Re: [PATCH 12/12] arm64: dts: rockchip: add support for NanoPi M6 board

["Diederik de Haas" <diederik@cknow-tech.com>]

Hi Joachim,

On Fri Jul 3, 2026 at 7:14 PM CEST, Joachim Eastwood wrote:
> On Fri, 3 Jul 2026 at 13:34, Diederik de Haas <diederik@cknow-tech.com> wrote:
>> On Fri Jul 3, 2026 at 12:07 AM CEST, Joachim Eastwood via B4 Relay wrote:
>> > From: Joachim Eastwood <joachim.eastwood@gmail.com>
>> >
>> > The NanoPi M6 board shares most of the features of the R6 boards.
>> >
>> > Main differences:
>> > * M.2 M-key slot with PCIe (Also present on R6C)
>> > * M.2 E-key slot with PCIe and USB (from hub)
>> > * 1 additional USB 2.0 port from an on-board USB hub
>> > * RT5616 audio CODEC
>> >
>> > Signed-off-by: Joachim Eastwood <joachim.eastwood@gmail.com>
>> > ---
>>
>> [...]
>>
>> > +&i2c7 {
>> > +     clock-frequency = <200000>;
>> > +     status = "okay";
>> > +
>> > +     rt5616: codec@1b {
>> > +             compatible = "realtek,rt5616";
>> > +             reg = <0x1b>;
>> > +             clocks = <&cru I2S0_8CH_MCLKOUT>;
>> > +             clock-names = "mclk";
>> > +             #sound-dai-cells = <0>;
>> > +             assigned-clocks = <&cru I2S0_8CH_MCLKOUT>;
>>
>> s/I2S0_8CH_MCLKOUT/I2S0_8CH_MCLKOUT_TO_IO/ ?
>>
>> See these links for details:
>> https://lore.kernel.org/linux-rockchip/DJGDSS875DDO.22TYPVYK5X8KZ@cknow-tech.com/
>> https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git/commit/?h=v7.2-clk/fixes&id=946352b2f88fd2378f0341312e47dff1e8dc2fac
>
> Thanks for the feedback and pointers.
> I will use I2S0_8CH_MCLKOUT_TO_IO for the next version.
>
> Grepping through the other boards there doesn't seem to any other users
> of I2S0_8CH_MCLKOUT_TO_IO right now. But I assume the other boards
> will be converted over later(?)

That is indeed the idea; boards can(/should) be converted, ideally by someone
who has the board and can verify its result.

There's at least 'one' (or 3) being worked on ;-)
https://git.sr.ht/~diederik/linux-kernel/log/rk3588-nanopc-t6-plus

But for new boards, it's better to use the new _TO_IO variant.

Cheers,
  Diederik