udev in initrd

[Jeremy Katz]

On Mon, 2004-07-05 at 11:44 +0200, Owen Taylor wrote:
> (*) I know some people (yes, you, Jeremy) have serious reservations
>     about the details of how udev is implemented. And I really
>     just mean "dynamic creation of device nodes on a ram filesystem".
>     But I don't think the fact that udev is too 
>     bloated/complex/policyless/whatever to run in Anaconda should keep
>     us from trying to start fixing these problems for installed
>     systems.

It's not anaconda that worries me the most.  If we're not doing any sort
of different device naming, then with my anaconda hat on, I'll just
stick my fingers in my ears and pretend it doesn't exist.  "Dynamic
creation of device nodes on a ram filesystem" actually already happens
in anaconda, so it's not like it's a foreign concept or something that
I'm completely against.  I have my doubts as to the value of it on a
general purpose system, but in your weird diskless case, sure, it makes
sense.  Granted, I can implement it as something far simpler than udev.

My bigger concern is that udev has _zero_ policy.  It basically is a
"well, we want to let people do what they want" system.  Which is no
better than doing nothing at all.  And then, when you try to put it into
initrds, you have to allow the full range of shell utilities which is
just absurdity.  If we're willing to say "this is our policy, if you
change it, you get to make changes to other things too so that they keep
working", that's fine and then udev could be almost reasonable (although
I still think it's overkill)

Jeremy