Re: [RFC PATCH 1/2] mempool: do not consume memory reserves from the reclaim path

From: Andrew Morton
Date: Fri Jul 22 2016 - 15:44:54 EST


On Fri, 22 Jul 2016 14:26:19 +0200 Vlastimil Babka <vbabka@xxxxxxx> wrote:

> On 07/22/2016 08:37 AM, Michal Hocko wrote:
> > On Thu 21-07-16 16:53:09, Michal Hocko wrote:
> >> From d64815758c212643cc1750774e2751721685059a Mon Sep 17 00:00:00 2001
> >> From: Michal Hocko <mhocko@xxxxxxxx>
> >> Date: Thu, 21 Jul 2016 16:40:59 +0200
> >> Subject: [PATCH] Revert "mm, mempool: only set __GFP_NOMEMALLOC if there are
> >> free elements"
> >>
> >> This reverts commit f9054c70d28bc214b2857cf8db8269f4f45a5e23.
> >
> > I've noticed that Andrew has already picked this one up. Is anybody
> > against marking it for stable?
>
> It would be strange to have different behavior with known regression in
> 4.6 and 4.7 stables. Actually, there's still time for 4.7 proper?
>

I added the cc:stable.

Do we need to bust a gut to rush it into 4.7? It sounds safer to let
it bake for a while, fix it in 4.7.1?