java 使用ConcurrentHashMap和計數器實現鎖

在某些場景下,我們想讓線程根據某些業務數據進行排隊,簡單代碼如下：

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{ //處理業務 } 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來排隊,如果每個線程處理數據后不釋放鎖的話，那么可以不利用計數器。但是加了釋放鎖的操作，則必須加上計算器。因為當線程把鎖釋放掉后，還沒來得及退出synchronized 代碼塊時，另外一個線程調用了tryLock方法，那該線程將拿到另外一個對象的鎖，導致利用synchronized 關鍵字進行userId排隊失敗。

也可以利用guava的API來實現。

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))){ //處理業務操作 } } }