linux-kernel - Re: [RFC 1/3] tick-sched: Remove last_tick and calculate next tick from now

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: <ZzNeeCvSFL7OzHKF@localhost.localdomain>
Date: Tue, 12 Nov 2024 14:56:08 +0100
From: Frederic Weisbecker <frederic@...nel.org>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: "Joel Fernandes (Google)" <joel@...lfernandes.org>,
	linux-kernel@...r.kernel.org,
	Anna-Maria Behnsen <anna-maria@...utronix.de>,
	Ingo Molnar <mingo@...nel.org>
Subject: Re: [RFC 1/3] tick-sched: Remove last_tick and calculate next tick
 from now

Le Tue, Nov 12, 2024 at 02:46:23PM +0100, Thomas Gleixner a écrit :
> On Tue, Nov 12 2024 at 00:43, Frederic Weisbecker wrote:
> > Le Fri, Nov 08, 2024 at 05:48:34PM +0000, Joel Fernandes (Google) a écrit :
> 
> >> During tick restart, we use last_tick and forward it past now.
> >>
> >> Since we are forwarding past now, we can simply use now as a reference
> >> instead of last_tick. This patch removes last_tick and does so.
> >>
> >> This patch potentially does more mul/imul than the existing code,
> >> as sometimes forwarding past now need not be done if last_tick > now.
> >> However, the patch is a cleanup which reduces LOC and reduces the size
> >> of struct tick_sched.
> 
> May I politely ask you to read and follow the Documentation
> vs. changelogs?
> 
>   https://www.kernel.org/doc/html/latest/process/maintainer-tip.html#changelog
> 
> Also
> 
> git grep 'This patch' Documentation/process
> 
> might give you a hint.
> 
> >> -	/* Forward the time to expire in the future */
> >> -	hrtimer_forward(&ts->sched_timer, now, TICK_NSEC);
> >> +	hrtimer_set_expires(&ts->sched_timer, DIV_ROUND_UP_ULL(now, TICK_NSEC) * TICK_NSEC);
> 
> How is a division and multiplication in this hotpath helpful? That's
> awfully slow on 32-bit machines and pointless on 64-bit too.
> 
> Using now is also wrong as it breaks the sched_skew_tick distribution by
> aligning the tick on all CPUs again.
> 
> IOW, this "cleanup" is making things worse.
> 
> > We don't want to rewrite hrtimer_forward() but, after all, the current expiry is
> > enough a relevant information.
> >
> > How about just this? It's worth it as it now forwards after the real last programmed
> > tick, which should be close enough from @now with a delta below TICK_NSEC, or even
> > better @now is below the expiry. Therefore it should resume as just a no-op
> > or at worst an addition within hrtimer_forward():
> >
> > diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
> > index 753a184c7090..ffd0c026a248 100644
> > --- a/kernel/time/tick-sched.c
> > +++ b/kernel/time/tick-sched.c
> > @@ -838,7 +838,6 @@ EXPORT_SYMBOL_GPL(get_cpu_iowait_time_us);
> >  static void tick_nohz_restart(struct tick_sched *ts, ktime_t now)
> >  {
> >  	hrtimer_cancel(&ts->sched_timer);
> > -	hrtimer_set_expires(&ts->sched_timer, ts->last_tick);
> >  
> >  	/* Forward the time to expire in the future */
> >  	hrtimer_forward(&ts->sched_timer, now, TICK_NSEC);
> 
> That's just wrong. ts->sched_timer.expires contains a tick in the
> future. If tick_nohz_stop_tick() set it to 10 ticks in the future and
> the CPU goes out of idle due to a device interrupt before the timer
> expires, then hrtimer_forward() will do nothing because expires is ahead
> of now.
> 
> Which means the CPU is not idle and has no tick until the delayed tick
> which was set by tick_nohz_stop_tick() expires. Not really correct.

Bah! Yes of course...

> 
> Thanks,
> 
>         tglx