On 15/01/2020 19:39, Niklas Söderlund wrote: > Hi Daniel, > > On 2020-01-15 19:15:11 +0100, Daniel Lezcano wrote: >> On 15/01/2020 14:45, Niklas Söderlund wrote: >>> Hi Daniel, >>> >>> Thanks for your feedback. >>> >>> On 2020-01-15 14:24:30 +0100, Daniel Lezcano wrote: >>>> On 14/01/2020 23:29, Niklas Söderlund wrote: >>>>> The hardware manual states that the operation of the sensor is not >>>>> guaranteed outside the range of -40°C to 125°C, not that the readings >>>>> are invalid. Remove the bound check and try to deliver temperature >>>>> readings even if we are outside the guaranteed operation range. >>>> >>>> And what if the sensor is returning crap in this out-of-range operation? >>> >>> I'm not sure what is worse, reporting an untrue (but still outside the >>> guaranteed operation range) extreme temperature or failing with -EIO. >>> The view of the hardware guys is that it's better to report what the >>> sensor indicates then to return -EIO. >> >> I don't get the point. >> >> What happens if we read the sensor while it is above or below the limits? > > The manual describes the read outs as being +/- 2°C from the true > temperature for the guaranteed operating range. Outside this range the > difference between the true temperature and the read out temperature > might be larger than 2°C. Ok, I see the point now. I guess in any case the SoC won't be working very well outside of these operating range also. It would be interesting to add in the core code a warning when reaching the sensor's operating range. Please replace the code deletion in your patches by a comment keeping the range values for reference if we add a sensor boundaries property later. Thanks -- Daniel -- <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog