On 27-03-23, 10:49, Konrad Dybcio wrote:
> On 27.03.2023 07:38, Vinod Koul wrote:
> > On 25-03-23, 13:34, Konrad Dybcio wrote:
> >>> + cpus {
> >>> + #address-cells = <2>;
> >>> + #size-cells = <0>;
> >>> +
> >>> + CPU0: cpu@0 {
> >>> + device_type = "cpu";
> >>> + compatible = "qcom,kryo485";
> >>> + reg = <0x0 0x0>;
> >> Please add clocks = <&cpufreq_hw n>;
> >>
> >>> + enable-method = "psci";
> >>> + capacity-dmips-mhz = <602>;
> >>> + next-level-cache = <&L2_0>;
> >>> + qcom,freq-domain = <&cpufreq_hw 0>;
> >
> > You mean this or something else?
> Check
>
> https://lore.kernel.org/lkml/20221102065448.GA10650@thinkpad/T/
Good point, thanks
> >>> + compatible = "qcom,sc8180x-camnoc-virt";
> >>> + #interconnect-cells = <2>;
> >>> + qcom,bcm-voters = <&apps_bcm_voter>;
> >>> + };
> >>> +
> >>> + mc_virt: interconnect-mc-virt {
> >> Please be consistent with your naming.
> >
> > Are you referring to adding -0 for this?
> I'm not sure which is preferred (-n vs -name), there's a mixed bag
> upstream.. Krzysztof?
Either ways this should be consistent, so camnoc_virt:
interconnect-camnoc_virt makes sense rather than arbitrary -0
> >>> + compatible = "qcom,sc8180x-mc-virt";
> >>> + #interconnect-cells = <2>;
> >>> + qcom,bcm-voters = <&apps_bcm_voter>;
> >>> + };
> >>> +
> >>> + qup_virt: interconnect-qup-virt {
> >>> + compatible = "qcom,sc8180x-qup-virt";
> >>> + #interconnect-cells = <2>;
> >>> + qcom,bcm-voters = <&apps_bcm_voter>;
> >>> + };
> >>> +
> >> [...]
> >>
> >>> + reserved-memory {
> >>> + #address-cells = <2>;
> >>> + #size-cells = <2>;
> >>> + ranges;
> >>> +
> >>> + hyp_mem: hyp-region@85700000 {
> >> the -region seems a bit unnecessary in all of these nodes
> >
> > This is reserved for hyp, I think we should add it here so that we dont
> > touch this piece..?
> I meant the '-region' bit in the node names
>
> (label: name@unit-address)
ack, thanks for clearing my misunderstanding
--
~Vinod
