*BSD News Article 34102


Return to BSD News archive

Xref: sserve comp.os.386bsd.questions:12222 comp.os.386bsd.misc:3108
Path: sserve!newshost.anu.edu.au!munnari.oz.au!bunyip.cc.uq.oz.au!harbinger.cc.monash.edu.au!msuinfo!agate!spool.mu.edu!bloom-beacon.mit.edu!ai-lab!life.ai.mit.edu!mycroft
From: mycroft@duality.gnu.ai.mit.edu (Charles M. Hannum)
Newsgroups: comp.os.386bsd.questions,comp.os.386bsd.misc
Subject: Re: BT445C == No FreeBSD???
Date: 09 Aug 1994 15:22:52 GMT
Organization: MIT Artificial Intelligence Lab
Lines: 38
Message-ID: <MYCROFT.94Aug9112252@duality.gnu.ai.mit.edu>
References: <31vs36$qgv@tekgen.bv.tek.com> <Cu4w68.99B@tfs.com>
	<DERAADT.94Aug6165915@newt.fsa.ca> <3272h7$9ip@u.cc.utah.edu>
NNTP-Posting-Host: duality.ai.mit.edu
In-reply-to: terry@cs.weber.edu's message of 9 Aug 1994 05:00:23 GMT


In article <3272h7$9ip@u.cc.utah.edu> terry@cs.weber.edu (Terry
Lambert) writes:

   Uh, what determines the device order?

The 3c509 specification determines the order.  I don't have it handy,
so I can't give you an exact quote.

   Specifically, it isn't sorted by port, and it isn't sorted by IRQ
   (that you can tell, anyway) and it isn't sorted by hardware address
   (unless it's reverse sorted order)... is it by who replies first?

All three of those (as well as the existing approach of not dealing
with it) have failure modes.  One case where they might all fail is
replacing a card, or moving it to a different I/O address or IRQ
because you need to add a SCSI card.

Currently, you can specific explicit addresses for any of the cards
and be guaranteed a specific binding, if that's important for you.
The existing mechanism is not a replacement for site-specific or
machine-specific configurations; it's mainly intended for generic
kernels that can be used for installation on a wide variety of
machines.  (And I might add that it works quite well for that.)

I don't think it's reasonable for the kernel to go out of its way to
avoid obvious failure modes, only to create new less-obvious ones.
The user is supposed to be intelligent; the kernel is a slave.

   Be a real bummer to make a mistake.

Yah, someone might die!

--
- Charles Hannum
  NetBSD group
  Working ports: i386, hp300, amiga, sun4c, mac68k, pc532, da30.
  In progress: sun3, pmax, vax, sun4m, alpha, sun4.