On Fri, Apr 14, 2017 at 5:48 AM, Rob Herring <r...@kernel.org> wrote:
> On Thu, Apr 13, 2017 at 4:07 PM, Guenter Roeck <li...@roeck-us.net> wrote:
>> On Thu, Apr 13, 2017 at 03:01:40PM -0500, Rob Herring wrote:
>>> On Fri, Apr 07, 2017 at 03:00:09PM -0700, Moritz Fischer wrote:
>>> > From: Moritz Fischer <m...@kernel.org>
>>> >
>>> > Add bindings for the Chromium EC HWMON. The Chromium EC HWMON
>>> > allows monitoring of temperature sensors and fans attached to the
>>> > EC.
>>> >
>>> > Signed-off-by: Moritz Fischer <m...@kernel.org>
>>> > ---
>>> > .../devicetree/bindings/hwmon/cros-ec-hwmon.txt | 25
>>> > ++++++++++++++++++++++
>>> > 1 file changed, 25 insertions(+)
>>> > create mode 100644
>>> > Documentation/devicetree/bindings/hwmon/cros-ec-hwmon.txt
>>> >
>>> > diff --git a/Documentation/devicetree/bindings/hwmon/cros-ec-hwmon.txt
>>> > b/Documentation/devicetree/bindings/hwmon/cros-ec-hwmon.txt
>>> > new file mode 100644
>>> > index 0000000..4c94869
>>> > --- /dev/null
>>> > +++ b/Documentation/devicetree/bindings/hwmon/cros-ec-hwmon.txt
>>> > @@ -0,0 +1,25 @@
>>> > +Chromium Embedded Controller EC temperature and fan control
>>> > +-----------------------------------------------------------
>>> > +
>>> > +Google's Chromium EC HWMON is a hwmon implemented byimplemented by the
>>> > Chromium EC
>>> > +firmware attached to the Embedded Controller (EC) and controlled via a
>>> > host-command
>>> > +interface.
>>> > +
>>> > +An EC HWMON node should be only found as a sub-node of the EC node (see
>>> > +Documentation/devicetree/bindings/mfd/cros-ec.txt).
>>> > +
>>> > +Required properties:
>>> > +- compatible: Must contain "google,cros-ec-hwmon"
>>> > +
>>> > +Example:
>>> > + embedded-controller@1e {
>>> > + reg = <0x1e>;
>>> > + compatible = "google,cros-ec-i2c";
>>> > + interrupts = <0 IRQ_TYPE_LEVEL_LOW>;
>>> > + interrupt-parent = <&gpio0>;
>>> > +
>>> > + hwmon {
>>> > + compatible = "google,cros-ec-hwmon";
>>>
>>> This is sufficient for all devices? I don't see that DT provides
>>> anything here other than instantiating a device, but the parent device
>>> can just as easily do that.
>>>
>> The parent driver (drivers/mfd/cros_ec_i2c.c) calls cros_ec_register(),
>> which uses uses of_platform_populate() to populate all sub-devices.
>> There are various examples in the dts files (look for "google,cros-ec").
>> Does it really make sense to start a second method for instantiating
>> sub-devices ?
>
> Okay, I suppose not. That wasn't clear from the example.
Do you want me to clarify that in the example somehow?
Moritz
