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

[Muchun Song]

> On Feb 21, 2023, at 13:17, Roman Gushchin <roman.gushchin@xxxxxxxxx> wrote:
> 
>> On Feb 20, 2023, at 3:06 PM, Shakeel Butt <shakeelb@xxxxxxxxxx> 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@xxxxxxxxx>
>>> 
>>> 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.

I think there is no difference in the assembly code between them in most
cases. The only intention that I can think of is to avoid the potential
complaint (data race) emitted by KCSAN.

> 
> Also there are other similar cgroup interfaces without READ_ONCE()/WRITE_ONCE().

If we decide to fix, then we should fix all.

Thanks.

> 
> Thanks!