linux-kernel - Re: [BUG] cgroup lockup while upgrading udev on debian

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: <20180122164636.GA24468@kroah.com>
Date:   Mon, 22 Jan 2018 17:46:36 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     Tejun Heo <tj@...nel.org>
Cc:     Steven Rostedt <rostedt@...dmis.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Johannes Weiner <hannes@...xchg.org>,
        Li Zefan <lizefan@...wei.com>, cgroups@...r.kernel.org
Subject: Re: [BUG] cgroup lockup while upgrading udev on debian

On Mon, Jan 22, 2018 at 08:30:51AM -0800, Tejun Heo wrote:
> Hello, Steven.
> 
> On Sun, Jan 21, 2018 at 09:31:08PM -0500, Steven Rostedt wrote:
> > While upgrading my system, it locked up while upgrading udev. It locked
> > up on vanilla v4.14.12, and it also locked up on debian's 4.14 kernel
> > as well. When I booted debian's 4.9 kernel, it upgraded fine.
> > 
> > Note, this is my main server and not a development box. It is not
> > something I'm going to run bisects or other tests on. I don't even have
> > it logging to serial console (hence the photo). It's serials are hooked
> > to my test boxes. That is, it receives logs, it's not set up to send
> > them.
> > 
> > Attached in the photo of the crash that happened when I did:
> > 
> >  # dpkg --configure udev
> 
> It looks like the same bug which was fixed by 74d0833c659a ("cgroup:
> fix css_task_iter crash on CSS_TASK_ITER_PROC").  The bug got
> introduced durinv v4.14 cycle and triggers when cgroup1 and cgroup2
> are used at the same time, which the recent versions of systemd does
> when in cgroup1 mode.

That fix ended up in 4.14.14, so 4.14.12 is a bit too old Steven :)

greg k-h