Re: çå: çå: [PATCH] mm/memcontrol.c: speed up to force empty a memory cgroup

[David Rientjes]

On Wed, 21 Mar 2018, Andrey Ryabinin wrote:

> > Is SWAP_CLUSTER_MAX the best answer if I'm lowering the limit by 1GB?
> > 
> 
> Absolutely not. I completely on your side here. 
> I've tried to fix this recently - http://lkml.kernel.org/r/20180119132544.19569-2-aryabinin@xxxxxxxxxxxxx
> I guess that Andrew decided to not take my patch, because Michal wasn't
> happy about it (see mail archives if you want more details).
> 

I unfortunately didn't see this patch in January, it seems very similar to 
what I was suggesting in this thread.  You do a page_counter_read() 
directly in mem_cgroup_resize_limit() where my suggestion was to have 
page_counter_limit() return the difference, but there's nothing 
significantly different about what you proposed and what I suggested.

Perhaps the patch would be better off as a compromise between what you, I, 
and Li RongQing have proposed/suggested: have page_counter_limit() return 
the difference, and clamp it to some value proportional to the new limit.