linux-kernel - Re: [BUG] workqueues and printk not playing nice since next-20240130

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: <a6762ad5-62f2-41d4-bba7-c2d9dcb12498@paulmck-laptop>
Date: Mon, 5 Feb 2024 09:45:53 -0800
From: "Paul E. McKenney" <paulmck@...nel.org>
To: Sergey Senozhatsky <senozhatsky@...omium.org>
Cc: Petr Mladek <pmladek@...e.com>, Tejun Heo <tj@...nel.org>,
	Jonas Oberhauser <jonas.oberhauser@...weicloud.com>,
	Lai Jiangshan <jiangshanlai@...il.com>,
	Steven Rostedt <rostedt@...dmis.org>,
	John Ogness <john.ogness@...utronix.de>,
	Stephen Rothwell <sfr@...b.auug.org.au>,
	linux-kernel@...r.kernel.org, rcu@...r.kernel.org
Subject: Re: [BUG] workqueues and printk not playing nice since next-20240130

On Mon, Feb 05, 2024 at 10:25:15PM +0900, Sergey Senozhatsky wrote:
> On (24/02/05 14:07), Petr Mladek wrote:
> > > Good point, if it does recur, I could try it on bare metal.
> > 
> > Please, me, John, and Sergey know if anyone see this again. I do not
> > feel comfortable when there is problem which might make consoles calm.
> 
> Agreed.
> 
> > Bisection identified this commit:
> > 5797b1c18919 ("workqueue: Implement system-wide nr_active enforcement for unbound workqueues")
> 
> That commit triggered early boot use-after-free (per kasan) on
> my system, which probably could derail some things.

And enabling KASAN on next-20240130 got me that same KASAN report and
also suppressed the misbehavior, which is not surprising given that
KASAN quarantines free memory for some time.  Plus enabling KASAN
on recent -next does not trigger that KASAN report.

So my guess is that we can attribute my oddball test failures to
that use after free.  But I will of course continue testing.

							Thanx, Paul