在Python中将multiprocessing.Array转换为ctypes.c_void_p

Question

I need to convert a multiprocessing.Array into a c_void_p . Currently this what I have:

shared_mem_bfr = multiprocessing.Array(c_uint8, img_width*img_height*3)
npyarr = np.frombuffer(shared_mem_bfr.get_obj(), c_uint8)
memory_buffer = npyarr.ctypes.data_as(c_void_p)

What bothers me is that I first convert it into a Numpy array and then into a c_void_p. Is there any way to remove the intermediate step?

Answer 1

Use ctypes.addressof :

import numpy as np
import multiprocessing as mp
import ctypes as c

shared_mem_bfr = mp.Array(c.c_uint8, 100)
print(c.c_void_p(c.addressof(shared_mem_bfr.get_obj())))

npyarr = np.frombuffer(shared_mem_bfr.get_obj(), c.c_uint8)
memory_buffer = npyarr.ctypes.data_as(c.c_void_p)
print(memory_buffer)

Output:

c_void_p(1552895115472)
c_void_p(1552895115472)