java并发：CopyOnWrite机制

CopyOnWriteArrayList

java.util.concurrent.CopyOnWriteArrayList是一个线程安全的 ArrayList，其修改操作是在底层的一个复制的数组(快照)上进行的，即用了写时复制策略。

其定义如下：

public class CopyOnWriteArrayList<E>
    implements List<E>, RandomAccess, Cloneable, java.io.Serializable {
    private static final long serialVersionUID = 8673264195747942595L;

    /**
     * The lock protecting all mutators.  (We have a mild preference
     * for builtin monitors over ReentrantLock when either will do.)
     */
    final transient Object lock = new Object();

    /** The array, accessed only via getArray/setArray. */
    private transient volatile Object[] array;

    /**
     * Gets the array.  Non-private so as to also be accessible
     * from CopyOnWriteArraySet class.
     */
    final Object[] getArray() {
        return array;
    }

其构造函数如下：

    /**
     * Creates an empty list.
     */
    public CopyOnWriteArrayList() {
        setArray(new Object[0]);
    }

    /**
     * Creates a list containing the elements of the specified
     * collection, in the order they are returned by the collection's
     * iterator.
     *
     * @param c the collection of initially held elements
     * @throws NullPointerException if the specified collection is null
     */
    public CopyOnWriteArrayList(Collection<? extends E> c) {
        Object[] es;
        if (c.getClass() == CopyOnWriteArrayList.class)
            es = ((CopyOnWriteArrayList<?>)c).getArray();
        else {
            es = c.toArray();
            if (c.getClass() != java.util.ArrayList.class)
                es = Arrays.copyOf(es, es.length, Object[].class);
        }
        setArray(es);
    }

    /**
     * Creates a list holding a copy of the given array.
     *
     * @param toCopyIn the array (a copy of this array is used as the
     *        internal array)
     * @throws NullPointerException if the specified array is null
     */
    public CopyOnWriteArrayList(E[] toCopyIn) {
        setArray(Arrays.copyOf(toCopyIn, toCopyIn.length, Object[].class));
    }

解释：

针对无参构造函数，其内部创建了一个大小为 0 的 Object 数组作为array 的初始值。

其添加元素的方法有多个，如下：

add(E e)的定义如下：

    /**
     * Appends the specified element to the end of this list.
     *
     * @param e element to be appended to this list
     * @return {@code true} (as specified by {@link Collection#add})
     */
    public boolean add(E e) {
        synchronized (lock) {
            Object[] es = getArray();
            int len = es.length;
            es = Arrays.copyOf(es, len + 1);
            es[len] = e;
            setArray(es);
            return true;
        }
    }

弱一致性问题

CopyOnWriteArrayList中获取指定位置的元素的方法如下：

    /**
     * {@inheritDoc}
     *
     * @throws IndexOutOfBoundsException {@inheritDoc}
     */
    public E get(int index) {
        return elementAt(getArray(), index);
    }

    static <E> E elementAt(Object[] a, int index) {
        return (E) a[index];
    }

解说：

从上述代码可以知道，当线程 x调用 get方法获取指定位置的元素时需要两步：

• 获取 array数组(步骤A)
• 通过下标访问指定位置的元素 (步骤B)

这个两步操作在整个过程中并没有加锁。

问题：假设线程 x执行完步骤 A 后，在执行步骤 B 前，线程 y进行了 remove操作，这时会发生什么？

remove方法的代码如下：

    /**
     * Removes the element at the specified position in this list.
     * Shifts any subsequent elements to the left (subtracts one from their
     * indices).  Returns the element that was removed from the list.
     *
     * @throws IndexOutOfBoundsException {@inheritDoc}
     */
    public E remove(int index) {
        synchronized (lock) {
            Object[] es = getArray();
            int len = es.length;
            E oldValue = elementAt(es, index);
            int numMoved = len - index - 1;
            Object[] newElements;
            if (numMoved == 0)
                newElements = Arrays.copyOf(es, len - 1);
            else {
                newElements = new Object[len - 1];
                System.arraycopy(es, 0, newElements, 0, index);
                System.arraycopy(es, index + 1, newElements, index,
                                 numMoved);
            }
            setArray(newElements);
            return oldValue;
        }
    }

remove操作获取独占锁，然后进行写时复制操作，在复制的数组里删除 index 处的元素，随后让 array 指向新复制的数组。  

  

        

上图指出了线程 x 执行步骤B时操作的数组是线程 y 删除元素之前线程 x 复制的数组。

所以，线程 y 虽然己经删除了 index 处的元素，但是线程 x 的步骤 B 还是会返回index 处的元素，这就是写时复制策略的弱一致性问题。

迭代器弱一致性

大家都知道可以使用法代器遍历List的元素，当调用 CopyOnWriteArrayList 的 iterator()方法获取法代器时，实际上会返回一个 COWiterator对象。

    /**
     * Returns an iterator over the elements in this list in proper sequence.
     *
     * <p>The returned iterator provides a snapshot of the state of the list
     * when the iterator was constructed. No synchronization is needed while
     * traversing the iterator. The iterator does <em>NOT</em> support the
     * {@code remove} method.
     *
     * @return an iterator over the elements in this list in proper sequence
     */
    public Iterator<E> iterator() {
        return new COWIterator<E>(getArray(), 0);
    }

    static final class COWIterator<E> implements ListIterator<E> {
        /** Snapshot of the array */
        private final Object[] snapshot;
        /** Index of element to be returned by subsequent call to next.  */
        private int cursor;

        COWIterator(Object[] es, int initialCursor) {
            cursor = initialCursor;
            snapshot = es;
        }

        public boolean hasNext() {
            return cursor < snapshot.length;
        }

        public boolean hasPrevious() {
            return cursor > 0;
        }

        @SuppressWarnings("unchecked")
        public E next() {
            if (! hasNext())
                throw new NoSuchElementException();
            return (E) snapshot[cursor++];
        }

        @SuppressWarnings("unchecked")
        public E previous() {
            if (! hasPrevious())
                throw new NoSuchElementException();
            return (E) snapshot[--cursor];
        }

        public int nextIndex() {
            return cursor;
        }

        public int previousIndex() {
            return cursor - 1;
        }

        /**
         * Not supported. Always throws UnsupportedOperationException.
         * @throws UnsupportedOperationException always; {@code remove}
         *         is not supported by this iterator.
         */
        public void remove() {
            throw new UnsupportedOperationException();
        }

        /**
         * Not supported. Always throws UnsupportedOperationException.
         * @throws UnsupportedOperationException always; {@code set}
         *         is not supported by this iterator.
         */
        public void set(E e) {
            throw new UnsupportedOperationException();
        }

        /**
         * Not supported. Always throws UnsupportedOperationException.
         * @throws UnsupportedOperationException always; {@code add}
         *         is not supported by this iterator.
         */
        public void add(E e) {
            throw new UnsupportedOperationException();
        }

        @Override
        public void forEachRemaining(Consumer<? super E> action) {
            Objects.requireNonNull(action);
            final int size = snapshot.length;
            int i = cursor;
            cursor = size;
            for (; i < size; i++)
                action.accept(elementAt(snapshot, i));
        }
    }

解说：

COWiterator对象的 snapshot变量保存了当前 array的内容，cursor是遍历数据的下标。

在线程使用返回的迭代器遍历元素的过程中，如果有其他线程对 list 进行了增删改，那么 snapshot就是快照了。

因为增删改后指向 array的变量（即本文最开始着色了的那行代码，private transient volatile Object[] array;）指向了新数组。 

所以在获取迭代器遍历元素时，其他线程对 list 进行的增删改对外不可见，因为它们操作的是两个不同的数组，这就是弱一致性。 

CopyOnWriteArraySet

java.util.concurrent.CopyOnWriteArraySet是在CopyOnWriteArrayList的基础上使用了装饰模式，其很多方法都是调用CopyOnWriteArrayList的API来实现的。

其添加元素的代码如下：

    /**
     * Adds the specified element to this set if it is not already present.
     * More formally, adds the specified element {@code e} to this set if
     * the set contains no element {@code e2} such that
     * {@code Objects.equals(e, e2)}.
     * If this set already contains the element, the call leaves the set
     * unchanged and returns {@code false}.
     *
     * @param e element to be added to this set
     * @return {@code true} if this set did not already contain the specified
     *         element
     */
    public boolean add(E e) {
        return al.addIfAbsent(e);
    }

解说：

从CopyOnWriteArraySet的定义以及构造函数可以看到其底层数据结构就是CopyOnWriteArrayList，代码如下：

public class CopyOnWriteArraySet<E> extends AbstractSet<E>
        implements java.io.Serializable {
    private static final long serialVersionUID = 5457747651344034263L;

    private final CopyOnWriteArrayList<E> al;

    /**
     * Creates an empty set.
     */
    public CopyOnWriteArraySet() {
        al = new CopyOnWriteArrayList<E>();
    }

    /**
     * Creates a set containing all of the elements of the specified
     * collection.
     *
     * @param c the collection of elements to initially contain
     * @throws NullPointerException if the specified collection is null
     */
    public CopyOnWriteArraySet(Collection<? extends E> c) {
        if (c.getClass() == CopyOnWriteArraySet.class) {
            @SuppressWarnings("unchecked") CopyOnWriteArraySet<E> cc =
                (CopyOnWriteArraySet<E>)c;
            al = new CopyOnWriteArrayList<E>(cc.al);
        }
        else {
            al = new CopyOnWriteArrayList<E>();
            al.addAllAbsent(c);
        }
    }