Re: [PATCH 2/2] mm, oom_reaper: do not mmput synchronously from the oom reaper context

[Michal Hocko]

On Fri 20-05-16 10:30:53, Minchan Kim wrote:
> Forking new thread because my comment is not related to this patch's
> purpose but found a thing during reading this patch.
> 
> On Tue, Apr 26, 2016 at 04:04:30PM +0200, Michal Hocko wrote:
> > From: Michal Hocko <mhocko@xxxxxxxx>
> > 
> > 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.
> 
> The mmput_async is used for only OOM reaper which is enabled on CONFIG_MMU.
> So until someone who want to use mmput_async in !CONFIG_MMU come out,
> we could save sizeof(struct work_struct) per mm in !CONFIG_MMU.

You are right. What about the following?
---