Re: [PATCH] mm, vmscan: do not loop on too_many_isolated for ever
From: Michal Hocko
Date: Mon Jul 10 2017 - 13:09:50 EST
On Mon 10-07-17 12:58:59, Johannes Weiner wrote:
> On Mon, Jul 10, 2017 at 09:58:03AM -0400, Rik van Riel wrote:
> > On Mon, 2017-07-10 at 09:48 +0200, Michal Hocko wrote:
> > > Johannes and Rik had some concerns that this could lead to premature
> > > OOM kills. I agree with them that we need a better throttling
> > > mechanism. Until now we didn't give the issue described above a high
> > > priority because it usually required a really insane workload to
> > > trigger. But it seems that the issue can be reproduced also without
> > > having an insane number of competing threads .
> > My worries stand, but lets fix the real observed bug, and not worry
> > too much about the theoretical bug for now.
> > Acked-by: Rik van Riel <riel@xxxxxxxxxx>
> I agree with this.
> Acked-by: Johannes Weiner <hannes@xxxxxxxxxxx>
Thanks to both of you. Just to make it clear. I really do want to
address the throttling problem longterm properly. I do not have any
great ideas to be honest. I am busy with other things so it might be
quite some time before I come up with something.