In my case I was interested in something that would allow a human to triage it after the fact and potentially many reboots later. In our use cases we assume the correct things will have been bound & its really the exception when they haven't been, but stil it was desireable information. In one sense it was just easier to add the logging to the kernel than it would be to get from udev. I think the kernel log must also be a more familiar source to most than udev.
