[As I'm not the submitter of this bug please Cc: me directly.]
Hi,
looking at the BTS about grub and mapping problems I found
this bug. I have a similar problem with the current grub
(0.5.96.1).
The scenario appears a bit complicated, but is usable with
lilo and works (somehow) with grub (although the latter not
quite correctly).
I have Windows 98 on my first harddisk, Linux on the second.
Normally, I use loadlin to start Linux after booting from the
first harddisk. However, sometimes (don't ask why :-) - I love
to experiment...) I let the BIOS switch the hard drives and
boot from the second. I created a lilo.conf, which manages
this quite well (i.e. boot linux or DOS/Windows after swapping
the drives). Now I wanted to test grub. I was able to install
grub after some time (btw, the documentation does not describe
why grub ever could write screen after screen full of 'GRUB' ;-).
Linux boots fine now, even Windows boots. But:
I have a primary DOS partition (i.e. Windows C:) on this first
hard disk and an extended partition with drives D:, E:, F:, G:.
After using grub, Windows fails to read anything from the drives
in the extended partition, but it appends new drives H:, I: a.s.o
which are actually D:, E: a.s.o. When I switch the drives back
in my BIOS everything works again. And it did work with lilo.
Well, two little details: I used the device command in the grub
shell to switch drives before installing, and the menu entry for
Win95 contains map (hd0) (hd1) and map (hd1) (hd0), just like it
is written in the docs. BTW, shouldn't map work from the commandline
in the grub shell and while booting? You can verify that it does not
by checking the file system as output of grub's root command.
I hope I was clear enough, feel free to ask otherwise.
Greetings,
Ulf
_______________________________________________
Bug-grub mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/bug-grub