Re: [PATCH net-next] selftests: net: update wait mechanism for relocating

[Thaison Phan <thaisonphan@google.com>]

> I don't get it. I never saw it hang. And if the event is scoped udevadm settle
> should return immediately, rather than hang. We generally run on latest
> systemd udev in Fedora and CentOS. Please explain your problem better.

Thanks for reviewing this. I looked into my issue further. I found out when I
run into the hanging problem, the machine that I run on is using an internal
Linux distro that turns out to be running an old version (182) of udev before it
was merged to be a part of systemd. The source code for udev I was referencing
was out of date where it looks like udevadm settle relied on an internal uevent
sequence number to match up with a kernel uevent sequence number. At some point
I think it was a known problem that these sequence numbers could become out of
sync because udev would never get uevents not tied to the default namespace. I
checked the upstream udev and saw that udevadm settle was refactored, so that
sequence numbers are now deprecated. This patch is only applicable for versions of
udev without the refactor, so there is probably no reason to upstream it.

> Should be a separate patch
Sounds good. I'll create a separate patch for the exception fix.