Re: [PATCH] mm/damon: validate addr_unit to be power of 2

[Liew Rui Yan <aethernet65535@gmail.com>]

Hi SeongJae,

> It would be better to not terminate.  But not a big deal.  Users could be wary
> of the consequence, as long as it is well documented.
> 
> From my perspective, therefore, code maintenance is more important.  If we can
> make DAMON not terminated by invalid user inputs caused damon_commit_ctx() but
> it introduces a significant amount of code complexity, I don't think that's a
> good deal.

Thank you so much for the thoughtful response! I completely agree that
code maintainability should be the priority and we should avoid
unnecessary complexity.

Just to make sure I captured the full context, in my earlier follow-ups
[1][2] (in case they got folded in your email client):

- Does the current kdamond termination on invalid user input align with
  your design goal mentioned in [3] (stopping only for "internal
  errors")?

- Would a lightweight fix be acceptable? For example, performing
  validation at the very beginning of damon_commit_ctx(), and returning
  -EINVAL before setting 'maybe_corrupted' to true. Since no
  modifications to 'dst' would have occurred, kdamond could safely
  continue with its old configuration.

Thank you again for your time, and guidance. :>

[1] https://lore.kernel.org/20260327064517.68131-1-aethernet65535@gmail.com
[2] https://lore.kernel.org/20260327121009.38374-1-aethernet65535@gmail.com
[3] https://lore.kernel.org/20260319151528.86490-1-sj@kernel.org

Best regards,
Rui Yan