Subject: Re: CVS commit: src/sys/arch/i386/conf
To: Manuel Bouyer <>
From: Julio M. Merino Vidal <>
List: source-changes
Date: 02/05/2006 17:17:54
El 05/02/2006, a las 17:06, Manuel Bouyer escribi=F3:

> On Sun, Feb 05, 2006 at 03:39:20PM +0000, Julio M. Merino Vidal wrote:
>> Module Name:	src
>> Committed By:	jmmv
>> Date:		Sun Feb  5 15:39:20 UTC 2006
>> Modified Files:
>> 	src/sys/arch/i386/conf: XEN0
>> Log Message:
>> Enable Multiboot support by default since this kernel is the one =20
>> to most
>> likely benefit from this feature.  Per anonymous suggestion in my =20
>> blog.
> Hum. Does it really work ? =46rom what I understood of multiboot, it =20=

> contains
> both a way to pass environnement such as memory maps, and command line
> variables. The environnement is different and is not passed using =20
> multiboot
> I think, and command line variables are already handled properly.

Which "environment"?  If you mean all the bootinfo stuff, that's the
whole point of the code I added.  It takes the information passed in
by the Multiboot information structure (the one that contains the
memory map, the command line, etc.9 and converts it to native bootinfo
data so the NetBSD kernel behaves as expected.  Take a look at
arch/i386/i386/multiboot.c (specially, the setup_* functions).

It works fine (except for a possible problem with LKMs, something that
has just been reported to me).


Julio M. Merino Vidal <>
The Julipedia -