Re: Regarding additional fixes for on node memcg quota metric issues

Re: Regarding additional fixes for on node memcg quota metric issues

[SJ Park]

Hello Harish,

On Mon, 13 Jul 2026 12:32:17 +0530 Harish CS <harish.cs.ss24@gmail.com> wrote:

> Hello SJ,
> 
> I came across the Sashiko review of damos_get_node_memcg_used_bp(), which
> identified possible unsigned underflow, division-by-zero, and 32-bit
> overflow issues in thenode_memcg_{used,free}_bp quota metric calculation.
> I saw your response that you planned to work on these findings separately.
> I am a newcomer to DAMON and upstream  kernel development, and I am
> interested in contributing to this work. Are these issues still open,and
> would it be okay for me to work on them?

Yes, please feel free to work on it and send patches! :)

> 
> And I would appreciate some initial direction on the approach on validating
> the fix.

I don't think it is easy to be tested.  Many Sashiko-found bugs are
theoretically possible but not very common.

Maybe you could run DAMON selftests, kunit tests and/or even DAMON test suite
for correctness [1] and ensure your change doesn't regress at least existing
tests?

> 
> Please let me know whether this would be a reasonable starting point or if
> you would recommend a different approach.

I think this is a reasonable starting point.  Thank you for considering
contribution!

> 
> Parent Bug Discussion:
> https://lore.kernel.org/20260329153425.47097-1-sj@kernel.org/
> Thank you for your time and guidance.
> 
> Regards,
> Harish C S
> 

[1] https://github.com/damonitor/damon-tests/tree/master/corr


Thanks,
SJ