Re: [lm-sensors] Lenovo Yoga 13 with SMSC EMC2104 chip

[Guenter Roeck <linux@roeck-us.net>]

On Wed, Oct 30, 2013 at 08:29:45AM +0100, Jean Delvare wrote:
> On Wed, 30 Oct 2013 07:52:51 +0100, Igor Popik wrote:
> > 2013/10/30 Guenter Roeck <linux@roeck-us.net>:
> > > As Jean mentioned, there may be another i2c bus, a hidden i2c multiplexer,
> > > or the chip may be connected to the IDT EC. The DSDT might give us
> > > some hints. You already know from the ACPI overlap that ACPI talks with the
> > > SMBus adapter; it might also talk with the EC. Can you extract the DSDT
> > > and provide it to us ?
> > >
> > >         sudo cat /sys/firmware/acpi/tables/DSDT > dsdt.aml
> > >
> > > Then, in case you want to learn things you never really wanted to know:
> > >
> > >         iasl -d dsdt.aml
> > 
> > Here it is. Hope You find something interesting there.
> 
> I see the following:
> 
>                 Scope (\_SB.PCI0.LPCB)
>                 {
>                     Device (H_EC)
>                     {
>                         Name (_HID, EisaId ("PNP0C09"))  // _HID: Hardware ID
>                         Name (_UID, One)  // _UID: Unique ID
>                         (...)
>                         OperationRegion (ECF2, EmbeddedControl, Zero, 0xFF)
>                         Field (ECF2, ByteAcc, Lock, Preserve)
>                         {
>                             (...)
>                             Offset (0xEE), 
>                             BHFW,   16, 
>                             F1PW,   8, 
>                             F2PW,   8, 
>                             F1RM,   8, 
>                             F2RM,   8, 
>                             FAMO,   8, 
>                             F1SP,   16, 
>                             F2SP,   16, 
>                             FAN1,   8, 
>                             FAN2,   8, 
>                             BCG1,   16, 
>                             BCG2,   16
> 
> This seems to imply there is an EC which knows something about fans.
> However there is no reference to these fields in the rest of the DSDT.
> There could be references in other ACPI tables though. Also note that
> PNP0C09 is referenced in driver thinkpad_acpi so maybe that driver can
> help. That being said, it seems that newer laptop models from Lenovo
> aren't supported by this driver as good as older models were.
> 
Given the "flexibility" in ACPI permitting vendors to change bindings
and underlying functionality at their leisure, that is hardly surprising.

Looking through the driver, there is lots of magic in there anyway.
I think the best shot would be to load the driver and hope that it works.
However, the following note on thinkwiki.org, related to Lenovo 3000, suggests
that it may not work: "The ACPI and EC firmware are completely incompatible
with ThinkPads, and thinkpad-acpi will not support it. The Lenovo 3000 are not
ThinkPads". Since the Yoga isn't even a laptop, that doesn't sound very 
encouraging.

> Also interesting is the SBUS device which defines a complete ACPI
> interface to the SMBus controller. Unfortunately it doesn't follow the
> SMBus CMI standard. Writing an ACPI driver for it would allow working
> around the i2c-i801 resource conflict. However this doesn't seem
> terribly useful as there's only the memory module SPD EEPROM visible on
> the SMBus.
> 
Also, the interface may change with the next laptop model, so a driver
would only have limited value.

> If it was me, I think I'd just leave the whole thing alone unless you
> have a real problem that needs to be solved.
> 
+1 

Guenter

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