In article <[EMAIL PROTECTED]> you write:
>From: Brian Brunswick <[EMAIL PROTECTED]>
>Subject: Re: My experience with grub
>Date: Thu, 13 May 1999 12:47:02 +0100 (BST)
>
>> Thats right. In fact, even the operating systems booted by stage2 still
>> have to be within the first 1024 cylinders.
>
> Yes or no. That depends on your operating system. In GNU/Hurd, to
>place the boot images (gnumach and serverboot) beyond 1024 cylinders
>should be fine. If it didn't work, that's a bug. Don't you mistake
>Large mode for LBA mode?
>
I may well be - I've managed to avoid having to get into this sort of
stuff before.
Is large mode where the disk has a fake 63/255 geometry?
On my system, the test for some extended bios call in disk_io.c
succeeds. (I put a trace on it - I think I removed it again for the
patch.) I didn't look closely at the code though - maybe the LBA
support wasn't actually there. It certainly had problems with
partitions past 1024 cyl, which I just assumed were the limit. In
/sbin/grub it would work.
BTW, does anybody know why difference versions of fdisk seem to
interpret cylinder boundaries completely differently (even for same
disk geometry)? When I was shuffling partitions about last night, the
only way I could get partitions back after deleting them was to use
exact sector numbers again. (Which thankfully I had cautiously backed
up)
