mm, memcg: fix inconsistent oom event behavior
Subsystems (3)
PATCH v2 1/1 mm, memcg: fix inconsistent oom event behavior
A recent commit 9852ae3fe529 ("mm, memcg: consider subtrees in memory.events") changes the behavior of memcg events, which will consider subtrees in memory.events. But oom_kill event is a special one as it is used in both cgroup1 and cgroup2. In cgroup1, it is displayed in memory.oom_control. The file memory.oom_control is in both root memcg and non root memcg, that is different with memory.event as it only in non-root memcg. That commit is okay for cgroup2, but it is not okay for cgroup1 as it will cause inconsistent behavior between root memcg and non-root memcg.
Here's an example on why this behavior is inconsistent in cgroup1.
root memcg
/ memcg foo
/
memcg bar
Suppose there's an oom_kill in memcg bar, then the oon_kill will be
root memcg : memory.oom_control(oom_kill) 0
/
memcg foo : memory.oom_control(oom_kill) 1
/
memcg bar : memory.oom_control(oom_kill) 1
For the non-root memcg, its memory.oom_control(oom_kill) includes its descendants' oom_kill, but for root memcg, it doesn't include its descendants' oom_kill. That means, memory.oom_control(oom_kill) has different meanings in different memcgs. That is inconsistent. Then the user has to know whether the memcg is root or not.
If we can't fully support it in cgroup1, for example by adding memory.events.local into cgroup1 as well, then let's don't touch its original behavior. So let's recover the original behavior for cgroup1.
--- include/linux/memcontrol.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
include/linux/memcontrol.h +2 -1
@@ -798,7 +798,8 @@ static inline void memcg_memory_event(struct mem_cgroup *memcg,atomic_long_inc (&memcg->memory_events[event]);cgroup_file_notify (&memcg->events_file); -if (cgrp_dfl_root.flags &CGRP_ROOT_MEMORY_LOCAL_EVENTS ) +if (cgrp_dfl_root.flags &CGRP_ROOT_MEMORY_LOCAL_EVENTS || + !cgroup_subsys_on_dfl (memory_cgrp_subsys))break ; }while ((memcg =parent_mem_cgroup (memcg)) && !mem_cgroup_is_root (memcg));
Signature · 1 line
2.18.2
Michal Hocko
maintainerQuoted 31 lines · Yafang Shao
A recent commit
9852ae3fe529("mm, memcg: consider subtrees in memory.events") changes the behavior of memcg events, which will consider subtrees in memory.events. Butoom_killevent is a special one as it is used in both cgroup1 and cgroup2. In cgroup1, it is displayed in memory.oom_control. The file memory.oom_control is in both root memcg and non root memcg, that is different with memory.event as it only in non-root memcg. That commit is okay for cgroup2, but it is not okay for cgroup1 as it will cause inconsistent behavior between root memcg and non-root memcg.Here's an example on why this behavior is inconsistent in cgroup1.
root memcg /memcg foo
/
memcg barSuppose there's an
oom_killin memcg bar, then theoon_killwill beroot memcg :
memory.oom_control(oom_kill)0
/
memcg foo :memory.oom_control(oom_kill)1
/
memcg bar :memory.oom_control(oom_kill)1For the non-root memcg, its
memory.oom_control(oom_kill)includes its descendants'oom_kill, but for root memcg, it doesn't include its descendants'oom_kill. That means,memory.oom_control(oom_kill)has different meanings in different memcgs. That is inconsistent. Then the user has to know whether the memcg is root or not.
Quoted 3 lines
If we can't fully support it in cgroup1, for example by adding memory.events.local into cgroup1 as well, then let's don't touch its original behavior. So let's recover the original behavior for cgroup1.
Wthe localevents was mostly cgroup v2 feature. I do not think there was an intention to have side effects on the legacy hierarchy. I thought this would be the case but it is not apparently. Would it make more sense to have CGRP_ROOT_MEMORY_LOCAL_EVENTS for legacy hierarchy by default rather than special casing it somewhere quite deep in the code?
Quoted 26 lines
Fixes 9852ae3fe529 ("mm, memcg: consider subtrees in memory.events")Cc Chris Down <[email protected]>Cc Johannes Weiner <[email protected]>Reviewed-by Shakeel Butt <[email protected]>Signed-off-by Yafang Shao <[email protected]>--- include/linux/memcontrol.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)include/linux/memcontrol.h +2
-1@@ -798,7 +798,8 @@ static inline void memcg_memory_event(struct mem_cgroup *memcg,atomic_long_inc (&memcg->memory_events[event]);cgroup_file_notify (&memcg->events_file); -if (cgrp_dfl_root.flags &CGRP_ROOT_MEMORY_LOCAL_EVENTS ) +if (cgrp_dfl_root.flags &CGRP_ROOT_MEMORY_LOCAL_EVENTS || + !cgroup_subsys_on_dfl (memory_cgrp_subsys))break ; }while ((memcg =parent_mem_cgroup (memcg)) && !mem_cgroup_is_root (memcg));Signature · 1 line
2.18.2
Signature · 2 lines
Michal Hocko SUSE Labs
Quoted 37 lines · Michal Hocko
Quoted 31 lines · Yafang Shao
A recent commit
9852ae3fe529("mm, memcg: consider subtrees in memory.events") changes the behavior of memcg events, which will consider subtrees in memory.events. Butoom_killevent is a special one as it is used in both cgroup1 and cgroup2. In cgroup1, it is displayed in memory.oom_control. The file memory.oom_control is in both root memcg and non root memcg, that is different with memory.event as it only in non-root memcg. That commit is okay for cgroup2, but it is not okay for cgroup1 as it will cause inconsistent behavior between root memcg and non-root memcg.Here's an example on why this behavior is inconsistent in cgroup1.
root memcg /memcg foo
/
memcg barSuppose there's an
oom_killin memcg bar, then theoon_killwill beroot memcg :
memory.oom_control(oom_kill)0
/
memcg foo :memory.oom_control(oom_kill)1
/
memcg bar :memory.oom_control(oom_kill)1For the non-root memcg, its
memory.oom_control(oom_kill)includes its descendants'oom_kill, but for root memcg, it doesn't include its descendants'oom_kill. That means,memory.oom_control(oom_kill)has different meanings in different memcgs. That is inconsistent. Then the user has to know whether the memcg is root or not.Quoted 3 lines
If we can't fully support it in cgroup1, for example by adding memory.events.local into cgroup1 as well, then let's don't touch its original behavior. So let's recover the original behavior for cgroup1.
Quoted 6 lines
Wthe localevents was mostly cgroup v2 feature. I do not think there was an intention to have side effects on the legacy hierarchy. I thought this would be the case but it is not apparently. Would it make more sense to have
CGRP_ROOT_MEMORY_LOCAL_EVENTSfor legacy hierarchy by default rather than special casing it somewhere quite deep in the code?
I had thought about setting CGRP_ROOT_MEMORY_LOCAL_EVENTS by defualt for cgroup1, but I was not sure whether we should also expose memory_localevents in cgroup1_show_options().
Quoted 18 lines
Quoted 18 lines
Fixes 9852ae3fe529 ("mm, memcg: consider subtrees in memory.events")Cc Chris Down <[email protected]>Cc Johannes Weiner <[email protected]>Reviewed-by Shakeel Butt <[email protected]>Signed-off-by Yafang Shao <[email protected]>--- include/linux/memcontrol.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)include/linux/memcontrol.h +0
-0@@ -798,7 +798,8 @@ static inline void memcg_memory_event(struct mem_cgroup *memcg,atomic_long_inc (&memcg->memory_events[event]);cgroup_file_notify (&memcg->events_file);
Quoted 12 lines
Quoted 8 lines
include/linux/memcontrol.h +2
-2-if (cgrp_dfl_root.flags &CGRP_ROOT_MEMORY_LOCAL_EVENTS ) +if (cgrp_dfl_root.flags &CGRP_ROOT_MEMORY_LOCAL_EVENTS || + !cgroup_subsys_on_dfl (memory_cgrp_subsys))break ; }while ((memcg =parent_mem_cgroup (memcg)) && !mem_cgroup_is_root (memcg)); --2.18.2
--
Michal Hocko
SUSE Labs
Thanks
Yafang