[PATCH] riscv: dts: spacemit: k3: add USB3 port A (DRD) nodes

[Chunyan Zhang]

Hi Aurelien,

On Sun, 1 Nov 2026 at 02:19, Aurelien Jarno <aurelien at aurel32.net> wrote:
>
> Hi,
>
> On 2026-10-30 17:40, Chunyan Zhang wrote:
> > Describe the DWC3 controller and its USB 2.0 PHY for USB3 port A
> > of the K3 SoC.
> >
> > Only the U2 lane is described: Port A's SuperSpeed PHY has no driver
> > yet, so the port runs at high-speed (maximum-speed = "high-speed").
> >
> > Enable USB3 port A with dr_mode = "peripheral", so it can be used as
> > a USB device.
> >
> > Signed-off-by: Chunyan Zhang <chunyan.zhang at linux.spacemit.com>
> > ---
> >  arch/riscv/boot/dts/spacemit/k3-pico-itx.dts |  9 ++++++
> >  arch/riscv/boot/dts/spacemit/k3.dtsi         | 30 ++++++++++++++++++++
> >  2 files changed, 39 insertions(+)
> >
> > diff --git a/arch/riscv/boot/dts/spacemit/k3-pico-itx.dts b/arch/riscv/boot/dts/spacemit/k3-pico-itx.dts
> > index b084a35aa00e..74d4fdf9e405 100644
> > --- a/arch/riscv/boot/dts/spacemit/k3-pico-itx.dts
> > +++ b/arch/riscv/boot/dts/spacemit/k3-pico-itx.dts
> > @@ -323,6 +323,15 @@ &usb2_phy {
> >       status = "okay";
> >  };
> >
> > +&usb3a_u2phy {
> > +     status = "okay";
> > +};
> > +
> > +&usb3a {
> > +     dr_mode = "peripheral";
> > +     status = "okay";
> > +};
> > +
> >  &usb3d_u2phy {
> >       status = "okay";
> >  };
> > diff --git a/arch/riscv/boot/dts/spacemit/k3.dtsi b/arch/riscv/boot/dts/spacemit/k3.dtsi
> > index c3f2dce0969c..220dfe29bb78 100644
> > --- a/arch/riscv/boot/dts/spacemit/k3.dtsi
> > +++ b/arch/riscv/boot/dts/spacemit/k3.dtsi
> > @@ -630,6 +630,36 @@ gmac2_axi_setup: stmmac-axi-config {
> >                       };
> >               };
> >
> > +             usb3a: usb at cad00000 {
>
> Please move this after the eth2 node, so that the node addresses are
> kept sorted.
>
> > +                     compatible = "spacemit,k3-dwc3";
> > +                     reg = <0x0 0xcad00000 0x0 0x10000>;
> > +                     interrupts = <118 IRQ_TYPE_LEVEL_HIGH>;
> > +                     interrupt-parent = <&saplic>;
> > +                     clocks = <&syscon_apmu CLK_APMU_USB3_PORTA_BUS>;
> > +                     clock-names = "usbdrd30";
> > +                     resets = <&syscon_apmu RESET_APMU_USB3_A_AHB>,
> > +                              <&syscon_apmu RESET_APMU_USB3_A_VCC>,
> > +                              <&syscon_apmu RESET_APMU_USB3_A_PHY>;
> > +                     reset-names = "ahb", "vcc", "phy";
> > +                     phys = <&usb3a_u2phy>;
> > +                     phy-names = "usb2-phy";
>
> Please move the PHYs definition to the board level, matching was it is
> done for the other USB3 ports, as other boards might have different PHY
> configuration.
>
> > +                     phy_type = "utmi";
> > +                     snps,dis_enblslpm_quirk;
> > +                     snps,dis_u2_susphy_quirk;
> > +                     snps,dis-del-phy-power-chg-quirk;
> > +                     snps,dis-tx-ipgap-linecheck-quirk;
> > +                     maximum-speed = "high-speed";
>
> I think the maximum-speed property should also be moved to the board
> level, as its value depends on the PHYs configuration.

Sure. I will address all the above comments.

>
> > +                     status = "disabled";
> > +             };
>
> Any reason not to define the snps,parkmode-disable-ss-quirk property? It
> is defined for the other USB3 ports, as well as for the USB3 A port in
> the vendor kernel

The original thought was that this property is SS-specific and thus unnecessary
without usb3-phy support. However, since it reflects a hardware characteristic
that applies regardless of the current PHY configuration, adding it makes sense
for consistency and forward compatibility.

>
> > +
> > +             usb3a_u2phy: phy at cad20000 {
> > +                     compatible = "spacemit,k3-usb2-phy";
> > +                     reg = <0x0 0xcad20000 0x0 0x200>;
> > +                     clocks = <&syscon_apmu CLK_APMU_USB3_PORTA_BUS>;
> > +                     #phy-cells = <0>;
> > +                     status = "disabled";
> > +             };
> > +
> >               pdma: dma-controller at d4000000 {
> >                       compatible = "spacemit,k3-pdma";
> >                       reg = <0x0 0xd4000000 0x0 0x4000>;
>
> Otherwise looks good.

Thanks for the review,
Chunyan