具有独立共享变量的同步方法

Question

I recently asked about if synchronized methods in Java shared the same lock key. The answer was yes. But the next thought I had was that this pattern is very restrictive and can cause unnecessary delays. For example suppose we have 2 synchronized methods and each of them process different and independent shared variables. What's the point of locking both methods at the same time if the processing of one variable does not affects the other? Is there a way to deal with this situation?

Answer 1

Suppose you have this class:

class C {
    public synchronized void foo(...) { ... }
    public synchronized void bar(...) { ... }
    ...
}

And suppose you have:

final C c_0 = new C(...);
final C c_1 = new C(...);

The fact that foo() and bar() both are synchronized does not prevent one thread from calling c_0.foo(...) while another thread simultaneously calls c_1.bar(...) . In fact, it does not even prevent one thread from calling c_0.foo(...) while another thread calls c_1.foo(...) .

A so-called "synchronized method"

synchronized void foo(...) { ... }

Really is just syntactic sugar for this:

void foo( ... ) {
    synchronized(this) {
        ...
    }
}

So, if one thread calls c_0.foo() it's "synchronizing" on the c_0 instance, and if another thread calls c_1.foo() , that thread is synchronizing on the c_1 instance--a different instance.

The synchronized keyword on the foo and bar methods only matters if two different threads both try to operate on the same instance at the same time.