Now that the general principles are explained the method of creating the loopback device can be explained.
To create the loopback root device will require a number of things.
Most important is access to an installed Linux system. This is because the loop device can only be created under Linux. This will mean that it is not possible to bootstrap a working system from nothing. The requirements of the Linux system that you use is that you can compile a kernel on it.
Once the loopback device is created it will be a large file. I have used an 80 MB files, but while this was sufficient for an X terminal it may not be enough if you want to use it for much else. This file must be copied onto the DOS partition, so either a network or a lot of floppy disks must be used.
The software that you will require includes
LOADLIN version 1.6 or abovemount that supports loopback devicesAll of these should be standard for recent Linux installations.
I created the loopback device using Linux kernel version 2.0.31, other versions should also work, but they must have at least the options listed below.
The kernel options that you will need to enable are the following:
CONFIG_BLK_DEV_RAM).CONFIG_BLK_DEV_INITRD).CONFIG_BLK_DEV_LOOP).CONFIG_FAT_FS).CONFIG_MSDOS_FS).The first two are for the RAM disk device itself and the initial ram disk device. The next one is the loop back filesystem option. The last two are the msdos filesystem support which is required to mount the DOS partitition.
Compiling a kernel without modules is the easiest option, although if you do want modules then it should be possible although I have not tried it. If modules are used then you should make sure that you have the options above compiled in and not as modules themselves.
Depending on the kernel version that you have you may need to apply a kernel patch. It is a very simple one that allows the loopback device to be used as the root filesystem.
For 2.0.x kernels the file /init/main.c needs to have a single line
added to it as shown by the modified version below. The line that says
"loop", 0x0700 is the one that was added.
static void parse_root_dev(char * line)
{
int base = 0;
static struct dev_name_struct {
const char *name;
const int num;
} devices[] = {
{ "nfs", 0x00ff },
{ "loop", 0x0700 },
{ "hda", 0x0300 },
...
{ "sonycd", 0x1800 },
{ NULL, 0 }
};
...
}
For 2.2.x kernels the file /init/main.c needs to have three lines added
to it as shown by the modified version below. The line that says "loop",
0x0700 and the ones either side of it are the ones that were added.
static struct dev_name_struct {
const char *name;
const int num;
} root_dev_names[] __initdata = {
#ifdef CONFIG_ROOT_NFS
{ "nfs", 0x00ff },
#endif
#ifdef CONFIG_BLK_DEV_LOOP
{ "loop", 0x0700 },
#endif
#ifdef CONFIG_BLK_DEV_IDE
{ "hda", 0x0300 },
...
{ "ddv", DDV_MAJOR << 8},
#endif
{ NULL, 0 }
};
Once the kernel is configured it should be compiled to produce a zImage
file (make zImage). This file will be arch/i386/boot/zImage
when compiled.
The initial ramdisk is most easily created as a loopback device from the start.
You will need to do this as root, the commands that you need to execute are
listed below, they are assumed to be run from root's home directory
(/root).
mkdir /root/initrd
dd if=/dev/zero of=initrd.img bs=1k count=1024
mke2fs -i 1024 -b 1024 -m 5 -F -v initrd.img
mount initrd.img /root/initrd -t ext2 -o loop
cd initrd
[create the files]
cd ..
umount /root/initrd
gzip -c -9 initrd.img > initrdgz.img
There are a number of steps to this, but they can be described as follows.
Contents Of The Initial Ramdisk
The files that you will need on the ramdisk are the minimum requirements to be able to execute any commands.
/linuxrc The script that is run to mount the msdos file system
(see below)./lib/* The dynamic linker and the libraries that the programs
need./etc/* The cache used by the dynamic linker (not strictly needed,
but does stop it complaining)./bin/* A shell interpreter (ash because it is smaller
than bash. The mount and losetup programs for
handling the DOS disk and setting up the loopback devices./dev/* The devices that will be used. You need
/dev/zero for ld-linux.so, /dev/hda* to mount the
msdos disk and /dev/loop* for the lopback device./mnt An empty directory to mount the msdos disk on.The initial ramdisk that I used is listed below, the contents come to about 800kB when the overhead of the filesystem are taken into account.
total 18
drwxr-xr-x 2 root root 1024 Jun 2 13:57 bin
drwxr-xr-x 2 root root 1024 Jun 2 13:47 dev
drwxr-xr-x 2 root root 1024 May 20 07:43 etc
drwxr-xr-x 2 root root 1024 May 27 07:57 lib
-rwxr-xr-x 1 root root 964 Jun 3 08:47 linuxrc
drwxr-xr-x 2 root root 12288 May 27 08:08 lost+found
drwxr-xr-x 2 root root 1024 Jun 2 14:16 mnt
./bin:
total 168
-rwxr-xr-x 1 root root 60880 May 27 07:56 ash
-rwxr-xr-x 1 root root 5484 May 27 07:56 losetup
-rwsr-xr-x 1 root root 28216 May 27 07:56 mount
lrwxrwxrwx 1 root root 3 May 27 08:08 sh -> ash
./dev:
total 0
brw-r--r-- 1 root root 3, 0 May 20 07:43 hda
brw-r--r-- 1 root root 3, 1 May 20 07:43 hda1
brw-r--r-- 1 root root 3, 2 Jun 2 13:46 hda2
brw-r--r-- 1 root root 3, 3 Jun 2 13:46 hda3
brw-r--r-- 1 root root 7, 0 May 20 07:43 loop0
brw-r--r-- 1 root root 7, 1 Jun 2 13:47 loop1
crw-r--r-- 1 root root 1, 3 May 20 07:42 null
crw-r--r-- 1 root root 5, 0 May 20 07:43 tty
crw-r--r-- 1 root root 4, 1 May 20 07:43 tty1
crw-r--r-- 1 root root 1, 5 May 20 07:42 zero
./etc:
total 3
-rw-r--r-- 1 root root 2539 May 20 07:43 ld.so.cache
./lib:
total 649
lrwxrwxrwx 1 root root 18 May 27 08:08 ld-linux.so.1 -> ld-linux.so.1.7.14
-rwxr-xr-x 1 root root 21367 May 20 07:44 ld-linux.so.1.7.14
lrwxrwxrwx 1 root root 14 May 27 08:08 libc.so.5 -> libc.so.5.3.12
-rwxr-xr-x 1 root root 583795 May 20 07:44 libc.so.5.3.12
./lost+found:
total 0
./mnt:
total 0
The only complex steps about this are the devices in dev. Use the
mknod program to create them, use the existing devices in /dev
as a template to get the required parameters.
The /linuxrc file
The /linuxrc file on the initial ramdisk is required to do all of the
preparations so that the loopback device can be used for the root partition when
it exits.
The example below tries to mount /dev/hda1 as an msdos partition and
if it succeeds then sets up the files /linux/linuxdsk.img as
/dev/loop0 and /linux/linuxswp.img as /dev/loop1.
#!/bin/sh
echo INITRD: Trying to mount /dev/hda1 as msdos
if /bin/mount -n -t msdos /dev/hda1 /mnt; then
echo INITRD: Mounted OK
/bin/losetup /dev/loop0 /mnt/linux/linuxdsk.img
/bin/losetup /dev/loop1 /mnt/linux/linuxswp.img
exit 0
else
echo INITRD: Mount failed
exit 1
fi
The first device /dev/loop0 will become the root device and the second
one /dev/loop1 will become the swap space.
If you want to be able to write to the DOS partition as a non-root user when you
have finished then you should use mount -n -t msdos /dev/hda1 /mnt -o
uid=0,gid=0,umask=000,quiet instead. This will map all accesses to the DOS
partition to root and set the permissions appropriately.
The root device that you will be using is the file linuxdsk.img. You
will need to create this in the same way that the initial ramdisk was created,
but bigger. You can install any Linux installation that you like onto this
disk.
The easiest way might be to copy an existing Linux installation into it. An alternative is to install a new Linux installation onto it.
Assuming that you have done this, there are some minor changes that you must make.
The /etc/fstab file must reference the root partition and the
swap using the two loopback devices that are setup on the initial ramdisk.
/dev/loop0 / ext2 defaults 1 1
/dev/loop1 swap swap defaults 1 1
This will ensure that when the real root device is used the kernel will not be confused about where the root device is. It will also allow the swap space to be added in the same way a swap partition is normally used. You should remove any other reference to a root disk device or swap partition.
If you want to be able to read the DOS partition after Linux has started then you will need to make a number of extra small changes.
Create a directory called /initrd, this is where the initial ramdisk
will be mounted once the loopback root filesystem is mounted.
Create a symbolic link called /DOS that points to /initrd/mnt
where the real DOS parition will be mounted.
Add a line into the rc file that mounts the disks. This should run the command
mount -f -t msdos /dev/hda1 /initrd/mnt, this will create a 'fake'
mount of the DOS partition so that all programs (like df) will know that
the DOS partition is mounted and where to find it. If you used different
options in the /linuxrc file that obviously you should use them here
also.
There is no need to have a Linux kernel on this root device since that is already loaded earlier. If you are using modules however then you should include them on this device as normal.
The root device that you will be using is the file linuxswap.img. The
swap device is very simple to create. Create an empty file as was done for the
initial ramdisk and then run mkswap linuxswap.img to intialise it.
The size of the swap space will depend on what you plan to do with the installed system, but I would recommend between 8MB and the amount of RAM that you have.
The files that are going to be used need to be moved onto the DOS partition.
The files that are required in the DOS directory called C:\LINUX are
the following:
LINUXDSK.IMG The disk image that will become the root device.LINUXSWP.IMG The swap space.
The boot floppy that is used is just a normal DOS format bootable floppy.
This is created using format a: /s from DOS.
Onto this disk you will need to create an AUTOEXEC.BAT file (as below)
and copy the kernel, compressed initial ramdisk and LOADLIN executable.
AUTOEXEC.BAT The DOS automatically executed batch file.LOADLIN.EXE The LOADLIN program executable.ZIMAGE The Linux kernel.INITRDGZ.IMG The compressed initial ramdisk image.The AUTOEXEC.BAT file should contain just one line as below.
\loadlin \zImage initrd=\initrdgz.img root=/dev/loop0 ro
This specifies the kernel image to use, the initial ramdisk image, the root device after the initial ramdisk has finished and that the root partition is to be mounted read-only.