On Thu, May 26, 2005 at 03:42:56PM +0200, steef wrote:
> Ionut Georgescu wrote:
> 
> >We should wake up people, both of them ar wrong!!! :-)
> >
> >First, we are loading the kernel itself as an initrd image :-))
> >Second, it cannot be append, because initrd is the boot loader's task.
> >
> >So:
> >
> >image=/boot/vmlinuz-2.4.27-2-686
> >    label=Linux-2.4.27
> >    initrd=/boot/initrd.img-2.4.27-2-686
> >    read-only
> >
> >lilo
> >
> >On Thu, May 26, 2005 at 08:59:39AM +0200, steef wrote:
> > 
> >
> no!  if you do a kernel-upgrade under woody p.e. from 2.2 to 2.4, and 
> you start with the kernel-headers you must after that of course install 
> the corresponding kernel-image. in my experience the chap who put this 
> 2.4 or other image in kpackage put a thread in it. this thread asks you 
> if: 'you have installed from scrap, uh?? and then 'debian/kpackage' asks 
> you if you will be so kind to add a line in lilo: something like 
> initrd=/initrd.img and asks again if it must make a symlinc from lilo to 
> the kernelimage in /boot. at that point. if i remember well, 
> initrd-tools-package is already installed.
> 
> kpackage proceeds asking if it will change lilo-make-up. i said allways 
> 'no' at that point.
> further try out the procedure for yourself to fresh up your memory 
> completely. last time i did this was i guess two months ago.
> 
> kind regards,
> 
> steef
> 
> 

You may have answered another message, but not mine :-)

Most probably you are right. Last time a made an upgrade from 2.2 to 2.4
was 3 years ago and I still don't like /vmlinuz and /initrd.img as symlinks to
/boot anyway.

But that was not my point. My point was that our proposal for the
contents of lilo.conf was wrong for pure technical reasons:

        1. you don't load a kernel, but an initrd.img as initrd image.
        This was on the line:
        initrd=/boot/vmlinuz-2.4.27-2-686
instead of
        initrd=/boot/initrd.img-2.4.27-2-686

        2. initrd is loaded by the boot loader into memory, so it
        cannot by passed as an argument to the kernel by append=""

Both of them were my mistake, but one has survived the list :-)

Cheers,
Ionut


-- 
***************
* Ionuţ Georgescu
* Max-Planck-Institut für Physik komplexer Systeme
* Noethnitzer Str. 38, D-01187 Dresden
* Phone: +49 (351) 871-2209
* Fax:   +49 (351) 871-1999 

Attachment: signature.asc
Description: Digital signature

Reply via email to