linux-kernel - Re: [PATCH] checkpatch: warn about flushing system-wide workqueues

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]

Date:   Thu, 12 May 2022 06:46:23 -1000
From:   Tejun Heo <tj@...nel.org>
To:     Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Joe Perches <joe@...ches.com>,
        Andy Whitcroft <apw@...onical.com>,
        Dwaipayan Ray <dwaipayanray1@...il.com>,
        Lukas Bulwahn <lukas.bulwahn@...il.com>
Subject: Re: [PATCH] checkpatch: warn about flushing system-wide workqueues

Hello,

On Fri, May 06, 2022 at 08:29:07AM +0900, Tetsuo Handa wrote:
> On 2022/05/06 2:32, Tejun Heo wrote:
> > Given that we'll need runtime check anyway, why not resurrect the original
> > runtime warning but exempt flush_schedule_work() if that's the only thing
> > remaining right now (using a special flag or whatever)?
> 
> Yes, we will also need runtime check for robustness, for we can't catch usage
> like
> 
> 	struct workqueue_struct *my_wq = alloc_workqueue();
> 	if (!my_wq)
> 		my_wq = system_long_wq;
> 	flush_workqueue(my_wq);
> 
> using compile time checks.
> 
> I found that it is not easy to trigger flush_workqueue() paths. For example,
> several modules are using flush_workqueue() only upon module unloading.

Ah, yeah, good point.

> Therefore, I'm trying to catch obvious flush_workqueue() paths at compile
> time when possible.
> 
> >                                                         If we're sure that
> > we aren't triggering it spuriously, we can ask Andrew to take the warning
> > patch into -mm so that it floats on top of everything else and gets pulled
> > into the trunk during the coming merge window.
> 
> OK, the coming merge window means 5.19.
> 
> The original runtime checking will be used anyway. Is "workqueue: Wrap
> flush_workqueue() using a macro" OK for you as a compile time check?

Sounds good to me.

Thanks.

-- 
tejun