深入理解ConcurrentHashMap

一, 什么是ConcurrentHashMap

ConcurrentHashMap和HashMap一样是一个用来存储键值对<key,value>的集合类,但和HashMap不同的是ConcurrentHashMap是线程安全的,也就是多个线程同时对ConcurrentHashMap进行修改或者删除增加操作不会出现数据错误的问题.

二, 实现原理

和HashMap一样采用数组+链表+红黑树实现

但和HashMap不同的是,数组中存储的节点类型有所增加,包括Node<key,value>,TreeNode<key,value>,ForwardingNode<key,value>,新增这个节点的目的就是为了线程并发协助扩容时使用

三, 基本属性介绍

//01111111111111111111111111111111 该值可以保证计算出来的哈希值为正数
static final int HASH_BITS = 0x7fffffff; // usable bits of normal node hash

//该属性用在扩容时生成一个负值,表示正在扩容
//The number of bits used for generation stamp in sizeCtl.
//sizeCtl中用于生成戳记的位数。
//Must be at least 6 for 32bit arrays.
//对于32位数组，必须至少为6。
private static int RESIZE_STAMP_BITS = 16;

//和上面一样,也是为了在扩容时生成一个负值,具体在代码中解释
//The bit shift for recording size stamp in sizeCtl.
//在sizeCtl中记录大小戳的位移位。
private static final int RESIZE_STAMP_SHIFT = 32 - RESIZE_STAMP_BITS;

//表示当前桶位正在被迁移
//Encodings for Node hash fields. See above for explanation.
static final int MOVED = -1;

//表示当前桶是以树来存储节点的
static final int TREEBIN = -2;

//Number of CPUS, to place bounds on some sizings
//cpu的数量,用来计算元素数量时限制CounterCell数组大小
static final int NCPU = Runtime.getRuntime().availableProcessors();

/**
 * The next table to use; non-null only while resizing.
 * 用来扩容的哈希表
 */
private transient volatile Node<K, V>[] nextTable;

/**
 * Base counter value, used mainly when there is no contention, but also as a fallback during table    initialization 
 * races. Updated via CAS.
 * 哈希表元素数量,通过longAdder来维护
 */
private transient volatile long baseCount;

/**
 * Table initialization and resizing control.
 * 哈希表初始化和扩容大小控制.
 * When negative, the table is being initialized or resized:
 * 当这个值为负数时,表示哈希表正在初始化或重新计算大小
 * -1 for initialization,
 * -1 表示正在初始化了
 * else -(1 + the number of active resizing threads).
 * 表示哈希表正在扩容,-(1+n),表示此时有n个线程正在共同完成哈希表的扩容
 * Otherwise, when table is null, holds the initial table size to use upon creation,or 0 for default.
 * 否则,当哈希表为空时, 保留要创建哈希表的大小0或默认(16)
 * After initialization, holds the next element count value upon which to resize the table.
 * 初始化完成之后,保留下一次需要扩容的阈值
 */
private transient volatile int sizeCtl;

/**
 * The next table index (plus one) to split while resizing.
 * 扩容时的当前转移下标
 */
private transient volatile int transferIndex;

/**
 * Spinlock (locked via CAS) used when resizing and/or creating CounterCells.
 * 获取计算集合元素容量的CounterCell对象的锁
 */
private transient volatile int cellsBusy;

/**
 * Table of counter cells. When non-null, size is a power of 2.
 * 计算元素数量的数组
 */
private transient volatile CounterCell[] counterCells;

四, 构造函数

/**
 * 和HashMap构造函数不同的是,数组容量的计算总是大于传入容量的2的幂
 * 即如果传入32则数组初始容量为64,而不是32,而HashMap计算出来为32
 */
public ConcurrentHashMap(int initialCapacity) {
    if (initialCapacity < 0)
        throw new IllegalArgumentException();
    int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
               MAXIMUM_CAPACITY : tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
    this.sizeCtl = cap;
}

五, 常用方法介绍

/**
 * 这个方法就是HashMap中的hash方法,用来计算哈希值
 */
static final int spread(int h) {
    return (h ^ (h >>> 16)) & HASH_BITS;
}

获取节点

public V get(Object key) {
    Node<K, V>[] tab;
    Node<K, V> e, p;
    int n, eh;
    K ek;
    //计算散列值
    int h = spread(key.hashCode());
    //计算下标(这一块同HashMap不再赘述)
    if ((tab = table) != null && (n = tab.length) > 0 && (e = tabAt(tab, (n - 1) & h)) != null) {
        if ((eh = e.hash) == h) {
            if ((ek = e.key) == key || (ek != null && key.equals(ek)))
                return e.val;
        } else if (eh < 0)
            //哈希值小于0,表示为树节点,从树中寻找,这一步和HashMap一致
            return (p = e.find(h, key)) != null ? p.val : null;
        //在链表中寻找
        while ((e = e.next) != null) {
            if (e.hash == h && ((ek = e.key) == key || (ek != null && key.equals(ek))))
                return e.val;
        }
    }
    return null;
}

插入节点

public V put(K key, V value) {
    return putVal(key, value, false);
}

final V putVal(K key, V value, boolean onlyIfAbsent) {
    if (key == null || value == null) throw new NullPointerException();
    //计算哈希值
    int hash = spread(key.hashCode());
    //插入桶的节点数量
    int binCount = 0;
    //使用死循环,目的是可能有的线程正在协助扩容,之后还需要插入或者更新,或者需要操作的节点所在的桶已经被其他线程锁定,需要等待其他线程执行完之后再执行
    for (Node<K, V>[] tab = table; ; ) {
        Node<K, V> f;
        int n, i, fh;
        if (tab == null || (n = tab.length) == 0)
            //初始化哈希表
            tab = initTable();
        else if
            //计算下标,并且计算该下标是否有元素
            ((f = tabAt(tab, i = (n - 1) & hash)) == null) {
            //cas插入,这一步不需要锁,因为当前桶为空
            if (casTabAt(tab, i, null, new Node<K, V>(hash, key, value, null)))
                break;
        } else if
            //代表当前节点已经被移动,正在扩容,需要当前线程协助扩容
            ((fh = f.hash) == MOVED)
            tab = helpTransfer(tab, f);
        else {
            V oldVal = null;
            //锁住头节点,保证所有线程的插入都是线程安全的
            synchronized (f) {
                //这一步判断的原因是,可能插入元素之后会造成链表树化,需要插入的位置已经发生了变化
                if (tabAt(tab, i) == f) {
                    if (fh >= 0) {
                        //当前桶上有多少个节点
                        binCount = 1;
                        for (Node<K, V> e = f; ; ++binCount) {
                            K ek;
                            //查找到了key相同的节点,直接修改值并返回旧的值
                            if (e.hash == hash && ((ek = e.key) == key || (ek != null && key.equals(ek)))) {
                                oldVal = e.val;
                                if (!onlyIfAbsent)
                                    e.val = value;
                                break;
                            }
                            Node<K, V> pred = e;
                            //没有找到相同的key,直接向链表尾部插入节点
                            if ((e = e.next) == null) {
                                pred.next = new Node<K, V>(hash, key, value, null);
                                break;
                            }
                        }
                    } else if (f instanceof TreeBin) {
                        //给树里面插入节点
                        Node<K, V> p;
                        binCount = 2;
                        if ((p = ((TreeBin<K, V>) f).putTreeVal(hash, key, value)) != null) {
                            oldVal = p.val;
                            if (!onlyIfAbsent)
                                p.val = value;
                        }
                    }
                }
            }
            if (binCount != 0) {
                //需要树化
                if (binCount >= TREEIFY_THRESHOLD)
                    treeifyBin(tab, i);
                //修改旧值,直接将旧值返回
                if (oldVal != null)
                    return oldVal;
                break;
            }
        }
    }
    //计算节点数量
    addCount(1L, binCount);
    return null;
}

初始化哈希表

private final Node<K, V>[] initTable() {
    Node<K, V>[] tab;
    int sc;
    while ((tab = table) == null || tab.length == 0) {
        //小于0表示正在初始化或者正在扩容,让出cpu
        if ((sc = sizeCtl) < 0)
            Thread.yield(); // lost initialization race; just spin
        else if
            //判断sc是否与SIZECTL是否相等,如果相等,则将SIZECTL设置为-1,表示当前正在初始化(只有一个线程能进行此操作,其他线程会被挡在前面的判断上)
            (U.compareAndSwapInt(this, SIZECTL, sc, -1)) {
            try {
                //防止有线程已经初始化了1
                if ((tab = table) == null || tab.length == 0) {
                    //该sc如果在构造器上传入了,则会被计算为大于其的2次幂,否则会按照默认值初始化
                    int n = (sc > 0) ? sc : DEFAULT_CAPACITY;
                    @SuppressWarnings("unchecked")
                    Node<K, V>[] nt = (Node<K, V>[]) new Node<?, ?>[n];
                    table = tab = nt;
                    //设置下一次扩容的阈值 n - (n >>> 2) = n - n / 4 = (3 / 4) * n = 0.75n,即下一次的扩容阈值为当前哈希表数量的0.75*n
                    sc = n - (n >>> 2);
                }
            } finally {
                //设置sizeCtl为-1,表示初始化动作已经有线程在执行了
                sizeCtl = sc;
            }
            break;
        }
    }
    return tab;
}

计算节点数量

private final void addCount(long x, int check) {
    CounterCell[] as;
    long b, s;
    /*
     * 维护数组长度
     */
    //尝试cas直接修改值,如果修改失败
    if ((as = counterCells) != null || !U.compareAndSwapLong(this, BASECOUNT, b = baseCount, s = b + x)) {
        CounterCell a;
        long v;
        int m;
        boolean uncontended = true;
        //数组为空或者长度小于0或者对应的位置为空或者直接修改数组对应位置上的值失败,则进行修改操作
        if (as == null || (m = as.length - 1) < 0 || (a = as[ThreadLocalRandom.getProbe() & m]) == null ||
            !(uncontended = U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))) {
            fullAddCount(x, uncontended);
            return;
        }
        //桶上的节点数量小于等于1,不需要判断扩容,直接退出
        if (check <= 1)
            return;
        //获取当前数组的元素数量
        s = sumCount();
    }
    /*
     * 判断是否需要扩容
     */
    if (check >= 0) {
        Node<K, V>[] tab, nt;
        int n, sc;
        //当前节点数量大于扩容阈值,并且数组不为空并且数组长度小于最大值则需要扩容
        while (s >= (long) (sc = sizeCtl) && (tab = table) != null && (n = tab.length) < MAXIMUM_CAPACITY) {
            //获取一个负值
            int rs = resizeStamp(n);
            //如果sc小于0,说明正在扩容,需要协助扩容
            if (sc < 0) {
                //判断扩容是否完成
                if ((sc >>> RESIZE_STAMP_SHIFT) != rs || sc == rs + 1 || sc == rs + MAX_RESIZERS || (nt = nextTable) == null || transferIndex <= 0)
                    break;
                //协助扩容,这里sc+1代表新加入一个线程协助扩容
                if (U.compareAndSwapInt(this, SIZECTL, sc, sc + 1))
                    transfer(tab, nt);
            } else if
                /*
                 * 假设 rs = 00000000 00000000 10000000 00000000
                 * 将其向左移16位结果为 10000000 00000000 00000000 00000000 可以看出该值为负
                 * 这一步尝试将sc设置为负数
                 */
                (U.compareAndSwapInt(this, SIZECTL, sc, (rs << RESIZE_STAMP_SHIFT) + 2))
                //将旧数组置空,里面会创建一个新的数组
                transfer(tab, null);
            //计算集合元素数量
            s = sumCount();
        }
    }
}

private final void fullAddCount(long x, boolean wasUncontended) {
    int h;
    //获取当前线程的hash值
    if ((h = ThreadLocalRandom.getProbe()) == 0) {
        ThreadLocalRandom.localInit();      // force initialization
        h = ThreadLocalRandom.getProbe();
        wasUncontended = true;
    }
    //检测是否有冲突,如果最后一个桶不为null,则为true
    boolean collide = false;
    for (; ; ) {
        CounterCell[] as;
        CounterCell a;
        int n;
        long v;
        //数组如果不为空,则优先对CounterCell里面的counterCell的value进行累加
        if ((as = counterCells) != null && (n = as.length) > 0) {
            //当前位置为空
            if ((a = as[(n - 1) & h]) == null) {
                //当前没有线程尝试修改该值
                if (cellsBusy == 0) {
                    CounterCell r = new CounterCell(x);
                    //抢占修改的锁
                    if (cellsBusy == 0 && U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
                        boolean created = false;
                        try {
                            CounterCell[] rs;
                            int m, j;
                            if ((rs = counterCells) != null && (m = rs.length) > 0 && rs[j = (m - 1) & h] == null) {
                                rs[j] = r;
                                created = true;
                            }
                        } finally {
                            //释放锁
                            cellsBusy = 0;
                        }
                        if (created)
                            break;
                        continue;           // Slot is now non-empty
                    }
                }
                //抢占失败
                collide = false;
            } else if
                //桶位不为空,重新计算线程hash值,继续循环
                (!wasUncontended)       // CAS already known to fail
                wasUncontended = true;      // Continue after rehash
            /*
                     * 重新计算hash值之后,对应的桶位还是不为空,对value进行累加
                     * 尝试cas对value加值
                     */
            else if (U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))
                break;
            //数组长度已经大于等于CPU的核数了,不需要再扩容了
            else if (counterCells != as || n >= NCPU)
                collide = false;
            //当没有冲突,修改为有冲突,重新计算hash值,继续循环
            else if (!collide)
                collide = true;
            else if
                //多次循环没有设置成功值,则对原数组进行扩容
                (cellsBusy == 0 && U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
                try {
                    if (counterCells == as) {// Expand table unless stale
                        //数组长度没有超过cpu核数,将数组扩容两倍
                        CounterCell[] rs = new CounterCell[n << 1];
                        for (int i = 0; i < n; ++i)
                            //扩容使用
                            rs[i] = as[i];
                        counterCells = rs;
                    }
                } finally {
                    cellsBusy = 0;
                }
                collide = false;
                continue;                   // Retry with expanded table
            }
            //重新计算随机值
            h = ThreadLocalRandom.advanceProbe(h);
        } else if
            //初始进来数组为空,需要初始化数组
            (cellsBusy == 0 && counterCells == as && U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
            boolean init = false;
            try {
                if (counterCells == as) {
                    CounterCell[] rs = new CounterCell[2];
                    rs[h & 1] = new CounterCell(x);
                    counterCells = rs;
                    init = true;
                }
            } finally {
                cellsBusy = 0;
            }
            if (init)
                break;
        } else if
            //数组为空并且有其他线程正在创建数组,尝试直接对baseCount进行累加
            (U.compareAndSwapLong(this, BASECOUNT, v = baseCount, v + x))
            break;                          // Fall back on using base
    }
}

扩容并迁移

private final void transfer(Node<K, V>[] tab, Node<K, V>[] nextTab) {
    //stride表示迁移数据的区间
    int n = tab.length, stride;
    /*
     * 这里计算每个CPU负责迁移元素的个数
     * 如果这里的跨度区间小于16,则按照最小区间16来计算
     */
    if ((stride = (NCPU > 1) ? (n >>> 3) / NCPU : n) < MIN_TRANSFER_STRIDE)
        stride = MIN_TRANSFER_STRIDE;
    //这里表示为第一个线程来扩容
    if (nextTab == null) {
        try {
            //扩容为两倍
            @SuppressWarnings("unchecked")
            Node<K, V>[] nt = (Node<K, V>[]) new Node<?, ?>[n << 1];
            nextTab = nt;
        } catch (Throwable ex) {
            sizeCtl = Integer.MAX_VALUE;
            return;
        }
        nextTable = nextTab;
        //迁移数据的index
        transferIndex = n;
    }
    //新扩容数组的长度
    int nextn = nextTab.length;
    //创建头节点,该节点会被标识为MOVED表示数据正在迁移中
    ForwardingNode<K, V> fwd = new ForwardingNode<K, V>(nextTab);

    boolean advance = true;
    boolean finishing = false; // to ensure sweep before committing nextTab
    //从后向前迁移
    for (int i = 0, bound = 0; ; ) {
        Node<K, V> f;
        int fh;
        while (advance) {
            int nextIndex, nextBound;
            //不属于自己的迁移位置或者已经迁移完成直接退出
            if (--i >= bound || finishing)
                advance = false;
            else if
                //下一个迁移位置小于等于0直接退出
                ((nextIndex = transferIndex) <= 0) {
                i = -1;
                advance = false;
            } else if
                //计算迁移位置(多线程会划分多个区间)
                (U.compareAndSwapInt(this, TRANSFERINDEX, nextIndex, nextBound = (nextIndex > stride ? nextIndex - stride : 0))) {
                bound = nextBound;
                i = nextIndex - 1;
                advance = false;
            }
        }
        if (i < 0 || i >= n || i + n >= nextn) {
            int sc;
            //判断是否所有的线程都做完了任务
            if (finishing) {
                nextTable = null;
                table = nextTab;
                //等于0.75 * 2n,也就是新数组扩容2倍*扩容因子
                sizeCtl = (n << 1) - (n >>> 1);
                return;
            }
            if (U.compareAndSwapInt(this, SIZECTL, sc = sizeCtl, sc - 1)) {
                //判断扩容是否成功
                if ((sc - 2) != resizeStamp(n) << RESIZE_STAMP_SHIFT)
                    return;
                finishing = advance = true;
                i = n; // recheck before commit
            }
        } else if
            //如果当前位置为空,直接插入fwd节点,表示当前节点正在被迁移
            ((f = tabAt(tab, i)) == null)
            advance = casTabAt(tab, i, null, fwd);
        else if ((fh = f.hash) == MOVED)
            //正在被迁移,需要重新计算位置
            advance = true; // already processed
        else {
            //迁移代码
            synchronized (f) {
                if (tabAt(tab, i) == f) {
                    Node<K, V> ln, hn;
                    if (fh >= 0) {
                        int runBit = fh & n;
                        Node<K, V> lastRun = f;
                        for (Node<K, V> p = f.next; p != null; p = p.next) {
                            int b = p.hash & n;
                            if (b != runBit) {
                                runBit = b;
                                lastRun = p;
                            }
                        }
                        if (runBit == 0) {
                            ln = lastRun;
                            hn = null;
                        } else {
                            hn = lastRun;
                            ln = null;
                        }
                        for (Node<K, V> p = f; p != lastRun; p = p.next) {
                            int ph = p.hash;
                            K pk = p.key;
                            V pv = p.val;
                            if ((ph & n) == 0)
                                ln = new Node<K, V>(ph, pk, pv, ln);
                            else
                                hn = new Node<K, V>(ph, pk, pv, hn);
                        }
                        setTabAt(nextTab, i, ln);
                        setTabAt(nextTab, i + n, hn);
                        //迁移完成,设置头节点为fwd
                        setTabAt(tab, i, fwd);
                        advance = true;
                    } else if (f instanceof TreeBin) {
                        TreeBin<K, V> t = (TreeBin<K, V>) f;
                        TreeNode<K, V> lo = null, loTail = null;
                        TreeNode<K, V> hi = null, hiTail = null;
                        int lc = 0, hc = 0;
                        for (Node<K, V> e = t.first; e != null; e = e.next) {
                            int h = e.hash;
                            TreeNode<K, V> p = new TreeNode<K, V>
                                (h, e.key, e.val, null, null);
                            if ((h & n) == 0) {
                                if ((p.prev = loTail) == null)
                                    lo = p;
                                else
                                    loTail.next = p;
                                loTail = p;
                                ++lc;
                            } else {
                                if ((p.prev = hiTail) == null)
                                    hi = p;
                                else
                                    hiTail.next = p;
                                hiTail = p;
                                ++hc;
                            }
                        }
                        ln = (lc <= UNTREEIFY_THRESHOLD) ? untreeify(lo) :
                        (hc != 0) ? new TreeBin<K, V>(lo) : t;
                        hn = (hc <= UNTREEIFY_THRESHOLD) ? untreeify(hi) :
                        (lc != 0) ? new TreeBin<K, V>(hi) : t;
                        setTabAt(nextTab, i, ln);
                        setTabAt(nextTab, i + n, hn);
                        //迁移完成,设置头节点为fwd
                        setTabAt(tab, i, fwd);
                        //重新计算位置继续迁移
                        advance = true;
                    }
                }
            }
        }
    }
}

获取节点数量

public int size() {
    long n = sumCount();
    return ((n < 0L) ? 0 : (n > (long) Integer.MAX_VALUE) ? Integer.MAX_VALUE : (int) n);
}

/**
 * 获取哈希表中节点的数量(非线程安全)
 * 这个方法返回的数据不一定准确,因为可能在调用该方法的时候,有其他线程正在尝试给数组中的节点加值
 */
final long sumCount() {
    CounterCell[] as = counterCells;
    CounterCell a;
    long sum = baseCount;
    if (as != null) {
        for (CounterCell counterCell : as) {
            if ((a = counterCell) != null)
                sum += a.value;
        }
    }
    return sum;
}