Re: [PATCH mm-unstable v2] mm/page_alloc: keep track of free highatomic

[Yu Zhao]

On Sun, Oct 27, 2024 at 1:40 PM Vlastimil Babka <vbabka@xxxxxxx> wrote:
>
> On 10/26/24 06:40, Yu Zhao wrote:
> > On Fri, Oct 25, 2024 at 10:24 PM Andrew Morton
> > <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> >>
> >> Is a -stable backport needed?
> >>
> >> If so, is a Fixes: target identifiable?
> >
> > The code has been there for many years, and we only recently noticed
> > the problem from Link's repro. So it doesn't look like a stable
> > material.
>
> The stable backportability would be limited as there's a prerequisity on the
> "mm: page_alloc: freelist migratetype hygiene" patchset from half a year ago
> - patch applicability wise and importantly functionality wise.

I don't see any dependency to that series.

> Otherwise the
> counter could drift easily.

Care to elaborate how that could happen?

> However we could perhaps mark it for a 6.12 stable backport as that's
> presumably the upcoming LTS.

Right.