Re: [RFC PATCH] tick/sched: update full_nohz status after SCHED dep is cleared

From: Juri Lelli
Date: Wed May 20 2020 - 12:49:41 EST

On 20/05/20 18:24, Frederic Weisbecker wrote:
> Hi Juri,
>
> On Wed, May 20, 2020 at 04:04:02PM +0200, Juri Lelli wrote:
> > After tasks enter or leave a runqueue (wakeup/block) SCHED full_nohz
> > dependency is checked (via sched_update_tick_dependency()). In case tick
> > can be stopped on a CPU (see sched_can_stop_tick() for details), SCHED
> > dependency for such CPU is cleared. However, this new information is not
> > used right away to actually stop the tick.
> >
> > In CONFIG_PREEMPT systems booted with threadirqs option, sched clock
> > tick is serviced by an actual task (ksoftirqd corresponding to the CPU
> > where tick timer fired).
>
> I must confess I haven't tested threaded IRQs but I was
> pretty sure that the timer tick is always serviced on hardirq.
>
> Now the timer list callbacks are executed on softirqs. So if the
> tick itself, executed on hardirq, sees pending timers, it raise
> the softirq which wakes up ksoftirqd on forced irq thread mode
> while calling irq_exit(). Then tick_nohz_irq_exit() sees ksoftirqd
> and the current task on the runqueue, which together can indeed prevent
> from turning off the tick. But then the root cause is pending timer
> list callbacks.
>
> > So, in case a CPU was running a single task,
> > servicing the timer involves exiting full nozh mode. Problem at this
> > point is that we might lose chances to enter back into full nozh mode,
> > since info about ksoftirqd thread going back to sleep is not used (as
> > mentioned above).
>
> It should enter into nohz_full mode in the next tick, which is usually
> a reasonable delay. If you need guarantee that the tick is stopped before
> resuming userspace, you need a stronger machinery such as the task isolation
> patchset.
>
> Let's have a look at the trace below:
>
> > ksoftirqd/19-125 [019] 170.700754: softirq_entry: vec=1 [action=TIMER]
> > ksoftirqd/19-125 [019] 170.700755: softirq_exit: vec=1 [action=TIMER]
> > ksoftirqd/19-125 [019] 170.700756: softirq_entry: vec=7 [action=SCHED]
> > ksoftirqd/19-125 [019] 170.700757: softirq_exit: vec=7 [action=SCHED]
> > ksoftirqd/19-125 [019] 170.700759: sched_switch: ksoftirqd/19:125 [120] S ==> sysjitter:2459 [120]
> > sysjitter-2459 [019] 170.701740: local_timer_entry: vector=236
> > sysjitter-2459 [019] 170.701742: softirq_raise: vec=1 [action=TIMER]
>
> See here the tick sees pending timer callbacks so it raises the timer softirq.
>
> > sysjitter-2459 [019] 170.701743: softirq_raise: vec=7 [action=SCHED]
>
> Oh and the scheduler tick activates the scheduler softirq as well.
>
> > sysjitter-2459 [019] 170.701744: local_timer_exit: vector=236
> > sysjitter-2459 [019] 170.701747: sched_wakeup: ksoftirqd/19:125 [120] success=1 CPU:019
> > sysjitter-2459 [019] 170.701748: tick_stop: success=0 dependency=SCHED
> > sysjitter-2459 [019] 170.701749: irq_work_entry: vector=246
> > sysjitter-2459 [019] 170.701750: irq_work_exit: vector=246
> > sysjitter-2459 [019] 170.701751: tick_stop: success=0 dependency=SCHED
> > sysjitter-2459 [019] 170.701753: sched_switch: sysjitter:2459 [120] R ==> ksoftirqd/19:125 [120]
> > ksoftirqd/19-125 [019] 170.701754: softirq_entry: vec=1 [action=TIMER]
> > ksoftirqd/19-125 [019] 170.701756: softirq_exit: vec=1 [action=TIMER]
> > ksoftirqd/19-125 [019] 170.701756: softirq_entry: vec=7 [action=SCHED]
> > ksoftirqd/19-125 [019] 170.701758: softirq_exit: vec=7 [action=SCHED]
> > ksoftirqd/19-125 [019] 170.701759: sched_switch: ksoftirqd/19:125 [120] S ==> sysjitter:2459 [120]
> > sysjitter-2459 [019] 170.702740: local_timer_entry: vector=236
> > sysjitter-2459 [019] 170.702742: softirq_raise: vec=1 [action=TIMER]
>
> A new tick but we still have pending timer callback so we'll need to wakeup ksoftirqd
> again.
>
> I think you should trace timers and check which one is concerned here.

Hummm, so I enabled 'timer:*', anything else you think I should be
looking at?

...
ksoftirqd/13-117 [013] 148.265945: softirq_entry: vec=1 [action=TIMER]
ksoftirqd/13-117 [013] 148.265947: softirq_exit: vec=1 [action=TIMER]
ksoftirqd/13-117 [013] 148.265948: softirq_entry: vec=7 [action=SCHED]
ksoftirqd/13-117 [013] 148.265950: softirq_exit: vec=7 [action=SCHED]
ksoftirqd/13-117 [013] 148.265952: sched_switch: ksoftirqd/13:117 [120] S ==> sysjitter:2536 [120]
sysjitter-2536 [013] 148.266912: local_timer_entry: vector=236
sysjitter-2536 [013] 148.266914: hrtimer_cancel: hrtimer=0xffff9807df91bee0
sysjitter-2536 [013] 148.266916: hrtimer_expire_entry: hrtimer=0xffff9807df91bee0 now=148249107837 function=tick_sched_timer/0x0
sysjitter-2536 [013] 148.266918: softirq_raise: vec=1 [action=TIMER]
sysjitter-2536 [013] 148.266920: softirq_raise: vec=7 [action=SCHED]
sysjitter-2536 [013] 148.266922: hrtimer_expire_exit: hrtimer=0xffff9807df91bee0
sysjitter-2536 [013] 148.266924: hrtimer_start: hrtimer=0xffff9807df91bee0 function=tick_sched_timer/0x0 expires=148250105422 softexpires=148250105422
sysjitter-2536 [013] 148.266926: local_timer_exit: vector=236
sysjitter-2536 [013] 148.266930: sched_wakeup: ksoftirqd/13:117 [120] success=1 CPU:013
sysjitter-2536 [013] 148.266932: tick_stop: success=0 dependency=SCHED
sysjitter-2536 [013] 148.266934: irq_work_entry: vector=246
sysjitter-2536 [013] 148.266936: irq_work_exit: vector=246
sysjitter-2536 [013] 148.266937: tick_stop: success=0 dependency=SCHED
sysjitter-2536 [013] 148.266940: sched_switch: sysjitter:2536 [120] R ==> ksoftirqd/13:117 [120]
ksoftirqd/13-117 [013] 148.266943: softirq_entry: vec=1 [action=TIMER]
ksoftirqd/13-117 [013] 148.266946: softirq_exit: vec=1 [action=TIMER]
ksoftirqd/13-117 [013] 148.266946: softirq_entry: vec=7 [action=SCHED]
ksoftirqd/13-117 [013] 148.266948: softirq_exit: vec=7 [action=SCHED]
ksoftirqd/13-117 [013] 148.266950: sched_switch: ksoftirqd/13:117 [120] S ==> sysjitter:2536 [120]
sysjitter-2536 [013] 148.267912: local_timer_entry: vector=236
sysjitter-2536 [013] 148.267914: hrtimer_cancel: hrtimer=0xffff9807df91bee0
sysjitter-2536 [013] 148.267916: hrtimer_expire_entry: hrtimer=0xffff9807df91bee0 now=148250107750 function=tick_sched_timer/0x0
sysjitter-2536 [013] 148.267918: softirq_raise: vec=1 [action=TIMER]
sysjitter-2536 [013] 148.267920: softirq_raise: vec=7 [action=SCHED]
sysjitter-2536 [013] 148.267922: hrtimer_expire_exit: hrtimer=0xffff9807df91bee0
sysjitter-2536 [013] 148.267924: hrtimer_start: hrtimer=0xffff9807df91bee0 function=tick_sched_timer/0x0 expires=148251105422 softexpires=148251105422
sysjitter-2536 [013] 148.267926: local_timer_exit: vector=236
sysjitter-2536 [013] 148.267929: sched_wakeup: ksoftirqd/13:117 [120] success=1 CPU:013
sysjitter-2536 [013] 148.267931: tick_stop: success=0 dependency=SCHED
sysjitter-2536 [013] 148.267934: irq_work_entry: vector=246
sysjitter-2536 [013] 148.267936: irq_work_exit: vector=246
sysjitter-2536 [013] 148.267937: tick_stop: success=0 dependency=SCHED
sysjitter-2536 [013] 148.267940: sched_switch: sysjitter:2536 [120] R ==> ksoftirqd/13:117 [120]
ksoftirqd/13-117 [013] 148.267943: softirq_entry: vec=1 [action=TIMER]
ksoftirqd/13-117 [013] 148.267945: softirq_exit: vec=1 [action=TIMER]
ksoftirqd/13-117 [013] 148.267946: softirq_entry: vec=7 [action=SCHED]
ksoftirqd/13-117 [013] 148.267947: softirq_exit: vec=7 [action=SCHED]
ksoftirqd/13-117 [013] 148.267949: sched_switch: ksoftirqd/13:117 [120] S ==> sysjitter:2536 [120]
sysjitter-2536 [013] 148.268865: call_function_single_entry: vector=251
sysjitter-2536 [013] 148.268867: call_function_single_exit: vector=251
sysjitter-2536 [013] 148.268870: tick_stop: success=1 dependency=NONE
sysjitter-2536 [013] 148.268872: hrtimer_cancel: hrtimer=0xffff9807df91bee0
sysjitter-2536 [013] 148.268876: call_function_single_entry: vector=251
sysjitter-2536 [013] 148.268878: call_function_single_exit: vector=251
...

it looks like tick sched timer is kept armed until tick_stop
succedes (because of the check performed after _single_exit)?

Thanks for your quick reply!