Re: [PATCHv2] ISP:BUILD:FIX: Move media_entity_init() and

[Mauro Carvalho Chehab]

Em 06-09-2011 11:47, Laurent Pinchart escreveu:
> Hi Vaibhav,
> 
> On Tuesday 06 September 2011 16:12:35 Hiremath, Vaibhav wrote:
>> On Monday, September 05, 2011 6:20 PM Laurent Pinchart wrote:
>>> On Sunday 04 September 2011 15:32:28 Mauro Carvalho Chehab wrote:
>>
>> <snip>
>>
>>> I don't mind splitting the config option. An alternative would be to
>>> compile media_entity_init() and media_entity_cleanup() based on
>>> CONFIG_MEDIA_SUPPORT instead of CONFIG_MEDIA_CONTROLLER, but that looks a
>>> bit hackish to me.
>>>
>>>> Also, I don't like the idea of increasing drivers complexity for the
>>>> existing drivers that work properly without MC. All those core
>>>> conversions that were done in the last two years caused already too much
>>>> instability to them.
>>>>
>>>> We should really avoid touching on them again for something that won't
>>>> be adding any new feature nor fixing any known bug.
>>>
>>> We don't have to convert them all in one go right now, we can implement
>>> pad-level operations support selectively when a subdev driver becomes used
>>> by an MC-enabled host/bridge driver.
>>
>> I completely agree that we should not be duplicating the code just for sake
>> of it.
>>
>> Isn't the wrapper approach seems feasible here?
> 
> As explained in a previous e-mail, a wrapper sounds like a good approach to 
> me, to emulate video::* operations based on pad::* operations. We want to move 
> to pad::* operations, so we should not perform emulation the other way around.

We have 300+ drivers under /drivers/media. Just a few of them need MC API.

Good sense says that we shouldn't touch on 300+ just because of half dozen drivers.

So, if a wrapper is needed, it should be done for the other side.

Mauro.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/