java 中锁你知道哪些?请手写一个自旋锁?
1 公平和非公平锁
1.1 是什么
- 公平锁:是指多个线程按照申请的顺序来获取锁
- 非公平锁:是指多个线程获取锁的顺序并不是按照申请锁的顺序,有可能后申请的线程比先申请的线程优先获取锁,在高并发的情况下,可能会造成优先级翻转或者饥饿现象
1.2 两者区别
- 公平锁:在并发环境中,每一个线程在获取锁时会先查看此锁维护的等待队列,如果为空,或者当前线程是等待队列的第一个就占有锁,否则就会加入到等待队列中,以后会按照 FIFO 的规则获取锁
- 非公平锁:一上来就尝试占有锁,如果失败再进行排队
2 可重入锁和不可重入锁
2.1 是什么
- 可重入锁:同一个线程在外层方法获取锁之后,进入内层方法会自动获得该锁
- 不可重入锁: 所谓不可重入锁,即若当前线程执行某个方法已经获取了该锁,那么在方法中尝试再次获取锁时,就会因获取不到而被阻塞
2.2 代码实现可重入锁
public class ReentrantLock {
boolean isLocked = false;
Thread lockedBy = null;
int lockedCount = 0;
public synchronized void lock() throws InterruptedException {
Thread thread = Thread.currentThread();
while (isLocked && lockedBy != thread) {
wait();
}
isLocked = true;
lockedCount++;
lockedBy = thread;
}
public synchronized void unlock() {
if (Thread.currentThread() == lockedBy) {
lockedCount--;
if (lockedCount == 0) {
isLocked = false;
lockedBy = null;
notify();
}
}
}
}
- 测试
class Data{
ReentrantLock lock = new ReentrantLock();
public void method1() {
lock.lock();
System.out.println(Thread.currentThread().getName() + "\t method1 in");
method2();
System.out.println(Thread.currentThread().getName() + "\t method1 out");
lock.unlock();
}
public void method2() {
lock.lock();
System.out.println(Thread.currentThread().getName() + "\t method2 in");
System.out.println(Thread.currentThread().getName() + "\t method2 out");
lock.unlock();
}
}
public class ReentrantLockDemo {
public static void main(String[] args) {
Data data = new Data();
ExecutorService executorService = Executors.newFixedThreadPool(2);
for(int i = 0; i < 2; i++) {
executorService.execute(()->{
data.method1();
});
}
}
}
我们设计两个线程调用 method1() 方法,第一个线程调用 method1() 方法获取锁,进入 lock() 方法,由于初始 lockedBy 是 null,所以不会进入 while 而挂起当前线程,而是增加 lockedCount 并记录 lockBy 为第一个线程。接着第一个线程进入 method2() 方法,由于同一进程,所以不会进入 while 而挂起,接着增加 lockedCount,当第二个线程尝试lock,由于 isLocked=true,所以他不会获取该锁,直到第一个线程调用两次 unlock() 将 lockCount 递减为0,才将标记为 isLocked 设置为 false。
2.3 代码实现不可重入锁
class NotReentrantLock{
boolean locked = false;
Thread lockedBy = null;
public synchronized void lock() {
Thread thread = Thread.currentThread();
while(locked == true) {
try {
this.wait();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
locked = true;
lockedBy = thread;
}
public synchronized void unlock() {
Thread thread = Thread.currentThread();
if(lockedBy == thread) {
locked = false;
lockedBy = null;
this.notify();
}
}
}
- 测试
class Data{
NotReentrantLock lock = new NotReentrantLock();
public void method1() {
lock.lock();
System.out.println(Thread.currentThread().getName() + "\t method1 in");
method2();
System.out.println(Thread.currentThread().getName() + "\t method1 out");
lock.unlock();
}
public void method2() {
lock.lock();
System.out.println(Thread.currentThread().getName() + "\t method2 in");
System.out.println(Thread.currentThread().getName() + "\t method2 out");
lock.unlock();
}
}
public class NotReentrantLockDemo {
public static void main(String[] args) {
Data data = new Data();
data.method1();
}
}
当前线程执行method1方法首先获取lock,接下来执行method2()方法就无法获取锁。这个例子很好的说明了不可重入锁。
2.4 synchronized 和 ReentrantLock 都是可重入锁
- synchronzied
public class SynchronziedDemo {
private synchronized void print() {
doAdd();
}
private synchronized void doAdd() {
System.out.println("doAdd...");
}
public static void main(String[] args) {
SynchronziedDemo synchronziedDemo = new SynchronziedDemo();
synchronziedDemo.print(); // doAdd...
}
}
上面可以说明 synchronized 是可重入锁。
- ReentrantLock
public class ReentrantLockDemo {
private Lock lock = new ReentrantLock();
private void print() {
lock.lock();
doAdd();
lock.unlock();
}
private void doAdd() {
lock.lock();
lock.lock();
System.out.println("doAdd...");
lock.unlock();
lock.unlock();
}
public static void main(String[] args) {
ReentrantLockDemo reentrantLockDemo = new ReentrantLockDemo();
reentrantLockDemo.print();
}
}
上面例子可以说明 ReentrantLock 是可重入锁,而且在 doAdd 方法中加两次锁和解两次锁也可以。
2.5 synchronized 和 ReentrantLock 有什么区别?
- 原始结构
synchronized
是关键字属于JVM
层面,反映在字节码上是monitorenter
和monitorexit
,其底层是通过monitor
对象来完成,其实wait/notify
等方法也是依赖monitor
对象,只有在同步块或方法中才能调用wait/notify
等方法。ReentrantLock
是具体类(java.util.concurrent.locks.ReentrantLock
),是api
层面的锁。
- 使用方法
synchronized
不需要用户手动去释放锁,当synchronized
代码执行完后系统会自动让线程释放对锁的占用。ReentrantLock
则需要用户手动的释放锁,若没有主动释放锁,可能导致出现死锁的现象,lock()
和unlock()
方法需要配合try/finally
语句来完成。
- 等待是否可中断
synchronized
不可中断,除非抛出异常或者正常运行完成。ReentrantLock
可中断,设置超时方法tryLock(long timeout, TimeUnit unit),lockInterruptibly()
放代码块中,调用interrupt()
方法可中断。
- 加锁是否公平
synchronized
非公平锁ReentrantLock
默认非公平锁,构造方法中可以传入boolean
值,true
为公平锁,false
为非公平锁。
- 锁可以绑定多个
Condition
synchronized
没有Condition
。ReentrantLock
用来实现分组唤醒需要唤醒的线程们,可以精确唤醒,而不是像synchronized
要么随机唤醒一个线程要么唤醒全部线程。
2.6 ReentrantLock的lock()、tryLock()、lockInterruptibly()有什么区别?
Lock接口的线程请求锁的几个方法:
- lock()忽视interrupt(), 拿不到锁就一直阻塞:
@Test
public void test1() {
ReentrantLock lock = new ReentrantLock();
lock.lock();
System.out.println(Thread.currentThread().getName() + " get lock");
Thread t = new Thread(()->{
lock.lock();
System.out.println(Thread.currentThread().getName() + " get lock");
lock.unlock();
},"t");
t.start();
System.out.println("t start");
try {
TimeUnit.SECONDS.sleep(1);
} catch (InterruptedException e) {
e.printStackTrace();
}
t.interrupt();
System.out.println("t was interrupted");
try {
t.join();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
输出:
main get lock
t start
t was interrupted
之后一直阻塞
- lockInterruptibly()是对lock()的改进,和lock()一样拿不到锁会一直阻塞,但是,阻塞期间如果别的进程调用此进程的interrupt()方法,此线程会被唤醒并被要求处理InterruptedException
@Test
public void test2() {
ReentrantLock lock = new ReentrantLock();
lock.lock();
System.out.println(Thread.currentThread().getName() + " get lock");
Thread t = new Thread(()->{
try {
lock.lockInterruptibly();
} catch (InterruptedException e) {
System.out.println(Thread.currentThread().getName() + " handle the interrupt");
}
System.out.println(Thread.currentThread().getName() + (lock.isHeldByCurrentThread() ? " get lock" : " don't get lock"));
try {
lock.unlock();
} catch (IllegalMonitorStateException e) {
}
},"t");
t.start();
System.out.println("t start");
try {
TimeUnit.SECONDS.sleep(1);
} catch (InterruptedException e) {
e.printStackTrace();
}
t.interrupt();
System.out.println("t was interrupted");
try {
t.join();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
输出:
main get lock
t start
t was interrupted
t handle the interrupte
t don't get lock
- tryLock(),马上返回,拿到lock就返回true,不然返回false。
@Test
public void test3() {
ReentrantLock lock = new ReentrantLock();
lock.lock();
System.out.println(Thread.currentThread().getName() + " get lock");
Thread t = new Thread(()->{
lock.tryLock();
System.out.println(Thread.currentThread().getName() + (lock.isHeldByCurrentThread() ? " get lock" : " don't get lock"));
try {
lock.unlock();
} catch (IllegalMonitorStateException e) {
}
},"t");
t.start();
System.out.println("t start");
try {
TimeUnit.SECONDS.sleep(1);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
输出:
main get lock
t start
t don't get lock
- 带时间限制的tryLock(),拿不到lock,就等一段时间,超时返回false。等待期间如果别的进程调用此进程的interrupt()方法,此线程会被唤醒并被要求处理InterruptedException
@Test
public void test4() {
ReentrantLock lock = new ReentrantLock();
lock.lock();
System.out.println(Thread.currentThread().getName() + " get lock");
Thread t = new Thread(()->{
try {
lock.tryLock(5,TimeUnit.SECONDS);
} catch (InterruptedException e) {
System.out.println(Thread.currentThread().getName() + " handle the interrupte");
}
System.out.println(Thread.currentThread().getName() + (lock.isHeldByCurrentThread() ? " get lock" : " don't get lock"));
try {
lock.unlock();
} catch (IllegalMonitorStateException e) {
}
},"t");
t.start();
System.out.println("t start");
try {
TimeUnit.SECONDS.sleep(1);
} catch (InterruptedException e) {
e.printStackTrace();
}
t.interrupt();
System.out.println("t was interrupted");
try {
t.join();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
输出:
main get lock
t start
t was interrupted
t handle the interrupte
t don't get lock
3 自旋锁
3.1 是什么
- 是指尝试获取锁的线程不会立即堵塞,而是采用循环的方式去尝试获取锁,这样的好处是减少线程上下文切换的消耗,缺点就是循环会消耗 CPU。
3.2 手动实现自旋锁
public class SpinLock {
private AtomicReference<Thread> atomicReference = new AtomicReference<>();
private void lock () {
while (!atomicReference.compareAndSet(null, Thread.currentThread())) {
// loop
}
System.out.println(Thread.currentThread() + " get lock...");
}
private void unlock() {
while (!atomicReference.compareAndSet(Thread.currentThread(), null)) {
// loop
}
System.out.println(thread + " unlock...");
}
public static void main(String[] args) throws InterruptedException {
SpinLock spinLock = new SpinLock();
new Thread(() -> {
spinLock.lock();
try {
Thread.sleep(3000);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println("hahaha");
spinLock.unlock();
}).start();
Thread.sleep(1);
new Thread(() -> {
spinLock.lock();
System.out.println("hehehe");
spinLock.unlock();
}).start();
}
}
- 输出:
Thread[Thread-0,5,main] get lock...
hahaha
Thread[Thread-0,5,main] unlock...
Thread[Thread-1,5,main] get lock...
hehehe
Thread[Thread-1,5,main] unlock...
获取锁的时候,如果原子引用为空就获取锁,不为空表示有人获取了锁,就循环等待。
4 独占锁(写锁)/共享锁(读锁)
4.1 是什么
- 独占锁:指该锁一次只能被一个线程持有
- 共享锁:该锁可以被多个线程持有
- ReentrantLock 和 synchronized 都是独占锁
- 对于ReentrantReadWriteLock,其读锁是共享锁而写锁是独占锁。读锁的共享可保证并发读是非常高效的,读写、写读和写写的过程是互斥的。
4.2 读写锁例子
public class MyCache {
private volatile Map<String, Object> map = new HashMap<>();
private ReentrantReadWriteLock lock = new ReentrantReadWriteLock();
WriteLock writeLock = lock.writeLock();
ReadLock readLock = lock.readLock();
public void put(String key, Object value) {
try {
writeLock.lock();
System.out.println(Thread.currentThread().getName() + " 正在写入...");
try {
Thread.sleep(1000);
} catch (InterruptedException e) {
e.printStackTrace();
}
map.put(key, value);
System.out.println(Thread.currentThread().getName() + " 写入完成,写入结果是 " + value);
} finally {
writeLock.unlock();
}
}
public void get(String key) {
try {
readLock.lock();
System.out.println(Thread.currentThread().getName() + " 正在读...");
try {
Thread.sleep(1000);
} catch (InterruptedException e) {
e.printStackTrace();
}
Object res = map.get(key);
System.out.println(Thread.currentThread().getName() + " 读取完成,读取结果是 " + res);
} finally {
readLock.unlock();
}
}
}
- 测试
public class ReadWriteLockDemo {
public static void main(String[] args) {
MyCache cache = new MyCache();
for (int i = 0; i < 5; i++) {
final int temp = i;
new Thread(() -> {
cache.put(temp + "", temp + "");
}).start();
}
for (int i = 0; i < 5; i++) {
final int temp = i;
new Thread(() -> {
cache.get(temp + "");
}).start();
}
}
}
- 输出结果
Thread-0 正在写入...
Thread-0 写入完成,写入结果是 0
Thread-1 正在写入...
Thread-1 写入完成,写入结果是 1
Thread-2 正在写入...
Thread-2 写入完成,写入结果是 2
Thread-3 正在写入...
Thread-3 写入完成,写入结果是 3
Thread-4 正在写入...
Thread-4 写入完成,写入结果是 4
Thread-5 正在读...
Thread-7 正在读...
Thread-8 正在读...
Thread-6 正在读...
Thread-9 正在读...
Thread-5 读取完成,读取结果是 0
Thread-7 读取完成,读取结果是 2
Thread-8 读取完成,读取结果是 3
Thread-6 读取完成,读取结果是 1
Thread-9 读取完成,读取结果是 4
能保证 **读写** 、 **写读** 和 **写写** 的过程是独享的, **读读** 的时候是共享的。