Re: [PATCH v2 5/6] hw/input/stellaris_input: Convert to qdev

["Philippe Mathieu-Daudé" <philmd@linaro.org>]

On 31/10/23 15:05, Peter Maydell wrote:
> On Tue, 31 Oct 2023 at 13:55, Peter Maydell <peter.maydell@linaro.org> wrote:
>>
>> On Mon, 30 Oct 2023 at 20:38, Mark Cave-Ayland
>> <mark.cave-ayland@ilande.co.uk> wrote:
>>>
>>> On 30/10/2023 11:48, Peter Maydell wrote:
>>> Is it worth converting this to use DEFINE_TYPES() during the conversion? I know Phil
>>> has considered some automation to remove the type_init() boilerplate for the majority
>>> of cases.
>>
>> I could, I guess. It seems a bit awkward that DEFINE_TYPES()
>> wants you to pass it an array even when you only have one type,
>> though, which is going to be a very common use case.

For single type, there is no point beside enforcing a QOM style.

I'll update docs/devel/qom.rst...

> I'm going to squash this into this patch:
> diff --git a/hw/input/stellaris_gamepad.c b/hw/input/stellaris_gamepad.c
> index 6ccf0e80adc..d42ba4f0582 100644
> --- a/hw/input/stellaris_gamepad.c
> +++ b/hw/input/stellaris_gamepad.c
> @@ -90,16 +90,13 @@ static void
> stellaris_gamepad_class_init(ObjectClass *klass, void *data)
>       device_class_set_props(dc, stellaris_gamepad_properties);
>   }
> 
> -static const TypeInfo stellaris_gamepad_info = {
> -    .name = TYPE_STELLARIS_GAMEPAD,
> -    .parent = TYPE_SYS_BUS_DEVICE,
> -    .instance_size = sizeof(StellarisGamepad),
> -    .class_init = stellaris_gamepad_class_init,
> +static const TypeInfo stellaris_gamepad_info[] = {
> +    {
> +        .name = TYPE_STELLARIS_GAMEPAD,
> +        .parent = TYPE_SYS_BUS_DEVICE,
> +        .instance_size = sizeof(StellarisGamepad),
> +        .class_init = stellaris_gamepad_class_init,
> +    },
>   };
> 
> -static void stellaris_gamepad_register_types(void)
> -{
> -    type_register_static(&stellaris_gamepad_info);
> -}
> -
> -type_init(stellaris_gamepad_register_types);
> +DEFINE_TYPES(stellaris_gamepad_info);
> 
> 
> The array is a bit awkward, but it's overall better than having
> to define the register-types function.

Thank you!