On 09/26/2013 02:33 PM, Kumar Gala wrote:
>
> On Sep 26, 2013, at 2:17 PM, Rohit Vaswani wrote:
>
>> On 9/26/2013 11:05 AM, Rohit Vaswani wrote:
>>> On 9/26/2013 9:37 AM, Kumar Gala wrote:
>>>> <snip>
>>>
>>>> +++ b/arch/arm/boot/dts/qcom-apq8074-dragonboard.dts @@ -0,0
>>>> +1,6 @@ +/include/ "qcom-msm8974.dtsi" + +/ { + model =
>>>> "Qualcomm APQ8074 Dragonboard"; + compatible =
>>>> "qcom,apq8074-dragonboard", "qcom,apq8074"; +}; diff --git
>>>> a/arch/arm/boot/dts/qcom-msm8974.dtsi
>>>> b/arch/arm/boot/dts/qcom-msm8974.dtsi new file mode 100644
>>>> index 0000000..f04b643 --- /dev/null +++
>>>> b/arch/arm/boot/dts/qcom-msm8974.dtsi @@ -0,0 +1,35 @@
>>>> +/dts-v1/; + +/include/ "skeleton.dtsi" + +/ { + model =
>>>> "Qualcomm MSM8974"; + compatible = "qcom,msm8974"; +
>>>> interrupt-parent = <&intc>; + + soc: soc { };
>>>>>> We should have a unit address here:
>>>>>>
>>>>>> soc: soc@FOOBAR {
>>>>>>
>>>>>> also, split out the curly braces so any future patches do
>>>>>> have to muck with that.
>>>>>>
>>>>>> };
>>>>>>
>>>>> Im not sure I understand the reasoning behind the unit
>>>>> address for soc ?
>>>> Its fairly standard practice and there is a fair amount of
>>>> discussion about the lack of a unit address for memory nodes.
>>>>
>>> That still doesn't really answer anything :) - and I couldn't
>>> find any discussions about this either. I don't see anybody in
>>> upstream adding an address to soc except sun. What is that
>>> address supposed to be for - what does it mean ? The soc is way
>>> of encapsulating meaningful blocks for the particular SoC.
>>
>> I see the mail from Stephen Warren for adding a check stating that
>>
>> "ePAPR 1.1 section 2.2.1.1 "Node Name Requirements" specifies that
>> any node that has a reg property must include a unit address in its
>> name with value matching the first entry in its reg property.
>> Conversely, if a node does not have a reg property, the node name
>> must not include a unit address."
>>
>> The soc node we have does not have a reg property ?
>
> Not 100% sure what people will decide on this. There are a number of
> examples on the PPC side (arch/powerpc/boot/dts) that are soc@ADDR,
> but they don't typically have "reg" properties at the soc level.
No, but you may have a ranges property which is related.
I've just hit this on highbank in needing to add a second bank of
peripherals for midway. So my vote would be to have unit address.
Rob
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
