Re: [PATCH] dt-bindings: arm: qcom,ids: Add SoC ID for Snapdragon SDA 850

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Missing SoC ID mapping in the qcom-socinfo driver causes incorrect sysfs machine attribute.
--

commit 40f8c20a8f811ef4d99535abc1be00a3d96ed7b4
Author: David Heidelberg <david@ixit.cz>

dt-bindings: arm: qcom,ids: Add SoC ID for Snapdragon SDA 850

This commit adds the numerical SoC ID (348) for the Qualcomm Snapdragon
SDA850 to the device tree bindings header.

> diff --git a/include/dt-bindings/arm/qcom,ids.h b/include/dt-bindings/arm/qcom,ids.h
> index 1af73c0ad41cd..4b3062f57eecd 100644
> --- a/include/dt-bindings/arm/qcom,ids.h
> +++ b/include/dt-bindings/arm/qcom,ids.h
> @@ -184,6 +184,7 @@
>  #define QCOM_ID_IPQ8078			344
>  #define QCOM_ID_SDM636			345
>  #define QCOM_ID_SDA636			346
> +#define QCOM_ID_SDA850			348

[Severity: Medium]
Should this ID also be added to the soc_id[] array in
drivers/soc/qcom/socinfo.c?

Without adding qcom_board_id(SDA850) to the array, socinfo_machine()
will return NULL when the qcom-socinfo driver probes a Snapdragon SDA850
device, because it fails to map the SMEM build ID (348) to a string.

As a result, soc_device_register() will fall back to using
of_machine_read_model(), which exposes the device tree's board model
string in the /sys/devices/soc0/machine sysfs attribute rather than
the expected SoC name.

Does this regression break user-space tools that rely on the
Qualcomm-specific ABI expectation for the machine attribute?

>  #define QCOM_ID_SDM632			349
>  #define QCOM_ID_SDA632			350
>  #define QCOM_ID_SDA450			351

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260623-sda850-v1-1-ddd8e62c85d8@ixit.cz?part=1