linux-kernel - Re: [syzbot] INFO: rcu detected stall in syscall_exit_to_user

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:   Wed, 15 Sep 2021 11:32:02 +0200
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Dmitry Vyukov <dvyukov@...gle.com>
Cc:     Hillf Danton <hdanton@...a.com>,
        syzbot <syzbot+0e964fad69a9c462bc1e@...kaller.appspotmail.com>,
        linux-kernel@...r.kernel.org, paulmck@...nel.org,
        syzkaller-bugs@...glegroups.com,
        Peter Zijlstra <peterz@...radead.org>,
        kasan-dev <kasan-dev@...glegroups.com>,
        Johannes Berg <johannes.berg@...el.com>,
        Kalle Valo <kvalo@...eaurora.org>,
        linux-wireless@...r.kernel.org
Subject: Re: [syzbot] INFO: rcu detected stall in syscall_exit_to_user_mode

On Wed, Sep 15 2021 at 11:14, Dmitry Vyukov wrote:
> On Wed, 15 Sept 2021 at 10:57, Thomas Gleixner <tglx@...utronix.de> wrote:
>> That made me actually look at that mac80211_hwsim callback again.
>>
>>         hrtimer_forward(&data->beacon_timer, hrtimer_get_expires(timer),
>>                         ns_to_ktime(bcn_int * NSEC_PER_USEC));
>>
>> So what this does is really wrong because it tries to schedule the timer
>> on the theoretical periodic timeline. Which goes really south once the
>> timer is late or the callback execution took longer than the
>> period. Hypervisors scheduling out a VCPU at the wrong place will do
>> that for you nicely.
>
> Nice!
>
> You mentioned that hrtimer_run_queues() may not return. Does it mean
> that it can just loop executing the same re-armed callback again and
> again? Maybe then the debug check condition should be that
> hrtimer_run_queues() runs the same callback more than N times w/o
> returning?

Something like that.