>A device node doing the translation is my best suggestion at this point, as
>it doesn't have the problems of Solaris's "update daemon":  it catches
>old-utmp accesses where they happen, and needs no "cleanup" procedure.

Do you have an estimate on how much work this is?

If its essentially a standalone feature - does not permiate throughout
the kernel, can be an option (on by default in GENERIC) and esp if 
it has some use other than dealing with utmp, I don't have a problem 
with it.

[sorry if you've covered all the above previously and I missed it,
or am just reading this thread out of order]

--sjg