Re: [PATCH v4 00/12] mm: establish const-correctness for pointer parameters

[David Hildenbrand <david@redhat.com>]

On 01.09.25 12:54, Max Kellermann wrote:
> On Mon, Sep 1, 2025 at 12:43 PM David Hildenbrand <david@redhat.com> wrote:
>> Max, I think this series here is valuable, and you can see that from the
>> engagement from reviewers (this is a *good* thing, I sometimes wish I
>> would get feedback that would help me improve my submissions).
>>
>> So if you don't want to follow-up on this series to polish the patch
>> descriptions etc,, let me now and I (or someone else around here) can
>> drag it over the finishing line.
> 
> Thanks David - I do want to finish this, if there is a constructive
> path ahead. I know what you want, but I'm not so sure about the
> others.

I think we primarily want to briefly explain the what, the why, and why it is okay.

For getter/test functions the "why it is okay" it's trivial -- test function.
Personally, I would not spell out the individual functions in that case, as
long as they logically belong together (like "shmem test functions"
describe what you did in that patch).

For anything beyond that people likely expect a different reasoning.

For example the following change:

-static inline void folio_migrate_refs(struct folio *new, struct folio *old)
+static inline void folio_migrate_refs(struct folio *const new,
+				      const struct folio *const old)

Adds two "const" ways of doing things. As a reviewer, seeing something like that
buried in a patch raises questionmarks.

-- 
Cheers

David / dhildenb