Re: [RFC][PATCH 9/9] arch/idle: Change arch_cpu_idle() IRQ behaviour

["Gautham R. Shenoy" <gautham.shenoy@amd.com>]

On Thu, May 19, 2022 at 11:27:59PM +0200, Peter Zijlstra wrote:
> Current arch_cpu_idle() is called with IRQs disabled, but will return
> with IRQs enabled.
> 
> However, the very first thing the generic code does after calling
> arch_cpu_idle() is raw_local_irq_disable(). This means that
> architectures that can idle with IRQs disabled end up doing a
> pointless 'enable-disable' dance.
> 
> Therefore, push this IRQ disabling into the idle function, meaning
> that those architectures can avoid the pointless IRQ state flipping.
> 
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> ---

[...]

> --- a/arch/x86/kernel/process.c
> +++ b/arch/x86/kernel/process.c
> @@ -699,6 +699,7 @@ EXPORT_SYMBOL(boot_option_idle_override)
>  void __cpuidle default_idle(void)
>  {
>  	raw_safe_halt();
> +	raw_local_irq_disable();
>  }
>  #if defined(CONFIG_APM_MODULE) || defined(CONFIG_HALTPOLL_CPUIDLE_MODULE)
>  EXPORT_SYMBOL(default_idle);
> @@ -804,13 +805,7 @@ static void amd_e400_idle(void)
>  
>  	default_idle();
>  
> -	/*
> -	 * The switch back from broadcast mode needs to be called with
> -	 * interrupts disabled.
> -	 */
> -	raw_local_irq_disable();
>  	tick_broadcast_exit();
> -	raw_local_irq_enable();
>  }
>  
>  /*
> @@ -849,12 +844,11 @@ static __cpuidle void mwait_idle(void)
>  		}
>  
>  		__monitor((void *)&current_thread_info()->flags, 0, 0);
> -		if (!need_resched())
> +		if (!need_resched()) {
>  			__sti_mwait(0, 0);
> -		else
> -			raw_local_irq_enable();
> +			raw_local_irq_disable();
> +		}
>  	} else {
> -		raw_local_irq_enable();
>  	}


We don't need the outer else clause anymore.

>  	__current_clr_polling();
>  }


--
Thanks and Regards
gautham.