Re: LF: Supervisor Stack


[Prev][Next][Index][Thread]

Re: LF: Supervisor Stack





On Fri, 6 Dec 1996, Rob Taylor wrote:

> 
> > 
> > On Thu, 5 Dec 1996, Shawn Walker wrote:
> > 
> > > 	The RAM.TXT file that comes with FARGO mentions a "Supervisor
> > > Stack."  Upon browsing the 68k programmers guide, I could not determine
> > > where, or what, the switch from the user's stack to the supervisor stack
> > > takes place.  Could anyone shed some light on this for me.  Also, other
> > > than exceptions and TRAP's, does the processor ever enter the supervisor
> > > state.  How?  Other than the ability to execute privileged instructions,
> > > and the setting of the supervisor flag in SR, does the change to
> > > supervisor state change anything else.
> > 
> > The stacks are switched whenever the mode is switched, what's strange about 
> > that? The SP-register (AKA A7) Holds the pointer of the current stack.
> > 
> > Exceptions, traps, interrupts - that's it. The state is meant for 
> > operating system code, to protect it from usermode programs.
> > 
> 
> Just thought I'd say.. Fargo programs are run in supervisor mode...
> 
> (I rember the good old days of teh ST, when every program switched to 
> supervisor mode as pectically teh first thing they did!!)
> 
You mean:
	move.l #start, $80.w
	trap #0
start:

That's what many people did on the Amiga as well... BTW Which of the 
trapvectors in the TI92 is "free"? They can be practical to use in some 
situations...

//Sasq (Jonas Minnberg)


References: