Re: [PATCH 7/7] mm,hwpoison: Remove drain_all_pages from shake_page

[Vlastimil Babka <vbabka@suse.cz>]

On 11/19/20 11:57 AM, Oscar Salvador wrote:
> get_hwpoison_page already drains pcplists, previously disabling
> them when trying to grab a refcount.
> We do not need shake_page to take care of it anymore.
> 
> Signed-off-by: Oscar Salvador <osalvador@suse.de>
> ---
>   mm/memory-failure.c | 7 ++-----
>   1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index 512613e9a1bd..ad976e1c3178 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -263,8 +263,8 @@ static int kill_proc(struct to_kill *tk, unsigned long pfn, int flags)
>   }
>   
>   /*
> - * When a unknown page type is encountered drain as many buffers as possible
> - * in the hope to turn the page into a LRU or free page, which we can handle.
> + * Unknown page type encountered. Try to check whether it can turn PageLRU by
> + * lru_add_drain_all, or a free page by reclaiming slabs when possible.
>    */
>   void shake_page(struct page *p, int access)
>   {
> @@ -275,9 +275,6 @@ void shake_page(struct page *p, int access)
>   		lru_add_drain_all();
>   		if (PageLRU(p))
>   			return;
> -		drain_all_pages(page_zone(p));
> -		if (PageLRU(p) || is_free_buddy_page(p))
> -			return;

I wonder if page in the lru pagevec can in fact become free after draining the 
lru - in that case we could keep the is_free_buddy_page() check.

>   	}
>   
>   	/*
>