No, a "privileged instruction fault" is something you'll find on any 
CPU that has more than one access mode (a "user" mode and a
"kernel" mode, a la VAX).  The x86 has two such spaces of which
I am aware when running in protected mode: user mode and
supervisor mode; the message that Jukka Marin sees is almost
redundant considering that the kernel is ONLY going to get
a "fatal privileged instruction fault" in supervisor mode;
if it happens in user mode, the CPU will probably trap on
it and return another error code of some sort; PrivInsFaults
don't typically crash the processor in user mode.

IIRC, "Illegal instruction" is generated when you tell
the processor to execute something it doesn't recognize, though
it might be what UN*X(-alikes) print when you try to
execute a privileged instruction in user mode as well.


> ----------
> From: 	Jason Thorpe[SMTP:thorpej@nas.nasa.gov]
> Sent: 	Saturday, May 17, 1997 9:55 AM
> To: 	Jukka Marin
> Cc: 	NetBSD i386
> Subject: 	Re: new kind of crash under 1.2 1/2 :) 
> 
> On Sat, 17 May 1997 18:37:21 +0300 
>  Jukka Marin <jmarin@pyy.jmp.fi> wrote:
> 
>  > This seems to be repeatable.  Whenever my wife is playing with
> Netscape,
>  > my machine crashes with the same eip and cs values:
>  > 
>  > fatal privileged instruction fault in supervisor mode
>  > trap type 0 code 0 eip f81b0cc0 cs f81b0008 eflags 10246 cr2 1b3348
> cpl e000d060
>  > panic: trap
>  > 
>  > The only running processes were des and netscape.
>  > 
>  > I understand that people aren't too interested in what's going on
> on a 1.2
>  > system, but I'd still like to know how to fix this problem.  I
> would upgrade
>  > to -current if I knew it worked well - but I see so many problem
> reports
>  > every day that I just can't do the upgrade yet. :-I
> 
> ...well, for starters, building a kernel with "makeoptions DEBUG=-g"
> would allow you to get some meaningful information with gdb
> post-mortem.
> 
> Privileged instruction fault... is that an illegal instruction in
> i386-speak?
> (Sigh, I find m68k traps so much more straight-forward :-)  It's not
> likely
> that something is scribbling over kernel text; that would cause a page
> fault trap.
> 
> Jason R. Thorpe
> thorpej@nas.nasa.gov
> NASA Ames Research Center                               Home:
> 408.866.1912
> NAS: M/S 258-6                                          Work:
> 415.604.0935
> Moffett Field, CA 94035                                Pager:
> 415.428.6939
>