Re: [PATCH v2 2/2] thermal: trip: Rework thermal_zone_set_trip() and its callers

[Lukasz Luba]

On 11/28/23 13:58, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
>
> Both trip_point_temp_store() and trip_point_hyst_store() use
> thermal_zone_set_trip() to update a given trip point, but none of them
> actually needs to change more than one field in struct thermal_trip
> representing it.  However, each of them effectively calls
> __thermal_zone_get_trip() twice in a row for the same trip index value,
> once directly and once via thermal_zone_set_trip(), which is not
> particularly efficient, and the way in which thermal_zone_set_trip()
> carries out the update is not particularly straightforward.
>
> Moreover, some checks done by them both need not go under the thermal
> zone lock and code duplication between them can be reduced quite a bit
> by moving the majority of logic into thermal_zone_set_trip().
>
> Rework all of the above functions to address the above.
>
> No intentional functional impact.
>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
> ---
>
> v1 -> v2:
>     * Fix 2 typos in the changelog (Lukasz).
>     * Split one change into the [1/2].
>
> ---
>   drivers/thermal/thermal_core.h  |    9 ++++++
>   drivers/thermal/thermal_sysfs.c |   52 ++++++++--------------------------
>   drivers/thermal/thermal_trip.c  |   60 +++++++++++++++++++++++++++-------------
>   include/linux/thermal.h         |    3 --
>   4 files changed, 62 insertions(+), 62 deletions(-)

[snip]

> Index: linux-pm/drivers/thermal/thermal_trip.c
> ===================================================================
> --- linux-pm.orig/drivers/thermal/thermal_trip.c
> +++ linux-pm/drivers/thermal/thermal_trip.c
> @@ -148,39 +148,61 @@ int thermal_zone_get_trip(struct thermal
>   EXPORT_SYMBOL_GPL(thermal_zone_get_trip);
>   
>   int thermal_zone_set_trip(struct thermal_zone_device *tz, int trip_id,
> -			  const struct thermal_trip *trip)
> +			  enum thermal_set_trip_target what, const char *buf)
>   {
> -	struct thermal_trip t;
> -	int ret;
> +	struct thermal_trip *trip;
> +	int val, ret = 0;
>   
> -	ret = __thermal_zone_get_trip(tz, trip_id, &t);
> +	if (trip_id < 0 || trip_id >= tz->num_trips)
> +		ret = -EINVAL;

That shouldn't progress forward IMO, but simply 'return -EINVAL;'...

> +
> +	ret = kstrtoint(buf, 10, &val);
>   	if (ret)
>   		return ret;
>   
> -	if (t.type != trip->type)
> -		return -EINVAL;
> +	mutex_lock(&tz->lock);
>   
> -	if (t.temperature != trip->temperature && tz->ops->set_trip_temp) {
> -		ret = tz->ops->set_trip_temp(tz, trip_id, trip->temperature);
> -		if (ret)
> -			return ret;
> -	}
> +	trip = &tz->trips[trip_id];

... because here we might get an issue.