> 
>   While I await data from Intel on the 82586, I thought I'd try to
> teach if_le.c to not find an AMD chip, since it isn't there.
>   Am I correct that it is not yet implemented? 

correct.

>   A way to check the contents of an address without incuring a page
> fault? 
> 

not exactly. i vaguely remember that if you setup a virtual mapping to a
"non-existing device" that you will receive a bus error.  If you
properly initialize p->p_addr->u_pcb.pcb_onfault (probably via
setjmp()) then i believe it will treat the fault as non-fatal and let
you know about it (longjmp().

The pcb_onfault stuff is from the hp300 port, so i suggest you look
there for references to it.

later,
Adam Glass

------------------------------------------------------------------------------