Re: [PATCH 4/3] mm: drop MMF_OOM_SKIP from exit_mmap

[Suren Baghdasaryan]

On Mon, Jan 3, 2022 at 1:16 PM Hugh Dickins <hughd@xxxxxxxxxx> wrote:
>
> On Mon, 3 Jan 2022, Michal Hocko wrote:
> > On Thu 30-12-21 09:29:40, Suren Baghdasaryan wrote:
> > > On Thu, Dec 30, 2021 at 12:24 AM Michal Hocko <mhocko@xxxxxxxx> wrote:
> > > >
> > > > It would be really great to make unlock_range oom reaper aware IMHO.
> > >
> > > What exactly do you envision? Say unlock_range() knows that it's
> > > racing with __oom_reap_task_mm() and that calling follow_page() is
> > > unsafe without locking, what should it do?
> >
> > My original plan was to make the page lock conditional and use
> > trylocking from the oom reaper (aka lockless context). It is OK to
> > simply bail out and leave some mlocked memory behind if there is a
> > contention on a specific page. The overall objective is to free as much
> > memory as possible, not all of it.
> >
> > IIRC Hugh was not a fan of this approach and he has mentioned that the
> > lock might not be even really needed and that the area would benefit
> > from a clean up rather than oom reaper specific hacks. I do tend to
> > agree with that. I just never managed to find any spare time for that
> > though and heavily mlocked oom victims tend to be really rare.
>
> I forget when that was, and what I had in mind at that time.
> But yes, by now I am very sure that munlocking needs a cleanup.
>
> And I do have that cleanup (against a much older tree), but never
> the time to rebase or post or shepherd it through N revisions.

How old was that tree?

>
> It was 22 files changed, 464 insertions, 706 deletions:
> which is too much to help with this immediate oom reaper question.
>
> I'd better not drive this discussion further off-course; but it pains
> me to see munlock_vma_pages obstructing, knowing there's a better way.
>
> I wonder: what if I were to steal time and promise to post a
> rebased series against 5.17-rc1 or rc2: not support it thereafter,
> but might there be someone to pick it up and shepherd it through?
> But there's no answer to that, without you seeing what it's like.

I would be interested in taking a look and see if it can be upstreamed
and supported without bugging you too much.
Thanks,
Suren.

>
> Hugh