ThreadLocal
能够理解为例如以下三点：
每一个线程都有一个独立于其它线程的上下文来保存这个变量，一个线程的本地变量对其它线程是不可见的。
2、独立于变量的初始化副本
ThreadLocal能够给一个初始值。而每一个线程都会获得这个初始化值的一个副本，这样才干保证不同的线程都有一份拷贝。
3、状态与某一个线程相关联
ThreadLocal 不是用于解决共享变量的问题的。不是为了协调线程同步而存在，而是为了方便每一个线程处理自己的私有状态而引入的一个机制，理解这点对正确使用ThreadLocal至关重要。
ThreadLocal的接口方法
public T get() { }
public void set(T value) { }
public void remove() { }
protected T initialValue() { } get()用来获取当前线程中变量的副本（保存在ThreadLocal中）。
set()用来设置当前线程中变量的副本。
remove()用来移除当前线程中变量的副本。
initialValue()是一个protected方法，用来给ThreadLocal变量提供初始值，每一个线程都会获取这个初始值的一个副本。
使用演示样例
public class ThreadLocalTest {
//创建一个Integer型的线程本地变量
public static final ThreadLocal<Integer> local = new ThreadLocal<Integer>() {
@Override
protected Integer initialValue() {
return 0;
}
};
//计数
static class Counter implements Runnable{
@Override
public void run() {
//获取当前线程的本地变量，然后累加100次
int num = local.get();
for (int i = 0; i < 100; i++) {
num++;
}
//又一次设置累加后的本地变量
local.set(num);
System.out.println(Thread.currentThread().getName() + " : "+ local.get());
}
}
public static void main(String[] args) throws InterruptedException {
Thread[] threads = new Thread[5];
for (int i = 0; i < 5; i++) {
threads[i] = new Thread(new Counter() ,"CounterThread-[" + i+"]");
threads[i].start();
}
}
} 输出：
CounterThread-[2] : 100
CounterThread-[0] : 100
CounterThread-[3] : 100
CounterThread-[1] : 100
CounterThread-[4] : 100
对initialValue函数的正确理解
public class ThreadLocalMisunderstand {
static class Index {
private int num;
public void increase() {
num++;
}
public int getValue() {
return num;
}
}
private static Index num=new Index();
//创建一个Index型的线程本地变量
public static final ThreadLocal<Index> local = new ThreadLocal<Index>() {
@Override
protected Index initialValue() {
return num;
}
};
//计数
static class Counter implements Runnable{
@Override
public void run() {
//获取当前线程的本地变量，然后累加10000次
Index num = local.get();
for (int i = 0; i < 10000; i++) {
num.increase();
}
//又一次设置累加后的本地变量
local.set(num);
System.out.println(Thread.currentThread().getName() + " : "+ local.get().getValue());
}
}
public static void main(String[] args) throws InterruptedException {
Thread[] threads = new Thread[5];
for (int i = 0; i < 5; i++) {
threads[i] = new Thread(new Counter() ,"CounterThread-[" + i+"]");
}
for (int i = 0; i < 5; i++) {
threads[i].start();
}
}
}输出：
CounterThread-[0] : 12019
CounterThread-[2] : 14548
CounterThread-[1] : 13271
CounterThread-[3] : 34069
CounterThread-[4] : 34069
如今得到的计数不一样了。而且每次执行的结果也不一样。说好的线程本地变量呢？
而如今我们的初始值是一个定义好的一个对象，num是这个对象的引用。
换句话说我们的初始值是一个引用。引用的副本和引用指向的不就是同一个对象吗？
假设我们想给每个线程都保存一个Index对象应该怎么办呢？那就是创建对象的副本而不是对象引用的副本。
private static ThreadLocal<Index> local = new ThreadLocal<Index>() {
@Override
protected Index initialValue() {
return new Index(); //注意这里，新建一个对象
}
}; 
ThreadLocal源代码分析
存储结构
public class Thread implements Runnable {
......
ThreadLocal.ThreadLocalMap threadLocals = null;//一个线程相应一个ThreadLocalMap
......
}public class ThreadLocal<T> {
......
static class ThreadLocalMap {//静态内部类
static class Entry extends WeakReference<ThreadLocal> {//键值对
//Entry是ThreadLocal对象的弱引用。this作为键（key）
/** The value associated with this ThreadLocal. */
Object value;//ThreadLocal关联的对象，作为值（value），也就是所谓的线程本地变量
Entry(ThreadLocal k, Object v) {
super(k);
value = v;
}
}
......
private Entry[] table;//用数组保存全部Entry，採用线性探測避免冲突
}
......
}
ThreadLocal源代码
public class ThreadLocal<T> {
/**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. */
//每一个线程相应一个基于线性探測的哈希表（ThreadLocalMap类型），通过Thread类的threadLocals属性关联。
//在该哈希表中，逻辑上key为ThreadLocal，实质上通过threadLocalHashCode属性作为哈希值查找。
private final int threadLocalHashCode = nextHashCode();
/** The next hash code to be given out. Updated atomically. Starts at zero.*/
private static AtomicInteger nextHashCode = new AtomicInteger();
private static final int HASH_INCREMENT = 0x61c88647;
/**Returns the next hash code.*/
private static int nextHashCode() {
return nextHashCode.getAndAdd(HASH_INCREMENT);
}
/**Returns the current thread's "initial value" for this thread-local variable.
This method will be invoked the first time a thread accesses the variable with the {@link #get} method,
unless the thread previously invoked the {@link #set} method, in which case the <tt>initialValue</tt>
method will not be invoked for the thread. Normally, this method is invoked at most once per thread,
but it may be invoked again in case of subsequent invocations of {@link #remove} followed by {@link #get}.
<p>This implementation simply returns <tt>null</tt>; if the programmer desires thread-local variables
to have an initial value other than <tt>null</tt>, <tt>ThreadLocal</tt> must be subclassed,
and this method overridden. Typically, an anonymous inner class will be used.
* @return the initial value for this thread-local
*/
//初始化线程本地变量。注意上面讲到的关于该方法的正确理解
protected T initialValue() {
return null;
}
/** Creates a thread local variable.*/
public ThreadLocal() {
}
/**Returns the value in the current thread's copy of this thread-local variable.
If the variable has no value for the current thread, it is first initialized to the value returned
by an invocation of the {@link #initialValue} method.
* @return the current thread's value of this thread-local
*/
public T get() {
Thread t = Thread.currentThread();//获取当前线程对象
ThreadLocalMap map = getMap(t);//获取当前线程对象关联的ThreadLocalMap
if (map != null) {
ThreadLocalMap.Entry e = map.getEntry(this);//以this作为key，查找线程本地变量
if (e != null)//假设该线程本地变量已经存在。返回就可以
return (T)e.value;
}
return setInitialValue();//假设该线程本地变量不存在，设置初始值并返回
}
/**Variant of set() to establish initialValue.
Used instead of set() in case user has overridden the set() method.
* @return the initial value
*/
private T setInitialValue() {
T value = initialValue();//获取线程本地变量的初始值
Thread t = Thread.currentThread();
ThreadLocalMap map = getMap(t);
if (map != null)//假设当前线程关联的ThreadLocalMap已经存在，将线程本地变量插入哈希表
map.set(this, value);
else
createMap(t, value);//否则，创建新的ThreadLocalMap并将<this,value>组成的键值对增加到ThreadLocalMap中
return value;
}
/**Sets the current thread's copy of this thread-local variableto 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);//获取当前线程的ThreadLocalMap
if (map != null)
map.set(this, value);
else
createMap(t, value);
}
/**Removes the current thread's value for this thread-local variable.
If this thread-local variable is subsequently {@linkplain #get read} by the current thread,
its value will be reinitialized by invoking its {@link #initialValue} method,
unless its value is {@linkplain #set set} by the current thread in the interim.
This may result in multiple invocations of the <tt>initialValue</tt> method in the current thread.
*
* @since 1.5
*/
public void remove() {//从当前线程的ThreadLocalMap中移除线程本地变量
ThreadLocalMap m = getMap(Thread.currentThread());
if (m != null)
m.remove(this);
}
/** Get the map associated with a ThreadLocal. Overridden in InheritableThreadLocal.*/
ThreadLocalMap getMap(Thread t) {
return t.threadLocals;
}
/**Create the map associated with a ThreadLocal. Overridden in InheritableThreadLocal.*/
//创建ThreadLocalMap后与当前线程关联，并将线程本地变量插入ThreadLocalMap
void createMap(Thread t, T firstValue) {
t.threadLocals = new ThreadLocalMap(this, firstValue);
}
//其它
static ThreadLocalMap createInheritedMap(ThreadLocalMap parentMap) {
return new ThreadLocalMap(parentMap);
}
T childValue(T parentValue) {
throw new UnsupportedOperationException();
}
static class ThreadLocalMap {//基于线性探測解决冲突的哈希映射表
......
}
......
}
内存泄露与WeakReference
static class Entry extends WeakReference<ThreadLocal> {
/** The value associated with this ThreadLocal. */
Object value;
Entry(ThreadLocal k, Object v) {
super(k);
value = v;
}
}对于键值对Entry。key为ThreadLocal实例，value为线程本地变量。不难发现，Entry继承自WeakReference<ThreadLocal>。WeakReference就是所谓的弱引用，也就是说Key是一个弱引用（引用ThreadLocal实例）。关于强引用、弱引用。參看：http://blog.csdn.net/sunxianghuang/article/details/52267282
public class Test {
public static class MyThreadLocal extends ThreadLocal {
private byte[] a = new byte[1024*1024*1];
@Override
public void finalize() {
System.out.println("My threadlocal 1 MB finalized.");
}
}
public static class My50MB {//占用内存的大对象
private byte[] a = new byte[1024*1024*50];
@Override
public void finalize() {
System.out.println("My 50 MB finalized.");
}
}
public static void main(String[] args) throws InterruptedException {
new Thread(new Runnable() {
@Override
public void run() {
ThreadLocal tl = new MyThreadLocal();
tl.set(new My50MB());
tl=null;//断开ThreadLocal的强引用
System.out.println("Full GC 1");
System.gc();
}
}).start();
System.out.println("Full GC 2");
System.gc();
Thread.sleep(1000);
System.out.println("Full GC 3");
System.gc();
Thread.sleep(1000);
System.out.println("Full GC 4");
System.gc();
Thread.sleep(1000);
}
}输出：
Full GC 2
Full GC 1
My threadlocal 1 MB finalized.
Full GC 3
My 50 MB finalized.
Full GC 4
从输出能够看出。一旦threadLocal的强引用断开，key的内存就能够得到释放。仅仅有当线程结束后，value的内存才释放。
每一个thread中都存在一个map， map的类型是ThreadLocal.ThreadLocalMap。
Map中的key为一个threadlocal实例。这个Map的确使用了弱引用，只是弱引用仅仅是针对key。每一个key都弱引用指向threadlocal。
当把threadlocal实例置为null以后。没有不论什么强引用指threadlocal实例，所以threadlocal将会被gc回收。可是，我们的value却不能回收。由于存在一条从current
thread连接过来的强引用。
仅仅有当前thread结束以后, current thread就不会存在栈中,强引用断开, Current Thread, Map, value将所有被GC回收.
注: 实线代表强引用,虚线代表弱引用.
所以得出一个结论就是仅仅要这个线程对象被gc回收，就不会出现内存泄露。可是value在threadLocal设为null和线程结束这段时间不会被回收，就发生了我们觉得的“内存泄露”。
因此，最要命的是线程对象不被回收的情况，这就发生了真正意义上的内存泄露。比方使用线程池的时候，线程结束是不会销毁的。会再次使用的，就可能出现内存泄露。
为了最小化内存泄露的可能性和影响，在ThreadLocal的get,set的时候。遇到key为null的entry就会清除相应的value。
所以最怕的情况就是，threadLocal对象设null了，開始发生“内存泄露”，然后使用线程池。这个线程结束，线程放回线程池中不销毁，这个线程一直不被使用，或者分配使用了又不再调用get,set方法。或者get,set方法调用时依旧没有遇到key为null的entry。那么这个期间就会发生真正的内存泄露。
使用ThreadLocal须要注意。每次运行完成后，要使用remove()方法来清空对象,否则 ThreadLocal 存放大对象后，可能会OMM。
为什么使用弱引用
To help deal with very large and long-lived usages, the hash table entries use WeakReferences for keys.
应用实例
Hibernate中使用Threadlocal实现线程相关的Session
參考：
JDK 1.7源代码
http://my.oschina.net/xpbug/blog/113444