Re: [PATCH v2 1/2] dt-binding: pinctrl: Add NPCM8XX pinctrl and GPIO documentation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, 20 Sept 2022 at 11:47, Krzysztof Kozlowski
<krzysztof.kozlowski@xxxxxxxxxx> wrote:
>
> On 20/09/2022 10:32, Tomer Maimon wrote:
> > On Tue, 20 Sept 2022 at 11:21, Krzysztof Kozlowski
> > <krzysztof.kozlowski@xxxxxxxxxx> wrote:
> >>
> >> On 20/09/2022 09:59, Tomer Maimon wrote:
> >>>>>>>>> +      pinctrl: pinctrl@f0800000 {
> >>>>>>>>> +        compatible = "nuvoton,npcm845-pinctrl";
> >>>>>>>>> +        ranges = <0x0 0x0 0xf0010000 0x8000>;
> >>>>>>>>> +        #address-cells = <1>;
> >>>>>>>>> +        #size-cells = <1>;
> >>>>>>>>> +        nuvoton,sysgcr = <&gcr>;
> >>>>>>>>> +
> >>>>>>>>> +        gpio0: gpio@f0010000 {
> >>>>>>>>
> >>>>>>>> gpio@0
> >>>>>>>>
> >>>>>>>> Is this really a child block of the pinctrl? Doesn't really look like it
> >>>>>>>> based on addressess. Where are the pinctrl registers? In the sysgcr? If
> >>>>>>>> so, then pinctrl should be a child of it. But that doesn't really work
> >>>>>>>> too well with gpio child nodes...
> >>>>>>> the pin controller mux is handled by sysgcr this is why the sysgcr in
> >>>>>>> the mother node,
> >>>>>>> and the pin configuration are handled by the GPIO registers.  each
> >>>>>>> GPIO bank (child) contains 32 GPIO.
> >>>>>>> this is why the GPIO is the child node.
> >>>>>>
> >>>>>> Then maybe pinctrl should be the sysgcr and expose regmap for other devices?
> >>>>> The pin controller using the sysgcr to handle the pinmux, this is why
> >>>>> the sysgcr is in the mother node, is it problematic?
> >>>>
> >>>> You said pin-controller mux registers are in sysgcr, so it should not be
> >>>> used via syscon.
> >>> Sorry but maybe I missed something.
> >>> the sysgcr is used for miscellaneous features and not only for the pin
> >>> controller mux, this is why it used syscon and defined in the dtsi:
> >>>                 gcr: system-controller@f0800000 {
> >>>                         compatible = "nuvoton,npcm845-gcr", "syscon";
> >>>                         reg = <0x0 0xf0800000 0x0 0x1000>;
> >>>                 };
> >>>>
> >>>> Please provide address map description to convince us that this is
> >>>> correct HW representation.
> >>> GCR (sysgcr) registers 0xf0800000-0xf0801000 - used for miscellaneous
> >>> features, not only pin mux.
> >>> GPIO0 0xf0010000-0xf0011000
> >>> GPIO1 0xf0011000-0xf0012000
> >>> ...
> >>> GPIO7 0xf0017000-0xf0018000
> >>>>
> >>
> >> Then why your pinctrl is in sysgcr IO range? (pinctrl@f0800000)
> > you suggest using pinctrl@0 or pinctrl@f0010000 and not
> > pinctrl@f0800000 because 0xf0800000 is the GCR address that serve
> > miscellaneous features and not only pinmux controller ?
>
> If you have a map like you pasted, then DTS like this:
>
> syscon@f0800000 {}
> pinctrl@f0800000 {
>   gpio@f0010000 {}
> }
>
> Is quite weird, don't you think? You have two devices on the same unit
> address which is not allowed. You have child of pinctrl with entirely
O.K.
> different unit address, so how is it its child?
The pinctrl node name will modify the pinctrl@f0010000 the same as the
range property and the start of the child registers,is it fine?
> Best regards,
> Krzysztof

Best regards,

Tomer



[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux