Its probably just that way for historical reasons - in much the
   	same way as md_root.c, for which an MI version has been created
   	and ports are moving across.
   
   	I'd submit a PR on it for now, and if noone fixes it before you
   	commit your 5MX port to the tree :) :), deal with it then...
   
   
   On 18 Apr 2001, John Fremlin wrote:
   
   > Hi!
   >
   > I'm fooling around with a port to the Psion 5 MX PDA. In doing so I
   > noticed that cpu_reboot looks pretty architecture independent and is
   > more or less the same on all the archs I looked at. Why is it in the
   > arch dependent code?
   >
   

i'm not so sure this is true!  cpu_reboot() is certainly quite MD as
far as i can tell.  i checked a few and while they mostly change the
same variables (which _are_ MD variables, and not all are the same),
the job of cpu_reboot() is the MD things required to do for a reboot.
it is _designed_ to be an MD routine.  it often has special asm()
calls, etc.


.mrg.