Re: [PATCH] nohz: Fix spurious warning when hrtimer and clockevent get out of sync

[Frederic Weisbecker <fweisbec@gmail.com>]

On Thu, Jun 08, 2017 at 10:13:38PM +0000, Levin, Alexander (Sasha Levin) wrote:
> On Thu, Jun 08, 2017 at 09:07:05PM +0200, Frederic Weisbecker wrote:
> > Awesome, these traces have been very helpful! So now I think I get what's going on.
> > Can you please test the following fix?
> 
> With the patch, I hit the warning early on boot:
> 
> [    1.423727] clocksource: Switched to clocksource kvm-clock
> [    1.429326] ------------[ cut here ]------------
> [    1.430234] WARNING: CPU: 1 PID: 0 at kernel/time/tick-sched.c:792 __tick_nohz_idle_enter+0xe1c/0x15c0

Oh sorry, I inverted some conditional. It warns as soon as the tick is completely stopped
and not just deferred, which seldom happen on my testbox.

I need to learn programming again.

Here is the fixed version:

---
>From f80041b5209aaf9d02ac25a29a248d0f214ba19f Mon Sep 17 00:00:00 2001
From: Frederic Weisbecker <fweisbec@gmail.com>
Date: Thu, 8 Jun 2017 16:32:58 +0200
Subject: [PATCH] nohz: Fix spurious warning when hrtimer and clocksource get
 out of sync

The sanity check ensuring that the tick expiry cache (ts->next_tick)
is actually in sync with the hardware clock (dev->next_event) makes the
wrong assumption that the clock can't be programmed later than the
hrtimer deadline.

In fact the clock hardware can be programmed later on some conditions
such as:

    * The hrtimer deadline is already in the past.
    * The hrtimer deadline is earlier than the minimum delay supported
      by the hardware.

Such conditions can be met when we program the tick, for example if the
last jiffies update hasn't been seen by the current CPU yet, we may
program the hrtimer to a deadline that is earlier than ktime_get()
because last_jiffies_update is our timestamp base to compute the next
tick.

As a result, we can randomly observe such warning:

	WARNING: CPU: 5 PID: 0 at kernel/time/tick-sched.c:794 tick_nohz_stop_sched_tick kernel/time/tick-sched.c:791 [inline]
	Call Trace:
	 tick_nohz_irq_exit
	 tick_irq_exit
	 irq_exit
	 exiting_irq
	 smp_call_function_interrupt
	 smp_call_function_single_interrupt
	 call_function_single_interrupt

Therefore, let's rather make sure that the tick expiry cache is sync'ed
with the tick hrtimer deadline, against which it is not supposed to
drift away. The clock hardware instead has its own will and can't be
used as a reliable comparison point.

Reported-by: Sasha Levin <alexander.levin@verizon.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: James Hartsock <hartsjc@redhat.com>
Cc: Tim Wright <tim@binbash.co.uk>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 kernel/time/tick-sched.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 9d31f1e..83c788e 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -768,7 +768,8 @@ static ktime_t tick_nohz_stop_sched_tick(struct tick_sched *ts,
 	/* Skip reprogram of event if its not changed */
 	if (ts->tick_stopped && (expires == ts->next_tick)) {
 		/* Sanity check: make sure clockevent is actually programmed */
-		if (likely(dev->next_event <= ts->next_tick))
+		if (tick == KTIME_MAX ||
+		    ts->next_tick == hrtimer_get_expires(&ts->sched_timer))
 			goto out;
 
 		WARN_ON_ONCE(1);
@@ -806,8 +807,10 @@ static ktime_t tick_nohz_stop_sched_tick(struct tick_sched *ts,
 		goto out;
 	}
 
+	hrtimer_set_expires(&ts->sched_timer, tick);
+
 	if (ts->nohz_mode == NOHZ_MODE_HIGHRES)
-		hrtimer_start(&ts->sched_timer, tick, HRTIMER_MODE_ABS_PINNED);
+		hrtimer_start_expires(&ts->sched_timer, HRTIMER_MODE_ABS_PINNED);
 	else
 		tick_program_event(tick, 1);
 out:
-- 
2.7.4