java 使用ConcurrentHashMap和计数器实现锁

在某些场景下,我们想让线程根据某些业务数据进行排队,简单代码如下：

import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.atomic.AtomicInteger;

public class TestServiceImpl {
  private static ConcurrentHashMap<Long, LockObj> lockMap = new ConcurrentHashMap<Long, LockObj>(40);

  public void test(Long userId){
    LockObj lock = tryLock(userId);
    synchronized (lock) {
      try{
        //处理业务
      }
      finally{
        unLock(lock);
      }
    }
  }

  private LockObj tryLock(Long key) {
    LockObj curVal = new LockObj(key);
    LockObj preVal = lockMap.putIfAbsent(key, curVal);
    if (null == preVal) {
      curVal.inc();
      return curVal;
    }
    else{
      preVal.inc();
    }
    return preVal;
  }

  private void unLock(LockObj lock){
    if (lock.dec() <= 0){
      lockMap.remove(lock.getKey());
    }
  }

  public class LockObj {
    private long key = 0;
    private AtomicInteger count = new AtomicInteger(0);

    public LockObj(long key){
      this.key = key;
    }

    public int inc(){
      return count.incrementAndGet();
    }
    public int dec(){
      return count.decrementAndGet();
    }

    public long getKey(){
      return key;
    }

    @Override
    public String toString() {
      return "LockObj [key=" + key + ", count=" + count + "]";
    }
  }

}

按照userId来排队,如果每个线程处理数据后不释放锁的话，那么可以不利用计数器。但是加了释放锁的操作，则必须加上计算器。因为当线程把锁释放掉后，还没来得及退出synchronized 代码块时，另外一个线程调用了tryLock方法，那该线程将拿到另外一个对象的锁，导致利用synchronized 关键字进行userId排队失败。

也可以利用guava的API来实现。

import com.google.common.collect.Interner;
import com.google.common.collect.Interners;


public class TestServiceImpl {

  Interner<String> pool = Interners.newWeakInterner();

  public void test(Long userId) throws OspException {

    synchronized ( pool.intern(String.valueOf(userId))){
      //处理业务操作
    }
  }
}