Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 24 Mar 2000 01:26:17 +0000 (GMT)
From:      Terry Lambert <tlambert@primenet.com>
To:        dcs@newsguy.com (Daniel C. Sobral)
Cc:        wes@softweyr.com (Wes Peters), mwlucas@blackhelicopters.org (Michael Lucas), advocacy@FreeBSD.ORG
Subject:   Re: New article
Message-ID:  <200003240126.SAA05518@usr02.primenet.com>
In-Reply-To: <38DAB25B.E2BBC400@newsguy.com> from "Daniel C. Sobral" at Mar 24, 2000 09:10:03 AM

next in thread | previous in thread | raw e-mail | index | archive | help
> > > Legacy hardware will still need to be hand configured (though not
> > > necessarily built in the kernel), and some kernel options are probably
> > > unavoidable.
> > 
> > But could potentially be configured through a loader script, rather
> > than compiled into the kernel.
> 
> The legacy stuff, yes. I said so. :-) The kernel options... As I said
> *some* as unavoidable. INVARIANTS?

Get rid of all invariants.

It has never made sense to me that the number of PCI busses or
the number of SMP per-CPU contexts, or the number of mbufs, or
any of dozens of other things were actually required to compile
the kernel.

For "I486_CPU" and similar things, well, I think this is really
a silly thing to do, instead of loading the generic code, and
then later loading the CPU specific code, after you know the
CPU type, and then _unloading_ the generic code it replaced.

The only real issues to work out are debugging options, which I
have no problem with forcing the debugging engineer to place on
the command line, or use a special build process.

The MATH_EMULATE, INET, NFS, etc., are a no-brainer.  The
boot filesystem has to be set, but that's always FFS (unless
you do a lot of fixing of the kernel, and then specify an
alternative on the command line).

VISUAL_USERCONFIG?  The loader should load it, if it is needed,
and unload it after it's done being used.

Hell, the root device driver stuff should be done in BIOS or
PPCBug or OpenBoot or ARCBIOS or SRM _or whatever_, and that
should get replaced by a dynamically loaded hardware specific
root device driver later... if there is one.  If not, it should
single thread  through the BIOS, or whatever it has to do, and
_just keep working_.

Psuedo devices?  If they are listed in a config file as being
permissable to load, then put a node in devfs in /dev, and if
it's referenced, load them then.


Invariants, inschmeriants.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-advocacy" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200003240126.SAA05518>