* [PATCH 1/2] mm, oom_reaper: hide oom reaped tasks from OOM killer more carefully
2016-04-26 14:04 [PATCH 0/2] last pile of oom_reaper patches for now Michal Hocko
@ 2016-04-26 14:04 ` Michal Hocko
2016-04-26 14:04 ` [PATCH 2/2] mm, oom_reaper: do not mmput synchronously from the oom reaper context Michal Hocko
1 sibling, 0 replies; 5+ messages in thread
From: Michal Hocko @ 2016-04-26 14:04 UTC (permalink / raw)
To: Andrew Morton; +Cc: Tetsuo Handa, David Rientjes, linux-mm, LKML, Michal Hocko
From: Michal Hocko <mhocko@suse.com>
36324a990cf5 ("oom: clear TIF_MEMDIE after oom_reaper managed to unmap
the address space") not only clears TIF_MEMDIE for oom reaped task
but also set OOM_SCORE_ADJ_MIN for the target task to hide it from
the oom killer. This works in simple cases but it is not sufficient
for (unlikely) cases where the mm is shared between independent
processes (as they do not share signal struct). If the mm had only
small amount of memory which could be reaped then another task
sharing the mm could be selected and that wouldn't help to move out from
the oom situation.
Introduce MMF_OOM_REAPED mm flag which is checked in oom_badness (same
as OOM_SCORE_ADJ_MIN) and task is skipped if the flag is set. Set the
flag after __oom_reap_task is done with a task. This will force the
select_bad_process() to ignore all already oom reaped tasks as well as
no such task is sacrificed for its parent.
Signed-off-by: Michal Hocko <mhocko@suse.com>
---
include/linux/sched.h | 1 +
mm/oom_kill.c | 9 +++++++--
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index acfc32b30704..7bd0fa9db199 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -512,6 +512,7 @@ static inline int get_dumpable(struct mm_struct *mm)
#define MMF_HAS_UPROBES 19 /* has uprobes */
#define MMF_RECALC_UPROBES 20 /* MMF_HAS_UPROBES can be wrong */
+#define MMF_OOM_REAPED 21 /* mm has been already reaped */
#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)
diff --git a/mm/oom_kill.c b/mm/oom_kill.c
index 415f7eb913fa..c0376efa79ec 100644
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -174,8 +174,13 @@ unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *memcg,
if (!p)
return 0;
+ /*
+ * Do not even consider tasks which are explicitly marked oom
+ * unkillable or have been already oom reaped.
+ */
adj = (long)p->signal->oom_score_adj;
- if (adj == OOM_SCORE_ADJ_MIN) {
+ if (adj == OOM_SCORE_ADJ_MIN ||
+ test_bit(MMF_OOM_REAPED, &p->mm->flags)) {
task_unlock(p);
return 0;
}
@@ -513,7 +518,7 @@ static bool __oom_reap_task(struct task_struct *tsk)
* This task can be safely ignored because we cannot do much more
* to release its memory.
*/
- tsk->signal->oom_score_adj = OOM_SCORE_ADJ_MIN;
+ set_bit(MMF_OOM_REAPED, &mm->flags);
out:
mmput(mm);
return ret;
--
2.8.0.rc3
^ permalink raw reply [flat|nested] 5+ messages in thread* [PATCH 2/2] mm, oom_reaper: do not mmput synchronously from the oom reaper context
2016-04-26 14:04 [PATCH 0/2] last pile of oom_reaper patches for now Michal Hocko
2016-04-26 14:04 ` [PATCH 1/2] mm, oom_reaper: hide oom reaped tasks from OOM killer more carefully Michal Hocko
@ 2016-04-26 14:04 ` Michal Hocko
2016-04-26 14:18 ` kbuild test robot
1 sibling, 1 reply; 5+ messages in thread
From: Michal Hocko @ 2016-04-26 14:04 UTC (permalink / raw)
To: Andrew Morton; +Cc: Tetsuo Handa, David Rientjes, linux-mm, LKML, Michal Hocko
From: Michal Hocko <mhocko@suse.com>
Tetsuo has properly noted that mmput slow path might get blocked waiting
for another party (e.g. exit_aio waits for an IO). If that happens the
oom_reaper would be put out of the way and will not be able to process
next oom victim. We should strive for making this context as reliable
and independent on other subsystems as much as possible.
Introduce mmput_async which will perform the slow path from an async
(WQ) context. This will delay the operation but that shouldn't be a
problem because the oom_reaper has reclaimed the victim's address space
for most cases as much as possible and the remaining context shouldn't
bind too much memory anymore. The only exception is when mmap_sem
trylock has failed which shouldn't happen too often.
The issue is only theoretical but not impossible.
Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Michal Hocko <mhocko@suse.com>
---
include/linux/mm_types.h | 1 +
include/linux/sched.h | 5 +++++
kernel/fork.c | 50 +++++++++++++++++++++++++++++++++---------------
mm/oom_kill.c | 8 ++++++--
4 files changed, 47 insertions(+), 17 deletions(-)
diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
index bf4f3ac5110c..f4f477244679 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -513,6 +513,7 @@ struct mm_struct {
#ifdef CONFIG_HUGETLB_PAGE
atomic_long_t hugetlb_usage;
#endif
+ struct work_struct async_put_work;
};
static inline void mm_init_cpumask(struct mm_struct *mm)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 7bd0fa9db199..df8778e72211 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -2604,6 +2604,11 @@ static inline void mmdrop(struct mm_struct * mm)
/* mmput gets rid of the mappings and all user-space */
extern void mmput(struct mm_struct *);
+/* same as above but performs the slow path from the async kontext. Can
+ * be called from the atomic context as well
+ */
+extern void mmput_async(struct mm_struct *);
+
/* Grab a reference to a task's mm, if it is not already going away */
extern struct mm_struct *get_task_mm(struct task_struct *task);
/*
diff --git a/kernel/fork.c b/kernel/fork.c
index accb7221d547..10b0f771d795 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -696,6 +696,26 @@ void __mmdrop(struct mm_struct *mm)
}
EXPORT_SYMBOL_GPL(__mmdrop);
+static inline void __mmput(struct mm_struct *mm)
+{
+ VM_BUG_ON(atomic_read(&mm->mm_users));
+
+ uprobe_clear_state(mm);
+ exit_aio(mm);
+ ksm_exit(mm);
+ khugepaged_exit(mm); /* must run before exit_mmap */
+ exit_mmap(mm);
+ set_mm_exe_file(mm, NULL);
+ if (!list_empty(&mm->mmlist)) {
+ spin_lock(&mmlist_lock);
+ list_del(&mm->mmlist);
+ spin_unlock(&mmlist_lock);
+ }
+ if (mm->binfmt)
+ module_put(mm->binfmt->module);
+ mmdrop(mm);
+}
+
/*
* Decrement the use count and release all resources for an mm.
*/
@@ -703,24 +723,24 @@ void mmput(struct mm_struct *mm)
{
might_sleep();
+ if (atomic_dec_and_test(&mm->mm_users))
+ __mmput(mm);
+}
+EXPORT_SYMBOL_GPL(mmput);
+
+static void mmput_async_fn(struct work_struct *work)
+{
+ struct mm_struct *mm = container_of(work, struct mm_struct, async_put_work);
+ __mmput(mm);
+}
+
+void mmput_async(struct mm_struct *mm)
+{
if (atomic_dec_and_test(&mm->mm_users)) {
- uprobe_clear_state(mm);
- exit_aio(mm);
- ksm_exit(mm);
- khugepaged_exit(mm); /* must run before exit_mmap */
- exit_mmap(mm);
- set_mm_exe_file(mm, NULL);
- if (!list_empty(&mm->mmlist)) {
- spin_lock(&mmlist_lock);
- list_del(&mm->mmlist);
- spin_unlock(&mmlist_lock);
- }
- if (mm->binfmt)
- module_put(mm->binfmt->module);
- mmdrop(mm);
+ INIT_WORK(&mm->async_put_work, mmput_async_fn);
+ schedule_work(&mm->async_put_work);
}
}
-EXPORT_SYMBOL_GPL(mmput);
/**
* set_mm_exe_file - change a reference to the mm's executable file
diff --git a/mm/oom_kill.c b/mm/oom_kill.c
index c0376efa79ec..c0e37dd1422f 100644
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -446,7 +446,6 @@ static DECLARE_WAIT_QUEUE_HEAD(oom_reaper_wait);
static struct task_struct *oom_reaper_list;
static DEFINE_SPINLOCK(oom_reaper_lock);
-
static bool __oom_reap_task(struct task_struct *tsk)
{
struct mmu_gather tlb;
@@ -520,7 +519,12 @@ static bool __oom_reap_task(struct task_struct *tsk)
*/
set_bit(MMF_OOM_REAPED, &mm->flags);
out:
- mmput(mm);
+ /*
+ * Drop our reference but make sure the mmput slow path is called from a
+ * different context because we shouldn't risk we get stuck there and
+ * put the oom_reaper out of the way.
+ */
+ mmput_async(mm);
return ret;
}
--
2.8.0.rc3
^ permalink raw reply [flat|nested] 5+ messages in thread