Re: [PATCH V1 2/2] arm64: dts: qcom: sc7280: Add UFS host controller and phy nodes

[Konrad Dybcio]

On 26.09.2023 18:24, Nitin Rawat wrote:
> 
> 
> On 8/22/2023 12:38 PM, Manivannan Sadhasivam wrote:
>> On Mon, Aug 21, 2023 at 03:19:37PM +0530, Nitin Rawat wrote:
>>> Add UFS host controller and PHY nodes for sc7280.
>>>
>>
>> You should split this patch into 2. One for SoC and another for board.
> Updated in Latest Patchset.
> 
>>
>>> Signed-off-by: Nitin Rawat <quic_nitirawa@xxxxxxxxxxx>
>>> ---
>>>   arch/arm64/boot/dts/qcom/sc7280-idp.dtsi | 19 +++++++
>>>   arch/arm64/boot/dts/qcom/sc7280.dtsi     | 64 ++++++++++++++++++++++++
>>>   2 files changed, 83 insertions(+)
>>>
>>> diff --git a/arch/arm64/boot/dts/qcom/sc7280-idp.dtsi b/arch/arm64/boot/dts/qcom/sc7280-idp.dtsi
>>> index 2ff549f4dc7a..c60cdd511222 100644
>>> --- a/arch/arm64/boot/dts/qcom/sc7280-idp.dtsi
>>> +++ b/arch/arm64/boot/dts/qcom/sc7280-idp.dtsi
>>> @@ -451,6 +451,25 @@
>>>       status = "okay";
>>>   };
>>>
>>> +&ufs_mem_hc {
>>> +    reset-gpios = <&tlmm 175 GPIO_ACTIVE_LOW>;
>>> +    vcc-supply = <&vreg_l7b_2p9>;
>>> +    vcc-max-microamp = <800000>;
>>> +    vccq-supply = <&vreg_l9b_1p2>;
>>> +    vccq-max-microamp = <900000>;
>>> +    vccq2-supply = <&vreg_l9b_1p2>;
>>> +    vccq2-max-microamp = <900000>;
>>> +
>>> +    status = "okay";
>>> +};
>>> +
>>> +&ufs_mem_phy {
>>> +    vdda-phy-supply = <&vreg_l10c_0p8>;
>>> +    vdda-pll-supply = <&vreg_l6b_1p2>;
>>> +
>>> +    status = "okay";
>>> +};
>>> +
>>>   &sdhc_1 {
>>>       status = "okay";
>>>
>>> diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
>>> index 925428a5f6ae..d4a15d56b384 100644
>>> --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
>>> +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
>>> @@ -908,6 +908,70 @@
>>>               };
>>>           };
>>>
>>> +        ufs_mem_phy: phy@1d87000 {
>>
>> Please sort the nodes in ascending order.
> Updated in Latest Patchset.
> 
>>
>>> +            compatible = "qcom,sc7280-qmp-ufs-phy";
>>> +            reg = <0x0 0x01d87000 0x0 0xe00>;
>>> +            clocks = <&rpmhcc RPMH_CXO_CLK>,
>>> +                 <&gcc GCC_UFS_PHY_PHY_AUX_CLK>,
>>> +                 <&gcc GCC_UFS_1_CLKREF_EN>;
>>> +            clock-names = "ref", "ref_aux", "qref";
>>> +
>>> +            resets = <&ufs_mem_hc 0>;
>>> +            reset-names = "ufsphy";
>>> +
>>> +            #clock-cells = <1>;
>>> +            #phy-cells = <0>;
>>> +
>>> +            status = "disabled";
>>> +
>>> +        };
>>> +
>>> +        ufs_mem_hc: ufs@1d84000 {
>>> +            compatible = "qcom,sc7280-ufshc", "qcom,ufshc",
>>> +                     "jedec,ufs-2.0";
>>> +            reg = <0x0 0x01d84000 0x0 0x3000>;
>>> +            interrupts = <GIC_SPI 265 IRQ_TYPE_LEVEL_HIGH>;
>>> +            phys = <&ufs_mem_phy>;
>>> +            phy-names = "ufsphy";
>>> +            lanes-per-direction = <2>;
>>> +            #reset-cells = <1>;
>>> +            resets = <&gcc GCC_UFS_PHY_BCR>;
>>> +            reset-names = "rst";
>>> +
>>> +            power-domains = <&gcc GCC_UFS_PHY_GDSC>;
>>> +            required-opps = <&rpmhpd_opp_nom>;
>>> +
>>> +            iommus = <&apps_smmu 0x80 0x0>;
>>> +            dma-coherent;
>>> +
>>> +            clock-names = "core_clk",
>>> +                      "bus_aggr_clk",
>>> +                      "iface_clk",
>>> +                      "core_clk_unipro",
>>> +                      "ref_clk",
>>> +                      "tx_lane0_sync_clk",
>>> +                      "rx_lane0_sync_clk",
>>> +                      "rx_lane1_sync_clk";
>>
>> "clocks" property should come first.
>  DT binding shows clock-names first followed by clocks.
>  Let me know if see still see concern, would update .
The dt bindings example is rarely useful.. perhaps we should
change that..

The general consensus there is to have

property
property-names

Konrad