[2008] in linux-scsi channel archive
Re: LILO, mixed SCSI and IDE
daemon@ATHENA.MIT.EDU (Michel LESPINASSE)
Sun Jun 15 09:08:35 1997
Date: Sun, 15 Jun 1997 13:51:49 +0200 (CEST)
From: Michel LESPINASSE <walken@via.ecp.fr>
To: Keith Owens <kaos@ocs.com.au>
cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <19970615072345.954.qmail@mail.ocs.com.au>
On Sun, 15 Jun 1997, Keith Owens wrote:
> Just a heads up about a potential LILO problem. A SCSI machine had an
> IDE drive added as a temporary storage relief. The motherboard (ASUS
> P/I P65UP5 P55T2D) was told to boot SCSI before IDE (which worked) but
> running LILO 0.19 to pick up a new kernel got warning messages. LILO
> thought the SCSI disk was not drive 0. Next reboot hung on "LI", reach
> for the emergency floppies.
I had this problem once, too...
> Turned out the bios was reporting the IDE drive as 0x80 and the SCSI
> boot as 0x81. Solution was to add
> disk=/dev/sda
> bios=0x80
> to lilo.conf. I still get warnings but at least it boots now.
Same here. But I dont think that Linux uses informations reported by the
BIOS : on my machine I wanted to boot on SCSI too, but my BIOS didnt
offered this option. The solution for me was to pretend to the BIOS that
I didnt had any IDE drive - like, "drive C: None" - so the BIOS boots on
SCSI and Linux is still able to see and mount my IDE drive, because he
does not use BIOS services.
I think the problem is in LILO : probably it just lists the available hd*
and sd* devices, and assigns numbers sequentially. Is there a better way
for an userspace program to find the device numbers used by the BIOS ????
Michel "Walken" LESPINASSE - Student at Ecole Centrale Paris (France)
www Email : walken@via.ecp.fr
(o o) VideoLan project : http://videolan.via.ecp.fr/
------oOO--(_)--OOo-------------------------------------------------------
Any sufficiently advanced bug is indistinguishable from a feature.
