This is a port of the OV13850 camera sensor (CAM1320) from the rockchip-kernel(https://github.com/friendlyarm/kernel-rockchip/) to the 5.10-rc6 Kernel.
I keep this repository to show the current development state, while I work on upstreaming the patch to the mainline kernel.
I required a system with a rk3399 SoC as it contains the rkisp1 Image Signal Processor. The NanoPC-T4 is a great Single Board Computer, it just has a few flaws, for example, a lot of functionality is found on the BSP(Board Specific Programming) Kernel from friendlyElec, which is based on the 4.4 Kernel (2016). This on its own is not a major problem, but it gets problematic as soon as you want to do upstream development for the Linux Kernel. The MIPI CSI-2 & CSI-1 sockets of the board have a special width, which makes it impossible to insert for example the raspberry pi camera module v2 with the Sony IMX219 image sensor. The only two cameras that I was able to find, which fit those sockets are the CAM1320 with the OV13850 image sensor & the MCAM-400 with the OV4689 image sensor and in contrast to the IMX219 these cameras have no upstream driver.
If you want more details, I have created a blog post about the whole endeavor, here TODO - not finished yet
In order to play around with this patch, you will have to apply it to the target kernel version tree and create an image for the NanoPC-T4. If you use the Armbian method described below, the procedure of adding this driver is simple just put the patch file into the userpatches/kernel/rockchip64-dev folder of the armbian build repository.
There are 3 major ways of creating an image for the NanoPC-T4 with an upstream kernel: Armbian, Buildroot, Yocto Project. My pick was armbian simply because I heard it was the fastest to set up. Now there is a little bit of learning involved in order to learn how to create an armbian image with upstream sources but it is possible. If you want to learn how: I describe that in this project together with the latest configurations and patches, that I use to build the image.