Re: timer-sp804: sched_clock: issue after first suspend

From: Thomas Gleixner
Date: Tue Nov 28 2017 - 11:43:44 EST


On Mon, 27 Nov 2017, Abhishek Shah wrote:
> Apart from this, I see following prints at times, can the above
> discussion be related to this behavior?
> [ 320.448348] INFO: rcu_sched detected stalls on CPUs/tasks:
> [ 320.454015] 3-...: (90 GPs behind) idle=162/140000000000001/0
> softirq=1220/1220 fqs=2462
> [ 320.462540] (detected by 1, t=5255 jiffies, g=518, c=517, q=37)
> [ 320.468738] Task dump for CPU 3:
> [ 320.472067] swapper/3 R running task 0 0 1 0x00000022
> [ 320.479341] Call trace:
> [ 320.481869] [<ffff000008085828>] __switch_to+0x98/0xb0

I dont think so. That's an issue which plagues Paul McKenney for quite some
time. It's either caused by a timer not firing or a wakeup being lost. We
have no idea yet how to debug that.

Thanks,

tglx