linux-kernel - Re: [PATCH V3 3/3] mm: page_alloc: drain pcp lists before oom kill

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 for Android: free password hash cracker in your pocket

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

Message-ID: <ZVYRJMUitykepLRy@tiehlicka>
Date:   Thu, 16 Nov 2023 13:55:00 +0100
From:   Michal Hocko <mhocko@...e.com>
To:     Charan Teja Kalla <quic_charante@...cinc.com>
Cc:     akpm@...ux-foundation.org, mgorman@...hsingularity.net,
        david@...hat.com, vbabka@...e.cz, hannes@...xchg.org,
        quic_pkondeti@...cinc.com, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH V3 3/3] mm: page_alloc: drain pcp lists before oom kill

On Thu 16-11-23 11:30:04, Charan Teja Kalla wrote:
> Thanks Michal.
> 
> On 11/15/2023 7:39 PM, Michal Hocko wrote:
> >> Also If you have any comments on [PATCH V2 2/3] mm: page_alloc: correct
> >> high atomic reserve calculations will help me.
> > I do not have a strong opinion on that one to be honest. I am not even
> > sure that reserving a full page block (4MB) on small systems as
> > presented is really a good use of memory.
> 
> May be other way to look at that patch is comment is really not being
> reflected in the code. It says, " Limit the number reserved to 1
> pageblock or roughly 1% of a zone.", but the current code is making it 2
> pageblocks. So, for 4M block size, it is > 1%.
> 
> A second patch, that I will post, like not reserving the high atomic
> page blocks on small systems -- But how to define the meaning of small
> systems is not sure. Instead will let the system administrators chose
> this through either:
> a) command line param, high_atomic_reserves=off, on by default --
> Another knob, so admins may really not like this?
> b) CONFIG_HIGH_ATOMIC_RESERVES, which if not defined, will not reserve.

Please don't! I do not see any admin wanting to care about this at all.
It just takes a lot of understanding of internal MM stuff to make an
educated guess. This should really be auto-tuned. And as responded in
other reply my take would be to reserve a page block on if it doesn't
consume more than 1% of memory to preserve the existing behavior yet not
overconsume on small systems.
 
> Please lmk If you have any more suggestions here?
> 
> Also, I am thinking to request Andrew to pick [PATCH V2 1/3] patch and
> take these discussions separately in a separate thread.

That makes sense as that is a clear bug fix.
-- 
Michal Hocko
SUSE Labs