MusePi Pro features a single RGMII Ethernet port with
PHY reset controlled by GPIO.
Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
---
arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
index 89a35914710bfd15c6ad27cd8d0de7ccf62e2309..e9a53dc7a417117c82f3e467677290bdaeffc845 100644
--- a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
+++ b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
@@ -47,6 +47,29 @@ &pdma {
status = "okay";
};
+ð0 {
+ phy-handle = <&rgmii0>;
+ phy-mode = "rgmii-id";
+ pinctrl-0 = <&gmac0_cfg>;
+ pinctrl-names = "default";
+ rx-internal-delay-ps = <0>;
+ tx-internal-delay-ps = <0>;
+ status = "okay";
+
+ mdio-bus {
+ #address-cells = <0x1>;
+ #size-cells = <0x0>;
+
+ reset-gpios = <&gpio K1_GPIO(110) GPIO_ACTIVE_LOW>;
+ reset-delay-us = <10000>;
+ reset-post-delay-us = <100000>;
+
+ rgmii0: phy@1 {
+ reg = <0x1>;
+ };
+ };
+};
+
&uart0 {
pinctrl-0 = <&uart0_2_cfg>;
pinctrl-names = "default";
--
2.51.0
On 10/10/25 09:56, Troy Mitchell wrote:
> MusePi Pro features a single RGMII Ethernet port with
> PHY reset controlled by GPIO.
>
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
> ---
> arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts | 23 +++++++++++++++++++++++
> 1 file changed, 23 insertions(+)
>
> diff --git a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> index 89a35914710bfd15c6ad27cd8d0de7ccf62e2309..e9a53dc7a417117c82f3e467677290bdaeffc845 100644
> --- a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> +++ b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> @@ -47,6 +47,29 @@ &pdma {
> status = "okay";
> };
>
> +ð0 {
For node orders in board DTS for SpacemiT devices, usually we follow
node name order because there's no unit address to go off of here, so
for what we have here it should be emmc, eth0, pdma, uart.
See e.g. arch/riscv/boot/dts/spacemit/k1-bananapi-f3.dts and previous
discussion [1].
Also, personally I think it would be fine to just add the whole
k1-musepi-pro.dts in the same patch instead of splitting it up like
this. I feel it would be easier to manage this way. For previous devices
these nodes have been added piecemeal mostly because the drivers were
being developed in parallel, but if you're going to base this on
v6.18-rc1 there shouldn't be dependency problems.
Maybe Yixun can chime in about this as well.
Thanks,
Vivian "dramforever" Wang
[1]: https://lore.kernel.org/linux-riscv/CAH1PCMa35n4dnJ94-EHGrJyU3kztQYO8_v2CkD=rid466zv3+A@mail.gmail.com
On 10/10/25 14:13, Vivian Wang wrote:
> On 10/10/25 09:56, Troy Mitchell wrote:
>> MusePi Pro features a single RGMII Ethernet port with
>> PHY reset controlled by GPIO.
>>
>> Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
>> ---
>> arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts | 23 +++++++++++++++++++++++
>> 1 file changed, 23 insertions(+)
>>
>> diff --git a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
>> index 89a35914710bfd15c6ad27cd8d0de7ccf62e2309..e9a53dc7a417117c82f3e467677290bdaeffc845 100644
>> --- a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
>> +++ b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
>> @@ -47,6 +47,29 @@ &pdma {
>> status = "okay";
>> };
>>
>> +ð0 {
> For node orders in board DTS for SpacemiT devices, usually we follow
> node name order because there's no unit address to go off of here, so
> for what we have here it should be emmc, eth0, pdma, uart.
>
> See e.g. arch/riscv/boot/dts/spacemit/k1-bananapi-f3.dts and previous
> discussion [1].
>
> Also, personally I think it would be fine to just add the whole
> k1-musepi-pro.dts in the same patch instead of splitting it up like
> this. I feel it would be easier to manage this way. For previous devices
> these nodes have been added piecemeal mostly because the drivers were
> being developed in parallel, but if you're going to base this on
> v6.18-rc1 there shouldn't be dependency problems.
... and also almost missed: this should have an ethernet0 alias. In the
most common boot flow it's read by U-Boot to set mac-address so that
both U-Boot and Linux use the same MAC address. This allows U-Boot to a
persistent MAC address rather than having to use a random one.
> Maybe Yixun can chime in about this as well.
>
> Thanks,
> Vivian "dramforever" Wang
>
> [1]: https://lore.kernel.org/linux-riscv/CAH1PCMa35n4dnJ94-EHGrJyU3kztQYO8_v2CkD=rid466zv3+A@mail.gmail.com
>
On Fri, Oct 10, 2025 at 02:20:50PM +0800, Vivian Wang wrote:
>
> On 10/10/25 14:13, Vivian Wang wrote:
> > On 10/10/25 09:56, Troy Mitchell wrote:
> >> MusePi Pro features a single RGMII Ethernet port with
> >> PHY reset controlled by GPIO.
> >>
> >> Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
> >> ---
> >> arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts | 23 +++++++++++++++++++++++
> >> 1 file changed, 23 insertions(+)
> >>
> >> diff --git a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> >> index 89a35914710bfd15c6ad27cd8d0de7ccf62e2309..e9a53dc7a417117c82f3e467677290bdaeffc845 100644
> >> --- a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> >> +++ b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> >> @@ -47,6 +47,29 @@ &pdma {
> >> status = "okay";
> >> };
> >>
> >> +ð0 {
> > For node orders in board DTS for SpacemiT devices, usually we follow
> > node name order because there's no unit address to go off of here, so
> > for what we have here it should be emmc, eth0, pdma, uart.
> >
> > See e.g. arch/riscv/boot/dts/spacemit/k1-bananapi-f3.dts and previous
> > discussion [1].
> >
> > Also, personally I think it would be fine to just add the whole
> > k1-musepi-pro.dts in the same patch instead of splitting it up like
> > this. I feel it would be easier to manage this way. For previous devices
> > these nodes have been added piecemeal mostly because the drivers were
> > being developed in parallel, but if you're going to base this on
> > v6.18-rc1 there shouldn't be dependency problems.
> ... and also almost missed: this should have an ethernet0 alias. In the
> most common boot flow it's read by U-Boot to set mac-address so that
> both U-Boot and Linux use the same MAC address. This allows U-Boot to a
> persistent MAC address rather than having to use a random one.
Yes, I forgot to add that.. I'll add it in the next version.
Thanks!
- Troy
> > Maybe Yixun can chime in about this as well.
> >
> > Thanks,
> > Vivian "dramforever" Wang
> >
> > [1]: https://lore.kernel.org/linux-riscv/CAH1PCMa35n4dnJ94-EHGrJyU3kztQYO8_v2CkD=rid466zv3+A@mail.gmail.com
> >
>
>
On Fri, Oct 10, 2025 at 02:13:31PM +0800, Vivian Wang wrote:
>
> On 10/10/25 09:56, Troy Mitchell wrote:
> > MusePi Pro features a single RGMII Ethernet port with
> > PHY reset controlled by GPIO.
> >
> > Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
> > ---
> > arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts | 23 +++++++++++++++++++++++
> > 1 file changed, 23 insertions(+)
> >
> > diff --git a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> > index 89a35914710bfd15c6ad27cd8d0de7ccf62e2309..e9a53dc7a417117c82f3e467677290bdaeffc845 100644
> > --- a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> > +++ b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> > @@ -47,6 +47,29 @@ &pdma {
> > status = "okay";
> > };
> >
> > +ð0 {
>
> For node orders in board DTS for SpacemiT devices, usually we follow
> node name order because there's no unit address to go off of here, so
> for what we have here it should be emmc, eth0, pdma, uart.
>
> See e.g. arch/riscv/boot/dts/spacemit/k1-bananapi-f3.dts and previous
> discussion [1].
Thanks for you point out!
>
> Also, personally I think it would be fine to just add the whole
> k1-musepi-pro.dts in the same patch instead of splitting it up like
> this. I feel it would be easier to manage this way. For previous devices
> these nodes have been added piecemeal mostly because the drivers were
> being developed in parallel, but if you're going to base this on
> v6.18-rc1 there shouldn't be dependency problems.
>
> Maybe Yixun can chime in about this as well.
We have talked here [1]
Personally, I think splitting them makes the commit history clearer.
I'm happy to merge them if others also think it's more appropriate.
Link: https://lore.kernel.org/linux-riscv/35C0EFBB84167959+aNj6K33fl7utuqcf@kernel.org/ [1]
- Troy
>
> Thanks,
> Vivian "dramforever" Wang
>
> [1]: https://lore.kernel.org/linux-riscv/CAH1PCMa35n4dnJ94-EHGrJyU3kztQYO8_v2CkD=rid466zv3+A@mail.gmail.com
>
>
On Fri, Oct 10, 2025 at 02:20:12PM +0800, Troy Mitchell wrote:
> On Fri, Oct 10, 2025 at 02:13:31PM +0800, Vivian Wang wrote:
> >
> > On 10/10/25 09:56, Troy Mitchell wrote:
> > > MusePi Pro features a single RGMII Ethernet port with
> > > PHY reset controlled by GPIO.
> > >
> > > Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
> > > ---
> > > arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts | 23 +++++++++++++++++++++++
> > > 1 file changed, 23 insertions(+)
> > >
> > > diff --git a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> > > index 89a35914710bfd15c6ad27cd8d0de7ccf62e2309..e9a53dc7a417117c82f3e467677290bdaeffc845 100644
> > > --- a/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> > > +++ b/arch/riscv/boot/dts/spacemit/k1-musepi-pro.dts
> > > @@ -47,6 +47,29 @@ &pdma {
> > > status = "okay";
> > > };
> > >
> > > +ð0 {
> >
> > For node orders in board DTS for SpacemiT devices, usually we follow
> > node name order because there's no unit address to go off of here, so
> > for what we have here it should be emmc, eth0, pdma, uart.
> >
> > See e.g. arch/riscv/boot/dts/spacemit/k1-bananapi-f3.dts and previous
> > discussion [1].
> Thanks for you point out!
>
> >
> > Also, personally I think it would be fine to just add the whole
> > k1-musepi-pro.dts in the same patch instead of splitting it up like
> > this. I feel it would be easier to manage this way. For previous devices
> > these nodes have been added piecemeal mostly because the drivers were
> > being developed in parallel, but if you're going to base this on
> > v6.18-rc1 there shouldn't be dependency problems.
> >
> > Maybe Yixun can chime in about this as well.
> We have talked here [1]
>
> Personally, I think splitting them makes the commit history clearer.
> I'm happy to merge them if others also think it's more appropriate.
tbh, I don't think it makes much sense to add it piecemeal, you'd then
end up potentially with dozens of commits for new platforms later in the
life of the SoC.
When you're adding new boards, add them in as complete as possible a
state from the beginning.
Hi Troy, Vivian: On 14:20 Fri 10 Oct , Troy Mitchell wrote: > On Fri, Oct 10, 2025 at 02:13:31PM +0800, Vivian Wang wrote: > > [snip].. > > > > Maybe Yixun can chime in about this as well. > We have talked here [1] > > Personally, I think splitting them makes the commit history clearer. > I'm happy to merge them if others also think it's more appropriate. > I do not have strong preference, either way is fine for me If you plan to split the DT into fine patches, then go ahead -- Yixun Lan (dlan)
Hi Troy, On 14:43 Fri 10 Oct , Yixun Lan wrote: > Hi Troy, Vivian: > > On 14:20 Fri 10 Oct , Troy Mitchell wrote: > > On Fri, Oct 10, 2025 at 02:13:31PM +0800, Vivian Wang wrote: > > > > [snip].. > > > > > > Maybe Yixun can chime in about this as well. > > We have talked here [1] > > > > Personally, I think splitting them makes the commit history clearer. > > I'm happy to merge them if others also think it's more appropriate. > > > I do not have strong preference, either way is fine for me > If you plan to split the DT into fine patches, then go ahead please squash the patches, see Conor's suggestion here: https://lore.kernel.org/all/20251010-confider-raven-0ad7a810e5de@spud/ Also, I'd plan to collect patches once v6.18-rc1 tagged, you can wait and do it after that.. -- Yixun Lan (dlan)
On Fri, Oct 10, 2025 at 02:43:10PM +0800, Yixun Lan wrote:
> Hi Troy, Vivian:
>
> On 14:20 Fri 10 Oct , Troy Mitchell wrote:
> > On Fri, Oct 10, 2025 at 02:13:31PM +0800, Vivian Wang wrote:
> > >
> [snip]..
> > >
> > > Maybe Yixun can chime in about this as well.
> > We have talked here [1]
> >
> > Personally, I think splitting them makes the commit history clearer.
> > I'm happy to merge them if others also think it's more appropriate.
> >
> I do not have strong preference, either way is fine for me
> If you plan to split the DT into fine patches, then go ahead
Thank you for your reconfirmation
- Troy
>
> --
> Yixun Lan (dlan)
>
© 2016 - 2025 Red Hat, Inc.