Your application using Hailo8 crashes and the inference returns some errors, and you are not able to restart the application because the Hailo-8 seems to be stuck. The only way seems to reboot the system.
What are the steps to take when you are in such a situation ?
Try first to run sudo dmesg | grep hailo
, and if there is any suspicious error such as segmentation fault error.
If yes, it is possible that there is a bug in your software implementation that is causing the issue. You can inspect the file ‘hailort.log’, which is generated in the same directory as your application and look for clues.
You are now wondering if there is a way to reset the Hailo-8 without rebooting so you can debug the issue more comfortably?
You can try running these commands:
sudo modprobe -r hailo_pci && sudo modprobe hailo_pci
It will remove the driver and reload it and in the process will reload the firmware.
Verify that the driver is corrected loaded and instantiated by running:
lspci -knn | grep hailo
You should see the following outputs:
Kernel driver in use: hailo
Kernel modules: hailo_pci
and lastly check the status of the dmesg by running sudo dmesg | grep hailo
again.
If there is no error, then you should be able to run your application to infer again.