[lm-sensors] [PATCH 4/4] hwmon: (pmbus) Auto-detect temp2 and temp3 registers/attributes

[Guenter Roeck]

Hi Jean,

On Sun, Jul 03, 2011 at 07:14:28AM -0400, Jean Delvare wrote:
> Hi Guenter,
> 
> On Thu, 30 Jun 2011 10:29:53 -0700, Guenter Roeck wrote:
> > Additional temperature attribute support is easy to detect, so do it.
> > 
> > Signed-off-by: Guenter Roeck <guenter.roeck at ericsson.com>
> > ---
> >  drivers/hwmon/pmbus.c |    4 ++++
> >  1 files changed, 4 insertions(+), 0 deletions(-)
> > 
> > diff --git a/drivers/hwmon/pmbus.c b/drivers/hwmon/pmbus.c
> > index b0ea00b..931d940 100644
> > --- a/drivers/hwmon/pmbus.c
> > +++ b/drivers/hwmon/pmbus.c
> > @@ -65,6 +65,10 @@ static void pmbus_find_sensor_groups(struct i2c_client *client,
> >  					      PMBUS_STATUS_TEMPERATURE))
> >  			info->func[0] |= PMBUS_HAVE_STATUS_TEMP;
> >  	}
> > +	if (pmbus_check_word_register(client, 0, PMBUS_READ_TEMPERATURE_2))
> > +		info->func[0] |= PMBUS_HAVE_TEMP2;
> > +	if (pmbus_check_word_register(client, 0, PMBUS_READ_TEMPERATURE_3))
> > +		info->func[0] |= PMBUS_HAVE_TEMP3;
> 
> Looks good, but is it not possible for a device to have temp2 (or
> temp3) but not temp1? If it is, then the test for
> PMBUS_STATUS_TEMPERATURE should be moved after the test for
> PMBUS_READ_TEMPERATURE_3, and should depend on any of PMBUS_HAVE_TEMP*
> being set. If it is not possible, then the code is fine as it is and:
> 
Yes, you are right, that condition can happen. I'll update the code and submit
another patch for it. Not that urgent, since I am not (yet) aware of a PMBus
device actually doing it, but given the inventiveness of PMBus device developers
I am sure we will hit that condition at some point.

Thanks,
Guenter