Re: [PATCH v3 08/12] arm64: dts: qcom: sm6350-lena: Enable PMK8350

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] arm64: dts: qcom: sm6350-lena: duplicate qcom-pon device from pmk8350.dtsi
--

commit ab0852d01a0d63b999b8519d90b2ed3ef4a0420c
Author: erebion <erebion@erebion.eu>

arm64: dts: qcom: sm6350-lena: Enable PMK8350

This commit includes pmk8350.dtsi and configures and enables the PMK8350 ADC,
thermal monitor, and RTC nodes on the Sony Xperia 10 III.

> diff --git a/arch/arm64/boot/dts/qcom/sm6350-sony-xperia-lena-pdx213.dts b/arch/arm64/boot/dts/qcom/sm6350-sony-xperia-lena-pdx213.dts
> index 61e191133cb4f..246a290e44f77 100644
> --- a/arch/arm64/boot/dts/qcom/sm6350-sony-xperia-lena-pdx213.dts
> +++ b/arch/arm64/boot/dts/qcom/sm6350-sony-xperia-lena-pdx213.dts
> @@ -8,7 +8,11 @@
>  #define PM7250B_SID 2
>  #define PM7250B_SID1 3
>  
> +/* PMK8350 (in reality a PMK8003) is configured to use SID6 instead of 0 */
> +#define PMK8350_SID 6
> +
>  #include <dt-bindings/firmware/qcom,scm.h>
> +#include <dt-bindings/iio/qcom,spmi-adc7-pmk8350.h>
>  #include <dt-bindings/leds/common.h>
>  #include <dt-bindings/pinctrl/qcom,pmic-gpio.h>
>  #include <dt-bindings/regulator/qcom,rpmh-regulator.h>
> @@ -16,6 +20,7 @@
>  #include "pm6150l.dtsi"
>  #include "pm6350.dtsi"
>  #include "pm7250b.dtsi"
> +#include "pmk8350.dtsi"

[Severity: Medium]
Does including pmk8350.dtsi alongside pm6350.dtsi without disabling or
deleting pmk8350_pon cause a probe conflict between two active qcom-pon
instances?

In arch/arm64/boot/dts/qcom/pm6350.dtsi, pm6350_pon is already enabled as the
primary PON device hosting reboot modes, pm6350_pwrkey, and pm6350_resin:

arch/arm64/boot/dts/qcom/pm6350.dtsi:
		pm6350_pon: pon@800 {
			compatible = "qcom,pm8998-pon";
			reg = <0x800>;
			mode-bootloader = <0x2>;
			mode-recovery = <0x1>;

In arch/arm64/boot/dts/qcom/pmk8350.dtsi, pmk8350_pon (and a root
/reboot-mode nvmem-reboot-mode node) is also enabled by default:

arch/arm64/boot/dts/qcom/pmk8350.dtsi:
		pmk8350_pon: pon@1300 {
			compatible = "qcom,pmk8350-pon";
			reg = <0x1300>, <0x800>;
			reg-names = "hlos", "pbs";

Both qcom,pm8998-pon and qcom,pmk8350-pon match qcom_pon_id_table with
reason_shift = GEN2_REASON_SHIFT in drivers/power/reset/qcom-pon.c:

qcom_pon_probe() {
	...
	if (reason_shift != NO_REASON_SHIFT) {
		pon->reboot_mode.dev = &pdev->dev;
		pon->reason_shift = reason_shift;
		pon->reboot_mode.write = qcom_pon_reboot_mode_write;
		error = devm_reboot_mode_register(&pdev->dev, &pon->reboot_mode);
		if (error) {
			dev_err(&pdev->dev, "can't register reboot mode\n");
			return error;
		}
	}

	platform_set_drvdata(pdev, pon);

	return devm_of_platform_populate(&pdev->dev);
}

In drivers/power/reset/reboot-mode.c, devm_reboot_mode_register() ->
reboot_mode_register() -> reboot_mode_create_device() registers a sysfs class
device named after reboot->dev->driver->name ("qcom-pon"):

reboot_mode_create_device() {
	...
	priv->reboot_mode_device = device_create(&reboot_mode_class, NULL, 0,
						 (void *)priv, "%s",
						 reboot->dev->driver->name);
	...
}

reboot_mode_register() {
	...
	ret = reboot_mode_create_device(reboot);
	if (ret)
		goto error;

	return 0;

error:
	kfree(info);
	reboot_mode_unregister(reboot);
	return ret;
}

When the second qcom-pon instance probes, device_create() fails with -EEXIST
and triggers a sysfs duplicate filename warning, and the error path in
reboot_mode_register() calls reboot_mode_unregister(), which looks up
"qcom-pon" by driver name via class_find_device(), unregisters the first
instance's sysfs device, and leaks the struct device reference returned by
class_find_device(). In addition, if pmk8350_pon probes before pm6350_pon,
qcom_pon_probe() for pm6350_pon returns an error before reaching
devm_of_platform_populate(), preventing pm6350_pwrkey and pm6350_resin from
being populated.

Should pmk8350_pon (and the root /reboot-mode node from pmk8350.dtsi) be
disabled or deleted in sm6350-sony-xperia-lena-pdx213.dts?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261102-pdx213-upstream-3-v3-0-cdc3be96d257@erebion.eu?part=8