phosphor-hwmon funcitonal interface

Matthew Barth msbarth at linux.ibm.com
Wed Jan 16 02:01:41 AEDT 2019



On 1/14/19 2:31 PM, Patrick Venture wrote:
> Right now, if a sensor has an explicit hwmon fault interface, then we
> have the notion of a status:
> 
> https://github.com/openbmc/phosphor-hwmon/blob/master/sensor.cpp#L205
> 
> However, I propose that if there isn't a fault file, we still add the
> interface, and set to functional=true.  This would allow us to change
> it to false in the following condition:
> 
Would this be optional so this interface/property doesnt show on sensors 
that dont want it?
> https://github.com/openbmc/phosphor-hwmon/blob/master/hwmonio.cpp#L127
> 
> So basically, instead of returning -rc on failure (optionally), it
> could optionally set functional to false, and phosphor-host-ipmid
> would need to know to check this additional interface.
> 
I'm ok with this. Where and how would this optionally be configured for 
a device and its sensors?

> Thoughts?  (IIRC< I've mentioned this idea before, but now i'm setting
> up some cycles to hammer it out, as I continue to implement unit-tests
> in phosphor-hwmon)
> 
> Patrick
> 



More information about the openbmc mailing list