Re: [PATCH] mm: Move check for SHRINKER_NUMA_AWARE to do_shrink_slab()

[Kirill Tkhai]

On 02.08.2018 23:47, Andrew Morton wrote:
> On Thu, 02 Aug 2018 14:00:52 +0300 Kirill Tkhai <ktkhai@xxxxxxxxxxxxx> wrote:
> 
>> In case of shrink_slab_memcg() we do not zero nid, when shrinker
>> is not numa-aware. This is not a real problem, since currently
>> all memcg-aware shrinkers are numa-aware too (we have two:
>> super_block shrinker and workingset shrinker), but something may
>> change in the future.
> 
> Fair enough.
> 
>> (Andrew, this may be merged to mm-iterate-only-over-charged-shrinkers-during-memcg-shrink_slab)
> 
> It got a bit messy so I got lazy and queued it as a separate patch.
> 
> btw, I have a note that https://lkml.org/lkml/2018/7/7/32 was caused by
> this patch series.  Is that the case and do you know if this was
> addressed?

It's not related to the patchset. Bisect leads to:

commit c6aeb9d4c351 (HEAD, refs/bisect/bad)
Author: David Howells <dhowells@xxxxxxxxxx>
Date:   Sun Jun 24 00:20:10 2018 +0100

kernfs, sysfs, cgroup, intel_rdt: Support fs_context

CC David.

David, please see reproducer at https://lkml.org/lkml/2018/7/7/32

Kirill