Re: [PATCH] arm64: dts: meson: add SM1 soundcard name to VIM3L

From: Kevin Hilman
Date: Fri Oct 02 2020 - 14:45:37 EST


Christian Hewitt <christianshewitt@xxxxxxxxx> writes:

>> On 2 Oct 2020, at 6:44 pm, Jerome Brunet <jbrunet@xxxxxxxxxxxx> wrote:
>>
>> On Fri 02 Oct 2020 at 16:16, Christian Hewitt <christianshewitt@xxxxxxxxx> wrote:
>>
>>> VIM3L now inherits the sound node from the VIM3 common dtsi but is
>>> an SM1 device, so label it as such, and stop users blaming future
>>> support issues on the distro/app "wrongly detecting" their device.
>>>
>>> Signed-off-by: Christian Hewitt <christianshewitt@xxxxxxxxx>
>>> ---
>>> arch/arm64/boot/dts/amlogic/meson-sm1-khadas-vim3l.dts | 4 ++++
>>> 1 file changed, 4 insertions(+)
>>>
>>> diff --git a/arch/arm64/boot/dts/amlogic/meson-sm1-khadas-vim3l.dts b/arch/arm64/boot/dts/amlogic/meson-sm1-khadas-vim3l.dts
>>> index 4b517ca72059..f46f0ecc37ec 100644
>>> --- a/arch/arm64/boot/dts/amlogic/meson-sm1-khadas-vim3l.dts
>>> +++ b/arch/arm64/boot/dts/amlogic/meson-sm1-khadas-vim3l.dts
>>> @@ -32,6 +32,10 @@
>>> regulator-boot-on;
>>> regulator-always-on;
>>> };
>>> +
>>> + sound {
>>> + model = "SM1-KHADAS-VIM3L";
>>> + };
>>
>> The sound card is the same so I don't see why the sm1 board should have
>> a different name. If you are not happy with the name, please update it
>> in the common file.
>
> It’s a nice-to-have not a must-have, but the current LE images that are
> in circulation use 5.7 with the previous board-correct name so I was
> looking for continuity. We do see user forum reports (infrequent but
> recurring) of wrongly detected hardware with other SoC platforms where
> similar name inheritance surfaces the ‘wrong’ device name in GUIs, and
> I like anything that avoids support work.
>
> I’d suggest KHADAS-VIM3-VIM3L as a common name, but then it’s the only
> device in the current device-tree set that is not prefixed with the SoC
> identifier, which (OCD) feels wrong.

True, but turns out there's nothing SoC specific about this sound block
since it's identical across SoCs, so specifying the SoC is being too
specific.

OTOH, while I agree it looks "wrong", it's pretty common in Linux DT to
have the SoC prefix to mean only that it's "compatible" with that SoC,
not that it *is* that SoC.

However, I agree that that can lead to confusion with end users, so
since this change has not functional change, and only a UX issue in
userspace, I'm fine to apply it.

Kevin