grub-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Chainloading GRUB2 from Lilo


From: Vladimir 'φ-coder/phcoder' Serbinenko
Subject: Re: Chainloading GRUB2 from Lilo
Date: Fri, 02 Apr 2010 21:53:03 +0200
User-agent: Mozilla-Thunderbird 2.0.0.22 (X11/20091109)

Tim Booth wrote:
> Hi All,
>
> Sorry to clog up the developer list with my questions but I've scoured
> the documentation and not found an answer.
>
> I've hit this known problem:
> http://sourceforge.net/apps/mediawiki/bootinfoscript/index.php?title=Boot_Problems:Windows_Writes_To_MBR
>
> Where broken software on Windows is clobbering GRUB2.  Modifying the
> corporate Windows installation is not an option, so I'm trying to use
> Lilo in the MBR and immediately chainload Grub.
>   
Asking such a lilo question on GRUB list is annoying, especially that
it's not a real solution.
Could you try installing grub to mbr, then dump the MBR gap with
dd if=/dev/sda of=before.bin count=64
<let windows do its anti-competitive job>
dd if=/dev/sda of=after.bin count=64
Then send both files to me? Or just say which sectors are modified.
It was reported that only sectore number 32 is modified. If it's
confirmed we can avoid using it unless we need the whole MBR gap (very
uncommon)

> Now I could install Grub into the (Ubuntu) Linux partition, but it seems
> rather simpler to just ask Lilo to chainload core.img.  As far as I can
> tell, the only way to do this is to use lnxboot.img, telling lilo that
> core.img is an initrd. - ie. like this:
>
> ---my /etc/lilo.conf---
> # Very basic LILO chainloader config
> boot = /dev/sda
> lba32
> default = chainload_grub2
> timeout = 0
>
> image = /boot/grub/lnxboot.img 
>   append = "prefix=(hd0,5)/boot/grub"
>   root=/dev/hda5
>   initrd = /boot/grub/core.img
>   label = chainload_grub2
> ---end---
>
> But when Grub loads the prefix is set to (hd0,1)/boot/grub and I just
> get a recovery prompt and have to do:
>
> set prefix=(hd0,5)/boot/grub
> insmod normal
> normal
>
> So, is there any way to get the correct prefix passed through?  I know
> there are other ways to attack this problem, but as the fix needs to be
> maintained on many machines I'm looking for something as simple as
> possible, and this is the simplest I can come up with, so I thought I'd
> ask this list before abandoning the idea completely.
>
> Cheers,
>
> TIM
>
>   


-- 
Regards
Vladimir 'φ-coder/phcoder' Serbinenko


Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]