Re: [PATCH RFC v3 01/19] mm: thread user_addr through page allocator for cache-friendly zeroing

[David Hildenbrand (Arm)]

On 4/23/26 16:46, Michael S. Tsirkin wrote:
> On Thu, Apr 23, 2026 at 04:13:50PM +0200, David Hildenbrand (Arm) wrote:
>> But really, that hugetlb code is rather messy. I'd vote for leaving hugetlb
>> alone on a v1, and focusing on non-hugetlb first.
> 
> I just dislike it when things are non orthogonal.
> People are used to: hugetlb = same perf as THP but more predictable at the cost
> of being harder to use and using more resources.
> Here, suddenly, we have an optimization but only for THP.

Note that we also didn't care about user_alloc_needs_zeroing() with hugetlb so far.

It's not that it cannot be done with hugetlb, it's just a question about how it
can be done more cleanly; and that requires more work to clean the hugetlb part up.

Best done separately.

-- 
Cheers,

David