Support non-contiguous planar data in QVideoFrame/QAbstractVideoBuffer

QVideoFrame and QAbstractVideoBuffer only returns a pointer to the first plane of YUV formats. It is then up to the user to calculate the offset of the subsquent planes based on the height and stride of that plane. Which gives the wrong results if the planes aren't packed together.

An option for the planes to be returned independently would be helpful in this instance.