synchronized(this) 用法详解

news/2025/11/27 17:42:55/文章来源:https://www.cnblogs.com/yppah/p/19278914

基本概念

synchronized(this) 是 Java 中的一种同步机制，通过对当前对象实例加锁来保证线程安全。当一个线程进入被 synchronized(this) 修饰的代码块时，会获取当前对象的锁，其他线程必须等待该线程释放锁后才能进入。

基本语法

public class MyClass {public void method() {synchronized(this) {// 需要同步保护的代码// 同一时刻只有一个线程能执行这里的代码}}
}

使用场景详解

1. 实例变量的线程安全操作

public class Counter {private int count = 0;// 不安全的递增方法public void unsafeIncrement() {count++; // 非原子操作，可能出现竞态条件}// 安全的递增方法public void safeIncrement() {synchronized(this) {count++;}}// 获取当前计数值public int getCount() {synchronized(this) {return count;}}
}

2. 复杂业务逻辑的原子性保证

public class BankAccount {private double balance;public BankAccount(double initialBalance) {this.balance = initialBalance;}public boolean withdraw(double amount) {synchronized(this) {// 检查余额是否充足if (balance >= amount) {// 模拟处理时间try {Thread.sleep(100);} catch (InterruptedException e) {Thread.currentThread().interrupt();}balance -= amount;return true;}return false;}}public void deposit(double amount) {synchronized(this) {// 模拟处理时间try {Thread.sleep(50);} catch (InterruptedException e) {Thread.currentThread().interrupt();}balance += amount;}}public double getBalance() {synchronized(this) {return balance;}}
}

3. 多步骤操作的原子性保证

public class InventoryManager {private Map<String, Integer> inventory = new HashMap<>();public InventoryManager() {// 初始化库存inventory.put("item1", 100);inventory.put("item2", 50);}// 批量扣减库存 - 需要保证原子性public boolean deductInventory(Map<String, Integer> itemsToDeduct) {synchronized(this) {// 第一步：检查所有商品库存是否充足for (Map.Entry<String, Integer> entry : itemsToDeduct.entrySet()) {String itemId = entry.getKey();Integer quantity = entry.getValue();Integer currentStock = inventory.get(itemId);if (currentStock == null || currentStock < quantity) {return false; // 库存不足，整个操作失败}}// 第二步：扣减所有商品库存for (Map.Entry<String, Integer> entry : itemsToDeduct.entrySet()) {String itemId = entry.getKey();Integer quantity = entry.getValue();Integer currentStock = inventory.get(itemId);inventory.put(itemId, currentStock - quantity);}return true;}}public int getStock(String itemId) {synchronized(this) {return inventory.getOrDefault(itemId, 0);}}
}

4. 缓存更新的线程安全

public class CachedDataService {private Map<String, Object> cache = new HashMap<>();private long lastUpdateTime = 0;private static final long CACHE_TIMEOUT = 60000; // 1分钟public Object getData(String key) {synchronized(this) {// 检查缓存是否过期if (System.currentTimeMillis() - lastUpdateTime > CACHE_TIMEOUT) {refreshCache();}return cache.get(key);}}public void putData(String key, Object data) {synchronized(this) {cache.put(key, data);lastUpdateTime = System.currentTimeMillis();}}private void refreshCache() {// 刷新缓存逻辑cache.clear();// 重新加载数据...lastUpdateTime = System.currentTimeMillis();}public void clearCache() {synchronized(this) {cache.clear();lastUpdateTime = 0;}}
}

与其他同步方式的比较

1. synchronized 方法 vs synchronized(this)

public class ComparisonExample {private int value = 0;// 方式1：同步整个方法public synchronized void method1() {value++;// 其他操作...}// 方式2：同步代码块public void method2() {// 一些不需要同步的操作System.out.println("不需要同步的操作");synchronized(this) {value++;// 需要同步的操作}// 更多不需要同步的操作System.out.println("更多不需要同步的操作");}
}

2. synchronized(this) vs synchronized(object)

public class LockGranularityExample {private final Object lock1 = new Object();private final Object lock2 = new Object();private int counter1 = 0;private int counter2 = 0;// 粗粒度锁定 - 锁定整个对象public void coarseGrainedOperation() {synchronized(this) {counter1++;counter2++;}}// 细粒度锁定 - 分别锁定不同的资源public void fineGrainedOperation() {synchronized(lock1) {counter1++;}synchronized(lock2) {counter2++;}}
}

实际应用场景

1. 单例模式中的懒加载

public class LazySingleton {private static LazySingleton instance;private LazySingleton() {}public static LazySingleton getInstance() {if (instance == null) {synchronized(LazySingleton.class) {if (instance == null) {instance = new LazySingleton();}}}return instance;}
}

2. 生产者-消费者模式

public class Buffer {private final Queue<Integer> queue = new LinkedList<>();private final int capacity = 10;public void produce(int item) throws InterruptedException {synchronized(this) {while (queue.size() == capacity) {// 缓冲区满，等待消费者消费wait();}queue.offer(item);System.out.println("生产: " + item);notifyAll(); // 通知等待的消费者}}public int consume() throws InterruptedException {synchronized(this) {while (queue.isEmpty()) {// 缓冲区空，等待生产者生产wait();}int item = queue.poll();System.out.println("消费: " + item);notifyAll(); // 通知等待的生产者return item;}}
}

3. 状态机的状态转换

public class StateMachine {private enum State {IDLE, PROCESSING, COMPLETED, ERROR}private State currentState = State.IDLE;public void startProcessing() {synchronized(this) {if (currentState == State.IDLE) {currentState = State.PROCESSING;System.out.println("开始处理...");} else {throw new IllegalStateException("不能从 " + currentState + " 状态开始处理");}}}public void completeProcessing() {synchronized(this) {if (currentState == State.PROCESSING) {currentState = State.COMPLETED;System.out.println("处理完成");} else {throw new IllegalStateException("不能在 " + currentState + " 状态完成处理");}}}public State getCurrentState() {synchronized(this) {return currentState;}}
}

注意事项和最佳实践

1. 避免死锁

public class DeadlockAvoidance {private final Object lockA = new Object();private final Object lockB = new Object();// 错误的做法 - 可能导致死锁public void badMethod1() {synchronized(lockA) {synchronized(lockB) {// 操作...}}}public void badMethod2() {synchronized(lockB) {synchronized(lockA) {// 操作...}}}// 正确的做法 - 统一加锁顺序public void goodMethod1() {synchronized(lockA) {synchronized(lockB) {// 操作...}}}public void goodMethod2() {synchronized(lockA) {  // 保持相同的加锁顺序synchronized(lockB) {// 操作...}}}
}

2. 减少锁的持有时间

public class LockHoldingOptimization {private List<String> dataList = new ArrayList<>();// 不好的做法 - 锁持有时间过长public void badPractice() {synchronized(this) {// 执行耗时操作expensiveOperation();// 添加数据dataList.add("data");}}// 好的做法 - 减少锁持有时间public void goodPractice() {// 先执行耗时操作String result = expensiveOperation();// 只在必要时加锁synchronized(this) {dataList.add(result);}}private String expensiveOperation() {// 模拟耗时操作try {Thread.sleep(1000);} catch (InterruptedException e) {Thread.currentThread().interrupt();}return "processed_data";}
}