On Mon, Jun 26, 2017 at 09:09:04AM -0700, Rick Moen wrote:
> Quite so. Reading netlink for kernel uevents seems like a solution for
> a problem I happily don't have.

I'm afraid that with this mindset you're acting like Lennart, who ignores
valid and widespread use cases that just happen to not affect his laptop.
Static devices is a setup that works on only straightforward systems.

Thus, while it is worth supporting that on minimal containers, those are not
systems that boot on their own. Even the crummiest semi-embedded machine
that has enough oomph to run a generic distribution has so much power that
trying to avoid netlink is a waste of time. It'd be like coding an init or
an editor in assembly.

Thus, it's better to simplify things and standardize on a solution that
strives to work for everyone (ie, udev/eudev/vdev/new mdev with netlink
support).


Meow!
--
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ A dumb species has no way to open a tuna can.
⢿⡄⠘⠷⠚⠋⠀ A smart species invents a can opener.
⠈⠳⣄⠀⠀⠀⠀ A master species delegates.