linux-kernel - Re: lockdep warns: cpu_hotplug_lock.rw_sem --> slab

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: <alpine.DEB.2.21.1912101829170.14347@www.lameter.com>
Date:   Tue, 10 Dec 2019 18:32:37 +0000 (UTC)
From:   Christopher Lameter <cl@...ux.com>
To:     Qian Cai <cai@....pw>
cc:     Daniel Wagner <dwagner@...e.de>, Pekka Enberg <penberg@...nel.org>,
        David Rientjes <rientjes@...gle.com>,
        Joonsoo Kim <iamjoonsoo.kim@....com>,
        Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: lockdep warns: cpu_hotplug_lock.rw_sem --> slab_mutex -->
 kn->count#39

On Mon, 9 Dec 2019, Qian Cai wrote:

> It happens to me too with a probably easier to trigger deadlock. Basically, we have,
>
> memcg_create_kmem_cache():
> cpu_hotplug_lock.rw_sem/mem_hotplug_lock.rw_sem —> kn->count

Right. Memcg runs this function at cgroup creation time. sysfs
initialization was designed to run a at boot. So the sysfs initialization
needs a rewrite of its serialization for the memcg case. We cannot loop
over the list of all kmem caches holding the slab_mutex anymore.