On 7/28/2023 9:27 PM, Krzysztof Kozlowski wrote:
On 28/07/2023 17:10, Krishna Chaitanya Chundru wrote:
On 7/28/2023 5:33 PM, Krzysztof Kozlowski wrote:
On 28/07/2023 12:39, Krishna chaitanya chundru wrote:
Add PCIe dtsi node for PCIe0 controller on sc7280 platform.
Signed-off-by: Krishna chaitanya chundru <quic_krichai@xxxxxxxxxxx>
Thank you for your patch. There is something to discuss/improve.
+ pcie0_phy: phy@1c06000 {
+ compatible = "qcom,sm8250-qmp-gen3x1-pcie-phy";
+ reg = <0 0x01c06000 0 0x1c0>;
+ #address-cells = <2>;
+ #size-cells = <2>;
+ ranges;
+ clocks = <&gcc GCC_PCIE_0_AUX_CLK>,
+ <&gcc GCC_PCIE_0_CFG_AHB_CLK>,
+ <&gcc GCC_PCIE_CLKREF_EN>,
+ <&gcc GCC_PCIE0_PHY_RCHNG_CLK>;
+ clock-names = "aux", "cfg_ahb", "ref", "refgen";
+
+ resets = <&gcc GCC_PCIE_0_PHY_BCR>;
+ reset-names = "phy";
+
+ assigned-clocks = <&gcc GCC_PCIE0_PHY_RCHNG_CLK>;
+ assigned-clock-rates = <100000000>;
+
+ status = "disabled";
+
+ pcie0_lane: phy@1c0e6200 {
Isn't this old-style of bindings? Wasn't there a change? On what tree
did you base it?
The work was here:
https://lore.kernel.org/all/20230324022514.1800382-5-dmitry.baryshkov@xxxxxxxxxx/
But I don't remember the status.
Let me rebase and send it again.
This anyway looks like wrong compatible. You used sm8250.
The patch was send on latest linux-next only and the above change is not
merged yet.
We are using the same compatible string as sm8250 because the phy is
same both from hardware and software perspective for sm8250.
that is why we are using the same compatible string.
Can you let me know if we want create a separate compatible string for
this even though we are using same phy?
- KC.
Best regards,
Krzysztof
