Re: [PATCH v1 6/6] ARM: dts: lan966x: add basic Kontron KSwitch D10 support

[<Claudiu.Beznea@microchip.com>]

On 07.03.2022 14:17, Michael Walle wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the
> content is safe
> 
> Am 2022-03-07 13:07, schrieb Claudiu.Beznea@microchip.com:
>> On 04.03.2022 13:15, Michael Walle wrote:
>>> EXTERNAL EMAIL: Do not click links or open attachments unless you know
>>> the
>>> content is safe
>>>
>>> Am 2022-03-04 09:31, schrieb Claudiu.Beznea@microchip.com:
>>>> On 03.03.2022 18:03, Michael Walle wrote:
>>>>> EXTERNAL EMAIL: Do not click links or open attachments unless you
>>>>> know
>>>>> the content is safe
>>>>>
>>>>> Add basic support for the Kontron KSwitch D10 MMT 6G-2GS which
>>>>> features 6 Gigabit copper ports and two SFP cages. For now the
>>>>> following is working:
>>>>>  - Kernel console
>>>>>  - SFP cages I2C bus and mux
>>>>>  - SPI
>>>>>  - SGPIO
>>>>>  - Watchdog
>>>>>
>>>>> Signed-off-by: Michael Walle <michael@walle.cc>
>>>>> ---
>>>>>  arch/arm/boot/dts/Makefile                    |   3 +-
>>>>>  ...lan966x-kontron-kswitch-d10-mmt-6g-2gs.dts | 159
>>>>> ++++++++++++++++++
>>>>>  2 files changed, 161 insertions(+), 1 deletion(-)
>>>>>  create mode 100644
>>>>> arch/arm/boot/dts/lan966x-kontron-kswitch-d10-mmt-6g-2gs.dts
>>>>>
>>>>> diff --git a/arch/arm/boot/dts/Makefile b/arch/arm/boot/dts/Makefile
>>>>> index 085c43649d44..86dd0f9804ee 100644
>>>>> --- a/arch/arm/boot/dts/Makefile
>>>>> +++ b/arch/arm/boot/dts/Makefile
>>>>> @@ -739,7 +739,8 @@ dtb-$(CONFIG_SOC_IMX7ULP) += \
>>>>>         imx7ulp-com.dtb \
>>>>>         imx7ulp-evk.dtb
>>>>>  dtb-$(CONFIG_SOC_LAN966) += \
>>>>> -       lan966x-pcb8291.dtb
>>>>> +       lan966x-pcb8291.dtb \
>>>>> +       lan966x-kontron-kswitch-d10-mmt-6g-2gs.dtb
>>>>>  dtb-$(CONFIG_SOC_LS1021A) += \
>>>>>         ls1021a-moxa-uc-8410a.dtb \
>>>>>         ls1021a-qds.dtb \
>>>>> diff --git
>>>>> a/arch/arm/boot/dts/lan966x-kontron-kswitch-d10-mmt-6g-2gs.dts
>>>>> b/arch/arm/boot/dts/lan966x-kontron-kswitch-d10-mmt-6g-2gs.dts
>>>>> new file mode 100644
>>>>> index 000000000000..958678dec7ad
>>>>> --- /dev/null
>>>>> +++ b/arch/arm/boot/dts/lan966x-kontron-kswitch-d10-mmt-6g-2gs.dts
>>>>> @@ -0,0 +1,159 @@
>>>>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
>>>>> +/*
>>>>> + * Device Tree file for the Kontron KSwitch D10 MMT 6G-2GS
>>>>> + */
>>>>> +
>>>>> +/dts-v1/;
>>>>> +#include "lan966x.dtsi"
>>>>> +
>>>>> +/ {
>>>>> +       model = "Kontron KSwitch D10 MMT 6G-2GS";
>>>>> +       compatible = "kontron,kswitch-d10-mmt-6g-2gs",
>>>>> "kontron,s1921",
>>>>> +                    "microchip,lan9668", "microchip,lan966";
>>>>> +
>>>>> +       aliases {
>>>>> +               serial0 = &usart0;
>>>>> +       };
>>>>> +
>>>>> +       chosen {
>>>>> +               stdout-path = "serial0:115200n8";
>>>>> +       };
>>>>> +
>>>>> +       gpio-restart {
>>>>> +               compatible = "gpio-restart";
>>>>> +               gpios = <&gpio 56 GPIO_ACTIVE_LOW>;
>>>>> +               priority = <200>;
>>>>> +       };
>>>>> +
>>>>> +       i2cmux {
>>>>> +               compatible = "i2c-mux-gpio";
>>>>> +               #address-cells = <1>;
>>>>> +               #size-cells = <0>;
>>>>> +               mux-gpios = <&sgpio_out 3 2 GPIO_ACTIVE_HIGH>,
>>>>> +                           <&sgpio_out 3 3 GPIO_ACTIVE_HIGH>;
>>>>> +               i2c-parent = <&i2c4>;
>>>>> +
>>>>> +               i2c4_0: i2c@1 {
>>>>> +                       reg = <1>;
>>>>> +                       #address-cells = <1>;
>>>>> +                       #size-cells = <0>;
>>>>> +               };
>>>>> +
>>>>> +               i2c4_1: i2c@2 {
>>>>> +                       reg = <2>;
>>>>> +                       #address-cells = <1>;
>>>>> +                       #size-cells = <0>;
>>>>> +               };
>>>>> +       };
>>>>> +
>>>>> +       sfp0: sfp0 {
>>>>> +               compatible = "sff,sfp";
>>>>> +               i2c-bus = <&i2c4_0>;
>>>>> +               los-gpios = <&sgpio_in 1 0 GPIO_ACTIVE_HIGH>;
>>>>> +               mod-def0-gpios = <&sgpio_in 1 1 GPIO_ACTIVE_LOW>;
>>>>> +               maximum-power-milliwatt = <2500>;
>>>>> +               tx-disable-gpios = <&sgpio_out 3 0 GPIO_ACTIVE_LOW>;
>>>>> +               tx-fault-gpios = <&sgpio_in 0 2 GPIO_ACTIVE_HIGH>;
>>>>> +               rate-select0-gpios = <&sgpio_out 2 0
>>>>> GPIO_ACTIVE_HIGH>;
>>>>> +               rate-select1-gpios = <&sgpio_out 2 1
>>>>> GPIO_ACTIVE_HIGH>;
>>>>> +       };
>>>>> +
>>>>> +       sfp1: sfp1 {
>>>>> +               compatible = "sff,sfp";
>>>>> +               i2c-bus = <&i2c4_1>;
>>>>> +               los-gpios = <&sgpio_in 1 2 GPIO_ACTIVE_HIGH>;
>>>>> +               mod-def0-gpios = <&sgpio_in 1 3 GPIO_ACTIVE_LOW>;
>>>>> +               maximum-power-milliwatt = <2500>;
>>>>> +               tx-disable-gpios = <&sgpio_out 3 1 GPIO_ACTIVE_LOW>;
>>>>> +               tx-fault-gpios = <&sgpio_in 0 3 GPIO_ACTIVE_HIGH>;
>>>>> +               rate-select0-gpios = <&sgpio_out 2 2
>>>>> GPIO_ACTIVE_HIGH>;
>>>>> +               rate-select1-gpios = <&sgpio_out 2 3
>>>>> GPIO_ACTIVE_HIGH>;
>>>>> +       };
>>>>> +};
>>>>> +
>>>>> +&flx0 {
>>>>> +       atmel,flexcom-mode = <ATMEL_FLEXCOM_MODE_USART>;
>>>>> +       status = "okay";
>>>>> +};
>>>>> +
>>>>> +&flx3 {
>>>>> +       atmel,flexcom-mode = <ATMEL_FLEXCOM_MODE_SPI>;
>>>>> +       status = "okay";
>>>>> +};
>>>>> +
>>>>> +&flx4 {
>>>>> +       atmel,flexcom-mode = <ATMEL_FLEXCOM_MODE_TWI>;
>>>>> +       status = "okay";
>>>>> +};
>>>>
>>>> Although there is 1:1 mapping b/w ids of flexcoms and the embedded
>>>> blocks
>>>> (flxX has usartX, i2cX, spiX) and there is nothing wrong with the
>>>> approach
>>>> here I found a bit hard to follow if the correspondent embedded block
>>>> (i2c, spi, usart) is enabled or not.
>>>
>>> I know and I had the same feeling, but I don't want to have the
>>> subnodes (matched by name) in these nodes.  I.e. I want to avoid
>>> something like:
>>>
>>> &flx4 {
>>>        atmel,flexcom-mode = <ATMEL_FLEXCOM_MODE_TWI>;
>>>        status = "okay";
>>>
>>>        i2c@600 {
>>>               pinctrl-0 = <&fc4_b_pins>;
>>>               pinctrl-names = "default";
>>>               status = "okay";
>>>        };
>>> };
>>
>> All the other AT91 DTs are using the above format + the specific label
>> in
>> front of the subnode, e.g:
>>
>>       i2cX: i2c@600 {
>>
>>>
>>> If someone renames the subnode in the dtsi, it might easily be
>>> overlooked in the board files. Having the handle will raise an
>>> error.
>>
>> If using label + node name as pointed above there will be an error
>> thrown
>> for your scenario.
> 
> Fair enough. You'll get a duplicate reference error as long as
> the reference itself isn't renamed either.
> 
> But to make it short, unless you force me too, I'd like
> to keep the child node as is and not as a subnode of
> the flexcom. I just don't want to repeat the name if
> there is no reason and I live with the fact that they
> are not near each other :)

To me it easy to follow with the current approach adopted by all AT91
targets. If you know that all flexcomX subnodes have the X in their naming
(e.g. i2c3, spi3, uart3 are childs of flexcom3) then its easy to follow it
also. But 1st time opening the dts file you probably don't know this.

Thank you,
Claudiu Beznea

> 
> That flexcom-mode could also be deduced from the enabled
> children, btw.
> 
>>> And because the node references should be sorted alphabetically
>>> it will be cluttered around in the file. You could rename the
>>> references to flx4_i2c though. But I don't know it its worth
>>> the efforts. Let me know what you think.
> 
> -michael