Re: [PATCH 2/2] arm64: dts: marvell: add support for Methode eDPU

[Robert Marko <robert.marko@sartura.hr>]

On Tue, May 10, 2022 at 1:46 PM Krzysztof Kozlowski
<krzysztof.kozlowski@linaro.org> wrote:
>
> On 10/05/2022 13:41, Robert Marko wrote:
> > On Tue, May 10, 2022 at 12:20 PM Krzysztof Kozlowski
> > <krzysztof.kozlowski@linaro.org> wrote:
> >>
> >> On 09/05/2022 13:00, Robert Marko wrote:
> >>> Methode eDPU is an Armada 3720 powered board based on the Methode uDPU.
> >>>
> >>> They feature the same CPU, RAM, and storage as well as the form factor.
> >>>
> >>> However, eDPU only has one SFP slot plus a copper G.hn port.
> >>>
> >>> In order to reduce duplication, split the uDPU DTS into a common one.
> >>>
> >>> Signed-off-by: Robert Marko <robert.marko@sartura.hr>
> >>> ---
> >>>  arch/arm64/boot/dts/marvell/Makefile          |   1 +
> >>>  .../boot/dts/marvell/armada-3720-eDPU.dts     |  14 ++
> >>>  .../boot/dts/marvell/armada-3720-uDPU.dts     | 148 +---------------
> >>>  .../boot/dts/marvell/armada-3720-uDPU.dtsi    | 163 ++++++++++++++++++
> >>>  4 files changed, 179 insertions(+), 147 deletions(-)
> >>>  create mode 100644 arch/arm64/boot/dts/marvell/armada-3720-eDPU.dts
> >>>  create mode 100644 arch/arm64/boot/dts/marvell/armada-3720-uDPU.dtsi
> >>>
> >>> diff --git a/arch/arm64/boot/dts/marvell/Makefile b/arch/arm64/boot/dts/marvell/Makefile
> >>> index 1c794cdcb8e6..104d7d7e8215 100644
> >>> --- a/arch/arm64/boot/dts/marvell/Makefile
> >>> +++ b/arch/arm64/boot/dts/marvell/Makefile
> >>> @@ -1,6 +1,7 @@
> >>>  # SPDX-License-Identifier: GPL-2.0
> >>>  # Mvebu SoC Family
> >>>  dtb-$(CONFIG_ARCH_MVEBU) += armada-3720-db.dtb
> >>> +dtb-$(CONFIG_ARCH_MVEBU) += armada-3720-eDPU.dtb
> >>>  dtb-$(CONFIG_ARCH_MVEBU) += armada-3720-espressobin.dtb
> >>>  dtb-$(CONFIG_ARCH_MVEBU) += armada-3720-espressobin-emmc.dtb
> >>>  dtb-$(CONFIG_ARCH_MVEBU) += armada-3720-espressobin-ultra.dtb
> >>> diff --git a/arch/arm64/boot/dts/marvell/armada-3720-eDPU.dts b/arch/arm64/boot/dts/marvell/armada-3720-eDPU.dts
> >>> new file mode 100644
> >>> index 000000000000..6b573a6854cc
> >>> --- /dev/null
> >>> +++ b/arch/arm64/boot/dts/marvell/armada-3720-eDPU.dts
> >>> @@ -0,0 +1,14 @@
> >>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
> >>> +
> >>> +/dts-v1/;
> >>> +
> >>> +#include "armada-3720-uDPU.dtsi"
> >>> +
> >>> +/ {
> >>> +     model = "Methode eDPU Board";
> >>> +     compatible = "methode,edpu", "marvell,armada3720";
> >>
> >> You need also bindings for the board compatible. Someone should convert
> >> the Documentation/devicetree/bindings/arm/marvell/armada-37xx.txt to YAML.
> >
> > Ok, I can convert the SoC compatibles at least for now.
> > Any advice you can give me on how the handle the Espressobin boards
> > having multiple board-specific compatibles?
> > For example, Espressobin V7 has:
> > "globalscale,espressobin-v7", "globalscale,espressobin"
> >
>
> Documentation/devicetree/bindings/arm/fsl.yaml

Thanks, now it makes sense.

>
> >>
> >>> +};
> >>> +> +  sfp_eth1: sfp-eth1 {
> >>
> >> Generic node names, please.
> >
> > Can you give me an example of what would be appropriate here because the SFP
> > bindings example utilizes the same naming scheme as used here?
>
> "sfp" if you have only one sfp node.

There are 2 SFP nodes in total, that is why they are named according
to the ethernet controller
to which they are connected.
uDPU has 2 SFP slots while eDPU has 1, so one was moved to uDPU DTS.

>
> >
> >>
> >>> +             compatible = "sff,sfp";
> >>> +             i2c-bus = <&i2c1>;
> >>> +             los-gpio = <&gpiosb 7 GPIO_ACTIVE_HIGH>;
> >>> +             mod-def0-gpio = <&gpiosb 8 GPIO_ACTIVE_LOW>;
> >>> +             tx-disable-gpio = <&gpiosb 9 GPIO_ACTIVE_HIGH>;
> >>> +             tx-fault-gpio = <&gpiosb 10 GPIO_ACTIVE_HIGH>;
> >>> +             maximum-power-milliwatt = <3000>;
> >>> +     };
> >>> +};
> >>> +
> >>> +&sdhci0 {
> >>> +     status = "okay";
> >>> +     bus-width = <8>;
> >>> +     mmc-ddr-1_8v;
> >>> +     mmc-hs400-1_8v;
> >>> +     marvell,pad-type = "fixed-1-8v";
> >>> +     non-removable;
> >>> +     no-sd;
> >>> +     no-sdio;
> >>> +};
> >>> +
> >>> +&spi0 {
> >>> +     status = "okay";
> >>> +     pinctrl-names = "default";
> >>> +     pinctrl-0 = <&spi_quad_pins>;
> >>> +
> >>> +     spi-flash@0 {
> >>
> >> Run dtbs_check and fix the errors.
> >
> > Ok, will split the DTSI and eDPU commits and fixup nodes in between.
> >>
> >>> +             compatible = "jedec,spi-nor";
> >>> +             reg = <0>;
> >>> +             spi-max-frequency = <54000000>;
> >>> +
> >>> +             partitions {
> >>> +                     compatible = "fixed-partitions";
> >>> +                     #address-cells = <1>;
> >>> +                     #size-cells = <1>;
> >>> +                     /* only bootloader is located on the SPI */
> >>> +                     partition@0 {
> >>> +                             label = "firmware";
> >>> +                             reg = <0x0 0x180000>;
> >>> +                     };
> >>> +
> >>> +                     partition@180000 {
> >>> +                             label = "u-boot-env";
> >>> +                             reg = <0x180000 0x10000>;
> >>> +                     };
> >>> +             };
> >>> +     };
> >>> +};
> >>> +
> >>> +&pinctrl_nb {
> >>> +     i2c2_recovery_pins: i2c2-recovery-pins {
> >>> +             groups = "i2c2";
> >>> +             function = "gpio";
> >>> +     };
> >>> +};
> >>> +
> >>> +&i2c1 {
> >>> +     status = "okay";
> >>> +     pinctrl-names = "default", "recovery";
> >>> +     pinctrl-0 = <&i2c2_pins>;
> >>> +     pinctrl-1 = <&i2c2_recovery_pins>;
> >>> +     /delete-property/mrvl,i2c-fast-mode;
> >>> +     scl-gpios = <&gpionb 2 (GPIO_ACTIVE_HIGH | GPIO_OPEN_DRAIN)>;
> >>> +     sda-gpios = <&gpionb 3 (GPIO_ACTIVE_HIGH | GPIO_OPEN_DRAIN)>;
> >>> +
> >>> +     nct375@48 {
> >>
> >> Generic node names, representing class of a device.
> > Ok, will rename in v2.
> >>
> >>> +             status = "okay";
> >>
> >> OK status is by default, why do you need it? Also, it goes as last property.
> >
> > It's not needed, I have not changed any nodes, they are just
> > copy/paste during the DTS split.
> > Will drop it in v2.
> >
>
> Hm, but the node names were different in original DTS, so this is not a
> simple split. In such case better to correct coding styles in one patch
> (node names, status etc) and then perform the split. The split should
> create the same output DTB, which is not the case here.

Understood, I did all of the fixups now before the split to make it clear.

Regards,
Robert
>
> Best regards,
> Krzysztof



-- 
Robert Marko
Staff Embedded Linux Engineer
Sartura Ltd.
Lendavska ulica 16a
10000 Zagreb, Croatia
Email: robert.marko@sartura.hr
Web: www.sartura.hr