On 2019/01/26 20:29, Arkadiusz Miśkiewicz wrote:
> On 26/01/2019 12:09, Tetsuo Handa wrote:
>> Arkadiusz, will you try this patch?
>
>
> Works. Several tries and always getting 0 pids.current after ~1s.
>
Thank you for testing.
I updated this patch to use tsk->signal->oom_mm (a snapshot of
tsk->mm saved by mark_oom_victim(tsk)) rather than raw tsk->mm
so that we don't need to worry about possibility of changing
tsk->mm across multiple wake_oom_reaper(tsk) calls.
>From 9c9e935fc038342c48461aabca666f1b544e32b1 Mon Sep 17 00:00:00 2001
From: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
Date: Sat, 26 Jan 2019 21:57:25 +0900
Subject: [PATCH v2] oom, oom_reaper: do not enqueue same task twice
Arkadiusz reported that enabling memcg's group oom killing causes
strange memcg statistics where there is no task in a memcg despite
the number of tasks in that memcg is not 0. It turned out that there
is a bug in wake_oom_reaper() which allows enqueuing same task twice
which makes impossible to decrease the number of tasks in that memcg
due to a refcount leak.
This bug existed since the OOM reaper became invokable from
task_will_free_mem(current) path in out_of_memory() in Linux 4.7,
but memcg's group oom killing made it easier to trigger this bug by
calling wake_oom_reaper() on the same task from one out_of_memory()
request.
Fix this bug using an approach used by commit 855b018325737f76
("oom, oom_reaper: disable oom_reaper for oom_kill_allocating_task").
As a side effect of this patch, this patch also avoids enqueuing
multiple threads sharing memory via task_will_free_mem(current) path.
Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
Reported-by: Arkadiusz Miśkiewicz <arekm@xxxxxxxx>
Tested-by: Arkadiusz Miśkiewicz <arekm@xxxxxxxx>
Fixes: af8e15cc85a25315 ("oom, oom_reaper: do not enqueue task if it is on the oom_reaper_list head")
---
mm/oom_kill.c | 17 +++++++----------
1 file changed, 7 insertions(+), 10 deletions(-)
diff --git a/mm/oom_kill.c b/mm/oom_kill.c
index f0e8cd9..057bfee 100644
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -505,14 +505,6 @@ bool __oom_reap_task_mm(struct mm_struct *mm)
struct vm_area_struct *vma;
bool ret = true;
- /*
- * Tell all users of get_user/copy_from_user etc... that the content
- * is no longer stable. No barriers really needed because unmapping
- * should imply barriers already and the reader would hit a page fault
- * if it stumbled over a reaped memory.
- */
- set_bit(MMF_UNSTABLE, &mm->flags);
-
for (vma = mm->mmap ; vma; vma = vma->vm_next) {
if (!can_madv_dontneed_vma(vma))
continue;
@@ -647,8 +639,13 @@ static int oom_reaper(void *unused)
static void wake_oom_reaper(struct task_struct *tsk)
{
- /* tsk is already queued? */
- if (tsk == oom_reaper_list || tsk->oom_reaper_list)
+ /*
+ * Tell all users of get_user/copy_from_user etc... that the content
+ * is no longer stable. No barriers really needed because unmapping
+ * should imply barriers already and the reader would hit a page fault
+ * if it stumbled over a reaped memory.
+ */
+ if (test_and_set_bit(MMF_UNSTABLE, &tsk->signal->oom_mm->flags))
return;
get_task_struct(tsk);
--
1.8.3.1
