I have two imx264 connected to our custom board using spi, named “imx264 spi0.0” and “imx264 spi1.0”. The first one is connected to /dev/video0, the second one to /dev/video1. I can verify using standard v4l2 commands like v4l2-ctl or v4l2src that I get different images on /dev/video0 and /dev/video1. I have described them in the dt as follows :
Unfortunately, when I use nvcamerasrc with sensor-id=0 or sensor-id=1 it always uses the sensor bound to /dev/video0, never the one connected to /dev/video1. And if I use “sensor-id=2”, nvcamera-daemon crashes.
How does nvcamera-daemon/nvcamerasrc decide which sensor is “sensor-id=0” and which one is “sensor-id=1”. Which entries does it look at in the dt ?
please update the camera-facing information in your sensor device tree.
please refer to [Sensor Driver Programming Guide] in the [L4T Documentation] and please read the [Module Properties] session for the position property.
thanks
I read there ‘The positions supported depends on the number of cameras in the system:
Two-camera system: rear and front’
Does that mean that I cannot choose freely that part of the ‘badge’ property or the ‘position’ property and that I must name them ‘rear’ and ‘front’ even if that does not match a physical reality ?
I would like to have sensor-id=0 use /dev/video0 and sensor-id=1 use /dev/video1, but that’s not what happens, regardless of if I put position=“rear” in module0 and position=“front” in module1, or the opposite.
How can I configure that sensor-id=0 uses /dev/video0 and sensor-id=1 uses /dev/video1 ?
You actually meant “Below case will set the spi0.0 as SENSOR-ID=0 and spi1.0 as SENSOR-ID=1 (module0 will be given sensor-id=1 and module1 will be given sensor-id=0)” ?