Re: [PATCH v2 4/6] mm/migrate_device: convert __migrate_device_pages() to folios

[Sidhartha Kumar <sidhartha.kumar@oracle.com>]

On 2/16/24 1:55 PM, Matthew Wilcox wrote:
> On Fri, Feb 16, 2024 at 01:13:18PM -0800, Sidhartha Kumar wrote:
>> Use migrate_pfn_to_folio() so we can work with folios directly in
>> __migrate_device_pages().
> 
> i don't understand why this would be correct if we have multipage
> folios.
> 

Alistair mentioned that he is working on order > 0 device page support so I was 
under the impression that currently device pages are only order 0.

Thanks,
Sid

>> @@ -719,33 +719,29 @@ static void __migrate_device_pages(unsigned long *src_pfns,
>>   					migrate->pgmap_owner);
>>   				mmu_notifier_invalidate_range_start(&range);
>>   			}
>> -			migrate_vma_insert_page(migrate, addr, newpage,
>> +			migrate_vma_insert_page(migrate, addr, &dst->page,
> 
> seems to me that a migration pfn is going to refer to a precise page.
> now you're telling it to insert the head page of the folio.  isn't this
> wrong?
> 
>> @@ -753,13 +749,11 @@ static void __migrate_device_pages(unsigned long *src_pfns,
>>   			continue;
>>   		}
>>   
>> -		if (migrate && migrate->fault_page == page)
>> -			r = migrate_folio_extra(mapping, page_folio(newpage),
>> -						page_folio(page),
>> -						MIGRATE_SYNC_NO_COPY, 1);
>> +		if (migrate && migrate->fault_page == &src->page)
> 
> shouldn't this rather be "page_folio(migrate->fault_page) == src"?
> ie we're looking for two pages from the same folio, rather than the page
> being the same as the head page of the folio?
>