Re: general protection fault in oom_unkillable_task

[Shakeel Butt]

On Mon, Jun 17, 2019 at 6:45 PM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> On Mon, 17 Jun 2019 06:23:07 -0700 Shakeel Butt <shakeelb@xxxxxxxxxx> wrote:
>
> > > Here is a patch to use CSS_TASK_ITER_PROCS.
> > >
> > > From 415e52cf55bc4ad931e4f005421b827f0b02693d Mon Sep 17 00:00:00 2001
> > > From: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
> > > Date: Mon, 17 Jun 2019 00:09:38 +0900
> > > Subject: [PATCH] mm: memcontrol: Use CSS_TASK_ITER_PROCS at mem_cgroup_scan_tasks().
> > >
> > > Since commit c03cd7738a83b137 ("cgroup: Include dying leaders with live
> > > threads in PROCS iterations") corrected how CSS_TASK_ITER_PROCS works,
> > > mem_cgroup_scan_tasks() can use CSS_TASK_ITER_PROCS in order to check
> > > only one thread from each thread group.
> > >
> > > Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
> >
> > Reviewed-by: Shakeel Butt <shakeelb@xxxxxxxxxx>
> >
> > Why not add the reproducer in the commit message?
>
> That would be nice.
>
> More nice would be, as always, a descriptoin of the user-visible impact
> of the patch.
>

This is just a cleanup and optimization where instead of traversing
all the threads in a memcg, we only traverse only one thread for each
thread group in a memcg. There is no user visible impact.

> As I understand it, it's just a bit of a cleanup against current
> mainline but without this patch in place, Shakeel's "mm, oom: refactor
> dump_tasks for memcg OOMs" will cause kernel crashes.  Correct?

No, the patch "mm, oom: refactor dump_tasks for memcg OOMs" is making
dump_stacks not depend on the memcg check within
oom_unkillable_task().

"mm, oom: fix oom_unkillable_task for memcg OOMs" is the actual fix
which is making oom_unkillable_task() correctly handle the memcg OOMs
code paths.

Shakeel