chewitt , you are 100% right, I guess you are talking about the vh264_4k2k.c in the vendor kernel, impressive amount of engineering in it. On the other hand does the cap in the current kernel actually helps in anyway? It seems the variables are only used to crop the output. But based on the content of vh264_4k2k.c one needs a different firmware to even decode the stream, so I assume 4K video will fail the same way with or without this constraint. On the other hand when the input is something the standard h264 code can handle, like 1080x1920, the cropping just output a little more than the upper half of the video, which hurts user experience.