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
| ||
|
Message-ID: <20200617141155.GQ9499@dhcp22.suse.cz> Date: Wed, 17 Jun 2020 16:11:55 +0200 From: Michal Hocko <mhocko@...nel.org> To: Chris Down <chris@...isdown.name> Cc: Naresh Kamboju <naresh.kamboju@...aro.org>, Yafang Shao <laoar.shao@...il.com>, Anders Roxell <anders.roxell@...aro.org>, "Linux F2FS DEV, Mailing List" <linux-f2fs-devel@...ts.sourceforge.net>, linux-ext4 <linux-ext4@...r.kernel.org>, linux-block <linux-block@...r.kernel.org>, Andrew Morton <akpm@...ux-foundation.org>, open list <linux-kernel@...r.kernel.org>, Linux-Next Mailing List <linux-next@...r.kernel.org>, linux-mm <linux-mm@...ck.org>, Arnd Bergmann <arnd@...db.de>, Andreas Dilger <adilger.kernel@...ger.ca>, Jaegeuk Kim <jaegeuk@...nel.org>, Theodore Ts'o <tytso@....edu>, Chao Yu <chao@...nel.org>, Hugh Dickins <hughd@...gle.com>, Andrea Arcangeli <aarcange@...hat.com>, Matthew Wilcox <willy@...radead.org>, Chao Yu <yuchao0@...wei.com>, lkft-triage@...ts.linaro.org, Johannes Weiner <hannes@...xchg.org>, Roman Gushchin <guro@...com>, Cgroups <cgroups@...r.kernel.org> Subject: Re: mm: mkfs.ext4 invoked oom-killer on i386 - pagecache_get_page [Our emails have crossed] On Wed 17-06-20 14:57:58, Chris Down wrote: > Naresh Kamboju writes: > > mkfs -t ext4 /dev/disk/by-id/ata-TOSHIBA_MG04ACA100N_Y8RQK14KF6XF > > mke2fs 1.43.8 (1-Jan-2018) > > Creating filesystem with 244190646 4k blocks and 61054976 inodes > > Filesystem UUID: 7c380766-0ed8-41ba-a0de-3c08e78f1891 > > Superblock backups stored on blocks: > > 32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, > > 4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968, > > 102400000, 214990848 > > Allocating group tables: 0/7453 done > > Writing inode tables: 0/7453 done > > Creating journal (262144 blocks): [ 51.544525] under min:0 emin:0 > > [ 51.845304] under min:0 emin:0 > > [ 51.848738] under min:0 emin:0 > > [ 51.858147] under min:0 emin:0 > > [ 51.861333] under min:0 emin:0 > > [ 51.862034] under min:0 emin:0 > > [ 51.862442] under min:0 emin:0 > > [ 51.862763] under min:0 emin:0 > > Thanks, this helps a lot. Somehow we're entering mem_cgroup_below_min even > when min/emin is 0 (which should indeed be the case if you haven't set them > in the hierarchy). > > My guess is that page_counter_read(&memcg->memory) is 0, which means > mem_cgroup_below_min will return 1. Yes this is the case because this is likely the root memcg which skips all charges. > However, I don't know for sure why that should then result in the OOM killer > coming along. My guess is that since this memcg has 0 pages to scan anyway, > we enter premature OOM under some conditions. I don't know why we wouldn't > have hit that with the old version of mem_cgroup_protected that returned > MEMCG_PROT_* members, though. Not really. There is likely no other memcg to reclaim from and assuming min limit protection will result in no reclaimable memory and thus the OOM killer. > Can you please try the patch with the `>=` checks in mem_cgroup_below_min > and mem_cgroup_below_low changed to `>`? If that fixes it, then that gives a > strong hint about what's going on here. This would work but I believe an explicit check for the root memcg would be easier to spot the reasoning. -- Michal Hocko SUSE Labs
Powered by blists - more mailing lists