Re: [PATCH v2 1/4] media: v4l: dev-decoder: Add source change V4L2_EVENT_SRC_CH_COLORSPACE

[Hans Verkuil]

On 4/8/25 10:45, Ming Qian(OSS) wrote:
> Hi Hans,
> 
> On 2025/4/8 16:30, Hans Verkuil wrote:
>> On 08/04/2025 08:34, Ming Qian(OSS) wrote:
>>> Hi Hans,
>>>
>>> On 2025/4/7 17:54, Hans Verkuil wrote:
>>>> On 17/01/2025 07:19, Ming Qian wrote:
>>>>> Add a new source change V4L2_EVENT_SRC_CH_COLORSPACE that
>>>>> indicates colorspace change in the stream.
>>>>> The change V4L2_EVENT_SRC_CH_RESOLUTION will always affect
>>>>> the allocation, but V4L2_EVENT_SRC_CH_COLORSPACE won't.
>>>>>
>>>>> Signed-off-by: Ming Qian <ming.qian@xxxxxxxxxxx>
>>>>> ---
>>>>>    Documentation/userspace-api/media/v4l/vidioc-dqevent.rst | 9 +++++++++
>>>>>    .../userspace-api/media/videodev2.h.rst.exceptions       | 1 +
>>>>>    include/uapi/linux/videodev2.h                           | 1 +
>>>>>    3 files changed, 11 insertions(+)
>>>>>
>>>>> diff --git a/Documentation/userspace-api/media/v4l/vidioc-dqevent.rst b/Documentation/userspace-api/media/v4l/vidioc-dqevent.rst
>>>>> index 8db103760930..91e6b86c976d 100644
>>>>> --- a/Documentation/userspace-api/media/v4l/vidioc-dqevent.rst
>>>>> +++ b/Documentation/userspace-api/media/v4l/vidioc-dqevent.rst
>>>>> @@ -369,6 +369,15 @@ call.
>>>>>    	loss of signal and so restarting streaming I/O is required in order for
>>>>>    	the hardware to synchronize to the video signal.
>>>>>    
>>>>> +    * - ``V4L2_EVENT_SRC_CH_COLORSPACE``
>>>>> +      - 0x0002
>>>>> +      - This event gets triggered when a colorsapce change is detected at
>>>>
>>>> colorsapce -> colorspace
>>>>
>>>
>>> Will fix in v3
>>>
>>>>> +	an input. This can come from a video decoder. Applications will query
>>>>
>>>> It can also come from a video receiver. E.g. an HDMI source changes colorspace
>>>> signaling, but not the resolution.
>>>>
>>>>> +	the new colorspace information (if any, the signal may also have been
>>>>> +	lost)
>>>>
>>>> Missing . at the end. Also, if the signal is lost, then that is a CH_RESOLUTION
>>>> change, not CH_COLORSPACE.
>>>>
>>> OK, will fix in v3
>>>>> +
>>>>> +	For stateful decoders follow the guidelines in :ref:`decoder`.
>>>>
>>>> I think this should emphasize that if CH_COLORSPACE is set, but not CH_RESOLUTION,
>>>> then only the colorspace changed and there is no need to reallocate buffers.
>>>>
>>>
>>> OK, will add in v3
>>>
>>>> I also wonder if the description of CH_RESOLUTION should be enhanced to explain
>>>> that this might also imply a colorspace change. I'm not sure what existing codec
>>>> drivers do if there is a colorspace change but no resolution change.
>>>
>>> I think there is no uniform behavior at the moment, it depends on the
>>> behavior of the decoder. Maybe most decoders ignore this.
>>
>> Can you try to do a quick analysis of this? Don't spend too much time on this,
>> but it is helpful to have an idea of how existing codecs handle this.
>>
>> Regards,
>>
>> 	Hans
>>
> 
> I checked the vpu used in our platforms,
> 1. amphion vpu, it will ignore the colorspace change.
> 2. hantro g1/g2 decoder, it also ignore the colorspace change.
> 3. chipsnmedia wave6 decoder, the firmware detect the colorspace change
> for HEVC format, but ignore for AVC format. But its driver just ignore
> it.
> 4. chipsnmedia wave511 decoder, same as wave6.

I meant stateful codec drivers that are in mainline. Out-of-tree drivers do not
concern me.

Regards,

	Hans

> 
> Regards,
> Ming
> 
>>>
>>>>
>>>> I'm a bit concerned about backwards compatibility issues: if a userspace application
>>>> doesn't understand this new flag and just honors CH_RESOLUTION, then it would
>>>> never react to just a colorspace change.
>>>>
>>>> Nicolas, does gstreamer look at these flags?
>>>
>>> I checked the gstreamer code, it does check this flag:
>>>
>>> if (event.type == V4L2_EVENT_SOURCE_CHANGE &&
>>>       (event.u.src_change.changes & V4L2_EVENT_SRC_CH_RESOLUTION)) {
>>>     GST_DEBUG_OBJECT (v4l2object->dbg_obj,
>>>         "Can't streamon capture as the resolution have changed.");
>>>     ret = GST_V4L2_FLOW_RESOLUTION_CHANGE;
>>> }
>>>
>>> Currently the gstreamer can't handle the CH_COLORSPACE flag.
>>>
>>> Thanks,
>>> Ming
>>>
>>>>
>>>> Regards,
>>>>
>>>> 	Hans
>>>>
>>>>> +
>>>>>    Return Value
>>>>>    ============
>>>>>    
>>>>> diff --git a/Documentation/userspace-api/media/videodev2.h.rst.exceptions b/Documentation/userspace-api/media/videodev2.h.rst.exceptions
>>>>> index 35d3456cc812..ac47c6d9448b 100644
>>>>> --- a/Documentation/userspace-api/media/videodev2.h.rst.exceptions
>>>>> +++ b/Documentation/userspace-api/media/videodev2.h.rst.exceptions
>>>>> @@ -526,6 +526,7 @@ replace define V4L2_EVENT_CTRL_CH_RANGE ctrl-changes-flags
>>>>>    replace define V4L2_EVENT_CTRL_CH_DIMENSIONS ctrl-changes-flags
>>>>>    
>>>>>    replace define V4L2_EVENT_SRC_CH_RESOLUTION src-changes-flags
>>>>> +replace define V4L2_EVENT_SRC_CH_COLORSPACE src-changes-flags
>>>>>    
>>>>>    replace define V4L2_EVENT_MD_FL_HAVE_FRAME_SEQ :c:type:`v4l2_event_motion_det`
>>>>>    
>>>>> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
>>>>> index c8cb2796130f..242242c8e57b 100644
>>>>> --- a/include/uapi/linux/videodev2.h
>>>>> +++ b/include/uapi/linux/videodev2.h
>>>>> @@ -2559,6 +2559,7 @@ struct v4l2_event_frame_sync {
>>>>>    };
>>>>>    
>>>>>    #define V4L2_EVENT_SRC_CH_RESOLUTION		(1 << 0)
>>>>> +#define V4L2_EVENT_SRC_CH_COLORSPACE		(1 << 1)
>>>>>    
>>>>>    struct v4l2_event_src_change {
>>>>>    	__u32 changes;
>>>>
>>>
>>
>