It's more like:

	1) Kernel input device driver notices a Magic Button was pressed.
	2) Kernel tells init about the Magic Button.

I don't think it's a good idea to add this mess to init.
Those magic buttons are console resources and should be handled
by whatever or whoever is controlling the console.

It would be better to have a different daemon sitting on the
device attempting to detect the magic button and have it
issue the required system calls.  That way it can pop up 
a message box to verify the action if X11 is running, or
ignore it of nobody's logged in, etc.


This also brings up the issue of detecting the magic
key sequence.  These codes vary by input device.  Come
to think of it, it may be better to have the whole issue
handled by a userland daemon that sniffs the console input 
stream and detects the sequence itself.  It's easier to do
in userland and results in much less kernel bloat.

Eduardo