linux-kernel - Re: Bug 4.9 and memorymanagement

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

Message-ID: <20161226110053.GA16042@dhcp22.suse.cz>
Date:   Mon, 26 Dec 2016 12:00:53 +0100
From:   Michal Hocko <mhocko@...nel.org>
To:     Klaus Ethgen <Klaus+lkml@...gen.de>
Cc:     linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: Bug 4.9 and memorymanagement

[CCing linux-mm]

On Sun 25-12-16 21:52:52, Klaus Ethgen wrote:
> Hello,
> 
> The last days I compiled version 4.9 for my i386 laptop. (Lenovo x61s)

Do you have memory cgroups enabled in runtime (aka does the same happen
with cgroup_disable=memory)?

> First, everything seems to be sane but after some sleep and awake
> (suspend to ram) cycles I seen some really weird behaviour ending in OOM
> or even complete freeze of the laptop.
> 
> What I was able to see is that it went to swap even if there is plenty
> of memory left. The OOMs was also with many memory left.

Could you paste those OOM reports from the kernel log?

> Once I also catched kswapd0 with running insane with 100% CPU
> utilization.
> 
> I first had in mind the CONFIG_SLAB_FREELIST_RANDOM setting and disabled
> it. This didn't made the problem to go away but it helped a little.
> Nevertheless, further OOM or other strange behaviour happened.
> 
> I went back to 4.8.15 now with the same config from 4.9 and everything
> gets back to normal.
> 
> So it seems for me that there are some really strange memory leaks in
> 4.9. The biggest problem is, that I do not know how to reproduce it
> reliable. The only what I know is that it happened after several
> suspends. (Not necessarily the first.)
> 
> Am I the only one seeing that behavior or do anybody have an idea what
> could went wrong?

no there were some reports recently and 32b with memory cgroups are
broken since 4.8 when the zone LRU's were moved to nodes.
>x 
> For the reference I put the .configs of the two compilings as attachment
> to this mail.
> 
> Please keep me in CC as I am not subscribed to LKML.
> 
> Regards
>    Klaus






-- 
Michal Hocko
SUSE Labs