hello!
On Sun, 2007-11-25 at 14:08 +0100, Jean Delvare wrote:
> > > +static ssize_t show_fan_input(struct device *dev,
> > > + struct device_attribute *attr, char *buf)
> > > +{
> > > + int nr = to_sensor_dev_attr(attr)->index;
> > > + struct gl518_data *data = gl518_update_device(dev);
> > > + return sprintf(buf, "%d\n", FAN_FROM_REG(data->fan_in[nr],
> > > + DIV_FROM_REG(data->fan_div[nr])));
> > > +}
> > > +
> > > +static ssize_t show_fan_min(struct device *dev,
> > > + struct device_attribute *attr, char *buf)
> > > +{
> > > + int nr = to_sensor_dev_attr(attr)->index;
> > > + struct gl518_data *data = gl518_update_device(dev);
> > > + return sprintf(buf, "%d\n", FAN_FROM_REG(data->fan_min[nr],
> > > + DIV_FROM_REG(data->fan_div[nr])));
> > > +}
one thing just sprang to my mind... is it safe to assume that the
expressions data->fan_min[nr] and data->fan_div[nr] are evaluated
atomically? or to put it differently, why don't we need a rw-lock here
to protect against interleaved updates?
cheers,
hvr
