Branches
Releases
Comments
[»]
how to install LILO boot loader into a disk image
by v01dp - Jul 4th 2007 02:28:59
It is possible to create a disk image using a file, partition it, create
file systems, install a Linux distro, and install the boot loader of
choice. This file can then be copied to actual disk, used as a virtual disk
with an emulator (such as bochs), etc.
Below is a short example of how to install LILO into such virtual disk,
one that I actually made and using currently.
Assuming you have created already the disk image file, have partitioned
it, and attached it to loopback device(s) appropriately, this is the
relevant part of lilo.conf:
========lilo.conf =====================
# You have attached your disk image to the loop0
# The disk geometry is that how you've defined it
# when creating your disk image (fdisk .... )
# Within my disk image, i have 3 partitions.
# Note the start offsets (fdisk -lu on image)
# The partitions are attached to other loops,
# if you need to use them.
disk=/dev/loop0
sectors=63
heads=255
cylinders=9
bios=0x80
partition=/dev/loop1
start=63
partition=/dev/loop2
start=16065
partition=/dev/loop3
start=128520
# Some standard lilo things ...
compact
lba32
# MBR to install LILO to: loop0 is the entire disk
# image, so we install into MBR of that image;
boot = /dev/loop0
# the boot partition I have mounted for working at
# lo_2/boot
map = /void/wrk/lo_2/boot/.map
install = /void/wrk/lo_2/boot/boot-menu.b
prompt
timeout=450
delay = 50
vga = normal
default=Linux
# I have vmlinuz on the boot partition,
# which is mounted at lo_2/boot
image=/void/wrk/lo_2/boot/vmlinuz
label=Linux
initrd=/void/wrk/lo_2/boot/initrd.img.gz
read-only
========end lilo.conf ==============
I have to note that inside my initrd image, my linuxrc
sets up appropriate root device, then does pivot.
(actual root device can be say /dev/sda, or /dev/hda).
In shells like nash, you can run mkrootdev, which can
create needed device node passed as root=/dev/XXX
during boot.
So then you only have to run lilo -v3 -C mylilo.conf -t,
and if all is oK, remove the '-t' :)
In case you wanted to see the mounts :
#loop0 is the entire disk, partitions are mounted
/dev/loop/2 on /void/wrk/lo_2 type squashfs (rw)
/dev/loop/1 on /void/wrk/lo_2/boot type ext2 (rw)
/dev/loop/3 on /void/wrk/lo_2/mnt type ext2 (rw)
This was my first adventure of this sort, so I can't guarantee
that installing LILO this way is best and most efficient.
If you know of a better / alternative / another useful way of installing
LILO into a file, or how to improve the above way,
please post a comment.
[reply]
[top]
[»]
Deadlocked Installations
by lalurus - Feb 26th 2007 09:20:58
Hi:
Many installation instructions advise the installer to produce a boot
disk. But not all Linux and BSD distros allow that. Using floppies is
not as common practice than that once used to be. In the case of new
installations however it can save a lot of headaches. What even further
complicates the situation is that the believed rescue options fail to
deliver.
I do not like to destroy what works, rather prefer to extend its scope.
But this cautious approach led to a situation that I have completed the
first phase of two installations neither of which can be restarted
now.
One of them is openSUSE 10.2. Based on some loosely-worded advises on
installing multiple Linux distros on one drive, and using lilo to boot them
just has not worked out. The used lilo.cnf is listed below. Executing
/sbin/lilo resulted in skiping the last entry, i.e., the newly installed
system.
(Please note that on the first IDE controller /dev/hda is the CDROM, and
/dev/hdb is the DVD unit; /dev/hdc, the first hard drive is master on the
secondary IDE controller.)
### LILO configuration file
### Global Section START
append="enableapic vga=0x0317"
boot=/dev/hdc
linear
vga=normal
message=/boot/message
menu-scheme=Wg:kw:Wg:Wg
read-only
prompt
timeout=200
### Global Section END
### Windows_and_DOS START
other=/dev/hdc1
label=Win_n_DOS
loader=/boot/chain.b
table=/dev/hdc
### Windows_and_DOS END
### SuSE_old kernel-2433 START
image=/boot/vmlinuz-2433
label=SuSE_old-2433
root=/dev/hdc2
optional
### SuSE_old kernel-2433 END
### SuSE_old_orig START
image=/boot/vmlinuz.orig
label=SuSE_old_orig
root=/dev/hdc2
optional
### SuSE_old_orig END
### openSUSE-10.2 Linux START
image=/boot/vmlinuz
label=openSUSE-10.2
root=/dev/hdc17
optional
### openSuSE-10.2 Linux END
The second system is FreeBSD located in the forth primary partition.
Supposedly it should be booted though the fixit.flp floppy using the
correct
n1 and n2 in the n1:ad(n2,a)/kernel command.
But, no matter what I use, it replies with "invalid label".
Also lilo should be able to handle it, but I have not been able to find any
specific example for it.
If anyone having closer familiarity with the subject has any specific
advice to resolve either case, or could point to an article on lilo that
deals with cases involving two Linux systems located in two different
directories, and/or has a specific description of booting a FreeBSD system
I would greatly appreciate it.
[reply]
[top]
[»]
Re: Deadlocked Installations
by John Coffman - Feb 26th 2007 21:03:00
To address just one issue, the skipping of the installation on /dev/hdc17.
The 'optional' keyword is used, meaning if LILO cannot find the specified
kernel, the entire descriptor will be skipped. Omit 'optional', and the
installation will probably fail with a 'file not found' complaint.
'linear' is used in the global section. This limits you to booting from
the first 1024 cylinders of your hard disk. Is this really what you want?
[reply]
[top]
[»]
Re: Deadlocked Installations
by John Coffman - Feb 26th 2007 21:09:03
I don't know if this will solve the problem of booting FreeBSD, but I would
chain boot this installation on /dev/hdc4 with: 'other = /dev/hdc4'. Then
install the FreeBSD boot loader on /dev/hdc4. The LILO installation can
chain to it through the 'other' descriptor.
[reply]
[top]
[»]
Incompatibility with 2.6 kernel
by John Coffman - Dec 16th 2006 21:43:14
LILO versions 22.7.3 and earlier will have an undesired interaction with
the 2.6.18 (and possibly earlier) kernels if a hidden directory (.udev) is
created in the /dev device hierarchy. The result is that LILO loops if an
undefined device is encountered. Updated source code is available in the
alpha directory at the developer's site, labeled 22.8-alpha3. It should
only be tried if you are experiencing the hang with a 2.6 kernel.
[reply]
[top]
[»]
lilo, FC5, raid and initrd
by roamdad - Oct 11th 2006 16:09:57
Just a note for cross-reference.
See bugzilla.redhat.com - Bugzilla Bug 197701: mkinitrd is not creating
the appropriate root mount point resulting in a /dev/root kernel panic
On upgrading a lilo booted raid1 system to FC5 and retaining lilo as the
bootloader, easy temporary fix is to replace
root=/dev/md0
with
append="root=/dev/md0"
and rerun lilo.
If one doesn't have a backup OS to boot into, run rescue mode of FC5 and
chroot.
Maybe this will save someone a couple days tracking this down.
[reply]
[top]
[»]
possible improvement: a maintenance shell
by xpol - Aug 19th 2006 00:25:24
Hi,
I have been a happy lilo user for a dozen years now;
although not a computer science professional, i have
found lilo easy to configure and a very dependable
program.
I would like to suggest a possible improvement: since
very often linux distributions prompt user for system
upgrades (i am upgrading my debian/testing distrib.
almost every day), it would be desirable upgrading to
be accomplished at a very early stage in the boot
process, so that every updated program can be started
in the current session. That would especially needed for
X/xorg related packages.
So, i am wondering about lilo to open a sort
of 'maintenance shell' over the default kernel with
network services enabled, getting info about new
packages available and prompting user about what to
do, before booting.
thank you
--
pol
[reply]
[top]
[»]
Re: possible improvement: a maintenance shell
by John Coffman - Aug 19th 2006 11:43:06
If I am reading your request correctly, this sounds more like an item to be
handled by the kernel startup scripts, not the boot loader. The network is
not up and usable until well into the startup process. It is certainly
possible for the boot loader to pass a parameter into the kernel by way of
the boot command line; viz., "boot: Linux updatecheck". It would be the
responsibility of the startup scripts to read the boot command line
(/proc/cmdline) and take appropriate action.
[reply]
[top]
[»]
Long install times with Reiser filesystem
by John Coffman - Mar 29th 2006 15:10:39
Two users have broght to my attention that recent releases of the Reiser 4
File System with the 2.6 kernels produce very slow LILO bootloader install
times. The mapping of kernel sectors to absolute disk addresses seems to
have slowed down considerably. I have one report of 20 minute per kernel
boot loader install times. Actual booting is not affected. However, the
impact on those folks who compile their own kernels is very serious. The
only workaround I can suggest at this time is to create a small (32MB)
/boot partition formatted for ext2 or ext3. Then place all kernels,
initrds, and the /boot/map file in this partition. With LBA32 addressing,
there should be no problem with this partition placed anywhere on the disk.
[reply]
[top]
[»]
Re: Long install times with Reiser filesystem
by John Coffman - Apr 3rd 2006 13:26:34
Patching 2.6.14.7 kernel with the following: reiser4-for-2.6.14-1.patch
will NOT exhibit the problem. However, patching the same kernel with:
reiser4-for-2.6.14.6-1.patch WILL exibit the problem. Kernels 2.6.15 and
later all exhibit the problem.
[reply]
[top]
[»]
Re: Long install times with Reiser filesystem
by Dark - Jan 3rd 2007 19:16:00
I can confirm that. Really long times (~ 10 minutes per image).
Is there another way to fix this, except creating another partition for
boot purposes? Currently, I don't have a way to do it (a lot of data needs
to be backed up, to create another small partition)
--
Avoid the Gates of Hell. Use Linux.
-Registered Linux user #374579-
[reply]
[top]
[»]
Re: Long install times with Reiser filesystem
by Dark - Jan 3rd 2007 19:19:03
I'm sorry, one correction:
time lilo
Added slack_11 *
Added 2.6.18
real 12m10.079s
user 0m0.044s
sys 0m5.840s
--
Avoid the Gates of Hell. Use Linux.
-Registered Linux user #374579-
[reply]
[top]
[»]
Re: Long install times with Reiser filesystem
by John Coffman - Jan 4th 2007 08:13:59
The change has occurred in the Reiser filesystem code, not in LILO. The
fix will have to come from the Reiser development team. The more people
they hear from, the quicker this problem will be addressed.
[reply]
[top]
[»]
i/o error
by stephen - Mar 9th 2006 08:20:37
hey i admitt im a total noob when it comes to lilo, i used a partition
manager from pcformats(uK) newest dvd issue running from lilo, cant
remember the name of it for th life of me because im in work and cant get
on my o/s to look at the disk again, basically all i done was format my
hdisk from ntsf to fat32 to install win98 from win xp, all went succesfully
and i shutdown the prog , wouldnt go further than ran out of threads when
shutting down so need to shutdown with power button, loaded back up and
recieved i/o error please replace system disk, when no boot disk is
present. it works fine with other boot disks i installed windows 98
succesfully but it wont let me past the i/o error to let me onto windows,
any help plz guys ive been at this for days to no success
[reply]
[top]
[»]
Disabling init=1 ?
by carcassonne - Jul 20th 2005 14:53:14
Hi,
Is it possible to disable this 'feature' that lets anyone armed with a
keyboard to run /bin/bash when the system starts instead of login in
properly by interrupting the boot sequence ? Is this part of lilo or is it
the kernel that's built like that? Looks like a major security flaw.
The system in question boots without prompting or using a lilo menu but
there's still a way to interrupt the boot sequence and specifying to run
/bin/bash.
Thanks for any suggestions.
[reply]
[top]
[»]
Re: Disabling init=1 ?
by John Coffman - Jul 20th 2005 19:16:52
LILO has three password options: bypass, restricted, and mandatory.
A mandatory password requires the user to ALWAYS enter a password before
booting is allowed. A restricted password allows booting with no password,
as long as no kernel parameters are passed (such as the above). If kernel
parameters are specified, then LILO will require the entering of a password
before booting can proceed. Password protection may be global, or per
descriptor (boot item selector). See the man pages for lilo.conf for more
information.
[reply]
[top]
[»]
Save LILO
by amxor - May 25th 2005 12:12:07
Hi,
I 'm new to installing LILO and quite new to linux at all.
I have an old Libretto 50 CT on wich I want to install slackware 10.1.
The problem is that I have no floppy drive or other possible boot devices
but my harddisk.
Can anyone explain me how I can install LILO to this system 'totaly save'.
So I want it to boot win98 in all cases even when LiLo doesn't work. I also
need to install LiLo from this DOS/Win platform.
I can use partitionmagic... if the C: partition containing DOS/Win98
boots. I don't think the thing supports lba32...
Thx, Benjamin
[reply]
[top]
[»]
Re: Save LILO
by amxor - May 29th 2005 02:38:33
I see I made a huge typo in the subject.
But my problem is solved!
I installed ZipSlack.
Loading it with loadlin, I made an ext2 part. for it.
Next I tried to get it loading from that partition, still using loadlin.
Since everyting worked fine I installed lilo for booting DOS or ZipSlack.
Only the screen resolution was not configured right.
Thanks lilo and SlackWare
[reply]
[top]
[»]
Transferring boot sequence from syslinux to lilo
by carcassonne - May 12th 2005 07:27:37
Hello,
What I'm trying to do is to boot a PC using a USB stick and then offer
to choice either to continue booting the USB stick or to boot the hard
disk.
Both the USB stick and the hard disk are booting fine on their own.
syslinux is used for the USB stick. lilo 22.5.7.2 is used to initialize
the hard disk. Please note that the hard disk is initialized on a host
machine and is seen as /dev/hdc when doing so. The following lilo command
is then used:
lilo -r /mnt/disk -b /dev/hdc
With the following lilo.conf:
boot=/dev/hda
map=/boot/map
install=/boot/boot.b
default=linux
vga=791
image=/boot/vmlinuz
label=linux
root=/dev/hda1
append="splash=silent devfs=mount"
initrd=/boot/initrd-linux"
Then this hard disk is put into another PC as sole hard disk (i.e. hda).
It boots fine on its own.
Now, using syslinux.cfg I can get lilo to output a 'L' without any error
code, but that's as far it can get. Peter Anvin (of syslinux) has
commented that what lilo needs is to trust DL and that perhaps there's an
option for this.
I've browsed through the lilo documentation but couldn't find something
pertaining to making it 'trust DL'. Perhaps it is expressed in a different
way. Could you help with this ?
Regards,
Al
[reply]
[top]
[»]
Re: Transferring boot sequence from syslinux to lilo
by John Coffman - May 12th 2005 09:41:38
> Now, using syslinux.cfg I can get lilo
> to output a 'L' without any error code,
> but that's as far it can get.
>
>
'L' with no error code is pretty unusual with versions of LILO from the
last year or so. Do try 22.7 (current) before you go any further; 22.5.7.2
is a year and a half out-of-date.
> Peter
> Anvin (of syslinux) has commented that
> what lilo needs is to trust DL and that
> perhaps there's an option for this.
>
>
Yes, there is. "bios-passes-DL=" is an option that is at least 2 years
old (22.5.1 is earliest reference in the changelog.) I would try 22.7
first, then this option. It is also accessible from the command line as:
"lilo -Z1" (capital-'Z'). It is documented in the man page for
"lilo.conf" (bios-passes-DL) and "lilo" (-Z).
--John
[reply]
[top]
[»]
Different keyboard makes lilo show prompt
by carcassonne - Mar 31st 2005 06:43:53
Hi,
I install systems by replicating a master hard disk using a dedicated
machine to replicate disks. The lilo configuration file is simple and must
show no prompt since these machines have usually (i.e. when not developing
directly on them) no keyboards. They uses a touch screen instead.
On one machine, lilo always shows a prompt (i.e. the red box) and since
there's no timeout, waits for a keypress. This machine uses a different
keyboard and we just got confirmation from the IT dept. that this keyboard
does not work.
Do you have any idea as to the manner in which a defective keyboard can
prevent lilo from booting and instead makes it show the red box ?
Regards,
Al
[reply]
[top]
[»]
Re: Different keyboard makes lilo show prompt
by John Coffman - Mar 31st 2005 08:05:46
I assume you are not using the "prompt" switch in the configuration file.
If the missing / malfunctioning keyboard appears to transmit a TAB
character, then the prompt screen will appear. Although intended for noisy
serial lines, the "unattended" keyword may be effective for you. Just be
sure to have set a very short "timeout=", say, about the same as your
"delay=" setting.
The "unattended" switch first appeared in LILO 22.6.
[reply]
[top]
[»]
Website
by shockzor - Feb 23rd 2005 05:22:17
It would be good if the project would have a decent website describing what
features it have, known problems, a F.A.Q, a online manual, etc.
[reply]
[top]
[»]
Preparing a hd[x] drive for another machine
by carcassonne - Feb 11th 2005 17:30:16
Is it possible with lilo to make bootable a drive that
will be put into another computer to boot it ? For
instance, computer A has drive hda. I connect a
hdb drive to it, copy system and application files,
then I want to make it bootable so that I can
afterwards disconect it and put it into computer B as
computer B only hda drive ?
If possible, can you tell me how to use lilo to do
this ?
Thanks !
[reply]
[top]
[»]
Re: Preparing a hd[x] drive for another machine
by John Coffman - Feb 11th 2005 19:31:51
Make sure you are using LILO 22.5.9 or newer, and that you are using
'lba32' disk addressing. A sector-by-sector copy of a bootable drive (same
size) will be bootable:
>dd if=/dev/hda of=/dev/hdb
Then move hdb to hda on the second machine. This, of course, assumes a
compatible BIOS.
If /dev/hda1 is the bootable partition on the first machine, and /dev/hdb1
is to be the bootable partition on the second machine (/dev/hda1 after
moving the disk), then a simple install of LILO on the second disk (mounted
as /mnt/b), would go as follows (identical /etc/lilo.conf files on both
disks):
> lilo -r /mnt/b -b /dev/hdb
At this point, dismount, power down, and move the hdb disk to hda on the
second machine. Presumably even the "root=/dev/hda1" specification for hdb
will now become valid.
There are other ways to create a bootable disk for a second machine, but
they quickly increase in complexity from this point. I suggest you contact
me directly if your situation is significantly more involved.
[reply]
[top]
[»]
Re: Preparing a hd[x] drive for another machine
by v01dp - Jun 18th 2007 08:24:17
Hello,
You wrote that it is possible to contact you directly for more involved
case.
I was wondering about the 'other' ways to create a bootable
disk for another machine, and what is / not possible with lilo here.
I was interested if its possible to install lilo without having actual
hard drive , only the disk image (which I then would copy to the disk
).
I have made a disk image, with partitions, file systems, and a linux
distro installed. I was trying to install lilo into it, but with no luck.
I was trying to install lilo through the use of loopback
devices, but that seems did not work (installed, but does not boot). Its
most likely wrong what I was trying to do.
I would really appreciate help on this.
Thank you
> Make sure you are using LILO 22.5.9 or
> newer, and that you are using 'lba32'
> disk addressing. A sector-by-sector
> copy of a bootable drive (same size)
> will be bootable:
>
>
> >dd if=/dev/hda of=/dev/hdb
>
>
> Then move hdb to hda on the second
> machine. This, of course, assumes a
> compatible BIOS.
>
>
> If /dev/hda1 is the bootable partition
> on the first machine, and /dev/hdb1 is
> to be the bootable partition on the
> second machine (/dev/hda1 after moving
> the disk), then a simple install of LILO
> on the second disk (mounted as /mnt/b),
> would go as follows (identical
> /etc/lilo.conf files on both disks):
>
>
> > lilo -r /mnt/b -b /dev/hdb
>
>
> At this point, dismount, power down, and
> move the hdb disk to hda on the second
> machine. Presumably even the
> "root=/dev/hda1" specification for hdb
> will now become valid.
>
>
> There are other ways to create a
> bootable disk for a second machine, but
> they quickly increase in complexity from
> this point. I suggest you contact me
> directly if your situation is
> significantly more involved.
>
[reply]
[top]
[»]
Re: Preparing a hd[x] drive for another machine
by John Coffman - Jun 18th 2007 12:32:11
If you wish to prepare a disk image and then copy it to a hard drive, I
suggest you follow the example in the script "mkrescue --iso --size HD".
Normally the "--iso" option creates a bootable floppy image; however, with
the 2.6 kernels >1.4Mb this is no longer possible. So the extra size (Hard
Disk), allows for very large kernels. Although this script ultimately
creates an ISO image, the hard disk image is created along the way.
The 'mkrescue' script is a part of the LILO source and binary
distributions.
[reply]
[top]
[»]
Incompatible S-ATA controllers
by John Coffman - Sep 27th 2004 12:27:24
Hardware:<br>
- Davicontrol 4-channel S-ATA with Silicon Image 3114
Chipset</br><br>
- Longshine 2-channel S-ATA with Silicon Image 3112
Chipset</br><br>
- possibly others</br><br>
<br>
<br>
Both controllers report a BIOS copyright by Silicon Image. The BIOS
allocates memory in the EBDA area but does not record this usage in the
Bios Data Area word at 40:13. This makes both of the above controllers,
and any other SI controllers using this BIOS incompatible with other add-in
PCI cards that use the EBDA. It is also incompatible with DOS
applications, including those that run under Windows 95/98, which use
memory just below the EBDA.</br><br>
<br>
These controllers are incompatible with all versions of the Linux Loader
(LILO), unless the loader is patched to avoid the memory conflict in the
EBDA area. Since this BIOS writearound is unsociable on other controllers,
it will not be made a part of the standard LILO
distribution.</br><br>
<br>
I have been able to contact Longshine about the problem, but so far I
cannot find any contact address for the other.
[reply]
[top]
[»]
Re: Incompatible S-ATA controllers
by John Coffman - Sep 29th 2004 06:42:53
My correspondent who initially uncovered the problem has informed me that
Davicontrol has a BIOS update that corrects the problem.
[reply]
[top]
[»]
Dual Boot LILO issue seemingly resistant to normal fixes
by Exitium - Sep 23rd 2004 14:17:40
I will be endlessly thankful if someone can provide me with a solution. My
system has two physical harddrives, one with XP Home and the other with
Debian; I need to clear the Debian distro from the second harddrive, but I
can't seem to shake LILO.
So far I've used fixmbr and fixboot in the Windows recovery console, and
used the LILO -u and LILO -U commands in Debian...all to no avail. LILO is
not impressed when I erase Debian, and I am then incapable of booting to XP
without first reinstalling Debian (and thus fixing LILO).
I don't have much experience with the mbr (or with Linux, for that
matter), and am perhaps missing something obvious...but I've been wrestling
with it fruitlessly for a couple of days now. Any help is greatly
appreciated.
[reply]
[top]
[»]
Re: Dual Boot LILO issue seemingly resistant to normal fixes
by John Coffman - Sep 23rd 2004 20:03:16
Assuming that XP is installed on 'hda1' and Debian on 'hdb2', you need to
make XP the only thing booted; viz.,
> lilo -A /dev/hda 1 # activate the XP partition (note space before
"1")
> lilo -M /dev/hda # install new Master Boot Record
The master boot loader boots the first active partition on the drive,
which will now be XP. The MBR overwrites the LILO boot record
("boot=/dev/hda") written when you run the boot installer (/sbin/lilo).
[reply]
[top]
[»]
Update to correct USB BIOS bug
by John Coffman - Aug 12th 2004 14:07:26
A BIOS bug that results in the unexpected clobber of the BX register has
been seen by a user in Europe. The symptom is the death of the first-stage
loader at just "L", with no diagnostic code printed. This occurs
when booting from a USB hard disk on a VIA Tech controller. Workarounds
include installing with the "-Z1" switch, or patching the LILO
source code with the latest update dated 2004/8/12. Please note that this
update is the second one issued for the current (22.5.9) release of LILO.
See the 'update' directory at the LILO developer's site to retrieve this
patch.
[reply]
[top]
[»]
Single HD, single boot...and LILO 01 01 error?
by Anne Kaelber - May 26th 2004 14:33:52
Hi,
It seems as though everyone encountering the
problem I have has a complex set-up---and
therefore the answers I'm finding are tailored to
that complex set-up. After reading documentation
and googling for answers, I'm stumped. My system
is a single-boot system using Linux. Here's what I
know (the short version):
IDE1, master: 100GB Western Digital Hard Drive
IDE2, master: CD-RW
I did a clean sweep on this drive after numerous
other attempts to figure this problem out. My
partitions now appear as follows (only listing
partial table:
hda1 bootable 128MB
hda5 /boot 128MB
hda7 / 11400MB
After a full Slackware 9.0 install and
configuration, I reboot and get:
L 01 01 01 01 01
until it says DISK BOOT FAILURE.....
I've put lba32 in my lilo.conf (and run
/sbin/lilo), but that didn't help. I'm at a loss
here. I don't even know what other information to
provide to help someone help me! What do I do now?
Frustrated on Day Three of this mess,
Anne.
[reply]
[top]
[»]
Re: Single HD, single boot...and LILO 01 01 error?
by Anne Kaelber - May 26th 2004 15:19:47
> After a full Slackware 9.0 install and
> configuration, I reboot and get:
>
> L 01 01 01 01 01
>
> until it says DISK BOOT FAILURE.....
>
> I've put lba32 in my lilo.conf (and run
> /sbin/lilo), but that didn't help.
Update:
Well, it never fails. Post a request for help and
I figure it out right after that!
I'm not sure if the fault was mine (improperly
reading instructions?) or if there's a bug/typo in
the LILO installer on the Slackware 9.0 distro. I
noticed that the "boot=" line said:
boot=/dev/hda7
When I took the 7 off, it booted fine. I could
have sworn I'd already tried that, but this time
it worked. Now to get the _second_ drive back in
and see if I'm back to my _original_ problem (not
being able to boot from hda at all).
Back to work,
Anne
[reply]
[top]
[»]
build error in file diagnose/test4.c bios.h missing
by Marc Kessels - Apr 14th 2004 14:20:57
Hi,
I am trying to compile lilo 22.5.8 on my mandrake 9.2 system. I have
installed all dependencies (I think..) but when I run make, I get the error
that bios.h is missing.
The reason for still using 22.5.8 is that there is no patch yet available
for using evms with the latest version.
Thanks for any help,
Marc
[reply]
[top]
[»]
Re: build error in file diagnose/test4.c bios.h missing
by John Coffman - Apr 14th 2004 16:15:39
Your installation of "bcc" is not complete or is out-of-date. "bios.h" is
part of the "bcc" distribution. Check the link for "bin86" in the LILO
dependencies here on freshmeat.net for a link to the bcc site.
[reply]
[top]
[»]
Re: build error in file diagnose/test4.c bios.h missing
by Marc Kessels - Apr 15th 2004 05:23:01
Thanks for pointing me in the right direction! problem was that urpmi
dev86 results in installation of bin86 without header files...
urpmi dev86-devel made bios.h available.
[reply]
[top]
[»]
Help me - help me, please ;_;!
by Oli-chan - Mar 17th 2004 19:40:41
I'm really good when it comes to Windows and DOS and the Microsoft
things. I decided to try some Linux. I have two hard drives. I use one for
Windows and the other was an extended partition.
I downloaded the SuSE ISOs and installed it on my second harddrive
(formatted the extended partition). It gave me the boot manager LILO. I was
gonna read some more about linux (file system, installing source, etc...)
but I didn't have time and i needed some HD space, so I deleted the Linux
partition.
I Didn't delete the LILO boot manager before doing so and now it says
SuSE is still there after I deleted the partition, and SuSE is the
master-boot. Now if I don't select Windows at bootup it will automatically
load SuSE (which isn't there).
So can someone please tell me how to get rid of it now that I have
uninstalled SuSE. I don't want to reinstall SuSE again (and that probably
won't solve the problem anyway).
Thanks, Oli ^_^
[reply]
[top]
[»]
Re: Help me - help me, please ;_;!
by John Coffman - Mar 17th 2004 21:15:20
Most Micro$oft products will restore the MBR with:
C:> fdisk /mbr
[reply]
[top]
[»]
Re: Help me - help me, please ;_;!
by Oli-chan - Mar 18th 2004 18:41:01
> Most Micro$oft products will restore the
> MBR with:
>
>
> C:> fdisk /mbr
>
>
thx alot ^_^! when i did some research, I read that in some occasions your
computer might not boot up after this command, so i'm backin up my
important files. and, off topic, does anybody reccomend any good books on
linux? (general info, file system, distros, compatibility); specifically on
SuSE? (don't have to answer)
[reply]
[top]
[»]
lilo dual boot problem
by Tim - Feb 28th 2004 11:34:09
I have a 15G hardisk on hda and a 4G hardisk on hdb.
The hda partitioning is :
hda1 Boot Primary Win95 FAT32 7.682.42
hda5 Logical Win95 Fat32 7674.19
Pri/Log Free Space 8.23
hdb partitioning is:
hdb5 Boot Logical Linux ext2 7.71
hdb6 Logical Win95 Fat32 1068.32
hdb7 Logical Linux swap 317.40
hdb8 Logical Linux ext2 2903.04
Some kind soul on the debian web site had the
excellent idea to put lilo on drive b, and change the
boot order in bios so that drive b becomes a in essence.
This basically worked after i put in the bios=0x80
command for /dev/hdb command.
The problem is that lilo won't boot the windows drive. It
just hangs. The other issue is that ez-bios is on
drive a. Is the problem that lilo is trying to boot windows
without starting ez-bios ? Is there a way to get ez-bios
going and then start windows. Below is my lilo.conf:
prompt
lba32
disk = /dev/hdb
bios = 0x80
disk = /dev/hda
bios = 0x81
boot=/dev/hdb
root=/dev/hdb8
install=/boot/boot.b
map=/boot/map
vga=normal
delay=20
image=/boot/vmlinuz-2.4.18-bf2.4
Label=Linux
other=/dev/hda1
Label=Win98
Thanks,
Tim
[reply]
[top]
[»]
Re: lilo dual boot problem
by John Coffman - Feb 28th 2004 15:59:35
First you have a Windows problem: Windows (all versions) boots ONLY from
drive 0x80. Using the current LILO:
other=/dev/hda1
label=Win98
master-boot
will take care of that problem. And yes, ez-bios will have to be
installed on /dev/hdb too, but set to control /dev/hda.
Personally, I'd suggest going back to booting from /dev/hda by installing
the LILO boot record on /dev/hda (boot=/dev/hda). Then Windows is on the
0x80 (C:) drive already.
[reply]
[top]
[»]
Re: lilo dual boot problem
by Tim - Feb 28th 2004 18:24:33
> First you have a Windows problem:
> Windows (all versions) boots ONLY from
> drive 0x80. Using the current LILO:
>
>
> other=/dev/hda1
>
> label=Win98
>
> master-boot
>
> will take care of that problem. And
> yes, ez-bios will have to be installed
> on /dev/hdb too, but set to control
> /dev/hda.
>
> Personally, I'd suggest going back to
> booting from /dev/hda by installing the
> LILO boot record on /dev/hda
> (boot=/dev/hda). Then Windows is on the
> 0x80 (C:) drive already.
>
Thanks John,
So I don't have to worry about ez-bios being overwritten by lilo when I
install lilo to hda ?
Tim
[reply]
[top]
[»]
Re: lilo dual boot problem
by John Coffman - Feb 29th 2004 07:25:41
The kernel is aware of ez-bios. LILO (/sbin/lilo) boot installer does all
I/O through the kernel; hence, there is no conflict with ez-bios. (EZ-BIOS
moves the actual boot sector from sector 0 to sector 1; the kernel will use
the correct boot sector.)
[reply]
[top]
[»]
Boot from an external RAID array
by Romain - Dec 11th 2003 05:23:12
Hi,
We would like to use LILO to start up a Linux server, but we didn't find
out how to do this. Here are some elements to understand how we are
disappointed!
- The system boot perfectly (default install settings) from the internal
SCSI hard drive (sdc)
- When Linux is installed on the RAID array, an "Operating System not
found" error message is displayed.
- The only solution is to boot from a floppy, which contains the kernel
and the initrd.
- Linux rescue perfectly detects the different partitions on the external
RAID (sda).
- LILO do not display any warning during MBR writing.
- We tried the "disk", "bios", "linear",
"lba32" options without succes.
Here is our /etc/lilo.conf:
prompt
timeout=50
default=linux
boot=/dev/sda
map=/boot/map
install=/boot/boot.b
linear
image=/boot/vmlinuz-2.4.21-4.ELsmp
label=linux
root=/dev/sda3
initrd=/boot/webs-initrd.img
read-only
Here is an extract of dmesg:
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
SCSI device sda: 71811072 512-byte hdwr sectors (36767 MB)
Partition check:
sda: sda1 sda2 sda3 sda4
Here is the layout of the different partitions:
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sda3 13013756 2067116 10285568 17% /
/dev/sda1 147766 16155 123982 12% /boot
We are using LILO version 22.5.8.
Anybody got an idea?
Thanks a lot,
Romain.
[reply]
[top]
[»]
Re: Boot from an external RAID array
by John Coffman - Dec 11th 2003 11:04:00
Have you run the LILO diagnostic disk #1? If the external RAID array has
no BIOS device code (as detected by the diagnostic), you will not be able
to boot from the drive.
Kernel drivers for hardware RAID arrays are in a state of flux. LILO
22.5.8 issues a cautionary message about installations when the "ataraid"
driver is in use. If the external RAID array is SCSI, a lot will depend on
the driver in use.
My homepage at http://lilo.go.dyndns.org/ contains questions to be
answered if you would e-mail me directly.
[reply]
[top]
[»]
LILO can't acces /dev/fd0
by Joel Kaasinen - Nov 16th 2003 03:39:10
Im working on a linux-on-a-disk-thingie, and ive got evrything ready,
boot.b, the kernel and etc/lilo.conf (BTW boot=/dev/fd0 and root=/dev/fd0)
in their places, but when i try to install lilo unto the disk with
"lilo -v -r /mnt/floppy"
lilo says: "cant acces dev/fd0: permission
defined"
As im doing this as root, the file permissions of /dev/fd0 shouldn't
be the cause
It would be nice, if someone could help...
--
cheers,
J
[reply]
[top]
[»]
Re: LILO can't acces /dev/fd0
by John Coffman - Nov 16th 2003 11:53:09
This is a well known gotcha when using "-r". It is expected that you have
a full Linux installation at the point you are 'chroot'ing to.
Specifically, you need a /dev directory, with all device permissions set
correctly. Sounds like you have omitted this step in creating your floppy
image. You might check the method used by the LILO utility "mkrescue", a
shell script, and adopt its methods of avoiding the need for a 'chroot'.
[reply]
[top]
[»]
Re: LILO can't acces /dev/fd0
by Joel Kaasinen - Nov 16th 2003 22:56:25
I have a dev directory with fd0, console and null in it. I got lilo to
eork by first mounting fd0 with:
mount -o dev /dev/fd0
I heard this from someone on the net, and I would like to know exactly
why did it work. The documentation for the dev option is quite
cryptic:
dev Interpret character or block special devices
on the file system.
thx in advance
J
--
cheers,
J
[reply]
[top]
[»]
Re: LILO can't acces /dev/fd0
by Kari Hurtta - Nov 29th 2003 12:01:35
> I have a dev directory with fd0, console
> and null in it. I got lilo to eork by
> first mounting fd0 with:
> mount -o dev /dev/fd0
> I heard this from someone on the net,
> and I would like to know exactly why did
> it work. The documentation for the dev
> option is quite cryptic:
>
> dev Interpret character or block
> special devices
> on the file system.
'nodev' tradiotionally prevents doing 'character'
or 'block' special files to filesystem. So 'dev'
is just opposite.
crw-rw-rw- 1 root root 5, 0 Nov 29 21:06 /dev/tty
That 'c' indicates that this is character special file.
brw-rw---- 1 root disk 3, 0 Apr 7 2003 /dev/hda
That 'b' indicates that this is block special file.
[reply]
[top]
[»]
LILO wants to assign a new Volume ID to this disk drive
by flobobo - Nov 13th 2003 07:22:12
hello, i am trying to use LILO version 22.5.8, but i got this when i type
"lilo" :
-----------------------------------------------------------------------
Reference: disk "/dev/ide/host0/bus0/target0/lun0/disc" (3,0)
0300
LILO wants to assign a new Volume ID to this disk drive. However,
changing
the Volume ID of a Windows NT, 2000, or XP boot disk is a fatal Windows
error.
This caution does not apply to Windows 95 or 98, or to NT data disks.
Is the above disk an NT boot disk? [Y/n]
------------------------------------------------------------------------
I have heard about an inaccessible option, but i have only one hard drive
... so ... what can i do ?
is it because my ext3 partition is out of the 1024 cylinders ? (i have a
XP partition before of 25Go)
thx, flo
[reply]
[top]
[»]
Re: LILO wants to assign a new Volume ID to this disk drive
by John Coffman - Nov 13th 2003 09:46:39
Single disk NT systems apparently don't receive a VolumeID from WinNT
(2000, XP). Before changing the volume ID, create a backup of the boot
sector (-S option: "lilo -S /boot"), ONCE only. Or create the backup file
manually:
"dd if=/dev/hda of=/boot/boot.0300 bs=512 count=1"
Then if WinNT hiccups on booting, you can restore the original boot sector
(/dev/hda); viz.,
"dd if=/boot/boot.0300 of=/dev/hda bs=512 count=1"
.
You will need a Volume ID written to the disk, to proceed with the 22.5.8
installation.
NT always uses Volume IDs on multi-disc systems, so LILO is usually not
the program that writes them on mixed NT/Linux systems.
[reply]
[top]
[»]
Re: LILO wants to assign a new Volume ID to this disk drive
by flobobo - Nov 19th 2003 09:12:04
thx a lot. i will try this.
bye, flo
[reply]
[top]
[»]
Hardware RAID workaround with LILO 22.5.8
by John Coffman - Oct 16th 2003 19:14:45
These comments apply only to HARDWARE raid controllers (Promise, HPT370,
etc.).
Due to the driver implementation a workaround for proper installation is
required when such a controller is in use. The driver for the hardware
controller registers a two drive array as 3 devices: two underlying
devices, and one RAID device. The underlying devices MUST not be touched
by LILO, and 22.5.8 is needed to recognize the following workaround:
disk=/dev/hdi inaccessible # ataraid disk 1
disk=/dev/hdk inaccessible
# ataraid disk 2
A warning is issued to remind the user that this workaroud is required.
This comment applies to both striped and mirrored arrays (Raid 0 or 1).
This workaround will be needed until the "ataraid" drivers provide proper
information about the underlying drives which comprise the array, as is the
case with the Linux software raid implementation (e.g., /dev/md0).
Software raid installations are on a per partition basis, and MUST NOT
attempt to use this workaround.
[reply]
[top]
[»]
Boot CD with lilo in MBR ?
by Uwe - Oct 8th 2003 16:08:10
Hi,
lilo is installed in MBR of hdb.
With this lines in lilo.conf I can boot from a
boot disk without changes bios settings
(1. boot dev.=D, hdb):
other = /dev/fd0
label = Boot_Disk
unsafe
Is it possible to boot from (any bootable)
cdrom with lilo installed on hdb (or hda) ?
I have this section in my lilo.conf:
other = /dev/hdc
label = Boot_CD
master-boot
unsafe
If Boot_CD is chosen in lilo-menu,
I can see Error 0x01 (illegal command...).
A bootable DVD is inserted.
I tried also different "disk=/dev/hdcI bios=..."
lines without success.
hdc is a DVD-ROM.If I change bios
settings the system will boot from (a bootable) DVD/CD.
I'm using LILO version 22.5.8-beta4
Thanks
Uwe
[reply]
[top]
[»]
Re: Boot CD with lilo in MBR ?
by John Coffman - Oct 8th 2003 16:34:47
DVDs & CDs do not have BIOS device codes, so LILO is unable to read them.
Thus they cannot be used for booting in the traditional way. However, many
newer BIOSs support the El Torito Bootable CD spec, which allows a CD to
emulate a floppy drive. With this emulation in place, it becomes possible
to boot from a CD. The method is not really straightforward, and it is
recommended that you create a bootable CD with the "mkrescue --iso"
command. This shell script comes with the lilo source distro, and is
installed with "make install". After creating an ISO image with mkrescue,
then "cdrecord" may be used to burn a bootable CD. If you have ever booted
a Linux distro install from a CD, then you will be able to boot from the
CD. I suggest using a re-writable CD, as you may make several attempts to
get the CD created properly.
[reply]
[top]
[»]
Lilo for USB bootable device
by mario - Sep 21st 2003 02:17:58
Hello,
I'm new to lilo and would like to use it as a bootloader for a USB
memory stick device (specifically a 256M Memorybird). The device is
/dev/sda1. I've created a single partition on it using cfdisk, tagged it
as bootable, and transferred some Linux necessary files. I've also copied
the kernel and boot.b to the USB /boot directory. I've made a lilo.conf in
/etc of the USB device that is like this:
boot=/dev/sda1
map=/boot/map
install=/boot/boot.b
vga=normal
default=linux
lba32
prompt
nowarn
timeout=100
message=/boot/message
menu-scheme=wb:bw:wb:bw
ignore-table
image=/boot/usbLinux
label=usblinux
root=/dev/sda1
read-only
Since this USB device will be used as boot device on another machine I
used the lilo documentation page 37 to install lilo somewhere else by doing
this (USB device is mounted as /mnt/usb on development system)
echo image=/mnt/usb/boot/usbLinux label=usbLinux | /sbin/lilo -C - -b
/dev/sda1 -i /mnt/usb//boot/boot.b -c -m /mnt/usb/boot/map
The target computer (VIA mainboard) has the USB-ZIP boot option activated.
When lilo starts it shows the following:
L 01 01 01 01 01 [....]
Can you share a light on what could happen in there ?
Cheers.
[reply]
[top]
[»]
Re: Lilo for USB bootable device
by John Coffman - Sep 21st 2003 08:43:52
Most likely the BIOS device code of the USB (boot) device is different on
the second system than on the first, where the boot loader was installed.
You may use a "disk= bios=" section in the config file to specify the
environment of the target system. Or, better, use the 22.5++ release of
LILO, which is not sensitive to BIOS device code changes from one system to
the next.
[reply]
[top]
[»]
LILO 22.5.7 fatal internal error
by John Coffman - Aug 16th 2003 20:56:23
Users running "devfs" with more than one SCSI disk are likely to
run into a fatal "LILO internal error". Unfortunately, the
developer does not have access to a hardware configuration on which to test
a proposed patch, so there will be a time delay before an update to correct
the problem will be available.
[reply]
[top]
[»]
Re: LILO 22.5.7 fatal internal error
by John Coffman - Aug 17th 2003 06:34:04
A patch to the 22.5.7 source (called 22.5.7.1) has been tested and uploaded
to the developer's website. The problem only affects systems running
-devfs- with multiple SCSI disks. The binary distribution will not be
updated until later this week, when a full release of LILO will be issued
to address this issue.
[reply]
[top]
[»]
URLs updated
by John Coffman - Aug 2nd 2003 08:43:16
Because of changes at dyndns.org, the project URLs for the developer's
site have all been updated. "brun.dyndns.org" is replaced by
"lilo.go.dyndns.org", a change dictated by the Web redirection
service.
[reply]
[top]
[»]
Thanks
by Lee Dunbar - Jul 23rd 2003 05:10:20
Everything has always worked fine for me - no complaints here! Just my
kudos!
--
Still a newbie, pardon the spew.
[reply]
[top]
[»]
Installing LILO on the hard disk from a bootable CD-ROM, problems going 22.5.1 => 22.5.6
by Jerome St-Louis - Jul 4th 2003 09:41:42
Hi,
I just tried upgrading my self-built distro's lilo to 22.5.6.
I ran into a problem which was not as bad in 22.5.1 (Now I get the L 40
40 40 syndrome, before I just had warnings)
The way my distro installs lilo is as follows:
1. I boot off a CD with ISOLINUX, with root=/dev/ram0
2. I mount the target partition in /setup, and install everything
there (including /setup/sbin/lilo)
3. I'm using devfsd, so I'm doing a cp -dpRf /dev/* in
/setup/dev/
4. Then I do chroot /setup lilo
This used to give me warnings about not being able to detect the video
adapter.
But now I also get a warning that the map file is not on the same disc
as something else
(which I don't fully understand since there's only one hard disc in
this machine) and also
says my BIOS might not recognize drive 0x307 or something like
that...
Now in order to get 22.5.6 to install properly, the only way I've found
is, after my setup has completed,
I reboot again with my setup CD, but pass root=/dev/discs/disc0/part7
as a boot option to the kernel.
If I do that, I won't get any of the warning and lilo will work fine.
That however prevents me from installing
lilo in one go as I could do with 22.5.1. I'm going back to 22.5.1, but
I'd like to know why this changed in 22.5.6,
and what is the proper way to achieve what I'm trying to do... Thanks
for any help in advance.
[reply]
[top]
[»]
Re: Installing LILO on the hard disk from a bootable CD-ROM, problems going 22.5.1 => 22.5.6
by John Coffman - Jul 4th 2003 16:13:02
Changes are detailed in the 'changelog', which would probably be easier for
me to interpret than you. However, general descriptions of what you
experience are of no help in tracking a problem. I need to see your
'lilo.conf' file, and the 'lilo -v5 >log.txt' installation log at the very
least. *** please contact the author directly ***
[reply]
[top]
[»]
Re: Installing LILO on the hard disk from a bootable CD-ROM, problems going 22.5.1 => 22.5.6
by John Coffman - Jul 5th 2003 21:10:33
Jerome kindly provided several install logs of the problem, and a patch has
been issued to resolve the issue. Only installations where a "chroot" has
been performed, either by the user, or using the "lilo -r" switch, will be
affected. The patch is not considered a high priority. It is available on
the developer's site for those interested. (22.5.6 -> 22.5.6.1)
[reply]
[top]
[»]
Please help me!
by Kaugex - Jun 11th 2003 12:10:05
This is not project-oriented but I have a big dilemma which I want to be
cleared...
So I want to reinstall my Linux. For that, I want to delete two ext3
partitions. On one of these there is the \boot, in which (I suppose) there
is LILO. How comes the hard part (for me at least :)):
If I delete the partition on which there is LILO, the boot loader will
survive? This is not clear for me: LILO is in \boot or in the MBR? Does it
need \boot to function?
Because I don't want to lose the data on my windows partition and I don't
want to remedy the situation with fdisk /mbr...
And if we are here allready: if I ovwerwrite the LILO in the new
installation, everything will be okay?
Thanking you for your patience,
Matthew Kosa
[reply]
[top]
[»]
Re: Please help me!
by John Coffman - Jun 11th 2003 12:50:49
1. Create a rescue disk, just in case. (mkrescue).
2. For the boot loader to survive, you need:
a. The boot sector
b. The sector map file (/boot/map)
c. Kernel & initrd files (usually in /boot)
If any of these files are moved, you will need the rescue disk.
3. /sbin/lilo is immaterial; it is not the boot loader. It is the boot
installer (it creates the /boot/map file of sector addresses.)
The kernels, map file &c. may live in a Windows partition just as well as
in a Linux partition. Just watch out for file naming conventions of other
filesystems.
[reply]
[top]
[»]
(another) [L 99 99 99 ... DISK BOOT FAILURE ... ] error
by Rheza - Apr 9th 2003 15:00:07
I'm having problems having LILO work. Basically, what I'm trying to do is
load Slackware 9.0 using Windows NT's boot.ini
I've configured the boot.ini ok, and when I select Slackware from the
choices, it'll go into LILO, but the L 99 99 99 ... DISK BOOT FAILURE
INSERT.... shows up.
I have searched about this in google, tried out the solutions, including
the one here, and having exhausted my options decided to also ask about it
here.
My Linux installation resides at /dev/hda6 while my boot.ini file resides
at /dev/hdc1
Anyway, here are the things you might want:
fdisk -l dev/hda
Device Boot Start End Blocks Id System
/dev/hda1 1 701 5630751 7 HPFS/NTFS
/dev/hda2 702 4997 34507620 f Win95 Ext'd (LBA)
/dev/hda5 702 1402 5630751 7 HPFS/NTFS
/dev/hda6 * 1403 1914 4112608+ 83 LInux
/dev/hda7 1915 2045 1052226 82 Linux Swap
/dev/hda8 2423 4997 20683656 b Win95 FAT32
fdisk -l dev/hdc
/dev/hdc1 * 1 702 5638783+ b Win95 FAT32
/dev/hdc2 703 4865 33439297+ f Win95 Ext'd (LBA)
/dev/hdc5 703 1212 4096543+ b Win95 FAT32
... (I don't think the rest are necessary)
lilo
Added SlackwareLinux*
/etc/lilo.conf
# LILO configuration file
# generated by 'liloconfig'
#
# Start LILO global section
boot = /dev/hda6
#compact # faster, but won't work on all systems.
#prompt
#timeout = 5
# Normal VGA console
vga = normal
lba32
# VESA framebuffer console @ 1024x768x64k
# vga=791
# VESA framebuffer console @ 1024x768x32k
# vga=790
# VESA framebuffer console @ 1024x768x256
# vga=773
# VESA framebuffer console @ 800x600x64k
# vga=788
# VESA framebuffer console @ 800x600x32k
# vga=787
# VESA framebuffer console @ 800x600x256
# vga=771
# VESA framebuffer console @ 640x480x64k
# vga=785
# VESA framebuffer console @ 640x480x32k
# vga=784
# VESA framebuffer console @ 640x480x256
# vga=769
# ramdisk = 0 # paranoia setting
# End LILO global section
# Linux bootable partition config begins
image = /vmlinuz
root = /dev/hda6
label = SlackwareLinux
read-only # Non-UMSDOS filesystems should be mounted read-only for
checking
# Linux bootable partition config ends
/etc/lilo.conf (without most of the comments)
# LILO configuration file
# generated by 'liloconfig'
# Start LILO global section
boot = /dev/hda6
vga = normal
lba32
image = /vmlinuz
root = /dev/hda6
label = SlackwareLinux
read-only
I've followed exactly what is written in
http://www.slackware.com/book/index.php?source=x1776.html
with regards to dual-booting
the dd ~~~, copying the bootsect.lnx, and editing boot.ini
Please help, thanks.
[reply]
[top]
[»]
Can't boot on old machine
by Matt Parlane - Apr 8th 2003 19:25:29
I've just installed Slackware 9.0 on an old AST P75 box, and LILO won't
boot - I just get "L 40 40 40..." which I understand is a seek
error.
The drive has 2099 cylinders, but the bios only lets me go up to 1023 - do
I need to get a BIOS extender of some sort? Windows 95 used to be on this
machine, and as far as I know it didn't need any other software to boot.
thanks,
Matt
[reply]
[top]
[»]
Report Bug
by Brett - Apr 5th 2003 11:38:25
1) Where do I report bugs?
2) Just upgraded to gcc 3.2.2 from 2.xsomthing
Now I get when building lilo 22.5.1 or 22.4.1
lilo.c:181:1: directives may not be used inside a macro argument
lilo.c:180:31: unterminated argument list invoking macro
"printf"
lilo.c: In function `show_images':
lilo.c:182: parse error before "descrs"
I can compile lilo 22.0.2 w/o errors.
Do I need to patch gcc or lilo?
[reply]
[top]
[»]
Re: Report Bug
by John Coffman - Apr 5th 2003 16:15:21
This is not a function of GCC, it is a function of the version of the
C-library in use. Recently, "printf" was changed from a function to a
macro. The syntax in question is legal for a function parameter list, but
not for a macro parameter list. Quick fix: change "printf" to
"fprint(stdout,". N.B. -- this has tripped me before. Sorry.
[reply]
[top]
[»]
Where to report a bug?
by Janfri Bogar - Mar 25th 2003 10:07:56
Hi all.
1st, now I am a happy user of lilo, and I've been using it since the
slackware 2.0 to boot my linux boxes.
I write here to tell My very first problem: I was trying to set up a RAID1
boot device in /dev/md0, and, of course I've read the docs, so i've
configurated lilo.conf as usual. But i've discovered that when I run lilo
-v /mnt (where i've mounted my xfs filesystem which I want boot) xfs gets
stuck and when the kernel tries to mount, panices saying that can't find
/dev/md0. So (i think) lilo breaks the superblock of the xfs partition. I'd
needed to xfs_repair the /dev/md0 partition before running lilo -v, and
lilo has been able to boot the system... ¿is this a lilo bug or a
xfs bug?
Comments and solutions, pleeeze ;)
cheers!
Thanks
Daniel
[reply]
[top]
[»]
Versions, and more info 8D
by Janfri Bogar - Mar 25th 2003 10:25:41
Vanilla kernel 2.4.20 with xfs patch from SGI
lilo v22.2
Debian Woody 3.0r1
I hope it should be enough.
Thanks
[reply]
[top]
[»]
Re: Where to report a bug?
by John Coffman - Mar 25th 2003 10:26:40
Certain filesystems use the first sector of a partition as a superblock
(just where a boot loader would be installed), among these are XFS, swap,
NTFS, MSDOS. There may be others I don't know about. Since version 22.3,
LILO checks for these "vulnerable" filesystems, and will not write a boot
record there, unless "forced" (-F flag). Presumably, you are using a
version of LILO that predates this check. With version 22, the "-x
mbr-only" will avoid the superblock on RAID1 partitions. [Anyone knowing
about other vulnerable filesystems, please alert the author -- as a
community service. I'll be glad to add to the list of known problematic
filesystems.]
[reply]
[top]
[»]
Odd booting problem.
by Francisco Moreno - Mar 7th 2003 13:53:55
I'd compiled and installed lilo using:
# make && make install
then I ran lilo -v , and apparently at this poit all is ok.
but when the machine did reboot, this appeared on the screen:
LILO
L 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40 40
by the way this is my lilo.conf file
prompt
timeout=50
default=RedHat_7.3
boot=/dev/hda
map=/boot/map
vga=normal
install=/boot/boot.b
#message=/boot/message
lba32
image=/boot/vmlinuz-2.4.18-3
label=RedHat_7.3
initrd=/boot/initrd-2.4.18-3.img
read-only
root=/dev/hda5
image=/boot/vmlinuz-2.4.19
label=LFS
read-only
root=/dev/hda8
ps.
I tried to do this using also the lilo's version 22.2, the same thing
happened but after a couple of second appearing 40s, a not ended pattern of
Ls (one per line) did show up.
I fixed the boot using a bootdisk and runing lilo from inside Redhat
(7.3) which has installed the 21.4 lilo's version.
Can anyone tell me what is happening? what I'm doing wrong???
Francisco.
[reply]
[top]
[»]
Re: Odd booting problem.
by John Coffman - Mar 7th 2003 14:16:02
The installation looks good to me. You are using "lba32" to access the
full range of cylinders on your disk. Unfortunately, "lba32" requires that
your BIOS support "LBA" (logical block addressing), which may not be turned
on -- it is a BIOS setup option. The "40" is the hex error code returned
from the disk access call. "40" == seek error, probably generated by LILO
falling back to Cylinder:Head:Sector addressing, and generating a cylinder
number greater than 1023. When you initially installed your system, the
disk was not full, and everything was below cylinder 1023. But now the
disk is filling up, and when the new map file was created, it lies above
cylinder 1023. If you can turn "LBA" on in your hard disk BIOS setup, do
so. If not, you will need to do one of two things: 1). Get a BIOS
extender, such as EZ-BIOS or DiskManager (both are software) which adds the
LBA support (EDD bios extensions), or 2). figure out how to reconfigure
your disk so that /boot can be a small partition below the 1024 cylinder
limit. /boot can be positioned within a Windows or DOS filesystem; just be
sure it is below cylinder 1024.
[reply]
[top]
[»]
Re: Odd booting problem.
by Mircea Baciu - Mar 8th 2003 18:53:18
I have the same problem here. LILO 22.5 displays L 40 40 40
40 40 40 40 ..., but 22.4.1 works perfectly (the same for
older versions).
I have tried the Slackware current package and also have
compiled the sources, with the same result.
My BIOS supports LBA and has it enabled.
Here's my lilo.conf:
boot = /dev/hda
delay = 50
timeout = 50
vga = 0x0123
lba32
prompt
root = /dev/hda3
read-only
image = /boot/vmlinuz-2.4.21-pre5
label = 2421pre5
image = /boot/vmlinuz-2.4.21-pre4
label = 2421pre4
image = /boot/vmlinuz-2.4.21-pre3
label = 2421pre3
image = /boot/vmlinuz-2.4.21-pre2
label = 2421pre2
image = /boot/vmlinuz-2.4.21-pre1
label = 2421pre1
image = /boot/vmlinuz-2.4.20
label = 2420
image = /boot/vmlinuz-2.4.19
label = 2419
other = /dev/hda1
label = dos
table = /dev/hda
Thanks,
Mircea
[reply]
[top]
[»]
Re: Odd booting problem.
by John Coffman - Mar 9th 2003 08:36:22
A sensitivity of 22.5 to the video BIOS unexpectedly trashing the DX
register has been tracked down on one system. The DX register contains the
boot device code, so this is a fatal error. Symptoms were very similar to
the two reports described here. It would be of great help if you contact
the author directly, to see if you have the same or different problem. The
e-mail address is easily generated from the project listing.
[reply]
[top]
[»]
boot on CD line?
by brokenclock - Mar 4th 2003 00:33:15
I tried to use Lilo to have the choice to boot on CD just like on floppy.
But it seems to be impossible. Could you tell me how to do this, or, if it
is impossible, may I suggest you to make it possible?
I think it might be powerful because you could disable the BIOS automatic
boot on CD, which is more secure.
Thanks
[reply]
[top]
[»]
Re: boot on CD line?
by John Coffman - Mar 4th 2003 08:12:28
>
First, no BIOS provides direct support for reading CD's in the fashion of
floppy disks or hard disks. CD's never receive a BIOS device code.
>
>
However, CD's in the El Torito format are bootable, because part of the CD
can be made to emulate a floppy disk. This is a function of your BIOS; and
not all BIOS's support the El Torito Bootable CD format.
>
>
There are two papers at the following site on making a bootable CD:
>
>
http://www.phoenix.com/en/customer+services/white+papers-specs/
>
>
One is a tutorial, and the other is the El Torito spec. The process
starts with making a bootable floppy, then transferring the image to an El
Torito CD.
>
>
'mkisofs' supports creating an ISO image of a CD in the El Torito format.
[reply]
[top]
[»]
"LI" error
by Bill Schoolcraft - Mar 3rd 2003 09:12:50
Hello Family,
The issue is the system will boot FINE with the floppy but not from the
standard startup UNLESS you softboot right after using the floppy..., then
the computers LILO will work.
Booting cold without the floppy gives you the "LI" error unless
you reboot after using the floppy.
This leads me to want to pipe the contents of the floppies data to the
desktops MBR if that is possible, maybe a restore from the floppy to the
mbr?
Thanks
[reply]
[top]
[»]
Re: "LI" error
by stuka - Mar 17th 2003 04:47:03
> Hello Family,
>
> The issue is the system will boot FINE
> with the floppy but not from the
> standard startup UNLESS you softboot
> right after using the floppy..., then
> the computers LILO will work.
>
> Booting cold without the floppy gives
> you the "LI" error unless you
> reboot after using the floppy.
>
> This leads me to want to pipe the
> contents of the floppies data to the
> desktops MBR if that is possible, maybe
> a restore from the floppy to the mbr?
>
> Thanks
Hi Bill
Sometimes I have this problem ....
Boot from diskette an run the installed system as superuser, then run
LILO. If u see the system don't have problems, reboot without diskette.
stuka
[reply]
[top]
[»]
Error 01 01...
by Bernhard Czech - Feb 22nd 2003 14:31:46
Hi,
I am sorry to come up with that topic again, but I have been looking for a
solution for a few hours, and was unable to find a solution.
I am working on a script for cloning Windows clients. Right now we are
useing floppies to boot the kernel in order to start the clone process. In
my new release I want to get rid of the need for floppies after the first
cloning. Therefore my script creates an ext2 partition containing the
kernel. It also extracts everything lilo needs into that partition. The
script also creates the partition for Win2K (ntfs). It is doing all this in
this order:
1. creates ntfs partition
2. creates and initalizes ext2 partition
3. writes an master image to the ntfs partition
4. mounts ext2 on /mnt
5. extracts kernel, /dev/hd*, /boot/* to /mnt
6. creates /mnt/etc/lilo.conf
7. runs "lilo -r /mnt"
8. unmounts /mnt
This is how my lilo.conf looks like:
lba32
boot=/dev/hda
install=/boot/boot-menu.b
map=/boot/map
default=Windows
image=/bzImage
label="Clone Suite"
append="root=/dev/nfs, ip=dhcp"
read-only
optional
other=/dev/hda1
label="Windows"
END OF lilo.conf
It worked before, but after a few tries it stopped working. The clients
are all IDE. Right now I have two sorts of them:
1) Intel Pentium 4 2000 MHz, HDD: 40GB
2) Intel Pentium 2 350 MHz,HDD: 5GB
The ntfs partition's size is about 5GB.
What am I doing wrong?
Thanks in advance
Bernhard Czech
P.S.: Please excuse my bad English.
[reply]
[top]
[»]
Re: Error 01 01...
by John Coffman - Feb 23rd 2003 09:31:21
As far as I can tell, neither the map file nor the second stage loader is
on the /mnt partition. There is no /boot directory there, according to
what you have said. I would suggest corresponding directly with the
project admin, as much more information may be needed to diagnose your
problem.
[reply]
[top]
[»]
A lilo problem?
by SNODX XDONS - Feb 20th 2003 22:44:21
Greetings everybody,
FACTS:
PROCESSOR=Intel Celeron
HDD SIZE=36 GB (3 partitions 10GB+13GB+13GB)
RAM=128 MB
OPERATING SYSTEM CURRENTLY INSTALLED=Windows Millenium Edition
OPERATING SYSTEM TO BE INSTALLED=RedHat Linux Personal edition 8.0
This is the information to start with. Windows ME recognizes the 3
partitions as C drive, D drive and E drive (10GB=C:\, 13GB=D:\=E:\)
I insert the RedHat CD, go through the various options step-by-step (.i.e
The Language, What is the Keyboard, The Mouse, what type of installation-
Custom/Typical/Desktop/.... e.t.c). Finally I encounter the partition
window. There are 3 partitioning tools that can be seen- Automatically
partition, Partition using Disk Druid and fdisk.
I want to keep Windows ME too, so ignore Automatically partition. FDisk
seems to be complicated, so I select Didk Druid.
Sure enough Disk Druid shows up 3 partitions, all of vfat type. I select
the last one /dev/hda3 (E:\ drive in Windows language) allocate around 12.7
GB for an ext3 filesystem and the rest (around 256 MB) for a swap
partition. I force the ext3 to be the PRIMARY partition. After the
partition stuff comes the most important stage- bootloader.
I select Lilo and can see 2 options- Install bootloader in /dev/hda (MBR),
Install in /dev/hda3. Do a quick thinking select /dev/hda.
After this I install RedHat successfully (ofcourse the 3rd partition is
formatted for an ext3 drive), and after the successful installation the CD
drive gets ejected automatically.
But what is this, at boot time I DONT see the lilo window. Instead Windows
ME boots up!
Whew after all that | 
