Re: [PATCH v1] mm/memory_hotplug: don't check the nid in find_(smallest|biggest)_section_pfn

[David Hildenbrand]

> Am 27.11.2019 um 20:03 schrieb Qian Cai <cai@xxxxxx>:
> 
> ï
> 
>> On Nov 27, 2019, at 12:42 PM, David Hildenbrand <david@xxxxxxxxxx> wrote:
>> 
>> Now that we always check against a zone, we can stop checking against
>> the nid, it is implicitly covered by the zone.
>> 
>> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
>> Cc: Michal Hocko <mhocko@xxxxxxxxxx>
>> Cc: Oscar Salvador <osalvador@xxxxxxx>
>> Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
>> ---
>> mm/memory_hotplug.c | 23 ++++++++---------------
>> 1 file changed, 8 insertions(+), 15 deletions(-)
>> 
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index 46b2e056a43f..602f753c662c 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -344,17 +344,14 @@ int __ref __add_pages(int nid, unsigned long pfn, unsigned long nr_pages,
>> }
>> 
>> /* find the smallest valid pfn in the range [start_pfn, end_pfn) */
>> -static unsigned long find_smallest_section_pfn(int nid, struct zone *zone,
>> -                     unsigned long start_pfn,
>> -                     unsigned long end_pfn)
>> +static unsigned long find_smallest_section_pfn(struct zone *zone,
>> +                           unsigned long start_pfn,
>> +                           unsigned long end_pfn)
>> {
>>   for (; start_pfn < end_pfn; start_pfn += PAGES_PER_SUBSECTION) {
>>       if (unlikely(!pfn_to_online_page(start_pfn)))
>>           continue;
>> 
>> -        if (unlikely(pfn_to_nid(start_pfn) != nid))
>> -            continue;
> 
> Are you sure? I thought this is to check against machines with odd layouts, no? 

The zone pointer is unique for every node. (in contrast to the zone index).

Thanks!

> 
>            /*
>             * Nodes's pfns can be overlapping.
>             * We know some arch can have a nodes layout such as
>             * -------------pfn-------------->
>             * N0 | N1 | N2 | N0 | N1 | N2|....
>             */
>