Hello Hans,
On 09/26/2011 04:25 PM, Hans Verkuil wrote:
> On Wednesday, September 14, 2011 16:30:47 Sylwester Nawrocki wrote:
>> On 09/14/2011 09:10 AM, Scott Jiang wrote:
>>>>> +static int bcap_qbuf(struct file *file, void *priv,
>>>>> + struct v4l2_buffer *buf)
>>>>> +{
>>>>> + struct bcap_device *bcap_dev = video_drvdata(file);
>>>>> + struct v4l2_fh *fh = file->private_data;
>>>>> + struct bcap_fh *bcap_fh = container_of(fh, struct bcap_fh, fh);
>>>>> +
>>>>> + if (!bcap_fh->io_allowed)
>>>>> + return -EACCES;
>>>>
>>>> I suppose -EBUSY would be more appropriate here.
>>>>
>>> no, io_allowed is to control which file instance has the right to do I/O.
>>
>> Looks like you are doing here what the v4l2 priority mechanism is meant for.
>> Have you considered the access priority (VIDIOC_G_PRIORITY/VIDIOC_S_PRIORITY
>> and friends)? Does it have any shortcomings?
>
> Sylwester, the priority handling doesn't take care of this particular case.
>
> When it comes to streaming you need to administrate which filehandle started
> the streaming and block any other filehandle from interfering with that.
>
> This check should really be done in vb2.
True, I've noticed QBUF/DQBUF are not touched by the priority handling.
Perhaps I didn't follow the discussions in this topic carefully enough.
Then we seem to have another feature request for vb2.
--
Thanks,
Sylwester
--
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
