Python多重处理Lock.acquire（）即使释放了Lock也会一直阻塞，甚至返回True

Question

我有以下一段代码。

import multiprocessing as mp
import time

def f(l,a):
    print("Acquiring lock in f")
    l.acquire()
    time.sleep(2)
    try:
        a.put([2,3])
        a.get()
        print("In f a is: ", a.get())
    finally:
        l.release()
        print("Just released lock in f")

def g(l,a):
    print("Acquiring lock in g")
    time.sleep(1)
    print(l.acquire())
    try:
        print("In g a is: ",a.get())
    finally:
        print("In finally")
        l.release()

if __name__ == '__main__':
    l = mp.RLock()
    data = mp.Queue()
    data.put([1])
    p1 = mp.Process(target=f, args=(l,data))
    p2 = mp.Process(target=g, args=(l,data))
    p1.start()
    p2.start()
    p2.join()
    p1.join()
    l.acquire()

输出显示g（）中的Lock.acquire（）不允许进一步执行，即使在其他地方释放该锁时它返回True。知道这里有什么问题吗？我使用了Lock而不是RLock，没有打印出yield（）返回值的结果，没有使用try和finally块，但是总会得到相同的结果。

$ python first.py 
Acquiring lock in f
Acquiring lock in g
In f a is:  [2, 3]
Just released lock in finally
True