ThreadLocal（TL）是Java中一种线程局部变量实现机制，他为每个线程提供一个单独的变量副本，保证多线程场景下，变量的线程安全。常常用于取代参数的显式传递。
InheritableThreadLocal（ITL）是JDK提供的TL增强版，而TransmittableThreadLocal（TTL）是阿里开源的ITL增强版
这些ThreadLocal在不同场景下有不同用途，我们来分析一下：
ThreadLocal主要的方法有四个：initialValue、set、get、remove
当线程首次访问该ThreadLocal时（ThreadLocal.get()），会进行初始化赋值。我们常用两种方法初始化ThreadLocal
ThreadLocal<String> threadLocal = new ThreadLocal<String>() { @Override protected String initialValue() { return "";
}
};ThreadLocal<String> threadLocal = ThreadLocal.withInitial(() -> "");
他会创建一个SuppliedThreadLocal内部类
public static <S> ThreadLocal<S> withInitial(Supplier<? extends S> supplier) { return new SuppliedThreadLocal<>(supplier);
}该类重写了initialValue方法
static final class SuppliedThreadLocal<T> extends ThreadLocal<T> { private final Supplier<? extends T> supplier;
SuppliedThreadLocal(Supplier<? extends T> supplier) { this.supplier = Objects.requireNonNull(supplier);
} @Override protected T initialValue() { //当该线程首次访问ThreadLocal时，会间接调用lambda表达式初始化 return supplier.get();
}
}⚠️ITL并没有重新实现withInitial，如果使用withInitial则会创建STL，失去自己增强的特性
public void set(T value) {
Thread t = Thread.currentThread();
ThreadLocalMap map = getMap(t); if (map != null)
map.set(this, value); else
createMap(t, value);
}这里出现了一个关键属性ThreadLocalMap，类定义在ThreadLocal中，是Thread的成员变量
ThreadLocalMap getMap(Thread t) { return t.threadLocals;
}ThreadLocalMap内部还有一个内部类Entry，是存值的地方
static class ThreadLocalMap { static class Entry extends WeakReference<ThreadLocal<?>> { Object value;
Entry(ThreadLocal<?> k, Object v) { //ThreadLocal的引用是“key” super(k); //线程局部变量是value
value = v;
}
} //Entry数组 //value具体放在哪个index下，是由ThreadLocal的hashCode算出来的
private Entry[] table;
}public T get() {
Thread t = Thread.currentThread(); //1、获取线程的ThreadLocalMap
ThreadLocalMap map = getMap(t); if (map != null) { //2、根据ThreadLocal的hashCode，获取对应Entry下的value
ThreadLocalMap.Entry e = map.getEntry(this); if (e != null) { @SuppressWarnings("unchecked")
T result = (T)e.value; return result;
}
} //3、如果没有赋过值，则初始化 return setInitialValue();
} public void remove() {
ThreadLocalMap m = getMap(Thread.currentThread()); if (m != null) //会将对应Entry、包括他的key、value手动置null
m.remove(this);
}我们先来看一个例子
public static void main(String[] args) throws InterruptedException {
ThreadLocal<String> threadLocal = ThreadLocal.withInitial(() -> "A");
threadLocal.set("B");
Thread thread = new Thread(() -> {
System.out.println("子线程ThreadLocal：" + threadLocal.get());
}, "子线程");
thread.start();
thread.join();
}打印结果如下，可见子线程的ThreadLocal是初始值，并没有使用父线程修改后的值：
子线程ThreadLocal：A
线程的ThreadLocalMap是首次访问时创建的，所以子线程使用ThreadLocal的时候，会初始化一个新的ThreadLocal，线程局部变量为默认值
⚠️所以，TL不具有遗传性
为了解决TL子线程遗传性的问题，JDK引入了ITL
他继承ThreadLocal，重写了childValue、getMap、createMap三个方法
public class InheritableThreadLocal<T> extends ThreadLocal<T> { protected T childValue(T parentValue) { return parentValue;
} ThreadLocalMap getMap(Thread t) { return t.inheritableThreadLocals;
} void createMap(Thread t, T firstValue) {
t.inheritableThreadLocals = new ThreadLocalMap(this, firstValue);
}
}这里出现了inheritableThreadLocals，他存储的就是从父线程拷贝过来的ThreadLocal，这个值是在父线程首次修改ThreadLocal的时候赋值的，然后在子线程创建时拷贝过来的
//父线程部分：public void set(T value) {
Thread t = Thread.currentThread(); //该方法被ITL重写，访问inheritableThreadLocals为null
ThreadLocalMap map = getMap(t); if (map != null)
map.set(this, value); else //该方法同样被ITL重写，创建一个ThreadLocalMap赋值给inheritableThreadLocals
createMap(t, value);
}//子线程部分：public Thread(Runnable target) {
init(null, target, "Thread-" + nextThreadNum(), 0);
}private void init(ThreadGroup g, Runnable target, String name, long stackSize, AccessControlContext acc,
boolean inheritThreadLocals) { //省略一些代码... //获取当前线程（父线程、也就是创建子线程的线程）
Thread parent = currentThread(); //1、允许ThreadLocal遗传（这个默认为true） //2、inheritableThreadLocals不为空，由于父线程调用set了 //父线程不调用set，那ThreadLocal就是初始值，那直接初始化就好了，也不用进该分支 if (inheritThreadLocals && parent.inheritableThreadLocals != null) this.inheritableThreadLocals =
ThreadLocal.createInheritedMap(parent.inheritableThreadLocals);
}//createInheritedMap使用该构造函数，根据父线程的inheritableThreadLocals进行深拷贝private ThreadLocalMap(ThreadLocalMap parentMap) {
Entry[] parentTable = parentMap.table; int len = parentTable.length;
setThreshold(len);
table = new Entry[len]; //深拷贝父线程ThreadLocalMap for (int j = 0; j < len; j++) {
Entry e = parentTable[j]; if (e != null) {
@SuppressWarnings("unchecked")
ThreadLocal<Object> key = (ThreadLocal<Object>) e.get(); if (key != null) { //childValue被ITL重写，返回父线程ThreadLocal的值
Object value = key.childValue(e.value);
Entry c = new Entry(key, value); int h = key.threadLocalHashCode & (len - 1); while (table[h] != null)
h = nextIndex(h, len);
table[h] = c;
size++;
}
}
}
}使用ITL的效果
public static void main(String[] args) throws InterruptedException {
ThreadLocal<String> threadLocal = new InheritableThreadLocal<String>() {
@Override protected String initialValue() { return "A";
}
};
threadLocal.set("B");
Thread thread = new Thread(() -> {
System.out.println("子线程ThreadLocal：" + threadLocal.get());
}, "子线程");
thread.start();
thread.join();
}打印结果如下，子线程拷贝了父线程ThreadLocal：
子线程ThreadLocal：B
总结一下，ITL解决父子线程遗传性的核心思路是，将可遗传的ThreadLocal放在父线程新的ThreadLocalMap中，在子线程首次使用时进行拷贝
我们再从一个简单的例子说起
public static void main(String[] args) throws InterruptedException, ExecutionException {
ThreadLocal<String> threadLocal = new InheritableThreadLocal<String>() {
@Override protected String initialValue() { return "A";
}
};
threadLocal.set("B");
ExecutorService executorService = Executors.newFixedThreadPool(1); //1、子线程第一次获取ThreadLocal
executorService.submit(() -> System.out.println("子线程ThreadLocal："+threadLocal.get())).get();
Thread.sleep(1000); //2、父线程修改ThreadLocal
threadLocal.set("C");
System.out.println("父线程修改ThreadLocal为"+threadLocal.get()); //3、子线程第二次获取ThreadLocal
executorService.submit(() -> System.out.println("子线程ThreadLocal："+threadLocal.get())).get();
}打印结果如下，子线程在第二次打印时，并没有拷贝父线程的ThreadLocal，使用的还是首次拷贝的值：
子线程ThreadLocal：B 父线程修改ThreadLocal为C 子线程ThreadLocal：B
⚠️可复用的子线程不会感知父线程ThreadLocal的变化
TTL在ITL上做了稍微复杂的封装，我们从使用开始了解
引入依赖
<dependency> <groupId>com.alibaba</groupId> <artifactId>transmittable-thread-local</artifactId> <version>latest</version></dependency>
在使用TTL时，线程需要经过TTL封装，线程池同理
public static void main(String[] args) throws InterruptedException, ExecutionException {
ThreadLocal<String> threadLocal = new TransmittableThreadLocal<String>() {
@Override protected String initialValue() { return "A";
}
};
threadLocal.set("B");
ExecutorService executorService = TtlExecutors.getTtlExecutorService(Executors.newFixedThreadPool(1));
executorService.submit(() -> System.out.println("子线程ThreadLocal：" + threadLocal.get())).get();
Thread.sleep(1000);
threadLocal.set("C");
System.out.println("父线程修改ThreadLocal为" + threadLocal.get());
executorService.submit(() -> System.out.println("子线程ThreadLocal：" + threadLocal.get())).get();
Thread.sleep(1000);
executorService.submit(() -> {
threadLocal.set("D");
System.out.println("子线程修改ThreadLocal为" + threadLocal.get());
});
Thread.sleep(1000);
executorService.submit(() -> System.out.println("子线程ThreadLocal：" + threadLocal.get()));
Thread.sleep(1000);
}打印结果如下，子线程每次都会获取父线程的ThreadLocal
子线程ThreadLocal：B 父线程修改ThreadLocal为C 子线程ThreadLocal：C 子线程修改ThreadLocal为D 子线程ThreadLocal：C
从使用上看，TTL要求将任务封装，那我们就从ThreadLocal和ExecutorService两部分入手
下面是TTL的取值和赋值逻辑，都涉及一个关键方法addThisToHolder，对应的属性holder会在线程池执行任务时用到
//TransmittableThreadLocal.addThisToHolder()private void addThisToHolder() { //InheritableThreadLocal<WeakHashMap<TransmittableThreadLocal<Object>, ?>> holder if (!holder.get().containsKey(this)) { //holder是静态变量，他会把TTL存到当前线程的map中 //value是null，他实则是把Map当Set用 //主线程赋值时，会获取主线程的holderMap，然后把TTL存进去
holder.get().put((TransmittableThreadLocal<Object>) this, null);
}
}
@Overridepublic final void set(T value) { if (!disableIgnoreNullValueSemantics && null == value) { remove();
} else {
super.set(value); //当主线程赋值时，会将自己的TTL放到自己的map中
addThisToHolder();
}
}
@Overridepublic final T get() {
T value = super.get(); if (disableIgnoreNullValueSemantics || null != value)
addThisToHolder(); return value;
}//我们通过TtlExecutors.getTtlExecutorService()对线程池进行封装public static ExecutorService getTtlExecutorService(@Nullable ExecutorService executorService) { if (TtlAgent.isTtlAgentLoaded() || executorService == null || executorService instanceof TtlEnhanced) { return executorService;
} //入参是线程池，通过包装类代理线程池的操作 return new ExecutorServiceTtlWrapper(executorService);
}//ExecutorServiceTtlWrapper.submit()public Future<?> submit(@NonNull Runnable task) { //将提交的任务进行封装 return executorService.submit(TtlRunnable.get(task));
}TtlRunnable构造方法
这里都是主线程在操作，由于任务是主线程提交的
private TtlRunnable(@NonNull Runnable runnable, boolean releaseTtlValueReferenceAfterRun) { this.capturedRef = new AtomicReference<Object>(capture()); this.runnable = runnable; this.releaseTtlValueReferenceAfterRun = releaseTtlValueReferenceAfterRun;
}这里有一个关键属性capturedRef，他是一个原子引用，存了TTL
//TrasmitableThreadLocal.Transmitterpublic static Object capture() { //获取ttl的值构建快照 return new Snapshot(captureTtlValues(), captureThreadLocalValues());
}private static HashMap<TransmittableThreadLocal<Object>, Object> captureTtlValues() {
HashMap<TransmittableThreadLocal<Object>, Object> ttl2Value = new HashMap<TransmittableThreadLocal<Object>, Object>(); for (TransmittableThreadLocal<Object> threadLocal : holder.get().keySet()) { //将主线程TTL的值存到当前任务中
ttl2Value.put(threadLocal, threadLocal.copyValue());
} return ttl2Value;
}任务执行的代码如下，在任务执行前回放ThreadLocal，在任务执行后恢复ThreadLocal：
这里都是子线程在操作，由于任务都是子线程执行的
@Overridepublic void run() { Object captured = capturedRef.get(); if (captured == null || releaseTtlValueReferenceAfterRun && !capturedRef.compareAndSet(captured, null)) { throw new IllegalStateException("TTL value reference is released after run!");
} //1、备份子线程ThreadLocal //2、使用主线程提交任务时构建的ThreadLocal副本，将子线程ThreadLocal覆盖 Object backup = replay(captured); try { //3、任务执行
runnable.run();
} finally { //3、使用之前备份的子线程ThreadLocal进行恢复
restore(backup);
}
}总结一下，TTL让子线程感知父线程变化的核心思路是，主线程在任务提交时构建ThreadLocal副本，在子线程执行任务时供其使用
⚠️提交和执行任务会对TTL进行若干操作，理论上对性能有一点点影响，官方性能测试结论说损耗可忽略
TTL官方性能测试
作者：京东物流 刘朝永
来源：京东云开发者 自猿其说Tech