Re: [PATCH v2 04/14] arm64: dts: qcom: ipq5332: Flatten usb controller node

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The flattened USB controller node specifies a 'reg' region that is too small, entirely excluding the QCOM wrapper register space that the driver relies on, which leads to probe failures.
--

--- Patch [4]: [PATCH v2 04/14] arm64: dts: qcom: ipq5332: Flatten usb controller node ---
commit 4c4cf3e503a7cae17e6181ac3e5f119bbfadc49e
Author: George Moussalem <[email protected]>

arm64: dts: qcom: ipq5332: Flatten usb controller node

This commit flattens the usb controller node for ipq5332. It updates the
node to use the latest bindings and the flattened driver approach.

> diff --git a/arch/arm64/boot/dts/qcom/ipq5332.dtsi b/arch/arm64/boot/dts/qcom/ipq5332.dtsi
> --- a/arch/arm64/boot/dts/qcom/ipq5332.dtsi
> +++ b/arch/arm64/boot/dts/qcom/ipq5332.dtsi
> @@ -456,16 +456,9 @@ qpic_nand: spi@79b0000 {
>  			status = "disabled";
>  		};
>  
> -		usb: usb@8af8800 {
> -			compatible = "qcom,ipq5332-dwc3", "qcom,dwc3";
> -			reg = <0x08af8800 0x400>;
> -
> -			interrupts = <GIC_SPI 62 IRQ_TYPE_LEVEL_HIGH>,
> -				     <GIC_SPI 53 IRQ_TYPE_LEVEL_HIGH>,
> -				     <GIC_SPI 52 IRQ_TYPE_LEVEL_HIGH>;
> -			interrupt-names = "pwr_event",
> -					  "dp_hs_phy_irq",
> -					  "dm_hs_phy_irq";
> +		usb: usb@8a00000 {
> +			compatible = "qcom,ipq5332-dwc3", "qcom,snps-dwc3";
> +			reg = <0x08a00000 0xe000>;

[Severity: High]
Does this reg size need to be larger to cover the QCOM wrapper registers?

When transitioning to the flattened qcom,snps-dwc3 binding, the reg size is
set to 0xe000, which only covers the DWC3 core registers.

The dwc3-qcom.c wrapper driver relies on the qscratch wrapper registers which
reside at an offset of 0xf8800. The driver modifies the requested resource end
address to include this offset, which means devm_ioremap_resource() will
request a region that exceeds the 0xe000 byte resource specified here.

Could this overlap mismatch cause a resource conflict during
__request_region() and fail the probe with -EBUSY, rendering USB unusable
on the SoC? Should the region be expanded to cover the wrapper registers?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4

-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy