Re: [PATCH v2 2/3] mm: huge_memory: refactor anon_enabled_store() with change_anon_orders()

[Baolin Wang <baolin.wang@linux.alibaba.com>]

On 3/6/26 7:22 PM, Lorenzo Stoakes (Oracle) wrote:
> On Fri, Mar 06, 2026 at 02:30:22AM -0800, Breno Leitao wrote:
>> On Fri, Mar 06, 2026 at 02:07:50PM +0800, Baolin Wang wrote:
>>>> +	mode = sysfs_match_string(enabled_mode_strings, buf);
>>>> +	if (mode < 0)
>>>> +		return -EINVAL;
>>>> -	if (ret > 0) {
>>>> -		int err;
>>>> +	if (change_anon_orders(order, mode)) {
>>>> +		int err = start_stop_khugepaged();
>>>
>>> Thanks for the cleanup, and the code looks better.
>>>
>>>> -		err = start_stop_khugepaged();
>>>>    		if (err)
>>>> -			ret = err;
>>>> +			return err;
>>>> +	} else {
>>>> +		/*
>>>> +		 * Recalculate watermarks even when the mode didn't
>>>> +		 * change, as the previous code always called
>>>> +		 * start_stop_khugepaged() which does this internally.
>>>> +		 */
>>>> +		set_recommended_min_free_kbytes();
>>>
>>> However, this won't fix your issue. You will still get lots of warning
>>> messages even if no hugepage options are changed.
>>
>> Correct — this was part of the earlier discussion, where the suggestion
>> was to retain the set_recommended_min_free_kbytes() call even when it
>> is a no-op.
>>
>> As for the warnings, I see a few possible approaches:
>>
>>   * Remove them entirely — are they providing any real value?
>>   * Apply rate limiting to reduce the noise
> 
> I think rate limiting is the sensible approach.

+1. Make sense to me.