[lm-sensors] Supermicro X7DWA-N All fans controlled by pwm3?

[lm-sensors] Supermicro X7DWA-N All fans controlled by pwm3?

[Andrew Lyon]

Hi,

I have a system with Supermicro X7DWA-N motherboard, according to the
manual it has "Winbond W83627HF w/Hardware Monitor support: W83793" so
I've loaded both the w83627hf and w83793 modules and several sensors
and fans are detected, however when I try to control fan speed I find
that the only pwm output which has any effect is pwm3 and it seems to
control all of the fans at once, also pwm1 is locked at 112:

ubermicro 0-002f # pwd
/sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f
ubermicro 0-002f # lspci | grep  1f.3
00:1f.3 SMBus: Intel Corporation 631xESB/632xESB/3100 Chipset SMBus
Controller (rev 09)
ubermicro 0-002f # ls pwm?
pwm1  pwm2  pwm3  pwm4  pwm5  pwm6  pwm7  pwm8
ubermicro 0-002f # cat pwm?
112
0
160
0
0
0
0
0


ubermicro 0-002f # echo 0 >pwm3
ubermicro 0-002f # sensors | grep fan
fan1:       1564 RPM  (min =  712 RPM)
fan2:       1652 RPM  (min =  712 RPM)
fan3:       1636 RPM  (min =  712 RPM)
fan4:       1638 RPM  (min =  712 RPM)
fan6:       2351 RPM  (min =  712 RPM)
fan7:       1622 RPM  (min =  712 RPM)
fan8:       1605 RPM  (min =  712 RPM)

ubermicro 0-002f # echo 140 >pwm3
ubermicro 0-002f # sensors | grep fan
fan1:       1973 RPM  (min =  712 RPM)
fan2:       2089 RPM  (min =  712 RPM)
fan3:       2064 RPM  (min =  712 RPM)
fan4:       2051 RPM  (min =  712 RPM)
fan6:       2755 RPM  (min =  712 RPM)
fan7:       2048 RPM  (min =  712 RPM)
fan8:       2002 RPM  (min =  712 RPM)


Is it possible to gain control of the fans individually?

Andy

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

Re: [lm-sensors] Supermicro X7DWA-N All fans controlled by pwm3?

[Jean Delvare]

Hi Andrew,

On Tue, 14 Sep 2010 09:52:12 +0100, Andrew Lyon wrote:
> I have a system with Supermicro X7DWA-N motherboard, according to the
> manual it has "Winbond W83627HF w/Hardware Monitor support: W83793" so
> I've loaded both the w83627hf and w83793 modules and several sensors
> and fans are detected, however when I try to control fan speed I find
> that the only pwm output which has any effect is pwm3 and it seems to
> control all of the fans at once, also pwm1 is locked at 112:

What's the value of pwm1_enable? Could be that this PWM output is not
currently in manual control mode.

Which kernel are you running?

The w83793 driver is still marked experimental, there could be a bug in
it.

> 
> ubermicro 0-002f # pwd
> /sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f
> ubermicro 0-002f # lspci | grep  1f.3
> 00:1f.3 SMBus: Intel Corporation 631xESB/632xESB/3100 Chipset SMBus
> Controller (rev 09)
> ubermicro 0-002f # ls pwm?
> pwm1  pwm2  pwm3  pwm4  pwm5  pwm6  pwm7  pwm8
> ubermicro 0-002f # cat pwm?
> 112
> 0
> 160
> 0
> 0
> 0
> 0
> 0
> 
> 
> ubermicro 0-002f # echo 0 >pwm3
> ubermicro 0-002f # sensors | grep fan
> fan1:       1564 RPM  (min =  712 RPM)
> fan2:       1652 RPM  (min =  712 RPM)
> fan3:       1636 RPM  (min =  712 RPM)
> fan4:       1638 RPM  (min =  712 RPM)
> fan6:       2351 RPM  (min =  712 RPM)
> fan7:       1622 RPM  (min =  712 RPM)
> fan8:       1605 RPM  (min =  712 RPM)
> 
> ubermicro 0-002f # echo 140 >pwm3
> ubermicro 0-002f # sensors | grep fan
> fan1:       1973 RPM  (min =  712 RPM)
> fan2:       2089 RPM  (min =  712 RPM)
> fan3:       2064 RPM  (min =  712 RPM)
> fan4:       2051 RPM  (min =  712 RPM)
> fan6:       2755 RPM  (min =  712 RPM)
> fan7:       2048 RPM  (min =  712 RPM)
> fan8:       2002 RPM  (min =  712 RPM)
> 
> 
> Is it possible to gain control of the fans individually?

Very unlikely. PWM signal routing to fans is a hardware thing, it's
been decided by your motherboard vendor and you have to live with it.

Here I have an Asus server board with 8 fan inputs, 2 PWM output
controls, one controls the 2 CPU fans, and the other one controls the 6
case fans. So it's similar to your board, only slightly better because
CPU fans have their own control.

It is possible (but again unlikely) that the hardware vendor
implemented a user-controllable PWM routing. In that case it would
certainly be advertised by said vendor, and exposed in the BIOS or by
some custom tool. But I don't think I've ever seen this in practice.
And I don't think it would make too much sense (it would be cheaper to
route PWM outputs individually and group them at the software level if
needed.)

-- 
Jean Delvare
http://khali.linux-fr.org/wishlist.html

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

Re: [lm-sensors] Supermicro X7DWA-N All fans controlled by pwm3?

[Andrew Lyon]

On Wed, Sep 15, 2010 at 5:13 PM, Jean Delvare <khali@linux-fr.org> wrote:
> Hi Andrew,
>
> On Tue, 14 Sep 2010 09:52:12 +0100, Andrew Lyon wrote:
>> I have a system with Supermicro X7DWA-N motherboard, according to the
>> manual it has "Winbond W83627HF w/Hardware Monitor support: W83793" so
>> I've loaded both the w83627hf and w83793 modules and several sensors
>> and fans are detected, however when I try to control fan speed I find
>> that the only pwm output which has any effect is pwm3 and it seems to
>> control all of the fans at once, also pwm1 is locked at 112:
>
> What's the value of pwm1_enable? Could be that this PWM output is not
> currently in manual control mode.

The only devices with pwm and enable in the name are:

/sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp1_pwm_enable
/sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp2_pwm_enable
/sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp3_pwm_enable
/sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp4_pwm_enable
/sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f/temp5_pwm_enable


All currently set to 2.

>
> Which kernel are you running?

2.6.34.7

>
> The w83793 driver is still marked experimental, there could be a bug in
> it.
>
>>
>> ubermicro 0-002f # pwd
>> /sys/devices/pci0000:00/0000:00:1f.3/i2c-0/0-002f
>> ubermicro 0-002f # lspci | grep  1f.3
>> 00:1f.3 SMBus: Intel Corporation 631xESB/632xESB/3100 Chipset SMBus
>> Controller (rev 09)
>> ubermicro 0-002f # ls pwm?
>> pwm1  pwm2  pwm3  pwm4  pwm5  pwm6  pwm7  pwm8
>> ubermicro 0-002f # cat pwm?
>> 112
>> 0
>> 160
>> 0
>> 0
>> 0
>> 0
>> 0
>>
>>
>> ubermicro 0-002f # echo 0 >pwm3
>> ubermicro 0-002f # sensors | grep fan
>> fan1:       1564 RPM  (min =  712 RPM)
>> fan2:       1652 RPM  (min =  712 RPM)
>> fan3:       1636 RPM  (min =  712 RPM)
>> fan4:       1638 RPM  (min =  712 RPM)
>> fan6:       2351 RPM  (min =  712 RPM)
>> fan7:       1622 RPM  (min =  712 RPM)
>> fan8:       1605 RPM  (min =  712 RPM)
>>
>> ubermicro 0-002f # echo 140 >pwm3
>> ubermicro 0-002f # sensors | grep fan
>> fan1:       1973 RPM  (min =  712 RPM)
>> fan2:       2089 RPM  (min =  712 RPM)
>> fan3:       2064 RPM  (min =  712 RPM)
>> fan4:       2051 RPM  (min =  712 RPM)
>> fan6:       2755 RPM  (min =  712 RPM)
>> fan7:       2048 RPM  (min =  712 RPM)
>> fan8:       2002 RPM  (min =  712 RPM)
>>
>>
>> Is it possible to gain control of the fans individually?
>
> Very unlikely. PWM signal routing to fans is a hardware thing, it's
> been decided by your motherboard vendor and you have to live with it.
>
> Here I have an Asus server board with 8 fan inputs, 2 PWM output
> controls, one controls the 2 CPU fans, and the other one controls the 6
> case fans. So it's similar to your board, only slightly better because
> CPU fans have their own control.
>
> It is possible (but again unlikely) that the hardware vendor
> implemented a user-controllable PWM routing. In that case it would
> certainly be advertised by said vendor, and exposed in the BIOS or by
> some custom tool. But I don't think I've ever seen this in practice.
> And I don't think it would make too much sense (it would be cheaper to
> route PWM outputs individually and group them at the software level if
> needed.)
>
> --
> Jean Delvare
> http://khali.linux-fr.org/wishlist.html
>

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