[lm-sensors] W83667HG-B testing

[Guenter Roeck]

On Tue, 2010-07-20 at 02:54 -0400, Jean Delvare wrote:
> On Mon, 19 Jul 2010 18:24:16 -0700, Guenter Roeck wrote:
> > On Sun, Jul 18, 2010 at 08:02:52AM -0400, Tobias Preclik wrote:
> > > Hi,
> > > 
> > > my Zotac board H55-ITX features the W83667HG-B chip. I downloaded and
> > > compiled the standalone kernel module today. Here's the sensors output:
> > > 
> > > The module loads fine:
> > > 
> > > # modprobe w83627ehf
> > > # tail -n 1 /var/log/messages
> > > Jul 18 13:42:58 leela kernel: [50378.028634] w83627ehf: Found W83667HG-B
> > > chip at 0xa10
> > > # sensors
> > > w83667hg-isa-0a10
> > > Adapter: ISA adapter
> > > in0:         +0.98 V  (min =  +0.00 V, max =  +1.74 V)  
> > > in1:         +1.11 V  (min =  +0.10 V, max =  +0.20 V)   ALARM
> > > in2:         +3.36 V  (min =  +3.95 V, max =  +3.28 V)   ALARM
> > > in3:         +3.36 V  (min =  +1.57 V, max =  +2.10 V)   ALARM
> > > in4:         +1.10 V  (min =  +1.74 V, max =  +1.87 V)   ALARM
> > > in5:         +1.18 V  (min =  +1.02 V, max =  +1.30 V)  
> > > in7:         +3.28 V  (min =  +3.23 V, max =  +2.62 V)   ALARM
> > > in8:         +3.34 V  (min =  +3.18 V, max =  +0.05 V)   ALARM
> > > fan1:          0 RPM  (min = 3515 RPM, div = 128)  ALARM
> > > fan2:       1022 RPM  (min = 1638 RPM, div = 8)  ALARM
> > > fan3:          0 RPM  (min =  958 RPM, div = 128)  ALARM
> > > fan4:          0 RPM  (min = 10546 RPM, div = 128)  ALARM
> > > fan5:          0 RPM  (min = 3515 RPM, div = 128)  ALARM
> > > temp1:       +30.0°C  (high = -56.0°C, hyst = -17.0°C)  ALARM  sensor =
> > > thermistor
> > > temp2:       +41.5°C  (high = +80.0°C, hyst = +75.0°C)  sensor = thermistor
> > > temp3:        -3.0°C  (high = +80.0°C, hyst = +75.0°C)  sensor = thermistor
> > > cpu0_vid:   +0.000 V
> > > 
> > Actual values look mostly good, but the limits are really off.
> > Difficult to say, though, if those are wrong readings or just badly configured.
> > 
> > Jean, any idea/thoughts ?
> 
> Only a register dump would say from sure. But given that the alarm
> flags all match out-of-limit readings, I would be inclined to say these
> are only misonconfigured limits. Tobias could set proper limits and
> check that all the alarms go away.
> 
> The only inconsistency is with temp3... the reading is out of limit but
> there is no alarm flag. You may want to double check the datasheet to
> ensure that the driver reads both the temp3 value and its alarm flag
> from the right register/bit.
> 
I checked the specification. Registers are the same for both chips,
unless I am missing something. No idea what is going on.

Guenter