On Mo, 2016-06-06 at 20:18 -0700, Guenter Roeck wrote:
> On 06/06/2016 02:21 AM, thilo.cestonaro@xxxxxxxxxxxxxx wrote:
> >
> > Hey Guenter!
> >
> > Thanks for your patience and very detailed review!
> > I will change the code to address all you hints where no discussion is needed.
> > The others follow down here.
> >
> > On Fr, 2016-06-03 at 22:51 -0700, Guenter Roeck wrote:
> > >
> > > >
> > > > +#define FTSTEUTATES_WATCHDOG_RESOLUTION 60
> > > > +
> > > 60 seconds (minimum) resolution ? Really ? This is very unusual.
> > Will double check and talk to the hw guy.
There is a register which sets the resolution to 1sec. but it wasn't documented :(.
So next "version" will have 1sec. resolution.
> > > > +static ssize_t show_fan_fault(struct device *dev,
> > > > + struct device_attribute *devattr, char *buf)
> > > > +{
> > > > + int index = to_sensor_dev_attr(devattr)->index;
> > > > + struct ftsteutates_data *data = ftsteutates_update_device(dev);
> > > > +
> > > > + return sprintf(buf, "%d\n", data->fan_present[index] == 1 ? 0 : 1);
> > > A non-present fan does not indicate a fan fault.
> > For me a fan fault was a not existing fan, as the alarm is handled via fan_alarm.
> > What is the understanding of fan_fault and fan_alarm?
> >
> Some fan controllers can detect if a fan is faulty, for example if it consumes
> power but does not turn. Unless you know for sure that a fan is faulty, don't
> report that it is. A non-existing fan is definitely not faulty.
About the presence register from Teutates Spec:
The present-detection is done by checking the Fan-Speed.
If the fan-speed is 0 or below fan-fault-speed all the time,
the fan will be handled as not present. If the fan-speed is
greater than fan-fault-speed, the fan is handled as present.
So the presence is the faulty fan detection?
Cheers,
Thilo
<<attachment: smime.p7s>>
