com.hazelcast.internal.util.ConcurrencyUtil Maven / Gradle / Ivy
/*
* Copyright (c) 2008-2024, Hazelcast, Inc. All Rights Reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.hazelcast.internal.util;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.Executor;
import java.util.concurrent.ForkJoinPool;
import java.util.concurrent.atomic.AtomicLongFieldUpdater;
import java.util.function.Function;
/**
* Utility methods to getOrPutSynchronized and getOrPutIfAbsent in a thread safe way
* from a {@link ConcurrentMap} with a {@link ConstructorFunction}.
*/
public final class ConcurrencyUtil {
/**
* The Caller runs executor is an Executor that executes the task on the calling thread.
* This is useful when an Executor is required, but offloading to a different thread
* is very costly and it is faster to run on the calling thread.
*/
public static final Executor CALLER_RUNS = new Executor() {
@Override
public void execute(Runnable command) {
command.run();
}
@Override
public String toString() {
return "CALLER_RUNS";
}
};
// Default executor for async callbacks: ForkJoinPool.commonPool() or a thread-per-task executor when
// the common pool does not support parallelism
private static final Executor DEFAULT_ASYNC_EXECUTOR;
static {
Executor asyncExecutor;
if (ForkJoinPool.getCommonPoolParallelism() > 1) {
asyncExecutor = ForkJoinPool.commonPool();
} else {
asyncExecutor = command -> new Thread(command).start();
}
DEFAULT_ASYNC_EXECUTOR = asyncExecutor;
}
private ConcurrencyUtil() {
}
/**
* WARNING this method should not be called from static context.
*/
public static Executor getDefaultAsyncExecutor() {
return DEFAULT_ASYNC_EXECUTOR;
}
/**
* Returns the {@code value} corresponding to {@code key} in the {@code map}. If {@code key} is not mapped in {@code map},
* then a {@code value} is computed using {@code func}, inserted into the map and returned.
*
* The behavior is equivalent to {@link ConcurrentMap#computeIfAbsent(K, Function)}, with the following exceptions:
*
* - If no mapping, the value of {@code func.createNew(K)} will be inserted into the {@code map} - even if {@code null}
*
- Instances of {@link ConcurrentMap} can override their implementation, but here the implementation can be assured
*
*
* The typical use case of this function over {@link ConcurrentMap#computeIfAbsent(K, Function)} would be in the case of a
* {@link ConcurrentHashMap}, where the implementation is overridden with the following guarantee:
* "The supplied function is invoked exactly once per invocation of this method if the key is absent, else not at all. Some
* attempted update operations on this map by other threads may be blocked while computation is in progress"
*
* By comparison, this implementation cannot guarantee {@code func.createNew(K)} will be executed a maximum of once, because
* no blocking if performed - if the {@code key} is absent from the {@code map}, and this method is called concurrently,
* {@code func.createNew(K)} may be invoked multiple times before any new value is inserted.
*
* This is a performance tradeoff - improve {@code map} throughput and reduce deadlocks, at the expense of redundant object
* instantiation.
*/
public static V getOrPutIfAbsent(ConcurrentMap map, K key, ConstructorFunction func) {
V value = map.get(key);
if (value == null) {
value = func.createNew(key);
V current = map.putIfAbsent(key, value);
value = current == null ? value : current;
}
return value;
}
/**
* Atomically sets the max value.
*
* If the current value is larger than the provided value, the call is ignored.
* So it will not happen that a smaller value will overwrite a larger value.
*/
public static void setMax(E obj, AtomicLongFieldUpdater updater, long value) {
for (; ; ) {
long current = updater.get(obj);
if (current >= value) {
return;
}
if (updater.compareAndSet(obj, current, value)) {
return;
}
}
}
public static V getOrPutSynchronized(ConcurrentMap map, K key, ContextMutexFactory contextMutexFactory,
ConstructorFunction func) {
if (contextMutexFactory == null) {
throw new NullPointerException();
}
V value = map.get(key);
if (value == null) {
ContextMutexFactory.Mutex mutex = contextMutexFactory.mutexFor(key);
try {
synchronized (mutex) {
value = map.get(key);
if (value == null) {
value = func.createNew(key);
map.put(key, value);
}
}
} finally {
mutex.close();
}
}
return value;
}
}