linux-kernel - Re: [PATCH] kswapd: no need reclaim cma pages triggered by unmovable allocation

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]

Date:   Mon, 15 Mar 2021 17:12:40 +0100
From:   David Hildenbrand <david@...hat.com>
To:     Michal Hocko <mhocko@...e.com>
Cc:     zhou <xianrong_zhou@....com>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org, akpm@...ux-foundation.org,
        mgorman@...e.de, willy@...ux.intel.com, rostedt@...dmis.org,
        mingo@...hat.com, vbabka@...e.cz, rientjes@...gle.com,
        pankaj.gupta.linux@...il.com, bhe@...hat.com, ying.huang@...el.com,
        iamjoonsoo.kim@....com, minchan@...nel.org,
        ruxian.feng@...nssion.com, kai.cheng@...nssion.com,
        zhao.xu@...nssion.com, zhouxianrong@....com,
        zhou xianrong <xianrong.zhou@...nssion.com>
Subject: Re: [PATCH] kswapd: no need reclaim cma pages triggered by unmovable
 allocation

On 15.03.21 17:09, Michal Hocko wrote:
> On Mon 15-03-21 16:46:33, David Hildenbrand wrote:
>> On 13.03.21 09:31, zhou wrote:
> [...]
>>> This optimization can avoid ~3% unnecessary isolations from cma
>>> (cma isolated / total isolated) with configuration of total 100Mb
>>> cma pages.
>>
>> Can you say a few words about interaction with ZONE_MOVABLE, which behaves
>> similar to CMA? I.e., does the same apply to ZONE_MOVABLE? Is it already
>> handled?
> 
> No, the movable zone shouldn't be affected as the reclaim is zone aware.
> The problem is that CMA doesn't belong to any particular zone. This is
> something Joonsoo worked in the past and I believe following up on that
> work has been recommended last time a similar/same approach like this
> patch was proposed.

Okay, thanks - that's what I expected.

-- 
Thanks,

David / dhildenb