linux-kernel - Re: Revert "leds: avoid races with workqueue"?

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.LSU.2.11.1905261318290.2394@eggly.anvils>
Date:   Sun, 26 May 2019 13:36:11 -0700 (PDT)
From:   Hugh Dickins <hughd@...gle.com>
To:     Pavel Machek <pavel@....cz>
cc:     Hugh Dickins <hughd@...gle.com>,
        Jacek Anaszewski <jacek.anaszewski@...il.com>,
        linux-leds@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: Revert "leds: avoid races with workqueue"?

On Sun, 26 May 2019, Pavel Machek wrote:
> On Sat 2019-05-25 10:32:31, Hugh Dickins wrote:
> > 
> > Thanks, Pavel: yes, that works fine for me on the T420s, no debug
> > complaints, good and silent; and the wifi LED is blinking as before.
> 
> I'd like to prevent recurrence of similar problem, and I wonder if you
> can give me a hint.
> 
> I can annotate code that can sleep with might_sleep().
> 
> How can I annotate code that can not sleep? I might do 
> 
> spin_lock(&dummy);
> this_should_not_sleep();
> spin_unlock(&dummy);
> 
> But I don't really need extra serialization. I just want annotations for
> lockdep. Any ideas?

I haven't tried to do that directly, so I'm likely to give bad advice:
in particular, I forget the limitations of checking in_atomic().

But very useful (and much cheaper than lockdep) debug options in
this area are CONFIG_DEBUG_PREEMPT and CONFIG_DEBUG_ATOMIC_SLEEP.

Hugh