ThreadLocal源碼深入分析
ThreadLocal ：線程本地存儲(chǔ)區(qū)（Thread Local Storage耀怜，簡(jiǎn)稱為T(mén)LS）袜茧，每個(gè)線程都有自己的私有的本地存儲(chǔ)區(qū)域爪瓜，不同線程之間彼此不能訪問(wèn)對(duì)方的TLS區(qū)域。它是一個(gè)線程內(nèi)部的數(shù)據(jù)類梯捕，通過(guò)它可以在指定線程中存儲(chǔ)數(shù)據(jù)厢呵，數(shù)據(jù)存儲(chǔ)以后，只有在指定線程中可以獲取到存儲(chǔ)的數(shù)據(jù)傀顾，對(duì)于其他線程來(lái)說(shuō)則無(wú)法獲取到數(shù)據(jù)襟铭。一般來(lái)說(shuō)，當(dāng)某些數(shù)據(jù)是以線程為作用域并且不同的線程具有不同的數(shù)據(jù)副本的時(shí)候，就可以采用ThreadLocal寒砖。比如對(duì)于Handler來(lái)說(shuō)赐劣，他需要獲取當(dāng)前線程的Looper，很顯然Looper的作用域就是線程入撒，并且不同線程具有不同的Looper隆豹，這個(gè)時(shí)候通過(guò)ThreadLocal就可以輕松實(shí)現(xiàn)Looper在線程中的獲取。下面通過(guò)實(shí)際的例子來(lái)演示ThreadLocal的真正含義茅逮。
1. 實(shí)例
public class TestThreadLocal {
private static final ThreadLocal mThreadLocal = new ThreadLocal();
public static void main(String[] args) {
mThreadLocal.set(true);
System.out.println("ThreadMain mThreadLocal = "+mThreadLocal.get());
new Thread("Thread1"){
@Override
public void run() {
mThreadLocal.set(false);
System.out.println("Thread1 mThreadLocal = "+mThreadLocal.get());
}
}.start();
new Thread("Thread2"){
@Override
public void run() {
System.out.println("Thread2 mThreadLocal = "+mThreadLocal.get());
}
}.start();
}
}
在上面的代碼中，在主線程中設(shè)置 mThreadLocal 的值為 true判哥，在子線程1中設(shè)置 mThreadLocal 的值為 false献雅，然后在子線程2中不設(shè)置 mThreadLocal 的值。然后分別在3個(gè)子線程中通過(guò) get 方法獲取 mThreadLocal 的值塌计，根據(jù)前面對(duì) ThreadLocal 的描述挺身，這個(gè)時(shí)候，主線程應(yīng)該是 true锌仅，子線程1中應(yīng)該是 false章钾，而子線程 2 中由于沒(méi)有設(shè)置值，所以應(yīng)該是 null热芹。
運(yùn)行結(jié)果為：
ThreadMain mThreadLocal = true
Thread1 mThreadLocal = false
Thread2 mThreadLocal = null
從結(jié)果可以看出贱傀， 雖然在不同線程中訪問(wèn)的是同一個(gè) ThreadLocal 對(duì)象，但是它們通過(guò) ThreadLocal 獲取到的值卻不同伊脓。下面從源碼的角度來(lái)分析它的工作原理府寒。
2. 類圖
假如讓我們來(lái)實(shí)現(xiàn)一個(gè)變量與線程相綁定的功能，我們可以很容易地想到用HashMap來(lái)實(shí)現(xiàn)报腔，Thread作為key株搔，變量作為value。事實(shí)上纯蛾，JDK中確實(shí)使用了類似Map的結(jié)構(gòu)存儲(chǔ)變量纤房，但不是像我們想的那樣。下面我們來(lái)探究OpenJDK 1.8中ThreadLocal的實(shí)現(xiàn)翻诉。
3. ThreadLocal的成員變量
我們從 ThreadLocal 的幾個(gè)成員變量入手：
/**
* ThreadLocals rely on per-thread linear-probe hash maps attached
* to each thread (Thread.threadLocals and
* inheritableThreadLocals). The ThreadLocal objects act as keys,
* searched via threadLocalHashCode. This is a custom hash code
* (useful only within ThreadLocalMaps) that eliminates collisions
* in the common case where consecutively constructed ThreadLocals
* are used by the same threads, while remaining well-behaved in
* less common cases.
*/
private final int threadLocalHashCode = nextHashCode();
/**
* The next hash code to be given out. Updated atomically. Starts at
* zero.
*/
private static AtomicInteger nextHashCode =
new AtomicInteger();
/**
* The difference between successively generated hash codes - turns
* implicit sequential thread-local IDs into near-optimally spread
* multiplicative hash values for power-of-two-sized tables.
*/
private static final int HASH_INCREMENT = 0x61c88647;
/**
* Returns the next hash code.
*/
private static int nextHashCode() {
return nextHashCode.getAndAdd(HASH_INCREMENT);
}
ThreadLocal通過(guò)threadLocalHashCode 來(lái)標(biāo)識(shí)每一個(gè)ThreadLocal 的唯一性炮姨。threadLocalHashCode 通過(guò) CAS操作 進(jìn)行更新，每次 hash 操作的增量為 0x61c88647 米丘。
4. ThreadLocal的重要方法
4.1 ThreadLocal#set方法
/**
* Sets the current thread's copy of this thread-local variable
* to the specified value. Most subclasses will have no need to
* override this method, relying solely on the {@link #initialValue}
* method to set the values of thread-locals.
*
* @param value the value to be stored in the current thread's copy of
* this thread-local.
*/
public void set(T value) {
Thread t = Thread.currentThread();
ThreadLocalMap map = getMap(t);
if (map != null)
map.set(this, value);
else
createMap(t, value);
}
可以看到通過(guò)Thread.currentThread()方法獲取了當(dāng)前的線程引用剑令，并傳給了getMap(Thread)方法獲取一個(gè)ThreadLocalMap的實(shí)例。我們繼續(xù)跟進(jìn)getMap(Thread)方法：
ThreadLocalMap getMap(Thread t) {
return t.threadLocals;
}
可以看到getMap(Thread)方法直接返回Thread實(shí)例的成員變量threadLocals拄查。它的定義在Thread內(nèi)部吁津，訪問(wèn)級(jí)別為package級(jí)別：
/* ThreadLocal values pertaining to this thread. This map is maintained
* by the ThreadLocal class. */
ThreadLocal.ThreadLocalMap threadLocals = null;
到了這里，我們可以看出，每個(gè)Thread里面都有一個(gè)ThreadLocal.ThreadLocalMap成員變量碍脏，也就是說(shuō)每個(gè)線程通過(guò)ThreadLocal.ThreadLocalMap與ThreadLocal相綁定梭依，這樣可以確保每個(gè)線程訪問(wèn)到的thread-local variable都是本線程的。
我們往下繼續(xù)分析典尾。獲取了ThreadLocalMap實(shí)例以后役拴，如果它不為空則調(diào)用ThreadLocalMap.ThreadLocalMap#set方法設(shè)值；若為空則調(diào)用ThreadLocal#createMap方法new一個(gè)ThreadLocalMap實(shí)例并賦給Thread.threadLocals钾埂。
ThreadLocal#createMap方法的源碼如下：
void createMap(Thread t, T firstValue) {
t.threadLocals = new ThreadLocalMap(this, firstValue);
}
4.2 ThreadLocal#get方法
ThreadLocal的get方法就是調(diào)用了ThreadLocalMap的getEntry方法：
public T get() {
Thread t = Thread.currentThread();
ThreadLocalMap map = getMap(t);
if (map != null) {
ThreadLocalMap.Entry e = map.getEntry(this);
if (e != null) {
@SuppressWarnings("unchecked")
T result = (T)e.value;
return result;
}
}
return setInitialValue();
}
可以看到 ThreadLocal 的 get 和 set 方法調(diào)用了 ThreadLocalMap 中的 set 和 getEntry 方法河闰。下面我們探究一下ThreadLocalMap的實(shí)現(xiàn)
5. ThreadLocalMap
5.1 ThreadLocal的成員變量
在上面的類圖中，我們可以看到 ThreadLocalMap 有一個(gè)常量和三個(gè)成員變量：
/**
* The initial capacity -- MUST be a power of two.
*/
private static final int INITIAL_CAPACITY = 16;
/**
* The table, resized as necessary.
* table.length MUST always be a power of two.
*/
private Entry[] table;
/**
* The number of entries in the table.
*/
private int size = 0;
/**
* The next size value at which to resize.
*/
private int threshold; // Default to 0
其中INITIAL_CAPACITY代表這個(gè)Map的初始容量褥紫；1是一個(gè)Entry類型的數(shù)組姜性，用于存儲(chǔ)數(shù)據(jù)；size代表表中的存儲(chǔ)數(shù)目髓考；threshold代表需要擴(kuò)容時(shí)對(duì)應(yīng)size的閾值部念。
Entry類是ThreadLocalMap的靜態(tài)內(nèi)部類，用于存儲(chǔ)數(shù)據(jù)氨菇。它的源碼如下：
/**
* The entries in this hash map extend WeakReference, using
* its main ref field as the key (which is always a
* ThreadLocal object). Note that null keys (i.e. entry.get()
* == null) mean that the key is no longer referenced, so the
* entry can be expunged from table. Such entries are referred to
* as "stale entries" in the code that follows.
*/
static class Entry extends WeakReference<ThreadLocal<?>> {
/** The value associated with this ThreadLocal. */
Object value;
Entry(ThreadLocal<?> k, Object v) {
super(k);
value = v;
}
}
Entry類繼承了WeakReference<ThreadLocal<?>>儡炼，即每個(gè)Entry對(duì)象都有一個(gè)ThreadLocal的弱引用（作為key）乌询，這是為了防止內(nèi)存泄露。一旦線程結(jié)束楣责，key變?yōu)橐粋€(gè)不可達(dá)的對(duì)象，這個(gè)Entry就可以被GC了秆麸。
5.2 ThreadLocalMap的構(gòu)造函數(shù)
ThreadLocalMap類有兩個(gè)構(gòu)造函數(shù)及汉，其中常用的是ThreadLocalMap(ThreadLocal<?> firstKey, Object firstValue)：
/**
* Construct a new map initially containing (firstKey, firstValue).
* ThreadLocalMaps are constructed lazily, so we only create
* one when we have at least one entry to put in it.
*/
ThreadLocalMap(ThreadLocal<?> firstKey, Object firstValue) {
table = new Entry[INITIAL_CAPACITY];
int i = firstKey.threadLocalHashCode & (INITIAL_CAPACITY - 1);
table[i] = new Entry(firstKey, firstValue);
size = 1;
setThreshold(INITIAL_CAPACITY);
}
構(gòu)造函數(shù)的第一個(gè)參數(shù)就是本ThreadLocal實(shí)例(this)，第二個(gè)參數(shù)就是要保存的線程本地變量房铭。構(gòu)造函數(shù)首先創(chuàng)建一個(gè)長(zhǎng)度為16的Entry數(shù)組，然后計(jì)算出firstKey對(duì)應(yīng)的哈希值缸匪，然后存儲(chǔ)到table中类溢，并設(shè)置size和threshold露懒。
注意一個(gè)細(xì)節(jié)，計(jì)算hash的時(shí)候里面采用了hashCode & (size - 1)的算法砂心，這相當(dāng)于取模運(yùn)算hashCode % size的一個(gè)更高效的實(shí)現(xiàn)（和HashMap中的思路相同）懈词。正是因?yàn)檫@種算法，我們要求size必須是2的指數(shù)辩诞，因?yàn)檫@可以使得hash發(fā)生沖突的次數(shù)減小坎弯。
5.3 ThreadLocalMap#set方法
接下來(lái)我們來(lái)看ThreadLocalMap#set方法的實(shí)現(xiàn)：
/**
* Set the value associated with key.
*
* @param key the thread local object
* @param value the value to be set
*/
private void set(ThreadLocal<?> key, Object value) {
// We don't use a fast path as with get() because it is at
// least as common to use set() to create new entries as
// it is to replace existing ones, in which case, a fast
// path would fail more often than not.
Entry[] tab = table;
int len = tab.length;
int i = key.threadLocalHashCode & (len-1);
//線性探測(cè)
for (Entry e = tab[i];
e != null;
e = tab[i = nextIndex(i, len)]) {
ThreadLocal<?> k = e.get();
//找到對(duì)應(yīng)的key
if (k == key) {
e.value = value;
return;
}
// 替換失效的key
if (k == null) {
//如果entry里對(duì)應(yīng)的key為null的話，表明此entry為staled entry (staled 舊的)译暂，就將其替換為 //當(dāng)前的key和value抠忘。詳情可見(jiàn)replaceStaleEntry 方法。
replaceStaleEntry(key, value, i);
return;
}
}
//若是經(jīng)歷了上面步驟沒(méi)有命中hash秧秉，也沒(méi)有發(fā)現(xiàn)無(wú)用的Entry褐桌，set方法就會(huì)創(chuàng)建一個(gè)新的Entry，并會(huì)進(jìn)行啟發(fā) //式的垃圾清理象迎，用于清理無(wú)用的Entry。如果size大于閾值呛踊，還會(huì)進(jìn)行rehash()算法砾淌。
tab[i] = new Entry(key, value);
int sz = ++size;
if (!cleanSomeSlots(i, sz) && sz >= threshold)
rehash();
}
如果沖突了，就會(huì)通過(guò)nextIndex方法再次計(jì)算哈希值：
/**
* Increment i modulo len.
*/
private static int nextIndex(int i, int len) {
return ((i + 1 < len) ? i + 1 : 0);
}
到這里谭网，我們看到ThreadLocalMap解決沖突的方法是線性探測(cè)法（不斷加1）汪厨，而不是HashMap的鏈地址法，這一點(diǎn)也能從Entry的結(jié)構(gòu)上推斷出來(lái)愉择。
5.4 ThreadLocalMap#replaceStaleEntry方法
/**
* Replace a stale entry encountered during a set operation
* with an entry for the specified key. The value passed in
* the value parameter is stored in the entry, whether or not
* an entry already exists for the specified key.
*
* As a side effect, this method expunges all stale entries in the
* "run" containing the stale entry. (A run is a sequence of entries
* between two null slots.)
*
* @param key the key
* @param value the value to be associated with key
* @param staleSlot index of the first stale entry encountered while
* searching for key.
*/
private void replaceStaleEntry(ThreadLocal<?> key, Object value, int staleSlot) {
Entry[] tab = table;
int len = tab.length;
Entry e;
// Back up to check for prior stale entry in current run.
// We clean out whole runs at a time to avoid continual
// incremental rehashing due to garbage collector freeing
// up refs in bunches (i.e., whenever the collector runs).
//向前掃描劫乱，查找最前的一個(gè)無(wú)效slot
int slotToExpunge = staleSlot;
for (int i = prevIndex(staleSlot, len);
(e = tab[i]) != null;
i = prevIndex(i, len))
if (e.get() == null)
slotToExpunge = i;
// Find either the key or trailing null slot of run, whichever
// occurs first
//向后遍歷table
for (int i = nextIndex(staleSlot, len);
(e = tab[i]) != null;
i = nextIndex(i, len)) {
ThreadLocal<?> k = e.get();
// If we find key, then we need to swap it
// with the stale entry to maintain hash table order.
// The newly stale slot, or any other stale slot
// encountered above it, can then be sent to expungeStaleEntry
// to remove or rehash all of the other entries in run.
//// 找到了key，將其與無(wú)效的slot交換
if (k == key) {
e.value = value;
tab[i] = tab[staleSlot];
tab[staleSlot] = e;
// Start expunge at preceding stale entry if it exists
if (slotToExpunge == staleSlot)
slotToExpunge = i;
cleanSomeSlots(expungeStaleEntry(slotToExpunge), len);
return;
}
// If we didn't find stale entry on backward scan, the
// first stale entry seen while scanning for key is the
// first still present in the run.
//// 如果當(dāng)前的slot已經(jīng)無(wú)效锥涕，并且向前掃描過(guò)程中沒(méi)有無(wú)效slot衷戈，則更新slotToExpunge為當(dāng)前位置
if (k == null && slotToExpunge == staleSlot)
slotToExpunge = i;
}
// If key not found, put new entry in stale slot
tab[staleSlot].value = null;
tab[staleSlot] = new Entry(key, value);
// If there are any other stale entries in run, expunge them
// 在探測(cè)過(guò)程中如果發(fā)現(xiàn)任何無(wú)效slot，則做一次清理（連續(xù)段清理+啟發(fā)式清理）
if (slotToExpunge != staleSlot)
cleanSomeSlots(expungeStaleEntry(slotToExpunge), len);
}
替換過(guò)程里面也進(jìn)行了不少的垃圾清理動(dòng)作以防止引用關(guān)系存在而導(dǎo)致的內(nèi)存泄露层坠。主要是cleanSomeSlots 。
5.5 ThreadLocalMap的清理算法#cleanSomeSlots和#expungeStaleEntry算法
/**
* 啟發(fā)式地清理slot,
* i對(duì)應(yīng)entry是非無(wú)效（指向的ThreadLocal沒(méi)被回收谦趣，或者entry本身為空）
* n是用于控制控制掃描次數(shù)的
* 正常情況下如果log n次掃描沒(méi)有發(fā)現(xiàn)無(wú)效slot前鹅，函數(shù)就結(jié)束了
* 但是如果發(fā)現(xiàn)了無(wú)效的slot峭梳，將n置為table的長(zhǎng)度len，做一次連續(xù)段的清理
* 再?gòu)南乱粋€(gè)空的slot開(kāi)始繼續(xù)掃描
*
* 這個(gè)函數(shù)有兩處地方會(huì)被調(diào)用叉橱，一處是插入的時(shí)候可能會(huì)被調(diào)用窃祝，另外個(gè)是在替換無(wú)效slot的時(shí)候可能會(huì)被調(diào)用粪小，
* 區(qū)別是前者傳入的n為元素個(gè)數(shù)抡句，后者為table的容量
*/
private boolean cleanSomeSlots(int i, int n) {
boolean removed = false;
Entry[] tab = table;
int len = tab.length;
do {
// i在任何情況下自己都不會(huì)是一個(gè)無(wú)效slot待榔，所以從下一個(gè)開(kāi)始判斷
i = nextIndex(i, len);
Entry e = tab[i];
if (e != null && e.get() == null) {
// 擴(kuò)大掃描控制因子
n = len;
removed = true;
// 清理一個(gè)連續(xù)段
i = expungeStaleEntry(i);
}
} while ( (n >>>= 1) != 0);
return removed;
}
/**
* 這個(gè)函數(shù)是ThreadLocal中核心清理函數(shù)腌闯，它做的事情很簡(jiǎn)單：
* 就是從staleSlot開(kāi)始遍歷雕憔，將無(wú)效（弱引用指向?qū)ο蟊换厥眨┣謇斫锉耍磳?duì)應(yīng)entry中的value置為null琉苇，將指向這 entry的table[i]置為null，直到掃到空entry趁冈。
* 另外渗勘，在過(guò)程中還會(huì)對(duì)非空的entry作rehash旺坠。
* 可以說(shuō)這個(gè)函數(shù)的作用就是從staleSlot開(kāi)始清理連續(xù)段中的slot（斷開(kāi)強(qiáng)引用取刃，rehash slot等）
*/
private int expungeStaleEntry(int staleSlot) {
Entry[] tab = table;
int len = tab.length;
// 因?yàn)閑ntry對(duì)應(yīng)的ThreadLocal已經(jīng)被回收，value設(shè)為null坯辩，顯式斷開(kāi)強(qiáng)引用
tab[staleSlot].value = null;
// 顯式設(shè)置該entry為null漆魔，以便垃圾回收
tab[staleSlot] = null;
size--;
Entry e;
int i;
for (i = nextIndex(staleSlot, len); (e = tab[i]) != null; i = nextIndex(i, len)) {
ThreadLocal<?> k = e.get();
// 清理對(duì)應(yīng)ThreadLocal已經(jīng)被回收的entry
if (k == null) {
e.value = null;
tab[i] = null;
size--;
} else {
/*
* 對(duì)于還沒(méi)有被回收的情況改抡，需要做一次rehash阿纤。
*
* 如果對(duì)應(yīng)的ThreadLocal的ID對(duì)len取模出來(lái)的索引h不為當(dāng)前位置i欠拾，
* 則從h向后線性探測(cè)到第一個(gè)空的slot清蚀，把當(dāng)前的entry給挪過(guò)去。
*/
int h = k.threadLocalHashCode & (len - 1);
if (h != i) {
tab[i] = null;
/*
* 在原代碼的這里有句注釋值得一提榛搔，原注釋如下：
*
* Unlike Knuth 6.4 Algorithm R, we must scan until
* null because multiple entries could have been stale.
*
* 這段話提及了Knuth高德納的著作TAOCP（《計(jì)算機(jī)程序設(shè)計(jì)藝術(shù)》）的6.4章節(jié)（散列）
* 中的R算法腹泌。R算法描述了如何從使用線性探測(cè)的散列表中刪除一個(gè)元素尔觉。
* R算法維護(hù)了一個(gè)上次刪除元素的index侦铜，當(dāng)在非空連續(xù)段中掃到某個(gè)entry的哈希值取模后的索 引
* 還沒(méi)有遍歷到時(shí)钉稍，會(huì)將該entry挪到index那個(gè)位置贡未，并更新當(dāng)前位置為新的index，
* 繼續(xù)向后掃描直到遇到空的entry嫩挤。
*
* ThreadLocalMap因?yàn)槭褂昧巳跻闷裾眩云鋵?shí)每個(gè)slot的狀態(tài)有三種也即
* 有效（value未回收），無(wú)效（value已回收）叼风，空（entry==null）无宿。
* 正是因?yàn)門(mén)hreadLocalMap的entry有三種狀態(tài)孽鸡，所以不能完全套高德納原書(shū)的R算法彬碱。
*
* 因?yàn)閑xpungeStaleEntry函數(shù)在掃描過(guò)程中還會(huì)對(duì)無(wú)效slot清理將之轉(zhuǎn)為空slot奥洼，
* 如果直接套用R算法灵奖，可能會(huì)出現(xiàn)具有相同哈希值的entry之間斷開(kāi)（中間有空entry）瓷患。
*/
while (tab[h] != null) {
h = nextIndex(h, len);
}
tab[h] = e;
}
}
}
// 返回staleSlot之后第一個(gè)空的slot索引
return i;
}
5.6 ThreadLocalMap#rehash()方法
/**
* Re-pack and/or re-size the table. First scan the entire
* table removing stale entries. If this doesn't sufficiently
* shrink the size of the table, double the table size.
*/
private void rehash() {
// 做一次全量清理
expungeStaleEntries();
// Use lower threshold for doubling to avoid hysteresis
/*
* 因?yàn)樽隽艘淮吻謇砩帽啵詓ize很可能會(huì)變小爱态。
* ThreadLocalMap這里的實(shí)現(xiàn)是調(diào)低閾值來(lái)判斷是否需要擴(kuò)容肢藐，
* threshold默認(rèn)為len*2/3吆豹，所以這里的threshold - threshold / 4相當(dāng)于len/2
*/
if (size >= threshold - threshold / 4)
resize();
}
/**
* Double the capacity of the table.
*/
/**
* 擴(kuò)容理盆，因?yàn)樾枰ＷCtable的容量len為2的冪猿规，所以擴(kuò)容即擴(kuò)大2倍
*/
private void resize() {
Entry[] oldTab = table;
int oldLen = oldTab.length;
int newLen = oldLen * 2;
Entry[] newTab = new Entry[newLen];
int count = 0;
for (int j = 0; j < oldLen; ++j) {
Entry e = oldTab[j];
if (e != null) {
ThreadLocal<?> k = e.get();
if (k == null) {
e.value = null; // Help the GC
} else {
int h = k.threadLocalHashCode & (newLen - 1);
while (newTab[h] != null)
h = nextIndex(h, newLen);
newTab[h] = e;
count++;
}
}
}
setThreshold(newLen);
size = count;
table = newTab;
}
/**
* Expunge all stale entries in the table.
*/
/*
* 做一次全量清理
*/
private void expungeStaleEntries() {
Entry[] tab = table;
int len = tab.length;
for (int j = 0; j < len; j++) {
Entry e = tab[j];
if (e != null && e.get() == null)
expungeStaleEntry(j);
}
}
我們來(lái)回顧一下 ThreadLocal的 set 方法可能會(huì)有的情況
- 探測(cè)過(guò)程中slot都不無(wú)效，并且順利找到key所在的slot环葵，直接替換即可
- 探測(cè)過(guò)程中發(fā)現(xiàn)有無(wú)效slot张遭，調(diào)用replaceStaleEntry地梨，效果是最終一定會(huì)把key和value放在這個(gè)slot宝剖，并且會(huì)盡可能清理無(wú)效slot
- 在replaceStaleEntry過(guò)程中万细，如果找到了key，則做一個(gè)swap把它放到那個(gè)無(wú)效slot中襟雷，value置為新值
- 在replaceStaleEntry過(guò)程中，沒(méi)有找到key卓缰，直接在無(wú)效slot原地放entry
- 探測(cè)沒(méi)有發(fā)現(xiàn)key征唬，則在連續(xù)段末尾的后一個(gè)空位置放上entry总寒，這也是線性探測(cè)法的一部分理肺。放完后，做一次啟發(fā)式清理炫欺，如果沒(méi)清理出去key熏兄，并且當(dāng)前table大小已經(jīng)超過(guò)閾值了摩桶，則做一次rehash硝清，rehash函數(shù)會(huì)調(diào)用一次全量清理slot方法也即expungeStaleEntries耍缴，如果完了之后table大小超過(guò)了threshold - threshold / 4袋马，則進(jìn)行擴(kuò)容2倍
5.7 ThreadLocalMap#getEntry方法
/**
* Get the entry associated with key. This method
* itself handles only the fast path: a direct hit of existing
* key. It otherwise relays to getEntryAfterMiss. This is
* designed to maximize performance for direct hits, in part
* by making this method readily inlinable.
*
* @param key the thread local object
* @return the entry associated with key, or null if no such
*/
private Entry getEntry(ThreadLocal<?> key) {
int i = key.threadLocalHashCode & (table.length - 1);
Entry e = table[i];
if (e != null && e.get() == key)
return e;
else
return getEntryAfterMiss(key, i, e);
}
/**
* Version of getEntry method for use when key is not found in
* its direct hash slot.
*
* @param key the thread local object
* @param i the table index for key's hash code
* @param e the entry at table[i]
* @return the entry associated with key, or null if no such
*/
private Entry getEntryAfterMiss(ThreadLocal<?> key, int i, Entry e) {
Entry[] tab = table;
int len = tab.length;
while (e != null) {
ThreadLocal<?> k = e.get();
if (k == key)
return e;
if (k == null)
expungeStaleEntry(i);
else
i = nextIndex(i, len);
e = tab[i];
}
return null;
}
從ThreadLocal讀一個(gè)值可能遇到的情況：
根據(jù)入?yún)hreadLocal的threadLocalHashCode對(duì)表容量取模得到index
- 如果index對(duì)應(yīng)的slot就是要讀的threadLocal，則直接返回結(jié)果
- 調(diào)用getEntryAfterMiss線性探測(cè)泄鹏，過(guò)程中每碰到無(wú)效slot番官，調(diào)用expungeStaleEntry進(jìn)行段清理徘熔；如果找到了key酷师，則返回結(jié)果entry
- 沒(méi)有找到key山孔，返回null荷憋。
5.8 ThreadLocalMap#remove方法
/**
* 從map中刪除ThreadLocal
*/
private void remove(ThreadLocal<?> key) {
Entry[] tab = table;
int len = tab.length;
int i = key.threadLocalHashCode & (len - 1);
for (Entry e = tab[i];
e != null;
e = tab[i = nextIndex(i, len)]) {
if (e.get() == key) {
// 顯式斷開(kāi)弱引用
e.clear();
// 進(jìn)行段清理
expungeStaleEntry(i);
return;
}
}
}
remove方法相對(duì)于getEntry和set方法比較簡(jiǎn)單串前，直接在table中找key，如果找到了减宣，把弱引用斷了做一次段清理玩荠。
6. ThreadLocal與內(nèi)存泄漏
關(guān)于ThreadLocal是否會(huì)引起內(nèi)存泄漏也是一個(gè)比較有爭(zhēng)議性的問(wèn)題阶冈，其實(shí)就是要看對(duì)內(nèi)存泄漏的準(zhǔn)確定義是什么女坑。
認(rèn)為T(mén)hreadLocal會(huì)引起內(nèi)存泄漏的說(shuō)法是因?yàn)槿绻粋€(gè)ThreadLocal對(duì)象被回收了匆骗，我們往里面放的value對(duì)于【當(dāng)前線程->當(dāng)前線程的threadLocals(ThreadLocal.ThreadLocalMap對(duì)象）->Entry數(shù)組->某個(gè)entry.value】這樣一條強(qiáng)引用鏈?zhǔn)强蛇_(dá)的碉就，因此value不會(huì)被回收。
認(rèn)為T(mén)hreadLocal不會(huì)引起內(nèi)存泄漏的說(shuō)法是因?yàn)門(mén)hreadLocal.ThreadLocalMap源碼實(shí)現(xiàn)中自帶一套自我清理的機(jī)制筋量。
之所以有關(guān)于內(nèi)存泄露的討論是因?yàn)樵谟芯€程復(fù)用如線程池的場(chǎng)景中桨武，一個(gè)線程的壽命很長(zhǎng)呀酸，大對(duì)象長(zhǎng)期不被回收影響系統(tǒng)運(yùn)行效率與安全。如果線程不會(huì)復(fù)用七咧，用完即銷(xiāo)毀了也不會(huì)有ThreadLocal引發(fā)內(nèi)存泄露的問(wèn)題”妫《Effective Java》一書(shū)中的第6條對(duì)這種內(nèi)存泄露稱為unintentional object retention(無(wú)意識(shí)的對(duì)象保留）先较。
當(dāng)我們仔細(xì)讀過(guò)ThreadLocalMap的源碼悼粮，我們可以推斷扣猫，如果在使用的ThreadLocal的過(guò)程中申尤，顯式地進(jìn)行remove是個(gè)很好的編碼習(xí)慣昧穿，這樣是不會(huì)引起內(nèi)存泄漏时鸵。
那么如果沒(méi)有顯式地進(jìn)行remove呢厅瞎？只能說(shuō)如果對(duì)應(yīng)線程之后調(diào)用ThreadLocal的get和set方法都有很高的概率會(huì)順便清理掉無(wú)效對(duì)象和簸，斷開(kāi)value強(qiáng)引用比搭，從而大對(duì)象被收集器回收。
但無(wú)論如何蜜托，我們應(yīng)該考慮到何時(shí)調(diào)用ThreadLocal的remove方法橄务。一個(gè)比較熟悉的場(chǎng)景就是對(duì)于一個(gè)請(qǐng)求一個(gè)線程的server如tomcat蜂挪，在代碼中對(duì)web api作一個(gè)切面嗓化，存放一些如用戶名等用戶信息刺覆，在連接點(diǎn)方法結(jié)束后，再顯式調(diào)用remove篇梭。
7. 總結(jié)
每個(gè)Thread里都含有一個(gè)ThreadLocalMap的成員變量酝枢，這種機(jī)制將ThreadLocal和線程巧妙地綁定在了一起帘睦，即可以保證無(wú)用的ThreadLocal被及時(shí)回收官脓，不會(huì)造成內(nèi)存泄露卑笨，又可以提升性能。假如我們把ThreadLocalMap做成一個(gè)Map<t extends Thread, ?>類型的Map妖滔，那么它存儲(chǔ)的東西將會(huì)非常多（相當(dāng)于一張全局線程本地變量表）座舍，這樣的情況下用線性探測(cè)法解決哈希沖突的問(wèn)題效率會(huì)非常差曲秉。而JDK里的這種利用ThreadLocal作為key承二，再將ThreadLocalMap與線程相綁定的實(shí)現(xiàn)纲爸，完美地解決了這個(gè)問(wèn)題。
本文參考了以下兩篇文章负蚊。特別是文章1家妆，里面對(duì)ThreadLocal中的清理算法做了詳細(xì)的注釋伤极。感謝兩篇文章作者的分享。
[1]http://www.cnblogs.com/micrari/p/6790229.html
[2]http://www.sczyh30.com/posts/Java/java-concurrent-threadlocal/
