On Mon, Oct 31, 2022 at 07:00:21AM -0700, Guenter Roeck wrote: > On Mon, Oct 31, 2022 at 01:25:23PM +0000, Sudeep Holla wrote: > > On Mon, Oct 31, 2022 at 11:40:18AM +0000, Cristian Marussi wrote: > > > Available sensors are enumerated and reported by the SCMI platform server > > > using a 16bit identification number; not all such sensors are of a type > > > supported by hwmon subsystem and, among the supported ones, only a subset > > > could be temperature sensors that have to be registered with the Thermal > > > Framework. > > > Potential clashes between hwmon channels indexes and the underlying real > > > sensors IDs do not play well with the hwmon<-->thermal bridge automatic > > > registration routines and could need a sensible number of fake dummy > > > sensors to be made up in order to keep indexes and IDs in sync. > > > > > > Avoid to use the hwmon<-->thermal bridge dropping the HWMON_C_REGISTER_TZ > > > attribute and instead explicit register temperature sensors directly with > > > the Thermal Framework. > > > > > > > Hi Guenter, > > > > FWIW from scmi perspective, > > > > Acked-by: Sudeep Holla <sudeep.holla@xxxxxxx> > > > > I was about to ask for your ack to pickup myself but I see there is no > > strict dependency for that. Not sure if you want to take this as fix for > > v6.1 as the thermal changes broke the existing support in SCMI hwmon > > or do you still prefer v6.2 ? > > > > It is a regression and should be applied to 6.1. I'll pick it up and send > a pull request to Linus later this week. Indeed, thanks for the confirmation. -- Regards, Sudeep
