Re: [PATCH] memcg: fix per_node_info cleanup

Re: [PATCH] memcg: fix per_node_info cleanup

[Sasha Levin]

Hi,

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag,
fixing commit: 00f3ca2c2d66 mm: memcontrol: per-lruvec stats infrastructure.

The bot has also determined it's probably a bug fixing patch. (score: 40.3266)

The bot has tested the following trees: v4.16.1, v4.15.16, v4.14.33.

v4.16.1: Build OK!
v4.15.16: Failed to apply! Possible dependencies:
    a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting")

v4.14.33: Failed to apply! Possible dependencies:
    a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting")


--
Thanks,
Sasha

Re: [PATCH] memcg: fix per_node_info cleanup

[Michal Hocko]

On Tue 10-04-18 13:50:30, Sasha Levin wrote:
> Hi,
> 
> [This is an automated email]
> 
> This commit has been processed because it contains a "Fixes:" tag,
> fixing commit: 00f3ca2c2d66 mm: memcontrol: per-lruvec stats infrastructure.

Yes and the changelog states conditions which are quite hard to achieve
to have this patch in the stable tree. So I am not really convinced it
is worth backporting even though the patch itself is trivial.
-- 
Michal Hocko
SUSE Labs

Re: [PATCH] memcg: fix per_node_info cleanup

[Sasha Levin]

On Tue, Apr 10, 2018 at 04:01:32PM +0200, Michal Hocko wrote:
>On Tue 10-04-18 13:50:30, Sasha Levin wrote:
>> Hi,
>>
>> [This is an automated email]
>>
>> This commit has been processed because it contains a "Fixes:" tag,
>> fixing commit: 00f3ca2c2d66 mm: memcontrol: per-lruvec stats infrastructure.
>
>Yes and the changelog states conditions which are quite hard to achieve
>to have this patch in the stable tree. So I am not really convinced it
>is worth backporting even though the patch itself is trivial.

Quite hard to achieve, but the changelog also indicated that this was
actually hit and is not just theoretical. Is there something that will
prevent this from happening on "real" workloads?

Re: [PATCH] memcg: fix per_node_info cleanup

[Michal Hocko]

On Tue 10-04-18 17:55:13, Sasha Levin wrote:
> On Tue, Apr 10, 2018 at 04:01:32PM +0200, Michal Hocko wrote:
> >On Tue 10-04-18 13:50:30, Sasha Levin wrote:
> >> Hi,
> >>
> >> [This is an automated email]
> >>
> >> This commit has been processed because it contains a "Fixes:" tag,
> >> fixing commit: 00f3ca2c2d66 mm: memcontrol: per-lruvec stats infrastructure.
> >
> >Yes and the changelog states conditions which are quite hard to achieve
> >to have this patch in the stable tree. So I am not really convinced it
> >is worth backporting even though the patch itself is trivial.
> 
> Quite hard to achieve, but the changelog also indicated that this was
> actually hit and is not just theoretical. Is there something that will
> prevent this from happening on "real" workloads?

Well, small allocations do not really fail and it would take to
configure a really large NUMA system to achieve the same. So I am
skeptical.

-- 
Michal Hocko
SUSE Labs