On 1/31/19 2:55 PM, Vlastimil Babka wrote:
> On 1/18/19 6:51 PM, Mel Gorman wrote:
> ...
>
>> + for (order = cc->order - 1;
>> + order >= PAGE_ALLOC_COSTLY_ORDER && pfn == cc->migrate_pfn && nr_scanned < limit;
>> + order--) {
>> + struct free_area *area = &cc->zone->free_area[order];
>> + struct list_head *freelist;
>> + unsigned long flags;
>> + struct page *freepage;
>> +
>> + if (!area->nr_free)
>> + continue;
>> +
>> + spin_lock_irqsave(&cc->zone->lock, flags);
>> + freelist = &area->free_list[MIGRATE_MOVABLE];
>> + list_for_each_entry(freepage, freelist, lru) {
>> + unsigned long free_pfn;
>> +
>> + nr_scanned++;
>> + free_pfn = page_to_pfn(freepage);
>> + if (free_pfn < high_pfn) {
>> + update_fast_start_pfn(cc, free_pfn);
>
> Shouldn't this update go below checking pageblock skip bit? We might be
> caching pageblocks that will be skipped, and also potentially going
Ah that move happens in the next patch.
