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

[Gutierrez Asier <gutierrez.asier@huawei-partners.com>]

On 6/5/2026 3:44 AM, SeongJae Park wrote:
> On Thu, 4 Jun 2026 15:03:34 +0000 <gutierrez.asier@huawei-partners.com> wrote:
> 
>> From: Asier Gutierrez <gutierrez.asier@huawei-partners.com>
>>
>> Introduce DAMOS_QUOTA_HUGEPAGE auto tuning Add a
>> new DAMOS quota goal metric to measure the amount of huge
>> page consumption to total anonymous memory consumption
>> ratio.
> 
> s/anonymous// ?
> 
>>
>> Signed-off-by: Asier Gutierrez <gutierrez.asier@huawei-partners.com>
>> ---
>>  include/linux/damon.h |  1 +
>>  mm/damon/core.c       | 14 ++++++++++++++
>>  2 files changed, 15 insertions(+)
>>
>> diff --git a/include/linux/damon.h b/include/linux/damon.h
>> index c7a31572689b..8e15a674e893 100644
>> --- a/include/linux/damon.h
>> +++ b/include/linux/damon.h
>> @@ -177,6 +177,7 @@ enum damos_quota_goal_metric {
>>  	DAMOS_QUOTA_ACTIVE_MEM_BP,
>>  	DAMOS_QUOTA_INACTIVE_MEM_BP,
>>  	DAMOS_QUOTA_NODE_ELIGIBLE_MEM_BP,
>> +	DAMOS_QUOTA_HUGEPAGE,
> 
> On the previous revision, I suggested renaming this to
> DAMOS_QUOTA_HUGEPAGE_MEM_BP for consistency, but it didn't got your reply.
> Could I ask your opinion?
> 
> Also, I asked you to add the kernel-doc comment before dropping RFC.  I know
> you didn't drop RFC tag yet, but could you consider adding it from the next
> revision?
Sorry, missed it. I will change it.
>>  	NR_DAMOS_QUOTA_GOAL_METRICS,
>>  };
>>  
>> diff --git a/mm/damon/core.c b/mm/damon/core.c
>> index 9f38deddcb30..7fc7477a353a 100644
>> --- a/mm/damon/core.c
>> +++ b/mm/damon/core.c
>> @@ -2536,6 +2536,17 @@ static unsigned int damos_get_in_active_mem_bp(bool active_ratio)
>>  	return mult_frac(inactive, 10000, total);
>>  }
>>  
>> +static unsigned int damos_hugepage_mem_bp(void)
>> +{
>> +	unsigned long thp, total;
>> +
>> +	thp = global_node_page_state(NR_ANON_THPS) +
>> +				global_node_page_state(NR_SHMEM_THPS) +
>> +				global_node_page_state(NR_FILE_THPS);
>> +	total = totalram_pages() - global_zone_page_state(NR_FREE_PAGES);
>> +	return mult_frac(thp, 10000, total);
>> +}
> 
> As Sashiko also [1] commented, let's avoid divide by zero by checking if total
> is zero, before calling mult_frac().
> 
> I wouldn't mind ignoring Sashiko's 32-bit concern, but do you have any opinion?
I just answered. I can cast 10.000 to unsigned long. I don't think many people
using 32 bit systems will use over 1.6GB of huge pages, but it makes sense to
be safe anyway.
>> +
>>  static void damos_set_quota_goal_current_value(struct damon_ctx *c,
>>  		struct damos *s, struct damos_quota_goal *goal)
>>  {
>> @@ -2567,6 +2578,9 @@ static void damos_set_quota_goal_current_value(struct damon_ctx *c,
>>  		goal->current_value = damos_get_node_eligible_mem_bp(c, s,
>>  				goal->nid);
>>  		break;
>> +	case DAMOS_QUOTA_HUGEPAGE:
>> +		goal->current_value = damos_hugepage_mem_bp();
>> +		break;
>>  	default:
>>  		break;
>>  	}
>> -- 
>> 2.43.0
> 
> Other than the above comments, looks good to me overall.
> 
> [1] https://lore.kernel.org/20260604151932.93C431F00893@smtp.kernel.org
> 
> 
> Thanks,
> SJ
> 

-- 
Asier Gutierrez
Huawei