Re: [PATCH v5] arm64: dts: qcom: qcom6490-idp: Add IPA node

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The IPA device will fail to probe unconditionally due to a missing `memory-region` property, contradicting the commit message claims.
--

--- Patch [1]: [PATCH v5] arm64: dts: qcom: qcom6490-idp: Add IPA node ---
commit 1670b1c514088d32ff0dcb8120b44156a4500dd0
Author: Pavan Kumar <pavan.m@oss.qualcomm.com>

arm64: dts: qcom: qcom6490-idp: Add IPA node

Enable the IPA block and provide the firmware name required for firmware
loading on qcm6490-idp.

> Changes in v3:
> - Rebased on top of IPA property consolidation series.
> - Dropped qcom,gsi-loader and memory-region properties as they are now
>   provided by kodiak.dtsi.

[Severity: High]
Is this description accurate regarding kodiak.dtsi providing these properties?

During system boot, the IPA driver in drivers/net/ipa/ipa_main.c probes the
ipa node. Without a qcom,gsi-loader property, ipa_firmware_loader() defaults
to IPA_LOADER_SELF. This triggers a call to ipa_firmware_load(), which
attempts to fetch the memory-region property via
of_reserved_mem_region_to_resource().

If memory-region is missing in both qcm6490-idp.dts and the included
kodiak.dtsi, the fetch will fail and abort the probe.

> diff --git a/arch/arm64/boot/dts/qcom/qcm6490-idp.dts b/arch/arm64/boot/dts/qcom/qcm6490-idp.dts
> index bdc02260f902b..7242695aed978 100644
> --- a/arch/arm64/boot/dts/qcom/qcm6490-idp.dts
> +++ b/arch/arm64/boot/dts/qcom/qcm6490-idp.dts

[ ... ]

> @@ -619,6 +619,12 @@ &gpu_zap_shader {
>  	firmware-name = "qcom/qcm6490/a660_zap.mbn";
>  };
>  
> +&ipa {
> +	firmware-name = "qcom/qcm6490/ipa_fws.mbn";
> +
> +	status = "okay";
> +};
> +

[Severity: High]
Will this node fail to probe unconditionally on the qcm6490-idp board due
to the missing memory-region property?

Additionally, looking earlier in arch/arm64/boot/dts/qcom/qcm6490-idp.dts,
there is an explicit deletion of the firmware memory region:

/delete-node/ &ipa_fw_mem;

If kodiak.dtsi did provide the memory-region property pointing to this
node, wouldn't deleting its target cause a Device Tree Compiler error?
Does the memory-region property need to be explicitly defined here to
allow successful probing?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260722082314.1502072-1-pavan.m@oss.qualcomm.com?part=1