Re: [PATCH v2 2/5] dt-bindings: PCI: qcom,shikra-pcie: Document the Shikra PCIe Controller

From: Sushrut Shree Trivedi

Date: Mon Sep 21 2026 - 10:06:33 EST



On 9/1/2026 2:12 PM, Krzysztof Kozlowski wrote:
On Mon, Aug 31, 2026 at 12:56:02PM +0530, Sushrut Shree Trivedi wrote:
Add a dedicated schema for the PCIe controller found on the Shikra
platform.

Shikra controller supports upto PCIe Gen3 but the PHY is
only capable of Gen2, hence the max link speed is limited
to Gen2 to avoid link-up failures.

Signed-off-by: Sushrut Shree Trivedi <sushrut.trivedi@xxxxxxxxxxxxxxxx>
---
.../devicetree/bindings/pci/qcom,shikra-pcie.yaml | 218 +++++++++++++++++++++
1 file changed, 218 insertions(+)

What is the difference in the binding against Eliza? IOW, why does this
need a separate file unlike Eliza?
Shikra does not have the Config NoC AXI clock unlike sm8550 / eliza. Apart
from that, minItems for reg  and interrupts need to be updated. I actually
didn't see any bindings with a clock constraint added for only a single compatible.

Do you suggest constraining shikra clock-names and other properties under an if
block inside sm8550 binding itself ?

...

+ reset-names = "pci";
+
+ power-domains = <&gcc GCC_PCIE_GDSC>;
+
+ max-link-speed = <2>;
+
+ operating-points-v2 = <&pcie_opp_table>;
+
+ pcie_opp_table: opp-table {
+ compatible = "operating-points-v2";
Are you sure dt-check-style does not complain?
ACK'd

+
+ /* GEN 1 x1 */
+ opp-2500000 {
+ opp-hz = /bits/ 64 <2500000>;
+ required-opps = <&rpmpd_opp_low_svs>;
+ opp-peak-kBps = <250000 1>;
+ opp-level = <1>;
+ };
+
Best regards,
Krzysztof