I'm concerned there might be issues with kernel 4.15 in bionic that would change the expected output.
Could you check and confirm that even the oldest kernel (in that case 4.15) is still delivering valid messages on the initial problem "loading a non existing module while aliases don't exist".
If it behaves different with 4.15 the solution for Bionic needs to be more complex :-/
I'm concerned there might be issues with kernel 4.15 in bionic that would change the expected output.
Could you check and confirm that even the oldest kernel (in that case 4.15) is still delivering valid messages on the initial problem "loading a non existing module while aliases don't exist".
If it behaves different with 4.15 the solution for Bionic needs to be more complex :-/