[lm-sensors] wrong readings for w83627ehf on p5mt-m

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Alexander Dobin]

Hello,
I am trying to get lm_sensors to work on the ASUS p5mt-m motherboard.
I can see the Winbond w83627ehf-A chip on the board.

I have kernel 2.6.14 and have installed lm_sensors 2.9.2

sensors-detect find the chip (see the output below), and after all the
suggested steps,
sensors output is apparently nothing for fans and temps (see below).

I will be happy to quickly provide any additional information if any of
developers decide to resolve this problem.

Thanks a lot in advance,
Alex Dobin

--------------------------------------------
sensors output:

w83627ehf-isa-0290
Adapter: ISA adapter
fan1:        0 RPM  (min = 10546 RPM, div = 128)
fan2:        0 RPM  (min =    0 RPM, div = 128)
fan3:        0 RPM  (min =    0 RPM, div = 128)
fan4:        0 RPM  (min = 10546 RPM, div = 128)
temp1:       -48?C  (high =    +0?C, hyst =   +64?C)
temp2:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)

-----------------------------------------------------------------------------
Here is the "summary of probes" of the sensors-detect:

Driver `eeprom' (should be inserted):
  Detects correctly:
  * Bus `SMBus I801 adapter at 0400'
    Busdriver `i2c-i801', I2C address 0x51
    Chip `SPD EEPROM' (confidence: 8)
  * Bus `SMBus I801 adapter at 0400'
    Busdriver `i2c-i801', I2C address 0x55
    Chip `SPD EEPROM' (confidence: 8)

Driver `w83627ehf' (should be inserted):
  Detects correctly:
  * ISA bus address 0x0290 (Busdriver `i2c-isa')
    Chip `Winbond W83627EHF/EHG Super IO Sensors' (confidence: 9)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.lm-sensors.org/pipermail/lm-sensors/attachments/20060114/6262c80e/attachment.html

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Rudolf Marek]

> I will be happy to quickly provide any additional information if any of
> developers decide to resolve this problem.


Hello, please send back output of:

isadump 0x295 0x296

Thanks

Regards
Rudolf

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Alexander Dobin]

Hello, Rudolf,
thanks for your interest in this problem!
Below is the output you asked for.
Alex

isadump 0x295 0x296
WARNING! Running this program can cause system crashes, data loss and worse!
I will probe address register 0x295 and data register 0x296.
Continue? [Y/n]
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: 01 bf 01 bf 01 00 00 00 01 01 01 01 3c 3c 0a 0a
10: 01 bf 00 00 00 01 01 3c 43 00 ff ff 24 32 00 de
20: ff ff d5 d3 d3 d9 f9 d0 ff ff ff e7 00 00 00 04
30: 02 00 80 10 48 28 00 00 00 00 00 02 00 00 01 ff
40: 03 00 00 de ff 00 03 04 2d 44 01 04 10 95 00 a3
50: ff ff 80 ff ff ff 00 80 88 73 ff ff 19 e0 ff 05
60: ff ff d5 d3 d3 d9 f9 d0 ff ff ff e7 00 00 00 04
70: 02 00 80 10 48 28 00 00 00 00 00 02 00 00 01 ff
80: 01 bf 01 bf 01 00 00 00 01 01 01 01 3c 3c 0a 0a
90: 01 bf 00 00 00 01 01 3c 43 00 ff ff 24 32 00 de
a0: ff ff d5 d3 d3 d9 f9 d0 ff ff ff e7 00 00 00 04
b0: 02 00 80 10 48 28 00 00 00 00 00 02 00 00 01 ff
c0: 03 00 00 de ff 00 03 04 2d 44 01 04 10 95 00 a3
d0: ff ff 80 ff ff ff 00 80 88 73 ff ff 19 e0 ff 05
e0: ff ff d5 d3 d3 d9 f9 d0 ff ff ff e7 00 00 00 04
f0: 02 00 80 10 48 28 00 00 00 00 00 02 00 00 01 ff


On 1/14/06, Rudolf Marek <r.marek at sh.cvut.cz> wrote:
>
> > I will be happy to quickly provide any additional information if any of
> > developers decide to resolve this problem.
>
>
> Hello, please send back output of:
>
> isadump 0x295 0x296
>
> Thanks
>
> Regards
> Rudolf
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.lm-sensors.org/pipermail/lm-sensors/attachments/20060117/98bab3d8/attachment.html

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Alexander Dobin]

Hello again,

I have some additional information - may be it will help.
SpeedFan (under Win2k) can detect the system and processor temperaure
correctly  - I compared with ASUS proprietary monitor.
Is it possible to figure out how this is done in SpeedFan?

Also, when I change the load, the isadump output changes in many places.
My humble guess is that the EHF datasheet does not work with EHF-A.

Thanks
Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.lm-sensors.org/pipermail/lm-sensors/attachments/20060118/6fc00fca/attachment.html

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Ymu]

Hello
 
> I have some additional information - may be it will help.
> SpeedFan (under Win2k) can detect the system and processor temperaure
> correctly  - I compared with ASUS proprietary monitor.
> Is it possible to figure out how this is done in SpeedFan?

Can you give me some pictures about SpeedFan?
 
> Also, when I change the load, the isadump output changes in many places.
> My humble guess is that the EHF datasheet does not work with EHF-A.
I have a board with W83627EHF-A chip and it works with lm_sensors. So I would believe the datasheet should work on EHF-A :)
 
Best Regards
Yuan Mu



=============================================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] wrong readings for w83627ehf on p5mt-m

[Jean Delvare]

Hi Alexander,

> I have kernel 2.6.14 and have installed lm_sensors 2.9.2
> 
> sensors-detect find the chip (see the output below), and after all the
> suggested steps,
> sensors output is apparently nothing for fans and temps (see below).
> 
> I will be happy to quickly provide any additional information if any of
> developers decide to resolve this problem.

Please provide the output of:
  isadump -k 0x87,0x87 0x2e 0x2f 0xb

I want to check the value of registers 0x22, which may be used to shut
down the hardware monitoring part of the device, and 0x30, which may
disable that part. I guess that you wouldn't even be able to dump the
device contents if it was shut down or disabled, but it doesn't hurt to
make sure.

> sensors output:
> 
> w83627ehf-isa-0290
> Adapter: ISA adapter
> fan1:        0 RPM  (min = 10546 RPM, div = 128)
> fan2:        0 RPM  (min =    0 RPM, div = 128)
> fan3:        0 RPM  (min =    0 RPM, div = 128)
> fan4:        0 RPM  (min = 10546 RPM, div = 128)
> temp1:       -48?C  (high =    +0?C, hyst =   +64?C)
> temp2:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
> temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)

For fans, try setting the limits to something reasonable like 1000 RPM.
It might help.

For temperatures, try:
  isaset 0x295 0x296 0x4e 0x00
  isaset 0x295 0x296 0x59 0x00
This will change the diode measurement type from Pentium II to 2N3904.
Maybe you'll have different readings after that. If it doesn't help,
try:
  isaset 0x295 0x296 0x5d 0x07
This will change the measurement mode from diode to thermistor. This
should change the reported values, although I can't promise the new
values will make more sense than the old ones.

Ideally, the driver should at least report the currently selected
thermal sensor type, and may even let the user change it to workaround
broken BIOSes. The w83627ehf driver is still in a preliminary stage
though.

We have a patch floating around which adds voltage input support to the
driver. Do you want to give it a try? Maybe you'll be more lucky with
voltages than with fans and temperatures. The dump you provided
elsewhere in this thread shows that at least some voltage values make
sense.

-- 
Jean Delvare

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Jean Delvare]

Hi again Alexander,

> I have some additional information - may be it will help.
> SpeedFan (under Win2k) can detect the system and processor temperaure
> correctly  - I compared with ASUS proprietary monitor.
> Is it possible to figure out how this is done in SpeedFan?

Not really, SpeedFan is closed source so it is of little use to us. It
can tell us what information is available, but not how to get it.

Please list everything that is reported by the Asus tool and/or
SpeedFan under Windows and/or your BIOS. If we know exactly what
information we are looking for, it'll help our investugation efforts.

Are you certain that these values are taken from the W83627EHF chip and
not another chip on your board? SpeedFan should tell you if I remember
correctly. If you only have temperature reported, it may even come from
ACPI.

> Also, when I change the load, the isadump output changes in many places.

Please provide consecutive dumps without then with load, so that we can
check which register values change and how.

> My humble guess is that the EHF datasheet does not work with EHF-A.

Very unlikely, as Yuan Mu from Winbond confirmed elsewhere in this
thread. Many Winbond chips have -A top markings and are still working
just like the datasheet says.

-- 
Jean Delvare

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Alexander Dobin]

Hello, Jean,

thanks a lot for your help!
No good news, I am afraid.

Here are some outputs you requested:

isadump -k 0x87,0x87 0x2e 0x2f 0xb

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: ff ff ff ff ff ff ff 0b ff ff ff ff ff ff ff ff
10: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
20: 88 54 ff 00 44 00 00 ff 50 04 00 00 fb 21 00 ff
30: 01 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
40: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
50: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
60: 02 90 ff ff ff ff ff ff ff ff ff ff ff ff ff ff
70: 00 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
90: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
a0: ff ff ff ff ff ff ff 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: 11 3f ff ff ff ff ff ff ff ff ff ff ff ff ff ff
I have added settings for fans in the sensors.config file, but to no avail:
w83627ehf-isa-0290
Adapter: ISA adapter
fan1:        0 RPM  (min = 1004 RPM, div = 16)
fan2:        0 RPM  (min = 1004 RPM, div = 16)
fan3:        0 RPM  (min = 1004 RPM, div = 16)
fan4:        0 RPM  (min = 1004 RPM, div = 16)
temp1:       -48?C  (high =    +0?C, hyst =   +64?C)
temp2:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
I have two fans, CPU Fan1 and CPU Fan2, with
BIOS reporting 2500 and 2300 rpms.

Oddly, when I run 'sensors' consequently the min and div values change until
they saturate on th 4th time:
fan1:        0 RPM  (min = 1054 RPM, div = 128)
fan2:        0 RPM  (min = 1054 RPM, div = 128)
fan3:        0 RPM  (min = 1054 RPM, div = 128)
fan4:        0 RPM  (min = 1054 RPM, div = 128)
temp1:       -48?C  (high =    +0?C, hyst =   +64?C)
temp2:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)

Also tried:
isaset 0x295 0x296 0x4e 0x00
isaset 0x295 0x296 0x59 0x00
1st line went in OK, for the second I got:
Data mismatch, wrote 0x00, read 0xff back.
I still tried sensors after that:

fan1:        0 RPM  (min = 1004 RPM, div = 32)
fan2:        0 RPM  (min = 1004 RPM, div = 32)
fan3:        0 RPM  (min = 1004 RPM, div = 32)
fan4:        0 RPM  (min = 1004 RPM, div = 32)
temp1:       -48?C  (high =    +0?C, hyst =   +64?C)
temp2:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)

Also tried:
isaset 0x295 0x296 0x5d 0x07
Got:
Data mismatch, wrote 0x07, read 0xff back.
This changed the sensors output like this:
fan1:        0 RPM  (min = 16071 RPM, div = 4)
fan2:        0 RPM  (min = 16071 RPM, div = 4)
fan3:        0 RPM  (min = 16071 RPM, div = 4)
fan4:        0 RPM  (min = 1004 RPM, div = 64)
temp1:      +127?C  (high =    +0?C, hyst =   +64?C)
temp2:    +127.5?C  (high = +80.0?C, hyst = +75.0?C)
temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
In this configuration, weirdly, as I put some load on, the temps did not
change, however the min and div values for fans jumped around.

I rebooted to get to the initial state.
and now while idling: isadump 0x295 0x296
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: 01 bf 01 bf 01 00 00 00 01 01 01 01 3c 3c 0a 0a
10: 01 bf 00 00 00 01 01 3c 43 00 ff ff 24 32 00 de
20: db a5 d4 d3 c7 c5 ef d0 ff ff ff e7 00 08 00 04
30: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
40: 03 cf 0f de ff 00 03 f4 2d 44 01 c4 90 95 00 a3
50: ff ff 80 ff ff ff 00 80 88 73 ff ff 19 e0 ff 05
60: db a5 d4 d3 c7 c5 ef d0 ff ff ff e7 00 08 00 04
70: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
80: 01 bf 01 bf 01 00 00 00 01 01 01 01 3c 3c 0a 0a
90: 01 bf 00 00 00 01 01 3c 43 00 ff ff 24 32 00 de
a0: db a5 d4 d3 c7 c5 ef d0 ff ff ff e7 00 08 00 04
b0: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
c0: 03 00 00 de ff 00 03 f4 2d 44 01 c4 90 95 00 a3
d0: ff ff 80 ff ff ff 00 80 88 73 ff ff 19 e0 ff 05
e0: db a5 d4 d3 c7 c5 ef d0 ff ff ff e7 00 08 00 04
f0: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
At full load:
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: 01 bf 01 bf 01 00 00 00 01 01 01 01 3c 3c 0a 0a
10: 01 bf 00 00 00 01 01 3c 43 00 ff ff 24 32 00 de
20: d9 a3 d4 d2 c6 c3 ed d0 ff ff ff e7 00 08 00 04
30: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
40: 03 00 00 de ff 00 03 f4 2d 44 01 c4 90 95 00 a3
50: ff ff 80 ff ff ff 00 80 88 73 ff ff 19 e0 ff 05
60: d9 a3 d4 d2 c6 c3 ed d0 ff ff ff e7 00 08 00 04
70: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
80: 01 bf 01 bf 01 00 00 00 01 01 01 01 3c 3c 0a 0a
90: 01 bf 00 00 00 01 01 3c 43 00 ff ff 24 32 00 de
a0: d9 a3 d4 d2 c6 c3 ed d0 ff ff ff e7 00 08 00 04
b0: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff
c0: 03 00 00 de ff 00 03 f4 2d 44 01 c4 90 95 00 a3
d0: ff ff 80 ff ff ff 00 80 88 73 ff ff 19 e0 ff 05
e0: d9 a3 d4 d2 c6 c3 ed d0 ff ff ff e7 00 08 00 04
f0: 02 00 80 90 48 29 00 01 02 00 40 0a 0a 0a 0a ff

I had to get rid of Windows - will put it back and report SpeedFan and ASUS
numbers hopefully tomorrow.
The BIOS actually supports ACPI and SpeedFan could only report CPU and MB
temperature, so it could have been ACPI - will check tomorrow.
ACPI under Linux does not work apparently - no information on T or fans.
The BIOS reports:

CPU Temp
MB  Temp

CPU Fan1
CPU Fan2
Front Fan1
Front Fan2
Rear Fan1
Rear Fan2

Smart Fan Disabled

VCORE
3.3V
5V
5VSB
VBA
12V

I will be happy to try the patch for voltages or any other things you
suggest,
but I will need instructions since I am not fluent in this stuff.

Thanks a lot again for your interest and help,

Have a good weekend,
Alex


8/06, Jean Delvare <khali at linux-fr.org> wrote:
> Hi Alexander,
>
> > I have kernel 2.6.14 and have installed lm_sensors 2.9.2
> >
> > sensors-detect find the chip (see the output below), and after all the
> > suggested steps,
> > sensors output is apparently nothing for fans and temps (see below).
> >
> > I will be happy to quickly provide any additional information if any of
> > developers decide to resolve this problem.
>
> Please provide the output of:
> isadump -k 0x87,0x87 0x2e 0x2f 0xb
>
> I want to check the value of registers 0x22, which may be used to shut
> down the hardware monitoring part of the device, and 0x30, which may
> disable that part. I guess that you wouldn't even be able to dump the
> device contents if it was shut down or disabled, but it doesn't hurt to
> make sure.
>
> > sensors output:
> >
> > w83627ehf-isa-0290
> > Adapter: ISA adapter
> > fan1:        0 RPM  (min = 10546 RPM, div = 128)
> > fan2:        0 RPM  (min =    0 RPM, div = 128)
> > fan3:        0 RPM  (min =    0 RPM, div = 128)
> > fan4:        0 RPM  (min = 10546 RPM, div = 128)
> > temp1:       -48?C  (high =    +0?C, hyst =   +64?C)
> > temp2:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
> > temp3:     -48.0?C  (high = +80.0?C, hyst = +75.0?C)
>
> For fans, try setting the limits to something reasonable like 1000 RPM.
> It might help.
>
> For temperatures, try:
> isaset 0x295 0x296 0x4e 0x00
> isaset 0x295 0x296 0x59 0x00
> This will change the diode measurement type from Pentium II to 2N3904.
> Maybe you'll have different readings after that. If it doesn't help,
> try:
> isaset 0x295 0x296 0x5d 0x07
> This will change the measurement mode from diode to thermistor. This
> should change the reported values, although I can't promise the new
> values will make more sense than the old ones.
>
> Ideally, the driver should at least report the currently selected
> thermal sensor type, and may even let the user change it to workaround
> broken BIOSes. The w83627ehf driver is still in a preliminary stage
> though.
>
> We have a patch floating around which adds voltage input support to the
> driver. Do you want to give it a try? Maybe you'll be more lucky with
> voltages than with fans and temperatures. The dump you provided
> elsewhere in this thread shows that at least some voltage values make
> sense.
>
> --
> Jean Delvare
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.lm-sensors.org/pipermail/lm-sensors/attachments/20060120/482acef9/attachment-0001.html

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Rudolf Marek]

Hello,


Please can you send me a copy of your DSDT acpi table?
cat /proc/acpi/dsdt > /tmp/dsdt.file

(enough is to me in private because mailing list is not allowing binary attachments)

Thanks

Regards,

Rudolf

[lm-sensors] wrong readings for w83627ehf on p5mt-m

[Rudolf Marek]

Rudolf Marek wrote:
> Hello,
> 
> 
> Please can you send me a copy of your DSDT acpi table?
> cat /proc/acpi/dsdt > /tmp/dsdt.file
> 
> (enough is to me in private because mailing list is not allowing binary attachments)

Thanks, for the record there is nothing interresting in ACPI.

Regards
Rudolf