source: postlfs/config/bootdisk.xml@ f8d632a

10.0 10.1 11.0 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 ken/refactor-virt krejzi/svn lazarus nosym perl-modules qt5new systemd-11177 systemd-13485 trunk xry111/git-date xry111/git-date-for-trunk xry111/git-date-test
Last change on this file since f8d632a was f8d632a, checked in by Bruce Dubbs <bdubbs@…>, 17 years ago

New XML Chapter 3

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

  • Property mode set to 100644
File size: 32.4 KB
1<?xml version="1.0" encoding="ISO-8859-1"?>
2<!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.3//EN"
3 "" [
4 <!ENTITY % general-entities SYSTEM "../../general.ent">
5 %general-entities;
8<sect1 id="postlfs-config-bootdisk">
9<?dbhtml filename="bootdisk.html"?>
10<title>Creating a Custom Boot Disk</title>
13<title>Decent Rescue Boot Disk Needs</title>
14<para>This section is really about creating a <emphasis>rescue</emphasis>
15diskette. As the name <emphasis>rescue</emphasis> implies, the host
16system has a problem, often lost partition information or corrupted file
17systems, that prevents it from booting and/or operating normally. For
18this reason, you <emphasis>must not</emphasis> depend on resources from
19the host being "rescued". To presume that any given partition or hard
20drive <emphasis>will</emphasis> be available is a risky presumption.</para>
22<para>Heeding the warning, the rescue disk created here has no
23dependency on the host system's resources, other than basic bootability
24and hardware soundness. At a minimum, the most common sorts of failures
25requiring a rescue boot disk should be addressed by the contents of the
26boot disk. This would include the common loss of partitioning (master
27boot record is lost or corrupted), file system corruption, and the need
28to allow creation and editing of files that may have been lost or
29corrupted, possibly as an effect of the other two problems.</para>
31<para>Additional utilities should be available to search for text or
32files, copy, move and remove files, and many other normal operations
33that might be expected to be needed when reconstructing.</para>
37<title>This Minimal Decent Rescue Disk</title>
39<para>The intent here is to create a "rescue boot disk" that will support
40the common operations listed above. These functions are provided by
41including selected executables from <application><ulink
43</ulink></application> and <application><ulink
45A basic editor and rudimentary disk partitioning utility may also be
46optionally included.</para>
48<para>This, however, is not the limit. A minimal disk is described
49here, but you can add anything you can fit on the floppy. Furthermore,
50if one floppy is not enough to meet your needs, you can make a
51multi-diskette rescue set that means, essentially, the sky is the limit.
52This is discussed below. The number of other possible variations are
53just too numerous to mention here.</para>
57<title>Build the Rescue Boot Disk</title>
61<para>You should have known-good floppy diskettes available. Some
62people prefer to use the <command>fdformat</command> command to prepare
63these because it also does a verification. See the man page for more
64details. Another good idea is to always prepare duplicates of the
65rescue diskette. Media does deteriorate.</para>
67<para>These instructions presume a base <acronym>LFS</acronym> install
68using ext2/ext3 file systems.</para>
70<para>You need to have loopback device support enabled in your host's
71kernel to use this procedure.</para>
73<para>You should make a custom kernel that includes only those features
74needed to rescue your system, so it will have the least size. No
75sense in building in support for things like <application>XFree86</application>,
76<application>DRI</application>, etc, as most rescues are performed from the
77command prompt. Along the same lines, if you have <xref linkend="gcc2"/>, it
78is known to produce smaller kernels. So you might want to use that
79compiler for this kernel. If you do so, don't overlook any loadable
80modules (which are not addressed here) you might need - they need to be
81compiled with the same compiler used to make the kernel.</para>
83<para>The rescue image must include support for the file system of your
84choice (we presume ext2/3 here), ramdisk and initial ramdisk (initrd).
85Disable everything that you can in the kernel configuration. You should
86keep support for the proc file system and tempfs file system enabled
87because of their general utility. The proc file system is needed for
88the <command>mount</command> to report properly.</para>
90<para>If you install <emphasis>only</emphasis> the minimal set of
91components shown in this document, you will need a kernel that is 643 or
92fewer blocks in size. If you want the optional programs - a very basic
93editor, like <command>ed</command>, and rudimentary disk partitioning, like
94<command>sfdisk</command> - the kernel will need to be 595 or fewer blocks in
95size. This should not be a major problem unless your needs are fairly
96esoteric. On the system used to develop this version of the procedures, using
97only ext2 file systems and not using networking or <acronym>CD</acronym>s for
98recovery, the kernel image is only 481 blocks. And there may be more to gain - it has not been closely examined for additional gains.</para>
100<para>This kernel image will be called "rescueimage" hereinafter. You
101can actually name your image anything you like and just use its name
102instead in any commands that include "rescueimage".</para>
104<para>If you can not get your rescueimage down to the size needed to allow
105all you need on the ramdisk image, don't fret. You can always build a
106two diskette set, one boot and one root diskette. The kernel will prompt
107you to insert the root file system diskette. This will allow room for a
108compressed ramdisk image of 1440 blocks and a rescueimage of the same
111<para>The rescueimage size limits given above are likely to vary as
112local system-specific configurations change. Use them only as a
113guideline and not as gospel. The size of rescueimage as shown by
114<command>ls -sk</command> is only an approximation because of some
115"overhead". On the system used to develop this version of these
116procedures, that command shows 488 blocks but the actual number of
117blocks written is only 480 and a fraction, which means that 481 blocks
118are actually used.</para>
122<title>Rescue Disk Build Process</title>
124<para>The basic process will be:</para>
126<listitem><para>make a mount point for a file system</para></listitem>
127<listitem><para>make an empty file to hold the file system</para></listitem>
128<listitem><para>bind the empty file to a loopback device</para></listitem>
129<listitem><para>make a 4MB file system</para></listitem>
130<listitem><para>mount the file system</para></listitem>
131<listitem><para>add components to the file system</para></listitem>
132<listitem><para>make the compressed initrd</para></listitem>
133<listitem><para>join rescueimage and initrd onto a diskette</para></listitem>
136<para>The initial ramdisk will be automatically loaded at boot time if
137setup is done correctly.</para>
139<para><emphasis>Make a mount point and an empty file to hold a file
142<screen><userinput><command>mkdir -p /mnt/loop1
143dd if=/dev/zero of=/tmp/rfloppy bs=1k count=4096</command></userinput></screen>
145<para><emphasis>Command explanations</emphasis></para>
147<para><command>dd</command>: This is a generalized input-to-output copy
148utility that also has many transformation capabilities.</para>
150<para><parameter>if=/dev/zero</parameter>: This parameter assigns
151<command>dd</command>'s input file to a device that returns an infinite
152stream of zeroes.</para>
154<para><parameter>of=/tmp/rfloppy</parameter>: This parameter directs
155<command>dd</command>'s output to <filename>/tmp/rfloppy</filename>.</para>
157<para><parameter>bs=1k count=4096</parameter>: These parameters tell
158<command>dd</command> to read and write in "chunks" of 1024 bytes and
159process 4096 "chunks".</para>
161<para><emphasis>Bind the file to a loopback device, make a file system and mount it.</emphasis></para>
163<para>The reason these commands are used is that they work regardless of
164the version of <command>mount</command> (older ones don't have the
165<userinput>-o loop</userinput> option) or if <filename>/etc/mtab</filename>
166is symlinked to <filename class="directory">/proc</filename> (which causes mount to be
167unable to properly "unbind" a loop device, due to "lost" information).
168An alternate set of commands is provided, after these three commands,
169that you can use if you don't have either of these situations.</para>
171<screen><userinput><command>losetup /dev/loop1 /tmp/rfloppy
172mke2fs -m 0 -N 504 /dev/loop1
173mount -t ext2 /dev/loop1 /mnt/loop1</command></userinput></screen>
175<para><emphasis>Command explanations</emphasis></para>
177<para><command>losetup /dev/loop1 /tmp/rfloppy</command>: This
178command "binds" a loopback device to the empty file.</para>
180<para><command>mke2fs -m 0 -N 504 /dev/loop1</command>: This
181command makes an ext2 file system on the loopback device (which really
182means it is created in the file to which the loopback device is bound)
183and reserves no blocks. The <userinput>-N 504</userinput> parameter causes
184only 504 inodes to be allocated, leaving more space for other things needed
185in the file system.</para>
187<para><command>mount -t ext2 /dev/loop1 /mnt/loop1</command>: This
188mounts the file system just created, just as if it were a real device,
189like a hard drive or diskette. This allows all the normal system I/O
190commands to operate as if a real device were present.</para>
192<para>If your <command>mount</command> supports the
193<option>-o loop</option> option <emphasis>and</emphasis> your
194<filename>/etc/mtab</filename> is a real file, rather than a symlink to
195<filename class="directory">/proc</filename>, the three above commands can be replaced
196by these next two commands.</para>
198<screen><userinput><command>mke2fs -F -m 0 -N 504 /tmp/rfloppy
199mount -o loop /tmp/rfloppy /mnt/loop1</command></userinput></screen>
201<para><emphasis>Command explanations</emphasis></para>
203<para><command>mke2fs -F -m 0 -N 504 /tmp/rfloppy</command>: As before,
204a file system is made, with only 504 inodes and no reserved blocks, that
205will be bound to a loopback device. The <userinput>-F</userinput> parameter
206just suppresses an irritating question issued when
207<command>mke2fs</command> realizes that you are not accessing a
210<para><command>mount -o loop /tmp/rfloppy /mnt/loop1</command>: This
211command tells <command>mount</command> to bind the named file to a
212loopback device it automatically selects (the first available) and mount
213the device on <filename class="directory">/mnt/loop1</filename>.</para>
215<para><emphasis>Add components to the file system</emphasis></para>
217<para><emphasis>A cautionary note:</emphasis> if you are not running in a
218<command>chroot</command> environment, be sure that you do not accidentally
219omit the <filename class="directory">/mnt/loop1</filename> reference in the commands. If
220you do so, you might replace the equivalent components on your host with
221the components that are installed by these procedures. Even if you are
222in a chroot environment, you may need to be careful if the environment
223is your freshly built <acronym>LFS</acronym> system which you intend to
224use as a host in the future.</para>
226<para>First, to have as much free space as possible, remove the
227<filename>lost+found</filename> directory, which is not needed because
228it is only used by <command>fsck</command>. Since <command>fsck</command>
229will never be run on this file system, it is unneeded.</para>
231<screen><userinput><command>rmdir /mnt/loop1/lost+found/</command></userinput></screen>
233<para>Now make a minimal set of directories.</para>
235<screen><userinput><command>mkdir /mnt/loop1/{dev,proc,etc,sbin,bin,lib,mnt,usr,var}</command></userinput></screen>
237<para>Add needed device files to the initrd image. If you use devfs,
238the following command works well, as you only have the devices you use
241<screen><userinput><command>cp -dpR /dev/* /mnt/loop1/dev</command></userinput></screen>
243<para>If you used <command>MAKEDEV</command> to create your devices on
244your host, you'll want to use something similar to this longer command,
245to minimize wasting space with unneeded inodes.</para>
247<para><emphasis>You must modify this to suit your rescueimage configuration and
248other needs.</emphasis> For example, you may need
249<acronym>SCSI</acronym> devices and may not need
250frame buffer devices or the pseudo-terminal directory. Also, the number
251of hard drives and partitions that you include should be the minimal
252that you need. Extensive analysis has not been done on the list below,
253so there are more inodes and space to be gained by "fine tuning" this
256<screen><userinput><command>mkdir /mnt/loop1/dev/pts
257cp -a \
258&nbsp;&nbsp;&nbsp;&nbsp;/dev/null /dev/console \
259&nbsp;&nbsp;&nbsp;&nbsp;/dev/fb[0-7] /dev/fd /dev/fd0 /dev/fd0h1440 /dev/full \
260&nbsp;&nbsp;&nbsp;&nbsp;/dev/hda* /dev/hdb* /dev/hdc* /dev/hdd* /dev/initctl /dev/kmem \
261&nbsp;&nbsp;&nbsp;&nbsp;/dev/loop[0-3] /dev/lp0 /dev/mem /dev/port \
262&nbsp;&nbsp;&nbsp;&nbsp;/dev/psaux /dev/ram \
263&nbsp;&nbsp;&nbsp;&nbsp;/dev/ram0 /dev/ram1 /dev/ram2 /dev/ram3 /dev/random /dev/rtc \
264&nbsp;&nbsp;&nbsp;&nbsp;/dev/shm /dev/stderr /dev/stdin /dev/stdout /dev/tty \
265&nbsp;&nbsp;&nbsp;&nbsp;/dev/tty[0-9] /dev/ttyS0 /dev/ttyS1 /dev/urandom /dev/zero \
268<para><emphasis>What is needed in the <filename class="directory">/etc</filename>
271<para>If you choose, you can copy all or selected parts of your
272<filename>/etc/passwd</filename> and <filename>/etc/group</filename>
273files. But even if each is less than 1024 bytes, you will lose two
274inodes and two blocks of space on the initial ramdisk. This only really
275matters because of trying to squeeze everything onto a 1.44MB
276diskette. Every little bit helps. The strategy taken here is to create
277these two files as part of the rescue boot and initialization process.
278The commands that make the two files will be embedded inside the
279<filename>rcS</filename> script that <filename>linuxrc</filename>
280(really <application><ulink
282</ulink></application>) invokes after the initrd is
283loaded. This way no more inodes or blocks are used on the
284diskette to carry these files.</para>
286<para>Some might like to copy their <filename class="directory">/etc/rc*</filename>
287directory into the ramdisk image, but this may have no value, other
288than archival use, in a worst-case recovery scenario. If you want
289automatic initialization of the system after repair, they may have some
290value. But few people need or want this to happen. If the file system
291on the hard drives are corrupted, what good will mount scripts do? Some
292scripts may be useful, like access to a network to copy over backup data
293when the hard drive's file systems are usable again. The point is that
294you should copy only the parts that you can use because space is at a
295premium on the diskette.</para>
297<para>Here, only the <filename>fstab</filename> will be included. This
298is handy because it eases mounting of partitions that may be useful and
299also can be examined and used as a guide as to what is available and
300what may need reconstruction. Because it may be larger than needed, you
301should edit it to remove any useless entries and minimize commentary.
302No other editing is needed because the boot scripts are not included and
303no automatic mounting will be done using the <filename>fstab</filename>.
304If you decide to include some boot scripts that might try to mount
305things, change the <filename>fstab</filename>'s entries to
306<command>noauto</command> in the options field so they don't cause an
307attempt to mount a potentially corrupt partition. Copy it to
308<filename class="directory">/tmp</filename>, edit it as desired and then:</para>
310<screen><userinput><command>cp -a /tmp/fstab /mnt/loop1/etc</command></userinput></screen>
312<para>Now the initialization script will be added. As mentioned above,
313<command>linuxrc</command> is symlinked to <application><ulink
316After the kernel and initial ramdisk have been loaded, the kernel gives
317control to <command>linuxrc</command> (<application><ulink
319</ulink></application>). It wants to run an <filename class="directory">/etc/init.d/rcS</filename>
320script to do any initial setup.</para>
322<para>If you use devfsd, you will need to set up the
323<filename>rcS</filename> script to handle the devfsd startup. Put the
324following commands in <filename class="directory">/mnt/loop1/etc/init.d/rcS</filename>.
325You may also want to add some of the processes shown in the non-devfs
326version that follows.</para>
329mount -t devfs devfs /dev
330/sbin/devfsd /dev</userinput></screen>
332<para>If you don't use devfsd, but created a static <filename
334directory using <command>MAKEDEV</command>, or any similar process, the
335<filename>rcS</filename> script will do slightly different things.
336Also, don't forget that it is creating the
337<filename>/etc/passwd</filename> and <filename>/etc/group</filename>
338files, thus saving space on the diskette.</para>
340<para>The script made next will mount <filename class="directory">/proc</filename>, turn
341on swap (no harm is done if it fails), make the
342<filename>/etc/passwd</filename> and <filename>/etc/group</filename>
343files, create a log directory and turn on swapping. Create the script
346<screen><userinput><command>mkdir -p /mnt/loop1/etc/init.d
347cat &gt;/mnt/loop1/etc/init.d/rcS &lt;&lt; EOD</command>
349mount -t proc proc /proc
350swapon -a
352echo "root:x:0:0:root:/root:/bin/bash" &gt; /etc/passwd
354<command>cat &gt; /etc/group &lt;&lt;EOF</command>
368chmod 644 /etc/passwd /etc/group</command>
370mkdir /var/log
373chmod u+x /mnt/loop1/etc/init.d/rcS</command></userinput></screen>
375<para>Unless you add a lot to this script, which <emphasis>is</emphasis>
376encouraged, the above should be reasonably close to what you need.</para>
378<para><emphasis>Install packages</emphasis></para>
380<para>There are two packages that must be installed. The <application><ulink
382package incorporates the core functions that provide a shell and many
383basic utilities. A file system package, like <application><ulink
384url="">e2fsprogs</ulink></application>, or
385a package for the file system you are using, will provide a minimal
386set of utilities for file system checking and reconstruction. The whole
387package will not be installed, but only certain needed components.</para>
389<para>If you use devfsd, you will also need to install that software.</para>
391<para>Install <application><ulink
393into the initial ramdisk image. Busybox incorporates many Unix utility
394program functions into a single small executable file.</para>
396<screen><userinput><command>make &amp;&amp;
397make PREFIX=/mnt/loop1 install &amp;&amp;
398&gt; /mnt/loop1/var/utmp</command></userinput></screen>
400<para>A <filename>var/utmp</filename> is made because <application><ulink
402needs it for the reboot command to work properly. If this file doesn't
403exist when <application><ulink
405is started, the reboot command will not work. This would be a bad thing
406for people that have no reset button available to them.</para>
408<para>If you use devfs to create devices on the fly and free up precious
409inodes on the floppy, you'll also install devfsd to facilitate the
410devices that <application><ulink
412expects to find. Use the following commands to do the install.</para>
414<screen><userinput><command>mv GNUmakefile Makefile &amp;&amp;
415make &amp;&amp;
416make PREFIX=/mnt/loop1 install</command></userinput></screen>
418<para><emphasis>Install part of <application>e2fsprogs</application></emphasis></para>
420<para>If you use the ext2 or ext3 (journaling) file system, you can use
421the commands below to install the minimal functionality that should
422allow you to get your hard drives usable again. If you use ext3, keep in
423mind that it is a part of the <application>e2fsprogs</application>
424package and you can get the components, which are mostly hard links,
425from the same places shown below. If you use some other file system,
426such as reiserfs, you should apply the <emphasis>principals</emphasis>
427you see here to install parts of that package instead.</para>
430mkdir build &amp;&amp;
431cd build &amp;&amp;
432../configure --prefix=/mnt/loop1/usr --with-root-prefix="" \
433 --disable-swapfs --disable-debugfs \
434 --enable-dynamic-e2fsck --disable-nls --disable-evms \
435 --disable-rpath &amp;&amp;
436make LDFLAGS="$LDFLAGS" &amp;&amp;
437strip -p --strip-unneeded --remove-section=.comment \
438 -o /mnt/loop1/sbin/mke2fs misc/mke2fs &amp;&amp;
439strip -p --strip-unneeded --remove-section=.comment \
440 -o /mnt/loop1/sbin/e2fsck e2fsck/e2fsck &amp;&amp;
441chmod 555 /mnt/loop1/sbin/{mke2fs,e2fsck}
444<para><emphasis>Two useful utilities</emphasis></para>
446<para>There are two very useful utilities that any rescue disk should
447have, to help in faster and more accurate recovery. The first is a
448partitioning utility. The <command>sfdisk</command> program is
449used here because of its small size and great power. Be warned though -
450it is not what is considered to be "user friendly". But the
451<command>fdisk</command> and <command>cfdisk</command> programs are
452substantially larger or require more shared objects, like <application>ncurses
455<para>The second utility is an editor. Most graphical editors are
456inherently too large and also require additional shared objects. For
457this reason, <command>ed</command> is used here. It is small, requires
458no additional shared objects and is a regex-based editor that is the
459ancestor to almost all subsequent editors that support regex-based
460editing, whether graphical or not. It is a "context editor" and offers
461powerful, but non-graphical, editing features. There are many other
462editors that may be suitable - feel free to use one of them instead.</para>
464<para>Read the <application>busybox</application>
465<filename>INSTALL</filename> and <filename>README</filename> files to
466see how to include a <command>vi</command> editor. It has not been
467investigated here yet, so it may or may not easily fit onto a single
468diskette image such as is made here.</para>
470<para>You can install these or not, but it is important for you to have
471some capability such as these offer. Exactly how you would install the
472utilities you choose will have to be determined by you.</para>
474<para><command>Sfdisk</command> and <command>ed</command> are installed
475by, essentially, copying them from your host. Strip is used, just to
476assure that they carry no "excess baggage", even though the base
477<acronym>LFS</acronym> install should have stripped them already. Use
478the following commands:</para>
480<screen><userinput><command>strip -p --strip-unneeded --remove-section=.comment \
481 -o /mnt/loop1/sbin/sfdisk /sbin/sfdisk
482strip -p --strip-unneeded --remove-section=.comment \
483 -o /mnt/loop1/bin/ed /bin/ed
484chmod 555 /mnt/loop1/sbin/sfdisk /mnt/loop1/bin/ed</command></userinput></screen>
486<para>Also, keeping in mind your space limitations, copy any other
487binaries and libraries you need to the image. Use the
488<command>ldd</command> command to see which libraries you will need to
489copy over for any executables. Don't forget to also strip them
490<emphasis>before</emphasis> copying them to the ramdisk image or use the
491<command>strip</command>, as above, to "copy" them.</para>
493<para><emphasis>Set up the lib directory</emphasis></para>
495<para>Once you have installed all the utilities from above and any
496additional ones you want, use the <command>ldd</command> command, as
497mentioned above, on those that were not listed in this document. If
498any additional libraries are needed, add them into the setup commands
499shown next.</para>
501<para>If you installed only those things shown above, the shared objects
502needed will be minimal. You can add them to the ramdisk image with:</para>
504<screen><userinput><command>strip -p --strip-unneeded --remove-section=.comment \
505&nbsp;&nbsp;&nbsp;&nbsp;-o /mnt/loop1/lib/ /lib/ &amp;&amp;
506strip -p --strip-unneeded --remove-section=.comment \
507&nbsp;&nbsp;&nbsp;&nbsp;-o /mnt/loop1/lib/ /lib/ &amp;&amp;
508strip -p --strip-unneeded --remove-section=.comment \
509&nbsp;&nbsp;&nbsp;&nbsp;-o /mnt/loop1/lib/ /lib/ &amp;&amp;
510chmod 555 /mnt/loop1/lib/{,,}</command></userinput></screen>
512<para>Note that the above commands change the names of the libraries,
513eliminating the need for the usual symlinks. If you add any additional
514shared objects, be alert for similar opportunities and also the pitfalls
515that may be present.</para>
517<para><emphasis>Make the compressed initrd</emphasis></para>
519<para>Unmount the loopback file. If you used <command>mount</command>'s
520<option>-o loop</option> option, the "bond" between the loop device
521and the file will be removed when the unmount is done. Just omit the
522<command>losetup -d /dev/loop1</command> from the following
523sequence. The <userinput>-9</userinput> parameter is used with
524<command>gzip</command> to make the smallest possible compressed image. To
525make sure it will fit on the diskette, list the file's size.</para>
527<screen><userinput><command>umount /mnt/loop1 &amp;&amp;
528losetup -d /dev/loop1 &amp;&amp; # Omit if mount's -o loop was used
529gzip -9 &lt; /tmp/rfloppy &gt; /tmp/rootfs.gz
530ls -l /tmp/rootfs.gz</command></userinput></screen>
532<para><emphasis>Join rescueimage and initrd onto a diskette</emphasis></para>
534<para>Now the rescueimage and initial ramdisk image will be written to
535the boot diskette. Before doing this, calculate the number of blocks
536needed for rescueimage and for <filename>/tmp/rootfs.gz</filename>
537(the initial ramdisk), individually, by dividing each
538size by 1024 and adding one if there is any remainder. Add these two
539results together. They must total 1,440 or fewer blocks. If they total
540more than this, don't worry too much. Changes to make a two-diskette
541set are presented later. Of course, you could reexamine your choices and
542try to shrink either the rescueimage or the initial ramdisk image.</para>
544<para>To make a single-floppy rescue, using devfs, use the following
545commands. If you use the static <filename class="directory">/dev</filename> setup, use
546<filename>/dev/fd0</filename> instead of the /dev/floppy/0.</para>
548<screen><userinput><command>dd if=rescueimage of=/dev/floppy/0 bs=1k
549rdev /dev/floppy/0 0,0
550rdev -R /dev/floppy/0 0</command></userinput></screen>
552<para><emphasis>Command explanations</emphasis></para>
554<para><command>rdev /dev/floppy/0 0,0</command>: sets the root file system
555the kernel will use when it boots. Because it loads an initrd, it will
556automatically set that as the root device, initially. So, the
557<option>0,0</option> gives it "no value", telling the kernel to not
558mount any other device. Some folks give <filename>/dev/fd0</filename> or
559something similar. But this has effect <emphasis>only</emphasis> when
560<command>linuxrc</command> (really <application><ulink
562</ulink></application>) exits and the normal <command>init</command>
563processes get invoked. Since this is not being done here, and the floppy
564is <emphasis>not</emphasis> a valid file system, it would be useless
565here. A hard drive would be a better choice if you are looking to
566automatically bring the system up after repair. Since <application><ulink
568</ulink></application> provides the <command>reboot</command> command,
569automatic initialization is not needed.</para>
571<para>The <command>rdev -R /dev/floppy/0 0</command> will set the
572"root flags" to zero. They have no use in this application.</para>
574<para>The <command>dd</command> from above showed some results, like</para>
576<screen> 480+1 records in
577 480+1 records out</screen>
579<para>In this example the rescueimage (kernel) was 480+1 blocks in size.
580Make sure that this number, which may be different for you, matches your
581calculations from above. You need to calculate a "magic number" now
582that will be inserted into rescueimage. The value consists of three
583significant parts. Two are discussed here. The third is touched upon
586<para>Bits 0 - 10 will contain the size of rescueimage, in blocks,
587that you calculated above, and which should match the results from the
588dd above. Bit 14 (the 15th bit, which is 2 to the 14th power, or 16,384)
589is a flag that, when set to 1, tells the kernel an initial ramdisk is to
590be loaded. So for the single-floppy rescue diskette, the two numbers
59116,384 and 481 (or whatever number is right for your rescueimage size) are
592added together to produce a decimal value, like 16865. This value is
593inserted into the proper place in rescueimage by the
594<command>rdev</command> command done next.</para>
596<para>Insert the "magic number" into rescueimage and then write the
597root file system right after rescueimage on the floppy by executing the
598following commands, with the proper numbers inserted. Notice that the
599<command>seek</command> parameter's number must be the size, in blocks,
600of your rescueimage. If you use the static <filename class="directory">/dev</filename>
601setup, use <filename>/dev/fd0</filename> in the commands below, instead
602of <filename>/dev/floppy/0</filename>.</para>
604<screen><userinput><command>rdev -r /dev/floppy/0 <replaceable>16865</replaceable>
605dd if=/tmp/rootfs.gz of=/dev/floppy/0 bs=1k seek=<replaceable>481</replaceable></command></userinput></screen>
607<para>In this command, <command>seek</command> was used to position to
608the block following the end of the rescueimage (480+1) and begin writing the
609root file system to the floppy.</para>
614<title>A Two-diskette Rescue Setup</title>
616<para>If you just can't live with a single-diskette rescue system, here
617is what to do to make a simple two-diskette system. Note that the
618endless possibilities presented by the availability of
619<command>linuxrc</command> and other components are not addressed
620here. Here you will just use the kernel's ability to prompt for a second
621diskette that contains the initrd image and load it.</para>
623<para>Modify the above instructions as follows. First a different magic
624number is needed. The 15th bit (bit 14) still needs to be set, but the
625size of the rescueimage, in blocks, is replaced with a zero. The third
626component, which was not discussed above, is now used. This is the 16th
627bit (bit 15) of the "magic number". When set, it tells the kernel to ask
628the user to insert the "root" floppy. It then loads the initrd image
629from that diskette. Because the size of the rescueimage was replaced
630by zero, the kernel starts loading from the "zero'th" block (the first
631one) on the second diskette.</para>
633<para>The 16th bit (bit 15) represents 2 raised to the 15th power, or
63432,768. So the new magic number is 32,768 + 16384, which is 49,152. This
635value tells the kernel to prompt for, and then load, an initial ramdisk
636image from the first block on the inserted floppy. So your first
637modification is to the command to write the "magic number" to the rescueimage
638image on the diskette.</para>
640<screen><userinput><command>rdev -r /dev/floppy/0 <replaceable>49152</replaceable></command></userinput></screen>
642<para>Note that the initrd image is <emphasis>not</emphasis> copied to
643the diskette yet. Remove the boot diskette and insert another diskette
644that will hold your root file system. Run this modified command (don't
645forget to use <filename>/dev/fd0</filename> if you don't use devfs).
646Note that no <command>seek</command> parameter is used.</para>
648<screen><userinput><command>dd if=/tmp/rootfs.gz of=/dev/floppy/0 bs=1k</command></userinput></screen>
650<para>That's all there is to it. The possibilities from here are limited only
651by your imagination and tenacity in pursuing enhancements. And your
652willingness to research available documentation. A good starting point
653is the "Documentation" directory in your kernel source tree. More help
654may be gained at
655<ulink url=""><acronym>LFS </acronym> Hints</ulink>
656(please use a mirror site that is suitable) and
657<ulink url="">TLDP</ulink>.</para>
Note: See TracBrowser for help on using the repository browser.