source: postlfs/config/bootdisk.xml@ 7b3d96c9

10.0 10.1 11.0 11.1 11.2 6.0 6.1 6.2 6.2.0 6.2.0-rc1 6.2.0-rc2 6.3 6.3-rc1 6.3-rc2 6.3-rc3 7.10 7.4 7.5 7.6 7.6-blfs 7.6-systemd 7.7 7.8 7.9 8.0 8.1 8.2 8.3 8.4 9.0 9.1 basic bdubbs/svn elogind gnome kde5-13430 kde5-14269 kde5-14686 krejzi/svn lazarus nosym perl-modules qt5new systemd-11177 systemd-13485 trunk upgradedb v5_0 v5_0-pre1 v5_1 v5_1-pre1 xry111/intltool xry111/soup3 xry111/test-20220226
Last change on this file since 7b3d96c9 was 7b3d96c9, checked in by Larry Lawrence <larry@…>, 19 years ago


git-svn-id: svn:// af4574ff-66df-0310-9fd7-8a98e5e911e0

  • Property mode set to 100644
File size: 31.7 KB
1<sect1 id="postlfs-config-bootdisk">
2<?dbhtml filename="bootdisk.html" dir="postlfs"?>
3<title>Creating a Custom Bootdisk</title>
6<title>Decent Rescue Bootdisk Needs</title>
7<para>This section is really about creating a <emphasis>rescue</emphasis>
8diskette. As the name <emphasis>rescue</emphasis> implies, the host
9system has a problem, often lost partition information or corrupted file
10systems, that prevents it from booting and/or operating normally. For
11this reason, you <emphasis>must not</emphasis> depend on resources from
12the host being "rescued". To presume that any given partition or hard
13drive <emphasis>will</emphasis> be available is a risky presumption.</para>
15<para>Heeding the warning, the rescue disk created here has no
16dependency on the host system's resources, other than basic bootability
17and hardware soundness. At a minimum, the most common sorts of failures
18requiring a rescue bootdisk should be addressed by the contents of the
19bootdisk. This would include the common loss of partitioning (master
20boot record is lost or corrupted), file system corruption, and the need
21to allow creation and editing of files that may have been lost or
22corrupted, possibly as an effect of the other two problems.</para>
24<para>Additional utilities should be available to search for text or
25files, copy, move and remove files, and many other normal operations
26that might be expected to be needed when reconstructing.</para>
30<title>This Minimal Decent Rescue Disk</title>
32<para>The intent here is to create a "rescue bootdisk" that will support
33the common operations listed above. These functions are provided by
34including selected executables from <application><ulink
36</ulink></application> and <application><ulink
38A basic editor and rudimentary disk partitioning utility may also be
39optionally included.</para>
41<para>This, however, is not the limit. A minimal disk is described
42here, but you can add anything you can fit on the floppy. Furthermore,
43if one floppy is not enough to meet your needs, you can make a
44multi-diskette rescue set that means, essentially, the sky is the limit.
45This is discussed below. The number of other possible variations are
46just too numerous to mention here.</para>
50<title>Build the Rescue Bootdisk</title>
54<para>You should have known-good floppy diskettes available. Some
55people prefer to use the <command>fdformat</command> command to prepare
56these because it also does a verification. See the man page for more
57details. Another good idea is to always prepare duplicates of the
58rescue diskette. Media does deteriorate.</para>
60<para>These instructions presume a base <acronym>LFS</acronym> install
61using ext2/ext3 file systems.</para>
63<para>You need to have loopback device support enabled in your host's
64kernel to use this procedure.</para>
66<para>You should make a custom kernel that includes only those features
67needed to rescue your system, so it will have the least size. No
68sense in building in support for things like <command>xfree86</command>,
69<command>dri</command>, etc, as most rescues are performed from the
70command prompt. Along the same lines, if you have GCC-2.95.*, it is
71known to produce smaller kernels. So you might want to use that
72compiler for this kernel. If you do so, don't overlook any loadable
73modules (which are not addressed here) you might need - they need to be
74compiled with same compiler used to make the kernel.</para>
76<para>The rescueimage must include support for the file system of your
77choice (we presume ext2/3 here), ramdisk and initial ramdisk (initrd).
78Disable everything that you can in the kernel configuration. You should
79keep support for the proc file system and tempfs file system enabled
80because of their general utility. The proc file system is needed for
81the <command>mount</command> to report properly.</para>
83<para>If you install <emphasis>only</emphasis> the minimal set of
84components shown in this document, you will need a kernel that is 643 or
85fewer blocks in size. If you want the optional programs - a very basic
86editor, like ed, and rudimentary disk partitioning, like sfdisk - the
87kernel will need to be 595 or fewer blocks in size. This should not be
88a major problem unless your needs are fairly esoteric. On the system
89used to develop this version of the procedures, using only ext2 file
90systems and not using networking or CDs for recovery, the kernel image
91is only 481 blocks. And there may be more to gain - it has not been
92closely examined for additional gains.</para>
94<para>This kernel image will be called "rescueimage" hereinafter. You
95can actually name your image anything you like and just use its name
96instead in any commands that include "rescueimage".</para>
98<para>If you can not get your kernel down to the size needed to allow
99all you need on the ramdisk image, don't fret. You can always build a
100two diskette set, one boot and one root diskette. The kernel will prompt
101you to insert the root file system diskette. This will allow room for a
102compressed ramdisk image of 1440 blocks and a kernel of the same
105<para>The kernel size limits given above are likely to vary as
106local system-specific configurations change. Use them only as a
107guideline and not as gospel. The size of the kernel image as shown by
108<command>ls -sk</command> is only an approximation because of some
109"overhead". On the system used to develop this version of these
110procedures, that command shows 488 blocks but the actual number of
111blocks written is only 480 and a fraction, which means that 481 blocks
112are actually used.</para>
116<title>Rescue Disk Build Process</title>
118<para>The basic process will be:</para>
120<listitem><para>make a mount point for a file system</para></listitem>
121<listitem><para>make an empty file to hold the file system</para></listitem>
122<listitem><para>bind the empty file to a loopback device</para></listitem>
123<listitem><para>make a 4MB file system</para></listitem>
124<listitem><para>mount the file system</para></listitem>
125<listitem><para>add components to the file system</para></listitem>
126<listitem><para>make the compressed initrd</para></listitem>
127<listitem><para>join a kernel image and initrd onto a diskette</para></listitem>
130<para>The initial ramdisk will be automatically loaded at boot time if
131setup is done correctly.</para>
133<para><emphasis>Make a mount point and an empty file to hold a file
136<screen><userinput><command>mkdir -p /mnt/loop1</command>
137<command>dd if=/dev/zero of=/tmp/rfloppy bs=1k count=4096</command></userinput></screen>
139<para><emphasis>Command explanations</emphasis></para>
141<para><command>dd</command>: This is a generalized input-to-output copy
142utility that also has many transformation capabilities.</para>
144<para><command>if=/dev/zero</command>: This parameter assigns
145<command>dd</command>'s input file to a device that returns an infinite
146stream of zeroes.</para>
148<para><command>of=/tmp/rfloppy</command>: This parameter directs
149<command>dd</command>'s output to <filename>/tmp/rfloppy</filename>.</para>
151<para><command>bs=1k count=4096</command>: These parameters tell
152<filename>dd</filename> to read and write in "chunks" of 1024 bytes and
153process 4096 "chunks".</para>
155<para><emphasis>Bind the file to a loopback device, make a file system and mount it.</emphasis></para>
157<para>The reason these commands are used is that they work regardless of
158the version of <command>mount</command> (older ones don't have the
159<command>-o loop</command> option) or if <filename>/etc/mtab</filename>
160is symlinked to <filename>/proc</filename> (which causes mount to be
161unable to properly "unbind" a loop device, due to "lost" information).
162An alternate set of commands is provided, after these three commands,
163that you can use if you don't have either of these situations.</para>
165<screen><userinput><command>losetup /dev/loop1 /tmp/rfloppy</command>
166<command>mke2fs -m 0 -N 504 /dev/loop1</command>
167<command>mount -t ext2 /dev/loop1 /mnt/loop1</command></userinput></screen>
169<para><emphasis>Command explanations</emphasis></para>
171<para><command>losetup /dev/loop1 /tmp/rfloppy</command>: This
172command "binds" a loopback device to the empty file.</para>
174<para><command>mke2fs -m 0 -N 504 /dev/loop1</command>: This
175command makes an ext2 file system on the loopback device (which really
176means it is created in the file to which the loopback device is bound)
177and reserves no blocks. The "-N 504" parameter causes only 504 inodes
178to be allocated, leaving more space for other things needed in the file
181<para><command>mount -t ext2 /dev/loop1 /mnt/loop1</command>: This
182mounts the file system just created, just as if it were a real device,
183like a hard drive or diskette. This allows all the normal system I/O
184commands to operate as if a real device were present.</para>
186<para>If your <command>mount</command> supports the
187<command>-o loop</command> option <emphasis>and</emphasis> your
188<filename>/etc/mtab</filename> is a real file, rather than a symlink to
189<filename>/proc</filename>, the three above commands can be replaced
190by these next two commands.</para>
192<screen><userinput><command>mke2fs -F -m 0 -N 504 /tmp/rfloppy</command>
193<command>mount -o loop /tmp/rfloppy /mnt/loop1</command></userinput></screen>
195<para><emphasis>Command explanations</emphasis></para>
197<para><command>mke2fs -F -m 0 -N 504 /tmp/rfloppy</command>: As before,
198a file system is made, with only 504 inodes and no reserved blocks, that
199will be bound to a loopback device. The <command>-F</command> parameter
200just suppresses an irritating question issued when
201<command>mke2fs</command> realizes that you are not accessing a
204<para><command>mount -o loop /tmp/rfloppy /mnt/loop1</command>: This
205command tells <command>mount</command> to bind the named file to a
206loopback device it automatically selects (the first available) and mount
207the device on <filename>/mnt/loop1</filename>.</para>
209<para><emphasis>Add components to the file system</emphasis></para>
211<para><emphasis>A cautionary note:</emphasis> if you are not running in a
212<command>chroot</command> environment, be sure that you do not accidently
213omit the <filename>/mnt/loop1</filename> reference in the commands. If
214you do so, you might replace the equivalent components on your host with
215the components that are installed by these procedures. Even if you are
216in a chroot environment, you may need to be careful if the environment
217is your freshly built <acronym>LFS</acronym> system which you intend to
218use as a host in the future.</para>
220<para>First, to have as much free space as possible, remove the
221<filename>lost+found</filename> directory, which is not needed because
222it is only used by fsck. Since fsck will never be run on this file
223system, it is unneeded.
225<screen><userinput><command>rmdir /mnt/loop1/lost+found/</command></userinput></screen></para>
227<para>Now make a minimal set of directories.</para>
229<screen><userinput><command>mkdir /mnt/loop1/{dev,proc,etc,sbin,bin,lib,mnt,usr,var}</command></userinput></screen>
231<para>Add needed device files to the initrd image. If you use devfs,
232the following command works well, as you only have the devices you use
235<screen><userinput><command>cp -dpR /dev/* /mnt/loop1/dev</command></userinput></screen>
237<para>If you used <command>MAKEDEV</command> to create your devices on
238your host, you'll want to use something similar to this longer command,
239to minimize wasting space with unneeded inodes.</para>
241<para><emphasis>You must modify this to suit your kernel configuration and
242other needs.</emphasis> For example, you may need scsi devices and may not need
243frame buffer devices or the pseudo-terminal directory. Also, the number
244of hard drives and partitions that you include should be the minimal
245that you need. Extensive analysis has not been done on the list below,
246so there are more inodes and space to be gained by "fine tuning" this
249<screen><userinput><command>mkdir /mnt/loop1/dev/pts
250cp -a \
251&nbsp;&nbsp;&nbsp;&nbsp;/dev/null /dev/console \
252&nbsp;&nbsp;&nbsp;&nbsp;/dev/fb[0-7] /dev/fd /dev/fd0 /dev/fd0h1440 /dev/full \
253&nbsp;&nbsp;&nbsp;&nbsp;/dev/hda* /dev/hdb* /dev/hdc* /dev/hdd* /dev/initctl /dev/kmem \
254&nbsp;&nbsp;&nbsp;&nbsp;/dev/loop[0-3] /dev/lp0 /dev/mem /dev/port \
255&nbsp;&nbsp;&nbsp;&nbsp;/dev/psaux /dev/ram \
256&nbsp;&nbsp;&nbsp;&nbsp;/dev/ram0 /dev/ram1 /dev/ram2 /dev/ram3 /dev/random /dev/rtc \
257&nbsp;&nbsp;&nbsp;&nbsp;/dev/shm /dev/stderr /dev/stdin /dev/stdout /dev/tty \
258&nbsp;&nbsp;&nbsp;&nbsp;/dev/tty[0-9] /dev/ttyS0 /dev/ttyS1 /dev/urandom /dev/zero \
261<para><emphasis>What is needed in the <filename>/etc</filename>
264<para>If you choose, you can copy all or selected parts of your
265<filename>/etc/passwd</filename> and <filename>/etc/group</filename>
266files. But even if each is less than 1024 bytes, you will lose two
267inodes and two blocks of space on the initial ramdisk. This only really
268matters because of trying to squeeze everything onto a 1.44MB
269diskette. Every little bit helps. The strategy taken here is to create
270these two files as part of the rescue boot and initialization process.
271The commands that make the two files will be imbedded inside the
272<filename>rcS</filename> script that <filename>linuxrc</filename>
273(really <application><ulink
275</ulink></application>) invokes after the initrd is
276loaded. This way no more inodes or blocks are used on the
277diskette to carry these files.</para>
279<para>Some might like to copy their <filename>/etc/rc*</filename>
280directory into the ramdisk image, but this may have no value, other
281than archival use, in a worst-case recovery scenario. If you want
282automatic initialization of the system after repair, they may have some
283value. But few people need or want this to happen. If the file system
284on the hard drives are corrupted, what good will mount scripts do? Some
285scripts may be useful, like access to a network to copy over backup data
286when the hard drive's file systems are usable again. The point is that
287you should copy only the parts that you can use because space is at a
288premium on the diskette.</para>
290<para>Here, only the <filename>fstab</filename> will be included. This
291is handy because it eases mounting of partitions that may be useful and
292also can be examined and used as a guide as to what is available and
293what may need reconstruction. Because it may be larger than needed, you
294should edit it to remove any useless entries and minimize commentary.
295No other editing is needed because the boot scripts are not included and
296no automatic mounting will be done using the <filename>fstab</filename>.
297If you decide to include some bootscripts that might try to mount
298things, change the <filename>fstab</filename>'s entries to
299<command>noauto</command> in the options field so they don't cause an
300attempt to mount a potentially corrupt partition. Copy it to
301<filename>/tmp</filename>, edit it as desired and then:</para>
303<screen><userinput><command>cp -a /tmp/fstab /mnt/loop1/etc</command></userinput></screen>
305<para>Now the initialization script will be added. As mentioned above,
306<command>linuxrc</command> is symlinked to <application><ulink
309After the kernel and initial ramdisk have been loaded, the kernel gives
310control to <command>linuxrc</command> (<application><ulink
312</ulink></application>). It wants to run an <filename>/etc/init.d/rcS</filename>
313script to do any initial setup.</para>
315<para>If you use devfsd, you will need to set up the
316<filename>rcS</filename> script to handle the devfsd startup. Put the
317following commands in <filename>/mnt/loop1/etc/init.d/rcS</filename>.
318You may also want to add some of the processes shown in the non-devfs
319version that follows.</para>
322mount -t devfs devfs /dev
323/sbin/devfsd /dev</userinput></screen>
325<para>If you don't use devfsd, but created a static /dev/directory
326using <command>MAKEDEV</command>, or any similar process, the
327<filename>rcS</filename> script will do slightly different things.
328Also, don't forget that it is creating the
329<filename>/etc/passwd</filename> and <filename>/etc/group</filename>
330files, thus saving space on the diskette.</para>
332<para>The script made next will mount <filename>/proc</filename>, turn
333on swap (no harm is done if it fails), make the
334<filename>/etc/passwd</filename> and <filename>/etc/group</filename>
335files, create a log directory and turn on swapping. Create the script
338<screen><userinput><command>mkdir -p /mnt/loop1/etc/init.d
339cat &gt;/mnt/loop1/etc/init.d/rcS &lt;&lt; EOD</command>
341mount -t proc proc /proc
342swapon -a
344echo "root:x:0:0:root:/root:/bin/bash" &gt; /etc/passwd
346<command>cat &gt; /etc/group &lt;&lt;EOF</command>
360chmod 644 /etc/passwd /etc/group</command>
362mkdir /var/log
365chmod u+x /mnt/loop1/etc/init.d/rcS</command></userinput></screen>
367<para>Unless you add a lot to this script, which <emphasis>is</emphasis>
368encouraged, the above should be reasonably close to what you need.</para>
370<para><emphasis>Install packages</emphasis></para>
372<para>There are two packages that must be installed. The <application><ulink
374package incorporates the core functions that provide a shell and many
375basic utilities. A file system package, like <application><ulink
376url="">e2fsprogs</ulink></application>, or
377a package for the file system you are using will provide a minimal
378set of utilities for file system checking and reconstruction. The whole
379package will not be installed, but only certain needed components.</para>
381<para>If you use devfsd, you will also need to install that software.</para>
383<para>Install <application><ulink
385into the initial ramdisk image. Busybox incorporates many UNIX utility
386program functions into a single small executable file.</para>
388<screen><userinput><command>make &amp;&amp;
389make PREFIX=/mnt/loop1 install &amp;&amp;
390&gt; /mnt/loop1/var/utmp</command></userinput></screen>
392<para>A <filename>var/utmp</filename> is made because <application><ulink
394needs it for the reboot command to work properly. If this file doesn't
395exist when <application><ulink
397is started, the reboot command will not work. This would be a bad thing
398for people that have no reset button available to them.</para>
400<para>If you use devfs to create devices on the fly and free up precious
401inodes on the floppy, you'll also install devfsd to facilitate the
402devices that <application><ulink
404expects to find. Use the following commands to do the install.</para>
406<screen><userinput><command>mv GNUmakefile Makefile &amp;&amp;
407make &amp;&amp;
408make PREFIX=/mnt/loop1 install &amp;&amp;</command></userinput></screen>
410<para><emphasis>Install part of <application>e2fsprogs</application></emphasis></para>
412<para>If you use the ext2 or ext3 (journaling) file system, you can use
413the commands below to install the minimal functionality that should
414allow you to get your hard drives usable again. If you use ext3, keep in
415mind that it is a part of the <application>e2fsprogs</application>
416package and you can get the components, which are mostly hard links,
417from the same places shown below. If you use some other file system,
418such as reiserfs, you should apply the <emphasis>principals</emphasis>
419you see here to install parts of that package instead.</para>
422mkdir build &amp;&amp;
423cd build &amp;&amp;
424../configure --prefix=/mnt/loop1/usr --with-root-prefix="" \
425 --disable-swapfs --disable-debugfs \
426 --enable-dynamic-e2fsck --disable-nls --disable-evms \
427 --disable-rpath &amp;&amp;
428make LDFLAGS="$LDFLAGS" &amp;&amp;
429strip -p --strip-unneeded --remove-section=.comment \
430 -o /mnt/loop1/sbin/mke2fs misc/mke2fs &amp;&amp;
431strip -p --strip-unneeded --remove-section=.comment \
432 -o /mnt/loop1/sbin/e2fsck e2fsck/e2fsck &amp;&amp;
433chmod 555 /mnt/loop1/sbin/{mke2f,e2fsck}
436<para><emphasis>Two useful utilities</emphasis></para>
438<para>There are two very useful utilities that any rescue disk should
439have to help in faster and more accurate recovery. The first is a
440partitioning utility. The <command>sfdisk</command> program is
441used here because of its small size and great power. Be warned though -
442it is not what is considered to be "user friendly". But the
443<command>fdisk</command> and <command>cfdisk</command> programs are
444substantially larger or require more shared objects, like ncurses.</para>
446<para>The second utility is an editor. Most graphical editors are
447inherently too large and also require additional shared objects. For
448this reason, <command>ed</command> is used here. It is small, requires
449no additional shared objects and is a regex-based editor that is the
450ancestor to almost all subsequent editors that support regex-based
451editing, whether graphical or not. It is a "context editor" and offers
452powerful, but non-graphical, editing features. There are many other
453editors that may be suitable - feel free to use one of them instead.</para>
455<para>Read the <application>busybox</application>
456<filename>INSTALL</filename> and <filename>README</filename> files to
457see how to include a <command>vi</command> editor. It has not been
458investigated here yet, so it may or may not easily fit onto a single
459diskette image such as is made here.</para>
461<para>You can install these or not, but it is important for you to have
462some capability such as these offer. Exactly how you would install the
463utilities you choose will have to be determined by you.</para>
465<para><command>Sfdisk</command> and <command>ed</command> are installed
466by, essentially, copying them from your host. Strip is used, just to
467assure that they carry no "excess baggage", even though the base
468<acronym>LFS</acronym> install should have stripped them already. Use
469the following commands:</para>
471<screen><userinput><command>strip -p --strip-unneeded --remove-section=.comment \
472 -o /mnt/loop1/sbin/sfdisk /sbin/sfdisk
473strip -p --strip-unneeded --remove-section=.comment \
474 -o /mnt/loop1/bin/ed /bin/ed
475chmod 555 /mnt/loop1/sbin/sfdisk /mnt/loop1/bin/ed
478<para>Also, keeping in mind your space limitations, copy any other
479binaries and libraries you need to the image. Use the
480<command>ldd</command> command to see which libraries you will need to
481copy over for any executables. Don't forget to also strip them
482<emphasis>before</emphasis> copying them to the ramdisk image or use the
483<command>strip</command>, as above, to "copy" them.</para>
485<para><emphasis>Set up the lib directory</emphasis></para>
487<para>Once you have installed all the utilities from above and any
488additional ones you want, use the <command>ldd</command> command, as
489mentioned above, on those that were not listed in this document. If
490any additional libraries are needed, add them into the setup commands
491shown next.</para>
493<para>If you installed only those things shown above, the shared objects
494needed will be minimal. You can add them to the ramdisk image with:</para>
496<screen><userinput><command>strip -p --strip-unneeded --remove-section=.comment \
497&nbsp;&nbsp;&nbsp;&nbsp;-o /mnt/loop1/lib/ /lib/ &amp;&amp;
498strip -p --strip-unneeded --remove-section=.comment \
499&nbsp;&nbsp;&nbsp;&nbsp;-o /mnt/loop1/lib/ /lib/ &amp;&amp;
500strip -p --strip-unneeded --remove-section=.comment \
501&nbsp;&nbsp;&nbsp;&nbsp;-o /mnt/loop1/lib/ /lib/ &amp;&amp;
502chmod 555 /mnt/loop1/lib/{,,}
505<para>Note that the above commands change the names of the libraries,
506eliminating the need for the usual symlinks. If you add any additional
507shared objects, be alert for similar opportunities and also the pitfalls
508that may be present.</para>
510<para><emphasis>Make the compressed initrd</emphasis></para>
512<para>Unmount the loopback file. If you used <command>mount</command>'s
513<command>-o loop</command> option, the "bond" between the loop device
514and the file will be removed when the unmount is done. Just omit the
515<command>losetup -d /dev/loop1</command> from the following
516sequence. The -9 parameter is used with gzip to make the smallest
517possible compressed image. To make sure it will fit on the diskette,
518list the file's size.</para>
520<screen><userinput><command>umount /mnt/loop1 &amp;&amp;
521losetup -d /dev/loop1 &amp;&amp; # Omit if mount's -o loop was used
522gzip -9 &lt; /tmp/rfloppy &gt; /tmp/rootfs.gz
523ls -l /tmp/rootfs.gz
526<para><emphasis>Join a kernel image and initrd onto a diskette</emphasis></para>
528<para>Now the kernel image and initial ramdisk image will be written to
529the boot diskette. Before doing this, calculate the number of blocks
530needed for the kernel and for the initrd, individually, by dividing each
531size by 1024 and adding one if there is any remainder. Add these two
532results together. They must total 1,440 or fewer blocks. If they total
533more than this, don't worry too much. Changes to make a two-diskette
534set are presented later. Of course, you could reexamine your choices and
535try to shrink either the kernel or the initial ramdisk image.</para>
537<para>To make a single-floppy rescue, using devfs, use the following
538commands. If you use the static <filename>/dev</filename> setup, use
539<filename>/dev/fd0</filename> instead of the /dev/floppy/0.</para>
541<screen><userinput><command>dd if=rescueimage of=/dev/floppy/0 bs=1k</command>
542<command>rdev /dev/floppy/0 0,0
543rdev -R /dev/floppy/0 0</command></userinput></screen>
545<para><emphasis>Command explanations</emphasis></para>
547<para><command>rdev /dev/floppy/0 0,0</command>: sets the root file system
548the kernel will use when it boots. Because it loads an initrd, it will
549automatically set that as the root device, initially. So, the
550<command>0,0</command> gives it "no value", telling the kernel to not
551mount any other device. Some folks give <filename>/dev/fd0</filename> or
552something similar. But this has effect <emphasis>only</emphasis> when
553<command>linuxrc</command> (really <application><ulink
555</ulink></application>) exits and the normal <command>init</command>
556processes get invoked. Since this is not being done here, and the floppy
557is <emphasis>not</emphasis> a valid file system, it would be useless
558here. A hard drive would be a better choice if you are looking to
559automatically bring the system up after repair. Since <application><ulink
561</ulink></application> provides the <command>reboot</command> command,
562automatic initialization is not needed.</para>
564<para>The <command>rdev -R /dev/floppy/0 0</command> will set the
565"root flags" to zero. They have no use in this application.</para>
567<para>The <command>dd</command> from above showed some results, like</para>
569<screen> 480+1 records in
570 480+1 records out</screen>
572<para>In this example the rescueimage (kernel) was 480+1 blocks in size.
573Make sure that this number, which may be different for you, matches your
574calculations from above. You need to calculate a "magic number" now
575that will be inserted into the kernel image. The value consists of three
576significant parts. Two are discussed here. The third is touched upon
579<para>Bits 0 - 10 will contain the size of the kernel image, in blocks,
580that you calculated above, and which should match the results from the
581dd above. Bit 14 (the 15th bit, which is 2 to the 14th power, or 16,384)
582is a flag that, when set to 1, tells the kernel an initial ramdisk is to
583be loaded. So for the single-floppy rescue diskette, the two numbers
58416,384 and 481 (or whatever number is right for your kernel size) are
585added together to produce a decimal value, like 16865. This value is
586inserted into the proper place in the kernel image by the
587<command>rdev</command> command done next.</para>
589<para>Insert the "magic number" into the kernel image and then write the
590root file system right after the kernel on the floppy by executing the
591following commands, with the proper numbers inserted. Notice that the
592<command>seek</command> parameter's number must be the size, in blocks,
593of your kernel image. If you use the static <filename>/dev</filename>
594setup, use <filename>/dev/fd0</filename> in the commands below, instead
595of <filename>/dev/floppy/0</filename>.</para>
597<screen><userinput><command>rdev -r /dev/floppy/0 16865
598dd if=/tmp/rootfs.gz of=/dev/floppy/0 bs=1k seek=481</command></userinput></screen>
600<para>In this command, <command>seek</command> was used to position to
601the block following the end of the kernel (480+1) and begin writing the
602root file system to the floppy.</para>
607<title>A Two-diskette Rescue Setup</title>
609<para>If you just can't live with a single-diskette rescue system, here
610is what to do to make a simple two-diskette system. Note that the
611endless possibilities presented by the availability of
612<command>linuxrc</command> and other components are not addressed
613here. Here you will just use the kernel's ability to prompt for a second
614diskette that contains the initrd image and load it.</para>
616<para>Modify the above instructions as follows. First a different magic
617number is needed. The 15th bit (bit 14) still needs to be set, but the
618size of the kernel image, in blocks, is replaced with a zero. The third
619component, which was not discussed above, is now used. This is the 16th
620bit (bit 15) of the "magic number". When set, it tells the kernel to ask
621the user to insert the "root" floppy. It then loads the initrd image
622from that diskette. Because the size of the kernel image was replaced
623by zero, the kernel starts loading from the "zero'th" block (the first
624one) on the second diskette.</para>
626<para>The 16th bit (bit 15) represents 2 raised to the 15th power, or
62732,768. So the new magic number is 32,768 + 16384, which is 49,152. This
628value tells the kernel to prompt for, and then load, an initial ramdisk
629image from the first block on the inserted floppy. So your first
630modification is to the command to write the "magic number" to the kernel
631image on the diskette.</para>
633<screen><userinput><command>rdev -r /dev/floppy/0 49152</command></userinput></screen>
635<para>Note that the initrd image is <emphasis>not</emphasis> copied to
636the diskette yet. Remove the boot diskette and insert another diskette
637that will hold your root file system. Run this modified command (don't
638forget to use <filename>/dev/fd0</filename> if you don't use devfs).
639Note that no <command>seek</command> parameter is used.</para>
641<screen><userinput><command>dd if=/tmp/rootfs.gz of=/dev/floppy/0 bs=1k</command></userinput></screen>
643<para>That's all there is to it. The possibilities from here are limited only
644by your imagination and tenacity in pursueing enhancements. And your
645willingness to research available documentation. A good starting point
646is the "Documentation" directory in your kernel source tree. More help
647may be gained at
648<ulink url=""><acronym>LFS </acronym> Hints</ulink>
649(please use a mirror site that is suitable) and
650<ulink url="">TLDP</ulink>.</para>
Note: See TracBrowser for help on using the repository browser.