lock()，目的就是当有竞争你的时候不会阻塞：拿到所就做下面的事情，拿不到就不做if (tryLock()) {try {// 如果key或者value的引用不是强引用，这个方法就是将那些实际内容被gc回收了的key/value对应的ReferenceEntry从链表中删除。drainReferenceQueues();// 1. 清空recencyQueue，去改变accessQueue// 2.清除根据最近写/最近访问时间，已经过期的缓存项，比如设置了accessAfter(5s)，则一个key最近访问5s后还没有访问，这里就会将其从缓存删除。expireEntries(now); // calls drainRecencyQueue// readCount这个变量的作用就是控制recencyQueue批量的大小的，没有其他的作用，这里已经清空了recencyQueue，这个变量就可以归零了。readCount.set(0);} finally {unlock();}}}void runUnlockedCleanup() {// locked cleanup may generate notifications we can send unlockedif (!isHeldByCurrentThread()) {map.processPendingNotifications();}}}static class LoadingValueReference<K, V> implements ValueReference<K, V> {volatile ValueReference<K, V> oldValue;// TODO(fry): rename get, then extend AbstractFuture instead of containing SettableFuturefinal SettableFuture<V> futureValue = SettableFuture.create();final Stopwatch         stopwatch   = Stopwatch.createUnstarted();public LoadingValueReference() {this(LocalCache.<K, V>unset());}public LoadingValueReference(ValueReference<K, V> oldValue) {this.oldValue = oldValue;}@Overridepublic boolean isLoading() {return true;}@Overridepublic boolean isActive() {return oldValue.isActive();}@Overridepublic int getWeight() {return oldValue.getWeight();}public boolean set(@Nullable V newValue) {return futureValue.set(newValue);}public boolean setException(Throwable t) {return futureValue.setException(t);}private ListenableFuture<V> fullyFailedFuture(Throwable t) {return Futures.immediateFailedFuture(t);}@Overridepublic void notifyNewValue(@Nullable V newValue) {if (newValue != null) {// The pending load was clobbered by a manual write.// Unblock all pending gets, and have them return the new value.set(newValue);} else {// The pending load was removed. Delay notifications until loading completes.oldValue = unset();}// TODO(fry): could also cancel loading if we had a handle on its future}public ListenableFuture<V> loadFuture(K key, CacheLoader<? super K, V> loader) {stopwatch.start();V previousValue = oldValue.get();try {// 如果老的值已经没有了，那其实跟缓存中没有，去加载逻辑是一样的。if (previousValue == null) {V newValue = loader.load(key);return set(newValue) ? futureValue : Futures.immediateFuture(newValue);}// 这里是缓存中的值存在，去重新加载的逻辑。其内部逻辑实际上就是直接去调用了CacheLoader#load()方法，所以这里其实是同步的。但是我们可以重写reload()方法，变成异步的。ListenableFuture<V> newValue = loader.reload(key, previousValue);if (newValue == null) {return Futures.immediateFuture(null);}// To avoid a race, make sure the refreshed value is set into loadingValueReference// *before* returning newValue from the cache query.return Futures.transform(newValue, new Function<V, V>() {@Overridepublic V apply(V newValue) {LoadingValueReference.this.set(newValue);return newValue;}});} catch (Throwable t) {if (t instanceof InterruptedException) {Thread.currentThread().interrupt();}return setException(t) ? futureValue : fullyFailedFuture(t);}}public long elapsedNanos() {return stopwatch.elapsed(NANOSECONDS);}@Overridepublic V waitForValue() throws ExecutionException {return getUninterruptibly(futureValue);}@Overridepublic V get() {return oldValue.get();}public ValueReference<K, V> getOldValue() {return oldValue;}@Overridepublic ReferenceEntry<K, V> getEntry() {return null;}@Overridepublic ValueReference<K, V> copyFor(ReferenceQueue<V> queue, @Nullable V value, ReferenceEntry<K, V> entry) {return this;}}// Queues/*** A custom queue for managing eviction order. Note that this is tightly integrated with {@code ReferenceEntry}, upon which it relies to* perform its linking.** <p>Note that this entire implementation makes the assumption that all elements which are in* the map are also in this queue, and that all elements not in the queue are not in the map.** <p>The benefits of creating our own queue are that (1) we can replace elements in the middle* of the queue as part of copyWriteEntry, and (2) the contains method is highly optimized for the current model.*/static final class WriteQueue<K, V> extends AbstractQueue<ReferenceEntry<K, V>> {final ReferenceEntry<K, V> head = new AbstractReferenceEntry<K, V>() {@Overridepublic long getWriteTime() {return Long.MAX_VALUE;}@Overridepublic void setWriteTime(long time) {}ReferenceEntry<K, V> nextWrite = this;@Overridepublic ReferenceEntry<K, V> getNextInWriteQueue() {return nextWrite;}@Overridepublic void setNextInWriteQueue(ReferenceEntry<K, V> next) {this.nextWrite = next;}ReferenceEntry<K, V> previousWrite = this;@Overridepublic ReferenceEntry<K, V> getPreviousInWriteQueue() {return previousWrite;}@Overridepublic void setPreviousInWriteQueue(ReferenceEntry<K, V> previous) {this.previousWrite = previous;}};// implements Queue@Overridepublic boolean offer(ReferenceEntry<K, V> entry) {// unlinkconnectWriteOrder(entry.getPreviousInWriteQueue(), entry.getNextInWriteQueue());// add to tailconnectWriteOrder(head.getPreviousInWriteQueue(), entry);connectWriteOrder(entry, head);return true;}@Overridepublic ReferenceEntry<K, V> peek() {ReferenceEntry<K, V> next = head.getNextInWriteQueue();return (next == head) ? null : next;}@Overridepublic ReferenceEntry<K, V> poll() {ReferenceEntry<K, V> next = head.getNextInWriteQueue();if (next == head) {return null;}remove(next);return next;}@Override@SuppressWarnings("unchecked")public boolean remove(Object o) {ReferenceEntry<K, V> e = (ReferenceEntry) o;ReferenceEntry<K, V> previous = e.getPreviousInWriteQueue();ReferenceEntry<K, V> next = e.getNextInWriteQueue();connectWriteOrder(previous, next);nullifyWriteOrder(e);return next != NullEntry.INSTANCE;}@Override@SuppressWarnings("unchecked")public boolean contains(Object o) {ReferenceEntry<K, V> e = (ReferenceEntry) o;return e.getNextInWriteQueue() != NullEntry.INSTANCE;}@Overridepublic boolean isEmpty() {return head.getNextInWriteQueue() == head;}@Overridepublic int size() {int size = 0;for (ReferenceEntry<K, V> e = head.getNextInWriteQueue(); e != head;e = e.getNextInWriteQueue()) {size++;}return size;}@Overridepublic void clear() {ReferenceEntry<K, V> e = head.getNextInWriteQueue();while (e != head) {ReferenceEntry<K, V> next = e.getNextInWriteQueue();nullifyWriteOrder(e);e = next;}head.setNextInWriteQueue(head);head.setPreviousInWriteQueue(head);}@Overridepublic Iterator<ReferenceEntry<K, V>> iterator() {return new AbstractSequentialIterator<ReferenceEntry<K, V>>(peek()) {@Overrideprotected ReferenceEntry<K, V> computeNext(ReferenceEntry<K, V> previous) {ReferenceEntry<K, V> next = previous.getNextInWriteQueue();return (next == head) ? null : next;}};}}/*** A custom queue for managing access order. Note that this is tightly integrated with {@code ReferenceEntry}, upon which it reliese to* perform its linking.** <p>Note that this entire implementation makes the assumption that all elements which are in* the map are also in this queue, and that all elements not in the queue are not in the map.** <p>The benefits of creating our own queue are that (1) we can replace elements in the middle* of the queue as part of copyWriteEntry, and (2) the contains method is highly optimized for the current model.*/static final class AccessQueue<K, V> extends AbstractQueue<ReferenceEntry<K, V>> {final ReferenceEntry<K, V> head = new AbstractReferenceEntry<K, V>() {@Overridepublic long getAccessTime() {return Long.MAX_VALUE;}@Overridepublic void setAccessTime(long time) {}ReferenceEntry<K, V> nextAccess = this;@Overridepublic ReferenceEntry<K, V> getNextInAccessQueue() {return nextAccess;}@Overridepublic void setNextInAccessQueue(ReferenceEntry<K, V> next) {this.nextAccess = next;}ReferenceEntry<K, V> previousAccess = this;@Overridepublic ReferenceEntry<K, V> getPreviousInAccessQueue() {return previousAccess;}@Overridepublic void setPreviousInAccessQueue(ReferenceEntry<K, V> previous) {this.previousAccess = previous;}};// implements Queue@Overridepublic boolean offer(ReferenceEntry<K, V> entry) {// unlinkconnectAccessOrder(entry.getPreviousInAccessQueue(), entry.getNextInAccessQueue());// add to tailconnectAccessOrder(head.getPreviousInAccessQueue(), entry);connectAccessOrder(entry, head);return true;}@Overridepublic ReferenceEntry<K, V> peek() {ReferenceEntry<K, V> next = head.getNextInAccessQueue();return (next == head) ? null : next;}@Overridepublic ReferenceEntry<K, V> poll() {ReferenceEntry<K, V> next = head.getNextInAccessQueue();if (next == head) {return null;}remove(next);return next;}@Override@SuppressWarnings("unchecked")public boolean remove(Object o) {ReferenceEntry<K, V> e = (ReferenceEntry) o;ReferenceEntry<K, V> previous = e.getPreviousInAccessQueue();ReferenceEntry<K, V> next = e.getNextInAccessQueue();connectAccessOrder(previous, next);nullifyAccessOrder(e);return next != NullEntry.INSTANCE;}@Override@SuppressWarnings("unchecked")public boolean contains(Object o) {ReferenceEntry<K, V> e = (ReferenceEntry) o;return e.getNextInAccessQueue() != NullEntry.INSTANCE;}@Overridepublic boolean isEmpty() {return head.getNextInAccessQueue() == head;}@Overridepublic int size() {int size = 0;for (ReferenceEntry<K, V> e = head.getNextInAccessQueue(); e != head;e = e.getNextInAccessQueue()) {size++;}return size;}@Overridepublic void clear() {ReferenceEntry<K, V> e = head.getNextInAccessQueue();while (e != head) {ReferenceEntry<K, V> next = e.getNextInAccessQueue();nullifyAccessOrder(e);e = next;}head.setNextInAccessQueue(head);head.setPreviousInAccessQueue(head);}@Overridepublic Iterator<ReferenceEntry<K, V>> iterator() {return new AbstractSequentialIterator<ReferenceEntry<K, V>>(peek()) {@Overrideprotected ReferenceEntry<K, V> computeNext(ReferenceEntry<K, V> previous) {ReferenceEntry<K, V> next = previous.getNextInAccessQueue();return (next == head) ? null : next;}};}}// Cache supportpublic void cleanUp() {for (Segment<?, ?> segment : segments) {segment.cleanUp();}}// ConcurrentMap methods@Overridepublic boolean isEmpty() {/** Sum per-segment modCounts to avoid mis-reporting when elements are concurrently added and* removed in one segment while checking another, in which case the table was never actually* empty at any point. (The sum ensures accuracy up through at least 1<<31 per-segment* modifications before recheck.)  Method containsValue() uses similar constructions for* stability checks.*/long sum = 0L;Segment<K, V>[] segments = this.segments;for (int i = 0; i < segments.length; ++i) {if (segments[i].count != 0) {return false;}sum += segments[i].modCount;}if (sum != 0L) { // recheck unless no modificationsfor (int i = 0; i < segments.length; ++i) {if (segments[i].count != 0) {return false;}sum -= segments[i].modCount;}if (sum != 0L) {return false;}}return true;}long longSize() {Segment<K, V>[] segments = this.segments;long sum = 0;for (int i = 0; i < segments.length; ++i) {sum += segments[i].count;}return sum;}@Overridepublic int size() {return Ints.saturatedCast(longSize());}@Override@Nullablepublic V get(@Nullable Object key) {if (key == null) {return null;}int hash = hash(key);return segmentFor(hash).get(key, hash);}@Nullablepublic V getIfPresent(Object key) {int hash = hash(checkNotNull(key));V value = segmentFor(hash).get(key, hash);if (value == null) {globalStatsCounter.recordMisses(1);} else {globalStatsCounter.recordHits(1);}return value;}V get(K key, CacheLoader<? super K, V> loader) throws ExecutionException {int hash = hash(checkNotNull(key));return segmentFor(hash).get(key, hash, loader);}V getOrLoad(K key) throws ExecutionException {return get(key, defaultLoader);}ImmutableMap<K, V> getAllPresent(Iterable<?> keys) {int hits = 0;int misses = 0;Map<K, V> result = Maps.newLinkedHashMap();for (Object key : keys) {V value = get(key);if (value == null) {misses++;} else {// TODO(fry): store entry key instead of query key@SuppressWarnings("unchecked")K castKey = (K) key;result.put(castKey, value);hits++;}}globalStatsCounter.recordHits(hits);globalStatsCounter.recordMisses(misses);return ImmutableMap.copyOf(result);}ImmutableMap<K, V> getAll(Iterable<? extends K> keys) throws ExecutionException {int hits = 0;int misses = 0;Map<K, V> result = Maps.newLinkedHashMap();Set<K> keysToLoad = Sets.newLinkedHashSet();for (K key : keys) {V value = get(key);if (!result.containsKey(key)) {result.put(key, value);if (value == null) {misses++;keysToLoad.add(key);} else {hits++;}}}try {if (!keysToLoad.isEmpty()) {try {Map<K, V> newEntries = loadAll(keysToLoad, defaultLoader);for (K key : keysToLoad) {V value = newEntries.get(key);if (value == null) {throw new InvalidCacheLoadException("loadAll failed to return a value for " + key);}result.put(key, value);}} catch (UnsupportedLoadingOperationException e) {// loadAll not implemented, fallback to loadfor (K key : keysToLoad) {misses--; // get will count this missresult.put(key, get(key, defaultLoader));}}}return ImmutableMap.copyOf(result);} finally {globalStatsCounter.recordHits(hits);globalStatsCounter.recordMisses(misses);}}/*** Returns the result of calling {@link CacheLoader#loadAll}, or null if {@code loader} doesn't implement {@code loadAll}.*/@NullableMap<K, V> loadAll(Set<? extends K> keys, CacheLoader<? super K, V> loader)throws ExecutionException {checkNotNull(loader);checkNotNull(keys);Stopwatch stopwatch = Stopwatch.createStarted();Map<K, V> result;boolean success = false;try {@SuppressWarnings("unchecked") // safe since all keys extend KMap<K, V> map = (Map<K, V>) loader.loadAll(keys);result = map;success = true;} catch (com.zj.cache.guavacache.CacheLoader.UnsupportedLoadingOperationException e) {success = true;throw e;} catch (InterruptedException e) {Thread.currentThread().interrupt();throw new ExecutionException(e);} catch (RuntimeException e) {throw new UncheckedExecutionException(e);} catch (Exception e) {throw new ExecutionException(e);} catch (Error e) {throw new ExecutionError(e);} finally {if (!success) {globalStatsCounter.recordLoadException(stopwatch.elapsed(NANOSECONDS));}}if (result == null) {globalStatsCounter.recordLoadException(stopwatch.elapsed(NANOSECONDS));throw new InvalidCacheLoadException(loader + " returned null map from loadAll");}stopwatch.stop();// TODO(fry): batch by segmentboolean nullsPresent = false;for (Entry<K, V> entry : result.entrySet()) {K key = entry.getKey();V value = entry.getValue();if (key == null || value == null) {// delay failure until non-null entries are storednullsPresent = true;} else {put(key, value);}}if (nullsPresent) {globalStatsCounter.recordLoadException(stopwatch.elapsed(NANOSECONDS));throw new InvalidCacheLoadException(loader + " returned null keys or values from loadAll");}// TODO(fry): record count of loaded entriesglobalStatsCounter.recordLoadSuccess(stopwatch.elapsed(NANOSECONDS));return result;}/*** Returns the internal entry for the specified key. The entry may be loading, expired, or partially collected.*/ReferenceEntry<K, V> getEntry(@Nullable Object key) {// does not impact recency orderingif (key == null) {return null;}int hash = hash(key);return segmentFor(hash).getEntry(key, hash);}void refresh(K key) {int hash = hash(checkNotNull(key));segmentFor(hash).refresh(key, hash, defaultLoader, false);}@Overridepublic boolean containsKey(@Nullable Object key) {// does not impact recency orderingif (key == null) {return false;}int hash = hash(key);return segmentFor(hash).containsKey(key, hash);}@Overridepublic boolean containsValue(@Nullable Object value) {// does not impact recency orderingif (value == null) {return false;}// This implementation is patterned after ConcurrentHashMap, but without the locking. The only// way for it to return a false negative would be for the target value to jump around in the map// such that none of the subsequent iterations observed it, despite the fact that at every point// in time it was present somewhere int the map. This becomes increasingly unlikely as// CONTAINS_VALUE_RETRIES increases, though without locking it is theoretically possible.long now = ticker.read();final Segment<K, V>[] segments = this.segments;long last = -1L;for (int i = 0; i < CONTAINS_VALUE_RETRIES; i++) {long sum = 0L;for (Segment<K, V> segment : segments) {// ensure visibility of most recent completed write@SuppressWarnings({"UnusedDeclaration", "unused"})int c = segment.count; // read-volatileAtomicReferenceArray<ReferenceEntry<K, V>> table = segment.table;for (int j = 0; j < table.length(); j++) {for (ReferenceEntry<K, V> e = table.get(j); e != null; e = e.getNext()) {V v = segment.getLiveValue(e, now);if (v != null && valueEquivalence.equivalent(value, v)) {return true;}}}sum += segment.modCount;}if (sum == last) {break;}last = sum;}return false;}@Overridepublic V put(K key, V value) {checkNotNull(key);checkNotNull(value);int hash = hash(key);return segmentFor(hash).put(key, hash, value, false);}@Overridepublic V putIfAbsent(K key, V value) {checkNotNull(key);checkNotNull(value);int hash = hash(key);return segmentFor(hash).put(key, hash, value, true);}@Overridepublic void putAll(Map<? extends K, ? extends V> m) {for (Entry<? extends K, ? extends V> e : m.entrySet()) {put(e.getKey(), e.getValue());}}@Overridepublic V remove(@Nullable Object key) {if (key == null) {return null;}int hash = hash(key);return segmentFor(hash).remove(key, hash);}@Overridepublic boolean remove(@Nullable Object key, @Nullable Object value) {if (key == null || value == null) {return false;}int hash = hash(key);return segmentFor(hash).remove(key, hash, value);}@Overridepublic boolean replace(K key, @Nullable V oldValue, V newValue) {checkNotNull(key);checkNotNull(newValue);if (oldValue == null) {return false;}int hash = hash(key);return segmentFor(hash).replace(key, hash, oldValue, newValue);}@Overridepublic V replace(K key, V value) {checkNotNull(key);checkNotNull(value);int hash = hash(key);return segmentFor(hash).replace(key, hash, value);}@Overridepublic void clear() {for (Segment<K, V> segment : segments) {segment.clear();}}void invalidateAll(Iterable<?> keys) {// TODO(fry): batch by segmentfor (Object key : keys) {remove(key);}}Set<K> keySet;@Overridepublic Set<K> keySet() {// does not impact recency orderingSet<K> ks = keySet;return (ks != null) ? ks : (keySet = new KeySet(this));}Collection<V> values;@Overridepublic Collection<V> values() {// does not impact recency orderingCollection<V> vs = values;return (vs != null) ? vs : (values = new Values(this));}Set<Entry<K, V>> entrySet;@Override@GwtIncompatible("Not supported.")public Set<Entry<K, V>> entrySet() {// does not impact recency orderingSet<Entry<K, V>> es = entrySet;return (es != null) ? es : (entrySet = new EntrySet(this));}// Iterator Supportabstract class HashIterator<T> implements Iterator<T> {int                                        nextSegmentIndex;int                                        nextTableIndex;Segment<K, V>                              currentSegment;AtomicReferenceArray<ReferenceEntry<K, V>> currentTable;ReferenceEntry<K, V>                       nextEntry;WriteThroughEntry                          nextExternal;WriteThroughEntry                          lastReturned;HashIterator() {nextSegmentIndex = segments.length - 1;nextTableIndex = -1;advance();}@Overridepublic abstract T next();final void advance() {nextExternal = null;if (nextInChain()) {return;}if (nextInTable()) {return;}while (nextSegmentIndex >= 0) {currentSegment = segments[nextSegmentIndex--];if (currentSegment.count != 0) {currentTable = currentSegment.table;nextTableIndex = currentTable.length() - 1;if (nextInTable()) {return;}}}}/*** Finds the next entry in the current chain. Returns true if an entry was found.*/boolean nextInChain() {if (nextEntry != null) {for (nextEntry = nextEntry.getNext(); nextEntry != null; nextEntry = nextEntry.getNext()) {if (advanceTo(nextEntry)) {return true;}}}return false;}/*** Finds the next entry in the current table. Returns true if an entry was found.*/boolean nextInTable() {while (nextTableIndex >= 0) {if ((nextEntry = currentTable.get(nextTableIndex--)) != null) {if (advanceTo(nextEntry) || nextInChain()) {return true;}}}return false;}/*** Advances to the given entry. Returns true if the entry was valid, false if it should be skipped.*/boolean advanceTo(ReferenceEntry<K, V> entry) {try {long now = ticker.read();K key = entry.getKey();V value = getLiveValue(entry, now);if (value != null) {nextExternal = new WriteThroughEntry(key, value);return true;} else {// Skip stale entry.return false;}} finally {currentSegment.postReadCleanup();}}@Overridepublic boolean hasNext() {return nextExternal != null;}WriteThroughEntry nextEntry() {if (nextExternal == null) {throw new NoSuchElementException();}lastReturned = nextExternal;advance();return lastReturned;}@Overridepublic void remove() {checkState(lastReturned != null);LocalCache.this.remove(lastReturned.getKey());lastReturned = null;}}final class KeyIterator extends HashIterator<K> {@Overridepublic K next() {return nextEntry().getKey();}}final class ValueIterator extends HashIterator<V> {@Overridepublic V next() {return nextEntry().getValue();}}/*** Custom Entry class used by EntryIterator.next(), that relays setValue changes to the underlying map.*/final class WriteThroughEntry implements Entry<K, V> {final K key; // non-nullV value; // non-nullWriteThroughEntry(K key, V value) {this.key = key;this.value = value;}@Overridepublic K getKey() {return key;}@Overridepublic V getValue() {return value;}@Overridepublic boolean equals(@Nullable Object object) {// Cannot use key and value equivalenceif (object instanceof Entry) {Entry<?, ?> that = (Entry<?, ?>) object;return key.equals(that.getKey()) && value.equals(that.getValue());}return false;}@Overridepublic int hashCode() {// Cannot use key and value equivalencereturn key.hashCode() ^ value.hashCode();}@Overridepublic V setValue(V newValue) {throw new UnsupportedOperationException();}/*** Returns a string representation of the form <code>{key}={value}</code>.*/@Overridepublic String toString() {return getKey() + "=" + getValue();}}final class EntryIterator extends HashIterator<Entry<K, V>> {@Overridepublic Entry<K, V> next() {return nextEntry();}}abstract class AbstractCacheSet<T> extends AbstractSet<T> {final ConcurrentMap<?, ?> map;AbstractCacheSet(ConcurrentMap<?, ?> map) {this.map = map;}@Overridepublic int size() {return map.size();}@Overridepublic boolean isEmpty() {return map.isEmpty();}@Overridepublic void clear() {map.clear();}}final class KeySet extends AbstractCacheSet<K> {KeySet(ConcurrentMap<?, ?> map) {super(map);}@Overridepublic Iterator<K> iterator() {return new KeyIterator();}@Overridepublic boolean contains(Object o) {return map.containsKey(o);}@Overridepublic boolean remove(Object o) {return map.remove(o) != null;}}final class Values extends AbstractCacheSet<V> {Values(ConcurrentMap<?, ?> map) {super(map);}@Overridepublic Iterator<V> iterator() {return new ValueIterator();}@Overridepublic boolean contains(Object o) {return map.containsValue(o);}}final class EntrySet extends AbstractCacheSet<Entry<K, V>> {EntrySet(ConcurrentMap<?, ?> map) {super(map);}@Overridepublic Iterator<Entry<K, V>> iterator() {return new EntryIterator();}@Overridepublic boolean contains(Object o) {if (!(o instanceof Entry)) {return false;}Entry<?, ?> e = (Entry<?, ?>) o;Object key = e.getKey();if (key == null) {return false;}V v = LocalCache.this.get(key);return v != null && valueEquivalence.equivalent(e.getValue(), v);}@Overridepublic boolean remove(Object o) {if (!(o instanceof Entry)) {return false;}Entry<?, ?> e = (Entry<?, ?>) o;Object key = e.getKey();return key != null && LocalCache.this.remove(key, e.getValue());}}// Serialization Support/*** Serializes the configuration of a LocalCache, reconsitituting it as a Cache using CacheBuilder upon deserialization. An instance of* this class is fit for use by the writeReplace of LocalManualCache.* <p>* Unfortunately, readResolve() doesn't get called when a circular dependency is present, so the proxy must be able to behave as the* cache itself.*/static class ManualSerializationProxy<K, V>extends ForwardingCache<K, V> implements Serializable {private static final long serialVersionUID = 1;final Strength                              keyStrength;final Strength                              valueStrength;final Equivalence<Object>                   keyEquivalence;final Equivalence<Object>                   valueEquivalence;final long                                  expireAfterWriteNanos;final long                                  expireAfterAccessNanos;final long                                  maxWeight;final Weigher<K, V>                         weigher;final int                                   concurrencyLevel;final RemovalListener<? super K, ? super V> removalListener;final Ticker                                ticker;final CacheLoader<? super K, V>             loader;transient Cache<K, V> delegate;ManualSerializationProxy(LocalCache<K, V> cache) {this(cache.keyStrength,cache.valueStrength,cache.keyEquivalence,cache.valueEquivalence,cache.expireAfterWriteNanos,cache.expireAfterAccessNanos,cache.maxWeight,cache.weigher,cache.concurrencyLevel,cache.removalListener,cache.ticker,cache.defaultLoader);}private ManualSerializationProxy(Strength keyStrength, Strength valueStrength,Equivalence<Object> keyEquivalence, Equivalence<Object> valueEquivalence,long expireAfterWriteNanos, long expireAfterAccessNanos, long maxWeight,Weigher<K, V> weigher, int concurrencyLevel,RemovalListener<? super K, ? super V> removalListener,Ticker ticker, CacheLoader<? super K, V> loader) {this.keyStrength = keyStrength;this.valueStrength = valueStrength;this.keyEquivalence = keyEquivalence;this.valueEquivalence = valueEquivalence;this.expireAfterWriteNanos = expireAfterWriteNanos;this.expireAfterAccessNanos = expireAfterAccessNanos;this.maxWeight = maxWeight;this.weigher = weigher;this.concurrencyLevel = concurrencyLevel;this.removalListener = removalListener;this.ticker = (ticker == Ticker.systemTicker() || ticker == NULL_TICKER)? null : ticker;this.loader = loader;}CacheBuilder<K, V> recreateCacheBuilder() {CacheBuilder<K, V> builder = CacheBuilder.newBuilder().setKeyStrength(keyStrength).setValueStrength(valueStrength).keyEquivalence(keyEquivalence).valueEquivalence(valueEquivalence).concurrencyLevel(concurrencyLevel).removalListener(removalListener);builder.strictParsing = false;if (expireAfterWriteNanos > 0) {builder.expireAfterWrite(expireAfterWriteNanos, TimeUnit.NANOSECONDS);}if (expireAfterAccessNanos > 0) {builder.expireAfterAccess(expireAfterAccessNanos, TimeUnit.NANOSECONDS);}if (weigher != CacheBuilder.OneWeigher.INSTANCE) {builder.weigher(weigher);if (maxWeight != UNSET_INT) {builder.maximumWeight(maxWeight);}} else {if (maxWeight != UNSET_INT) {builder.maximumSize(maxWeight);}}if (ticker != null) {builder.ticker(ticker);}return builder;}private void readObject(ObjectInputStream in) throws IOException, ClassNotFoundException {in.defaultReadObject();com.google.common.cache.CacheBuilder<K, V> builder = recreateCacheBuilder();this.delegate = builder.build();}private Object readResolve() {return delegate;}@Overrideprotected Cache<K, V> delegate() {return delegate;}}/*** Serializes the configuration of a LocalCache, reconsitituting it as an LoadingCache using CacheBuilder upon deserialization. An* instance of this class is fit for use by the writeReplace of LocalLoadingCache.* <p>* Unfortunately, readResolve() doesn't get called when a circular dependency is present, so the proxy must be able to behave as the* cache itself.*/static final class LoadingSerializationProxy<K, V>extends ManualSerializationProxy<K, V> implements LoadingCache<K, V>, Serializable {private static final long serialVersionUID = 1;transient LoadingCache<K, V> autoDelegate;LoadingSerializationProxy(LocalCache<K, V> cache) {super(cache);}private void readObject(ObjectInputStream in) throws IOException, ClassNotFoundException {in.defaultReadObject();com.google.common.cache.CacheBuilder<K, V> builder = recreateCacheBuilder();this.autoDelegate = builder.build(loader);}@Overridepublic V get(K key) throws ExecutionException {return autoDelegate.get(key);}@Overridepublic V getUnchecked(K key) {return autoDelegate.getUnchecked(key);}@Overridepublic ImmutableMap<K, V> getAll(Iterable<? extends K> keys) throws ExecutionException {return autoDelegate.getAll(keys);}@Overridepublic final V apply(K key) {return autoDelegate.apply(key);}@Overridepublic void refresh(K key) {autoDelegate.refresh(key);}private Object readResolve() {return autoDelegate;}}static class LocalManualCache<K, V> implements Cache<K, V>, Serializable {final LocalCache<K, V> localCache;LocalManualCache(com.google.common.cache.CacheBuilder<? super K, ? super V> builder) {this(new LocalCache<K, V>(builder, null));}private LocalManualCache(LocalCache<K, V> localCache) {this.localCache = localCache;}// Cache methods@Override@Nullablepublic V getIfPresent(Object key) {return localCache.getIfPresent(key);}@Overridepublic V get(K key, final Callable<? extends V> valueLoader) throws ExecutionException {checkNotNull(valueLoader);return localCache.get(key, new CacheLoader<Object, V>() {@Overridepublic V load(Object key) throws Exception {return valueLoader.call();}});}@Overridepublic ImmutableMap<K, V> getAllPresent(Iterable<?> keys) {return localCache.getAllPresent(keys);}@Overridepublic void put(K key, V value) {localCache.put(key, value);}@Overridepublic void putAll(Map<? extends K, ? extends V> m) {localCache.putAll(m);}@Overridepublic void invalidate(Object key) {checkNotNull(key);localCache.remove(key);}@Overridepublic void invalidateAll(Iterable<?> keys) {localCache.invalidateAll(keys);}@Overridepublic void invalidateAll() {localCache.clear();}@Overridepublic long size() {return localCache.longSize();}@Overridepublic ConcurrentMap<K, V> asMap() {return localCache;}@Overridepublic CacheStats stats() {SimpleStatsCounter aggregator = new SimpleStatsCounter();aggregator.incrementBy(localCache.globalStatsCounter);for (Segment<K, V> segment : localCache.segments) {aggregator.incrementBy(segment.statsCounter);}return aggregator.snapshot();}@Overridepublic void cleanUp() {localCache.cleanUp();}// Serialization Supportprivate static final long serialVersionUID = 1;Object writeReplace() {return new ManualSerializationProxy<K, V>(localCache);}}static class LocalLoadingCache<K, V>extends LocalManualCache<K, V> implements LoadingCache<K, V> {LocalLoadingCache(com.google.common.cache.CacheBuilder<? super K, ? super V> builder,CacheLoader<? super K, V> loader) {super(new LocalCache<K, V>(builder, checkNotNull(loader)));}// LoadingCache methods@Overridepublic V get(K key) throws ExecutionException {return localCache.getOrLoad(key);}@Overridepublic V getUnchecked(K key) {try {return get(key);} catch (ExecutionException e) {throw new UncheckedExecutionException(e.getCause());}}@Overridepublic ImmutableMap<K, V> getAll(Iterable<? extends K> keys) throws ExecutionException {return localCache.getAll(keys);}@Overridepublic void refresh(K key) {localCache.refresh(key);}@Overridepublic final V apply(K key) {return getUnchecked(key);}// Serialization Supportprivate static final long serialVersionUID = 1;@OverrideObject writeReplace() {return new LoadingSerializationProxy<K, V>(localCache);}}
}

guava之guava cache相关推荐

1. [Google Guava] 排序: Guava强大的”流畅风格比较器”

   原文链接 译者: 沈义扬 排序器[Ordering]是Guava流畅风格比较器[Comparator]的实现,它可以用来为构建复杂的比较器,以完成集合排序的功能. 从实现上说,Ordering实例就是 ...

2. Guava 系列 - Guava基础

   文章目录 Guava 系列 - Guava基础 1.是什么 2.如何构建 3.用户文档 4.为什么选择使用guava Guava 系列 - Guava基础 1.是什么 guava 是一个开源的java ...

3. java guava eventbus_JAVA | Guava EventBus 使用

   JAVA | Guava EventBus 使用 系列文章目录 Table of Contents [TOC] 前言 EventBus 是 Guava 的事件处理机制,是观察者模式(生产/消费模型)的 ...

4. 【Guava】Guava Cache的refresh和expire刷新机制

   1.概述 转载:https://www.cnblogs.com/liuxiaochong/p/13613071.html 总览参考:[Guava]Google Guava本地高效缓存 案例参考:[gu ...

5. Guava (一)Guava Cache进阶之同步/异步load

   使用guava cache的时候,在cache中没有值或者值需要更新的时候,都需要去load,而这个load往往对应从数据库或者远程接口拿数据并缓存下来的操作.在高qps场景的服务中,这个load可能 ...

6. Guava系列之Cache

   缓存是日常开发中使用很频繁的一种提升性能的方式,它其实解决的是硬件层面性能不对等的问题,比如CPU.内存.硬盘之间性能的巨大差异,会严重影响数据的读取与传输,而缓存就是用来平衡这种性能差异的手段 缓存 ...

7. 【guava】guava 11.0.2 版本 key 肯能丢数据的bug

   文章目录 1.概述 1.1 11版本丢数据 1.1 20版本正常 3.小结 1.概述 在开发的时候,我把数据放到guava中,但是却发现数据没有存储进去,开始怀疑是Key的hash相同了,后来调试发现 ...

8. 学习Guava Cache知识汇总

   (一)MapMaker 在软件开发中,Cache缓存技术一直是非常重要的主题,不管我们正在进行任何简单的编程工作,我们总能在编程过程中找到一些缓存机制,即使是你使用一个 Map获取静态的值,它也是缓存 ...

9. 高并发之--Guava Cache

   最近需要用到缓存来存放临时数据,又不想采用Redis,Java自带的Map功能太少,发现google的Guava提供的Cache模块功能很强大,于是选择使用它. 本地缓存 本地缓存作用就是提高系统的运 ...

最新文章

1. 编译ONNX模型Compile ONNX Models
2. 腾讯广告广点通API接入文档(Android)
3. C语言与C++不得不说的那点事
4. 想学python从哪里入手-想要学习python，如何入手学习？
5. Linux下搭建jenkins+svn+http+maven自动化部署
6. 深度案例 | 微车：数据驱动价值，建立 1.3 亿车主的汽车生活平台
7. django前端模板循环多个list
8. B2C网站转化率优化
9. Java保存class文件,[转载]Class文件在JVM中如何存储
10. 在eclipse中运行 carrot2 workbench
11. Python的张量运算
12. Unity3d 2019室内光照贴图lightmap快速烘焙（转载）
13. mac上php的集成环境
14. R语言TCGA数据下载及处理biolinks包的学习与使用（一）数据下载
15. Oracle 绑定变量详解
16. 【内网学习笔记】25、Exchange 邮件服务器
17. 5种方案实现订单30分钟未支付，则自动取消
18. avplayer视频重复循环播放，视频重播
19. Muli3D 8 计算Shader中顶点属性相对于屏幕坐标的偏导数
20. android开关机日志_Android使用代码实现关机/重启

热门文章

1. 〖产品思维训练白宝书 - 核心竞争力篇⑦〗- 产品经理核心竞争力解读之如何提升执行力
2. 猴痘与非猴痘(水痘、麻疹)的二元分类数据集
3. html中加定时器,怎么用HTML设置一个定时器
4. Xamarin Forms Api请求开源框架Refit
5. 怎样下载百度文库文章
6. iOS头像选择控件 — 云头像 AvatarSDK
7. Java后台数据自动备份的实现
8. 【代码质量】C/C++静态检测/静态分析|TscanCode|cppcheck
9. excel宏教程_利用Excel和Power BI,打造自己的BI数据模型
10. 天鸟技术：项目结构-架构设计