Re: [lm-sensors] extreme fan rpms with atk0110

Re: [lm-sensors] extreme fan rpms with atk0110

[Jean Delvare]

Hi Anders,

On Sat, 12 Mar 2011 20:25:16 +0100, Anders Kullenberg wrote:
> I have extremely high fan reading on my ubuntu 10.10 server installation
> (AMD64). Any clues to what could be the reason?
> 
> Asus m2v motherboard
> Lm-sensors 3.1.2 kernel 2.6.35-22
> 
> Sensors output:
> atk0110-acpi-0
> Adapter: ACPI interface
> Vcore Voltage:      +1.14 V  (min =  +0.85 V, max =  +1.60 V)
>  +3.3 Voltage:      +3.31 V  (min =  +2.97 V, max =  +3.63 V)
>  +5 Voltage:        +5.00 V  (min =  +4.50 V, max =  +5.50 V)
>  +12 Voltage:      +11.31 V  (min = +10.20 V, max = +13.80 V)
> CPU FAN Speed:     18000 RPM  (min =  800 RPM)
> CHASSIS FAN Speed: 14835 RPM  (min =  800 RPM)
> CHASSIS2 FAN Speed:   0 RPM  (min =  800 RPM)
> CPU Temperature:    +25.0°C  (high = +60.0°C, crit = +95.0°C)
> MB Temperature:     +35.0°C  (high = +45.0°C, crit = +95.0°C)

Do you see the same crazy values in the BIOS? I guess so.

There can be two reasons:
* You are using special fans with more than 2 poles, so they emit more
  than 2 pulses per revolution. If this is the case, you have to
  correct the values using compute statements in your configuration
  file.
* You are using potentiometers or a so-called "fanbus" to control the
  speed of 3-wire fans. These tend to degrade the tachometer signal to
  a point where monitored values get completely wrong. In this case,
  you have to increase the speed to get valid readings again.

-- 
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] extreme fan rpms with atk0110

[Anders Kullenberg]

Hi,

Thanks for the quick reply.

I checked the BIOS readings. They are normal, even below 800rpm. I have the
asus q-fan enabled in the bios that should control fan speeds. I tried to
disable it, didn't change anything.
It is possible to select ACPI 1.0, 2.0 or 3.0 in the bios, any
recommendations on that?

The fans are not any special stuff as far as I'm aware. They are connected
to the motherboard directly, no potentiometers used in the installation.

I found no chip "atk0110-*" section in /etc/sensors3.conf, what config is
used then, where should I put the compute statements?

BR
Anders

-----Ursprungligt meddelande-----
Från: Jean Delvare [mailto:khali@linux-fr.org] 
Skickat: den 12 mars 2011 21:28
Till: Anders Kullenberg
Kopia: lm-sensors@lm-sensors.org
Ämne: Re: [lm-sensors] extreme fan rpms with atk0110

Hi Anders,

On Sat, 12 Mar 2011 20:25:16 +0100, Anders Kullenberg wrote:
> I have extremely high fan reading on my ubuntu 10.10 server installation
> (AMD64). Any clues to what could be the reason?
> 
> Asus m2v motherboard
> Lm-sensors 3.1.2 kernel 2.6.35-22
> 
> Sensors output:
> atk0110-acpi-0
> Adapter: ACPI interface
> Vcore Voltage:      +1.14 V  (min =  +0.85 V, max =  +1.60 V)
>  +3.3 Voltage:      +3.31 V  (min =  +2.97 V, max =  +3.63 V)
>  +5 Voltage:        +5.00 V  (min =  +4.50 V, max =  +5.50 V)
>  +12 Voltage:      +11.31 V  (min = +10.20 V, max = +13.80 V)
> CPU FAN Speed:     18000 RPM  (min =  800 RPM)
> CHASSIS FAN Speed: 14835 RPM  (min =  800 RPM)
> CHASSIS2 FAN Speed:   0 RPM  (min =  800 RPM)
> CPU Temperature:    +25.0°C  (high = +60.0°C, crit = +95.0°C)
> MB Temperature:     +35.0°C  (high = +45.0°C, crit = +95.0°C)

Do you see the same crazy values in the BIOS? I guess so.

There can be two reasons:
* You are using special fans with more than 2 poles, so they emit more
  than 2 pulses per revolution. If this is the case, you have to
  correct the values using compute statements in your configuration
  file.
* You are using potentiometers or a so-called "fanbus" to control the
  speed of 3-wire fans. These tend to degrade the tachometer signal to
  a point where monitored values get completely wrong. In this case,
  you have to increase the speed to get valid readings again.

-- 
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] extreme fan rpms with atk0110

[Jean Delvare]

Anders,

On Sat, 12 Mar 2011 22:39:42 +0100, Anders Kullenberg wrote:
> Thanks for the quick reply.

Please don't top-post.

> I checked the BIOS readings. They are normal, even below 800rpm. I have the
> asus q-fan enabled in the bios that should control fan speeds. I tried to
> disable it, didn't change anything.
> It is possible to select ACPI 1.0, 2.0 or 3.0 in the bios, any
> recommendations on that?

I'd go for ACPI 2.0 as the best compromise between recentness and
support level, but in the end it really depends on what the BIOS adds
in each case, and whether Linux properly supports that yet or not. But
anyway it probably doesn't matter for the problem at hand.

> The fans are not any special stuff as far as I'm aware. They are connected
> to the motherboard directly, no potentiometers used in the installation.

OK. Maybe you should send your board's DSDT to Luca for analysis then.

> I found no chip "atk0110-*" section in /etc/sensors3.conf, what config is
> used then, where should I put the compute statements?

The ATK0110 usually doesn't need any configuration, as the driver knows
all the scaling factors and labels. But you are free to add a new
configuration section for it if you really have to.

I'm a little surprised as the M2V isn't a recent board, so it should be
already supported, or more users would have reported by now. Did you
look for a BIOS update for your board?

-- 
Jean Delvare

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

Re: [lm-sensors] extreme fan rpms with atk0110

[Luca Tettamanti]

On Sat, Mar 12, 2011 at 11:45 PM, Anders Kullenberg <akullen@swipnet.se> wrote:
> Hi,
>
> Sorry about the top-post, but I don't know what it means. Never used a
> mail-list before and I guess it has to with that.

No, it means that you should not write your answer at the top of the
email. Instead reply inline, addressing each point separately (like
I'm doing right now).
See for example this link:
http://www.dickgaughan.co.uk/usenet/guide/faq08-topp.html

> Should I reply only to your address?

No, and keep the list in CC.

> I have attached what I think is the DSDT, used this method:"sudo cat
> /sys/firmware/acpi/tables/DSDT > /tmp/DSDT" is that correct?

Yes, that's correct. I'll take a look tomorrow.

Luca

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

Re: [lm-sensors] extreme fan rpms with atk0110

[Anders Kullenberg]

Hi Jean,

> Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> 
> Anders,
> 
> On Sat, 12 Mar 2011 22:39:42 +0100, Anders Kullenberg wrote:
> > Thanks for the quick reply.
> 
> Please don't top-post.

Sorry for that, now I know how to do it.

> 
> > I checked the BIOS readings. They are normal, even below 800rpm. I
> have the
> > asus q-fan enabled in the bios that should control fan speeds. I
> tried to
> > disable it, didn't change anything.
> > It is possible to select ACPI 1.0, 2.0 or 3.0 in the bios, any
> > recommendations on that?
> 
> I'd go for ACPI 2.0 as the best compromise between recentness and
> support level, but in the end it really depends on what the BIOS adds
> in each case, and whether Linux properly supports that yet or not. But
> anyway it probably doesn't matter for the problem at hand.

I switched the setting to ACPI 2.0

> 
> > The fans are not any special stuff as far as I'm aware. They are
> connected
> > to the motherboard directly, no potentiometers used in the
> installation.
> 
> OK. Maybe you should send your board's DSDT to Luca for analysis then.

Done.

> 
> > I found no chip "atk0110-*" section in /etc/sensors3.conf, what
> config is
> > used then, where should I put the compute statements?
> 
> The ATK0110 usually doesn't need any configuration, as the driver knows
> all the scaling factors and labels. But you are free to add a new
> configuration section for it if you really have to.
> 
> I'm a little surprised as the M2V isn't a recent board, so it should be
> already supported, or more users would have reported by now. Did you
> look for a BIOS update for your board?

Have the latest bios available from asus homepage.

> 
> --
> Jean Delvare

Some more info. Don't know if it is of use but I have configured munin to
read the sensor output.
Over the last ~24hour both the fan RPM readings has decreased with ~4000 rpm
in a very stable decrease over time.
During the same period the temp readings have been very stable on the same
level.

Thanks
Anders


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

Re: [lm-sensors] extreme fan rpms with atk0110

[Luca Tettamanti]

On Sat, Mar 12, 2011 at 11:54 PM, Luca Tettamanti <kronos.it@gmail.com> wrote:
> On Sat, Mar 12, 2011 at 11:45 PM, Anders Kullenberg <akullen@swipnet.se> wrote:
>> I have attached what I think is the DSDT, used this method:"sudo cat
>> /sys/firmware/acpi/tables/DSDT > /tmp/DSDT" is that correct?
>
> Yes, that's correct. I'll take a look tomorrow.

I don't see anything suspicious in the ACPI code, but the relevant
method (HWF1) scales the value read from the sensor using the divider
(also read from the monitoring chip) and a built-in table.
Can you identify the chip using sensors-detect?
If possible load the native driver (blacklist asus_atk0110 and boot
with acpi_enforce_resources=lax) and compare the fan readings.

Luca

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

Re: [lm-sensors] extreme fan rpms with atk0110

[Jean Delvare]

On Tue, 15 Mar 2011 15:01:24 +0100, Luca Tettamanti wrote:
> On Sat, Mar 12, 2011 at 11:54 PM, Luca Tettamanti <kronos.it@gmail.com> wrote:
> > On Sat, Mar 12, 2011 at 11:45 PM, Anders Kullenberg <akullen@swipnet.se> wrote:
> >> I have attached what I think is the DSDT, used this method:"sudo cat
> >> /sys/firmware/acpi/tables/DSDT > /tmp/DSDT" is that correct?
> >
> > Yes, that's correct. I'll take a look tomorrow.
> 
> I don't see anything suspicious in the ACPI code, but the relevant
> method (HWF1) scales the value read from the sensor using the divider
> (also read from the monitoring chip) and a built-in table.
> Can you identify the chip using sensors-detect?

He did already, before you got Cc'd:

Probing for Super-I/O at 0x2e/0x2f
Trying family `National Semiconductor'...                   No
Trying family `SMSC'...                                     No
Trying family `VIA/Winbond/Nuvoton/Fintek'...               No
Trying family `ITE'...                                      Yes
Found `ITE IT8712F Super IO Sensors'                        Success!
    (address 0xd00, driver `it87')


> If possible load the native driver (blacklist asus_atk0110 and boot
> with acpi_enforce_resources=lax) and compare the fan readings.

Would be interesting, yes. Some IT8712F chips support 16-bit fan speed
readings (without divisor) and others support 8-bit fan speed readings
(with divisor). Maybe the chip and its configuration disagree with what
the DSDT code expects.

-- 
Jean Delvare

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

Re: [lm-sensors] extreme fan rpms with atk0110

[Luca Tettamanti]

On Tue, Mar 15, 2011 at 3:58 PM, Jean Delvare <khali@linux-fr.org> wrote:
> On Tue, 15 Mar 2011 15:01:24 +0100, Luca Tettamanti wrote:
[cut]
> Found `ITE IT8712F Super IO Sensors'                        Success!
>    (address 0xd00, driver `it87')
>
>> If possible load the native driver (blacklist asus_atk0110 and boot
>> with acpi_enforce_resources=lax) and compare the fan readings.
>
> Would be interesting, yes. Some IT8712F chips support 16-bit fan speed
> readings (without divisor) and others support 8-bit fan speed readings
> (with divisor). Maybe the chip and its configuration disagree with what
> the DSDT code expects.

I've cross checked and the DSDT matches the code from it87 for 8 bit readings.

Luca

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

Re: [lm-sensors] extreme fan rpms with atk0110

[Anders Kullenberg]

> Från: Luca Tettamanti [mailto:kronos.it@gmail.com]
> Skickat: den 15 mars 2011 17:01
> Till: Jean Delvare
> Kopia: Anders Kullenberg; LM Sensors
> Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> 
> On Tue, Mar 15, 2011 at 3:58 PM, Jean Delvare <khali@linux-fr.org>
> wrote:
> > On Tue, 15 Mar 2011 15:01:24 +0100, Luca Tettamanti wrote:
> [cut]
> > Found `ITE IT8712F Super IO Sensors'                        Success!
> >    (address 0xd00, driver `it87')
> >
> >> If possible load the native driver (blacklist asus_atk0110 and boot
> >> with acpi_enforce_resources=lax) and compare the fan readings.

Blacklisting is done in "/usr/src/linux-headers-2.6.35-27-server/.config" correct?
In the ACPI drivers section I have:
CONFIG_SENSORS_ATK0110=m
CONFIG_SENSORS_LIS3LV02D=m

Should I comment both those out?

Where do I put the "acpi_enforce_resources=lax"

Thanks


> > Would be interesting, yes. Some IT8712F chips support 16-bit fan
> speed
> > readings (without divisor) and others support 8-bit fan speed
> readings
> > (with divisor). Maybe the chip and its configuration disagree with
> what
> > the DSDT code expects.
> 
> I've cross checked and the DSDT matches the code from it87 for 8 bit
> readings.
> 
> Luca


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

Re: [lm-sensors] extreme fan rpms with atk0110

[Jean Delvare]

Anders,

On Tue, 15 Mar 2011 20:39:12 +0100, Anders Kullenberg wrote:
> > Från: Luca Tettamanti [mailto:kronos.it@gmail.com]
> > Skickat: den 15 mars 2011 17:01
> > Till: Jean Delvare
> > Kopia: Anders Kullenberg; LM Sensors
> > Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> > 
> > On Tue, Mar 15, 2011 at 3:58 PM, Jean Delvare <khali@linux-fr.org>
> > wrote:
> > > On Tue, 15 Mar 2011 15:01:24 +0100, Luca Tettamanti wrote:
> > [cut]
> > > Found `ITE IT8712F Super IO Sensors'                        Success!
> > >    (address 0xd00, driver `it87')
> > >
> > >> If possible load the native driver (blacklist asus_atk0110 and boot
> > >> with acpi_enforce_resources=lax) and compare the fan readings.
> 
> Blacklisting is done in "/usr/src/linux-headers-2.6.35-27-server/.config" correct?
> In the ACPI drivers section I have:
> CONFIG_SENSORS_ATK0110=m
> CONFIG_SENSORS_LIS3LV02D=m
> 
> Should I comment both those out?

No, blacklisting is done with statements in /etc/modprobe.d/*. On my
distribution, I'd add:

blacklist asus_atk0110

to file etc/modprobe.d/99-local.conf. Or you can create a brand new
file in this directory if you want.

> 
> Where do I put the "acpi_enforce_resources=lax"

It has to go on the boot command line, typically you add it when grub
presents the boot options at system start-up. Alternatively, you can
edit /boot/grub/menu.lst or /boot/grub/grub.cfg (depending on your
distribution) manually to add it at the end of the relevant "linux" or
"kernel" line.

-- 
Jean Delvare

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

Re: [lm-sensors] extreme fan rpms with atk0110

[Anders Kullenberg]

> Från: Jean Delvare [mailto:khali@linux-fr.org]
> Skickat: den 16 mars 2011 12:38
> Till: Anders Kullenberg
> Kopia: 'Luca Tettamanti'; 'LM Sensors'
> Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> 
> Anders,
> 
> On Tue, 15 Mar 2011 20:39:12 +0100, Anders Kullenberg wrote:
> > > Från: Luca Tettamanti [mailto:kronos.it@gmail.com]
> > > Skickat: den 15 mars 2011 17:01
> > > Till: Jean Delvare
> > > Kopia: Anders Kullenberg; LM Sensors
> > > Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> > >
> > > On Tue, Mar 15, 2011 at 3:58 PM, Jean Delvare <khali@linux-fr.org>
> > > wrote:
> > > > On Tue, 15 Mar 2011 15:01:24 +0100, Luca Tettamanti wrote:
> > > [cut]
> > > > Found `ITE IT8712F Super IO Sensors'
> Success!
> > > >    (address 0xd00, driver `it87')
> > > >
> > > >> If possible load the native driver (blacklist asus_atk0110 and
> boot
> > > >> with acpi_enforce_resources=lax) and compare the fan readings.
> >
> > Blacklisting is done in "/usr/src/linux-headers-2.6.35-27-
> server/.config" correct?
> > In the ACPI drivers section I have:
> > CONFIG_SENSORS_ATK0110=m
> > CONFIG_SENSORS_LIS3LV02D=m
> >
> > Should I comment both those out?
> 
> No, blacklisting is done with statements in /etc/modprobe.d/*. On my
> distribution, I'd add:
> 
> blacklist asus_atk0110
> 
> to file etc/modprobe.d/99-local.conf. Or you can create a brand new
> file in this directory if you want.
> 
> >
> > Where do I put the "acpi_enforce_resources=lax"
> 
> It has to go on the boot command line, typically you add it when grub
> presents the boot options at system start-up. Alternatively, you can
> edit /boot/grub/menu.lst or /boot/grub/grub.cfg (depending on your
> distribution) manually to add it at the end of the relevant "linux" or
> "kernel" line.
> 
> --
> Jean Delvare

Followed the instructions and it worked until next reboot, then the CPU fan
jumped up to ~3000 rpm's. I went back to using the atk0110 again.
I have computed the CPU fan with @/32,32*@ and the chassis fan with @/8,8*@,
seams to give relatively correct figures. Tried using the set fanN_div
statement to get correct min values but I get an error when running sensors
afterwards. Any recommendations on that?

Thanks
Anders


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

Re: [lm-sensors] extreme fan rpms with atk0110

[Jean Delvare]

Hi Anders,

On Sun, 20 Mar 2011 23:36:32 +0100, Anders Kullenberg wrote:
> > Från: Jean Delvare [mailto:khali@linux-fr.org]
> > Skickat: den 16 mars 2011 12:38
> > Till: Anders Kullenberg
> > Kopia: 'Luca Tettamanti'; 'LM Sensors'
> > Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> > On Tue, 15 Mar 2011 20:39:12 +0100, Anders Kullenberg wrote:
> > > Blacklisting is done in "/usr/src/linux-headers-2.6.35-27-
> > server/.config" correct?
> > > In the ACPI drivers section I have:
> > > CONFIG_SENSORS_ATK0110=m
> > > CONFIG_SENSORS_LIS3LV02D=m
> > >
> > > Should I comment both those out?
> > 
> > No, blacklisting is done with statements in /etc/modprobe.d/*. On my
> > distribution, I'd add:
> > 
> > blacklist asus_atk0110
> > 
> > to file etc/modprobe.d/99-local.conf. Or you can create a brand new
> > file in this directory if you want.
> > 
> > >
> > > Where do I put the "acpi_enforce_resources=lax"
> > 
> > It has to go on the boot command line, typically you add it when grub
> > presents the boot options at system start-up. Alternatively, you can
> > edit /boot/grub/menu.lst or /boot/grub/grub.cfg (depending on your
> > distribution) manually to add it at the end of the relevant "linux" or
> > "kernel" line.
> 
> Followed the instructions and it worked until next reboot, then the CPU fan

Boot parameters typed on the grub command line are only for the current
boot. If you want them to be permanent, you have to edit the grub
configuration files instead.

> jumped up to ~3000 rpm's. I went back to using the atk0110 again.
> I have computed the CPU fan with @/32,32*@ and the chassis fan with @/8,8*@,
> seams to give relatively correct figures. Tried using the set fanN_div
> statement to get correct min values but I get an error when running sensors
> afterwards. Any recommendations on that?

The asus_atk0110 driver doesn't support changing the divisor values,
because the ACPI interface implemented by the BIOS doesn't expose these
settings. So the error you hit is expected.

Please provide the output of the following command:
# isadump 0xd05 0xd06

This will show the IT8712F hardware monitoring registers, and in
particular register 0x0c which contains the 8/16-bit fan configuration
bits. Luca says that the ATK0110 code expects them to be set to 8-bit.

Additionally, I would like to know the exact revision of your IT8712F
chip. The it87 driver logged the information when you loaded it, so you
may be able to find the information in your logs. If not, the following
dump will tell:
# isadump -k 0x87,0x01,0x55,0x55 0x2e 0x2f 4

-- 
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] extreme fan rpms with atk0110

[Anders Kullenberg]

Hi Jean,

> -----Ursprungligt meddelande-----
> Från: Jean Delvare [mailto:khali@linux-fr.org]
> Skickat: den 24 mars 2011 09:40
> Till: Anders Kullenberg
> Kopia: 'Luca Tettamanti'; 'LM Sensors'
> Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> 
> Hi Anders,
> 
> On Sun, 20 Mar 2011 23:36:32 +0100, Anders Kullenberg wrote:
> > > Från: Jean Delvare [mailto:khali@linux-fr.org]
> > > Skickat: den 16 mars 2011 12:38
> > > Till: Anders Kullenberg
> > > Kopia: 'Luca Tettamanti'; 'LM Sensors'
> > > Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> > > On Tue, 15 Mar 2011 20:39:12 +0100, Anders Kullenberg wrote:
> > > > Blacklisting is done in "/usr/src/linux-headers-2.6.35-27-
> > > server/.config" correct?
> > > > In the ACPI drivers section I have:
> > > > CONFIG_SENSORS_ATK0110=m
> > > > CONFIG_SENSORS_LIS3LV02D=m
> > > >
> > > > Should I comment both those out?
> > >
> > > No, blacklisting is done with statements in /etc/modprobe.d/*. On
> my
> > > distribution, I'd add:
> > >
> > > blacklist asus_atk0110
> > >
> > > to file etc/modprobe.d/99-local.conf. Or you can create a brand new
> > > file in this directory if you want.
> > >
> > > >
> > > > Where do I put the "acpi_enforce_resources=lax"
> > >
> > > It has to go on the boot command line, typically you add it when
> grub
> > > presents the boot options at system start-up. Alternatively, you
> can
> > > edit /boot/grub/menu.lst or /boot/grub/grub.cfg (depending on your
> > > distribution) manually to add it at the end of the relevant "linux"
> or
> > > "kernel" line.
> >
> > Followed the instructions and it worked until next reboot, then the
> CPU fan
> 
> Boot parameters typed on the grub command line are only for the current
> boot. If you want them to be permanent, you have to edit the grub
> configuration files instead.

I did added  GRUB_CMDLINE_LINUX="acpi_enforce_resources=lax" to
/etc/default/grub and ran update-grub2
That added it to the grub.cfg file for me. Does not that mean that it should
stick after reboot?

> 
> > jumped up to ~3000 rpm's. I went back to using the atk0110 again.
> > I have computed the CPU fan with @/32,32*@ and the chassis fan with
> @/8,8*@,
> > seams to give relatively correct figures. Tried using the set
> fanN_div
> > statement to get correct min values but I get an error when running
> sensors
> > afterwards. Any recommendations on that?
> 
> The asus_atk0110 driver doesn't support changing the divisor values,
> because the ACPI interface implemented by the BIOS doesn't expose these
> settings. So the error you hit is expected.
> 
> Please provide the output of the following command:
> # isadump 0xd05 0xd06

akullen@ubuntu64:~$ sudo isadump 0xd05 0xd06
[sudo] password for akullen:
WARNING! Running this program can cause system crashes, data loss and worse!
I will probe address register 0xd05 and data register 0xd06.
Continue? [Y/n] y
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: 19 00 ff 03 00 00 00 00 00 80 7f 3b 07 80 ff ff
10: ff ff ff 77 87 80 7f 7f 00 ff ff ff ff ff ff ff
20: 46 ff cf ba ba ff ff ff cc 1e 2e 80 0e fe fe fe
30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
40: ff ff ff ff ff ff ff ff 2d ff ff ff ff ff ff ff
50: 1d 18 7f 7f 7f 00 56 00 90 56 ff 12 80 00 00 00
60: 00 23 7f 37 18 03 ff ff 00 19 7f 28 18 03 ff ff
70: 7f 7f 7f 00 00 7f ff ff ff ff ff ff ff ff ff ff
80: 00 00 00 00 ff ff ff ff 00 00 ff ca 02 00 99 99
90: 7f 7f 7f 00 00 7f ff ff 7f 7f 7f 00 00 7f ff ff
a0: 00 00 00 00 00 00 00 ff ff ff ff ff ff ff ff ff
b0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
c0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

> 
> This will show the IT8712F hardware monitoring registers, and in
> particular register 0x0c which contains the 8/16-bit fan configuration
> bits. Luca says that the ATK0110 code expects them to be set to 8-bit.
> 
> Additionally, I would like to know the exact revision of your IT8712F
> chip. The it87 driver logged the information when you loaded it, so you
> may be able to find the information in your logs. If not, the following
> dump will tell:
> # isadump -k 0x87,0x01,0x55,0x55 0x2e 0x2f 4

akullen@ubuntu64:/var/log$ sudo isadump -k 0x87,0x01,0x55,0x55 0x2e 0x2f 4
[sudo] password for akullen:
WARNING! Running this program can cause system crashes, data loss and worse!
I will probe address register 0x2e and data register 0x2f.
Probing bank 4 using bank register 0x07.
Continue? [Y/n] y
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: 00 00 00 00 00 00 00 04 00 00 00 00 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 87 12 08 01 00 00 00 00 00 40 00 00 1d 00 00 00
30: 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
40: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
50: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
60: 0d 00 0c 00 00 00 00 00 00 00 00 00 00 00 00 00
70: 00 02 00 00 04 04 00 00 00 00 00 00 00 00 00 00
80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
f0: 00 00 00 00 a0 00 fd 00 00 00 00 00 00 00 00 00

Also found this in the logs
messages:Mar 21 22:18:38 ubuntu64 kernel: [    6.528399] it87: Found IT8712F
chip at 0xd00, revision 8
messages:Mar 21 22:18:38 ubuntu64 kernel: [    6.528413] it87: in3 is VCC
(+5V)
messages:Mar 21 22:18:38 ubuntu64 kernel: [    6.528454] ACPI: resource it87
[io  0x0d05-0x0d06] conflicts w                       ith ACPI region ECRE
[mem 0x00000d00-0x00000d1f 64bit pref window disabled]

Is this something to worry about?

Thanks
Anders

> 
> --
> 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] extreme fan rpms with atk0110

[Jean Delvare]

Hi Anders,

On Thu, 24 Mar 2011 18:23:08 +0100, Anders Kullenberg wrote:
> Hi Jean,
> 
> > -----Ursprungligt meddelande-----
> > Från: Jean Delvare [mailto:khali@linux-fr.org]
> > Skickat: den 24 mars 2011 09:40
> > Till: Anders Kullenberg
> > Kopia: 'Luca Tettamanti'; 'LM Sensors'
> > Ämne: Re: [lm-sensors] extreme fan rpms with atk0110
> >
> > Boot parameters typed on the grub command line are only for the current
> > boot. If you want them to be permanent, you have to edit the grub
> > configuration files instead.
> 
> I did added  GRUB_CMDLINE_LINUX="acpi_enforce_resources=lax" to
> /etc/default/grub and ran update-grub2
> That added it to the grub.cfg file for me. Does not that mean that it should
> stick after reboot?

I'm not familiar with grub2, but yes, it certainly looks like it should
stick after reboot.

> > (...)
> > Please provide the output of the following command:
> > # isadump 0xd05 0xd06
> 
> akullen@ubuntu64:~$ sudo isadump 0xd05 0xd06
> [sudo] password for akullen:
> WARNING! Running this program can cause system crashes, data loss and worse!
> I will probe address register 0xd05 and data register 0xd06.
> Continue? [Y/n] y
>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> 00: 19 00 ff 03 00 00 00 00 00 80 7f 3b 07 80 ff ff

You can see the value of register 0x0c: 0x07. This means fan1, fan2 and
fan3 in 16-bit mode.

> > This will show the IT8712F hardware monitoring registers, and in
> > particular register 0x0c which contains the 8/16-bit fan configuration
> > bits. Luca says that the ATK0110 code expects them to be set to 8-bit.

According to Luca, the ATK0110 code on this board expects all fans to
be in 8-bit mode (which according to the IT8712F datasheet is the
default, but the datasheet isn't necessarily correct on this.)

I don't know who enabled 16-bit mode... undocumented hardware default,
faulty BIOS, Linux kernel driver (the it87 driver would, but is
normally not allowed to load on ATK0110-enabled machines)? But
certainly the atk0110 driver would work fine if we the register value
was reverted to 8-bit mode.

So you could do the following:
* Unload the it87 driver if it is loaded.
* Temporarily comment out the fan compute statements in the
  asus_atk0110 section of your libsensors configuration file.
* As root, run:
  # isaset 0xd05 0xd06 0x0c 0x00 0x07
  This will set fan1, fan2 and fan3 to 8-bit mode.
* Load the asus_atk0110 driver if it's not already loaded.
* Check the fan speeds. I expect them to be back to normal.

Now there is a potential problem: the datasheet for the revision J of
the IT8712F (which is the one you have) does _not_ list fan clock
divisor register which 8-bit fan mode needs. If the register in
question is really no longer implemented in the hardware, this makes
the 8-bit fan mode pretty much useless, as speeds below ~2650 RPM can't
be monitored.

So, here's my theory: I think Asus has been using the previous revision
(I) of the IT8712F for this board originally, and it supported 8-bit
fan mode just fine, and they wrote their BIOS for it. Then ITE upgraded
their line to revision J and Asus used this as a replacement of the
revision I, without realizing it wasn't fully compatible. Now they may
have two variants of the board in the wild.

> (...)
> Also found this in the logs
> messages:Mar 21 22:18:38 ubuntu64 kernel: [    6.528399] it87: Found IT8712F
> chip at 0xd00, revision 8
> messages:Mar 21 22:18:38 ubuntu64 kernel: [    6.528413] it87: in3 is VCC
> (+5V)
> messages:Mar 21 22:18:38 ubuntu64 kernel: [    6.528454] ACPI: resource it87
> [io  0x0d05-0x0d06] conflicts with ACPI region ECRE
> [mem 0x00000d00-0x00000d1f 64bit pref window disabled]
> 
> Is this something to worry about?

No, this is expected. This is the I/O resource conflict which forces
you to choose between asus_atk0110 and it87.

-- 
Jean Delvare

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