linux-kernel - Re: [patch V2 9/9] tasklets: Prevent kill/unlock

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:   Mon, 07 Dec 2020 18:50:56 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Frederic Weisbecker <frederic@...nel.org>,
        Paul McKenney <paulmck@...nel.org>
Subject: Re: [patch V2 9/9] tasklets: Prevent kill/unlock_wait deadlock on RT

On Mon, Dec 07 2020 at 18:49, Thomas Gleixner wrote:

> On Mon, Dec 07 2020 at 16:39, Sebastian Andrzej Siewior wrote:
>> On 2020-12-07 16:22:07 [+0100], Thomas Gleixner wrote:
>>> On Mon, Dec 07 2020 at 15:00, Sebastian Andrzej Siewior wrote:
>>> > So we keep the RT part as-is and replace the non-RT bits with this?
>>> 
>>> No. It would work for both.
>>
>> So instead of boosting our way through we simply wait until the tasklet
>> completes. Given that canceling is usually done on start/stop events, it
>> shouldn't matter if the RT priority is lost.
>
> That was my reasoning. The only thing we need to figure out whether
> there are callers on !RT which invoke that muck from non-sleepable
> context.

The kill() variant is fine. It must be called from sleepable
context. The otherone can be invoked everywhere except hard interrupt
context IIRC.

Thanks,

        tglx