Resizing the memcg limit for cgroup-v2 drains the stocks before
triggering the memcg reclaim. Do the same for cgroup-v1 to make the
behavior consistent.
Signed-off-by: Shakeel Butt <[email protected]>
---
mm/memcontrol.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 25b148c2d222..e2d33a37f971 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -2463,6 +2463,7 @@ static int mem_cgroup_resize_max(struct mem_cgroup *memcg,
unsigned long max, bool memsw)
{
bool enlarge = false;
+ bool drained = false;
int ret;
bool limits_invariant;
struct page_counter *counter = memsw ? &memcg->memsw : &memcg->memory;
@@ -2493,6 +2494,12 @@ static int mem_cgroup_resize_max(struct mem_cgroup *memcg,
if (!ret)
break;
+ if (!drained) {
+ drain_all_stock(memcg);
+ drained = true;
+ continue;
+ }
+
if (!try_to_free_mem_cgroup_pages(memcg, 1,
GFP_KERNEL, !memsw)) {
ret = -EBUSY;
--
2.17.0.441.gb46fe60e1d-goog
On Fri, May 04, 2018 at 01:55:48PM -0700, Shakeel Butt wrote:
> Resizing the memcg limit for cgroup-v2 drains the stocks before
> triggering the memcg reclaim. Do the same for cgroup-v1 to make the
> behavior consistent.
>
> Signed-off-by: Shakeel Butt <[email protected]>
Acked-by: Johannes Weiner <[email protected]>
On Fri 04-05-18 13:55:48, Shakeel Butt wrote:
> Resizing the memcg limit for cgroup-v2 drains the stocks before
> triggering the memcg reclaim. Do the same for cgroup-v1 to make the
> behavior consistent.
>
> Signed-off-by: Shakeel Butt <[email protected]>
Acked-by: Michal Hocko <[email protected]>
> ---
> mm/memcontrol.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 25b148c2d222..e2d33a37f971 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -2463,6 +2463,7 @@ static int mem_cgroup_resize_max(struct mem_cgroup *memcg,
> unsigned long max, bool memsw)
> {
> bool enlarge = false;
> + bool drained = false;
> int ret;
> bool limits_invariant;
> struct page_counter *counter = memsw ? &memcg->memsw : &memcg->memory;
> @@ -2493,6 +2494,12 @@ static int mem_cgroup_resize_max(struct mem_cgroup *memcg,
> if (!ret)
> break;
>
> + if (!drained) {
> + drain_all_stock(memcg);
> + drained = true;
> + continue;
> + }
> +
> if (!try_to_free_mem_cgroup_pages(memcg, 1,
> GFP_KERNEL, !memsw)) {
> ret = -EBUSY;
> --
> 2.17.0.441.gb46fe60e1d-goog
--
Michal Hocko
SUSE Labs