On Wed, Nov 13, 2024 at 07:18:04PM +0100, Daniel Lezcano wrote: > > Hi Ansuel, > > On 13/11/2024 16:56, Christian Marangi wrote: > > On Fri, Oct 18, 2024 at 12:48:04PM +0200, Christian Marangi wrote: > > > Add support for Airoha EN7581 thermal sensor and monitor. This is a > > > simple sensor for the CPU or SoC Package that provide thermal sensor and > > > trip point for hot low and critical condition to fire interrupt and > > > react on the abnormal state. > > > > > > Signed-off-by: Christian Marangi <ansuelsmth@xxxxxxxxx> > > > Reviewed-by: Rob Herring (Arm) <robh@xxxxxxxxxx> > > > > Any news with this series? Everything wrong with the thermal core small > > patch? > > I understand why you are trying to achieve this but usually it is the kernel > which overloads the firmware description, not the opposite, no? > > Either way, we ignore the offset/slope from tzp and use a couple of private > variables offset/slope in the driver (iow do not call > thermal_zone_get_offset() thermal_zone_get_slope()). Or add the > thermal_zone_set_offset() and thermal_zone_set_slope() helpers. > > I would prefer the first solution as for today I can not see any DT for > ARM64 with the coefficients set. So may be we can consider the slope and the > offset as a legacy which should be removed from sysfs and the thermal zone > device parameters in a near future. > Hi Daniel, Having set OPs is problematic as that would diverge from what is set in DT that should always have priority. Well yes my idea was trying to make use of them as currently there are many driver that set these values but have the slope and offset in thermal core always set to 0 and 1. Thing is that reading temp with ADC is very common and in some way or another you always have a slope and an offset so it makese sense to permit to have those values preallocated instead of handling them in priv struct. Anyway if the idea is to drop that, I will happly move those values handling back in the driver. Just need a confirm on that. Also thanks for the feedback! -- Ansuel
