iOS数据锁

简介

当一个线程访问数据时，而其他数据不能进行访问，保证线程安全或者可以理解为执行多线程，对于共享资源访问时保证互斥的要求

文章

不再安全的 OSSpinLock

iOS开发中的11种锁以及性能对比

iOS 十种线程锁

iOS多线程篇-NSThread-synchronized(互斥锁)

分类

1.自旋锁：是用于多线程同步的一种锁，线程反复检查锁变量是否可用（一直进行do while忙等）

2.信号量：可以有更多的取值空间，用来实现更加复杂的同步，而不单单是线程间互斥

3.互斥锁：是一种用于多线程编程中，防止两条线程同时对同一公共资源（比如全局变量）进行读写的机制

4.条件锁：当进程的某些资源要求不满足时就进入休眠，也就是锁住了。当资源被分配到了，条件锁打开，进程继续运行

5.递归锁：在同一线程上该锁是可重入的，对于不同线程则相当于普通的互斥锁

6.读写锁：对共享资源的访问者划分成读者和写者，读者只对共享资源进行读访问，写者则需要对共享资源进行写操作

OSSpinLock

- (void)osspinlock:(int)count{
    NSTimeInterval begin, end;
    OSSpinLock lock = OS_SPINLOCK_INIT;
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        OSSpinLockLock(&lock);
        OSSpinLockUnlock(&lock);
    }
    end = CACurrentMediaTime();
    printf("OSSpinLock:               %8.2f msn", (end - begin) * 1000);
}

os_unfair_lock（iOS10之后替代OSSPinLock的锁，解决了优先级反转的问题）

- (void)os_unfair_lock:(int)count{
    if (@available(iOS 10.0, *)) {
        NSTimeInterval begin, end;
        os_unfair_lock_t unfairLock;
        unfairLock = &(OS_UNFAIR_LOCK_INIT);
        begin = CACurrentMediaTime();
        for (int i = 0; i < count; i++) {
            os_unfair_lock_lock(unfairLock);
            os_unfair_lock_unlock(unfairLock);
        }
        end = CACurrentMediaTime();
        printf("os_unfair_lock:           %8.2f msn", (end - begin) * 1000);
    }
}

信号量

dispatch_semaphore

- (void)dispatch_semaphore:(int)count{
    NSTimeInterval begin, end;
    dispatch_semaphore_t lock =  dispatch_semaphore_create(1);
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        dispatch_semaphore_wait(lock, DISPATCH_TIME_FOREVER);
        dispatch_semaphore_signal(lock);
    }
    end = CACurrentMediaTime();
    printf("dispatch_semaphore:       %8.2f msn", (end - begin) * 1000);
}

互斥锁

NSLock

• NSLock是最常使用的互斥锁，遵循NSLocking协议，通过lock和unlock来完成锁定和解锁*

- (void)nslock:(int)count{
    NSTimeInterval begin, end;
    NSLock *lock = [NSLock new];
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        [lock lock];
        [lock unlock];
    }
    end = CACurrentMediaTime();
    printf("NSLock:                   %8.2f msn", (end - begin) * 1000);
    
}

pthread_mutex

- (void)pthread_mutex:(int)count{
    NSTimeInterval begin, end;
    pthread_mutex_t lock;
    pthread_mutex_init(&lock,  大专栏  iOS数据锁NULL);
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        pthread_mutex_lock(&lock);
        pthread_mutex_unlock(&lock);
    }
    end = CACurrentMediaTime();
    pthread_mutex_destroy(&lock);
    printf("pthread_mutex:            %8.2f msn", (end - begin) * 1000);
    
}

synchronized

- (void)synchronized:(int)count{
    NSTimeInterval begin, end;
    NSObject *lock = [NSObject new];
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        @synchronized(lock) {}
    }
    end = CACurrentMediaTime();
    printf("@synchronized:            %8.2f msn", (end - begin) * 1000);
    
}

条件锁

NSCondition

- (void)_NSCondition:(int)count{
    NSTimeInterval begin, end;
    NSCondition *lock = [NSCondition new];
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        [lock lock];
        [lock unlock];
    }
    end = CACurrentMediaTime();
    printf("NSCondition:              %8.2f msn", (end - begin) * 1000);
}

NSConditionLock

- (void)_NSConditionLock:(int)count{
    NSTimeInterval begin, end;
    NSConditionLock *lock = [[NSConditionLock alloc] initWithCondition:1];
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        [lock lock];
        [lock unlock];
    }
    end = CACurrentMediaTime();
    printf("NSConditionLock:          %8.2f msn", (end - begin) * 1000);
    
}

递归锁

NSRecursiveLock

- (void)_NSRecursiveLock:(int)count{
    NSTimeInterval begin, end;
    NSRecursiveLock *lock = [NSRecursiveLock new];
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        [lock lock];
        [lock unlock];
    }
    end = CACurrentMediaTime();
    printf("NSRecursiveLock:          %8.2f msn", (end - begin) * 1000);
}

pthread_mutex_recursive

- (void)pthread_mutex_recursive:(int)count{
    NSTimeInterval begin, end;
    pthread_mutex_t lock;
    pthread_mutexattr_t attr;
    pthread_mutexattr_init(&attr);
    pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE);
    pthread_mutex_init(&lock, &attr);
    pthread_mutexattr_destroy(&attr);
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        pthread_mutex_lock(&lock);
        pthread_mutex_unlock(&lock);
    }
    end = CACurrentMediaTime();
    pthread_mutex_destroy(&lock);
    printf("pthread_mutex(recursive): %8.2f msn", (end - begin) * 1000);

}

读写锁

pthread_rwlock

- (void)pthread_rwlock:(int)count{
    NSTimeInterval begin, end;
    pthread_rwlock_t rwlock;
    pthread_rwlock_init(&rwlock,NULL);
    begin = CACurrentMediaTime();
    for (int i = 0; i < count; i++) {
        pthread_rwlock_rdlock(&rwlock);
        pthread_rwlock_unlock(&rwlock);
    }
    end = CACurrentMediaTime();
    printf("pthread_rwlock:           %8.2f msn", (end - begin) * 1000);
}

总结

通过各种常见锁介绍和性能评测，可以看出要是没有优先级反转的问题的话，osspinlock为最优，其次就是dispatch_semaphore，dispatch_semaphore和os_unfair_lock差距很小，然后就是pthread_mutex。

代码下载

TJLock

---