线程锁（Mutex）互斥锁
以上演示的是GIL锁，GIL锁只能保证同一时刻，CPU上只有一条线程在运行，但不能保证同一时刻只有一条线程在修改数据，要想实现这个效果，我们要加上另一把锁，线程锁。示例如下：
import threading
import time
start_time = time.time()
def run(n):
lock.acquire() #获取线程锁对象
global num
# time.sleep(1)
num += 1
lock.release() #释放线程锁
lock = threading.Lock() #示例化一个线程锁对象
num = 0
t_l = [] #存放线程
for i in range(10):
t = threading.Thread(target=run, args=('t-%s' % i, ))
t.start()
t_l.append(t) #为了不阻塞其他线程的启动，先放在列表里，最后join
for t in t_l: #循环线程列表，执行所有线程
t.join()
print('All threads have ran down!', threading.current_thread(), threading.activeCount())
time.sleep(0)
#current_thread显示当前线程
#activeCount显示活动的线程数
print('Num: ', num)
print('Total Cost: ', time.time() - start_time)
运行结果
All threads have ran down! <_MainThread(MainThread, started 7268)> 1 Num: 10 Total Cost: 0.0026278495788574
这次结果就对了，但此时实际上是串行运行。
递归锁
说白了就是在一个大锁中还要再包含子锁
示例
import threading
def run1():
print("grab the first part data")
lock.acquire() #第二把锁
global num #声明全局变量，让全局变量num可以在run1()内进行修改
num += 1
lock.release() #run1()执行完毕，释放第二把锁
return num #返回run3()num的值
def run2():
print("grab the second part data")
lock.acquire()
global num2
num2 += 2
lock.release()
return num2
def run3():
lock.acquire() #加递归锁，第一把锁
res = run1() #执行run1()函数
print('--------between run1 and run2-----')
res2 = run2()
lock.release() #释放第一把锁
print(res, res2)
if __name__ == '__main__':
num, num2 = 0, 0
lock = threading.RLock() #RLock递归锁，如果是Lock，程序会陷入死循环
for i in range(10):
t = threading.Thread(target=run3) #每循环一次，启动一条线程执行run3()
t.start()
while threading.active_count() != 1: #循环条件，活动线程数为1时循环结束
print('From Primary Thread: Thread left: ', threading.active_count())
else:
print('All threads are done!')
print('Num1: %d, Num2: %d' %(num, num2))
输出结果
grab the first part data --------between run1 and run2----- grab the second part data 1 2 grab the first part data --------between run1 and run2----- grab the second part data 2 4 grab the first part data --------between run1 and run2----- grab the second part data 3 6 grab the first part data --------between run1 and run2----- grab the second part data 4 grab the first part data --------between run1 and run2----- 8grab the second part data 5grab the first part data 10--------between run1 and run2----- grab the second part data 6 grab the first part data12 --------between run1 and run2----- grab the second part data 7grab the first part data 14From Primary Thread: Thread left: --------between run1 and run2----- 5grab the second part data From Primary Thread: Thread left: 4 From Primary Thread: Thread left: 8grab the first part data4 16--------between run1 and run2-----From Primary Thread: Thread left: grab the second part data4 9grab the first part data From Primary Thread: Thread left: 18 --------between run1 and run2----- 3 grab the second part dataFrom Primary Thread: Thread left: 102 From Primary Thread: Thread left: 202 All threads are done! Num1: 10, Num2: 20
上面示例中
lock = threading.RLock()
这里如果使用Lock，而不是RLock，程序会陷入死循环。