Comment 10 for bug 1729145

So, the kernel fix could look like this:

* instead of just failing with "device already registered"
* the kernel would emit another uevent with CACHED_UUID which would be processed another time to set up a symlink

It seems like this approach with "synthetic uevents" may be used for other purposes as well with changes that landed in 4.13

http://elixir.free-electrons.com/linux/v4.14.2/source/Documentation/ABI/testing/sysfs-uevent
"Enable passing additional variables for synthetic uevents that are generated by writing /sys/.../uevent file

Recognized extended format is ACTION [UUID [KEY=VALUE ...]

Users: udev, userspace tools generating synthetic uevents"