Re: [PATCH 0/9] TRACE_IRQFLAGS wreckage

[Steven Rostedt]

On Thu, 20 Aug 2020 16:58:21 +0200
peterz@xxxxxxxxxxxxx wrote:

> @@ -91,11 +91,11 @@ void __cpuidle default_idle_call(void)
>  	if (current_clr_polling_and_test()) {
>  		local_irq_enable();
>  	} else {
> -		rcu_idle_enter();
>  		stop_critical_timings();
> +		rcu_idle_enter();
>  		arch_cpu_idle();
> -		start_critical_timings();
>  		rcu_idle_exit();
> +		start_critical_timings();
>  	}
>  }
>  

tip/master didn't have the above function and instead I had this:

@@ -93,11 +93,11 @@ void __cpuidle default_idle_call(void)
        } else {
 
                trace_cpu_idle(1, smp_processor_id());
-               rcu_idle_enter();
                stop_critical_timings();
+               rcu_idle_enter();
                arch_cpu_idle();
-               start_critical_timings();
                rcu_idle_exit();
+               start_critical_timings();
                trace_cpu_idle(PWR_EVENT_EXIT, smp_processor_id());
        }
 }

But it booted to completion without warning with this patch applied.

Reviewed-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>

For the entire series, with this update.

-- Steve