Re: OOM Killer don't works at all if the system have >gigabytesmemory (was Re: [PATCH] mm: check zone->all_unreclaimable inall_unreclaimable())

[David Rientjes]

On Thu, 12 May 2011, Minchan Kim wrote:

> > processes a 1% bonus for every 30% of memory they use as proposed
> > earlier.)
> 
> I didn't follow earlier your suggestion.
> But it's not formal patch so I expect if you send formal patch to
> merge, you would write down the rationale.
> 

Yes, I'm sure we'll still have additional discussion when KOSAKI-san 
replies to my review of his patchset, so this quick patch was written only 
for CAI's testing at this point.

In reference to the above, I think that giving root processes a 3% bonus 
at all times may be a bit aggressive.  As mentioned before, I don't think 
that all root processes using 4% of memory and the remainder of system 
threads are using 1% should all be considered equal.  At the same time, I 
do not believe that two threads using 50% of memory should be considered 
equal if one is root and one is not.  So my idea was to discount 1% for 
every 30% of memory that a root process uses rather than a strict 3%.

That change can be debated and I think we'll probably settle on something 
more aggressive like 1% for every 10% of memory used since oom scores are 
only useful in comparison to other oom scores: in the above scenario where 
there are two threads, one by root and one not by root, using 50% of 
memory each, I think it would be legitimate to give the root task a 5% 
bonus so that it would only be selected if no other threads used more than 
44% of memory (even though the root thread is truly using 50%).

This is a heuristic within the oom killer badness scoring that can always 
be debated back and forth, but I think a 1% bonus for root processes for 
every 10% of memory used is plausible.

Comments?

> > diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> > --- a/mm/oom_kill.c
> > +++ b/mm/oom_kill.c
> > @@ -160,7 +160,7 @@ unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> > Â Â Â Â */
> > Â Â Â Âif (p->flags & PF_OOM_ORIGIN) {
> > Â Â Â Â Â Â Â Âtask_unlock(p);
> > - Â Â Â Â Â Â Â return 1000;
> > + Â Â Â Â Â Â Â return 10000;
> > Â Â Â Â}
> >
> > Â Â Â Â/*
> > @@ -177,32 +177,32 @@ unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> > Â Â Â Âpoints = get_mm_rss(p->mm) + p->mm->nr_ptes;
> > Â Â Â Âpoints += get_mm_counter(p->mm, MM_SWAPENTS);
> >
> > - Â Â Â points *= 1000;
> > + Â Â Â points *= 10000;
> > Â Â Â Âpoints /= totalpages;
> > Â Â Â Âtask_unlock(p);
> >
> > Â Â Â Â/*
> > - Â Â Â Â* Root processes get 3% bonus, just like the __vm_enough_memory()
> > - Â Â Â Â* implementation used by LSMs.
> > + Â Â Â Â* Root processes get 1% bonus per 30% memory used for a total of 3%
> > + Â Â Â Â* possible just like LSMs.
> > Â Â Â Â */
> > Â Â Â Âif (has_capability_noaudit(p, CAP_SYS_ADMIN))
> > - Â Â Â Â Â Â Â points -= 30;
> > + Â Â Â Â Â Â Â points -= 100 * (points / 3000);
> >
> > Â Â Â Â/*
> > Â Â Â Â * /proc/pid/oom_score_adj ranges from -1000 to +1000 such that it may
> > Â Â Â Â * either completely disable oom killing or always prefer a certain
> > Â Â Â Â * task.
> > Â Â Â Â */
> > - Â Â Â points += p->signal->oom_score_adj;
> > + Â Â Â points += p->signal->oom_score_adj * 10;
> >
> > Â Â Â Â/*
> > Â Â Â Â * Never return 0 for an eligible task that may be killed since it's
> > - Â Â Â Â* possible that no single user task uses more than 0.1% of memory and
> > + Â Â Â Â* possible that no single user task uses more than 0.01% of memory and
> > Â Â Â Â * no single admin tasks uses more than 3.0%.
> > Â Â Â Â */
> > Â Â Â Âif (points <= 0)
> > Â Â Â Â Â Â Â Âreturn 1;
> > - Â Â Â return (points < 1000) ? points : 1000;
> > + Â Â Â return (points < 10000) ? points : 10000;
> > Â}
> >
> > Â/*
> > @@ -314,7 +314,7 @@ static struct task_struct *select_bad_process(unsigned int *ppoints,
> > Â Â Â Â Â Â Â Â Â Â Â Â */
> > Â Â Â Â Â Â Â Â Â Â Â Âif (p == current) {
> > Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Âchosen = p;
> > - Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â *ppoints = 1000;
> > + Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â *ppoints = 10000;
> 
> Scattering constant value isn't good.
> You are proving it now.
> I think you did it since this is not a formal patch.
> I expect you will define new value (ex, OOM_INTERNAL_MAX_SCORE or whatever)
> 

Right, we could probably do something like

	#define OOM_SCORE_MAX_FACTOR	10
	#define OOM_SCORE_MAX		(OOM_SCORE_ADJ_MAX * OOM_SCORE_MAX_FACTOR)

in mm/oom_kill.c, which would then be used to replace all of the constants 
above since OOM_SCORE_ADJ_MAX is already defined to be 1000 in 
include/linux/oom.h.