分段锁的原理

import java.util.Map;

/**
 * Created by louyuting on 17/1/10.
 */
public class StripedMap {
  //同步策略: buckets[n]由 locks[n%N_LOCKS] 来保护

  private static final int N_LOCKS = 16;//分段锁的个数
  private final Node[] buckets;
  private final Object[] locks;

  /**
   * 结点
   * @param <K>
   * @param <V>
   */
  private static class Node<K,V> implements Map.Entry<K,V>{
    final K key;//key
    V value;//value
    Node<K,V> next;//指向下一个结点的指针
    int hash;//hash值

    //构造器，传入Entry的四个属性
    Node(int h, K k, V v, Node<K,V> n) {
      value = v;
      next = n;//该Entry的后继
      key = k;
      hash = h;
    }

    public final K getKey() {
      return key;
    }

    public final V getValue() {
      return value;
    }

    public final V setValue(V newValue) {
      V oldValue = value;
      value = newValue;
      return oldValue;
    }

  }

  /**
   * 构造器: 初始化散列桶和分段锁数组
   * @param numBuckets
   */
  public StripedMap(int numBuckets) {
    buckets = new Node[numBuckets];
    locks = new Object[N_LOCKS];

    for(int i=0; i<N_LOCKS; i++){
      locks[i] = new Object();
    }

  }

  /**
   * 返回散列之后在散列桶之中的定位
   * @param key
   * @return
   */
  private final int hash(Object key){
    return Math.abs(key.hashCode() % N_LOCKS);
  }


  /**
   * 分段锁实现的get
   * @param key
   * @return
   */
  public Object get(Object key){
    int hash = hash(key);//计算hash值

    //获取分段锁中的某一把锁
    synchronized (locks[hash% N_LOCKS]){
      for(Node m=buckets[hash]; m!=null; m=m.next){
        if(m.key.equals(key)){
          return m.value;
        }
      }
    }

    return null;
  }

  /**
   * 清除整个map
   */
  public void clear() {
    //分段获取散列桶中每个桶地锁，然后清除对应的桶的锁
    for(int i=0; i<buckets.length; i++){
      synchronized (locks[i%N_LOCKS]){
        buckets[i] = null;
      }
    }
  }
}