[lm-sensors] f71889ed temp alarm polarity reversed.

[lm-sensors] f71889ed temp alarm polarity reversed.

[James Cloos]

As of Linus' v2.6.39-rc4-89-g2f666bc I still get:

fan1:         871 RPM
fan2:           0 RPM  ALARM
fan3:           0 RPM  ALARM
temp1:        +35.0°C  (high = +255.0°C, hyst = +251.0°C)  ALARM (CRIT)
                       (crit = +255.0°C, hyst = +251.0°C)  sensor = transistor
temp2:          FAULT  (high = +255.0°C, hyst = +251.0°C)
                       (crit = +255.0°C, hyst = +251.0°C)  sensor = transistor
temp3:        +38.0°C  (high = +255.0°C, hyst = +253.0°C)  ALARM (CRIT)
                       (crit = +255.0°C, hyst = +253.0°C)  sensor = transistor

The polarity for the fan alarms is correct, but it is reversed for the
temp alarms.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] f71889ed temp alarm polarity reversed.

[Hans de Goede]

Hi,

On 04/24/2011 12:39 AM, James Cloos wrote:
> As of Linus' v2.6.39-rc4-89-g2f666bc I still get:
>
> fan1:         871 RPM
> fan2:           0 RPM  ALARM
> fan3:           0 RPM  ALARM
> temp1:        +35.0°C  (high = +255.0°C, hyst = +251.0°C)  ALARM (CRIT)
>                         (crit = +255.0°C, hyst = +251.0°C)  sensor = transistor
> temp2:          FAULT  (high = +255.0°C, hyst = +251.0°C)
>                         (crit = +255.0°C, hyst = +251.0°C)  sensor = transistor
> temp3:        +38.0°C  (high = +255.0°C, hyst = +253.0°C)  ALARM (CRIT)
>                         (crit = +255.0°C, hyst = +253.0°C)  sensor = transistor
>
> The polarity for the fan alarms is correct, but it is reversed for the
> temp alarms.

I just checked the data sheet and it disagrees with your observation. It is
possible that the temp has been outside alarm boundaries before they
were set to their current settings, and the alarm register has never
been read before. Try running the sensors command again (after waiting
a couple of seconds otherwise you will get cached readings).

Or maybe the problem is that 255 gets seen as -1 ? The F71889ED does
have signed values in certain temp registers (but not the alarm thresholds
according to the datasheet). Try writing 100 to the temp#_max and
temp#_crit sysfs attributes, you can find these files under
/sys/class/hwmon#, or /sys/class/hwmon#/device.

Regards,

Hans

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] f71889ed temp alarm polarity reversed.

[Jean Delvare]

On Sun, 24 Apr 2011 16:57:12 +0200, Hans de Goede wrote:
> Hi,
> 
> On 04/24/2011 12:39 AM, James Cloos wrote:
> > As of Linus' v2.6.39-rc4-89-g2f666bc I still get:
> >
> > fan1:         871 RPM
> > fan2:           0 RPM  ALARM
> > fan3:           0 RPM  ALARM
> > temp1:        +35.0°C  (high = +255.0°C, hyst = +251.0°C)  ALARM (CRIT)
> >                         (crit = +255.0°C, hyst = +251.0°C)  sensor = transistor
> > temp2:          FAULT  (high = +255.0°C, hyst = +251.0°C)
> >                         (crit = +255.0°C, hyst = +251.0°C)  sensor = transistor
> > temp3:        +38.0°C  (high = +255.0°C, hyst = +253.0°C)  ALARM (CRIT)
> >                         (crit = +255.0°C, hyst = +253.0°C)  sensor = transistor
> >
> > The polarity for the fan alarms is correct, but it is reversed for the
> > temp alarms.
> 
> I just checked the data sheet and it disagrees with your observation. It is
> possible that the temp has been outside alarm boundaries before they
> were set to their current settings, and the alarm register has never
> been read before. Try running the sensors command again (after waiting
> a couple of seconds otherwise you will get cached readings).

Note that only the critical limit alarms appear to be set.

> Or maybe the problem is that 255 gets seen as -1 ? The F71889ED does
> have signed values in certain temp registers (but not the alarm thresholds
> according to the datasheet). Try writing 100 to the temp#_max and
> temp#_crit sysfs attributes, you can find these files under
> /sys/class/hwmon#, or /sys/class/hwmon#/device.

Actually value 100000 and directory /sys/class/hwmon/hwmon#,
or /sys/class/hwmon/hwmon#/device. But it may be easier to add:

chip "f71889ed-*"

   set temp1_max   90
   set temp1_crit 100
   set temp3_max   90
   set temp3_crit 100

to /etc/sensors.d/f71889ed.conf and run "sensors -s".

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] f71889ed temp alarm polarity reversed.

[James Cloos]

>>>>> "JD" = Jean Delvare <khali@linux-fr.org> writes:

JD> Actually value 100000 and directory /sys/class/hwmon/hwmon#,
JD> or /sys/class/hwmon/hwmon#/device. But it may be easier to add:

JD> chip "f71889ed-*"

JD>    set temp1_max   90
JD>    set temp1_crit 100
JD>    set temp3_max   90
JD>    set temp3_crit 100

JD> to /etc/sensors.d/f71889ed.conf and run "sensors -s".

That alleviated the false alarms.  Thanks.

The fact that the temp sensor which is not connected was not in alarm,
whereas the two which are were, led to my false diagnosis.

Thanks both!

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors