Received-SPF: pass (google.com: domain of linux-kernel+bounces-189869-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Sun, 26 May 2024 22:52:30 +0200
From: Frederic Weisbecker <frederic@kernel.org>
To: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@redhat.com>,
	Nicholas Piggin <npiggin@gmail.com>,
	Peter Zijlstra <peterz@infradead.org>, Phil Auld <pauld@redhat.com>,
	Chris von Recklinghausen <crecklin@redhat.com>,
	linux-kernel@vger.kernel.org
Subject: Re: sched/isolation: tick_take_do_timer_from_boot() calls
 smp_call_function_single() with irqs disabled
Message-ID: <ZlOhDiqomHlumd-u@pavilion.home>
References: <20240522151742.GA10400@redhat.com>
 <20240523132358.GA1965@redhat.com>
 <87h6eneeu7.ffs@tglx>
 <ZlCwKk65-eL0FrKX@pavilion.home>
 <20240524183700.GA17065@redhat.com>
 <87v832dfw1.ffs@tglx>
 <20240526192758.GA21193@redhat.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20240526192758.GA21193@redhat.com>

Le Sun, May 26, 2024 at 09:27:58PM +0200, Oleg Nesterov a ?crit :
> > So up to this point the tick is never stopped neither on housekeeping
> > nor on NOHZ FULL CPUs:
> >
> > tick_nohz_full_update_tick()
> >   if (!tick_sched_flag_test(ts, TS_FLAG_NOHZ))
> >     return;
> 
> OK... But tick_nohz_idle_update_tick() doesn't check TS_FLAG_NOHZ and
> the tick_nohz_full_cpu() check can't help at boot time.

Yes but tick_nohz_idle_update_tick() is only called when the tick is already
stopped. And for the tick to be already stopped, TS_FLAG_NOHZ must have been
set.

> And I still don't understand why we can rely on can_stop_idle_tick() even
> in tick_nohz_idle_stop_tick().

Not sure I follow you on this one...

> 
> I'll try to read this code again tomorrow, but it will never fit my poor
> little brain ;)

You understood more than I ever did in just a few hours :-)

But yes the tick code was much more simple before I put my own hands
inside. nohz_full and cpuidle did not arrange it.

Thanks.