On Mon, 3 Apr 2000, Andrew Brown wrote:

# >Now, if we can do a skeletal runlevel implementation (i.e. we handle things
# >as we do now:  We have a single-user runlevel and multi-user runlevel with
# >the other levels to be filled in as needed), it probably won't be so bad.
# >If we can do it *right*.  IF.  BIG - I say, BIG IF, there.
# 
# allow me to interject here for a second.  as far as i can
# "pedantically" count them, we currently have four (er...five) run levels:
# 
#  a) single user
#  b) multi user (which include networking and nfs stuff)
#  c) hardware (or "press any key to reboot") [1]
#  d) power off (for people for whom apm or whatever can turn the machine off)
#  e) rebooting
# 
#  [1] do sparcs go to the ok prompt here?  i don't have one, so i don't
#  know.

I think that state is better referred to as 'halt'.  i386s don't halt,
really, they pause, but: powerdown/halt/reboot/single/multi, yeah, that's
about right.  Oh, yeah: [1]: Yes.

# i never really understood the distinction between S (and s) and 1 in
# solaris.  calling them different "run-levels" seems pedantic in and of
# itself.

I'd like to see us skip the single-letter naming convention for run-levels
and just run symbolically.

I think, by the way, that 1 will actually shut some other things down before
going to S.

# in the spirit of bsd...i think it's fair to say that you can do that if
# you want to.  hey!  why not roll up some init patches to make it spew
# the contents of various files at various "levels"?  :)
# 
#    printf("New run level: %ld\n", random());

Well, :-), not quite.

				--*greywolf;
--
BSD: For The Network Generation