Re: [PATCH v2] cpuidle: Add sanity check for exit latency and target residency

[Christian Loehle <christian.loehle@arm.com>]

On 11/7/25 19:07, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Make __cpuidle_driver_init() fail if the exit latency of one of the
> driver's idle states is less than its exit latency which would break

s/exit latency/target residency/
for the second one here.

Apart from that
Reviewed-by: Christian Loehle <christian.loehle@arm.com>

> cpuidle assumptions.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
> 
> v1 -> v2: Make __cpuidle_driver_init() fail if the check is not passed (Artem).
> 
> ---
>  drivers/cpuidle/driver.c |   16 ++++++++++++++--
>  1 file changed, 14 insertions(+), 2 deletions(-)
> 
> --- a/drivers/cpuidle/driver.c
> +++ b/drivers/cpuidle/driver.c
> @@ -152,7 +152,7 @@ static void cpuidle_setup_broadcast_time
>   * __cpuidle_driver_init - initialize the driver's internal data
>   * @drv: a valid pointer to a struct cpuidle_driver
>   */
> -static void __cpuidle_driver_init(struct cpuidle_driver *drv)
> +static int __cpuidle_driver_init(struct cpuidle_driver *drv)
>  {
>  	int i;
>  
> @@ -193,7 +193,17 @@ static void __cpuidle_driver_init(struct
>  			s->exit_latency_ns =  0;
>  		else
>  			s->exit_latency = div_u64(s->exit_latency_ns, NSEC_PER_USEC);
> +
> +		/*
> +		 * Ensure that the exit latency of a CPU idle state does not
> +		 * exceed its target residency which is assumed in cpuidle in
> +		 * multiple places.
> +		 */
> +		if (s->exit_latency_ns > s->target_residency_ns)
> +			return -EINVAL;
>  	}
> +
> +	return 0;
>  }
>  
>  /**
> @@ -223,7 +233,9 @@ static int __cpuidle_register_driver(str
>  	if (cpuidle_disabled())
>  		return -ENODEV;
>  
> -	__cpuidle_driver_init(drv);
> +	ret = __cpuidle_driver_init(drv);
> +	if (ret)
> +		return ret;
>  
>  	ret = __cpuidle_set_driver(drv);
>  	if (ret)
> 
> 
> 
>