Hi,

Oops, forgot the attachment it is here now...

Regards,

Hans


On 05/05/2012 05:05 PM, Hans de Goede wrote:
> Hi,
>
> On 05/05/2012 04:44 PM, Hans de Goede wrote:
>> Hi,
>>
>> On 05/05/2012 11:14 AM, Hans Verkuil wrote:
>>> So you get:
>>>
>>> vidioc_foo()
>>> lock(mylock)
>>> v4l2_ctrl_s_ctrl(ctrl, val)
>>> s_ctrl(ctrl, val)
>>> lock(mylock)
>>
>> Easy solution here, remove the first lock(mylock), since we are not using v4l2-dev's
>> locking, we are the one doing the first lock, and if we are going to call v4l2_ctrl_s_ctrl
>> we should simply not do that!
>>
>> Now I see that we are doing exactly that in for example vidioc_g_jpegcomp in gspca.c, so
>> we should stop doing that. We can make vidioc_g/s_jpegcomp only do the usb locking if
>> gspca_dev->vdev.ctrl_handler == NULL, and once all sub drivers are converted simply remove
>> it. Actually I'm thinking about making the jpegqual control part of the gspca_dev struct
>> itself and move all handling of vidioc_g/s_jpegcomp out of the sub drivers and into
>> the core.
>
> Here is an updated version of this patch implementing this approach for
> vidioc_g/s_jpegcomp. We may need to do something similar in other places, although I cannot
> think of any such places atm,
>
> Regards,
>
> Hans
> --
> To unsubscribe from this list: send the line "unsubscribe linux-media" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html