Re: [RFC 1/2] mm: Uncharge poisoned pages

[Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>]

On Thu, Apr 20, 2017 at 11:26:01AM +0200, Laurent Dufour wrote:
> When page are poisoned, they should be uncharged from the root memory
> cgroup.

Could you include some information about what problem this patch tries
to solve?
# I know that you already explain it in patch 0/2, so you can simply
# copy from it.

>
> Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
> ---
>  mm/memory-failure.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index 27f7210e7fab..00bd39d3d4cb 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -530,6 +530,7 @@ static const char * const action_page_types[] = {
>  static int delete_from_lru_cache(struct page *p)
>  {
>  	if (!isolate_lru_page(p)) {
> +		memcg_kmem_uncharge(p, 0);

This function is supposed to be called with if (memcg_kmem_enabled()) check,
so could you do like below?

+		if (memcg_kmem_enabled())
+			memcg_kmem_uncharge(p, 0);


And I feel that we can call this function outside if (!isolate_lru_page(p))
block, because isolate_lru_page could fail and then the error page is left
incompletely isolated. Such error page has PageHWPoison set, so I guess that
the reported bug still triggers on such case.

Thanks,
Naoya Horiguchi
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>