Re: [PATCH] gpio: shared: call gpio_chip::of_xlate() if set

[Jon Hunter <jonathanh@nvidia.com>]

On 17/03/2026 13:43, Bartosz Golaszewski wrote:
> On Tue, Mar 17, 2026 at 1:53 PM Jon Hunter <jonathanh@nvidia.com> wrote:
>>
>>>
>>> I suppose this is not a reset-gpio-like use-case? Could you point me
>>> to the DTS you're using? I've reproduced the bug with a dummy dts and
>>> xlate function in gpio-sim and this patch fixes it but maybe I'm
>>> missing something.
>>
>> So the board I originally observed this on is a farm board with camera
>> hardware I don't have. So like you I started off reproducing locally
>> with a dummy dts by making the following change ...
>>
>> diff --git a/arch/arm64/boot/dts/nvidia/tegra234-p3737-0000+p3701.dtsi b/arch/arm64/boot/dts/nvidia/tegra234-p3737-0000+p3701.dtsi
>> index f6cad29355e6..5e62ffb425f4 100644
>> --- a/arch/arm64/boot/dts/nvidia/tegra234-p3737-0000+p3701.dtsi
>> +++ b/arch/arm64/boot/dts/nvidia/tegra234-p3737-0000+p3701.dtsi
>> @@ -389,14 +389,14 @@ gpio-keys {
>>
>>                   key-force-recovery {
>>                           label = "Force Recovery";
>> -                       gpios = <&gpio TEGRA234_MAIN_GPIO(G, 0) GPIO_ACTIVE_LOW>;
>> +                       gpios = <&gpio TEGRA234_MAIN_GPIO(AF, 2) GPIO_ACTIVE_LOW>;
> 
> The difference between my test and this is that I used foo-gpios vs
> gpios here. I was thinking this would be the culprit but when I
> changed my setup to replicate it, it works fine...
> 
>>                           linux,input-type = <EV_KEY>;
>>                           linux,code = <BTN_1>;
>>                   };
>>
>>                   key-power {
>>                           label = "Power";
>> -                       gpios = <&gpio_aon TEGRA234_AON_GPIO(EE, 4) GPIO_ACTIVE_LOW>;
>> +                       gpios = <&gpio TEGRA234_MAIN_GPIO(AF, 2) GPIO_ACTIVE_LOW>;
>>                           linux,input-type = <EV_KEY>;
>>                           linux,code = <KEY_POWER>;
>>                           wakeup-event-action = <EV_ACT_ASSERTED>;
>>
>>
>> With this I see ...
>>
>>    gpiolib_shared: GPIO 154 owned by tegra234-gpio is shared by multiple consumers
>>    gpiolib_shared: Setting up a shared GPIO entry for key-force-recovery (con_id: '(none)')
>>    gpiolib_shared: Created an auxiliary GPIO proxy gpiolib_shared.proxy.3 for GPIO device tegra234-gpio
>>    gpiolib_shared: Setting up a shared GPIO entry for key-power (con_id: '(none)')
>>    gpiolib_shared: Created an auxiliary GPIO proxy gpiolib_shared.proxy.4 for GPIO device tegra234-gpio
> 
> This looks correct.
> 
>>    ...
>>    gpio-keys gpio-keys: error -ENXIO: Unable to determine IRQ# for button #0
>>
> 
> Is the above error linked to the issue?

Yes, that the final error message I see after the WARNING splat, hence, 
I was thinking that this is somehow linked to the above.

>> I am not sure if this is because these are child nodes of gpio-keys?
>> Obviously this is not a proper example, but something quick and dirty
>> for local testing :-)
>>
> 
> Let me create a setup with GPIO keys then.

Thanks!
Jon

-- 
nvpublic