Re: [PATCH v2] mm/hotplug: treat CMA pages as unmovable

[Michal Hocko]

On Fri 12-04-19 11:26:59, Qian Cai wrote:
[...]
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index d96ca5bc555b..a9d2b0236167 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -8005,7 +8005,10 @@ void *__init alloc_large_system_hash(const char *tablename,
>  bool has_unmovable_pages(struct zone *zone, struct page *page, int count,
>  			 int migratetype, int flags)
>  {
> -	unsigned long pfn, iter, found;
> +	unsigned long found;
> +	unsigned long iter = 0;
> +	unsigned long pfn = page_to_pfn(page);
> +	char reason[] = "unmovable page";

	const char *reason = "unovable page";
>  
>  	/*
>  	 * TODO we could make this much more efficient by not checking every
> @@ -8015,17 +8018,20 @@ bool has_unmovable_pages(struct zone *zone, struct page *page, int count,
>  	 * can still lead to having bootmem allocations in zone_movable.
>  	 */
>  
> -	/*
> -	 * CMA allocations (alloc_contig_range) really need to mark isolate
> -	 * CMA pageblocks even when they are not movable in fact so consider
> -	 * them movable here.
> -	 */
> -	if (is_migrate_cma(migratetype) &&
> -			is_migrate_cma(get_pageblock_migratetype(page)))
> -		return false;
> +	if (is_migrate_cma(get_pageblock_migratetype(page))) {
> +		/*
> +		 * CMA allocations (alloc_contig_range) really need to mark
> +		 * isolate CMA pageblocks even when they are not movable in fact
> +		 * so consider them movable here.
> +		 */
> +		if (is_migrate_cma(migratetype))
> +			return false;
> +
> +		strscpy(reason, "CMA page", 9);

		reason = "CMA page";

> +		goto unmovable;
> +	}

Other than that looks good. After fixing the above, feel free to add
Acked-by: Michal Hocko <mhocko@xxxxxxxx>

-- 
Michal Hocko
SUSE Labs