Subject: Re: putting functionality in firmware vs OS
To: David Seifert <seifert@sequent.com>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: tech-kern
Date: 07/29/1998 13:05:00
On Wed, 29 Jul 98 12:28:21 PDT 
 David Seifert <seifert@sequent.com> wrote:

 > Why does the PCI spec care whether the code that mucks with these
 > registers is linked into the lump called firmware or the lump called
 > the OS?

Because the system firmware knows best which devices are present, etc.
The system firmware also knows best about e.g. interrupt routing; I don't
want to program the core logic of a system to properly route the interrupts.

Take, for example, an Alpha PC164 workstation.  There are several flavors,
the EB164, PC164, PC164SX, and PC164LX.  Each of these have slight differences
in how interrupts, etc. are routed, devices are mapped, what have you.  But
because the console firmware does its job, they present the same software
interface.

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                            Home: +1 408 866 1912
NAS: M/S 258-5                                       Work: +1 650 604 0935
Moffett Field, CA 94035                             Pager: +1 650 940 5942