Skip to contentSkip to conversation

mm/damon/core: add short circuit for esz<min_region_sz in damos_adjust_quota()

PATCH In review Liew Rui Yan 3 participants 4 messages Sep 13 – Sep 14

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm/damon/core: add short circuit for esz<min_region_sz in damos_adjust_quota()

After damos_set_effective_quota(), the effective size quota may be lower than min_region_sz (e.g., when the temporal quota goal is [over-]achieved).

In this situation, damos_adjust_quota() will still calculates quota->min_score. However, this min_score will not be used in this window, because in damon_do_apply_schemes(), damos_quota_is_full() will always returns true, preventing the scheme from being applied to any region.

Therefore, add a short circuit for 'esz < min_region_sz' schemes to early return from damos_adjust_quota() before calculating min_score.

---

Changes from RFC v1:
- RFC v1: https://lore.kernel.org/damon/[email protected]
- Make patch more complete by covering 'esz < min_region_sz' instead of covering "temporal quota goal is [over-]achieved" only.

---
 mm/damon/core.c | 2 ++
 1 file changed, 2 insertions(+)
mm/damon/core.c +2 -0
index ce8c6f99106e..15758525765e 100644
@@ -3318,6 +3318,8 @@ static void damos_adjust_quota(struct damon_ctx *c, struct damos *s)
 			damos_trace_esz(c, s, quota);
 	}
 
+	if (damos_quota_is_full(quota, c->min_region_sz))
+		return;
 	if (!c->ops.get_scheme_score)
 		return;
 
Signature · 1 line
2.55.0
#1 lore raw
Sashiko has reviewed this patch and found no issues. It looks great!
Quoted 10 lines

After damos_set_effective_quota(), the effective size quota may be lower than min_region_sz (e.g., when the temporal quota goal is [over-]achieved).

In this situation, damos_adjust_quota() will still calculates quota->min_score. However, this min_score will not be used in this window, because in damon_do_apply_schemes(), damos_quota_is_full() will always returns true, preventing the scheme from being applied to any region.

Quoted 4 lines

Therefore, add a short circuit for 'esz < min_region_sz' schemes to early return from damos_adjust_quota() before calculating min_score.

Sashiko has reviewed this patch and found no issues. It looks great!

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
maintainer Reviewed-by on v1 1/1
#2 lore raw
In reply to Liew Rui Yan

Hi Liew,

From the next time, please complete discussions on the previous version before posting a new version. For example, in this case, you could replied to my last reply to the previous version of this patch saying you will accept my suggested changes and wait a reasonable time to decide there is no other opinions.

I think the subject is not very clear. How about something like, "skip quota score setup when quota is full"?

Quoted 2 lines · On Sun, 13 Sep 2026 19:01:38 +0800 Liew Rui Yan <[email protected]>…

After damos_set_effective_quota(), the effective size quota may be lower than min_region_sz (e.g., when the temporal quota goal is [over-]achieved).

This sounds unnecessarily specific and doesn't cover all the case. What about

"In damos_adjust_quota(), the quota could already be full."

Quoted 11 lines

In this situation, damos_adjust_quota() will still calculates quota->min_score. However, this min_score will not be used in this window, because in damon_do_apply_schemes(), damos_quota_is_full() will always returns true, preventing the scheme from being applied to any region.

Therefore, add a short circuit for 'esz < min_region_sz' schemes to early return from damos_adjust_quota() before calculating min_score.

Assuming you agree to above suggestions,

With the above assumption, I modified this patch with the suggestions and applied to damon/next [1] tree. I will repost it for mm.git merge after a week. So, if you are ok with the plan, no action from your side is needed.

Please let me know if you have different opinions.

[1] https://origin.kernel.org/doc/html/latest/mm/damon/maintainer-profile.html#scm-trees

Thanks,
SJ

[...]

#3 lore raw
Quoted 7 lines · On Sun, 13 Sep 2026 09:20:45 -0700 SJ Park <[email protected]> wrote:

Hi Liew,

From the next time, please complete discussions on the previous version before posting a new version. For example, in this case, you could replied to my last reply to the previous version of this patch saying you will accept my suggested changes and wait a reasonable time to decide there is no other opinions.

Thank you for your guidance. From the next time, I will reply and say if I will accept your suggestions.

Quoted 3 lines

I think the subject is not very clear. How about something like, "skip quota score setup when quota is full"?

Sure, it is shorter and more understandable.

Quoted 9 lines · On Sun, 13 Sep 2026 19:01:38 +0800 Liew Rui Yan <[email protected]>…
Quoted 3 lines

After damos_set_effective_quota(), the effective size quota may be lower than min_region_sz (e.g., when the temporal quota goal is [over-]achieved).

This sounds unnecessarily specific and doesn't cover all the case. What about

"In damos_adjust_quota(), the quota could already be full."

Agree, that makes sense. Since users who is using temporal quota auto-tuning should know what happens when goal is achieved.

Quoted 8 lines
Quoted 8 lines

In this situation, damos_adjust_quota() will still calculates quota->min_score. However, this min_score will not be used in this window, because in damon_do_apply_schemes(), damos_quota_is_full() will always returns true, preventing the scheme from being applied to any region.

Quoted 6 lines
Quoted 4 lines

Therefore, add a short circuit for 'esz < min_region_sz' schemes to early return from damos_adjust_quota() before calculating min_score.

Assuming you agree to above suggestions,

Yes, I agree those all suggestions.

Quoted 10 lines

With the above assumption, I modified this patch with the suggestions and applied to damon/next [1] tree. I will repost it for mm.git merge after a week. So, if you are ok with the plan, no action from your side is needed.

Please let me know if you have different opinions.

[1] https://origin.kernel.org/doc/html/latest/mm/damon/maintainer-profile.html#scm-trees

Best regards,
Rui Yan