Re: [PATCH] mm: add `const` to lots of pointer parameters

[David Hildenbrand <david@redhat.com>]

On 28.08.25 14:40, Max Kellermann wrote:
>> On 28.08.25 14:24, Lorenzo Stoakes wrote:
>>> On Wed, Aug 27, 2025 at 09:22:33PM +0200, Max Kellermann wrote:
>>>> For improved const-correctness.
>>>
>>> 'const-correctness' in C is extremely weak.
> 
> "const correctness" is a commonly used term for this concept, and I
> find your arguments against const-correctness "extremely weak". I
> don't think discussing the benefits of const-correctness is worth the
> time; the concept is trivial enough, and if you don't already agree,
> there is no way I can convince you. Let's agree to disagree.
> 
> On Thu, Aug 28, 2025 at 2:29 PM David Hildenbrand <david@redhat.com> wrote:
>> I recall that Willy did some more targeted conversions of that kind in the past regarding folios.
> 
> And I found many more commits adding "const" to pointer parameters.
> That shows this kind of patch is acceptable in the Linux kernel.
> 
>> I would appreciate similarly doing it in logical chunks.
> 
> So you suggest splitting the patch into many?

Yes, for example focusing on a specific class of functions like Willy did.

That makes my life as a maintainer and reviewer easier, and allows for 
holding back some of these patches if we expect conflicts in that area soon.

/me stares helplessly at his overflowing inbox where #submitters is 
significantly larger than #reviewers, and the majority of patches are 
not in a form that would allow for any kinds of easy review.

-- 
Cheers

David / dhildenb