[PATCH] sched/core: Fix kick offline cpu to do nohz idle load balance

[PATCH] sched/core: Fix kick offline cpu to do nohz idle load balance

[Wanpeng Li]

From: Wanpeng Li <wanpeng.li@hotmail.com>

 WARNING: CPU: 0 PID: 3404 at arch/x86/kernel/smp.c:125 native_smp_send_reschedule+0x3f/0x50
 CPU: 0 PID: 3404 Comm: qemu-system-x86 Not tainted 4.8.0+ #21
 Call Trace:
   __warn+0xd1/0xf0
   warn_slowpath_null+0x1d/0x20
   native_smp_send_reschedule+0x3f/0x50
   trigger_load_balance+0x29c/0x4a0
   ? trigger_load_balance+0x72/0x4a0
   scheduler_tick+0x9f/0xd0
   ? tick_sched_do_timer+0x50/0x50
   update_process_times+0x47/0x60
   tick_sched_handle.isra.24+0x25/0x60
   tick_sched_timer+0x3d/0x70
   __hrtimer_run_queues+0xf4/0x510
   hrtimer_interrupt+0xb7/0x1d0
   local_apic_timer_interrupt+0x35/0x60
   smp_apic_timer_interrupt+0x3d/0x50
   apic_timer_interrupt+0x96/0xa0

If there is a need to kick the idle load balancer, an ILB will be selected 
to perform nohz idle load balance, however, if the selected ILB is in the 
process of offline, smp_sched_reschedule() which generates a sched IPI will 
splat as above.

           CPU0                      CPU1 
                         
                                 find_new_ilb() 
    set_rq_offline() 
                                 smp_sched_reschedule()  Oops
    nohz_balance_exit_idle() 

This patch fix it by exiting nohz idle balance before set cpu offline.

Cc: Ingo Molnar <mingo@kernel.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
---
 kernel/sched/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 94732d1..7c83f99 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7412,6 +7412,7 @@ int sched_cpu_dying(unsigned int cpu)
 
 	/* Handle pending wakeups and then migrate everything off */
 	sched_ttwu_pending();
+	nohz_balance_exit_idle(cpu);
 	raw_spin_lock_irqsave(&rq->lock, flags);
 	if (rq->rd) {
 		BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
@@ -7422,7 +7423,6 @@ int sched_cpu_dying(unsigned int cpu)
 	raw_spin_unlock_irqrestore(&rq->lock, flags);
 	calc_load_migrate(rq);
 	update_max_interval();
-	nohz_balance_exit_idle(cpu);
 	hrtick_clear(rq);
 	return 0;
 }
-- 
1.9.1

Re: [PATCH] sched/core: Fix kick offline cpu to do nohz idle load balance

[Wanpeng Li]

2016-10-10 12:10 GMT+08:00 Wanpeng Li <kernellwp@gmail.com>:
> From: Wanpeng Li <wanpeng.li@hotmail.com>
>
>  WARNING: CPU: 0 PID: 3404 at arch/x86/kernel/smp.c:125 native_smp_send_reschedule+0x3f/0x50
>  CPU: 0 PID: 3404 Comm: qemu-system-x86 Not tainted 4.8.0+ #21
>  Call Trace:
>    __warn+0xd1/0xf0
>    warn_slowpath_null+0x1d/0x20
>    native_smp_send_reschedule+0x3f/0x50
>    trigger_load_balance+0x29c/0x4a0
>    ? trigger_load_balance+0x72/0x4a0
>    scheduler_tick+0x9f/0xd0
>    ? tick_sched_do_timer+0x50/0x50
>    update_process_times+0x47/0x60
>    tick_sched_handle.isra.24+0x25/0x60
>    tick_sched_timer+0x3d/0x70
>    __hrtimer_run_queues+0xf4/0x510
>    hrtimer_interrupt+0xb7/0x1d0
>    local_apic_timer_interrupt+0x35/0x60
>    smp_apic_timer_interrupt+0x3d/0x50
>    apic_timer_interrupt+0x96/0xa0
>
> If there is a need to kick the idle load balancer, an ILB will be selected
> to perform nohz idle load balance, however, if the selected ILB is in the
> process of offline, smp_sched_reschedule() which generates a sched IPI will
> splat as above.
>
>            CPU0                      CPU1
>
>                                  find_new_ilb()
>     set_rq_offline()
>                                  smp_sched_reschedule()  Oops
>     nohz_balance_exit_idle()
>
> This patch fix it by exiting nohz idle balance before set cpu offline.

             CPU 0                       CPU1

                                          find_new_ilb()
    nohz_balance_exit_idle()
    set_rq_offline()
                                          smp_sched_reschedule()

It seems that the patch still can't avoid this race, so any proposal
is a great appreciated. :)

Regards,
Wanpeng Li

Re: [PATCH] sched/core: Fix kick offline cpu to do nohz idle load balance

[Peter Zijlstra]

On Mon, Oct 10, 2016 at 04:34:48PM +0800, Wanpeng Li wrote:
> > If there is a need to kick the idle load balancer, an ILB will be selected
> > to perform nohz idle load balance, however, if the selected ILB is in the
> > process of offline, smp_sched_reschedule() which generates a sched IPI will
> > splat as above.
> >
> >            CPU0                      CPU1
> >
> >                                  find_new_ilb()
> >     set_rq_offline()
> >                                  smp_sched_reschedule()  Oops
> >     nohz_balance_exit_idle()
> >
> > This patch fix it by exiting nohz idle balance before set cpu offline.
> 
>              CPU 0                       CPU1
> 
>                                           find_new_ilb()
>     nohz_balance_exit_idle()
>     set_rq_offline()
>                                           smp_sched_reschedule()
> 
> It seems that the patch still can't avoid this race, so any proposal
> is a great appreciated. :)


Not sure how this can happen, scheduler_tick() -> trigger_load_balance()
-> nohz_balancer_kick() is called with IRQs disabled, this too implies a
RCU-sched read side section.

And hotplug explicitly includes a rcu_sync_sched().

It would be find_new_ilb() is 'broken' in that it considers !active
CPUs. That's not immediately obvious.