Kernel Panic: It won't mount the root filesystem (2.6.8 and 2.6.9) and LILO hates me

edman007 · 10-29-2004, 09:43 PM

can anyone help my try to get this to work, please give me any suggestion on what to do, first, i have compiled and successfully run all the 2.6.x kernels the same way but when i got to 2.6.8 and 2.6.9 it stopped working (and yes i do have the stuff for the filesystem, i copied the config from /proc and used it as a base when configuring a new kernel)

first some info, i have 2 hard drives, one IDE at /dev/hda* and one sata at /dev/hde, each has a swap as the first partition and a mountable filesystem on the second partition, i also have an ASUS p4c800-E dlx motherboard, which has a promise SATA controller for my sata drive, i'm currently stuck with the 2.6.7 kernel and am running slackware-current

first lilo hates me, when booted normally, lilo will not work go on any drive (either hda or hde), if try i install it (which works fine), i then boot and just get "L", so what i have to do is boot with a slackware cd, neither of them have the promise sata drivers so i have to copy the System.map, kernel, and copy of lilo.conf to the IDE drive and then unplug the sata drive from the computer (the boot cd does not have the promise drivers so it hangs), i have to use the 9.0 install cd to then run lilo with the new conf file on the IDE drive and put lilo on that drive, (the slackware 10 cd says something like it doesn't know how to use the device 0x0100), so with this setup i have to turn off the comp, unplug the sata drive, boot off a CD, install lilo onto the IDE drive, turn the comp back off, and then boot normally but using the IDE drive for LILO (meaning i can't have just my SATA drive in there, the IDE drive has to take up space on that IDE cable [which i want to use for more cd drives]), and this setup is very painful, and a note: i've tried grub, but its config stuff is impossible

second, the new kernels (2.6.8 and 2.6.9) won't boot, it goes fine recognizing all the drives and other things, then just says (paraphrasing) Kernel panic: VFS: cannot mount root device (some_number) use the right root= thing

and if you want to see the lilo conf files, here they are (a different one for each drive i have tried to put it on), also i have removed a few comments to save space

Code:

# LILO configuration file
# generated by 'liloconfig'
#
# Start LILO global section
boot = /dev/hde
#compact        # faster, but won't work on all systems.
prompt
timeout = 50
#lba32#could be causing the problem
# VESA framebuffer console @ 1024x768x64k
vga = normal
map=/boot/map
# Normal VGA console
# Linux bootable partition config begins
image = /boot/linux-2.6.9
  root = /dev/hde2   
  label = Linux-2.6.9
  read-only
image = /boot/linux-2.6.7
  root = /dev/hde2
  label = Linux-2.6.7
  read-only

# Linux bootable partition config ends

Code:

# LILO configuration file
# generated by 'liloconfig'
#
# Start LILO global section
boot = /dev/hda
#compact        # faster, but won't work on all systems.
prompt
timeout = 50
#lba32#might be a problem
# VESA framebuffer console @ 1024x768x64k
vga = normal
# Normal VGA console
# End LILO global section
# Linux bootable partition config begins
image = /boot/linux-2.6.9
  root = /dev/hde2
  label = Linux-2.6.9
  read-only
image = /boot/linux-2.6.7
  root = /dev/hde2
  label = Linux-2.6.7
  read-only
# Linux bootable partition config ends

Bruce Hill · 10-30-2004, 04:32 AM

Quote:

Originally posted by edman007
can anyone help my try to get this to work, please give me any suggestion on what to do, first, i have compiled and successfully run all the 2.6.x kernels the same way but when i got to 2.6.8 and 2.6.9 it stopped working (and yes i do have the stuff for the filesystem, i copied the config from /proc and used it as a base when configuring a new kernel)

first some info, i have 2 hard drives, one IDE at /dev/hda* and one sata at /dev/hde, each has a swap as the first partition and a mountable filesystem on the second partition, i also have an ASUS p4c800-E dlx motherboard, which has a promise SATA controller for my sata drive, i'm currently stuck with the 2.6.7 kernel and am running slackware-current

So you're currently running a good, working 2.6.7 kernel?

I have an IDE drive (/dev/hda) and a SATA drive (/dev/sda) in this comp
running Slack-10.0 with a 2.6.7 kernel. You have (dev/hde) because your
sata drive is recognized as PATA - I think - will have to search to be sure.
But it should be acting like a SCSI and be /dev/sda.

/me thinks this is a major part of the problem.

Actually for the Intel ICH5R you need CONFIG_BLK_DEV_PIIX=y
and for your Promise SATA controller CONFIG_SCSI_SATA=y and
CONFIG_SCSI_SATA_PROMISE=y plus all the stuff for SCSI

Code:

#
# SCSI device support
#
CONFIG_SCSI=y
CONFIG_SCSI_PROC_FS=y

#
# SCSI support type (disk, tape, CD-ROM)
#
CONFIG_BLK_DEV_SD=y

Quote:

first lilo hates me, when booted normally, lilo will not work go on any drive (either hda or hde), if try i install it (which works fine), i then boot and just get "L", so what i have to do is boot with a slackware cd, neither of them have the promise sata drivers so i have to copy the System.map, kernel, and copy of lilo.conf to the IDE drive and then unplug the sata drive from the computer (the boot cd does not have the promise drivers so it hangs), i have to use the 9.0 install cd to then run lilo with the new conf file on the IDE drive and put lilo on that drive, (the slackware 10 cd says something like it doesn't know how to use the device 0x0100), so with this setup i have to turn off the comp, unplug the sata drive, boot off a CD, install lilo onto the IDE drive, turn the comp back off, and then boot normally but using the IDE drive for LILO (meaning i can't have just my SATA drive in there, the IDE drive has to take up space on that IDE cable [which i want to use for more cd drives]), and this setup is very painful, and a note: i've tried grub, but its config stuff is impossible

Are you having to do all this with your current 2.6.7 kernel?
That should not be - it can be fixed! And what is the rest of the
LILO error code? That L means the first stage boot loader
has been loaded and started, but it can't load the second stage
boot loader. The two-digit error codes which come after the L
indicate the type of problem.

Quote:

second, the new kernels (2.6.8 and 2.6.9) won't boot, it goes fine recognizing all the drives and other things, then just says (paraphrasing) Kernel panic: VFS: cannot mount root device (some_number) use the right root= thing

It is important what (some_number) and thing are - the comp is hung
and not going any further - just right them down with pencil and paper.
And look right up before that and see what drive the kernel is calling
it - whether /dev/hde or /dev/sda or something else.

Quote:

and if you want to see the lilo conf files, here they are (a different one for each drive i have tried to put it on), also i have removed a few comments to save space

I still prefer to have a bootable partition first on the drive,
and swap later, but shouldn't matter if LILO is put in MBR.

Just for kicks how about posting the output of "fdisk -l" so we can
get a look at the partitions of those drives.

edman007 · 10-30-2004, 09:54 AM

Quote:

Originally posted by Chinaman
So you're currently running a good, working 2.6.7 kernel?

yes

Quote:

I have an IDE drive (/dev/hda) and a SATA drive (/dev/sda) in this comp
running Slack-10.0 with a 2.6.7 kernel. You have (dev/hde) because your
sata drive is recognized as PATA - I think - will have to search to be sure.
But it should be acting like a SCSI and be /dev/sda.

that explains a lot, i will try it today

Quote:

And what is the rest of the
LILO error code? That L means the first stage boot loader
has been loaded and started, but it can't load the second stage
boot loader. The two-digit error codes which come after the L
indicate the type of problem.

thats is, just L, i get a black screen with an L in the corner and a flashing cursor

Quote:

It is important what (some_number) and thing are - the comp is hung
and not going any further - just right them down with pencil and paper.
And look right up before that and see what drive the kernel is calling
it - whether /dev/hde or /dev/sda or something else.

well since i don't really want to restart, i'll say that i did post it somewhere on these forums....here is the thread http://www.linuxquestions.org/questi...hreadid=217710 and it said last time (which i'm pretty sure that its the same this time)

Code:

VFS: Cannot Open device "2102" or unknown-block(33,2)

Quote:

Just for kicks how about posting the output of "fdisk -l" so we can
get a look at the partitions of those drives.

here it is

Code:

root@mr-slack:~# fdisk -l

Disk /dev/hda: 20.4 GB, 20485785600 bytes
255 heads, 63 sectors/track, 2490 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hda1               1         243     1951866   82  Linux swap
/dev/hda2             244        2490    18049027+  83  Linux

Disk /dev/hde: 81.9 GB, 81964302336 bytes
16 heads, 63 sectors/track, 158816 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hde1               1        3875     1952968+  83  Linux
/dev/hde2            3876      158816    78090264   83  Linux

hmmmm.....says /dev/hde1 isn't swap, all well, its getting used as a swap, thats what fstab says, should i reformat that part to swap? will it run better

Bruce Hill · 10-30-2004, 09:46 PM

Quote:

Originally posted by edman007
yes

Could you please answer if you have to do all the stuff you described
in your first post to mount your 2.6.7 kernel? Where you said you have
to boot with a Slack CD and unplug the SATA drive and all?

If this is correct, I recommend you backup your data and start over
and install this correctly. LILO does not hate you - it is a computer
program not capable of emotions. It executes the commands you give
it. As the old saying goes, "garbage in - garbage out." So if we give
LILO incorrect data, it won't function correctly. Read this from "man lilo"

Quote:

BOOT ERRORS
The boot process takes place in two stages. The first stage loader is a single sector, and is loaded by the BIOS or by the loader in the MBR. It loads the multi-sector second stage loader, but is very space limited. When the first stage loader gets control, it types the letter "L"; when it is ready to transfer control to the second stage loader it types the letter "I". If any error occurs, like a disk read error, it will put out a hexadecimil error code, and then it will re-try the operation.

So there should be a number after the L. If you install LILO to the
MBR, I think it will work okay. That's all I've ever done, and have never
had any problems with LILO - even with as many as 3 hard drives in one
computer. And you don't need map=/boot/map in LILO, cause if `map'
is omitted, the file /boot/map is used .

I suspect from reading your posts that the SATA drive was never
properly partitioned. There isn't an official Slackware CD with SATA
support, but there are unofficial ones. Now Pat Volkerding has put
a SATA kernel in Slackware/current.

If my assumptions are true, and you'd like to start over and get it
done correctly, you have two choices:
1) if you know how to boot from a floppy, and desire that method, get
it from /pub/slackware/slackware-current/kernels/sata.i
2) if you'd like a CD, which is my preferred method, there is one
available (which I used) here -> ftp://ftp.readmore.biz/
NB: This is the Slackware-10.0 CD1 with a sata.i kernel added.
If you'd like to use it, there are a couple of instructions you'll need
to save yourself some headaches when installing with it. I won't
spend the time to post that now, but if you'd like to d/l this and have
a Slack-10.0 install CD with a sata kernel, let me know.

Quote:

here it is

Quote:

Code:

root@mr-slack:~# fdisk -l

Disk /dev/hda: 20.4 GB, 20485785600 bytes
255 heads, 63 sectors/track, 2490 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hda1               1         243     1951866   82  Linux swap
/dev/hda2             244        2490    18049027+  83  Linux

Disk /dev/hde: 81.9 GB, 81964302336 bytes
16 heads, 63 sectors/track, 158816 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hde1               1        3875     1952968+  83  Linux
/dev/hde2            3876      158816    78090264   83  Linux

hmmmm.....says /dev/hde1 isn't swap, all well, its getting used as a swap, thats what fstab says, should i reformat that part to swap? will it run better

You should not put swap as your first partition. You should put the
swap partition on the outer tracks of your drive, and on the opposite
drive if possible. Your boot partition must be of type 0x83 "Linux native"
so you cannot boot from /dev/hda1 because it's 0x82 "Linux swap".

vectordrake · 11-13-2004, 05:41 PM

One thing that hung me on that boot message "kernel panic...." was not having the /dev flesystem compiled into the kernel (not module) and the option to load on boot checked as well. Its an easy thing to overlook. be sure all the filesystems you mount and use at boot are compiled right into the kernel (so if "/" is reiserfs, be sure its not a module - and be sure /dev is compiled in too). Good luck.