linux-kernel - RE: [PATCH v2] page_alloc: fix invalid watemark check on a negative value

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: <20220727074619epcms1p7285bceab31c4e04406094eee84862803@epcms1p7>
Date:   Wed, 27 Jul 2022 16:46:19 +0900
From:   Jaewon Kim <jaewon31.kim@...sung.com>
To:     Andrew Morton <akpm@...ux-foundation.org>,
        Jaewon Kim <jaewon31.kim@...sung.com>
CC:     "minchan@...nel.org" <minchan@...nel.org>,
        "bhe@...hat.com" <bhe@...hat.com>,
        "vbabka@...e.cz" <vbabka@...e.cz>,
        "mgorman@...hsingularity.net" <mgorman@...hsingularity.net>,
        "hannes@...xchg.org" <hannes@...xchg.org>,
        "mhocko@...nel.org" <mhocko@...nel.org>,
        "linux-mm@...ck.org" <linux-mm@...ck.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        GyeongHwan Hong <gh21.hong@...sung.com>,
        YongTaek Lee <ytk.lee@...sung.com>,
        "jaewon31.kim@...il.com" <jaewon31.kim@...il.com>
Subject: RE: [PATCH v2] page_alloc: fix invalid watemark check on a negative
 value

> 
> 
>--------- Original Message ---------
>Sender : Andrew Morton <akpm@...ux-foundation.org>
>Date : 2022-07-26 03:37 (GMT+9)
>Title : Re: [PATCH v2] page_alloc: fix invalid watemark check on a negative value
> 
>On Mon, 25 Jul 2022 18:52:12 +0900 Jaewon Kim <jaewon31.kim@...sung.com> wrote:
>
>> There was a report that a task is waiting at the
>> throttle_direct_reclaim. The pgscan_direct_throttle in vmstat was
>> increasing.
>> 
>> This is a bug where zone_watermark_fast returns true even when the free
>> is very low. The commit f27ce0e14088 ("page_alloc: consider highatomic
>> reserve in watermark fast") changed the watermark fast to consider
>> highatomic reserve. But it did not handle a negative value case which
>> can be happened when reserved_highatomic pageblock is bigger than the
>> actual free.
>> 
>> If watermark is considered as ok for the negative value, allocating
>> contexts for order-0 will consume all free pages without direct reclaim,
>> and finally free page may become depleted except highatomic free.
>> 
>> Then allocating contexts may fall into throttle_direct_reclaim. This
>> symptom may easily happen in a system where wmark min is low and other
>> reclaimers like kswapd does not make free pages quickly.
>> 
>> Handle the negative case by using MIN.
>> 
>
>Thanks, I added cc:stable to this.
>

Sorry for your inconvenience.

Could you fix a typo by adding r to watermark on title

- page_alloc: fix invalid watemark check on a negative value
+ page_alloc: fix invalid watermark check on a negative value


I've appreciated it in advance.