linux-kernel - Re: mount cgroup with "already mounted or cgroup busy"

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:   Fri, 22 Nov 2019 16:19:48 +0100
From:   Michal Koutný <mkoutny@...e.com>
To:     "Li,Rongqing" <lirongqing@...du.com>
Cc:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "tj@...nel.org" <tj@...nel.org>,
        "lizefan@...wei.com" <lizefan@...wei.com>,
        "hannes@...xchg.org" <hannes@...xchg.org>,
        "cgroups@...r.kernel.org" <cgroups@...r.kernel.org>
Subject: Re: mount cgroup with "already mounted or cgroup busy"

Hello Li.

On Wed, Oct 30, 2019 at 09:32:41AM +0000, "Li,Rongqing" <lirongqing@...du.com> wrote:
> I meet a issue, and not sure if it is normal
It is a consequence of how cgroup state is retained.

I considered a "trivial" fix, however, I learnt later it'd be a revert
(and hence regression) of 
	3c606d35fe97 "cgroup: prevent mount hang due to memory controller lifetime"

> Using the below script, the last line mount will fail after unmount
> And it will fail until reboot system
The cause is that controllers remained pinned to the same root.
You can unpin them by mounting/unmounting again attached to the same
root (i.e. cpu,cpuset again).

(Perhaps the logic in cgroup1_root_to_use may be changed to cope better
with the "stateless" controllers.)

Michal

Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)