Trying to get my Roboteq 2130 controller hooked up. lsmod shows cdc_acm device and dmesg shows it connected to ttyACM1 but there is no ttyACM1 device in dev. There is a directory /sys/class/tty/ttyACM1. But for some reason its not getting created in /dev.
The first ttyACM would normally be “/dev/ttyACM0”. What do you see from “ls /dev/ttyACM*”?
In case it failed to use udev to rename the device it would have a standard “/dev/ttyS#” syntax. “udev” is what renames devices to alternate names.
Before connecting the device, run this in a couple of terminals (have both running at the same time from different terminals…it’ll tell you about changes):
watch -d -n 1 ls /dev/ttyS* watch -d -n 1 find /dev -name 'ttyACM*'
…see if maybe it is there but not using a udev rule.
ok I’ll try that. Currently there are no ttyACMx devices in /dev I have two devices the imu and roboteq are both acm devices.
strange when I did the watch commands (I had to use -n 1 it said -s was not recognized) they both showed up. Weird.
Anything going on with udev renaming doesn’t surprise me too much. Udev has resulted in much confusion (but I am a fan of udev, it was a good idea).
Yea I like it too. A bit hard at first but its very useful. The only thing I can think of is I added nvidia user to dialout before I ran those commands. New rover is up and turning wheels under ROS now. Might be able to test it outside tomorrow if it doesn’t rain some more. Had to make some changes to stack for ROS melodic and compile a few nodes from source but other than that it was straightforward to port it over to the Xavier.
Oops…corrected the “-s 1” to “-n 1”.
I just reflashed with 4.1 and the issue reappeared. using the two watch commands brought them back. So something isn’t quite right.
Perhaps the kernel module is not loading until some event occurs (assuming it is in the form of a module). You might see if the module shows up with “lsmod” when the device appears to be missing, and then again after the device seems to appear.