Re: [PATCH v2 3/7] arm64: dts: qcom: lemans: Move platform resources into lemans-embedded.dtsi

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] dts: Incorrect removal of 'compatible' from base DTSI nodes
--

commit f8600c852f6da5600fb82650c04f9b466cb80549
Author: Shawn Guo <shengchao.guo@oss.qualcomm.com>

arm64: dts: qcom: lemans: Move platform resources into lemans-embedded.dtsi

This commit moves platform resources like clocks, regulators, and interconnects
from the base lemans.dtsi to the lemans-embedded.dtsi file. This isolates
these managed configurations so the base devicetree can be shared with the
GearVM variant.

> diff --git a/arch/arm64/boot/dts/qcom/lemans.dtsi b/arch/arm64/boot/dts/qcom/lemans.dtsi
> index 7547042d8e851..821c86409c5da 100644
> --- a/arch/arm64/boot/dts/qcom/lemans.dtsi
> +++ b/arch/arm64/boot/dts/qcom/lemans.dtsi
[ ... ]
> @@ -4662,49 +3089,9 @@ gpu: gpu@3d00000 {
[ ... ]
>  		gmu: gmu@3d6a000 {
> -			compatible = "qcom,adreno-gmu-663.0", "qcom,adreno-gmu";

[Severity: Medium]
Does this leave the GMU node structurally incomplete for other variants?

A compatible string is typically a fundamental description of the hardware IP
rather than a dynamically managed platform resource. Removing it from the base
lemans.dtsi means that any environment (like the incoming GearVM variant)
including this file will inherit a node that has "reg" and "interrupts" but
lacks a compatible string.

Will this break devicetree schema validation and prevent proper driver binding
for variants that do not include lemans-embedded.dtsi?

>  			reg = <0x0 0x03d6a000 0x0 0x34000>,
>  			      <0x0 0x03de0000 0x0 0x10000>,
>  			      <0x0 0x0b290000 0x0 0x10000>;
[ ... ]
> -		pmu@9091000 {
> -			compatible = "qcom,sa8775p-llcc-bwmon", "qcom,sc7280-llcc-bwmon";
> +		llcc_bwmon: pmu@9091000 {

[Severity: Medium]
Similarly, does removing the compatible string for the llcc_bwmon node here
leave it incomplete in the base devicetree?

>  			reg = <0x0 0x9091000 0x0 0x1000>;
>  			interrupts = <GIC_SPI 620 IRQ_TYPE_LEVEL_HIGH>;
[ ... ]
> -		pmu@90b5400 {
> -			compatible = "qcom,sa8775p-cpu-bwmon", "qcom,sdm845-bwmon";
> +		cpu_bwmon0: pmu@90b5400 {

[Severity: Medium]
Does this also incorrectly strip the hardware description from the cpu_bwmon0
node for environments relying only on lemans.dtsi?

>  			reg = <0x0 0x90b5400 0x0 0x600>;
>  			interrupts = <GIC_SPI 581 IRQ_TYPE_LEVEL_HIGH>;
[ ... ]
> -		pmu@90b6400 {
> -			compatible = "qcom,sa8775p-cpu-bwmon", "qcom,sdm845-bwmon";
> +		cpu_bwmon1: pmu@90b6400 {

[Severity: Medium]
Does this also incorrectly strip the hardware description from the cpu_bwmon1
node for environments relying only on lemans.dtsi?

>  			reg = <0x0 0x90b6400 0x0 0x600>;
>  			interrupts = <GIC_SPI 581 IRQ_TYPE_LEVEL_HIGH>;
[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261015023711.721837-1-shengchao.guo@oss.qualcomm.com?part=3