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

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 20 Sep 2022 17:16:51 +0200
From:   Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To:     Tomer Maimon <tmaimon77@...il.com>
Cc:     Rob Herring <robh@...nel.org>,
        Avi Fishman <avifishman70@...il.com>,
        Tali Perry <tali.perry1@...il.com>,
        Joel Stanley <joel@....id.au>,
        Patrick Venture <venture@...gle.com>,
        Nancy Yuen <yuenn@...gle.com>,
        Benjamin Fair <benjaminfair@...gle.com>,
        Linus Walleij <linus.walleij@...aro.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Jonathan Neuschäfer <j.neuschaefer@....net>,
        zhengbin13@...wei.com, OpenBMC Maillist <openbmc@...ts.ozlabs.org>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        devicetree <devicetree@...r.kernel.org>
Subject: Re: [PATCH v2 1/2] dt-binding: pinctrl: Add NPCM8XX pinctrl and GPIO
 documentation

On 20/09/2022 11:27, Tomer Maimon wrote:
> On Tue, 20 Sept 2022 at 11:47, Krzysztof Kozlowski
> <krzysztof.kozlowski@...aro.org> wrote:
>>
>> On 20/09/2022 10:32, Tomer Maimon wrote:
>>> On Tue, 20 Sept 2022 at 11:21, Krzysztof Kozlowski
>>> <krzysztof.kozlowski@...aro.org> wrote:
>>>>
>>>> On 20/09/2022 09:59, Tomer Maimon wrote:
>>>>>>>>>>> +      pinctrl: pinctrl@...00000 {
>>>>>>>>>>> +        compatible = "nuvoton,npcm845-pinctrl";
>>>>>>>>>>> +        ranges = <0x0 0x0 0xf0010000 0x8000>;
>>>>>>>>>>> +        #address-cells = <1>;
>>>>>>>>>>> +        #size-cells = <1>;
>>>>>>>>>>> +        nuvoton,sysgcr = <&gcr>;
>>>>>>>>>>> +
>>>>>>>>>>> +        gpio0: gpio@...10000 {
>>>>>>>>>>
>>>>>>>>>> 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@...00000 {
>>>>>                         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@...00000)
>>> you suggest using pinctrl@0 or pinctrl@...10000 and not
>>> pinctrl@...00000 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@...00000 {}
>> pinctrl@...00000 {
>>   gpio@...10000 {}
>> }
>>
>> 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@...10000 the same as the
> range property and the start of the child registers,is it fine?

We are all busy, so I don't have that much bandwidth to review each of
your many solutions and instead poking me with every possible solution,
I would prefer if you think a bit how this all should work and look.

I don't know if it is fine. Why you should have two devices like this:
pinctrl@...10000 {
gpio@...10000 {}
}

???
Instead of one device? Answer such questions to yourself before asking
me. Please come with reasonable DTS describing the hardware.

Best regards,
Krzysztof