2021-02-04 11:00:03

by Muchun Song

[permalink] [raw]
Subject: [PATCH] mm: memcontrol: replace the loop with a list_for_each_entry()

The rule of list walk has gone since:

commit a9d5adeeb4b2 ("mm/memcontrol: allow to uncharge page without using page->lru field")

So remove the strange comment and replace the loop with a
list_for_each_entry().

Signed-off-by: Muchun Song <[email protected]>
---
mm/memcontrol.c | 17 ++---------------
1 file changed, 2 insertions(+), 15 deletions(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 6c7f1ea3955e..43341bd7ea1c 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -6891,24 +6891,11 @@ static void uncharge_page(struct page *page, struct uncharge_gather *ug)
static void uncharge_list(struct list_head *page_list)
{
struct uncharge_gather ug;
- struct list_head *next;
+ struct page *page;

uncharge_gather_clear(&ug);
-
- /*
- * Note that the list can be a single page->lru; hence the
- * do-while loop instead of a simple list_for_each_entry().
- */
- next = page_list->next;
- do {
- struct page *page;
-
- page = list_entry(next, struct page, lru);
- next = page->lru.next;
-
+ list_for_each_entry(page, page_list, lru)
uncharge_page(page, &ug);
- } while (next != page_list);
-
uncharge_batch(&ug);
}

--
2.11.0


2021-02-04 15:23:39

by Johannes Weiner

[permalink] [raw]
Subject: Re: [PATCH] mm: memcontrol: replace the loop with a list_for_each_entry()

On Thu, Feb 04, 2021 at 06:53:20PM +0800, Muchun Song wrote:
> The rule of list walk has gone since:
>
> commit a9d5adeeb4b2 ("mm/memcontrol: allow to uncharge page without using page->lru field")
>
> So remove the strange comment and replace the loop with a
> list_for_each_entry().
>
> Signed-off-by: Muchun Song <[email protected]>
> ---
> mm/memcontrol.c | 17 ++---------------
> 1 file changed, 2 insertions(+), 15 deletions(-)
>
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 6c7f1ea3955e..43341bd7ea1c 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -6891,24 +6891,11 @@ static void uncharge_page(struct page *page, struct uncharge_gather *ug)
> static void uncharge_list(struct list_head *page_list)
> {
> struct uncharge_gather ug;
> - struct list_head *next;
> + struct page *page;
>
> uncharge_gather_clear(&ug);
> -
> - /*
> - * Note that the list can be a single page->lru; hence the
> - * do-while loop instead of a simple list_for_each_entry().
> - */
> - next = page_list->next;
> - do {
> - struct page *page;
> -
> - page = list_entry(next, struct page, lru);
> - next = page->lru.next;
> -
> + list_for_each_entry(page, page_list, lru)
> uncharge_page(page, &ug);
> - } while (next != page_list);
> -
> uncharge_batch(&ug);

Good catch, this makes things much simpler.

Looking at the surrounding code, there also seems to be no reason
anymore to have uncharge_list() as a separate function: there is only
one caller after the mentioned commit, and it's trivial after your
change. Would you mind folding it into mem_cgroup_uncharge_list()?

The list_empty() check in that one is also unnecessary now: the
do-while loop required at least one page to be on the list or it would
crash, but list_for_each() will be just fine on an empty list.

Thanks

2021-02-04 16:01:15

by Muchun Song

[permalink] [raw]
Subject: Re: [External] Re: [PATCH] mm: memcontrol: replace the loop with a list_for_each_entry()

On Thu, Feb 4, 2021 at 11:14 PM Johannes Weiner <[email protected]> wrote:
>
> On Thu, Feb 04, 2021 at 06:53:20PM +0800, Muchun Song wrote:
> > The rule of list walk has gone since:
> >
> > commit a9d5adeeb4b2 ("mm/memcontrol: allow to uncharge page without using page->lru field")
> >
> > So remove the strange comment and replace the loop with a
> > list_for_each_entry().
> >
> > Signed-off-by: Muchun Song <[email protected]>
> > ---
> > mm/memcontrol.c | 17 ++---------------
> > 1 file changed, 2 insertions(+), 15 deletions(-)
> >
> > diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> > index 6c7f1ea3955e..43341bd7ea1c 100644
> > --- a/mm/memcontrol.c
> > +++ b/mm/memcontrol.c
> > @@ -6891,24 +6891,11 @@ static void uncharge_page(struct page *page, struct uncharge_gather *ug)
> > static void uncharge_list(struct list_head *page_list)
> > {
> > struct uncharge_gather ug;
> > - struct list_head *next;
> > + struct page *page;
> >
> > uncharge_gather_clear(&ug);
> > -
> > - /*
> > - * Note that the list can be a single page->lru; hence the
> > - * do-while loop instead of a simple list_for_each_entry().
> > - */
> > - next = page_list->next;
> > - do {
> > - struct page *page;
> > -
> > - page = list_entry(next, struct page, lru);
> > - next = page->lru.next;
> > -
> > + list_for_each_entry(page, page_list, lru)
> > uncharge_page(page, &ug);
> > - } while (next != page_list);
> > -
> > uncharge_batch(&ug);
>
> Good catch, this makes things much simpler.
>
> Looking at the surrounding code, there also seems to be no reason
> anymore to have uncharge_list() as a separate function: there is only
> one caller after the mentioned commit, and it's trivial after your
> change. Would you mind folding it into mem_cgroup_uncharge_list()?

Will do. Thanks.

>
> The list_empty() check in that one is also unnecessary now: the
> do-while loop required at least one page to be on the list or it would
> crash, but list_for_each() will be just fine on an empty list.

Right. It makes things more simple.

>
> Thanks