linux-kernel - Re: [PATCH v3] mm, sysctl: make NUMA stats configurable

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <d65b7725-238b-a92c-6ea7-6621696d0711@suse.cz>
Date:   Wed, 11 Oct 2017 08:16:24 +0200
From:   Vlastimil Babka <vbabka@...e.cz>
To:     Dave Hansen <dave.hansen@...ux.intel.com>,
        Michal Hocko <mhocko@...nel.org>
Cc:     kemi <kemi.wang@...el.com>,
        "Luis R . Rodriguez" <mcgrof@...nel.org>,
        Kees Cook <keescook@...omium.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Jonathan Corbet <corbet@....net>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Johannes Weiner <hannes@...xchg.org>,
        Christopher Lameter <cl@...ux.com>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Tim Chen <tim.c.chen@...el.com>,
        Andi Kleen <andi.kleen@...el.com>,
        Jesper Dangaard Brouer <brouer@...hat.com>,
        Ying Huang <ying.huang@...el.com>,
        Aaron Lu <aaron.lu@...el.com>,
        Proc sysctl <linux-fsdevel@...r.kernel.org>,
        Linux MM <linux-mm@...ck.org>,
        Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3] mm, sysctl: make NUMA stats configurable

On 10/10/2017 05:39 PM, Dave Hansen wrote:
> On 10/10/2017 07:57 AM, Michal Hocko wrote:
>>> But, let's be honest, this leaves us with an option that nobody is ever
>>> going to turn on.  IOW, nobody except a very small portion of our users
>>> will ever see any benefit from this.
>> But aren't those small groups who would like to squeeze every single
>> cycle out from the page allocator path the targeted audience?
> 
> They're the reason we started looking at this.  They also care the most.
> 
> But, the cost of these stats, especially we get more and more cores in a
> NUMA node is really making them show up in profiles.  It would be nice
> to get rid of them there, too.

Furthermore, the group that actually looks at those stats, could be also
expected to be quite small. The group that cares neither about the
stats, nor relies on top allocator performance, might still arguably
benefit from improved allocator performance, but won't for sure benefit
from the stats.

> Aaron, do you remember offhand how much of the allocator overhead was
> coming from NUMA stats?
>