[PATCH] timers/nohz: Low-res tick handler switch to ONESHOT_STOPPED if tick stops

Nicholas Piggin posted 1 patch 4 years ago
kernel/time/tick-sched.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
[PATCH] timers/nohz: Low-res tick handler switch to ONESHOT_STOPPED if tick stops
Posted by Nicholas Piggin 4 years ago
When tick_nohz_stop_tick() stops the tick, the the clock event device
is not put into ONESHOT_STOPPED mode. This can lead to spurious timer
interrupts with some clock event device drivers that don't shut down
entirely after firing.

Eliminate these by putting the device into ONESHOT_STOPPED mode at
points where it is not being reprogrammed. When there are no timers
active, then tick_program_event() with KTIME_MAX can be used to stop the
device. When there is a timer active, the device can be stopped at the
next tick (any new timer added by timers will reprogram the tick).

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 kernel/time/tick-sched.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index d257721c68b8..da1a7efa45a4 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -928,6 +928,8 @@ static void tick_nohz_stop_tick(struct tick_sched *ts, int cpu)
 	if (unlikely(expires == KTIME_MAX)) {
 		if (ts->nohz_mode == NOHZ_MODE_HIGHRES)
 			hrtimer_cancel(&ts->sched_timer);
+		else
+			tick_program_event(KTIME_MAX, 1);
 		return;
 	}
 
@@ -1364,9 +1366,14 @@ static void tick_nohz_handler(struct clock_event_device *dev)
 	tick_sched_do_timer(ts, now);
 	tick_sched_handle(ts, regs);
 
-	/* No need to reprogram if we are running tickless  */
-	if (unlikely(ts->tick_stopped))
+	if (unlikely(ts->tick_stopped)) {
+		/*
+		 * If we are tickless, no need to reprogram, so change the
+		 * clock event device to ONESHOT_STOPPED.
+		 */
+		tick_program_event(KTIME_MAX, 1);
 		return;
+	}
 
 	hrtimer_forward(&ts->sched_timer, now, TICK_NSEC);
 	tick_program_event(hrtimer_get_expires(&ts->sched_timer), 1);
-- 
2.35.1
[tip: timers/core] timers/nohz: Switch to ONESHOT_STOPPED in the low-res handler when the tick is stopped
Posted by tip-bot2 for Nicholas Piggin 4 years ago
The following commit has been merged into the timers/core branch of tip:

Commit-ID:     62c1256d544747b38e77ca9b5bfe3a26f9592576
Gitweb:        https://git.kernel.org/tip/62c1256d544747b38e77ca9b5bfe3a26f9592576
Author:        Nicholas Piggin <npiggin@gmail.com>
AuthorDate:    Sat, 23 Apr 2022 00:14:46 +10:00
Committer:     Thomas Gleixner <tglx@linutronix.de>
CommitterDate: Mon, 25 Apr 2022 14:45:22 +02:00

timers/nohz: Switch to ONESHOT_STOPPED in the low-res handler when the tick is stopped

When tick_nohz_stop_tick() stops the tick and high resolution timers are
disabled, then the clock event device is not put into ONESHOT_STOPPED
mode. This can lead to spurious timer interrupts with some clock event
device drivers that don't shut down entirely after firing.

Eliminate these by putting the device into ONESHOT_STOPPED mode at points
where it is not being reprogrammed. When there are no timers active, then
tick_program_event() with KTIME_MAX can be used to stop the device. When
there is a timer active, the device can be stopped at the next tick (any
new timer added by timers will reprogram the tick).

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20220422141446.915024-1-npiggin@gmail.com
---
 kernel/time/tick-sched.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 2d76c91..b1b105d 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -928,6 +928,8 @@ static void tick_nohz_stop_tick(struct tick_sched *ts, int cpu)
 	if (unlikely(expires == KTIME_MAX)) {
 		if (ts->nohz_mode == NOHZ_MODE_HIGHRES)
 			hrtimer_cancel(&ts->sched_timer);
+		else
+			tick_program_event(KTIME_MAX, 1);
 		return;
 	}
 
@@ -1364,9 +1366,15 @@ static void tick_nohz_handler(struct clock_event_device *dev)
 	tick_sched_do_timer(ts, now);
 	tick_sched_handle(ts, regs);
 
-	/* No need to reprogram if we are running tickless  */
-	if (unlikely(ts->tick_stopped))
+	if (unlikely(ts->tick_stopped)) {
+		/*
+		 * The clockevent device is not reprogrammed, so change the
+		 * clock event device to ONESHOT_STOPPED to avoid spurious
+		 * interrupts on devices which might not be truly one shot.
+		 */
+		tick_program_event(KTIME_MAX, 1);
 		return;
+	}
 
 	hrtimer_forward(&ts->sched_timer, now, TICK_NSEC);
 	tick_program_event(hrtimer_get_expires(&ts->sched_timer), 1);
Re: [tip: timers/core] timers/nohz: Switch to ONESHOT_STOPPED in the low-res handler when the tick is stopped
Posted by Frederic Weisbecker 3 years ago
On Mon, Apr 25, 2022 at 12:50:56PM -0000, tip-bot2 for Nicholas Piggin wrote:
> The following commit has been merged into the timers/core branch of tip:
> 
> Commit-ID:     62c1256d544747b38e77ca9b5bfe3a26f9592576
> Gitweb:        https://git.kernel.org/tip/62c1256d544747b38e77ca9b5bfe3a26f9592576
> Author:        Nicholas Piggin <npiggin@gmail.com>
> AuthorDate:    Sat, 23 Apr 2022 00:14:46 +10:00
> Committer:     Thomas Gleixner <tglx@linutronix.de>
> CommitterDate: Mon, 25 Apr 2022 14:45:22 +02:00
> 
> timers/nohz: Switch to ONESHOT_STOPPED in the low-res handler when the tick is stopped
> 
> When tick_nohz_stop_tick() stops the tick and high resolution timers are
> disabled, then the clock event device is not put into ONESHOT_STOPPED
> mode. This can lead to spurious timer interrupts with some clock event
> device drivers that don't shut down entirely after firing.
> 
> Eliminate these by putting the device into ONESHOT_STOPPED mode at points
> where it is not being reprogrammed. When there are no timers active, then
> tick_program_event() with KTIME_MAX can be used to stop the device. When
> there is a timer active, the device can be stopped at the next tick (any
> new timer added by timers will reprogram the tick).

I'm confused by the above, why are we handling the timer active part here?

> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Link: https://lore.kernel.org/r/20220422141446.915024-1-npiggin@gmail.com
> ---
>  kernel/time/tick-sched.c | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
> index 2d76c91..b1b105d 100644
> --- a/kernel/time/tick-sched.c
> +++ b/kernel/time/tick-sched.c
> @@ -928,6 +928,8 @@ static void tick_nohz_stop_tick(struct tick_sched *ts, int cpu)
>  	if (unlikely(expires == KTIME_MAX)) {
>  		if (ts->nohz_mode == NOHZ_MODE_HIGHRES)
>  			hrtimer_cancel(&ts->sched_timer);
> +		else
> +			tick_program_event(KTIME_MAX, 1);
>  		return;
>  	}
>  
> @@ -1364,9 +1366,15 @@ static void tick_nohz_handler(struct clock_event_device *dev)
>  	tick_sched_do_timer(ts, now);
>  	tick_sched_handle(ts, regs);
>  
> -	/* No need to reprogram if we are running tickless  */
> -	if (unlikely(ts->tick_stopped))
> +	if (unlikely(ts->tick_stopped)) {
> +		/*
> +		 * The clockevent device is not reprogrammed, so change the
> +		 * clock event device to ONESHOT_STOPPED to avoid spurious
> +		 * interrupts on devices which might not be truly one shot.
> +		 */
> +		tick_program_event(KTIME_MAX, 1);

More specifically why are we stopping the tick here entirely and
unconditionally? If the tick is stopped (actually meaning it is delayed
or _might_ be totally stopped), then the next tick is going to be re-evaluated
shortly after:

* On the idle loop if within idle
* On IRQ exit if nohz_full

And then tick_nohz_stop_tick() will be called and stop the tick entirely
if necessary.

Am I missing something else?

Thanks.