JAVA对象在线程间是共享的，某些场景下，我们期望对象在线程间隔离，即一个对象只对一个线程可见，其他线程无法访问。ThreadLocal提供了线程内局部变量的机制，这个变量在线程的生命周期中存活。
下面是官方文档中的ThreadLocal的api：
代码示例
public class ThreadLocalTest { private static final ThreadLocal<Integer> stringThreadLocal = new ThreadLocal<>(); public static void main(String[] args) { CyclicBarrier barrier = new CyclicBarrier(4); for (int i = 0; i < 4; i ++) { new Thread(new MyThread(barrier)).start(); } } static class MyThread implements Runnable { private CyclicBarrier barrier; public MyThread(CyclicBarrier barrier) { this.barrier = barrier; } @Override public void run() { try { barrier.await(); for (int i = 0; i < 100; i++) { Integer value = stringThreadLocal.get(); if (value == null) { value = 0; } Integer sum = value + i; stringThreadLocal.set(sum); } System.out.println(Thread.currentThread().getName() + " sum is " + stringThreadLocal.get()); } catch (Exception e) { e.printStackTrace(); } } } } /** 运行结果 Thread-1 sum is 4950 Thread-0 sum is 4950 Thread-3 sum is 4950 Thread-2 sum is 4950 **/从运行结果可以看出，每个线程的变量是独有的，线程之间不会相互影响。
最初的想法就是维护一个线程到值得映射map，如下：
public class MyThreadLocal<T> { private Map<Thread, T> keyValueMap = new WeakHashMap<>(); public synchronized void set(T value) { Thread thread = Thread.currentThread(); keyValueMap.put(thread, value); } public synchronized T get() { Thread thread = Thread.currentThread(); return keyValueMap.get(thread); } public synchronized void remove() { Thread thread = Thread.currentThread(); keyValueMap.remove(thread); } }自己实现存在的问题：
1、synchronized 导致并发性能较差
带着问题，我们看下ThreadLocal是怎么实现的。
先来看set方法，如下：
public void set(T value) { Thread t = Thread.currentThread(); ThreadLocalMap map = getMap(t); if (map != null) map.set(this, value); else createMap(t, value); } void createMap(Thread t, T firstValue) { t.threadLocals = new ThreadLocalMap(this, firstValue); } ThreadLocalMap getMap(Thread t) { return t.threadLocals; }可以看到，set方法将set值的任务委托给了ThreadLocalMap，而ThreadLocalMap居然是Thread中的变量，这跟假想的变量保存在ThreadLocal中是不一样的，稍后对ThreadLocalMap进行深入分析。set方法的逻辑也很清晰，如果ThreadLocalMap不为空则直接设置值，否则创建新的ThreadLocalMap。接下来继续看get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(); } private T setInitialValue() { T value = initialValue(); Thread t = Thread.currentThread(); ThreadLocalMap map = getMap(t); if (map != null) map.set(this, value); else createMap(t, value); return value; } protected T initialValue() { return null; }这里get方法也是将操作委托给了ThreadLocalMap，通过最终获得ThreadLocalMap.Entry来获取最终的值。提一下setInitialValue()方法，用来设置初始值，而initialValue()方法是protected，子类继承ThreadLocal可以实现初始值的设置。
下面来看remove方法，同样是交给ThreadLocalMap来处理。
public void remove() { ThreadLocalMap m = getMap(Thread.currentThread()); if (m != null) m.remove(this); }ThreadLocal的实现重点在ThreadLocalMap中的实现。先看ThreadLocalMap的存储位置，它是存在Thread对象中的，而非存在ThreadLocal对象中。我理解这样做是为了避免多线程竞争，因为放在Thread对象中就相当于线程私有了，处理的时候不需要加锁，这样就避免了我们前面自己实现的代码中加锁导致效率低的问题。由于ThreadLocal本身的设计就是变量不与其他线程共享，不需要其他线程访问本对象的变量，放在Thread对象中不会有问题。
继续看ThreadLocalMap源码。ThreadLocalMap是ThreadLocal的一个静态内部类。
1、Entry数据结构
ThreadLocalMap维护了一个Entry类型的table数据
/** * 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; } } /** * The table, resized as necessary. * table.length MUST always be a power of two. */ private Entry[] table;Entry是一个以ThreadLocal为key，Object为value的键值对，需要注意的是，threadLocal是弱引用，当外部的threadLocal被置为null时，threadLocal会被回收。
2、set源码
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. ThreadLocal.ThreadLocalMap.Entry[] tab = table; int len = tab.length; // 定位Entry存放的位置 int i = key.threadLocalHashCode & (len-1); // 处理hash冲突的情况，这里采用的是开放地址法 for (ThreadLocal.ThreadLocalMap.Entry e = tab[i]; e != null; e = tab[i = nextIndex(i, len)]) { ThreadLocal<?> k = e.get(); //更新 if (k == key) { e.value = value; return; } if (k == null) { replaceStaleEntry(key, value, i); return; } } // 新建entry并插入 tab[i] = new ThreadLocal.ThreadLocalMap.Entry(key, value); int sz = ++size; // 清除脏数据，扩容 if (!cleanSomeSlots(i, sz) && sz >= threshold) rehash(); }3、get源码
private ThreadLocal.ThreadLocalMap.Entry getEntry(ThreadLocal<?> key) { // 确定entry位置 int i = key.threadLocalHashCode & (table.length - 1); ThreadLocal.ThreadLocalMap.Entry e = table[i]; // 命中 if (e != null && e.get() == key) return e; else // 存在hash冲突，继续查找 return getEntryAfterMiss(key, i, e); } private ThreadLocal.ThreadLocalMap.Entry getEntryAfterMiss(ThreadLocal<?> key, int i, ThreadLocal.ThreadLocalMap.Entry e) { ThreadLocal.ThreadLocalMap.Entry[] tab = table; int len = tab.length; while (e != null) { ThreadLocal<?> k = e.get(); //找到entry if (k == key) return e; // 脏数据处理 if (k == null) expungeStaleEntry(i); else //遍历 i = nextIndex(i, len); e = tab[i]; } return null; }4、remove源码
private void remove(ThreadLocal<?> key) { ThreadLocal.ThreadLocalMap.Entry[] tab = table; int len = tab.length; int i = key.threadLocalHashCode & (len-1); for (ThreadLocal.ThreadLocalMap.Entry e = tab[i]; e != null; e = tab[i = nextIndex(i, len)]) { if (e.get() == key) { //清空key e.clear(); //清空value expungeStaleEntry(i); return; } } }先来看下ThreadLocal,ThreadLcoalMap,Entry之间的关系图
图中实线表示强引用，虚线表示弱引用。如果threadLocal外部强引用被设置为null，内存中的threadLocal对象会被回收（虚引用不影响回收），但是value值的强引用仍然存在，所以value不会被回收，存在内存泄漏。当然线程如果执行结束，threadLocalMap、Entry都会被回收掉，但是实际开发中我们为了线程复用，大都用线程池维护线程，内存泄漏的问题依然会产生。当然，TheadLocal的实现已经对这种情况进行了处理。简单来说就是在set、get、remove等操作的时候，会将key=null && value != null的entry的value设置为null，清理掉脏数据。
