Re: [RFC PATCH v1.3 14/19] mm/damon/core: get merge threshold from probe hits when weights are set

[SJ Park <sj@kernel.org>]

On Fri, 10 Jul 2026 00:38:39 +0000 sashiko-bot@kernel.org wrote:

> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] Premature termination of regions merging when using probe weights, breaking `max_nr_regions` enforcement.
> --
> 
> --- Patch [14]: [RFC PATCH v1.3 14/19] mm/damon/core: get merge threshold from probe hits when weights are set ---
> commit c2e77863ffba068b4f2b7db5f45690d846ec2ad7
> Author: SJ Park <sj@kernel.org>
> 
> mm/damon/core: get merge threshold from probe hits when weights are set
> 
> This commit updates DAMON to calculate the merge threshold based on the
> maximum probe hits weighted sum when probe weights are used. It updates
> kdamond_fn() to retrieve this max sum via the apply_probes() callback.
> 
> > diff --git a/mm/damon/core.c b/mm/damon/core.c
> > index f7782c69037da..91964336ce31e 100644
> > --- a/mm/damon/core.c
> > +++ b/mm/damon/core.c
> [ ... ]
> > @@ -3776,9 +3777,18 @@ static int kdamond_fn(void *data)
> >  
> >  		if (!access_check_disabled && ctx->ops.check_accesses)
> >  			max_merge_score = ctx->ops.check_accesses(ctx);
> > -		if (ctx->ops.apply_probes)
> > -			ctx->ops.apply_probes(ctx, access_check_disabled,
> > -					false);
> > +		if (ctx->ops.apply_probes) {
> > +			if (time_after_eq(ctx->passed_sample_intervals,
> > +						next_aggregation_sis) &&
> > +					access_check_disabled)
> > +				get_max_wsum = true;
> > +			else
> > +				get_max_wsum = false;
> > +			max_wsum = ctx->ops.apply_probes(ctx,
> > +					access_check_disabled, get_max_wsum);
> > +			if (get_max_wsum)
> > +				max_merge_score = max_wsum;
> 
> [Severity: High]
> Does this regression break the max_nr_regions enforcement in
> kdamond_merge_regions()?
> 
> When probe weights are set, max_merge_score is derived from the user-provided
> probe weights and passed as the initial threshold to kdamond_merge_regions():
> 
> mm/damon/core.c:kdamond_merge_regions() {
>     ...
>     max_thres = c->attrs.aggr_interval /
>         (c->attrs.sample_interval ?  c->attrs.sample_interval : 1);
>     do {
>         ...
>         threshold = max(1, threshold * 2);
>     } while (nr_regions > c->attrs.max_nr_regions &&
>             threshold / 2 < max_thres);
>     ...
> }
> 
> However, max_thres is hardcoded to the maximum unweighted access frequency.
> 
> If the dynamically weighted max_wsum is large enough, the initial threshold
> passed to kdamond_merge_regions() could exceed max_thres * 2.
> 
> If that happens, wouldn't the loop terminate prematurely on the first pass
> even when nr_regions > max_nr_regions, allowing the number of regions to
> grow unboundedly?

No issue.  The later patch of this series will set max_thres with weighted sum.

> 
> > +		}
> >  
> >  		if (time_after_eq(ctx->passed_sample_intervals,
> >  					next_aggregation_sis)) {
> 
> -- 
> Sashiko AI review · https://sashiko.dev/#/patchset/20260710002349.111414-1-sj@kernel.org?part=14


Thanks,
SJ