[lm-sensors] sensors.conf for supermicro x7dbe+

[lm-sensors] sensors.conf for supermicro x7dbe+

[Noah Romer]

Support at SuperMicro provided a list of offsets for the sensors on the 
x7dbe+, but they don't look like a complete match with the settings in 
sensors.conf from lm_sensors 2.10.1 . If I can get these to play nice, the 
only other thing I'll need is a released kernel with the w83793 driver in it.

Bus Type = SMBus
One W83793G, One W83792D

Windbond W83793G, Slave Address=0x2f (0x5E in 8-Bit format)
==============================Fan1 Fan Speed, Offset 0x23, 0x24
Fan2 Fan Speed, Offset 0x25, 0x26
Fan3 Fan Speed, Offset 0x27, 0x28
Fan4 Fan Speed, Offset 0x29, 0x2a
Fan5 Fan Speed, Offset 0x2b, 0x2c
Fan6 Fan Speed, Offset 0x2d, 0x2e
Fan7/CPU1 Fan Speed, Offset 0x2f, 0x30
Fan8/CPU2 Fan Speed, Offset 0x31, 0x32
CPU1 Core Voltage, Offset 0x10          
CPU2 Core Voltage, Offset 0x11
-12V Voltage, Offset 0x14
+1.5V Voltage, Offset 0x15
+3.3V Voltage, Offset 0x16
+12V Voltage, Offset 0x17
+5V Voltage, Offset 0x18
5Vsb Voltage, Offset 0x19
Battery Voltage, Offset 0x1a
CPU1 CoreA (PECI Agent1) Temperature, Offset 0x1c
CPU1 CoreB (PECI Agent2) Temperature, Offset 0x1d
CPU2 CoreA (PECI Agent3) Temperature, Offset 0x1e
CPU1 CoreB (PECI Agent4) Temperature, Offset 0x1f
System Temperature, Offset 0x20


Windbond W83792D, Slave Address=0x2e (0x5C in 8-Bit format)
==============================Chassis Intrusion, Offset 0xaa, BitMask 0x20


Power Supply Failure, GP11

-- 
Noah Romer <nromer at arcmailtech.com>
GPG Key Fingerprint: 944E C2F8 EBEC BCC7 EEA9  9381 1B21 0098 CC8D 45BA

[lm-sensors] sensors.conf for supermicro x7dbe+

[Rudolf Marek]

Hi,

> Support at SuperMicro provided a list of offsets for the sensors on the 
> x7dbe+, but they don't look like a complete match with the settings in 
> sensors.conf from lm_sensors 2.10.1 . If I can get these to play nice, the 
> only other thing I'll need is a released kernel with the w83793 driver in it.

You may use the 2.6.19. Or a patch for earlier versions. Maybe it will apply to
previus recent kernels too. Can you use 2.6.19 or need any earlier kernels?

> Bus Type = SMBus
> One W83793G, One W83792D
> 
> Windbond W83793G, Slave Address=0x2f (0x5E in 8-Bit format)
> ==============================> Fan1 Fan Speed, Offset 0x23, 0x24
> Fan2 Fan Speed, Offset 0x25, 0x26
> Fan3 Fan Speed, Offset 0x27, 0x28
> Fan4 Fan Speed, Offset 0x29, 0x2a
> Fan5 Fan Speed, Offset 0x2b, 0x2c
> Fan6 Fan Speed, Offset 0x2d, 0x2e
> Fan7/CPU1 Fan Speed, Offset 0x2f, 0x30
> Fan8/CPU2 Fan Speed, Offset 0x31, 0x32
> CPU1 Core Voltage, Offset 0x10          
> CPU2 Core Voltage, Offset 0x11
> -12V Voltage, Offset 0x14
> +1.5V Voltage, Offset 0x15
> +3.3V Voltage, Offset 0x16
> +12V Voltage, Offset 0x17
> +5V Voltage, Offset 0x18
> 5Vsb Voltage, Offset 0x19
> Battery Voltage, Offset 0x1a
> CPU1 CoreA (PECI Agent1) Temperature, Offset 0x1c
> CPU1 CoreB (PECI Agent2) Temperature, Offset 0x1d
> CPU2 CoreA (PECI Agent3) Temperature, Offset 0x1e
> CPU1 CoreB (PECI Agent4) Temperature, Offset 0x1f
> System Temperature, Offset 0x20

I think even the labels should be correct.You will just need to change the label
of the temps maybe. Please get the w83793 driver and run sensors command to see
what we get.


> Windbond W83792D, Slave Address=0x2e (0x5C in 8-Bit format)
> ==============================> Chassis Intrusion, Offset 0xaa, BitMask 0x20

Wow I dont believe this. I think it must be wrong... Why to put there the other
chip just for this... Please can you run there a sensors-detect and post it here?
> Power Supply Failure, GP11

Thanks,

Rudolf

[lm-sensors] sensors.conf for supermicro x7dbe+

[Noah Romer]

On Friday 05 January 2007 14:19, Rudolf Marek <r.marek at assembler.cz> wrote:
> You may use the 2.6.19. Or a patch for earlier versions. Maybe it will apply
> to previus recent kernels too. Can you use 2.6.19 or need any earlier
> kernels? 

Had to go with 2.6.20-rc3. 2.6.19 doesn't have the w83793 driver in it. The 
output of sensors looks good for the w83793, just need to add a few "ignores" 
to sensors.conf for the bits of hardware that don't exist in this chasis. The 
output for the w83627hf that sensors-detect finds looks off, though (see 
below). SuperMicro didn't mention anything about it.

[offset values snipped] 

> > Windbond W83792D, Slave Address=0x2e (0x5C in 8-Bit format)
> > ==============================> > Chassis Intrusion, Offset 0xaa, BitMask 0x20
> 
> Wow I dont believe this. I think it must be wrong... Why to put there the 
> other chip just for this... Please can you run there a sensors-detect and
> post it here? 
 
Here we are. No W83792D found.

	Probing for `Winbond W83792D'...                            No
[snip]
	Now follows a summary of the probes I have just done.
	Just press ENTER to continue: 
	
	Driver `w83781d' (should be inserted):
	  Detects correctly:
	  * Bus `SMBus I801 adapter at 1100'
	    Busdriver `i2c-i801', I2C address 0x2d
	    Chip `w83627hf' (confidence: 6)
	  * Bus `SMBus I801 adapter at 1100'
	    Busdriver `i2c-i801', I2C address 0x48
	    Chip `w83627hf subclient' (confidence: 6)
	  * Bus `SMBus I801 adapter at 1100'
	    Busdriver `i2c-i801', I2C address 0x49
	    Chip `w83627hf subclient' (confidence: 6)
	  * ISA bus address 0x0290 (Busdriver `i2c-isa')
	    Chip `Winbond W83627HF' (confidence: 8)
	
	Driver `w83793' (should be inserted):
	  Detects correctly:
	  * Bus `SMBus I801 adapter at 1100'
	    Busdriver `i2c-i801', I2C address 0x2f
	    Chip `Winbond W83793R/G' (confidence: 8)
	
	Driver `w83627hf' (should be inserted):
	  Detects correctly:
	  * ISA bus address 0x0295 (Busdriver `i2c-isa')
	    Chip `Winbond W83627HF Super IO Sensors' (confidence: 9)


`sensors` says:

w83627hf-i2c-0-2d
Adapter: SMBus I801 adapter at 1100
in0:       +4.08 V  (min =  +0.00 V, max =  +4.08 V)              
in1:       +4.08 V  (min =  +0.00 V, max =  +4.08 V)              
in2:       +4.08 V  (min =  +2.82 V, max =  +3.79 V)       ALARM  
in3:       +3.09 V  (min =  +2.82 V, max =  +3.12 V)              
in4:       +3.04 V  (min =  +2.85 V, max =  +3.47 V)              
in5:       +3.09 V  (min =  +0.34 V, max =  +0.78 V)       ALARM  
in6:       +3.09 V  (min =  +0.78 V, max =  +0.94 V)       ALARM  
in7:       +3.38 V  (min =  +2.82 V, max =  +3.12 V)       ALARM  
in8:       +0.06 V  (min =  +2.40 V, max =  +3.60 V)       ALARM  
fan1:        0 RPM  (min =    0 RPM, div = 2)                     
fan2:        0 RPM  (min =    0 RPM, div = 2)                     
fan3:        0 RPM  (min =    0 RPM, div = 2)                     
temp1:       -48?C  (high =    -1?C, hyst =    -3?C)   sensor = thermistor           
temp2:     -48.0?C  (high =   +80?C, hyst =   +75?C)   sensor = thermistor           
temp3:     -48.0?C  (high =   +80?C, hyst =   +75?C)   sensor = thermistor           
vid:      +0.000 V  (VRM Version 10.0)
alarms:   
beep_enable:
          Sound alarm enabled

w83627hf-isa-0290
Adapter: ISA adapter
in0:       +4.08 V  (min =  +0.00 V, max =  +4.08 V)              
in1:       +4.08 V  (min =  +0.00 V, max =  +4.08 V)              
in2:       +4.08 V  (min =  +2.82 V, max =  +3.79 V)       ALARM  
in3:       +3.04 V  (min =  +2.82 V, max =  +3.12 V)              
in4:       +3.04 V  (min =  +2.85 V, max =  +3.47 V)              
in5:       +3.09 V  (min =  +0.34 V, max =  +0.78 V)       ALARM  
in6:       +3.07 V  (min =  +0.78 V, max =  +0.94 V)       ALARM  
in7:       +3.39 V  (min =  +2.82 V, max =  +3.12 V)       ALARM  
in8:       +0.06 V  (min =  +2.40 V, max =  +3.60 V)       ALARM  
fan1:        0 RPM  (min =    0 RPM, div = 2)                     
fan2:        0 RPM  (min =    0 RPM, div = 2)                     
fan3:        0 RPM  (min =    0 RPM, div = 2)                     
temp1:       -48?C  (high =    -1?C, hyst =    -3?C)   sensor = thermistor           
temp2:     -48.0?C  (high =   +80?C, hyst =   +75?C)   sensor = thermistor           
temp3:     -48.0?C  (high =   +80?C, hyst =   +75?C)   sensor = thermistor           
vid:      +0.000 V  (VRM Version 10.0)
alarms:   
beep_enable:
          Sound alarm enabled

w83793-i2c-0-2f
Adapter: SMBus I801 adapter at 1100
VCoreA:    +1.22 V  (min =  +0.92 V, max =  +1.49 V)   
VCoreB:    +1.26 V  (min =  +0.92 V, max =  +1.49 V)   
-12V:      +1.49 V  (min =  +1.34 V, max =  +1.65 V)   
+1.5V:     +3.30 V  (min =  +2.96 V, max =  +3.63 V)   
+3.3V:    +11.90 V  (min = +10.75 V, max = +13.25 V)   
+12V:      +4.97 V  (min =  +4.49 V, max =  +5.50 V)   
+5V:       +4.92 V  (min =  +4.49 V, max =  +5.50 V)   
5VSB:      +3.25 V  (min =  +2.99 V, max =  +3.66 V)   
fan1:        0 RPM  (min =  712 RPM)                   ALARM
fan2:     4981 RPM  (min =  712 RPM)                   
fan3:     5192 RPM  (min =  712 RPM)                   
fan4:     4821 RPM  (min =  712 RPM)                   
fan5:        0 RPM  (min =  712 RPM)                   ALARM
fan6:        0 RPM  (min =  712 RPM)                   ALARM
fan7:        0 RPM  (min =  712 RPM)                   ALARM
fan8:        0 RPM  (min =  712 RPM)                   ALARM
fan9:        0 RPM  (min =  712 RPM)                   ALARM
fan10:       0 RPM  (min =  712 RPM)                   ALARM
CPU1 Temp: +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
CPU2 Temp: +23.8?C  (high = +80.0?C, hyst = +75.0?C)   
temp3:     +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
temp4:     +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
temp5:       +28?C  (high =   +50?C, hyst =   +45?C)   
temp6:       +92?C  (high =  +100?C, hyst =   +95?C)   
cpu0_vid: +1.075 V  (VRM Version 10.0)
cpu1_vid: +0.000 V  (VRM Version 10.0)

Thanks,
Noah
-- 
Noah Romer <nromer at arcmailtech.com>
GPG Key Fingerprint: 944E C2F8 EBEC BCC7 EEA9  9381 1B21 0098 CC8D 45BA

[lm-sensors] sensors.conf for supermicro x7dbe+

[Rudolf Marek]

> Had to go with 2.6.20-rc3. 2.6.19 doesn't have the w83793 driver in it. The 
> output of sensors looks good for the w83793, just need to add a few "ignores" 
> to sensors.conf for the bits of hardware that don't exist in this chasis. The 
> output for the w83627hf that sensors-detect finds looks off, though (see 
> below). SuperMicro didn't mention anything about it.

They did the mistake. It should be w83627hf instead of W83792D.
It should contain the power failure on some GPIO pin, plus chassis intrusion.

If you don't need this. Do not load the hf driver. It is twice there because it
can be accessed via i2c or via isa.

CPU1 Core Voltage, Offset 0x10  -> in0
CPU2 Core Voltage, Offset 0x11  -> in1
-12V Voltage, Offset 0x14 -> in3
+1.5V Voltage, Offset 0x15 -> in4
+3.3V Voltage, Offset 0x16 -> in5
+12V Voltage, Offset 0x17 -> in6
+5V Voltage, Offset 0x18 ->in7
5Vsb Voltage, Offset 0x19 -> in8
Battery Voltage, Offset 0x1a -> in9

Our mapping:
       /*Read, High, Low */

        {0x10, 0x60, 0x61},     /* Vcore A      */

        {0x11, 0x62, 0x63},     /* Vcore B      */

        {0x12, 0x64, 0x65},     /* Vtt          */

        {0x14, 0x6a, 0x6b},     /* VSEN1        */

        {0x15, 0x6c, 0x6d},     /* VSEN2        */

        {0x16, 0x6e, 0x6f},     /* +3VSEN       */

        {0x17, 0x70, 0x71},     /* +12VSEN      */

        {0x18, 0x72, 0x73},     /* 5VDD         */

        {0x19, 0x74, 0x75},     /* 5VSB         */

        {0x1a, 0x76, 0x77},     /* VBAT         */

Your labels should be:

   label in0 "VCoreA"
    label in1 "VCoreB"
    label in2 "Vtt - Unknown"
    label in3 "-12V"
    label in4 "+1.5V"
    label in5 "+3.3V"
    label in6 "+12V"
    label in7 "+5V"
    label in8 "5VSB"
    label in9 "VBAT"

In your dump you have it off by one, please fix the configuration file and post
the sensors command again. You will need to ask for the -12V formula to
Supermicro (or try some from sensors.conf)

> fan10:       0 RPM  (min =  712 RPM)                   ALARM
> CPU1 Temp: +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
> CPU2 Temp: +23.8?C  (high = +80.0?C, hyst = +75.0?C)   
> temp3:     +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
> temp4:     +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
> temp5:       +28?C  (high =   +50?C, hyst =   +45?C)   
> temp6:       +92?C  (high =  +100?C, hyst =   +95?C)   

I'm suspecting that this temperature channel is disabled. Please can you check
the temp6_mode file?

The path to file is something like this, please fix it.

cat /sys/class/hwmon/hwmon?/device/temp6_mode
> cpu0_vid: +1.075 V  (VRM Version 10.0)
> cpu1_vid: +0.000 V  (VRM Version 10.0)

This is also bit strange, maybe it is not connected.

Rudolf

[lm-sensors] sensors.conf for supermicro x7dbe+

[Rudolf Marek]

Hi all,

One more thing.

> w83793-i2c-0-2f
> Adapter: SMBus I801 adapter at 1100
> VCoreA:    +1.22 V  (min =  +0.92 V, max =  +1.49 V)   
> VCoreB:    +1.26 V  (min =  +0.92 V, max =  +1.49 V)   
> -12V:      +1.49 V  (min =  +1.34 V, max =  +1.65 V)   
> +1.5V:     +3.30 V  (min =  +2.96 V, max =  +3.63 V)   
> +3.3V:    +11.90 V  (min = +10.75 V, max = +13.25 V)   
> +12V:      +4.97 V  (min =  +4.49 V, max =  +5.50 V)   
> +5V:       +4.92 V  (min =  +4.49 V, max =  +5.50 V)   
> 5VSB:      +3.25 V  (min =  +2.99 V, max =  +3.66 V)   

Why has your board 5VSB connected to 3.3V? The chip specs say it needs this to
be 5V? Maybe Supermicro did this wrong??? (I checked different board and it has
there 5V)

Please can you ask them if pin7 5VSB is connected to 5VSB?

Thanks,

Rudolf

[lm-sensors] sensors.conf for supermicro x7dbe+

[JGong at winbond.com]

Hi, Rudolf, Noah

>> w83793-i2c-0-2f
>> Adapter: SMBus I801 adapter at 1100
>> VCoreA:    +1.22 V  (min =  +0.92 V, max =  +1.49 V)   
>> VCoreB:    +1.26 V  (min =  +0.92 V, max =  +1.49 V)   
>> -12V:      +1.49 V  (min =  +1.34 V, max =  +1.65 V)   
>> +1.5V:     +3.30 V  (min =  +2.96 V, max =  +3.63 V)   
>> +3.3V:    +11.90 V  (min = +10.75 V, max = +13.25 V)   
>> +12V:      +4.97 V  (min =  +4.49 V, max =  +5.50 V)   
>> +5V:       +4.92 V  (min =  +4.49 V, max =  +5.50 V)   
>> 5VSB:      +3.25 V  (min =  +2.99 V, max =  +3.66 V)   


>Why has your board 5VSB connected to 3.3V? The chip specs say it needs
this >to be 5V? Maybe Supermicro did this wrong??? (I checked different
board and >it has there 5V)
>Please can you ask them if pin7 5VSB is connected to 5VSB?

According to the voltage readings, I think Noah's sensor.conf may be
something wrong like this:

label in0 "VcoreA"
label in1 "VcoreB"
label in3 "-12V"
label in5 "+1.5V"
label in6 "+3.3V"
label in7 "+12V"
label in8  "+5V"
label in9  "5VSB"

So the output seems a little messy, try this one, Noah.

    label in0 "VCoreA"
    label in1 "VCoreB"
    label in2 "Vtt"
    label in3 "-12V"
    label in4 "+1.5V"
    label in5 "+3.3V"
    label in6 "+12V"
    label in7 "+5V"
    label in8 "5VSB"
    label in9 "VBAT"

>> cpu0_vid: +1.075 V (VRM Version 10.0)
>> cpu1_vid: +0.000 V (VRM Version 10.0)
>This is also bit strange, maybe it is not connected.

Well, Rudolf, usually the motherboard manufacturer uses the w83793's PIN
49-56 for fan usage, and in fact we can see the register of VIDB latch
value to judge whether the VIDB has been connected. On my w83793 board,
the VIDB has not been connected...

Best regards,
PI14
Gong Jun


=============================================The privileged confidential information contained in this email is intended for use only by the addressees as indicated by the original sender of this email. If you are not the addressee indicated in this email or are not responsible for delivery of the email to such  a person, please kindly reply to the sender indicating this fact and delete all copies of it from your computer and network server immediately. Your cooperation is highly appreciated. It is advised that any unauthorized use of confidential information of Winbond is strictly prohibited; and any information in this email irrelevant to the official business of Winbond shall be deemed as neither given nor endorsed by Winbond.

[lm-sensors] sensors.conf for supermicro x7dbe+

[Noah Romer]

On Sunday 07 January 2007 03:08, Rudolf Marek <r.marek at assembler.cz> wrote:
> They did the mistake. It should be w83627hf instead of W83792D.
> It should contain the power failure on some GPIO pin, plus chassis 
> intrusion. 
> 
> If you don't need this. Do not load the hf driver. It is twice there because 
> it can be accessed via i2c or via isa.

Well, there's no module loaded for it and commenting out the w83627hf line 
in /etc/sysconfig/lm_sensors has no impact. I'll have to check if it's 
compiled into the kernel. What sort of power failure is the sensor detecting? 
These boxes usually have redundant power supplies, and it would be cool if we 
could pick up a failure off via lm_sensors.
 
> CPU1 Core Voltage, Offset 0x10  -> in0
> CPU2 Core Voltage, Offset 0x11  -> in1
> -12V Voltage, Offset 0x14 -> in3
> +1.5V Voltage, Offset 0x15 -> in4
> +3.3V Voltage, Offset 0x16 -> in5
> +12V Voltage, Offset 0x17 -> in6
> +5V Voltage, Offset 0x18 ->in7
> 5Vsb Voltage, Offset 0x19 -> in8
> Battery Voltage, Offset 0x1a -> in9

Ah. I was reading the 0x11 - > 0x14 jump in offsets as meaning that I needed 
to skip in2 and in3. Fixed that, the output looks more reasonable:

w83793-i2c-0-2f
Adapter: SMBus I801 adapter at 1100
VCoreA:    +1.22 V  (min =  +0.92 V, max =  +1.49 V)   
VCoreB:    +1.26 V  (min =  +0.92 V, max =  +1.49 V)   
Vtt:       +1.18 V  (min =  +1.08 V, max =  +1.33 V)   
-12V:      +0.50 V  (min =  +0.38 V, max =  +0.69 V)   
+1.5V:     +1.49 V  (min =  +1.34 V, max =  +1.65 V)   
+3.3V:     +3.30 V  (min =  +2.96 V, max =  +3.63 V)   
+12V:     +11.90 V  (min = +10.75 V, max = +13.25 V)   
+5V:       +4.97 V  (min =  +4.49 V, max =  +5.50 V)   
5VSB:      +4.92 V  (min =  +4.49 V, max =  +5.50 V)   
VBAT:      +3.25 V  (min =  +2.99 V, max =  +3.66 V)   
fan2:     5000 RPM  (min =  712 RPM)                   
fan3:     5192 RPM  (min =  712 RPM)                   
fan4:     4804 RPM  (min =  712 RPM)                   
CPU1 Temp: +25.2?C  (high = +80.0?C, hyst = +75.0?C)   
CPU2 Temp: +24.2?C  (high = +80.0?C, hyst = +75.0?C)   
temp3:     +26.2?C  (high = +80.0?C, hyst = +75.0?C)   
temp4:     +24.8?C  (high = +80.0?C, hyst = +75.0?C)   
temp5:       +28?C  (high =   +50?C, hyst =   +45?C)   
temp6:       +92?C  (high =  +100?C, hyst =   +95?C)   
cpu0_vid: +1.075 V  (VRM Version 10.0)
cpu1_vid: +0.000 V  (VRM Version 10.0)

> > temp6:       +92?C  (high =  +100?C, hyst =   +95?C)   
> 
> I'm suspecting that this temperature channel is disabled. Please can you 
> check the temp6_mode file?
> 
> The path to file is something like this, please fix it.
> 
> cat /sys/class/hwmon/hwmon?/device/temp6_mode

There is no temp6_mode file. Only three files in /sys/class/hwmon:
/sys/class/hwmon/hwmon2/uevent
/sys/class/hwmon/hwmon1/uevent
/sys/class/hwmon/hwmon0/uevent

There's group of temp6 files in the the devices 
heirarchy: /sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp6*
-- 
Noah Romer <nromer at arcmailtech.com>
GPG Key Fingerprint: 944E C2F8 EBEC BCC7 EEA9  9381 1B21 0098 CC8D 45BA

[lm-sensors] sensors.conf for supermicro x7dbe+

[JGong at winbond.com]

Hi, Noah

>There is no temp6_mode file. Only three files in /sys/class/hwmon:
>/sys/class/hwmon/hwmon2/uevent
>/sys/class/hwmon/hwmon1/uevent
>/sys/class/hwmon/hwmon0/uevent
>There's group of temp6 files in the the devices 
>heirarchy: /sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp6*

En, you can try this one to see the temp6 mode:

cat /sys/bus/i2c/drivers/w83793/0-002f/temp6_type

Best regards,
Gong Jun



=============================================The privileged confidential information contained in this email is intended for use only by the addressees as indicated by the original sender of this email. If you are not the addressee indicated in this email or are not responsible for delivery of the email to such  a person, please kindly reply to the sender indicating this fact and delete all copies of it from your computer and network server immediately. Your cooperation is highly appreciated. It is advised that any unauthorized use of confidential information of Winbond is strictly prohibited; and any information in this email irrelevant to the official business of Winbond shall be deemed as neither given nor endorsed by Winbond.

[lm-sensors] sensors.conf for supermicro x7dbe+

[Noah Romer]

On Monday 08 January 2007 22:18, JGong at winbond.com wrote:
> 
> En, you can try this one to see the temp6 mode:
> 
> cat /sys/bus/i2c/drivers/w83793/0-002f/temp6_type

'0'.
-- 
Noah Romer <nromer at arcmailtech.com>
GPG Key Fingerprint: 944E C2F8 EBEC BCC7 EEA9  9381 1B21 0098 CC8D 45BA

[lm-sensors] sensors.conf for supermicro x7dbe+

[JGong at winbond.com]

Hi, Noah 
>> En, you can try this one to see the temp6 mode:
>> 
>> cat /sys/bus/i2c/drivers/w83793/0-002f/temp6_type

>'0'

Ok, that is to say, the temp6 channel is disabled, and you can ignore
the temp6 readings by adding one line "ignore temp6" into your
sensors.conf.

By now, I am sorry that the w83793 driver has done nothing with the
disabled temperature channel situation. :(

Best regards,
Gong Jun



=============================================The privileged confidential information contained in this email is intended for use only by the addressees as indicated by the original sender of this email. If you are not the addressee indicated in this email or are not responsible for delivery of the email to such  a person, please kindly reply to the sender indicating this fact and delete all copies of it from your computer and network server immediately. Your cooperation is highly appreciated. It is advised that any unauthorized use of confidential information of Winbond is strictly prohibited; and any information in this email irrelevant to the official business of Winbond shall be deemed as neither given nor endorsed by Winbond.

[lm-sensors] sensors.conf for supermicro x7dbe+

[Jean Delvare]

Hi Gong,

On Wed, 10 Jan 2007 16:59:05 +0800, JGong at winbond.com wrote:
> 
> Hi, Noah 
> >> En, you can try this one to see the temp6 mode:
> >> 
> >> cat /sys/bus/i2c/drivers/w83793/0-002f/temp6_type
> 
> >'0'
> 
> Ok, that is to say, the temp6 channel is disabled, and you can ignore
> the temp6 readings by adding one line "ignore temp6" into your
> sensors.conf.
> 
> By now, I am sorry that the w83793 driver has done nothing with the
> disabled temperature channel situation. :(

I agree, it's not handled properly; let's make it better. I see 3
possibilities:

1* We display the sensor type in "sensors". This is what we do for the
IT87xxF devices. When the sensor type is "disabled", the user should
realize that the reading is meaningless.

2* We change sensors to not even display tempN if that sensor is
disabled. This is probably even more user-friendly.

3* We change the w83793 driver to not present the temperature value
when the sensor is disabled. A bit more difficult, but doable. There's
something of this kind done in the f71805f hardware monitoring driver.

A combination of 2 and 3 sounds best to me. Gong, what do you think?
Can you please propose patches?

Thanks,
-- 
Jean Delvare

[lm-sensors] sensors.conf for supermicro x7dbe+

[Rudolf Marek]

> I agree, it's not handled properly; let's make it better. I see 3
> possibilities:
> 

Yes I proposed similar, but nobody noticed that :)

> 1* We display the sensor type in "sensors". This is what we do for the
> IT87xxF devices. When the sensor type is "disabled", the user should
> realize that the reading is meaningless.
> 
> 2* We change sensors to not even display tempN if that sensor is
> disabled. This is probably even more user-friendly.

You mean during init phase?

> 3* We change the w83793 driver to not present the temperature value
> when the sensor is disabled. A bit more difficult, but doable. There's
> something of this kind done in the f71805f hardware monitoring driver.

Yes but we need to keep the control file for that.

> A combination of 2 and 3 sounds best to me. Gong, what do you think?

I like it. Do not create tempX_input files for disabled channels, but create 
channel mode files. When enabled the tempX_input will appear.

Rudolf

[lm-sensors] sensors.conf for supermicro x7dbe+

[Jean Delvare]

Hi Rudolf,

On Wed, 10 Jan 2007 23:34:50 +0100, Rudolf Marek wrote:
> > 1* We display the sensor type in "sensors". This is what we do for the
> > IT87xxF devices. When the sensor type is "disabled", the user should
> > realize that the reading is meaningless.
> > 
> > 2* We change sensors to not even display tempN if that sensor is
> > disabled. This is probably even more user-friendly.
> 
> You mean during init phase?

I mean that the sensors tool would first read temp1_type, if it's 0
skip temp1 altogether, else read and display all temp1 attributes as
before. Same for temp2 etc.

> > 3* We change the w83793 driver to not present the temperature value
> > when the sensor is disabled. A bit more difficult, but doable. There's
> > something of this kind done in the f71805f hardware monitoring driver.
> 
> Yes but we need to keep the control file for that.

Depends on how we implement it. We may decide that the BIOS must have
set the types properly, make tempN_type read-only (the f71805f driver
does that), and simply not create any file for disabled temperature
channels. Or indeed we can let the user override what the BIOS has set,
in which case we must create and delete the tempN_input files
dynamically (or change their permissions dynamically, that works too).
This second option is more complex, but seems to have your preference.
As for me, I'm fine with both, as I'm not the one writing the code ;)

-- 
Jean Delvare

[lm-sensors] sensors.conf for supermicro x7dbe+

[Rudolf Marek]

> What sort of power failure is the sensor detecting? 
> These boxes usually have redundant power supplies, and it would be cool if we 
> could pick up a failure off via lm_sensors.

Hmm I dont know. Maybe it is 1 if power failed. They should provide 
accurate/fixed info, so we know if it is really GP11 if so, I can tell you how 
you can read this bit though some script. (last time they answered wrong so 
please ask them once again. What chip, what bus, what GPIO what register)

thanks,
Rudolf