[PATCH v3 2/3] arm: dts: add Nuvoton NPCM750 device tree

Brendan Higgins brendanhiggins at google.com
Wed Sep 6 18:19:11 AEST 2017


On Tue, Sep 5, 2017 at 10:23 PM, Joel Stanley <joel at jms.id.au> wrote:
> On Wed, Sep 6, 2017 at 10:00 AM, Brendan Higgins
> <brendanhiggins at google.com> wrote:
>> +++ b/Documentation/devicetree/bindings/arm/npcm/npcm.txt
>> @@ -0,0 +1,6 @@
>> +NPCM Platforms Device Tree Bindings
>> +-----------------------------------
>> +NPCM750 SoC
>> +Required root node properties:
>> +       - compatible = "nuvoton,npcm750";
>> +
>
> This is minimal. I assume there will be more content added as more
> support is added?

Yep, that's the plan. They have a number of similar BMCs, both those
based on different ARM cores and some with different peripheral sets,
so we will probably want to have different compat strings for those.

>
> Does it need it's own directory?

Not sure, I saw that some of the other architectures did it, some did
not. I don't feel strongly about it.

>
>
>> diff --git a/arch/arm/boot/dts/nuvoton-npcm750-evb.dts b/arch/arm/boot/dts/nuvoton-npcm750-evb.dts
>> new file mode 100644
>> index 000000000000..54df32cff21b
>> --- /dev/null
>> +++ b/arch/arm/boot/dts/nuvoton-npcm750-evb.dts
>
>> +
>> +/dts-v1/;
>> +#include "nuvoton-npcm750.dtsi"
>> +
>> +/ {
>> +       model = "Nuvoton npcm750 Development Board (Device Tree)";
>> +       compatible = "nuvoton,npcm750";
>> +
>> +       chosen {
>> +               stdout-path = &serial3;
>> +               bootargs = "earlyprintk=serial,serial3,115200";
>> +       };
>> +
>> +       memory {
>> +               reg = <0 0x40000000>;
>> +       };
>> +
>> +       cpus {
>> +               enable-method = "nuvoton,npcm7xx-smp";
>> +       };
>> +
>> +       clk: clock-controller at f0801000 {
>> +               status = "okay";
>> +       };
>> +
>> +       apb {
>> +               watchdog1: watchdog at f0009000 {
>> +                       status = "okay";
>> +               };
>
> You've already got the label for the node, is there are reason you
> don't use a phandle to set the status?

Addressed in v4.

>
> &watchdog1 {
>    status = "okay";
> };
>
> Same with the serial nodes below.
>
>> +
>> +               serial0: serial0 at f0001000 {
>> +                       status = "okay";
>> +               };
>> +
>> +               serial1: serial1 at f0002000 {
>> +                       status = "okay";
>> +               };
>> +
>> +               serial2: serial2 at f0003000 {
>> +                       status = "okay";
>> +               };
>> +
>> +               serial3: serial3 at f0004000 {
>> +                       status = "okay";
>> +               };
>> +       };


More information about the openbmc mailing list