Re: [RFC PATCH 0/6] mm: swap support for hugetlb folios

[David Hildenbrand (Arm)]

On 9/29/26 11:26, Zongkun Lei wrote:
> On Tue, 29 Sep 2026 10:02:11 +0200, David Hildenbrand wrote:
>> Hugetlb is effectively in a feature freeze. There is no interest in
>> maintaining even more of the special hugetlb sauce, sorry. :/
> 
> Understood -- thank you for the direct answer.
> 
> My mistake was misreading the subsystem's direction.  I had read the
> earlier list discussions and concluded that the red line was touching
> hugetlb's reservation semantics, so I designed the whole series to
> keep them intact -- user-driven only, no LRU, no reclaim hooks.  I
> see now that the actual position is a feature freeze: not "new
> features are welcome if reservations stay sacred", but no new
> features at all.
> 
> Going forward I will stick to work that reduces hugetlb's divergence
> from generic code rather than adding to it, and I will float ideas
> on the list before putting serious effort into anything big.

Yes, that's usually the best approach: if you intend to do a rather invasive
and/or likely controversial work, better reach out to the community+maintainers
early.

Most of us prefer a well written problem statement + high-level design idea, as
it rather focuses on requirements + exploring solution space vs. a concrete
solution.

-- 
Cheers,

David