28. Mai 2020 21:22, "Andrzej Pietrasiewicz" <andrzej.p@xxxxxxxxxxxxx> schrieb:
> Use thermal_zone_device_{en|dis}able() and thermal_zone_device_is_enabled().
>
> Consequently, all set_mode() implementations in drivers:
>
> - can stop modifying tzd's "mode" member,
> - shall stop taking tzd's lock, as it is taken in the helpers
> - shall stop calling thermal_zone_device_update() as it is called in the
> helpers
> - can assume they are called when the mode truly changes, so checks to
> verify that can be dropped
>
> Not providing set_mode() by a driver no longer prevents the core from
> being able to set tzd's mode, so the relevant check in mode_store() is
> removed.
>
> Other comments:
>
> - acpi/thermal.c: tz->thermal_zone->mode will be updated only after we
> return from set_mode(), so use function parameter in thermal_set_mode()
> instead, no need to call acpi_thermal_check() in set_mode()
> - thermal/imx_thermal.c: regmap writes and mode assignment are done in
> thermal_zone_device_{en|dis}able() and set_mode() callback
> - thermal/intel/intel_quark_dts_thermal.c: soc_dts_{en|dis}able() are a
> part of set_mode() callback, so they don't need to modify tzd->mode, and
> don't need to fall back to the opposite mode if unsuccessful, as the return
> value will be propagated to thermal_zone_device_{en|dis}able() and
> ultimately tzd's member will not be changed in thermal_zone_device_set_mode().
> - thermal/of-thermal.c: no need to set zone->mode to DISABLED in
> of_parse_thermal_zones() as a tzd is kzalloc'ed so mode is DISABLED anyway
>
> Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@xxxxxxxxxxxxx>
> ---
[...]
> drivers/platform/x86/acerhdf.c | 17 +++++----
Acked-by: Peter Kaestle <peter@xxxxxxxx>
