ThreadLocal

ThreadLocal(线程变量).每个线程获取与填充都操作当前线程对应的变量副本. 不同线程间的变量副本是隔离开的.

csharp 复制代码

static ThreadLocal<String> threadLocal = new ThreadLocal<>();
public static void main(String[] args) {
    threadLocal.set("main");
    new Thread(()->{
        threadLocal.set("thread1");
    }).start();

    new Thread(()->{
        threadLocal.set("thread2");
        System.out.println("thread2:"+threadLocal.get());
    }).start();
    System.out.println("main:"+threadLocal.get());
}


控制台:
thread2:thread2
main:main

从功能上来ThreadLocal或许是一个容器,key是线程id,value是线程的拥有的对象数据. 如果我设计的话可能ThreadLocal会设计成上述这种.

MyThreadLocal:

ini 复制代码

  public class MyThreadLocal<T>{
    Map<Thread, T> local= new HashMap<Thread, T>();
    T get(){
        Thread thread = Thread.currentThread();
        T o = local.get(thread);
        return o;
    }
    void set(T t){
        Thread thread = Thread.currentThread();
        local.put(thread, t);
    }

}

csharp 复制代码

public static void main(String[] args) throws InterruptedException {
    MyThreadLocal<String> threadLocalUtil = new MyThreadLocal<>();
    threadLocalUtil.set("main");
    System.out.println(threadLocalUtil.get());

    new Thread(()->{
        threadLocalUtil.set("thread2");
        System.out.println("thread2:"+threadLocalUtil.get());
    }).start();

}

来看看这种实现的缺陷:

ThreadLocal就是为多线程所设计的,所以一定涉及并发问题.HashMap线程并不是安全的,所以如果这样实现也要用ConcurrentHashMap<Thread, String> map = new ConcurrentHashMap<>();
ThreadLocal的思想是每个线程持有自己的副本,不存在竞争关系. 上面的实现无论是用hashMap,还是ConcurrentHashMap,或者是锁都会造成资源的竞争.
如果在主线程中new MyThreadLocal();. 只要主线程一直存在,MyThreadLocal引用就一直在,即使使用过MyThreadLocal的子线程已经结束了 .map容器中子线程持有的内容也不会被GC掉. local对象会越来越大.

点进原码看看真正的ThreadLocal是怎么设计的:

ThreadLocalMap: ThreadLocalMap是ThreadLocal的静态内部类,继承了WeakReference的一个map结构.(WeakReference:弱引用只要发生的gc 对象就会被回收.具体可以看下blog.csdn.net/javaphpsqlm...)
threadLocals: ThreadLocal的私有对象,类型是ThreadLocalMap. 数据都存放在各自线程的私有对象的这个map里边,所有称之为数据副本,线程间互相隔离.

从set方法开始,线程首次调用ThreadLocal的set方法时threadLocals是空的,走到createMap. new了一个ThreadLocalMap.

用ThreadLocal对象set值:

scss 复制代码

public void set(T value) {
    Thread t = Thread.currentThread();
    ThreadLocalMap map = getMap(t);
    if (map != null)
        map.set(this, value);
    else
        createMap(t, value);
}

获取当前线程的私有对象threadLocals

javascript 复制代码

ThreadLocalMap getMap(Thread t) {
    return t.threadLocals;
}

创建一个map 并将值存到这个map中

javascript 复制代码

void createMap(Thread t, T firstValue) {
    t.threadLocals = new ThreadLocalMap(this, firstValue);
}

new Entry(firstKey, firstValue)可以看到这个map的key就是ThreadLocal对象本身. value就是需要保存的值 key是ThreadLocal对象本身是因为每个线程可以使用多个ThreadLocal去保存值(

ThreadLocal threadLocal0 = new ThreadLocal<>();
ThreadLocal threadLocal1 = new ThreadLocal<>();
threadLocal0.set(int)
threadLocal0.set(string)

)

ini 复制代码

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);
}

set方法搞清楚了 get方法也都能想到.就是去thread的私有对象threadLocalMap中通过当前调用的threadLocal对象作为key取值.

总体了解了ThreadLocal我们再回顾看下实现细节

为什么要继承弱引用WeakReference.

scala 复制代码

static class ThreadLocalMap {
    static class Entry extends WeakReference<ThreadLocal<?>> {
       
        Object value;

        Entry(ThreadLocal<?> k, Object v) {
            super(k);
            value = v;
        }
    }

那我们反过来看下强引用会发生什么,我们系统中可能会持续运行多个线程,系统中的线程又大部分来源于线程池的复用,线程的周期很长.每个线程中都运行如下代码ThreadLocalMap会存放了大量数据.

csharp 复制代码

public  void systemOut() {
    ThreadLocal<String> threadLocal = new ThreadLocal<>();
    threadLocal.set("BIG_DATA");
    System.out.println(threadLocal.get());
}

方法执行结束了 jvm的栈中threadLocal对象已经没了. 但是ThreadLocalMap仍然持有key和"BIG_DATA"的强引用, 所有不会被GC掉.ThreadLocalMap数据就会越来越大

Entry继承了WeakReference就是为了解决这个问题,把key设置成弱引用.一旦threadLocal的强引用消失,下一次垃圾回收ThreadLocalMap中的key就会被gc掉.注意这里只有key是弱引用. value不是,value不会被gc.

后续get set时会触发一些清除操作,将key为null的value remove掉. 但是这个并不是固定的,在使用ThreadLocal时大家还是要在不需要的时候手动remove掉,尽可能的阻止内存溢出.

从图来看会清晰一点,看下弱引用的作用就是为了当threadLocal为null也就是强引用1不存在的时候.能够吧key1垃圾回收起来

来思考个小问题

value不会被gc可能会造成内存溢出, 为什么不参考key把key、value都搞成弱引用.没有强引用时都给Gc掉

csharp 复制代码

public static void main(String[] args) {
    //假设key value都是弱引用
    ThreadLocal<String> threadLocal = new ThreadLocal<>();
    threadLocal.set("main");
    new Thread(()->{
        threadLocal.set("thread1");
    }).start();
    //模拟jvm发生了垃圾回收
    System.gc();
    System.out.println("main:"+threadLocal.get());
}

控制台:
main:null

上面的代码如果value也是弱引用的话,threadLocal.get()取出来的就是空的了. 因为main方法没结束ThreadLocal threadLocal对象的引用还在栈中,key不会被回收. 但是value也就是字符串"main"没强引用指向他了直接被垃圾回收了!

下面我们继续看下用于储存的数据结构ThreadLocalMap中的细节

ThreadLocalMap的内部类:Entry

scala 复制代码

static class ThreadLocalMap {
    static class Entry extends WeakReference<ThreadLocal<?>> {

ThreadLocalMap的构造函数

new一个初始大小为Entry16数组
计算key的hashCode与数组长度与运算得到下标. threadLocalHashCode里边的内容大致和.hashCode差不多,只不过threadLocalHashCode的散列性更好
设置大小为初始值1
设置阀值,调整大小阈值以在最坏情况下保持2/3的负载系数。

ini 复制代码

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);
}

arduino 复制代码

/**
 * Set the resize threshold to maintain at worst a 2/3 load factor.
 */
private void setThreshold(int len) {
    threshold = len * 2 / 3;
}

hash算法大概是每次使用AtomicInteger原子性的的增加0x61c8864.能够使生成的哈希值足够分散

arduino 复制代码

 private static final int HASH_INCREMENT = 0x61c88647;
     private static int nextHashCode() {
        return nextHashCode.getAndAdd(HASH_INCREMENT);
    }

ThreadLocal的set方法

Entry e = tab[i]; e != null;命中了这个条件说明尽管hash足够分散仍然发生了hash碰撞.
e = tab[i = nextIndex(i, len)]通过线性寻址法找到下一个空的位置.把数据存进去 e.value = value;
线性寻址set和get都会用到.在寻址的过程中会识别到null key的数据,就是前边提到的弱引用.将key为空的元素删除,其余数据向前移动.

ini 复制代码

private void set(ThreadLocal<?> key, Object value) {
    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)]) {
        ThreadLocal<?> k = e.get();

        if (k == key) {
            e.value = value;
            return;
        }

        if (k == null) {
            replaceStaleEntry(key, value, i);
            return;
        }
    }

    tab[i] = new Entry(key, value);
    int sz = ++size;
    if (!cleanSomeSlots(i, sz) && sz >= threshold)
        rehash();
}

线性寻址:当前位置有冲突,就去按顺序依次寻找当前位置+1、+2、+3的地方是否是空的, 如果到了数组的末尾就从0开始直到找到空余位置. 线性寻址是最为简单直接的hash冲突解决方式.其他还有二次寻址、链表、双重散列等.

arduino 复制代码

private static int nextIndex(int i, int len) {
    return ((i + 1 < len) ? i + 1 : 0);
}

ThreadLocal没有继承关系,也就是子线程拿不到父线程的副本. InheritableThreadLocal补充了这个功能

csharp 复制代码

ThreadLocal<String> inheritableThreadLocal = new InheritableThreadLocal<>();
inheritableThreadLocal.set("main");
new Thread(()->{
    
    System.out.println("thread2:"+inheritableThreadLocal.get());
}).start();

控制台输出 main




ThreadLocal<String> inheritableThreadLocal = new InheritableThreadLocal<>();
inheritableThreadLocal.set("main");
new Thread(()->{
    //覆盖父线程内容
    inheritableThreadLocal.set("thread2");
    System.out.println("thread2:"+inheritableThreadLocal.get());
}).start();


控制台输出 thread2

InheritableThreadLocal继承了ThreadLocal 重写了createMap等方法.inheritableThreadLocals取代了ThreadLocal中的对象

javascript 复制代码

void createMap(Thread t, T firstValue) {
    t.inheritableThreadLocals = new ThreadLocalMap(this, firstValue);
}

看一眼Thread的构造函数 ,其中有段逻辑是判断当前线程inheritThreadLocals是否有值有的话调用ThreadLocal.createInheritedMap批量复制过去

csharp 复制代码

public Thread(Runnable target) {
    init(null, target, "Thread-" + nextThreadNum(), 0);
}

ini 复制代码

if (inheritThreadLocals && parent.inheritableThreadLocals != null)
    this.inheritableThreadLocals =
        ThreadLocal.createInheritedMap(parent.inheritableThreadLocals);

因为是创建线程时复制进去的,在子线程new出来之后,父线程向inheritableThreadLocal录入的数据子线程是不会拿到的

csharp 复制代码

ThreadLocal<String> inheritableThreadLocal = new InheritableThreadLocal<>();
new Thread(()->{
    try {
        Thread.sleep(2000);
    } catch (InterruptedException e) {
        throw new RuntimeException(e);
    }
    System.out.println("thread2:"+inheritableThreadLocal.get());
}).start();
inheritableThreadLocal.set("main");

控制台:
thread2:null

InheritableThreadLocal的应用:

子线程需要父线程的登录态信息
日志链路id等

Java基础-ThreadLocal

ThreadLocal