CODEWITHSUNDEEP
javaconcurrenthashmapjava.util.concurrent
khotyn
khotyn

Reputation: 954

Why the "next" field in ConcurrentHashMap$HashEntry is final

I'm reading reading the source code of java.util.ConcurrentHashMap and find that the next field in ConcurrentHashMap$HashEntry is final. There are two operations that is possible to modify the value of next:add and remove. But those two operations cat be done thread safely even though the next field is not final. So I cann't understand why the next field is final, can anyone tell me why? thanks.

Upvotes: 1

Views: 284

Answers (1)

axtavt
axtavt

Reputation: 242716

final on next is needed to ensure that reading threads see the initialized value of the field, because (most of) reads of ConcurrentHashMap happen without synchronization.

Note that, for example, value is not final, therefore reading thread can see a non-initialized value (null) in that field, and have to recheck it under synchronization in that case:

    V get(Object key, int hash) {
        if (count != 0) { // read-volatile
            HashEntry<K,V> e = getFirst(hash);
            while (e != null) {
                if (e.hash == hash && key.equals(e.key)) {
                    V v = e.value;
                    if (v != null)
                        return v;
                    return readValueUnderLock(e); // recheck
                }
                e = e.next;
            }
        }
        return null;
    }

Upvotes: 1

Related Questions