Re: [RFC PATCH v1 4/4] mm, memory_hotplug: fix inconsistent num_poisoned_pages on memory hotremove

[Miaohe Lin <linmiaohe@huawei.com>]

On 2022/4/27 12:28, Naoya Horiguchi wrote:
> From: Naoya Horiguchi <naoya.horiguchi@nec.com>
> 
> When offlining memory section with hwpoisoned pages, the hwpoisons are
> canceled. But num_poisoned_pages is not updated for that event, so the
> counter becomes inconsistent.

IIUC, this work is already done via clear_hwpoisoned_pages when __remove_pages.
Or am I miss something?

Thanks!

> 
> Add num_poisoned_pages_dec() in __offline_isolated_pages(). PageHWPoison
> can be set on a tail page of some high order buddy page, so we need check
> PageHWPoison on each subpage.
> 
> Signed-off-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
> ---
>  mm/page_alloc.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 6e5b4488a0c5..dcd962855617 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -9487,12 +9487,15 @@ void __offline_isolated_pages(unsigned long start_pfn, unsigned long end_pfn)
>  	zone = page_zone(pfn_to_page(pfn));
>  	spin_lock_irqsave(&zone->lock, flags);
>  	while (pfn < end_pfn) {
> +		int i;
> +
>  		page = pfn_to_page(pfn);
>  		/*
>  		 * The HWPoisoned page may be not in buddy system, and
>  		 * page_count() is not 0.
>  		 */
>  		if (unlikely(!PageBuddy(page) && PageHWPoison(page))) {
> +			num_poisoned_pages_dec();
>  			pfn++;
>  			continue;
>  		}
> @@ -9510,6 +9513,9 @@ void __offline_isolated_pages(unsigned long start_pfn, unsigned long end_pfn)
>  		BUG_ON(page_count(page));
>  		BUG_ON(!PageBuddy(page));
>  		order = buddy_order(page);
> +		for (i = 0; i < 1 << order; i++)
> +			if (PageHWPoison(page + i))
> +				num_poisoned_pages_dec();
>  		del_page_from_free_list(page, zone, order);
>  		pfn += (1 << order);
>  	}
>