Describe the issue:

All arrays from other libraries imported with from_dlpack come out as non-writable.

#28599 fixed a similar bug, but only for passing a numpy array into from_dlpack. The issue still exists when passing an external writable array, such as torch.tensor. Note that pytorch actually dos not support non-writable tensors.

Reproduce the code example:

import torch
import numpy as np
x = torch.arange(1)
y = np.from_dlpack(x)
y[0] = 1
y.flags.writeable = True

Error message:

Traceback (most recent call last):
  File "<python-input-4>", line 1, in <module>
    y[0] = 1
    ~^^^
ValueError: assignment destination is read-only

Traceback (most recent call last):
  File "<python-input-5>", line 1, in <module>
    y.flags.writeable = True
    ^^^^^^^^^^^^^^^^^
ValueError: cannot set WRITEABLE flag to True of this array

Python and NumPy Versions:

2.3.2
3.13.5 (main, Jun 21 2025, 09:35:00) [GCC 15.1.1 20250425]

Runtime Environment:

[{'numpy_version': '2.3.2',
'python': '3.13.5 (main, Jun 21 2025, 09:35:00) [GCC 15.1.1 20250425]',
'uname': uname_result(system='Linux', node='node', release='6.15.5-arch1-1', version='#1 SMP PREEMPT_DYNAMIC Sun, 06 Jul 2025 11:14:36 +0000', machine='x86_64')},
{'simd_extensions': {'baseline': ['SSE', 'SSE2', 'SSE3'],
'found': ['SSSE3',
'SSE41',
'POPCNT',
'SSE42',
'AVX',
'F16C',
'FMA3',
'AVX2'],
'not_found': ['AVX512F',
'AVX512CD',
'AVX512_KNL',
'AVX512_KNM',
'AVX512_SKX',
'AVX512_CLX',
'AVX512_CNL',
'AVX512_ICL',
'AVX512_SPR']}}]

Context for the issue:

This issue prevents the use of numpy for writing into aliased data from other tensor libraries.