Quoting Gwendal Grignou (2021-01-26 12:59:50) > On Mon, Jan 25, 2021 at 3:44 PM Stephen Boyd <swboyd@xxxxxxxxxxxx> wrote: > > > > instead of having to do > > an offset after the fact. For now I'll use ec_dev->last_event_time > > because this is only used on chromeos and that should work until > > userspace is changed, but in the future I think we'll need to have a way > > for this IIO device to be notified when the clock base changes in > > iio_device_set_clock() and then have this driver call into cros_ec to > > request that such a timestamp be made when this event is seen. Or even > > better have a way to request that cros_ec timestamp the event itself on > > the EC side, but I don't know if that's possible. > One way would be use the EC sensor stack that collect such timestamp, > but that would be more firmware changes. > > On second thought, to keep it simple and consistent with other IIO > drivers, I suggest to keep using iio_get_time_ns() when the sensor > clock is not CLOCK_BOOTTIME, ec_dev->last_event_time when it is. Ok I will make that change and send v3.
