Re: [PATCH net-next 1/3] mlxsw: core_thermal: Use static trip points for transceiver modules

[Simon Horman <simon.horman@corigine.com>]

On Fri, Mar 31, 2023 at 04:17:30PM +0200, Petr Machata wrote:
> From: Ido Schimmel <idosch@nvidia.com>
> 
> The driver registers a thermal zone for each transceiver module and
> tries to set the trip point temperatures according to the thresholds
> read from the transceiver. If a threshold cannot be read or if a
> transceiver is unplugged, the trip point temperature is set to zero,
> which means that it is disabled as far as the thermal subsystem is
> concerned.
> 
> A recent change in the thermal core made it so that such trip points are
> no longer marked as disabled, which lead the thermal subsystem to
> incorrectly set the associated cooling devices to the their maximum
> state [1]. A fix to restore this behavior was merged in commit
> f1b80a3878b2 ("thermal: core: Restore behavior regarding invalid trip
> points"). However, the thermal maintainer suggested to not rely on this
> behavior and instead always register a valid array of trip points [2].
> 
> Therefore, create a static array of trip points with sane defaults
> (suggested by Vadim) and register it with the thermal zone of each
> transceiver module. User space can choose to override these defaults
> using the thermal zone sysfs interface since these files are writeable.
> 
> Before:
> 
>  $ cat /sys/class/thermal/thermal_zone11/type
>  mlxsw-module11
>  $ cat /sys/class/thermal/thermal_zone11/trip_point_*_temp
>  65000
>  75000
>  80000
> 
> After:
> 
>  $ cat /sys/class/thermal/thermal_zone11/type
>  mlxsw-module11
>  $ cat /sys/class/thermal/thermal_zone11/trip_point_*_temp
>  55000
>  65000
>  80000
> 
> Also tested by reverting commit f1b80a3878b2 ("thermal: core: Restore
> behavior regarding invalid trip points") and making sure that the
> associated cooling devices are not set to their maximum state.
> 
> [1] https://lore.kernel.org/linux-pm/ZA3CFNhU4AbtsP4G@shredder/
> [2] https://lore.kernel.org/linux-pm/f78e6b70-a963-c0ca-a4b2-0d4c6aeef1fb@linaro.org/
> 
> Signed-off-by: Ido Schimmel <idosch@nvidia.com>
> Reviewed-by: Vadim Pasternak <vadimp@nvidia.com>
> Signed-off-by: Petr Machata <petrm@nvidia.com>

Reviewed-by: Simon Horman <simon.horman@corigine.com>