Re: [PATCH 1/3] powerpc/dts: mpc5121: Move GPIO controller properties to SoC dtsi

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The ads7846 touchscreen driver and gpio-mpc8xxx driver both claim IPIC IRQ 78, leading to an IRQ request conflict and masked interrupts.
--

--- Patch [1]: [PATCH 1/3] powerpc/dts: mpc5121: Move GPIO controller properties to SoC dtsi ---
commit 9f0c3f1fd5a726a9eddf4e07512d2ccf88203d01
Author: Dmitry Torokhov <dmitry.torokhov@gmail.com>

powerpc/dts: mpc5121: Move GPIO controller properties to SoC dtsi

Define gpio@1100 with label gpio_pic as a GPIO controller and
interrupt controller directly in mpc5121.dtsi so all MPC5121 boards
inherit the correct controller capabilities.

> diff --git a/arch/powerpc/boot/dts/mpc5121.dtsi b/arch/powerpc/boot/dts/mpc5121.dtsi
> index a278fb7b9e711..9c5dacb94b924 100644
> --- a/arch/powerpc/boot/dts/mpc5121.dtsi
> +++ b/arch/powerpc/boot/dts/mpc5121.dtsi

[ ... ]

> @@ -146,10 +146,14 @@ pmc@1000 {
>  			interrupts = <83 0x8>;
>  		};
>  
> -		gpio@1100 {
> +		gpio_pic: gpio@1100 {
>  			compatible = "fsl,mpc5121-gpio";
>  			reg = <0x1100 0x100>;
>  			interrupts = <78 0x8>;
> +			gpio-controller;
> +			#gpio-cells = <2>;
> +			interrupt-controller;
> +			#interrupt-cells = <2>;

[Severity: High]
This is a pre-existing issue, but does centralizing this interrupt controller
declaration leave an unresolved IRQ request conflict on the pdm360ng board?

In arch/powerpc/boot/dts/pdm360ng.dts, the touchscreen node directly claims
IPIC IRQ 78 instead of routing through the GPIO controller:

ts@0 {
        ...
        interrupts = <78 0x8>;
};

Additionally, in arch/powerpc/platforms/512x/pdm360ng.c,
pdm360ng_penirq_init() directly manipulates the GPIO controller's IMR
register:

setbits32(pdm360ng_gpio_base + 0x10, 0x40);

When both the ads7846 touchscreen driver and gpio-mpc8xxx driver try to
claim IRQ 78, could the request fail with -EBUSY? 

Even if they could share the IRQ, the gpio-mpc8xxx driver clears the GPIO
IMR register on probe, which would silently overwrite the earlier setup
done by pdm360ng_penirq_init() and mask the touchscreen's interrupt.

Could this cause a permanent loss of touch input on boot for the pdm360ng
when both drivers are enabled?

>  		};
>  
>  		can@1300 {

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260718-ads7846-pdm360ng-v1-0-f318f85cda57@gmail.com?part=1