2017-03-22 00:51:25

by Johannes Weiner

[permalink] [raw]
Subject: [PATCH] mm: rmap: fix huge file mmap accounting in the memcg stats

Huge pages are accounted as single units in the memcg's "file_mapped"
counter. Account the correct number of base pages, like we do in the
corresponding node counter.

Signed-off-by: Johannes Weiner <[email protected]>
---
include/linux/memcontrol.h | 6 ++++++
mm/rmap.c | 4 ++--
2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index baa274150210..c5ebb32fef49 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -741,6 +741,12 @@ static inline bool mem_cgroup_oom_synchronize(bool wait)
return false;
}

+static inline void mem_cgroup_update_page_stat(struct page *page,
+ enum mem_cgroup_stat_index idx,
+ int nr)
+{
+}
+
static inline void mem_cgroup_inc_page_stat(struct page *page,
enum mem_cgroup_stat_index idx)
{
diff --git a/mm/rmap.c b/mm/rmap.c
index 1d82057144ba..f514cdd84482 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -1154,7 +1154,7 @@ void page_add_file_rmap(struct page *page, bool compound)
goto out;
}
__mod_node_page_state(page_pgdat(page), NR_FILE_MAPPED, nr);
- mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED);
+ mem_cgroup_update_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED, nr);
out:
unlock_page_memcg(page);
}
@@ -1194,7 +1194,7 @@ static void page_remove_file_rmap(struct page *page, bool compound)
* pte lock(a spinlock) is held, which implies preemption disabled.
*/
__mod_node_page_state(page_pgdat(page), NR_FILE_MAPPED, -nr);
- mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED);
+ mem_cgroup_update_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED, -nr);

if (unlikely(PageMlocked(page)))
clear_page_mlock(page);
--
2.12.0


2017-03-24 11:08:13

by Kirill A. Shutemov

[permalink] [raw]
Subject: Re: [PATCH] mm: rmap: fix huge file mmap accounting in the memcg stats

On Tue, Mar 21, 2017 at 08:51:11PM -0400, Johannes Weiner wrote:
> Huge pages are accounted as single units in the memcg's "file_mapped"
> counter. Account the correct number of base pages, like we do in the
> corresponding node counter.
>
> Signed-off-by: Johannes Weiner <[email protected]>

Sorry for missing that:

Reviewed-by: Kirill A. Shutemov <[email protected]>

Do we want it into stable?

Cc: <[email protected]> [4.8+]

--
Kirill A. Shutemov

2017-03-24 13:46:36

by Johannes Weiner

[permalink] [raw]
Subject: Re: [PATCH] mm: rmap: fix huge file mmap accounting in the memcg stats

On Fri, Mar 24, 2017 at 02:07:55PM +0300, Kirill A. Shutemov wrote:
> On Tue, Mar 21, 2017 at 08:51:11PM -0400, Johannes Weiner wrote:
> > Huge pages are accounted as single units in the memcg's "file_mapped"
> > counter. Account the correct number of base pages, like we do in the
> > corresponding node counter.
> >
> > Signed-off-by: Johannes Weiner <[email protected]>
>
> Sorry for missing that:
>
> Reviewed-by: Kirill A. Shutemov <[email protected]>

Thanks!

> Do we want it into stable?
>
> Cc: <[email protected]> [4.8+]

Yep, that makes sense.

2017-03-28 15:13:51

by Michal Hocko

[permalink] [raw]
Subject: Re: [PATCH] mm: rmap: fix huge file mmap accounting in the memcg stats

On Tue 21-03-17 20:51:11, Johannes Weiner wrote:
> Huge pages are accounted as single units in the memcg's "file_mapped"
> counter. Account the correct number of base pages, like we do in the
> corresponding node counter.
>
> Signed-off-by: Johannes Weiner <[email protected]>

with the CC: stable
Acked-by: Michal Hocko <[email protected]>

thanks!
> ---
> include/linux/memcontrol.h | 6 ++++++
> mm/rmap.c | 4 ++--
> 2 files changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
> index baa274150210..c5ebb32fef49 100644
> --- a/include/linux/memcontrol.h
> +++ b/include/linux/memcontrol.h
> @@ -741,6 +741,12 @@ static inline bool mem_cgroup_oom_synchronize(bool wait)
> return false;
> }
>
> +static inline void mem_cgroup_update_page_stat(struct page *page,
> + enum mem_cgroup_stat_index idx,
> + int nr)
> +{
> +}
> +
> static inline void mem_cgroup_inc_page_stat(struct page *page,
> enum mem_cgroup_stat_index idx)
> {
> diff --git a/mm/rmap.c b/mm/rmap.c
> index 1d82057144ba..f514cdd84482 100644
> --- a/mm/rmap.c
> +++ b/mm/rmap.c
> @@ -1154,7 +1154,7 @@ void page_add_file_rmap(struct page *page, bool compound)
> goto out;
> }
> __mod_node_page_state(page_pgdat(page), NR_FILE_MAPPED, nr);
> - mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED);
> + mem_cgroup_update_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED, nr);
> out:
> unlock_page_memcg(page);
> }
> @@ -1194,7 +1194,7 @@ static void page_remove_file_rmap(struct page *page, bool compound)
> * pte lock(a spinlock) is held, which implies preemption disabled.
> */
> __mod_node_page_state(page_pgdat(page), NR_FILE_MAPPED, -nr);
> - mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED);
> + mem_cgroup_update_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED, -nr);
>
> if (unlikely(PageMlocked(page)))
> clear_page_mlock(page);
> --
> 2.12.0
>

--
Michal Hocko
SUSE Labs