Re: [PATCH v3 2/2] arm64: dts: allwinner: h6: Add Orange Pi 3 LTS

From: Chen-Yu Tsai

Date: Mon Nov 16 2026 - 11:57:18 EST


On Tue, Nov 17, 2026 at 12:44 AM Cristian Filipescu
<filipescu.cristian@xxxxxxxxx> wrote:
>
> On Mon, Nov 16, 2026 at 7:38 AM Chen-Yu Tsai <wens@xxxxxxxxxx> wrote:
> >
> > On Mon, Nov 16, 2026 at 7:47 AM Cristian Filipescu
> > <filipescu.cristian@xxxxxxxxx> wrote:
> > >
> > > Hi Chen-Yu,
> > >
> > > On Sun, Nov 15, 2026 at 3:14 AM Chen-Yu Tsai <wens@xxxxxxxxxx> wrote:
> > > >
> > > > On Tue, Oct 27, 2026 at 7:34 AM Cristian Filipescu via B4 Relay
> > > > <devnull+filipescu.cristian.gmail.com@xxxxxxxxxx> wrote:
> > > > >
> > > > > From: Jernej Skrabec <jernej.skrabec@xxxxxxxxx>
> > > > >
> > > > > Add the board description for the Allwinner H6-based Orange Pi 3 LTS,
> > > > > including its regulators, storage, Ethernet PHY, LEDs and USB ports.
> > > > >
> > > > > Specify the Motorcomm YT8531C PHY ID to avoid early ID autodetection,
> > > > > which failed on MAC driver rebind. Reset the sole PHY at MDIO bus
> > > > > registration and keep reset outside interface detach so its configured
> > > > > 125 MHz clock output is retained. Use PHY-internal RGMII delays.
> > > > >
> > > > > Select the weakest GPIO drive level for both LEDs. Enable 1.8 V eMMC
> > > > > DDR52 and HS200 modes; retain the H6 maximum frequency of 150 MHz.
> > > > >
> > > > > Link: https://lore.kernel.org/linux-sunxi/20250413134318.66681-3-jernej.skrabec@xxxxxxxxx/
> > > > > Signed-off-by: Jernej Skrabec <jernej.skrabec@xxxxxxxxx>
> > > > > Signed-off-by: Cristian Filipescu <filipescu.cristian@xxxxxxxxx>
> > > > > ---
> > > > > arch/arm64/boot/dts/allwinner/Makefile | 1 +
> > > > > .../dts/allwinner/sun50i-h6-orangepi-3-lts.dts | 379 +++++++++++++++++++++
> > > > > 2 files changed, 380 insertions(+)
> > > > >
> > > > > diff --git a/arch/arm64/boot/dts/allwinner/Makefile b/arch/arm64/boot/dts/allwinner/Makefile
> > > > > index aa21f58a4..7ea136380 100644
> > > > > --- a/arch/arm64/boot/dts/allwinner/Makefile
> > > > > +++ b/arch/arm64/boot/dts/allwinner/Makefile
> > > > > @@ -41,6 +41,7 @@ dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h5-orangepi-zero-plus.dtb
> > > > > dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h5-orangepi-zero-plus2.dtb
> > > > > dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h6-beelink-gs1.dtb
> > > > > dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h6-orangepi-3.dtb
> > > > > +dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h6-orangepi-3-lts.dtb
> > > > > dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h6-orangepi-lite2.dtb
> > > > > dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h6-orangepi-one-plus.dtb
> > > > > dtb-$(CONFIG_ARCH_SUNXI) += sun50i-h6-pine-h64.dtb
> > > > > diff --git a/arch/arm64/boot/dts/allwinner/sun50i-h6-orangepi-3-lts.dts b/arch/arm64/boot/dts/allwinner/sun50i-h6-orangepi-3-lts.dts
> > > > > new file mode 100644
> > > > > index 000000000..e2ca41b94
> > > > > --- /dev/null
> > > > > +++ b/arch/arm64/boot/dts/allwinner/sun50i-h6-orangepi-3-lts.dts
> > > > > @@ -0,0 +1,379 @@
> > > > > +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
> > > > > +// Copyright (C) 2025 Jernej Skrabec <jernej.skrabec@xxxxxxxxx>
> > > > > +// Based on sun50i-h6-orangepi-3.dts, which is:
> > > > > +// Copyright (C) 2019 Ondřej Jirman <megous@xxxxxxxxxx>
> > > >
> > > > We now have a common sun50i-h6-orangepi.dtsi file. You really should
> > > > rewrite this patch to just use that as a base. It is mostly the same
> > > > across all the OrangePi H6 boards.
> > >
> > > v4 now uses that include and keeps the LTS-specific additions
> > > and overrides. I compared the resolved DTBs to check the inherited
> > > properties and preserved the LTS electrical settings, except for the
> > > review changes described below. Common naming cleanups are deferred.
> >
> > [...]
> >
> > > > Set the supply on all cores.
> > >
> > > Done. All four running CPU nodes reference DCDCA, and all four CPUs are
> > > online with the shared cpufreq policy.
> >
> > I should've mentioned that since this is already enabled at the .dtsi
> > level, you shouldn't add it anymore. If anything, it's the .dtsi file
> > that should have all cores referencing the regulator.
>
> Understood on the CPU supply. I have rebased onto sunxi/dt-for-7.5 at
> 098728144545 and removed the board-local CPU0-CPU3 supply declarations
> and CPU/GPU OPP includes. These are now inherited from the common
> Orange Pi H6 description. I also removed the redundant ramp-delay and
> CPU maximum-voltage overrides, keeping the existing LTS 800 mV minimum.
> Any all-core supply cleanup belongs in the common include.
>
> >
> > [...]
> >
> > > > > +&mmc1 {
> > > > > + vmmc-supply = <&reg_wifi_3v3>;
> > > > > + vqmmc-supply = <&reg_bldo3>;
> > > >
> > > > Something is fishy here. The schematic also shows vcc-wifi-io (on the aw859
> > > > side) directly derived from vcc-wifi (reg_wifi_3v3). I suggest at least
> > > > checking with a multimeter to confirm what the voltage is on the VDDIO
> > > > pin of the module.
> > > >
> > >
> > > Regarding the AW859A I/O supply, I traced U60 pin 22 (VDDIO) to
> > > VCC-WIFI-IO, which is connected to BLDO3 in the schematic. I measured
> > > 1.8 V at an RW1 pad relative to board ground. This supports keeping the
> > > existing BLDO3 vqmmc assignment; I have left it unchanged.
> >
> > You traced it on the board or on the schematic?
> >
> > Can you test at U60? The schematic has two options for I/O voltage from
> > the LDOs, and then another from VCC-WIFI, hence I would like to be
> > certain. I imagine the SMT components aren't labeled on the board, so
> > it would be somewhat hard / eye-hurting to trace the board directly.
> >
>
> The Wi-Fi connection trace was from the schematic. Initially I was worried I
> would short some pin, but I see pin 22 is at the edge. I have now measured
> 1.8 V directly at U60/AW859A pin 22 (VDDIO), relative to board ground after
> the kernel boots.

Cool! Thanks for confirming.


ChenYu

> > [...]
> >
> > > > > +&usb2otg {
> > > > > + dr_mode = "host";
> > > > > + status = "okay";
> > > > > +};
> > > > > +
> > > > > +&usb2phy {
> > > > > + usb0_id_det-gpios = <&pio 2 15 GPIO_ACTIVE_HIGH>; /* PC15 */
> > > >
> > > > This isn't connected to anything but a pull-up. And you already specified
> > > > host mode only above.
> > >
> > > Removed usb0_id_det-gpios with /delete-property/ so the common include's
> > > PC6 setting is removed too. USB0 remains fixed to host mode. The running
> > > DT confirms the ID GPIO property is absent and dr_mode is host.
> >
> > Thanks!
> > ChenYu
>
>
> Regards,
> Cristian Filipescu
> ------------------------------------