linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] mm: add lazyfree folio to lru tail
@ 2024-08-15 12:33 gaoxu
  2024-08-15 18:10 ` Yu Zhao
  0 siblings, 1 reply; 3+ messages in thread
From: gaoxu @ 2024-08-15 12:33 UTC (permalink / raw)
  To: Andrew Morton
  Cc: linux-mm, linux-kernel, Suren Baghdasaryan, Yu Zhao, Shaohua Li,
	yipengxiang, fengbaopeng

Replace lruvec_add_folio with lruvec_add_folio_tail in the lru_lazyfree_fn:
1. The lazy-free folio is added to the LRU_INACTIVE_FILE list. If it's
   moved to the LRU tail, it allows for faster release lazy-free folio and
   reduces the impact on file refault.
2. When mglru is enabled, the lazy-free folio is reclaimabled and should be
   added using lru_gen_add_folio(lruvec, folio, true) instead of
   lru_gen_add_folio(lruvec, folio, false) for adding to gen.

With the change in place, workingset_refault_file is reduced by 33% in the
continuous startup testing of the applications in the Android system.

Signed-off-by: gao xu <gaoxu2@hihonor.com>
---
 mm/swap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/swap.c b/mm/swap.c
index 67786cb77..ed49643ec 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -635,7 +635,7 @@ static void lru_lazyfree_fn(struct lruvec *lruvec, struct folio *folio)
 		 * anonymous folios
 		 */
 		folio_clear_swapbacked(folio);
-		lruvec_add_folio(lruvec, folio);
+		lruvec_add_folio_tail(lruvec, folio);
 
 		__count_vm_events(PGLAZYFREE, nr_pages);
 		__count_memcg_events(lruvec_memcg(lruvec), PGLAZYFREE,
-- 
2.17.1

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] mm: add lazyfree folio to lru tail
  2024-08-15 12:33 [PATCH] mm: add lazyfree folio to lru tail gaoxu
@ 2024-08-15 18:10 ` Yu Zhao
  2024-08-16  1:41   ` 回复: " gaoxu
  0 siblings, 1 reply; 3+ messages in thread
From: Yu Zhao @ 2024-08-15 18:10 UTC (permalink / raw)
  To: gaoxu, Kalesh Singh
  Cc: Andrew Morton, linux-mm, linux-kernel, Suren Baghdasaryan,
	Shaohua Li, yipengxiang, fengbaopeng

On Thu, Aug 15, 2024 at 6:33 AM gaoxu <gaoxu2@honor.com> wrote:
>
> Replace lruvec_add_folio with lruvec_add_folio_tail in the lru_lazyfree_fn:
> 1. The lazy-free folio is added to the LRU_INACTIVE_FILE list. If it's
>    moved to the LRU tail, it allows for faster release lazy-free folio and
>    reduces the impact on file refault.
> 2. When mglru is enabled, the lazy-free folio is reclaimabled and should be
>    added using lru_gen_add_folio(lruvec, folio, true) instead of
>    lru_gen_add_folio(lruvec, folio, false) for adding to gen.
>
> With the change in place, workingset_refault_file is reduced by 33% in the
> continuous startup testing of the applications in the Android system.
>
> Signed-off-by: gao xu <gaoxu2@hihonor.com>
> ---
>  mm/swap.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/mm/swap.c b/mm/swap.c
> index 67786cb77..ed49643ec 100644
> --- a/mm/swap.c
> +++ b/mm/swap.c
> @@ -635,7 +635,7 @@ static void lru_lazyfree_fn(struct lruvec *lruvec, struct folio *folio)

Please work against the latest mm-unstable.

>                  * anonymous folios
>                  */
>                 folio_clear_swapbacked(folio);
> -               lruvec_add_folio(lruvec, folio);
> +               lruvec_add_folio_tail(lruvec, folio);
>
>                 __count_vm_events(PGLAZYFREE, nr_pages);
>                 __count_memcg_events(lruvec_memcg(lruvec), PGLAZYFREE,

Kalesh has been working on something similar for lru_deactivate(), and
he can help test and review this patch.


^ permalink raw reply	[flat|nested] 3+ messages in thread

* 回复: [PATCH] mm: add lazyfree folio to lru tail
  2024-08-15 18:10 ` Yu Zhao
@ 2024-08-16  1:41   ` gaoxu
  0 siblings, 0 replies; 3+ messages in thread
From: gaoxu @ 2024-08-16  1:41 UTC (permalink / raw)
  To: Yu Zhao, Kalesh Singh
  Cc: Andrew Morton, linux-mm, linux-kernel, Suren Baghdasaryan,
	Shaohua Li, yipengxiang, fengbaopeng

> 
> On Thu, Aug 15, 2024 at 6:33 AM gaoxu <gaoxu2@honor.com> wrote:
> >
> > Replace lruvec_add_folio with lruvec_add_folio_tail in the lru_lazyfree_fn:
> > 1. The lazy-free folio is added to the LRU_INACTIVE_FILE list. If it's
> >    moved to the LRU tail, it allows for faster release lazy-free folio and
> >    reduces the impact on file refault.
> > 2. When mglru is enabled, the lazy-free folio is reclaimabled and should be
> >    added using lru_gen_add_folio(lruvec, folio, true) instead of
> >    lru_gen_add_folio(lruvec, folio, false) for adding to gen.
> >
> > With the change in place, workingset_refault_file is reduced by 33% in the
> > continuous startup testing of the applications in the Android system.
> >
> > Signed-off-by: gao xu <gaoxu2@hihonor.com>
> > ---
> >  mm/swap.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/mm/swap.c b/mm/swap.c
> > index 67786cb77..ed49643ec 100644
> > --- a/mm/swap.c
> > +++ b/mm/swap.c
> > @@ -635,7 +635,7 @@ static void lru_lazyfree_fn(struct lruvec *lruvec, struct
> folio *folio)
> 
> Please work against the latest mm-unstable.
This was my mistake. Thank you for the reminder.
I will resend the patch based on the mm-unstable branch
> 
> >                  * anonymous folios
> >                  */
> >                 folio_clear_swapbacked(folio);
> > -               lruvec_add_folio(lruvec, folio);
> > +               lruvec_add_folio_tail(lruvec, folio);
> >
> >                 __count_vm_events(PGLAZYFREE, nr_pages);
> >                 __count_memcg_events(lruvec_memcg(lruvec),
> PGLAZYFREE,
> 
> Kalesh has been working on something similar for lru_deactivate(), and
> he can help test and review this patch.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2024-08-16  1:41 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-08-15 12:33 [PATCH] mm: add lazyfree folio to lru tail gaoxu
2024-08-15 18:10 ` Yu Zhao
2024-08-16  1:41   ` 回复: " gaoxu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox