No, raspivid/raspistill rely on the legacy stack, and the legacy stack only supports the v1, v2, and HQ cameras.How would I determine if our camera is recognized by the Pi then and could we still use raspivid for capture?"vcgencmd get_camera" is ONLY valid for the legacy camera stack. It is irrelevant for anything using the kernel drivers.
They have been replaced by rpicam-apps for use with libcamera.
If the sensor is on that I2C bus, then yes.How would we go about fixing the DTS file then. I noticed most (if not all) of the existing DTS files use <&i2c_csi_dsi> as the target for fragment0. Is this what we should be using?Both fragments are targeting csi0. The endpoint of that node is then pointed effectively at itself, so that isn't going to work.
If it isn't (which if you have no control path is the case) then it should go in a sensible location. I've pointed you at my dummy sensor branch, and the overlay there puts it in the root of device tree.
That would be why I've already pointed you at my branch where I'd started looking at that. It'll need a little tweaking for the latest kernel as APIs have changed, but would be a good basis for your system. Search the forums for "dummy sensor" and you'll find a number of threads discussing it in the past. I'll see about updating it for the 6.6 kernel.Okay say I write a driver, I'm guessing this needs to be compiled into some other file. How would we go about this? Is there any documentation on what needs to be included?Even if you have no control over the CSI source, you need a driver for it to define things like the resolution to be expected. That is what the dummy_sensor driver I was tinkering with on https://github.com/6by9/linux/tree/rpi- ... mmy_sensor was looking to do.
Building the kernel is covered in the documentation - https://www.raspberrypi.com/documentati ... the-kernel
Statistics: Posted by 6by9 — Sat Mar 02, 2024 7:07 am