Kernel LISP - how low down can it go?

William A. Barnett-Lewis wlewis@mailbag.com
Wed, 21 May 1997 23:12:47 -0500


At 07:45 PM 5/21/97 -0700, you wrote:
>
>>From: "Dwight Hughes" <dhughes@intellinet.com>
>>To: <lispos@math.gatech.edu>, "BRIAN SPILSBURY" <zhivago@iglou.com>
>>Subject: Re: Kernel LISP - how low down can it go?
>>Date: Wed, 21 May 1997 21:19:57 -0500
>>
>>| From: BRIAN SPILSBURY <zhivago@iglou.com>
>>| 
>>  [  snip  ]
>>| 
>>| I've been thinking about this for a while, I think the best solution
>>| may be to extend the lisp-set into covering the assembly-level
>>operations.
>>| 
>>| (bnz #r0)
>>| 
>>| etc (while keeping this to a minimum)
>>
>>
>>CMU-CL defines some assembly level operations but I've just begun to look
>>at what they have done.
>>
>>If we do this we might want to define a nice generic set of assembly
>>language instructions with mapping to the particular CPU and
>>architecture being used, though this might be more trouble than
>>it's worth. Where has this been done before?
>>
>>-- Dwight
>>
>
>  Why do you think you need access to the underlying machine's
>individual instructions? About the only thing I can think of that's
>written in assembler in the version of Unix I'm familiar with is the
>initialization code (setting up the CPU) and the first level TLB MISS
>handler. Nothing else needs to be in assembler.
>
>  Mike McDonald
>  mikemac@engr.sgi.com
>
>
William A. Barnett-Lewis
wlewis@mailbag.com
-------------------------------------------------------------------------
"We are artists.  Poets paint motion  and light.  Historians paint stills.
It can be dangerous to get history from a poet.  It can also be the greatest
blessing."
						Larry Miller Murdock
-------------------------------------------------------------------------