Danny Kukawka <danny.kukawka@xxxxxxxxx> wrote:
>On Dienstag, 31. Januar 2012, Andy Walls wrote:
>> On Mon, 2012-01-30 at 20:40 +0100, Danny Kukawka wrote:
>> > Fixed handling of 'radio' module parameter from module_param_array
>> > to module_param_named to fix these compiler warnings in
>ivtv-driver.c:
>> >
>> > In function ‘__check_radio’:
>> > 113:1: warning: return from incompatible pointer type [enabled by
>> > default] At top level:
>> > 113:1: warning: initialization from incompatible pointer type
>[enabled by
>> > default] 113:1: warning: (near initialization for
>> > ‘__param_arr_radio.num’) [enabled by default]
>> >
>> > Set initial state of radio_c to true instead of 1.
>>
>> NACK.
>>
>> "radio" is an array of tristate values (-1, 0, 1) per installed card:
>>
>> static int radio[IVTV_MAX_CARDS] = { -1, -1,
>>
>> and must remain an array or you will break the driver.
>>
>> Calling "radio_c" a module parameter named "radio" is wrong.
>>
>> The correct fix is to reverse Rusty Russel's patch to the driver in
>> commit 90ab5ee94171b3e28de6bb42ee30b527014e0be7
>> to change the "bool" to an "int" as it should be in
>> "module_param_array(radio, ...)"
>
>Overseen this. But wouldn't be the correct fix in this case to:
>a) reverse the part of 90ab5ee94171b3e28de6bb42ee30b527014e0be7 to get:
> static unsigned int radio_c = 1;
>
>b) change the following line:
> module_param_array(radio, bool, &radio_c, 0644);
> to:
> module_param_array(radio, int, &radio_c, 0644);
>
>Without b) you would get a warning from the compiler again.
>
>Danny
Yes both need to happen.
I mentioned b) in my original email.
Regards,
Andy
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
