This patchset makes the OOM killer cgroup-aware.
Patch 1 causes out_of_memory() look at the oom_victim counter
to decide if a new victim is required.
Patch 2 is main patch which implements cgroup-aware OOM killer.
Patch 3 adds some debug output, which can be refined later.
Patch 4 introduces per-cgroup oom_score_adj knob.
Patch 5 fixes a problem with too many processes receiving an
access to the memory reserves.
Patch 6 is docs update.
v1:
https://lkml.org/lkml/2017/5/18/969
v2:
- Reworked victim selection based on feedback
from Michal Hocko, Vladimir Davydov and Johannes Weiner
- "Kill all tasks" is now an opt-in option, by default
only one process will be killed
- Added per-cgroup oom_score_adj
- Refined oom score calculations, suggested by Vladimir Davydov
- Converted to a patchset
v3:
- Fixed swap accounting
- Switched to use oom_victims counter to prevent unnecessary kills
- TIF_MEMDIE is set only when necessary
- Moved all oom victim killing code into oom_kill.c
- Merged commits 1-4 into 6
- Separated oom_score_adj logic into a separate commit 4
- Separated debug output into a separate commit 3
Roman Gushchin (6):
mm, oom: use oom_victims counter to synchronize oom victim selection
mm, oom: cgroup-aware OOM killer
mm, oom: cgroup-aware OOM killer debug info
mm, oom: introduce oom_score_adj for memory cgroups
mm, oom: don't mark all oom victims tasks with TIF_MEMDIE
mm,oom,docs: describe the cgroup-aware OOM killer
Documentation/cgroup-v2.txt | 44 ++++++++++
include/linux/memcontrol.h | 23 +++++
include/linux/oom.h | 3 +
kernel/exit.c | 2 +-
mm/memcontrol.c | 209 ++++++++++++++++++++++++++++++++++++++++++++
mm/oom_kill.c | 202 ++++++++++++++++++++++++++++--------------
6 files changed, 416 insertions(+), 67 deletions(-)
--
2.7.4
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>
