Subject: Re: More VS2000 booting problems
To: None <paule@shadowfax.martex.GEN.OH.US>
From: Ignatios Souvatzis <>
List: port-vax
Date: 06/02/1997 11:11:15
   Date: Sat, 31 May 1997 14:05:53 -0400 (EDT)
   From: Paul Evans <paule@shadowfax.martex.GEN.OH.US>
   X-Sender: paule@shadowfax
   cc: Boris Gjenero <>, port-vax@NetBSD.ORG
   MIME-Version: 1.0
   Content-Type: TEXT/PLAIN; charset=US-ASCII
   Sender: port-vax-owner@NetBSD.ORG
   Precedence: list
   Delivered-To: port-vax@NetBSD.ORG

   On Sat, 31 May 1997, Jason Thorpe wrote:

   > On Sat, 31 May 1997 10:27:17 -0400 
   >  Boris Gjenero <> wrote:
   >  > > - a boot partition selector
   >  > > - a boot loader
   >  > > - a program to configure both former components.
   >  > 
   >  > LILO isn't that bad.  Just because it isn't your traditional Unix
   >  > bootloader and it does multiple things doesn't mean that it's a horrible
   >  > mess.  I haven't ever had problems using it as my OS selector and it has
   >  > some neat features for x86 such as allowing you to boot from the second
   >  > floppy drive.
   > Err, given how the bootloader/configurer work, it certainly is pretty
   > bad.  As I recall, LILO has no standalone file system reader package.
   > This means that the configurator has to hard-code the actual disk block
   > numbers of the kernel into the bootloader.  That is a serious "ick"
   > in my book, and makes it totally useless as a bootloader.  (So much
   > for booting alternate/test kernels!)

   This is one of the few things I always found attractive about LILO: dosen't
   this allow chain-booting to another loader or OS -- without having to
   know the details of it?

   >  > compared to LILO.  I like the fact that it understands the filesystem
   >  > and can load any file.  I don't think that more time should be wasted on
   > ...ah, you see my point :-)

   Maybe for the VAX this is inconvient, but because LILO is PC based I
   think it does it's job quite well. Although not as well as the Solaris x86
   boot loader.. (MDB ? I think?)

   Anyway the feature I'm always missing on non-sun3 machines is the ability
   to do some fairly rudimentry hardware checks. I know DEC's pmax machines
   have almost decent boot strapper that does this, but what about the VAXen?

   And, wouldn't be nice for all NetBSD machines to have an identical
   "phase 1" boot loader with diagnostics?


Well, in theory, maybe.

But many NetBSD machines provide pretty nice boot loader & selector
semantics from the boot rom. (Eg., Sun3, Sparc, Amiga ).

Only think a bootloader needs to do there is to boot the kernel, maybe
telling it about single/multiuser.