> -----Original Message----- > From: Aisheng Dong > Sent: Thursday, May 16, 2019 6:40 PM > To: Anson Huang <anson.huang@xxxxxxx>; robh+dt@xxxxxxxxxx; > mark.rutland@xxxxxxx; shawnguo@xxxxxxxxxx; s.hauer@xxxxxxxxxxxxxx; > kernel@xxxxxxxxxxxxxx; festevam@xxxxxxxxx; catalin.marinas@xxxxxxx; > will.deacon@xxxxxxx; rui.zhang@xxxxxxxxx; edubezval@xxxxxxxxx; > daniel.lezcano@xxxxxxxxxx; ulf.hansson@xxxxxxxxxx; Daniel Baluta > <daniel.baluta@xxxxxxx>; Peng Fan <peng.fan@xxxxxxx>; > heiko@xxxxxxxxx; horms+renesas@xxxxxxxxxxxx; agross@xxxxxxxxxx; > olof@xxxxxxxxx; bjorn.andersson@xxxxxxxxxx; jagan@xxxxxxxxxxxxxxxxxxxx; > enric.balletbo@xxxxxxxxxxxxx; marc.w.gonzalez@xxxxxxx; > devicetree@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; linux-arm- > kernel@xxxxxxxxxxxxxxxxxxx; linux-pm@xxxxxxxxxxxxxxx > Cc: dl-linux-imx <linux-imx@xxxxxxx> > Subject: RE: [PATCH V12 3/5] thermal: imx_sc: add i.MX system controller > thermal support > > [...] > > > > > + if (!sensor) > > > > + return -ENOMEM; > > > > + > > > > + ret = thermal_zone_of_get_sensor_id(child, > > > > + &sensor_specs, > > > > > > It looks a bit strange why need sensor_specs as you even did not use it. > > > > It is because parsing the phandle arguments needs it, such as the > > thermal_zone_of_sensor_register() function needs this pointer to be > > passed back to check the elements are valid or NOT, so the API does > > NOT define it as local structure. The API NOT just return the sensor > > id, but also the of_phandle_args address, although we do NOT need it at all > here. > > The main problem is this one introduced in 2/5 which needs get ack from > maintainer. > For me, I would suggest to make sensor_specs transparent to user if it's > really not needed by users. > > Please try to make it either optional or hide to core users as well. > And I guess the later one is doable after a glance at the core code. OK, I will hide it to user if no other different comments. Anson. > > Regards > Dong Aisheng
