Re: [PATCH 1/2] arm64: dts: qcom: sm8550-qrd: add PCIe0

[Krzysztof Kozlowski]

On 16/05/2023 19:15, Dmitry Baryshkov wrote:
> On Tue, 16 May 2023 at 19:43, Krzysztof Kozlowski
> <krzysztof.kozlowski@xxxxxxxxxx> wrote:
>>
>> On 16/05/2023 18:39, Dmitry Baryshkov wrote:
>>> On Tue, 16 May 2023 at 16:30, Krzysztof Kozlowski
>>> <krzysztof.kozlowski@xxxxxxxxxx> wrote:
>>>>
>>>> Add PCIe0 nodes used with WCN7851 device.  The PCIe1 is not connected,
>>>> thus skip pcie_1_phy_aux_clk input clock to GCC.
>>>>
>>>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
>>>> ---
>>>>  arch/arm64/boot/dts/qcom/sm8550-qrd.dts | 32 +++++++++++++++++++++++++
>>>>  1 file changed, 32 insertions(+)
>>>>
>>>> diff --git a/arch/arm64/boot/dts/qcom/sm8550-qrd.dts b/arch/arm64/boot/dts/qcom/sm8550-qrd.dts
>>>> index ccc58e6b45bd..e7a2bc5d788b 100644
>>>> --- a/arch/arm64/boot/dts/qcom/sm8550-qrd.dts
>>>> +++ b/arch/arm64/boot/dts/qcom/sm8550-qrd.dts
>>>> @@ -385,6 +385,38 @@ vreg_l3g_1p2: ldo3 {
>>>>         };
>>>>  };
>>>>
>>>> +&gcc {
>>>> +       clocks = <&bi_tcxo_div2>, <&sleep_clk>,
>>>> +                <&pcie0_phy>,
>>>> +                <&pcie1_phy>,
>>>> +                <0>,
>>>> +                <&ufs_mem_phy 0>,
>>>> +                <&ufs_mem_phy 1>,
>>>> +                <&ufs_mem_phy 2>,
>>>> +                <&usb_dp_qmpphy QMP_USB43DP_USB3_PIPE_CLK>;
>>>> +};
>>>
>>> Is there any reason to disable the PCIe1 PHY AUX clock here? I mean,
>>> the PCIe1 is still enabled in the hardware.
>>
>> I was thinking about this. The AUX clock seems to be an external clock,
>> although I could not find it in schematics. I assume that on QRD8550 it
>> could be missing, if it is really external. OTOH, downstream DTS did not
>> seem to care...
> 
> I might be completely wrong here, but I think that AUX clock is yet
> another clock provided by the PHY to the GCC, which we were just
> ignoring for now. For example, for sm8450 we have <0> there. I don't
> see it as an external clock, so I think it is internal to the SoC.

Hm, in that case it would make sense to keep it here. It's frequency,
with some safe choice, could also go to DTSI.

Best regards,
Krzysztof