Hi Dmitry,
On 1/14/2025 2:12 AM, Dmitry Baryshkov wrote:
> On Mon, Jan 13, 2025 at 01:52:08PM -0800, Melody Olvera wrote:
>> From: Wesley Cheng <quic_wcheng@xxxxxxxxxxx>
>>
>> On SM8750, the M31 eUSB2 PHY is being used to support USB2. Add the
>> binding definition for the PHY driver.
>>
>> Signed-off-by: Wesley Cheng <quic_wcheng@xxxxxxxxxxx>
>> Signed-off-by: Melody Olvera <quic_molvera@xxxxxxxxxxx>
>> ---
>> .../bindings/phy/qcom,m31-eusb2-phy.yaml | 84 ++++++++++++++++++++++
>> 1 file changed, 84 insertions(+)
>>
>> +examples:
>> + - |
>> + #include <dt-bindings/clock/qcom,sm8750-gcc.h>
>> + #include <dt-bindings/clock/qcom,rpmh.h>
>> + #include <dt-bindings/clock/qcom,sm8750-tcsr.h>
>
> This a typical comment, please actually update your internal
> documentation: don't use GCC's and other clock controller's bindings in
> examples for other blocks.
>
I can see that other examples also reference bindings from external
drivers, otherwise I think it would throw an error for not being able to
find the parents in properties such as, resets and clocks. I tried to
update my dtschema to ensure that this failure was not missed.
I checked and made sure that no errors were not seen, so I started to
wonder if maybe when this series was submitted the qcom,sm8750-gcc.h wasn't
yet present on the kernel tree. To confirm this, I removed the
aforementioned header file, and got the same error/issue. Now that the
sm8750 gcc header has landed upstream, I believe that this error should be
resolved.
Thanks
Wesley Cheng
>> +
>> + usb_1_hsphy: phy@88e3000 {
>> + compatible = "qcom,sm8750-m31-eusb2-phy";
>> + reg = <0x88e3000 0x29c>;
>> +
>> + clocks = <&tcsrcc TCSR_USB2_CLKREF_EN>;
>> + clock-names = "ref";
>> +
>> + resets = <&gcc GCC_QUSB2PHY_PRIM_BCR>;
>> +
>> + #phy-cells = <0>;
>> +
>> + vdd-supply = <&vreg_l2d_0p88>;
>> + vdda12-supply = <&vreg_l3g_1p2>;
>> +
>> + };
>>
>> --
>> 2.46.1
>>
>
