Java concurrent development-sample code of built-in lock Synchronized-javaTutorial-php.cn

Abstract:

In multi-threaded programming, thread safety is one of the most critical issues. Its core concept lies in correctness, that is, when multiple threads access a shared , variable data will never lead to data corruption or other unexpected results. When all concurrency modes solve this problem, they use serialized access to critical resources. In Java, two ways are provided to implement synchronized mutually exclusive access: synchronized and Lock. This article discusses in detail the application of synchronized built-in locks in Java concurrency, including its specific usage scenarios (synchronized methods, synchronized code blocks, instance object locks and Class object locks), reentrancy and precautions.

1. Thread safety issues

Thread safety issues will not occur in single threads, but in multi-thread programming, it is possible to access the same shared and variable resources at the same time. , this resource can be: a variable, an object, a file, etc. Pay special attention to two points,

Shared: means that the resource can be accessed by multiple threads at the same time;
Variable: means that the resource Can be modified during its lifetime. Therefore, when multiple threads access this kind of resource at the same time, there will be a problem: since the execution process of each thread is uncontrollable, a synchronization mechanism needs to be used to coordinate access to the variable state of the object.

Give an example of dirty reading of data:

//资源类
class PublicVar {

    public String username = "A";
    public String password = "AA";

    //同步实例方法
    public synchronized void setValue(String username, String password) {
        try {
            this.username = username;
            Thread.sleep(5000);
            this.password = password;

            System.out.println("method=setValue " +"\t" + "threadName="
                    + Thread.currentThread().getName() + "\t" + "username="
                    + username + ", password=" + password);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }

    //非同步实例方法
    public void getValue() {
        System.out.println("method=getValue " + "\t" +  "threadName="
                + Thread.currentThread().getName()+ "\t" + " username=" + username
                + ", password=" + password);
    }
}

//线程类
class ThreadA extends Thread {

    private PublicVar publicVar;

    public ThreadA(PublicVar publicVar) {
        super();
        this.publicVar = publicVar;
    }

    @Override
    public void run() {
        super.run();
        publicVar.setValue("B", "BB");
    }
}

//测试类
public class Test {

    public static void main(String[] args) {
        try {
            //临界资源
            PublicVar publicVarRef = new PublicVar();

            //创建并启动线程
            ThreadA thread = new ThreadA(publicVarRef);
            thread.start();

            Thread.sleep(200);// 打印结果受此值大小影响

            //在主线程中调用
            publicVarRef.getValue();

        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}/* Output ( 数据交叉 ): 
        method=getValue     threadName=main         username=B, password=AA
        method=setValue     threadName=Thread-0     username=B, password=BB
 *///:~

Copy after login

It can be seen from the program output that although the write operation is synchronized, there may still be some unexpected events in the read operation. Unexpected situations, such as the dirty read shown above. Dirty reading occurs when the corresponding data has been partially modified by other threads when performing a read operation, resulting in data crossover.

This is actually a thread safety issue, that is, when multiple threads access a resource at the same time, the program running result will not be the result you want to see. Here, this resource is called: critical resource. In other words, when multiple threads access critical resources (an object, a property in an object, a file, a database, etc.) at the same time, thread safety issues may arise.

However, when multiple threads execute a method, the local variables inside the method are not critical resources, because these local variables are in the private stack of each thread, so they are not shared and will not Lead to thread safety issues.

2. How to solve thread safety issues

In fact, all concurrency modes use serialized access to critical resources when solving thread safety issues. That is, at the same time, only one thread can access critical resources, also called synchronous mutually exclusive access. In other words, add a lock in front of the code that accesses the critical resource. After accessing the critical resource, release the lock to allow other threads to continue accessing.

In Java, two ways are provided to implement synchronous mutually exclusive access: synchronized and Lock. This article mainly talks about the use of synchronized. The use of Lock is explained in my other blog post "Java Concurrency: Detailed Explanation of Lock Framework".

3. synchronized synchronization method or synchronization block

Before understanding how to use the synchronized keyword, let’s first look at a concept: mutual exclusion lock, which can achieve the purpose of mutual exclusion access Lock. To give a simple example, if a mutex lock is added to a critical resource, when one thread accesses the critical resource, other threads can only wait.

In Java, you can use the synchronized keyword to mark a method or code block. When a thread calls the synchronized method of the object or accesses the synchronized code block, the thread obtains the lock of the object. Other threads cannot access this method temporarily. Only when this method or code block is completed will this thread release the object's lock and other threads can execute this method or code block.

In the following code, the two threads call the insertData object respectivelyInsert data:

1) synchronized method

public class Test {

    public static void main(String[] args)  {
        final InsertData insertData = new InsertData();
        // 启动线程 1  
        new Thread() {
            public void run() {
                insertData.insert(Thread.currentThread());
            };
        }.start();

        // 启动线程 2
        new Thread() {
            public void run() {
                insertData.insert(Thread.currentThread());
            };
        }.start();
    }  
}

class InsertData {

    // 共享、可变资源
    private ArrayList<Integer> arrayList = new ArrayList<Integer>();

    //对共享可变资源的访问
    public void insert(Thread thread){
        for(int i=0;i<5;i++){
            System.out.println(thread.getName()+"在插入数据"+i);
            arrayList.add(i);
        }
    }
}/* Output: 
        Thread-0在插入数据0
        Thread-1在插入数据0
        Thread-0在插入数据1
        Thread-0在插入数据2
        Thread-1在插入数据1
        Thread-1在插入数据2
 *///:~

Copy after login

According to the running results, it can be seen that these two threads are executing the insert() method at the same time. And if the keyword synchronized is added in front of the insert() method, the running result is:

class InsertData {
    private ArrayList<Integer> arrayList = new ArrayList<Integer>();

    public synchronized void insert(Thread thread){
        for(int i=0;i<5;i++){
            System.out.println(thread.getName()+"在插入数据"+i);
            arrayList.add(i);
        }
    }
}/* Output: 
        Thread-0在插入数据0
        Thread-0在插入数据1
        Thread-0在插入数据2
        Thread-1在插入数据0
        Thread-1在插入数据1
        Thread-1在插入数据2
 *///:~

Copy after login

From the above output results, it can be seen that Thread-1 inserts data after Thread-0 has inserted the data. of. Note that Thread-0 and Thread-1 execute the insert() method sequentially. This is what the synchronized keyword does for methods.

However, you need to pay attention to the following three points:

1）当一个线程正在访问一个对象的 synchronized 方法，那么其他线程不能访问该对象的其他 synchronized 方法。这个原因很简单，因为一个对象只有一把锁，当一个线程获取了该对象的锁之后，其他线程无法获取该对象的锁，所以无法访问该对象的其他synchronized方法。

2）当一个线程正在访问一个对象的 synchronized 方法，那么其他线程能访问该对象的非 synchronized 方法。这个原因很简单，访问非 synchronized 方法不需要获得该对象的锁，假如一个方法没用 synchronized 关键字修饰，说明它不会使用到临界资源，那么其他线程是可以访问这个方法的，

3）如果一个线程 A 需要访问对象 object1 的 synchronized 方法 fun1，另外一个线程 B 需要访问对象 object2 的 synchronized 方法 fun1，即使 object1 和 object2 是同一类型），也不会产生线程安全问题，因为他们访问的是不同的对象，所以不存在互斥问题。

2) synchronized 同步块

synchronized 代码块类似于以下这种形式：

synchronized (lock){
    //访问共享可变资源
    ...
}

Copy after login

当在某个线程中执行这段代码块，该线程会获取对象lock的锁，从而使得其他线程无法同时访问该代码块。其中，lock 可以是 this，代表获取当前对象的锁，也可以是类中的一个属性，代表获取该属性的锁。特别地，实例同步方法与 synchronized(this)同步块是互斥的，因为它们锁的是同一个对象。但与 synchronized(非this)同步块是异步的，因为它们锁的是不同对象。

比如上面的insert()方法可以改成以下两种形式：

// this 监视器
class InsertData {
    private ArrayList<Integer> arrayList = new ArrayList<Integer>();

    public void insert(Thread thread){
        synchronized (this) {
            for(int i=0;i<100;i++){
                System.out.println(thread.getName()+"在插入数据"+i);
                arrayList.add(i);
            }
        }
    }
}

// 对象监视器
class InsertData {
    private ArrayList<Integer> arrayList = new ArrayList<Integer>();
    private Object object = new Object();

    public void insert(Thread thread){
        synchronized (object) {
            for(int i=0;i<100;i++){
                System.out.println(thread.getName()+"在插入数据"+i);
                arrayList.add(i);
            }
        }
    }
}

Copy after login

从上面代码可以看出，synchronized代码块比 synchronized方法的粒度更细一些，使用起来也灵活得多。因为也许一个方法中只有一部分代码只需要同步，如果此时对整个方法用synchronized进行同步，会影响程序执行效率。而使用synchronized代码块就可以避免这个问题，synchronized代码块可以实现只对需要同步的地方进行同步。

3) class 对象锁

特别地，每个类也会有一个锁，静态的 synchronized方法就是以Class对象作为锁。另外，它可以用来控制对 static 数据成员（static 数据成员不专属于任何一个对象，是类成员）的并发访问。并且，如果一个线程执行一个对象的非static synchronized 方法，另外一个线程需要执行这个对象所属类的 static synchronized 方法，也不会发生互斥现象。因为访问 static synchronized 方法占用的是类锁，而访问非 static synchronized 方法占用的是对象锁，所以不存在互斥现象。例如，

public class Test {

    public static void main(String[] args)  {
        final InsertData insertData = new InsertData();
        new Thread(){
            @Override
            public void run() {
                insertData.insert();
            }
        }.start(); 
        new Thread(){
            @Override
            public void run() {
                insertData.insert1();
            }
        }.start();
    }  
}

class InsertData { 

    // 非 static synchronized 方法
    public synchronized void insert(){
        System.out.println("执行insert");
        try {
            Thread.sleep(5000);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        System.out.println("执行insert完毕");
    }

    // static synchronized 方法
    public synchronized static void insert1() {
        System.out.println("执行insert1");
        System.out.println("执行insert1完毕");
    }
}/* Output: 
        执行insert
        执行insert1
        执行insert1完毕
        执行insert完毕
 *///:~

Copy after login

根据执行结果，我们可以看到第一个线程里面执行的是insert方法，不会导致第二个线程执行insert1方法发生阻塞现象。下面，我们看一下 synchronized 关键字到底做了什么事情，我们来反编译它的字节码看一下，下面这段代码反编译后的字节码为：

public class InsertData {
    private Object object = new Object();

    public void insert(Thread thread){
        synchronized (object) {}
    }

    public synchronized void insert1(Thread thread){}

    public void insert2(Thread thread){}
}

Copy after login

Java concurrent development-sample code of built-in lock Synchronized

从反编译获得的字节码可以看出，synchronized 代码块实际上多了 monitorenter 和 monitorexit 两条指令。 monitorenter指令执行时会让对象的锁计数加1，而monitorexit指令执行时会让对象的锁计数减1，其实这个与操作系统里面的PV操作很像，操作系统里面的PV操作就是用来控制多个进程对临界资源的访问。对于synchronized方法，执行中的线程识别该方法的 method_info 结构是否有 ACC_SYNCHRONIZED 标记设置，然后它自动获取对象的锁，调用方法，最后释放锁。如果有异常发生，线程自动释放锁。

有一点要注意：对于 synchronized方法或者 synchronized代码块，当出现异常时，JVM会自动释放当前线程占用的锁，因此不会由于异常导致出现死锁现象。

四. 可重入性

一般地，当某个线程请求一个由其他线程持有的锁时，发出请求的线程就会阻塞。然而，由于 Java 的内置锁是可重入的，因此如果某个线程试图获得一个已经由它自己持有的锁时，那么这个请求就会成功。可重入锁最大的作用是避免死锁。例如：

public class Test implements Runnable {

    // 可重入锁测试
    public synchronized void get() {
        System.out.println(Thread.currentThread().getName());
        set();
    }

    public synchronized void set() {
        System.out.println(Thread.currentThread().getName());
    }

    @Override
    public void run() {
        get();
    }

    public static void main(String[] args) {
        Test test = new Test();
        new Thread(test,"Thread-0").start();
        new Thread(test,"Thread-1").start();
        new Thread(test,"Thread-2").start();
    }
}/* Output: 
        Thread-1
        Thread-1
        Thread-2
        Thread-2
        Thread-0
        Thread-0
 *///:~

Copy after login

五. 注意事项

1). 内置锁与字符串常量

由于字符串常量池的原因，在大多数情况下，同步synchronized代码块都不使用 String 作为锁对象，而改用其他，比如 new Object() 实例化一个 Object 对象，因为它并不会被放入缓存中。看下面的例子：

//资源类
class Service {
    public void print(String stringParam) {
        try {
            synchronized (stringParam) {
                while (true) {
                    System.out.println(Thread.currentThread().getName());
                    Thread.sleep(1000);
                }
            }
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}

//线程A
class ThreadA extends Thread {
    private Service service;

    public ThreadA(Service service) {
        super();
        this.service = service;
    }

    @Override
    public void run() {
        service.print("AA");
    }
}

//线程B
class ThreadB extends Thread {
    private Service service;

    public ThreadB(Service service) {
        super();
        this.service = service;
    }

    @Override
    public void run() {
        service.print("AA");
    }
}

//测试
public class Run {
    public static void main(String[] args) {

        //临界资源
        Service service = new Service();

        //创建并启动线程A
        ThreadA a = new ThreadA(service);
        a.setName("A");
        a.start();

        //创建并启动线程B
        ThreadB b = new ThreadB(service);
        b.setName("B");
        b.start();

    }
}/* Output (死锁): 
        A
        A
        A
        A
        ...
 *///:~

Copy after login

出现上述结果就是因为 String 类型的参数都是 “AA”，两个线程持有相同的锁，所以线程B 始终得不到执行，造成死锁。进一步地，所谓死锁是指：

不同的线程都在等待根本不可能被释放的锁，从而导致所有的任务都无法继续完成。

b). 锁的是对象而非引用

在将任何数据类型作为同步锁时，需要注意的是，是否有多个线程将同时去竞争该锁对象：

1).若它们将同时竞争同一把锁，则这些线程之间就是同步的；
2).否则，这些线程之间就是异步的。

看下面的例子：

//资源类
class MyService {
    private String lock = "123";

    public void testMethod() {
        try {
            synchronized (lock) {
                System.out.println(Thread.currentThread().getName() + " begin "
                        + System.currentTimeMillis());
                lock = "456";
                Thread.sleep(2000);
                System.out.println(Thread.currentThread().getName() + "   end "
                        + System.currentTimeMillis());
            }
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
    }
}

//线程B
class ThreadB extends Thread {

    private MyService service;

    public ThreadB(MyService service) {
        super();
        this.service = service;
    }

    @Override
    public void run() {
        service.testMethod();
    }
}

//线程A
class ThreadA extends Thread {

    private MyService service;

    public ThreadA(MyService service) {
        super();
        this.service = service;
    }

    @Override
    public void run() {
        service.testMethod();
    }
}

//测试
public class Run1 {
    public static void main(String[] args) throws InterruptedException {

        //临界资源
        MyService service = new MyService();

        //线程A
        ThreadA a = new ThreadA(service);
        a.setName("A");

        //线程B
        ThreadB b = new ThreadB(service);
        b.setName("B");

        a.start();
        Thread.sleep(50);// 存在50毫秒
        b.start();
    }
}/* Output(循环): 
       A begin 1484319778766
       B begin 1484319778815
       A   end 1484319780766
       B   end 1484319780815
 *///:~

Copy after login

由上述结果可知，线程 A、B 是异步的。因为50毫秒过后，线程B 取得的锁对象是 “456”，而线程A 依然持有的锁对象是 “123”。所以，这两个线程是异步的。若将上述语句 “Thread.sleep(50);” 注释，则有：

//测试
public class Run1 {
    public static void main(String[] args) throws InterruptedException {

        //临界资源
        MyService service = new MyService();

        //线程A
        ThreadA a = new ThreadA(service);
        a.setName("A");

        //线程B
        ThreadB b = new ThreadB(service);
        b.setName("B");

        a.start();
        // Thread.sleep(50);// 存在50毫秒
        b.start();
    }
}/* Output(循环): 
       B begin 1484319952017
       B   end 1484319954018
       A begin 1484319954018
       A   end 1484319956019
 *///:~

Copy after login

由上述结果可知，线程 A、B 是同步的。因为线程 A、B 竞争的是同一个锁“123”，虽然先获得运行的线程将 lock 指向了对象“456”，但结果还是同步的。因为线程 A 和 B 共同争抢的锁对象是“123”，也就是说，锁的是对象而非引用。

六. 总结

用一句话来说，synchronized 内置锁是一种对象锁 (锁的是对象而非引用)，作用粒度是对象，可以用来实现对临界资源的同步互斥访问，是可重入的。特别地，对于临界资源有：

若该资源是静态的，即被 static 关键字修饰，那么访问它的方法必须是同步且是静态的，synchronized 块必须是 class锁；
若该资源是非静态的，即没有被 static 关键字修饰，那么访问它的方法必须是同步的，synchronized 块是实例对象锁；

实质上，关键字synchronized 主要包含两个特征：

互斥性：保证在同一时刻，只有一个线程可以执行某一个方法或某一个代码块；
可见性：保证线程工作内存中的变量与公共内存中的变量同步，使多线程读取共享变量时可以获得最新值的使用。

The above is the detailed content of Java concurrent development-sample code of built-in lock Synchronized. For more information, please follow other related articles on the PHP Chinese website!