Re: [RFC PATCH v3 1/4] mm/damon: Introduce DAMOS_QUOTA_HUGEPAGE auto tuning

[SeongJae Park <sj@kernel.org>]

On Tue, 9 Jun 2026 14:01:07 +0300 Gutierrez Asier <gutierrez.asier@huawei-partners.com> wrote:

> 
> 
> On 6/6/2026 3:12 AM, SeongJae Park wrote:
> > On Fri, 5 Jun 2026 13:57:47 +0300 Gutierrez Asier <gutierrez.asier@huawei-partners.com> wrote:
> > 
> >>
> >>
> >> On 6/4/2026 6:19 PM, sashiko-bot@kernel.org wrote:
[...]
> >>> [Severity: Medium]
> >>> Is there a risk of a 32-bit overflow in this calculation?
> >>>
> >>> On 32-bit architectures, mult_frac() preserves types, and computing
> >>> thp * 10000 could exceed the maximum unsigned long value if there are
> >>> more than ~1.7 GB of THPs, producing an arbitrarily small quota target.
> >>>
> >> Good catch! I will cast 10000 to unsigned long to preserve the types.
> > 
> > But still it is 'unsigned long' that has only 32 bits on 32-bit architectures?
> > So it shouldn't make any change here.  Or, am I missing something?
> How about just a warning? Something like
> 
> WARN_ONCE(!IS_ENABLED(CONFIG_64BIT),
> 	"In 32bit systems DAMOS_QUOTA_HUGEPAGE_MEM_BP may not be accurate!\n");

I'm not sure if this is a real problem for our real users.  I'd suggest to
ignore Sashiko's concern for now and revisit if real users find and report this
as a problem.


Thanks,
SJ

[...]