On 02/24/16 09:39, Geert Uytterhoeven wrote:
> Hi Hans,
>
> On Wed, Feb 24, 2016 at 8:20 AM, Hans Verkuil <hverk...@xs4all.nl> wrote:
>> On 02/24/2016 07:17 AM, Simon Horman wrote:
>>> On Wed, Feb 24, 2016 at 07:59:57AM +0200, Laurent Pinchart wrote:
>>>> Hi Simon,
>>>>
>>>> Thank you for the patch.
>>>>
>>>> On Wednesday 24 February 2016 11:07:59 Simon Horman wrote:
>>>>> This driver does not appear to have ever been used by any SoC's defconfig
>>>>> and does not appear to support DT. In sort it seems unused an unlikely
>>>>> to be used.
>>>>>
>>>>> Signed-off-by: Simon Horman <horms+rene...@verge.net.au>
>>>>> ---
>>>>> drivers/media/platform/soc_camera/Kconfig | 7 -
>>>>> drivers/media/platform/soc_camera/Makefile | 1 -
>>>>> drivers/media/platform/soc_camera/sh_mobile_csi2.c | 400
>>>>> ------------------
>>>>
>>>> Shouldn't you also remove include/media/drv-intf/sh_mobile_csi2.h ? You
>>>> would
>>>> then need to update drivers/media/platform/soc_camera/sh_mobile_ceu.c
>>>> accordingly, or remove it altogether.
>>>
>>> Thanks.
>>>
>>> sh_mobile_ceu appears to be used by several SH boards so I'd rather
>>> not remove it, at least not for this reason.
>>>
>>> So I'd prefer to look into updating sh_mobile_ceu.c and removing
>>> sh_mobile_csi2.h.
>>
>> Last time I checked the ceu driver failed to work (missing clock). I'll test
>> again this weekend with the latest kernel. See what the status is of this
>> driver.
>
> I don't know when you tested last time, but as Simon postponed "[PATCH 3/4]
> drivers: sh: Stop using the legacy clock domain on ARM"
> (http://www.spinics.net/lists/arm-kernel/msg483561.html), clocks may still be
> broken.
>
> I'll send a simple fix for the regression only.
Thanks! I'll test it this weekend.
Regards,
Hans
