Re: [PATCH v4 1/5] hugetlb: add demote hugetlb page sysfs interfaces

[Mike Kravetz]

On 10/18/21 12:35 AM, Oscar Salvador wrote:
> On Fri, Oct 08, 2021 at 01:24:28PM -0700, Mike Kravetz wrote:
>> In general, the resize_lock prevents unexpected consequences when
>> multiple users are modifying the number of pages in a pool concurrently
>> from the proc/sysfs interfaces.  The mutex is acquired here because we
>> are modifying (decreasing) the pool size.
> 
> Yes, I got that. My question was wrt. n_mask initialization:
> 
>  +	if (nid != NUMA_NO_NODE) {
>  +		init_nodemask_of_node(&nodes_allowed, nid);
>  +		n_mask = &nodes_allowed;
>  +	} else {
>  +		n_mask = &node_states[N_MEMORY];
>  +	}
> 
> AFAICS, this does not need to be protected.
> 

Sorry that I misunderstood your quesion!

You are correct, the n_mask initialization does not need to be protected
by the mutex.  Thanks for pointing that out.

The updated patch below simply moves taking the mutex after this
initialization code.

Andrew, please let me know if you want something else to make this
update simpler for you.