Re: [PATCH] mm: change memcg->oom_group access with atomic operations

[Roman Gushchin <roman.gushchin@linux.dev>]

On Wed, Feb 22, 2023 at 01:00:00AM +0800, Martin Zhao wrote:
> On Tue, Feb 21, 2023 at 1:17 PM Roman Gushchin <roman.gushchin@linux.dev> wrote:
> >
> > > On Feb 20, 2023, at 3:06 PM, Shakeel Butt <shakeelb@google.com> wrote:
> > >
> > > On Mon, Feb 20, 2023 at 01:09:44PM -0800, Roman Gushchin wrote:
> > >>> On Mon, Feb 20, 2023 at 11:16:38PM +0800, Yue Zhao wrote:
> > >>> The knob for cgroup v2 memory controller: memory.oom.group
> > >>> will be read and written simultaneously by user space
> > >>> programs, thus we'd better change memcg->oom_group access
> > >>> with atomic operations to avoid concurrency problems.
> > >>>
> > >>> Signed-off-by: Yue Zhao <findns94@gmail.com>
> > >>
> > >> Hi Yue!
> > >>
> > >> I'm curious, have any seen any real issues which your patch is solving?
> > >> Can you, please, provide a bit more details.
> > >>
> > >
> > > IMHO such details are not needed. oom_group is being accessed
> > > concurrently and one of them can be a write access. At least
> > > READ_ONCE/WRITE_ONCE is needed here.
> >
> > Needed for what?
> >
> > I mean it’s obviously not a big deal to put READ_ONCE()/WRITE_ONCE() here, but I struggle to imagine a scenario when it will make any difference. IMHO it’s easier to justify a proper atomic operation here, even if it’s most likely an overkill.
> >
> > My question is very simple: the commit log mentions “… to avoid concurrency problems”, so I wonder what problems are these.
> 
> Thanks for your watching!
> This topic is found in code review by coincidence, so no real issues
> recorded for now. I checked other read/write callbacks about other knobs,
> most of them use READ_ONCE/WRITE_ONCE on the user setting variable.

Sorry, which knobs are you talking about? I actually don't see any user knobs
in mm/memcontrol.c which are using WRITE_ONCE(). I see some of them using
xchg(), but it's a different thing.

> Actually I am curious as well why this interface doesn't use
> READ_ONCE/WRITE_ONCE, is there any other synchronization mechanism I
> didn't notice yet?

Because nobody saw any issues with the current code?

And again if it's something that makes any automated verifiers/tooling unhappy,
I'm totally fine for fixing it, just let make it clear (and also fix the commit
title, which is not true).

Thanks!