Subject: kernel vs boot-program
To: None <>
From: Gordon W. Ross <>
List: tech-kern
Date: 03/14/1994 16:49:32
> Date: Mon, 14 Mar 1994 10:59:55 -0800
> From: Adam Glass <>
> the net-2/4.4 diskless implementation works as follows:
[ boot program patches all needed diskless info into the kernel ]

> [Adam]/Theo? architecture:
[ kernel gets diskless info itself, without help from boot program ]

> advantages:
> 	you can strip your kernels :)
> 	bootstrapping program doesn't have to understand symbol table
> 	your bootstrap program can do something awful (like tftp)
> 		if you don't want to implement nfs (see libnetboot)
> 	source of kernel does not inherently imply root/swap source.
> 		maybe easier to do dataless.
> 	boot -a can  do the right thing.

> your thoughts?

I too would prefer that the boot program not be expected to patch
data in the kernel.  One handy result is that I can use Sun's boot
program with a netbsd kernel.  At the moment my Sun3 loads its
kernel from disk and then does nfs_mountroot (asks for the info).

I wonder if it might be convenient to have something in-between
the two approaches (kernel vs boot-program) for getting the NFS
diskless info:

For example, if the boot program could pass either some argument
strings or and environment to the kernel's main, then the kernel
would not need to know how it was loaded or where its diskless
information came from.  I favor strings for the interface between
the boot program and the kernel because it makes "boot -a" easy.

Is it generally feasible to pass either argument or environment
strings to the kernel?  (I know some systems support this.)

Gordon W. Ross  		Mercury Computer Systems      		199 Riverneck Road
508-256-1300			Chelmsford, MA 01824-2820