On Tue, 2019-08-06 at 08:57 -0700, Guenter Roeck wrote: > > The hwmon subsystem is not the appropriate subsystem to report charge > levels > (measured in Ampere-seconds as far as I can see). hwmon is about > hardware > monitoring. > > I am not in favor of supporting single-shot measurements in the first > place. Most chips support it, but it is hardly ever supported by > drivers, > for the most part because power savings are negligible and the > downside > of supporting it (programming complexity, delayed readings) is > substantial. I see. Then I guess I will (at least for now) drop single shot mode since it's just adding complexity and I actually though the power savings were better. > What is your use case for both requirements (reporting charge levels > and single-shot mode) ? Both seem mutually exclusive. I don't have, yet, a use case for charge reporting. I was just asking because the device supports it and I saw it's not a standard hwmon type (although almost all the other measured types are). But I guess it can be used to monitor a battery attached on the ltc2947. So, if I end up adding a charge attribute, would it be a "no go" for accepting the driver upstream? Thanks, Nuno Sá
