Re: [PATCH v4 2/2] HWPOISON: soft offlining for non-lru movable page

[Michal Hocko]

On Wed 25-01-17 23:05:38, ysxie@xxxxxxxxxxx wrote:
> From: Yisheng Xie <xieyisheng1@xxxxxxxxxx>
> 
> This patch is to extends soft offlining framework to support
> non-lru page, which already support migration after
> commit bda807d44454 ("mm: migrate: support non-lru movable page
> migration")
> 
> When memory corrected errors occur on a non-lru movable page,
> we can choose to stop using it by migrating data onto another
> page and disable the original (maybe half-broken) one.
> 
> Signed-off-by: Yisheng Xie <xieyisheng1@xxxxxxxxxx>
> Suggested-by: Michal Hocko <mhocko@xxxxxxxxxx>
> Suggested-by: Minchan Kim <minchan@xxxxxxxxxx>
> Reviewed-by: Minchan Kim <minchan@xxxxxxxxxx>
> Acked-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> CC: Vlastimil Babka <vbabka@xxxxxxx>
> ---
>  mm/memory-failure.c | 26 ++++++++++++++++----------
>  1 file changed, 16 insertions(+), 10 deletions(-)
> 
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index f283c7e..56e39f8 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -1527,7 +1527,8 @@ static int get_any_page(struct page *page, unsigned long pfn, int flags)
>  {
>  	int ret = __get_any_page(page, pfn, flags);
>  
> -	if (ret == 1 && !PageHuge(page) && !PageLRU(page)) {
> +	if (ret == 1 && !PageHuge(page) &&
> +	    !PageLRU(page) && !__PageMovable(page)) {
>  		/*
>  		 * Try to free it.
>  		 */

Is this sufficient? Not that I am familiar with get_any_page() but
__get_any_page doesn't seem to be aware of movable pages and neither
shake_page is.

> @@ -1649,7 +1650,10 @@ static int __soft_offline_page(struct page *page, int flags)
>  	 * Try to migrate to a new page instead. migrate.c
>  	 * handles a large number of cases for us.
>  	 */
> -	ret = isolate_lru_page(page);
> +	if (PageLRU(page))
> +		ret = isolate_lru_page(page);
> +	else if (!isolate_movable_page(page, ISOLATE_UNEVICTABLE))
> +		ret = -EBUSY;

As pointed out in the previous response isolate_movable_page should
really have the same return value contract as [__]isolate_lru_page

>  	/*
>  	 * Drop page reference which is came from get_any_page()
>  	 * successful isolate_lru_page() already took another one.
-- 
Michal Hocko
SUSE Labs