linux-kernel - Re: commit 3d5bfbd97163 versus -rt

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: <b66ff695-7f75-2a70-7feb-0df4f5ea0608@prevas.dk>
Date:   Tue, 15 Jun 2021 17:57:25 +0200
From:   Rasmus Villemoes <rasmus.villemoes@...vas.dk>
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     Song Hui <hui.song_1@....com>,
        Bartosz Golaszewski <bgolaszewski@...libre.com>,
        linux-rt-users@...r.kernel.org,
        LKML <linux-kernel@...r.kernel.org>,
        Linus Walleij <linus.walleij@...aro.org>,
        Vladimir Oltean <vladimir.oltean@....com>,
        Esben Haabendal <esben@...nix.com>
Subject: Re: commit 3d5bfbd97163 versus -rt

On 15/06/2021 17.33, Steven Rostedt wrote:
> On Tue, 15 Jun 2021 14:35:27 +0200
> Rasmus Villemoes <rasmus.villemoes@...vas.dk> wrote:
> 
>> Reverting commit 3d5bfbd9716318b1ca5c38488aa69f64d38a9aa5 (gpio:
>> mpc8xxx: change the gpio interrupt flags.) makes it go away, as does
>> disabling CONFIG_PREEMPT_RT or simply booting a vanilla v5.10.42 (where
>> that option exists but cannot be selected).
> 
> I'm curious if it will also trigger on vanilla v5.10.42 but add to the
> kernel command line: threadirqs
> 
> Make sure you have CONFIG_IRQ_FORCED_THREADING set too.
> 
> Because it appears to be an issue with that being called by the generic
> threaded irq infrastructure, which PREEMPT_RT enables automatically.

It doesn't:

~ # uname -r
5.10.42-00001-g10216cf63a12
~ # grep -ow threadirqs /proc/cmdline
threadirqs
~ # zcat /proc/config.gz | grep FORCED_THREADING
CONFIG_IRQ_FORCED_THREADING=y
~ # dmesg | grep WARNING
~ #

(the one patch on top of 5.10.42 is a fixup of ls1021a.dtsi that I
should get upstream some day, but not something that should affect this
issue in any way).

Thanks for the suggestion.

Rasmus