Re: [PATCH -mm] memcg: do not trigger OOM from add_to_page_cache_locked

[Johannes Weiner <hannes@cmpxchg.org>]

On Mon, Nov 26, 2012 at 09:08:48PM +0100, Michal Hocko wrote:
> On Mon 26-11-12 14:29:41, Johannes Weiner wrote:
> > On Mon, Nov 26, 2012 at 08:03:29PM +0100, Michal Hocko wrote:
> > > On Mon 26-11-12 13:24:21, Johannes Weiner wrote:
> > > > On Mon, Nov 26, 2012 at 07:04:44PM +0100, Michal Hocko wrote:
> > > > > On Mon 26-11-12 12:46:22, Johannes Weiner wrote:
> > > [...]
> > > > > > I think global oom already handles this in a much better way: invoke
> > > > > > the OOM killer, sleep for a second, then return to userspace to
> > > > > > relinquish all kernel resources and locks.  The only reason why we
> > > > > > can't simply change from an endless retry loop is because we don't
> > > > > > want to return VM_FAULT_OOM and invoke the global OOM killer.
> > > > > 
> > > > > Exactly.
> > > > > 
> > > > > > But maybe we can return a new VM_FAULT_OOM_HANDLED for memcg OOM and
> > > > > > just restart the pagefault.  Return -ENOMEM to the buffered IO syscall
> > > > > > respectively.  This way, the memcg OOM killer is invoked as it should
> > > > > > but nobody gets stuck anywhere livelocking with the exiting task.
> > > > > 
> > > > > Hmm, we would still have a problem with oom disabled (aka user space OOM
> > > > > killer), right? All processes but those in mem_cgroup_handle_oom are
> > > > > risky to be killed.
> > > > 
> > > > Could we still let everybody get stuck in there when the OOM killer is
> > > > disabled and let userspace take care of it?
> > > 
> > > I am not sure what exactly you mean by "userspace take care of it" but
> > > if those processes are stuck and holding the lock then it is usually
> > > hard to find that out. Well if somebody is familiar with internal then
> > > it is doable but this makes the interface really unusable for regular
> > > usage.
> > 
> > If oom_kill_disable is set, then all processes get stuck all the way
> > down in the charge stack.  Whatever resource they pin, you may
> > deadlock on if you try to touch it while handling the problem from
> > userspace.
> 
> OK, I guess I am getting what you are trying to say. So what you are
> suggesting is to just let mem_cgroup_out_of_memory send the signal and
> move on without retry (or with few charge retries without further OOM
> killing) and fail the charge with your new FAULT_OOM_HANDLED (resp.
> something like FAULT_RETRY) error code resp. ENOMEM depending on the
> caller.  OOM disabled case would be "you are on your own" because this
> has been dangerous anyway. Correct?

Yes.

> I do agree that the current endless retry loop is far from being ideal
> and can see some updates but I am quite nervous about any potential
> regressions in this area (e.g. too aggressive OOM etc...). I have to
> think about it some more.

Agreed on all points.  Maybe we can keep a couple of the oom retry
iterations or something like that, which is still much more than what
global does and I don't think the global OOM killer is overly eager.

Testing will show more.

> Anyway if you have some more specific ideas I would be happy to review
> patches.

Okay, I just wanted to check back with you before going down this
path.  What are we going to do short term, though?  Do you want to
push the disable-oom-for-pagecache for now or should we put the
VM_FAULT_OOM_HANDLED fix in the next version and do stable backports?

This issue has been around for a while so frankly I don't think it's
urgent enough to rush things.