linux-kernel - Re: [RFC][PATCH 2/2] time: alarmtimer: Use TASK

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <874jrfq3jw.ffs@tglx>
Date:   Tue, 21 Feb 2023 01:12:51 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Michael Nazzareno Trimarchi <michael@...rulasolutions.com>
Cc:     John Stultz <jstultz@...gle.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Stephen Boyd <sboyd@...nel.org>, Arnd Bergmann <arnd@...db.de>,
        Michael <michael@...isi.de>, kernel-team@...roid.com,
        Peter Zijlstra <peterz@...radead.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>
Subject: Re: [RFC][PATCH 2/2] time: alarmtimer: Use TASK_FREEZABLE to
 cleanup freezer handling

Michael!

On Mon, Feb 20 2023 at 22:32, Michael Nazzareno Trimarchi wrote:
> On Mon, Feb 20, 2023 at 10:18 PM Thomas Gleixner <tglx@...utronix.de> wrote:
>>   * alarmtimer_fired - Handles alarm hrtimer being fired.
>> @@ -194,6 +196,8 @@ static enum hrtimer_restart alarmtimer_f
>>         int ret = HRTIMER_NORESTART;
>>         int restart = ALARMTIMER_NORESTART;
>>
>> +       atomic_inc(&alarmtimer_wakeup);
>> +
>
>        ptr->it_active = 0;
>         if (ptr->it_interval) {
>                 atomic_inc(&alarmtimer_wakeup);
>                 si_private = ++ptr->it_requeue_pending;
>         }
>
> Should I not go to the alarm_handle_timer? and only if it's a periodic
> one?

Why?

Any alarmtimer which hits that window has exactly the same problem.

It's not restricted to periodic timers. Why would a dropped one-shot
wakeup be acceptable?

It's neither restricted to posix timers. If a clock_nanosleep(ALARM)
expires in that window then the task wake up will just end up in the
/dev/null bucket for the very same reason. Why would this be correct?

Hmm?

<GRMBL>
> Michael
>
>>         spin_lock_irqsave(&base->lock, flags);

<SNIP>Tons of wasted electrons</SNIP>

Can you please trim your replies?

</GRMBL>

Thanks,

        tglx