TLDR;不要使用synchronized修饰符或synchronized(this){…}表达式but synchronized(myLock){…其中myLock是一个持有私有对象的最终实例字段。

在方法声明中使用synchronized修饰符与在方法主体中使用synchronized(..){}表达式的区别如下:

The synchronized modifier specified on the method's signature is visible in the generated JavaDoc, is programmatically determinable via reflection when testing a method's modifier for Modifier.SYNCHRONIZED, requires less typing and indention compared to synchronized(this) { .... }, and (depending on your IDE) is visible in the class outline and code completion, uses the this object as lock when declared on non-static method or the enclosing class when declared on a static method. The synchronized(...){...} expression allows you to only synchronize the execution of parts of a method's body, to be used within a constructor or a (static) initialization block, to choose the lock object which controls the synchronized access.

然而，使用synchronized修饰符或synchronized(…){…}使用this作为锁对象(如synchronized(this){…})，也有同样的缺点。两者都使用它自己的实例作为锁对象进行同步。这是很危险的，因为不仅对象本身，而且任何其他持有该对象引用的外部对象/代码也可以将其用作同步锁，这可能会产生严重的副作用(性能下降和死锁)。

因此，最佳实践是既不使用synchronized修饰符，也不使用synchronized(…)表达式作为锁对象，而是使用该对象的私有锁对象。例如:

public class MyService {
    private final lock = new Object();

    public void doThis() {
       synchronized(lock) {
          // do code that requires synchronous execution
        }
    }

    public void doThat() {
       synchronized(lock) {
          // do code that requires synchronous execution
        }
    }
}

您也可以使用多个锁对象，但是需要特别注意，以确保在嵌套使用时不会导致死锁。

public class MyService {
    private final lock1 = new Object();
    private final lock2 = new Object();

    public void doThis() {
       synchronized(lock1) {
          synchronized(lock2) {
              // code here is guaranteed not to be executes at the same time
              // as the synchronized code in doThat() and doMore().
          }
    }

    public void doThat() {
       synchronized(lock1) {
              // code here is guaranteed not to be executes at the same time
              // as the synchronized code in doThis().
              // doMore() may execute concurrently
        }
    }

    public void doMore() {
       synchronized(lock2) {
              // code here is guaranteed not to be executes at the same time
              // as the synchronized code in doThis().
              // doThat() may execute concurrently
        }
    }
}

同步的方法

优点:

您的IDE可以指示同步方法。 语法更加紧凑。 强制将同步块分割为单独的方法。

缺点:

与此同步，因此外部人员也可以与之同步。 将代码移到同步块之外更加困难。

同步块

优点:

允许为锁使用私有变量，从而将锁强制留在类内部。 同步块可以通过搜索变量的引用来找到。

缺点:

语法更复杂，因此使代码更难阅读。

就我个人而言，我更喜欢使用同步方法，类只关注需要同步的东西。这样的类应该尽可能小，所以应该很容易检查同步。其他人不需要关心同步。

同步的方法

同步方法有两个效果。 首先，当一个线程正在为一个对象执行同步方法时，所有为同一对象调用同步方法的其他线程将阻塞(暂停执行)，直到第一个线程处理完该对象。

其次，当同步方法退出时，它自动与同一对象的同步方法的任何后续调用建立happens-before关系。这保证了对对象状态的更改对于所有线程都是可见的。

注意，构造函数不能同步——在构造函数中使用synchronized关键字是一个语法错误。同步构造函数没有意义，因为只有创建对象的线程才能在构造对象时访问它。

同步语句

与同步方法不同，同步语句必须指定提供内在锁的对象:大多数情况下，我使用它来同步对列表或映射的访问，但我不想阻塞对对象的所有方法的访问。

Q:内在锁和同步 同步是围绕一个称为内在锁或监视器锁的内部实体构建的。(API规范通常将此实体简单地称为“监视器”)内在锁在同步的两个方面都发挥作用:强制独占访问对象的状态，并建立对可见性至关重要的happens-before关系。

Every object has an intrinsic lock associated with it. By convention, a thread that needs exclusive and consistent access to an object's fields has to acquire the object's intrinsic lock before accessing them, and then release the intrinsic lock when it's done with them. A thread is said to own the intrinsic lock between the time it has acquired the lock and released the lock. As long as a thread owns an intrinsic lock, no other thread can acquire the same lock. The other thread will block when it attempts to acquire the lock.

package test;

public class SynchTest implements Runnable {  
    private int c = 0;

    public static void main(String[] args) {
        new SynchTest().test();
    }

    public void test() {
        // Create the object with the run() method
        Runnable runnable = new SynchTest();
        Runnable runnable2 = new SynchTest();
        // Create the thread supplying it with the runnable object
        Thread thread = new Thread(runnable,"thread-1");
        Thread thread2 = new Thread(runnable,"thread-2");
//      Here the key point is passing same object, if you pass runnable2 for thread2,
//      then its not applicable for synchronization test and that wont give expected
//      output Synchronization method means "it is not possible for two invocations
//      of synchronized methods on the same object to interleave"

        // Start the thread
        thread.start();
        thread2.start();
    }

    public synchronized  void increment() {
        System.out.println("Begin thread " + Thread.currentThread().getName());
        System.out.println(this.hashCode() + "Value of C = " + c);
//      If we uncomment this for synchronized block, then the result would be different
//      synchronized(this) {
            for (int i = 0; i < 9999999; i++) {
                c += i;
            }
//      }
        System.out.println("End thread " + Thread.currentThread().getName());
    }

//    public synchronized void decrement() {
//        System.out.println("Decrement " + Thread.currentThread().getName());
//    }

    public int value() {
        return c;
    }

    @Override
    public void run() {
        this.increment();
    }
}

用同步方法，块和不同步交叉检查不同的输出。

来自Java规范摘要: http://www.cs.cornell.edu/andru/javaspec/17.doc.html

The synchronized statement (§14.17) computes a reference to an object; it then attempts to perform a lock action on that object and does not proceed further until the lock action has successfully completed. ... A synchronized method (§8.4.3.5) automatically performs a lock action when it is invoked; its body is not executed until the lock action has successfully completed. If the method is an instance method, it locks the lock associated with the instance for which it was invoked (that is, the object that will be known as this during execution of the body of the method). If the method is static, it locks the lock associated with the Class object that represents the class in which the method is defined. ...

基于这些描述，我想说以前的大多数答案都是正确的，同步方法可能对静态方法特别有用，否则您必须弄清楚如何获得“表示定义方法的类的Class对象”。

编辑:我原本以为这些是对实际Java规范的引用。澄清一下，本页只是对规范的总结/解释

通常在方法级别上使用锁是不礼貌的。为什么要通过锁定整个方法来锁定一段不能访问任何共享资源的代码呢?因为每个对象都有一个锁，所以可以创建虚拟对象来实现块级同步。 块级的效率更高，因为它不锁定整个方法。

这里有一些例子

方法级

class MethodLevel {

  //shared among threads
SharedResource x, y ;

public void synchronized method1() {
   //multiple threads can't access
}
public void synchronized method2() {
  //multiple threads can't access
}

 public void method3() {
  //not synchronized
  //multiple threads can access
 }
}

块级别

class BlockLevel {
  //shared among threads
  SharedResource x, y ;

  //dummy objects for locking
  Object xLock = new Object();
  Object yLock = new Object();

    public void method1() {
     synchronized(xLock){
    //access x here. thread safe
    }

    //do something here but don't use SharedResource x, y
    // because will not be thread-safe
     synchronized(xLock) {
       synchronized(yLock) {
      //access x,y here. thread safe
      }
     }

     //do something here but don't use SharedResource x, y
     //because will not be thread-safe
    }//end of method1
 }

(编辑)

对于像Vector和Hashtable这样的集合，当ArrayList或HashMap不同步时，它们是同步的，你需要设置synchronized关键字或调用Collections synchronized方法:

Map myMap = Collections.synchronizedMap (myMap); // single lock for the entire map
List myList = Collections.synchronizedList (myList); // single lock for the entire list

关于使用同步块的重要提示:小心你使用的锁对象!

上面user2277816的代码片段说明了这一点，将字符串字面值的引用用作锁定对象。 意识到字符串字面值在Java中是自动互缩的，您应该开始看到问题所在:在字面值“锁”上同步的每段代码都共享同一个锁!这很容易导致完全不相关的代码段发生死锁。

您需要注意的不仅仅是String对象。装箱的原语也是一种危险，因为autoboxing和valueOf方法可以重用相同的对象，这取决于值。

有关更多信息，请参阅: https://www.securecoding.cert.org/confluence/display/java/LCK01-J.+Do+not+synchronize+on+objects+that+may+be+reused