Having used a number of different Linux distributions, I was never fully satisfied with either of those. I didn't like the way the bootscripts were arranged, or I didn't like the way certain programs were configued by default and more of those things. I came to realize that when I want to be totally satisfied with a Linux system, I have to build my own Linux system from scratch. Ideally only using the source code. No pre-compiled packages of any kind. No help from some sort of cdrom or bootdisk that would install some basic utilities. You would use your current Linux system and use that one to build your own.
This, at one time, wild idea seemed very difficult and at times almost impossible. The reason for most problems were due to my lack of knowledge about certain programs and procedures. After sorted out all kinds of dependency problems, compilation problems, etcetera, a manually Linux system was created and fully operational. I called this system and LFS system which stands for LinuxFromScratch.
The latest version of the document can always be found at http://huizen.dds.nl/~glb/
2.0-pre1 - Unknown date, sorry folks
This is not a full list of modified things. Because v2.0 is a major release, only the major changes are mentioned and not the minor ones.
There are two mailinglists you can subscribe to. The lfs-discuss and the lfs-announce list. The former is an open non-moderated list discussing anything that has got anything to do with this document. The latter is an open moderated list. Anybody can subscribe to it, but you cannot post messages to it (only the moderator(s) can do this). This list is primarily used for announcements of new versions of this document.
If you're subscribed to the lfs-discuss list you don't need to be subscribed to the lfs-announce list as well. Everything that is sent over the lfs-announce list is also sent over the lfs-discuss list.
To subscribe to a list, send an email to [email protected] and type in the body either subscribe lfs-discuss or subscribe lfs-announce
Majordomo will send you a confirmation-request email. This email will contain an authentication code. Once you send this email back to Majordomo (instructions are provided in that email) you will be subscribed.
To unsubscribe from a list, send an email to [email protected] and type in the the body either unsubscribe lfs-discuss or unsubscribe lfs-announce
Direct all your questions preferably to the mailinglist. If you need to reach me personally, send an email to [email protected]
Please read the following carefully: throughout this document you will
frequently see the variable name $LFS. $LFS must at all times
be replaced by the directory where the partition that contains the LFS system
is mounted. How to create and where to mount the partition will be explained
later on in full detail in chatper 4. In my case the LFS partition is mounted
on /mnt/hda5. If I read this document myself and I see $LFS somewhere,
I will pretend that I read /mnt/hda5. If I read that I have to run this command:
cp inittab $LFS/etc I actually will run this: cp inittab /mnt/hda5/etc
It's important that you do this no matter where you read it; be it in commands you enter on the prompt, or in some file you edit or create.
If you want, you can set the environment variable LFS. This way you can literally enter $LFS in stead of replacing it by something like /mnt/hda5. This is accomplished by running: export LFS=/mnt/hda5
If I read cp inittab $LFS/etc, I literally can type cp inittab $LFS/etc and the shell will replace this command by cp inittab /mnt/hda5/etc automatically.
Do not forget to set the LFS variable at all times. If you haven't set the variable and you use it in a command, $LFS will be ignored and whatever is left will be executed. The command cp inittab $LFS/etc without the LFS variable set, will result in copying the inittab file to the /etc directory which will overwrite your system's inittab. A file like inittab isn't that big a problem as it can easily be restored, but if you would make this mistake during the installation of the C Library, you can break your system badly and might have to reinstall it if you don't know how to repair it. So that's why I strongly advise against using the LFS variable. You better replace $LFS yourself by something like /mnt/hda5. If you make a typo while entering /mnt/hda5, the worst thing that can happen is that you'll get an error saying "no such file or directory" but it won't break your system. Don't say I didn't warn you ;)
Throughout this document I will assume that you have stored all the packages you have downloaded in a subdirectory under $LFS/usr/src.
I myself have use the convention of having a $LFS/usr/src/sources directory. Under sources you'll find the directory 0-9 and the directories a through z. A package as sysvinit-2.78.tar.gz is stored under $LFS/usr/src/sources/s/ A package as bash-3.02.tar.gz is stored under $LFS/usr/src/sources/b/ and so forth. You don't have to follow this convention of course, I was just giving an example. It's better to keep the packages out of $LFS/usr/src and move them to a subdirectory, so we'll have a clean $LFS/usr/src directory in which we will unpack the packages and work with them.
The next chapter contains the list of all the packages you need to download, but the partition that is going to contain our LFS system isn't created yet. Therefore store the files temporarily somewhere where you want and remember to copy them to $LFS/usr/src/<somesubdirectory> when you have finished chapter 4.
Before you can actually start doing something with a package, you need to unpack it first. Often you will find the package files being tar'ed and gzip'ed (you can see this from a .tar.gz or .tgz extension). I'm not going to write down every time how to ungzip and how to untar an archive. I will tell you how to that once, in this paragraph. There is also the possibility that you have the possibility of downloading a .tar.bz2 file. Such a file is tar'ed and compressed with the bzip2 program. Bzip2 achieves a better compression than the commonly used gzip does. In order to use bz2 archives you need to have the bzip2 program installed. Most if not every distribution comes with this program so chances are high it is already installed on your system. If not, install it using your distribution's installation tool.
tar xvfz filename.tar.gz; rm filename.tar.gz or tar xvfz filename.tgz; rm filename.tgztar --use-compress-prog=bzip2 -xvf filename.tar.bz2; rm filename.tar.bz2tar
xvf filename.tar; rm filename.tarNote that immediately after we have unpacked the archive, we delete the package file as we don't need it anymore. That's why you have to copy the file and not move it. If you move it and then delete it, you will need to re-download it when you need it again.
When the archive is unpacked a new directory will be created under the current directory (and this howto assumes that you unpack the archives under the $LFS/usr/src directory). You have to enter that new directory before you continue with the installation instructions. All the above will be sumarized as 'Unpack the xxx archive'. So, when you read it, you copy the package to $LFS/usr/src, you run the tar program to ungzip/unbzip and untar it, then you enter the directory that was created and then you read the next line of the installation instructions.
After you have installed a package you can do two things with it. You can
either delete the directroy that contains the sources or you can keep it. If
you decide to keep it, that's fine by me. But if you need the same package
again in a later chapter (all software upto chapter 7.2 will be re-installed
in chapter 7.3) you need to delete the directory first before using it again.
If you don't do this, you might end up in trouble because old settings will
be used (settings that apply to your normal Linux system but which don't apply
anymore when you have restarted your computer into the LFS system). Doing a
simple make clean does not always guarantee a totally clean source tree. The
configure script also has files lying around in various subdirectories which
are rearely removed by the make clean process.
Below is a list of all the software that you need to download for use in this document. I display the sites and directories where you can download the software, but it is up to you to make sure you download the source archive and the latest version. The version numbers correspondent to versions of the software that is known to work and which this HOWTO is going to be based on. If you experience problems which you can't solve yourself, download the version that is assumed in this HOWTO (in case you download a newer version).
Sysvinit (2.78) : ftp://ftp.cistron.nl/pub/people/miquels/sysvinit/
Bash (2.03) : ftp://ftp.gnu.org/gnu/bash/
Linux Kernel (2.2.14) : ftp://ftp.kernel.org/
Binutils (2.9.1) : ftp://ftp.gnu.org/gnu/binutils/
Bzip2 (0.9.5d) : http://www.muraroa.demon.co.uk
Diff Utils (2.7) : ftp://ftp.gnu.org/gnu/diffutils/
File Utils (4.0) : ftp://ftp.gnu.org/gnu/fileutils/
GCC (2.95.2) : ftp://ftp.gnu.org/gnu/gcc/
Glibc (2.1.3) : ftp://ftp.gnu.org/gnu/glibc/
Glibc-crypt (2.1.2) : ftp://ftp.gwdg.de/pub/linux/glibc/
Glibc-linuxthreads (2.1.3) : ftp://ftp.gnu.org/gnu/glibc/
Grep (2.4) : ftp://ftp.gnu.org/gnu/grep/
Gzip (1.2.4) : ftp://ftp.gnu.org/gnu/gzip/
Make (3.78.1) : ftp://ftp.gnu.org/gnu/make/
Sed (3.02) : ftp://ftp.gnu.org/gnu/sed/
Shell Utils (2.0) : ftp://ftp.gnu.org/gnu/sh-utils/
Tar (1.13) : ftp://ftp.gnu.org/gnu/tar/
Text Utils (2.0) : ftp://ftp.gnu.org/gnu/textutils/
Util Linux (2.10f) : ftp://ftp.win.tue.nl/pub/linux/utils/util-linux/
Bison (1.28) : ftp://ftp.gnu.org/gnu/bison/
Mawk (1.3.3) : ftp://ftp.whidbey.net/pub/brennan/
Find Utils (4.1) : ftp://ftp.gnu.org/gnu/findutils/
Ncurses (5.0) : ftp://ftp.gnu.org/gnu/ncurses/
Less (340) : ftp://ftp.gnu.org/gnu/less/
Perl (5.005_03) : ftp://ftp.gnu.org/gnu/perl/
M4 (1.4) : ftp://ftp.gnu.org/gnu/m4/
Texinfo (4.0) : ftp://ftp.gnu.org/gnu/texinfo/
Autoconf (2.13) : ftp://ftp.gnu.org/gnu/autoconf/
Automake (1.4) : ftp://ftp.gnu.org/gnu/automake/
Flex (2.5.4a) : ftp://ftp.gnu.org/gnu/flex/
E2fsprogs (1.18) : ftp://tsx-11.mit.edu/pub/linux/packages/ext2fs/
File (3.26) : http://tts.ookhoi.dds.nl/lfs-howto/download/file-3.26-lfs.tar.gz
Groff (1.15) : ftp://ftp.gnu.org/gnu/groff/
Ld.so (1.9.9) : ftp://tsx-11.mit.edu/pub/linux/packages/GCC/
Libtool (1.3.4) : ftp://ftp.gnu.org/gnu/libtool/
Linux86 (0.14.3) : http://tts.ookhoi.dds.nl/lfs-howto/download/linux86-0.14.3-lfs.tar.gz
Lilo (21) : ftp://sunsite.unc.edu/pub/Linux/system/boot/lilo/
Shadow Password Suite (19990827) : ftp://piast.t19.ds.pwr.wroc.pl/pub/linux/shadow/
Man-db (2.3.10) : http://tts.ookhoi.dds.nl/lfs-howto/download/man-db-2.3.13-lfs.tar.gz
Modutils (2.3.9) : ftp://ftp.ocs.com.au/pub/modutils/
Termcap (1.3) : ftp://ftp.gnu.org/gnu/termcap/
Procinfo (17) : ftp://ftp.cistron.nl/pub/people/svm/
Procps (2.0.6) : ftp://people.redhat.com/johnsonm/procps/
Psmisc (19) : ftp://lrcftp.epfl.ch/pub/linux/local/psmisc/
Start-stop-daemon (0.4.1) : http://tts.ookhoi.dds.nl/lfs-howto/download/ssd-0.4.1-lfs.tar.gz
Sysklogd (1.3.31) : ftp://sunsite.unc.edu/pub/Linux/system/daemons/
Vim (5.6) : ftp://ftp.vim.org/pub/editors/vim/
All software below is used in sections 13 and above and are not strictly necessary. You have to determine for yourself if you want to install certain packages. If, for example, you don't intend to go online with the LFS system, you might not want to install the email, telnet, ftp, www, etc. utilities.
Netkit-base (0.17) : ftp://ftp.uk.linux.org/pub/linux/Networking/netkit-devel/
Net-tools (1.54) : http://www.tazenda.demon.co.uk/phil/net-tools/
Procmail (3.14) : ftp://ftp.procmail.org/pub/procmail/
Sendmail (8.9.3) : ftp://ftp.sendmail.org/pub/sendmail/
Mailx (8.1.1) : http://tts.ookhoi.dds.nl/lfs-howto/download/mailx-8.1.1-lfs.tar.gz
Mutt (1.0i) : ftp://ftp.mutt.org/pub/mutt/
Fetchmail (5.2.0) : http://www.tuxedo.org/~esr/fetchmail/
Netkit-telnet (0.17) : ftp://ftp.uk.linux.org/pub/linux/Networking/netkit-devel/
Proftpd (1.2.0pre9) : ftp://ftp.tos.net/pub/proftpd/
Netkit-ftp (0.17) : ftp://ftp.uk.linux.org/pub/linux/Networking/netkit-devel/
Apache (1.3.11) : http://www.apache.org/dist/
Zlib Library (1.1.3) : http://www.cdrom.com/pub/infozip/zlib/
Lynx (2.8.2) : http://www.slcc.edu/lynx/release/
PPP (2.3.11) : ftp://cs.anu.edu.au/pub/software/ppp/
Xfree86 (3.3.5) : ftp://ftp.xfree86.org/pub/XFree86/
libPropList (0.9.1) : ftp://ftp.windowmaker.org/pub/libs/
libXpm (4.7) : ftp://sunsite.unc.edu/pub/Linux/libs/X/
libpng (1.0.3) : http://www.cdrom.com/pub/png/
libtiff (3.4) : ftp://ftp.sgi.com/graphics/tiff/
libjpeg (6b) : http://www.ijg.org/
libungif (4.1.0) : ftp://prtr-13.ucsc.edu/pub/libungif/
WindowMaker (0.61.1) : ftp://ftp.windowmaker.org/pub/release/
We are going to build the LFS system using an already installed Linux distribution such as Debian, SuSe, Slackware, Mandrake, RedHat, etc. You don't need to have any kind of bootdisk. We will use an existing Linux system as the base (since we need a compiler, linker, text editor and other tools).
If you don't have Linux installed yet, you won't be able to put this HOWTO to use right away. I suggest you first install a Linux distribution. It really doesn't matter which one you install. It also doesn't need to be the latest version, though it shouldn't be a too old one. If it is about a year old or newer it'll do just fine. You will safe yourself a lot of trouble if your normal system uses glibc-2.0 or newer. Libc5 can cause some problems and is not supported in this document as I don't have access to such a machine anymore.
Before we can build our new Linux system, we need to have an empty Linux partition on which we can build our new system. I recommend a partition size of at least 5 00 MB. You can get away with around 250MB for a bare system with no extra whistles and bells (such as software for emailing, networking, Internet, X Window System and such). If you already have a Linux Native partition available, you can skip this subsection.
Start the fdisk program (or some other fdisk program if you prefer) with
the appropriate hard disk as the option (like /dev/hda if you want to create
a new partition on the primary master IDE disk). Create a Linux Native partition,
write the partition table and exit the fdisk program. If you get the message
that you need to reboot your system to ensure that that partition table is
updated, then please reboot your system now before continuing. Remember what
your new parition's designation is. It could be something like hda5 (as it
is in my case). This newly created partition will be refered to as the LFS
partition in this document.
Once the partition is created, we have to create a new ext2 file system
on that partition. To create a new ext2 file system we use the mke2fs command.
Enter the new partition as the only option and the file system will be created.
If your partition was hda5, you would run the command as mke2fs /dev/hda5
Once we have created the ext2 file system, it is ready for use. All we have to do to be able to access it (as in reading from and writing date to it) is mounting it. If you mount it under /mnt/hda5, you can access this partition by going to the /mnt/hda5 directory and then do whatever you need to do. This document will assume that you have mounted the partition on a subdirectory under /mnt. It doesn't matter which subdirectory you choose (or you can use just the /mnt directory as the mounting point), but a good practise is to create a directory with the same name as the partition's designation. In my case the LFS partition is called hda5 and therefore I mount it on /mnt/hda5
This directory (/mnt/xxx) is the $LFS you have read about earlier. So if you read somewhere to "cp inittab $LFS/etc" you actually will type "cp inittab /mnt/xxx/etc" where xxx is replaced by your partition's designation.
Let's create the directory tree on the LFS partition according to the FHS standard which can be found at http://www.pathname.com/fhs/. Issuing the following commands will create the necessary directories.
cd $LFS mkdir bin boot dev etc home lib mnt proc root sbin tmp usr var cd $LFS/usr mkdir bin include lib sbin share src ln -s share/man man ln -s share/doc doc ln -s . local ln -s ../etc etc ln -s ../var var cd $LFS/usr/share mkdir dict doc info locale man nls misc terminfo zoneinfo cd $LFS/usr/share/man mkdir man1 man2 man3 man4 man5 man6 man7 man8 cd $LFS/var mkdir lock log run spool tmp
Now that the directories are created, copy the source files you have downloaded in chapter 3 to some subdirectory under $LFS/usr/src (you will need to create this subdirectory yourself).
We can create every single file that we need to be in the $LFS/dev directory using the mknod command, but that just takes up a lot of time. I choose to just simply copy the current /dev directory to the $LFS partition. Use this command to copy the entire directory while preserving original rights, symlinks and ownerships:
cp -av /dev $LFS
Under normal circumstances, after the kernel is done loading and initializing
various system components, it attempts to load a program called init which
will finalize the system boot process. The package found on most if not every
single Linux system is called Sysvinit and that's the program we're going to
install on our LFS system.
Makefile fileAfter applying the $(ROOT) parts to the last four lines, they should look like this:
@if [! -p $(ROOT)/dev/initctl ]; then \ echo "Creating $(ROOT)/dev/initctl"; \ rm -f $(ROOT)/dev/initctl; \ mknod -m 600 $(ROOT)/dev/initctl p; fi
make -e LDFLAGS=-static; make install
In order for Sysvinit to work, we need to create it's configuration file.
Create the $LFS/etc/inittab file containing the following:
# Begin /etc/inittab id:2:initdefault: ca:12345:ctrlaltdel:/sbin/shutdown -t1 -a -r now 1:2345:respawn:/sbin/sulogin # End /etc/inittab
As you can see from the inittab file, when we boot the system, init will start the sulogin program and sulogin will ask you for user root's password. This means we need to have at least a passwd file present on the LFS system.
The encoded password string above is: lfs123
When you logon to your LFS system, enter lfs123 when asked to enter user root's password.
When sulogin asks you for the root password and you've entered the password, a shell needs to be started. Usually this is the bash shell. Since there are no libraries installed yet, we need to link bash statically, just like we did with Sysvinit.
./configure --enable-static-link make; make -e prefix=$LFS/usr install mv $LFS/usr/bin/bash $LFS/bin cd $LFS/bin; ln -s bash sh
In order to being able to boot from this partition, we need to update our
/etc/lilo.conf file. Add the following lines to lilo.conf:
image=<currently used image> label=<label> root=$LFS read-only
Replace <currently used image> by the kernel image file that you are using to boot your normal Linux system. <label> can be anything you want it to be. I named the label "lfs" What you enter as <label> is what you enter at the LILO-prompt when you choose with system to boot.
Now run the lilo program to update the boot loader.
After you've completed this section, we can test the system by rebooting into LFS and see if we can log on to it. When you reboot and are at the LILO prompt, enter the label you have entered in the lilo.conf file to start the LFS system. Then enter root's password and you should be on the bash-prompt now. You won't be able to shutdown the system with a program like shutdown. Although the program is present, it will give you the following error: "You don't exist. Go away." when you try to use the program. The meaning of this error is that the system isn't able to locate the password file. Although the shutdown program is statically linked against the libraries it needs, it still depends on the NSS Library (Name Server Switch) which is part of the GNU C Library, which will be installed in a later chapter. This NSS library passes on information where (in this case) the passwd file can be found.
For now you can reboot the system using the reboot -f command. This will
bypass shutting down the system using the shutdown program and reboot immediately.
Since the file system is mounted read-only this will not harm our system in
any way (though you might get a warning next time you try to mount the system
that it wasn't unmounted cleanly the last time and that you should run e2fsck
to make sure the file system is still intact).
In section 2 above I mentioned you can download a new kernel from ftp://ftp.kernel.org/ However, this site is often too busy to get through and the maintainers of this site encourage you to download the kernel from a location near you. You can access a mirror site by going to ftp://ftp.<country code>.kernel.org/ (like ftp.ca.kernel.org).
/usr/src/ directory (this will create
a new /usr/src/linux directory) make depmake bzImagearch/<cpu>/boot/bzImage file to the /boot directory (or
some place else if your Linux system uses a different convention where kernel
images and the like are stored)/boot/bzImage file to something like /boot/lfskernelcp
-av /usr/src/linux $LFS/usr/src
cd $LFS/usr/include ln -s ../src/linux/include/linux linux ln -s ../src/linux/include/asm asm
/etc/lilo.conf file and go to the LFS sectionlfskernel (or whatever you've named the originally
called bzImage file)lilo to update the boot loader.Reboot your system and start your LFS system. Verify that the newly installed kernel doesn't perform out-of-the-ordinary actions (such as crashing).
In this chapter we will install all the software that belongs to a basic Linux system. After you're done with this chapter you have a fully working Linux system. The remaining chapters deals with optional issues such as setting up networking, internet servers + clients (telnet, ftp, http, email), setting up Internet itself and the X Window System. You can skip chapters at your own discretion. If you don't plan on going online with the LFS system there's little use to setup Internet for example.
There are a number of packages that need to be already installed before we can start installing all the basic system software. A typical configure scripts needs programs like rm, grep, sed, mv, cat, cp, diff. You need to be able to ungzip and untar archives, you need to link programs after you have compiled the objects files. All these (and a few more) programs needs to be available before we can install anything else. These programs are going to be linked statically. The reasoning behind this is that your normal Linux system may have a different C Library version than the LFS system is going to have. The programs you install in this section will be linked against the C Library of your normal Linux system. This may cause library conflicts if you run those programs on the LFS system. Therefore we have to link those programs statically. During the installation of the basic system software set, we will re-install the statically linked software so that they are linked dynamically against the C library on the LFS system.
Every program and library is default compiled with debugging symbols. This
means you can run a program or library through a debugger and the degugger's
output will be more user friendly. These debugging symbols also enlarge the
program or binary significantly. This HOWTO will not install software without
debugging symbols (as I don't know if the majority of readers do or don't debug
software). In stead, you can remove those symbols manually if you want with
the strip program.
To remove debugging symbols from a binary (must be an a.out or ELF binary)
run strip --strip-debug filename You can use wildcards if you need to strip
debugging symbols from multiple files (use something like strip --strip-debug
$LFS/usr/bin/*).
Before you wonder if these debugging symbols would make a big difference, here are some statistics:
Sizes may vary depending on which compiler has been used and which C library version is used to link dynamic programs against, but your results will be very similar if you compare programs with and without debugging symbols. After I was done with this chapter and stripped all debugging symbols from all LFS binaries and libraries I regained a little over 102 MB of diskspace. Quite the difference. The difference would be even greater when I would do this at the end of this HOWTO when everything is installed.
./configure make -e LDFLAGS=-all-static make -e prefix=$LFS/usr install cd $LFS/usr/bin; mv bunzip2 bzip2 $LFS/bin
make -e LDFLAGS=-static make -e PREFIX=$LFS/usr install cd $LFS/usr/bin mv $LFS/usr/bin/*zip2* $LFS/bin
./configure make -e LDFLAGS=-static make -e prefix=$LFS/usr install
This package is known to cause static linking problems on certain platforms. If you're having trouble compiling this package as well, you can download a fixed package from http:/tts.ookhoi.dds.nl/lfs-howto/download/diffutils-2.7-fixed.tar.gz
./configure --disable-nls make -e LDFLAGS=-static make -e prefix=$LFS/usr install cd $LFS/usr/bin mv chgrp chmod chown cp dd df ln ls mkdir mknod mv rm rmdir sync $LFS/bin
In order to compile Glibc-2.1.3 you need to have gcc-2.95.2 installed.
Any version from 2.8 and up would do, but 2.95.2 is recommended. Many glibc-2.0
based systems have gcc-2.7.2.3 installed and you can't compile glibc-2.1 with
that compiler. Therefore we will install gcc-2.95.2. also on the normal system,
but without overwriting the existing compiler. Before you install gcc on your
normal system, make sure wether you need it or not. Run gcc --version and check
if the version number it reports equals or is higher than 2.8. If not, you
need to install gcc-2.95.2. If you experience difficulties compiling glibc
later on, you might want to install gcc-2.95.2 anyways.
mkdir $LFS/usr/src/gcc-build; cd $LFS/usr/src/gcc-build ../gcc-2.95.2/configure --prefix=/usr/gcc2952 --with-local-prefix=/usr/gcc2952 --with-gxx-include-dir=/usr/gcc2952/include/g++ --enable-shared --enable-languages=c,c++ make bootstrap; make install
mkdir $LFS/usr/src/gcc-build;cd $LFS/usr/src/gcc-build ../gcc-2.95.2/configure --with-gxx-include-dir=/usr/include/g++ --enable-shared --enable-languages=c,c++ --disable-nls make -e LDFLAGS=-static bootstrap make -e prefix=$LFS/usr local_prefix=$LFS/usr gxx_include_dir=$LFS/usr/include/g++ install
The system needs a few symlinks to ensure every program is able to find the compiler and the pre-processor. Some programs run the 'cc' program, others run the 'gcc' program, some programs expect the cpp program to be in /lib (which is /usr/lib on the LFS system) and others expect to find it in /usr/bin.
cd $LFS/usr/lib; ln -s gcc-lib/<host>/2.95.2/cpp cpp cd $LFS/usr/bin; ln -s gcc cc
Replace <host> with the directory where the gcc-2.95.2 files were installed (i686-unknown-linux in my case). You will most likely find two different directories.
A note on the glibc-crypt package:
-*-*-*-*-*- The add-on is not included in the main distribution of the GNU C library because some governments, mostly notable those of France, Russia and the US, have very restrictive rules governing the distribution and use of encryption software. Please read the node "Legal Problems" in the manual for more details. In particular, the US does not allow export of this software without a license, including via the Internet. So please do not download it from the main FSF FTP site at ftp.gnu.org if you are outside of the US. This software was completely developed outside the US. -*-*-*-*-*-
"This software" refers to the glibc-crypt package at ftp://ftp.gwdg.de/pub/linux/glibc/. This law only affects people who don't live in the US. It's not prohibited to import DES software, so if you live in the US you can import it from that German site.
mkdir $LFS/usr/src/glibc-build;cd $LFS/usr/src/glibc-build ../glibc-2.1.3/configure --enable-add-ons make; make install_root=$LFS install cd $LFS/usr/lib; mv *so* $LFS/lib
mkdir $LFS/usr/src/glibc-build; cd $LFS/usr/src/glibc-build CC=/usr/gcc2952/bin/gcc ../glibc-2.1.3/configure --enable-add-ons make; make install_root=$LFS install cd $LFS/usr/lib; mv *so* $LFS/lib cd $LFS/lib; mv libstdc++* $LFS/usr/lib
If your normal Linux system runs libc-2.0.x, you need to copy the NSS library
files to the LFS partition. Certain statically linked programs still depend
on the NSS library, especially programs that need to lookup usernames, userid's
and groupid's. You can check which C Library version your normal Linux system
uses by running: ls -l libc.so.*
Your system uses glibc-2.0 if the output looks like: /lib/libc.so.6 -> libc-2.0.7.so
Your system uses glibc-2.1 is the output looks like: /lib/libc.so.6 -> libc-2.1.2.so
If your have a libc-2.0.x.so file (where x is the micro version number
such as 7) copy the NSS Library files by running: cp -av /lib/*nss* $LFS/lib
./configure --disable-nls make -e LDFLAGS=-static make -e prefix=$LFS/usr install
This package is known to cause static linking problems on certain platforms. If you're having trouble compiling this package as well, you can download a fixed package from http:/tts.ookhoi.dds.nl/lfs-howto/download/grep-2.4-fixed.tar.gz
./configure make -e LDFLAGS=-static make -e prefix=$LFS/usr install cd $LFS/usr/bin mv gunzip gzip $LFS/bin
This package is known to cause compilation problems on certain platforms. If you're having trouble compiling this package as well, you can download a fixed package from http:/tts.ookhoi.dds.nl/lfs-howto/download/gzip-1.2.4-fixed.tar.gz
./configure make -e LDFLAGS=-static make -e prefix=$LFS/usr install
./configure make -e LDFLAGS=-static make -e prefix=$LFS/usr install cd $LFS/usr/bin; mv sed $LFS/bin
This package is known to cause static linking problems on certain platforms. If you're having trouble compiling this package as well, you can download a fixed package from http:/tts.ookhoi.dds.nl/lfs-howto/download/sed-3.02-fixed.tar.gz
./configure --disable-nls make -e LDFLAGS=-static make -e prefix=$LFS/usr install cd $LFS/usr/bin mv date echo false pwd stty su true uname hostname $LFS/bin
./configure --disable-nls make -e LDFLAGS=-static make -e prefix=$LFS/usr install cd $LFS/usr/bin; mv tar $LFS/bin
./configure --disable-nls make -e LDFLAGS=-static make -e prefix=$LFS/usr install mv $LFS/usr/bin/cat $LFS/bin
./configure cd lib;make cd ../mount;make -e LDFLAGS=-static mount umount cp mount umount $LFS/bin
Before we install the rest of the basic system software, you first have
to reboot into the LFS system before continuing. Once you're rebooted and logged
in, remount the root partition in read-write mode by running: /sbin/mount -n
-o remount,rw / /
The installation of all the software is pretty straightforward and you'll think it's so much easier and shorter to give the generic installation instructions for each package and only explain how to install something if a certain package requires an alternate installation method. Although I agree with you on this aspect, I, however, choose to give the full instructions for each and every package. This is simply to avoid any possible confusion and errors.
./configure make; make install
./configure make; make install cd /usr/bin; ln -s mawk awk
./configure make; make install
This package is known to cause compilation problems. If you're having trouble compiling this package as well, you can download a fixed package from http:/tts.ookhoi.dds.nl/lfs-howto/download/findutils-4.1-fixed.tar.gz
Unpack the Termcap archive and install it by running:
./configure make; make install
./configure --with-shared make; make install
./configure make; make install mv /usr/bin/less /bin
./Configure make; make install
Note that we skip the 'make test' step. This is because at this moment the system isn't ready yet for running the perl test. At this time we'll trust that perl compiled fine.
./configure make; make install
./configure make; make install
./configure make; make install
./configure make install
./configure make; make install mv /usr/bin/bash /bin
./configure make; make install
./configure make; make install
make; make install cd /usr/bin; mv bunzip2 bzip2 /bin
./configure make; make install
./configure make; make install mv /usr/sbin/mklost+found /sbin
./configure make; make install
./configure make; make install cd /usr/bin mv chgrp chmod chown cp dd df ln ls mkdir mknod mv rm rmdir sync /bin
mkdir $LFS/usr/src/gcc-build;cd $LFS/usr/src/gcc-build ../gcc-2.95.2/configure --with-gxx-include-dir=/usr/include/g++ \ --enable-shared --enable-languages=c,c++ make bootstrap; make install
./configure make; make install
./configure make; make install
./configure make; make install cd /usr/bin; mv z* gunzip gzip /bin
cd util; make ldd ldconfig cp ldd /bin; cp ldconfig /sbin rm /usr/bin/ldd
./configure make; make install
cd as make; make install cd ../ld make; make install
make; make install
./configure make; make install
./configure make; make install cd /usr/bin mv date echo false pwd stty su true uname hostname /bin
./configure make; make install cd etc cp limits login.access login.defs.linux shells suauth /etc mv /etc/login.defs.linux /etc/login.defs cd /usr/sbin mv chpasswd dpasswd groupadd groupdel groupmod logoutd mkpasswd \ newusers useradd userdel usermod grpck pwck vipw grpconv grpunconv \ pwconv pwunconv /sbin
groupadd -g 6 man useradd -u 6 -g man man ./configure make all; make install
./configure make; make install
make; make install
gcc -O3 -Wall -Wno-unused -c watch.c make; make -e XSCPT="" install mv /usr/bin/kill /bin
make; make install
./configure make; make install mv /usr/bin/sed /bin
make start-stop-daemon cp start-stop-daemon /sbin cp start-stop-daemon.8 /usr/share/man/man8
make; make install
cd src make; make install
./configure make; make install mv /usr/bin/tar /bin
./configure make; make install mv /usr/bin/cat /bin
./configure make; make install
HAVE_PASSWD=yes HAVE_SLN=yes HAVE_TSORT=yes
groupadd -g 5 tty ./configure make; make install
If you have copied the NSS Library files from your normal Linux system to the LFS system (because your normal system runs glibc-2.0) it's time to remove them now by running:
rm /lib/libnss*.so.1 /lib/libnss*2.0*
Now that all software is installed, all that we need to do to get a few programs running properly is to create their configuration files.
We /need to create the /etc/nsswitch.conf file. Although glibc should provide defaults when this file is missing or corrupt, it's defaults don't work work well with networking which will be dealt with in a later chapter. Also, our timezone needs to be setup.
# Begin /etc/nsswitch.conf passwd: files group: files shadow: files hosts: files dns networks: files protocols: db files services: db files ethers: db files rpc: db files netgroup: db files # End /etc/nsswitch.conf
tzselect's output can be something like "EST5EDT" or "Canada/Eastern". The symlink you would create with that information would be ln -s /usr/share/zoneinfo/EST5EDT /etc/localtime or ln -s /usr/share/zoneinfo/Canada/Eastern /etc/localtime
We're not going to create lilo's configuration file from scratch, but we'll use the file from your normal Linux system. This file is differnet on every machine and thus I can't create it here. Since you would want to have the same options regarding lilo as you have when you're using your normal Linux system you would create the file exactly as it is on the normal system.
cp /mnt/original/etc/lilo.conf /etc cp /mnt/original/boot/* /boot
If your normal Linux system does not have (all of) it's kernel images in /mnt/original/boot, then check your /etc/lilo.conf file for the location of those files and copy those as well to the location where /etc/lilo.conf expects them to be. Or you can copy them to /boot regardless and modify the /etc/lilo.conf file so it contains the new paths for the images as you have them on the LFS system. Either way works fine, it's up to you how you want to do it.
# Begin /etc/syslog.conf auth,authpriv.* -/var/log/auth.log *.*;auth,authpriv.none -/var/log/sys.log daemon.* -/var/log/daemon.log kern.* -/var/log/kern.log mail.* -/var/log/mail.log user.* -/var/log/user.log *.emerg * # End /etc/syslog.conf
This package contains the utilities to modify user's passwords, add new users/groups, delete users/groups and more. I'm not going to explain to you what 'password shadowing' means. You can read all about that in the doc/HOWTO file. There's one thing you should keep in mind, if you decide to use shadow support, that programs that need to verify passwords (examples are xdm, ftp daemons, pop3d, etc) need to be 'shadow-compliant', eg. they need to be able to work with shadowed passwords.
If you decide you don't want to use shadowed passwords (after you're read the doc/HOWTO document), you still use this archive since the utilities in this archive are also used on system which have shadowed passwords disabled. You can read all about this in the HOWTO. Also note that you can switch between shadow and non-shadow at any point you want.
Now is a very good moment to read section #5 of the doc/HOWTO file. You can read how you can test if shadowing works and if not, how to disable it. If it doesn't work and you haven't tested it, you'll end up with an unusable system after you logout of all your consoles, since you won't be able to login anymore. You can easily fix this by passing the init=/sbin/sulogin parameter to the kernel, unpack the util-linux archive, go to the login-utils directory, build the login program and replace the /bin/login by the one in the util-linux package. Things are never hopelessly messed up (at least not under Linux), but you can avoid a hassle by testing properly and reading manuals ;)
After you have made the following modification to the /etc/inittab file, you will be able to logon to it as you are used to (using the agetty and login programs). Sulogin won't be used anymore for normal logins.
# Begin /etc/inittab id:2:initdefault: si::sysinit:/etc/init.d/rcS su:S:wait:/sbin/sulogin l0:0:wait:/etc/init.d/rc 0 l1:1:wait:/etc/init.d/rc 1 l2:2:wait:/etc/init.d/rc 2 l3:3:wait:/etc/init.d/rc 3 l4:4:wait:/etc/init.d/rc 4 l5:5:wait:/etc/init.d/rc 5 l6:6:wait:/etc/init.d/rc 6 ft:6:respawn:/sbin/sulogin ca:12345:ctrlaltdel:/sbin/shutdown -t1 -a -r now 1:2345:respawn:/sbin/agetty /dev/tty1 9600 2:2345:respawn:/sbin/agetty /dev/tty2 9600 3:2345:resapwn:/sbin/agetty /dev/tty3 9600 4:2345:respawn:/sbin/agetty /dev/tty4 9600 5:2345:respawn:/sbin/agetty /dev/tty5 9600 6:2345:respawn:/sbin/agetty /dev/tty6 9600 # End /etc/inittab
Programs like login, shutdown and others want to write to the /var/run/utmp file. This file contains information about who is currently logged in. It also contains information on when the computer was last shutdown.
These bootscripts are started at system boot time. The scripts are responsible for mounting the root file system in read-write mode, activating swap, setting up some system settings and starting the various daemons that our system needs.
You need the Sysvinit package again for this section.
cd /etc mkdir rc0.d rc1.d rc2.d rc3.d rc4.d rc5.d rc6.d init.d rcS.d
debian/etc/init.d/rc file to: /etc/init.drcS containing the following:
#!/bin/sh
# Begin /etc/init.d/rcS
runlevel=S
prevlevel=N
umask 022
export runlevel prevlevel
trap ":" INT QUIT TSTP
for i in /etc/rcS.d/S??*
do
[ ! -f "$i" ] && continue;
$i start
done
# End /etc/init.d/rcS
reboot containing the following:
#!/bin/sh # Begin /etc/init.d/reboot echo -n "System reboot in progress..." /sbin/reboot -d -f -i # End /etc/init.d/reboot
halt containing the following:
#!/bin/sh # Begin /etc/init.d/halt /sbin/halt -d -f -i -p # End /etc/init.d/halt
mountfs containing the following:
#!/bin/sh
# Begin /etc/init.d/mountfs
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
echo -n "Remounting root file system in read-write mode..."
/sbin/mount -n -o remount,rw /
check_status
> /etc/mtab
/sbin/mount -f -o remount,rw /
echo -n "Mounting proc file system..."
/sbin/mount proc
check_status
# End /etc/init.d/mountfs
umountfs containing the following:
#!/bin/sh
# Begin /etc/init.d/umountfs
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
echo "Deactivating swap..."
/sbin/swapoff -av
check_status
echo -n "Unmounting file systems..."
/sbin/umount -a -r
check_status
# End /etc/init.d/umountfs
sendsignals containing the following:
#!/bin/sh
# Begin /etc/init.d/sendsignals
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
echo -n "Sending all processes the TERM signal..."
/sbin/killall5 -15
check_status
echo -n "Sending all processes the KILL signal..."
/sbin/killall5 -9
check_status
/etc/init.d/checkroot containing the following:
#!/bin/sh
# Begin /etc/init.d/checkroot
echo "Activating swap..."
/sbin/swapon -av
if [ -f /fastboot ]
then
echo "Fast boot, no file system check"
else
mount -n -o remount,ro /
if [ $? = 0 ]
then
if [ -f /forcecheck ]
then
force="-f"
else
force=""
fi
echo "Checking root file system..."
fsck $force -a /
if [ $? -gt 1 ]
then
echo
echo "fsck failed. Please repair your file system manually by"
echo "running fsck without the -a option"
echo "Please note that the file system is currently mounted in"
echo "read-only mode."
echo
echo "I will start sulogin now. CTRL+D will reboot your system."
/sbin/sulogin
/reboot -f
fi
else
echo "Cannot check root file system because it is not mounted in"
echo "read-only mode."
fi
fi
# End /etc/init.d/checkroot
/etc/init.d/sysklogd containing the following:
#!/bin/sh
# Begin /etc/init.d/sysklogd
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
case "$1" in
start)
echo -n "Starting system log daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/syslogd -- -m 0
check_status
echo -n "Starting kernel log daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/klogd
check_status
;;
stop)
echo -n "Stopping kernel log daemon..."
start-stop-daemon -K -q -o -p /var/run/klogd.pid
check_status
echo -n "Stopping system log daemon..."
start-stop-daemon -K -q -o -p /var/run/syslogd.pid
check_status
;;
reload)
echo -n "Reloading system load daemon configuration file..."
start-stop-daemon -K -q -o -s 1 -p /var/run/syslogd.pid
check_status
;;
restart)
echo -n "Stopping kernel log daemon..."
start-stop-daemon -K -q -o -p /var/run/klogd.pid
check_status
echo -n "Stopping system log daemon..."
start-stop-daemon -K -q -o -p /var/run/syslogd.pid
check_status
sleep 1
echo -n "Starting system log daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/syslogd -- -m 0
check_status
echo -n "Starting kernel log daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/klogd
check_status
;;
*)
echo "Usage: $0 {start|stop|reload|restart}"
exit 1
;;
esac
# End /etc/init.d/sysklogd
:
chmod 755 rcS reboot halt mountfs umountfs sendsignals checkroot sysklogd cd ../rc0.d ln -s ../init.d/sysklogd K90sysklogd ln -s ../init.d/sendsignals S80sendsignals ln -s ../init.d/umountfs S90umountfs ln -s ../init.d/halt S99halt cd ../rc6.d ln -s ../init.d/sysklogd K90sysklogd ln -s ../init.d/sendsignals S80sendsignals ln -s ../init.d/umountfs S90umountfs ln -s ../init.d/reboot S99reboot cd ../rcS.d ln -s ../init.d/checkroot S05checkroot ln -s ../init.d/mountfs S10mountfs cd /etc/rc2.d ln -s ../init.d/sysklogd S03sysklogd
/etc/fstab containing the following:
/dev/<LFS-partition designation> / ext2 defaults 0 1 /dev/<swap-partition designation> none swap sw 0 0 proc /proc proc defaults 0 0
./configure make; make install cd etc.sample; cp services protocols /etc mv /usr/bin/ping /bin
make; make install mv /usr/bin/netstat /bin cd /usr/sbin; mv ifconfig route /sbin
/etc/init.d/localnet containing the following:
#!/bin/sh
# Begin /etc/init.d/localnet
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
echo -n "Setting up loopback device..."
/sbin/ifconfig lo 127.0.0.1
check_status
echo -n "Setting up hostname..."
/bin/hostname --file /etc/hostname
check_status
# End /etc/init.d/localnet
chmod 755 /etc/init.d/localnetcd /etc/rcS.d; ln -s ../init.d/localnet
S03localnetCreate a new file /etc/hostname and put the hostname in it. This is not
the FQDN (Fully Qualified Domain Name). This is the name you wish to call your
computer in a network.
If you want to configure a network card, you have to decide on the IP-address, FQDN and possible aliases for use in the /etc/hosts file. An example is:
<myip> myhost.mydomain.org somealiases
Make sure the IP-address is in the private network IP-address range. Valid ranges are:
Class Networks A 10.0.0.0 B 172.16.0.0 through 172.31.0.0 C 192.168.0.0 through 192.168.255.0
A valid IP address could be 192.168.1.1. A valid FQDN for this IP could be me.linuxfromscratch.org
If you're not going to use a network card, you still need to come up with a FQDN. This is necessary for programs like Sendmail to operate correctly (in fact; Sendmail won't run when it can't determine the FQDN).
Here's the /etc/hosts file if you don't configure a network card:
# Begin /etc/hosts (no network card version) 127.0.0.1 me.lfs.org <contents of /etc/hostname> localhost # End /etc/hosts (no network card version)
Here's the /etc/hosts file if you do configure a network card:
# Begin /etc/hosts (network card version) 127.0.0.1 localhost 192.168.1.1 me.lfs.org <contents of /etc/hostname> # End /etc/hosts (network card version)
Of course, change the 192.168.1.1 and me.lfs.org to your own liking (or requirements if you are assigned an IP-address by a network/system administrator and you plan on connecting this machine to that network).
This sub section only applies if you are going to configure a network card. If not, skip this sub section and read on.
Create a new file /etc/init.d/ethnet containing the following:
#!/bin/sh
# Begin /etc/init.d/ethnet
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
/sbin/ifconfig eth0 <ipaddress>
check_status
# End /etc/init.d/ethnet
chmod 755 ethnetcd ../rc2.d; ln -s ../init.d/ethnet
S10ethnet
/etc/init.d/localnet/etc/init.d/ethnet
ping <your FQDN> ping <what you choose for hostname> ping localhost ping 127.0.0.1 ping 192.168.1.1 (only when you configured your network card)
All these five ping command's should work without failures. If so, the basic network is working.
For a far more detailed guide on how to set up Internet, I refer to Kirch's Linux Network Administrator's Guide which is published by O'Reilly and Asociates. The LDP also has a copy of this guide available for download at http://www.linuxdoc.org/
Now that all software has been installed, bootscripts have been written and the local network is setup, it's time for you to reboot your computer and test these new scripts to verify that they actually work. You first want to execute them manually from the /etc/init.d directory so you can fix the most obvious problems (typos, wrong paths and such). When those scripts seem to work just fine manually they should also work during a system start or shutdown. There's only one way to test that. Shutdown your system with shutdown -r now and reboot into LFS. After the reboot you will have a normal login prompt like you have on your normal Linux system (unless you use XDM or some sort of other Display Manger (like KDM - KDE's version of XDM).
We need to add a few groups and a user which will be used by the email utilities.
groupadd -g 1 bin groupadd -g 2 kmemgroupadd -g 3 mailuseradd -u 1 -g bin -d /bin -s /bin/sh binThere are two directories used by the email sub system, thus we need to create them and give them the proper permissions.
mkdir /var/spool mkdir /var/mail cd /var/spool; ln -s ../mail mail chmod 700 /var/spool/mqueue chmod 775 /var/mail chgrp mail /var/mail chmod 1777 /tmp
cd src ./Build; ./Build install
Configuring Sendmail isn't as easily said as done. There are a lot of things you need to consider while configuring Sendmail and I can't take everything into account. That's why at this time we'll create a very basic and standard setup. If you want to tweak Sendmail to your own liking, go right ahead, but this is not the right article. You could always use your existing /etc/sendmail.cf (or /etc/mail/sendmail.cf) file if you need to use certain features.
cf/lfs.mc containing the following:
OSTYPE(LFS) FEATURE(nouucp) define(`LOCAL_MAILER_PATH', /usr/bin/procmail) MAILER(local) MAILER(smtp)
ostype/lfs.m4 by running touch ostype/LFS.m4m4 m4/cf.m4 cf/lfs.cf > cf/lfs.cf/etc/sendmail.cftouch /etc/aliasessendmail -v -bi
make; make install; make install-suid
gcc *.c -o mail chgrp mail mail; chmod 2755 mail cp mail /usr/bin cp mail.1 /usr/man/man1
/etc/init.d/sendmail containing the following:
#!/bin/sh
# Begin /etc/init.d/sendmail
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
case "$1" in
start)
echo -n "Starting Sendmail..."
start-stop-daemon -S -q -o -x /usr/sbin/sendmail -- -bd
check_status
;;
stop)
echo -n "Stopping Sendmail..."
start-stop-daemon -K -q -o -p /var/run/sendmail.pid
check_status
;;
reload)
echo -n "Reloading Sendmail configuration file..."
start-stop-daemon -K -q -s 1 -p /var/run/sendmail.pid
check_status
;;
restart)
echo -n "Stopping Sendmail..."
start-stop-daemon -K -q -o -p /var/run/sendmail.pid
check_status
sleep 1
echo -n "Starting Sendmail..."
start-stop-daemon -S -q -o -x /usr/sbin/sendmail -- -bd
check_status
;;
*)
echo "Usage: $0 {start|stop|reload|restart}"
exit 1
;;
esac
# End /etc/init.d/sendmail
chmod 755 /etc/init.d/sendmail
cd /etc/init.d/rc2.d; ln -s ../init.d/sendmail S20sendmail cd ../rc0.d; ln -s ../init.d/sendmail K20sendmail cd ../rc6.d; ln -s ../init.d/sendmail K20sendmail
My favorite email client is Mutt, so that's why we're installing this one. Feel free to skip the installation of Mutt and install your own favorite client. After all, this is going to be your system. Not mine.
If your favorite client is an X Window client (such as Netscape Mail) then you'll have to sit tight a little while till we've installed X.
./configure make; make install
./configure make; make install
It's time to test the email system now.
/usr/sbin/sendmail -bd (you need to start sendmail
using the full path. If you don't, you can't let sendmail reload the sendmail.cf
by with kill -1 <sendmail pid>).echo "this is an email test"
| mail -s test rootmail program and you should see your email there.useradd -m testuserecho "test mail to testuser"
| mail -s test testuser If this all worked just fine, you have a working email system for local email. It's not necessarily ready for Internet yet. You can remove the testuser by running userdel -r testuser
In this section we're going to install three of the most used Internet servers, together with the necessary clients. These are going to be installed:
telnetd with the standard telnet client
proftpd with the standard ftp client
apache with lynx as client
./configure make; make install
./configure make; make install
./configure make; make install
Apache isn't that easily configured. Like with Sendmail, a lot depends on your own preference and system setup. Therefore, once I again I stick with a very basic installation. If this doesn't work well enough for you, read the documentation and modify whatever you need to.
./configure make; make install
Zlib is a compression library, used by programs like PKware's zip and unzip utilities. Lynx can use this library to compress certain files.
./configure make; make install
./configure --libdir=/etc --with-zlib make; make install make install-help; make install-doc
It's possible to run the daemons in either stand-alone mode or via the Internet Server daemon (inetd). Where possible, I choose to run the daemons in stand-alone mode. This makes it easier to start and stop individual processes without modifying the /etc/inetd.conf file constantly.
However, in the telnetd case it's better to run it via inetd, since telnetd doesn't seem to respawn itself when the last user logs out. This would mean as soon as the last person logs out from the telnet session, the telnet daemon stops as well. This isn't desirable, so we let telnetd run using inetd to spawn a telnet process whenever somebody logs on.
/etc/inetd.conf containing the following:
# Begin /etc/inetd.conf telnet stream tcp nowait root /usr/sbin/in.telnetd # End /etc/inetd.conf
/etc/init.d/inetd containing the following:
#!/bin/sh
# Begin /etc/init.d/inetd
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
case "$1" in
start)
echo -n "Starting Internet Server daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/inetd
check_status
;;
stop)
echo -n "Stopping Internet Server daemon..."
start-stop-daemon -K -q -o -p /var/run/inetd.pid
check_status
;;
reload)
echo -n "Reloading Internet Server configuration file..."
start-stop-daemon -K -q -s 1 -p /var/run/inetd.pid
check_status
;;
restart)
echo -n "Stopping Internet Server daemon..."
start-stop-daemon -K -q -o -p /var/run/inetd.pid
check_status
sleep 1
echo -n "Starting Internet Server daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/inetd
check_status
;;
*)
echo "Usage: $0 {start|stop|reload|restart}"
;;
esac
# End /etc/init.d/inetd
chmod 755 /etc/init.d/inetd
cd /etc/rc2.d; ln -s ../init.d/inetd S30inetd cd ../rc0.d; ln -s ../init.d/inetd K30inetd cd ../rc6.d; ln -s ../init.d/inetd K30 inetd
groupadd -g 65534 nogroup groupadd -g 4 ftp
useradd -u 65534 -g nogroup -d /home nobody useradd -u 4 -g ftp -m ftp
/etc/init.d/proftpd containing the following:
#!/bin/sh
# Begin /etc/init.d/proftpd
check_status()
{
if [ $? = 0 ]
then
echo "OK"
else
echo "FAILED"
fi
}
case "$1" in
start)
echo -n "Starting Pro FTP daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/proftpd
check_status
;;
stop)
echo -n "Stopping Pro FTP daemon..."
start-stop-daemon -K -q -o -x /usr/sbin/proftpd
check_status
;;
restart)
echo -n "Stopping Pro FTP daemon..."
start-stop-daemon -K -q -o -x /usr/sbin/proftpd
check_status
sleep 1
echo -n "Starting Pro FTP daemon..."
start-stop-daemon -S -q -o -x /usr/sbin/proftpd
check_status
;;
*)
echo "Usage: $0 {start|stop|restart}"
;;
esac
# End /etc/init.d/proftpd
chmod 755 /etc/init.d/proftpd
cd /etc/rc2.d; ln -s ../init.d/proftpd S40proftpd cd ../rc0.d; ln -s ../init.d/proftpd K40proftpd cd ../rc6.d; ln -s ../init.d/proftpd K40proftpd
Edit the files in the /usr/apache/etc directory and modify them according to your own needs.
httpd.conf file and find the variable: Group
/etc/init.d/apache containing the following:
#!/bin/sh
# Begin /etc/init.d/apache
case "$1" in
start)
echo -n "Starting Apache HTTP daemon..."
/usr/apache/sbin/apachectl start
;;
stop)
echo -n "Stopping Apache HTTP daemon..."
/usr/apache/sbin/apachectl stop
;;
restart)
echo -n "Restarting Apache HTTP daemon..."
/usr/apache/sbin/apachectl restart
;;
force-restart)
echo -n "Stopping Apache HTTP daemon..."
/usr/apache/sbin/apachectl stop
sleep 1
echo -n "Starting Apache HTTP daemon..."
/usr/apache/sbin/apachectl start
;;
*)
echo "Usage: $0 {start|stop|restart|force-restart}"
;;
esac
# End /etc/init.d/apache
chmod 755 /etc/init.d/apache
cd /etc/rc2.d; ln -s ../init.d/apache S50apache cd ../rc0.d; ln -s ../init.d/apache K50apache cd ../rc6.d; ln -s ../init.d/apache K50apache
The last step in this section is testing the just installed and configured daemons.
/etc/init.d/inetd
starttelnet localhost/etc/init.d/proftpd startftp localhost/etc/init.d/apache startlynx http://localhostIf these tests ran without trouble, the daemons are all working fine.
Before you can logon to the Internet, the kernel must be ppp-aware. You can accomplish this by compiling ppp-support directly into the kernel, or compiling the ppp drivers are modules which you load when you need them. Whatever you prefer, do it now by re-configuring the kernel if necessary. If your LFS kernel is already ppp-aware than you don't have to re-configure the kernel.
groupadd -g 5 daemonmkdir /var/lock
./configure make; make install
/etc/resolv.conf containing the following:
# Begin /etc/resolv.conf nameserver <IP address of your ISP's primairy DNS server> nameserver <IP address of your ISP's secundairy DNS server> # End /etc/resolv.conf
/usr/bin/pon file containing the following:
#!/bin/sh # Begin /usr/bin/pon /usr/sbin/pppd call provider # End /usr/bin/pon
/usr/bin/poff file containing the following:
#!/bin/sh
# Begin /usr/bin/poff
set -- `cat /var/run/ppp*.pid`
case $# in
0)
kill -15 `ps axw|grep "pppd call [[allnum:]]+"|grep -v grep|awk '{print $1}'`
exit 0
;;
1)
kill -15 $1
exit 0
;;
esac
# End /usr/bin/poff
/etc/ppp/peers directory/etc/ppp/peers/provider containing the following:
# Begin /etc/ppp/peers/provider noauth connect "/usr/sbin/chat -v -f /etc/chatscripts/provider" /dev/ttyS1 115200 defaultroute noipdefault # End /etc/ppp/peers/provider
/etc/chatscripts directory/etc/chatscripts/provider containing the following:
# Begin /etc/chatscripts/provider ABORT BUSY ABORT "NO CARRIER" ABORT VOICE ABORT "NO DIALTONE" ABORT "NO ANSWER" "" ATZ OK ATDT <ISP's phonenumber> TIMEOUT 35 CONNECT '' TIMEOUT 10 ogin: \q<username> TIMEOUT 10 assword: \q<mysecretpassword> # End /etc/chatscripts/provider
As you see from the sample scripts (these are the actual scripts I use when I'm not using X) above I logon to my ISP using this chatscripts in stead of authenticating via pap or chap. Though my ISP supports pap, I choose to do it this slightly different way which has it's disadvantages and advantages. In my case the advantages outweigh the disadvantages. This way I have more control over my logon procedure and I can see closer what is happening when.
For example most times when I connect I have a window running tail -f /var/log/syslog
so I can keep an eye on when (with my provider it's mostly 'if') things like
the username and password are sent.
make World make install; make install.man
During the compilation process you will encounter a few errors about the "makedepend" script not being able to find the stddef.h stdarg.h and float.h header files. The script just isn't as smart as the compiler is apparently, since the compilation itself does work fine without compilation errors. Though, creating a few temporary symlinks won't solve the problem; they only will cause more problems.
So you just ignore the many makedepend errors you most likely will be getting. Also errors similar to "pointer targets in passing arg x of somefunction differ in signedness". You can rewrite those files if you feel like it. I won't do it.
Create a new file /etc/ld.so.conf containing the following:
# Begin /etc/ld.so.conf /lib /usr/lib /usr/X11R6/lib # End /etc/ld.so.conf
ldconfig
ln
-s /usr/X11R6/include/X11 /usr/include/X11Often software copies files to /usr/X11 so it doesn't have to know which release of X you are using. This symlink hasn't been created by the X installation, so we have to create it by ourselves.
ln -s /usr/X11R6 /usr/X11There are a few ways to add the /usr/X11/bin path to the $PATH environment variable. One way of doing so is the following:
/root/.bashrc with it's contents as follows: export PATH=$PATH:/usr/X11/binYou need to login again for this change to become effective. Or you can
update the path by running export PATH=$PATH:/usr/X11/bin manually
xf86configIf the XF86Config file created by xf86config doesn't suffice, then you better copy the already existing XF86Config from your normal Linux system to /etc. Cases wherein you need to make special changes to the file which aren't supported by the xf86config program force you to do this. You can always modify the created XF86Config file by hand. This can be very time consuming, especially if you don't quite remember what needs to be changed.
Now that X is properly configured it's time for our first test run.
startxThe X server should start and display 3 xterm's on your screen. If this is true in your case, X is running fine.
I choose to install Window Maker as the Window Manager. This is because I've used WindowMaker for quite a while now and I'm very satisfied with it. As usual, you don't have to do what I'm doing; install whatever you want. As you might know, you can install several Window Managers simultaneously and choose which one to start by specifying it in the $HOME/.xinitrc (or $HOME/.xsession in case you decide to use xdm) file.
./configure make; make install
xmkmf; make Makefiles; make includes; make depend cd lib; make; make install cd ..; make; make install
This slightly different installation is necessary due to a bug in one of the Makefiles. It depends on files in the lib directory which aren't installed yet and it's not searching for them in the lib directory, so we have to install those files first before compiling the actual package.
make -f scripts/makefile.lnx; make -f scripts/makefile.lnx install
./configure make; make instal
./configure --enable-shared --enable-static make libjpb.a; make install
./configure make; make install
./configure make; make install
ldconfigEvery user who wishes to use WindowMaker has to run the wmaker.inst script before he or she can use it. This script will copy the necessary files into the user's home directory and modify the $HOME/.xinitrc file (or create it if it's not there yet).
wmaker.inst
startxYou have reached the end of the Linux From Scratch HOWTO. I hope this experience helped you getting to know Linux better. If you have anything that you think needs to be mentioned in here (be it a bugfix, extra software which has been forgotten but which you concider important) let us know. Together with your help and suggestions this HOWTO can be improved significantly.
Copyright (C) 1999 by Gerard Beekmans. This document may be distributed only subject to the terms and conditions set forth in the LDP License at http://www.linuxdoc.org/COPYRIGHT.html.
It is not necessary to display the license notice, as described in the LDP License, when only a small part of this document (the HOWTO) is quoted for informational or similar purposes. However, I do require you to display with the quotation(s) a line similar to the following line: "Quoted from the LFS-HOWTO at http://huizen.dds.nl/~glb/