You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
openwrt/openwrt/docs/buildroot-documentation.html

599 lines
26 KiB
HTML

<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>OpenWrt Buildroot - Usage and documentation</title>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" />
<link rel="stylesheet" type="text/css" href="stylesheet.css" />
</head>
<body>
<div class="main">
<div class="titre">
<h1>OpenWrt Buildroot</h1>
</div>
<p>Usage and documentation by Felix Fietkau, based on uClibc Buildroot
documentation by Thomas Petazzoni. Contributions from Karsten Kruse,
Ned Ludd, Martin Herren.</p>
<p><small>Last modification : $Id$</small></p>
<ul>
<li><a href="#about">About OpenWrt Buildroot</a></li>
<li><a href="#download">Obtaining OpenWrt Buildroot</a></li>
<li><a href="#using">Using OpenWrt Buildroot</a></li>
<li><a href="#custom_targetfs">Customizing the target filesystem</a></li>
<li><a href="#custom_busybox">Customizing the Busybox
configuration</a></li>
<li><a href="#custom_uclibc">Customizing the uClibc
configuration</a></li>
<li><a href="#buildroot_innards">How OpenWrt Buildroot works</a></li>
<li><a href="#using_toolchain">Using the uClibc toolchain</a></li>
<li><a href="#toolchain_standalone">Using the uClibc toolchain
outside of Buildroot</a></li>
<li><a href="#downloaded_packages">Location of downloaded packages</a></li>
<li><a href="#add_software">Extending OpenWrt with more Software</a></li>
<li><a href="#links">Ressources</a></li>
</ul>
<h2><a name="about" id="about"></a>About OpenWrt Buildroot</h2>
<p>OpenWrt Buildroot is a set of Makefiles and patches that allows to easily
generate both a cross-compilation toolchain and a root filesystem for your
Wireless Router. The cross-compilation toolchain uses uClibc (<a href=
"http://www.uclibc.org/">http://www.uclibc.org/</a>), a tiny C standard
library.</p>
<p>A compilation toolchain is the set of tools that allows to
compile code for your system. It consists of a compiler (in our
case, <code>gcc</code>), binary utils like assembler and linker
(in our case, <code>binutils</code>) and a C standard library (for
example <a href="http://www.gnu.org/software/libc/libc.html">GNU
Libc</a>, <a href="http://www.uclibc.org">uClibc</a> or <a
href="http://www.fefe.de/dietlibc/">dietlibc</a>). The system
installed on your development station certainly already has a
compilation toolchain that you can use to compile application that
runs on your system. If you're using a PC, your compilation
toolchain runs on an x86 processor and generates code for a x86
processor. Under most Linux systems, the compilation toolchain
uses the GNU libc as C standard library. This compilation
toolchain is called the "host compilation toolchain", and more
generally, the machine on which it is running, and on which you're
working is called the "host system". The compilation toolchain is
provided by your distribution, and OpenWrt Buildroot has nothing to do
with it.</p>
<p>As said above, the compilation toolchain that comes with your system
runs and generates code for the processor of your host system. As your
embedded system has a different processor, you need a cross-compilation
toolchain: it's a compilation toolchain that runs on your host system but
that generates code for your target system (and target processor). For
example, if your host system uses x86 and your target system uses MIPS, the
regular compilation toolchain of your host runs on x86 and generates code
for x86, while the cross-compilation toolchain runs on x86 and generates
code for MIPS.</p>
<p>You might wonder why such a tool is needed when you can compile
<code>gcc</code>, <code>binutils</code>, uClibc and all the tools by hand.
Of course, doing so is possible. But dealing with all configure options,
with all problems of every <code>gcc</code> or <code>binutils</code>
version it very time-consuming and uninteresting. OpenWrt Buildroot automates this
process through the use of Makefiles, and has a collection of patches for
each <code>gcc</code> and <code>binutils</code> version to make them work
on the MIPS architecture of most Broadcom based Wireless Routers.</p>
<h2><a name="download" id="download"></a>Obtaining OpenWrt Buildroot</h2>
<p>OpenWrt Buildroot is currently available as experimental snapshots</p>
<p>The latest snapshot is always available at <a
href="http://openwrt.org/downloads/experimental/">http://openwrt.org/downloads/experimental/</a>,
<h2><a name="using" id="using"></a>Using OpenWrt Buildroot</h2>
<p>OpenWrt Buildroot has a nice configuration tool similar to the one you can find
in the Linux Kernel (<a href="http://www.kernel.org/">http://www.kernel.org/</a>)
or in Busybox (<a href="http://www.busybox.org/">http://www.busybox.org/</a>).
Note that you can run everything as a normal user. There is no need to be root to
configure and use the Buildroot. The first step is to run the configuration
assistant:</p>
<pre>
$ make menuconfig
</pre>
<p>For each entry of the configuration tool, you can find associated help
that describes the purpose of the entry.</p>
<p>Once everything is configured, the configuration tool has generated a
<code>.config</code> file that contains the description of your
configuration. It will be used by the Makefiles to do what's needed.</p>
<p>Let's go:</p>
<pre>
$ make
</pre>
<p>This command will download, configure and compile all the selected
tools, and finally generate target firmware images and additional packages
(depending on your selections in <code>make menuconfig</code>.
All the target files can be found in the <code>bin/</code> subdirectory.
You can compile firmware images containing two different filesystem types:
<ul>
<li>jffs2</li>
<li>squashfs</li>
</ul>
<p><code>jffs2</code> contains a writable root filesystem, which will expand to
the size of your flash image. Note that you if you use the generic firmware
Image, you need to pick the right image for your Flash size, because of different
eraseblock sizes.</p>
<p><code>squashfs</code> contains a read-only root filesystem using a modified
<code>squashfs</code> filesystem with LZMA compression. When booting it, you can
create a writable second filesystem, which will contain your modifications to
the root filesystem, including the packages you install.
<h2><a name="custom_targetfs" id="custom_targetfs"></a>Customizing the
target filesystem</h2>
<p>There are two ways to customize the resulting target filesystem:</p>
<ul>
<li>Customize the target filesystem directly, and rebuild the image. The
target filesystem is available under <code>build_ARCH/root/</code> where
<code>ARCH</code> is the chosen target architecture, usually mipsel.
You can simply make your changes here, and run make target_install afterwards,
which will rebuild the target filesystem image. This method allows to do
everything on the target filesystem, but if you decide to rebuild your toolchain,
tools or packages, these changes will be lost.</li>
<li>Customize the target filesystem skeleton, available under
<code>target/default/target_skeleton/</code>. You can customize
configuration files or other stuff here. However, the full file hierarchy
is not yet present, because it's created during the compilation process.
So you can't do everything on this target filesystem skeleton, but
changes to it remains even when you completely rebuild the cross-compilation
toolchain and the tools.<br />
</ul>
<h2><a name="custom_busybox" id="custom_busybox"></a>Customizing the
Busybox configuration</h2>
<p>Busybox is very configurable, and you may want to customize it.
Its configuration is completely integrated into the main menuconfig system.
You can find it under "OpenWrt Package Selection" =&gt; "Busybox Configuration"</p>
<h2><a name="custom_uclibc" id="custom_uclibc"></a>Customizing the uClibc
configuration</h2>
<p>Just like <a href="#custom_busybox">BusyBox</a>, <a
href="http://www.uclibc.org">uClibc</a> offers a lot of
configuration options. They allow to select various
functionalities, depending on your needs and limitations.</p>
<p>The easiest way to modify the configuration of uClibc is to
follow these steps :</p>
<ol>
<li>Make a first compilation of buildroot without trying to
customize uClibc.</li>
<li>Go into the directory
<code>toolchain_build_ARCH/uClibc/</code> and run <code>make
menuconfig</code>. The nice configuration assistant, similar to
the one used in the Linux Kernel appears. Make
your configuration as appropriate.</li>
<li>Copy the <code>.config</code> file to
<code>toolchain/uClibc/uClibc.config</code> or
<code>toolchain/uClibc/uClibc.config-locale</code>. The former
is used if you haven't selected locale support in the Buildroot
configuration, and the latter is used if you have selected
locale support.</li>
<li>Run the compilation again</li>
</ol>
<p>Otherwise, you can simply change
<code>toolchain/uClibc/uClibc.config</code> or
<code>toolchain/uClibc/uClibc.config-locale</code> without running
the configuration assistant.</p>
<h2><a name="buildroot_innards" id="buildroot_innards"></a>How OpenWrt Buildroot
works</h2>
<p>As said above, OpenWrt is basically a set of Makefiles that download,
configure and compiles software with the correct options. It also includes
some patches for various software, mainly the ones involved in the
cross-compilation tool chain (<code>gcc</code>, <code>binutils</code> and
uClibc).</p>
<p>There is basically one Makefile per software, and they are named <code>Makefile</code>.
Makefiles are split into three sections:</p>
<ul>
<li><b>package</b> (in the <code>package/</code> directory) contains the
Makefiles and associated files for all user-space tools that Buildroot
can compile and add to the target root filesystem. There is one
sub-directory per tool.</li>
<li><b>toolchain</b> (in the <code>toolchain/</code> directory) contains
the Makefiles and associated files for all software related to the
cross-compilation toolchain : <code>binutils</code>, <code>ccache</code>,
<code>gcc</code>, <code>gdb</code>, <code>kernel-headers</code> and
<code>uClibc</code>.</li>
<li><b>target</b> (in the <code>target</code> directory) contains the
Makefiles and associated files for software related to the generation of
the target root filesystem image. Two types of filesystems are supported
: jffs2 and squashfs.
</ul>
<p>Each directory contains at least 2 files :</p>
<ul>
<li><code>Makefile</code> is the Makefile that downloads, configures,
compiles and installs the software <code>something</code>.</li>
<li><code>Config.in</code> is a part of the configuration tool
description file. It describes the option related to the current
software.</li>
</ul>
<p>The main Makefile do the job through the following steps (once the
configuration is done):</p>
<ol>
<li>Create the download directory (<code>dl/</code> by default). This is
where the tarballs will be downloaded. It is interesting to know that the
tarballs are in this directory because it may be useful to save them
somewhere to avoid further downloads.</li>
<li>Create the build directory (<code>build_ARCH/</code> by default,
where <code>ARCH</code> is your architecture). This is where all
user-space tools while be compiled.</li>
<li>Create the toolchain build directory
(<code>toolchain_build_ARCH/</code> by default, where <code>ARCH</code>
is your architecture). This is where the cross compilation toolchain will
be compiled.</li>
<li>Setup the staging directory (<code>staging_dir_ARCH/</code> by
default). This is where the cross-compilation toolchain will be
installed. If you want to use the same cross-compilation toolchain for
other purposes, such as compiling third-party applications, you can add
<code>staging_dir_ARCH/bin</code> to your PATH, and then use
<code>arch-linux-gcc</code> to compile your application. In order to
setup this staging directory, it first removes it, and then it creates
various subdirectories and symlinks inside it.</li>
<li>Create the target directory (<code>build_ARCH/root/</code> by
default) and the target filesystem skeleton. This directory will contain
the final root filesystem. To setup it up, it first deletes it, then it
copies the skeleton available in <code>target/default/target_skeleton</code>
and then removes useless <code>CVS/</code> directories.</li>
<li>Call the <code>prepare</code>, <code>compile</code> and <code>install</code>
targets for the subdirectories <code>toolchain</code>, <code>package</code>
and <code>target</code></li>
</ol>
<h2><a name="using_toolchain" id="using_toolchain"></a>Using the
uClibc toolchain</h2>
<p>You may want to compile your own programs or other software
that are not packaged in OpenWrt. In order to do this, you can
use the toolchain that was generated by the Buildroot.</p>
<p>The toolchain generated by the Buildroot by default is located in
<code>staging_dir_ARCH</code>. The simplest way to use it
is to add <code>staging_dir_ARCH/bin/</code> to your PATH
environment variable, and then to use
<code>arch-linux-gcc</code>, <code>arch-linux-objdump</code>,
<code>arch-linux-ld</code>, etc.</p>
<p>For example, you may add the following to your
<code>.bashrc</code> (considering you're building for the MIPS
architecture and that Buildroot is located in
<code>~/buildroot/</code>) :</p>
<pre>
export PATH=$PATH:~/buildroot/staging_dir_mipsel/bin/
</pre>
<p>Then you can simply do :</p>
<pre>
mipsel-linux-uclibc-gcc -o foo foo.c
</pre>
<p><b>Important</b> : do not try to move the toolchain to an other
directory, it won't work. There are some hard-coded paths in the
<i>gcc</i> configuration. If the default toolchain directory
doesn't suit your needs, please refer to the <a
href="#toolchain_standalone">Using the uClibc toolchain outside of
buildroot</a> section.</p>
<h2><a name="toolchain_standalone" id="toolchain_standalone"></a>Using the
uClibc toolchain outside of buildroot</h2>
<p>By default, the cross-compilation toolchain is generated inside
<code>staging_dir_ARCH/</code>. But sometimes, it may be useful to
install it somewhere else, so that it can be used to compile other programs
or by other users. Moving the <code>staging_dir_ARCH/</code>
directory elsewhere is <b>not possible</b>, because they are some hardcoded
paths in the toolchain configuration.</p>
<p>If you want to use the generated toolchain for other purposes,
you can configure Buildroot to generate it elsewhere using the
option of the configuration tool : <code>Build options -&gt;
Toolchain and header file location</code>, which defaults to
<code>staging_dir_ARCH/</code>.</p>
<h2><a name="downloaded_packages"
id="downloaded_packages"></a>Location of downloaded packages</h2>
<p>It might be useful to know that the various tarballs that are
downloaded by the <i>Makefiles</i> are all stored in the
<code>DL_DIR</code> which by default is the <code>dl</code>
directory. It's useful for example if you want to keep a complete
version of Buildroot which is know to be working with the
associated tarballs. This will allow you to regenerate the
toolchain and the target filesystem with exactly the same
versions.</p>
<h2><a name="add_software" id="add_software"></a>Extending OpenWrt with
more software</h2>
<p>This section will only consider the case in which you want to
add user-space software.</p>
<h3>Package directory</h3>
<p>First of all, create a directory under the <code>package</code>
directory for your software, for example <code>foo</code>.</p>
<h3><code>Config.in</code> file</h3>
<p>Then, create a file named <code>Config.in</code>. This file
will contain the portion of options description related to our
<code>foo</code> software that will be used and displayed in the
configuration tool. It should basically contain :</p>
<pre>
config BR2_PACKAGE_FOO
tristate "foo"
default n
help
This is a comment that explains what foo is.
</pre>
<p>Of course, you can add other options to configure particular
things in your software.</p>
<h3><code>Makefile</code> in the package directory</h3>
<p>To add your package to the build process, you need to edit
the Makefile in the <code>package/</code> directory. Locate the
lines that look like the following:</p>
<pre>
package-$(BR2_PACKAGE_FOO) += foo
</pre>
<p>As you can see, this short line simply adds the target
<code>foo</code> to the list of targets handled by OpenWrt Buildroot.</p>
<p>In addition to the default dependencies, you make your package
depend on another package (e.g. a library) by adding a line:
<pre>
foo-compile: bar-compile
</pre>
<h3>The <i>.control</i> file</h3>
<p>Additionally, you need to create a control file which contains
information about your package, readable by the <i>ipkg</i> package
utility.</p>
<p>The file looks like this</p>
<pre>
1 Package: foo
2 Priority: optional
3 Section: net
4 Maintainer: Foo Software &lt;foo@foosoftware.com&gt;
5 Source: http://foosoftware.com
6 Description: Your Package Description
</pre>
<p>You can skip the usual <code>Version:</code> and <code>Architecture</code>
fields, as they will be generated by the <code>make-ipkg-dir.sh</code> script
called from your Makefile</p>
<h3>The real <i>Makefile</i></h3>
<p>Finally, here's the hardest part. Create a file named
<code>Makefile</code>. It will contain the <i>Makefile</i> rules that
are in charge of downloading, configuring, compiling and installing
the software. Below is an example that we will comment
afterwards.</p>
<pre>
1 # $Id$
2
3 PKG_NAME:=foo
4 PKG_VERSION:=1.0
5 PKG_RELEASE:=1
6 PKG_MD5SUM:=4584f226523776a3cdd2fb6f8212ba8d
7
8 PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.gz
9 PKG_SOURCE_URL:=http://www.foosoftware.org/downloads
10 PKG_DIR:=$(BUILD_DIR)/$(PKG_NAME)-$(PKG_VERSION)
11 PKG_IPK:=$(PACKAGE_DIR)/$(PKG_NAME)_$(PKG_VERSION)-$(PKG_RELEASE)_$(ARCH).ipk
12 PKG_IPK_DIR:=$(PKG_DIR)/ipkg
13
14 $(DL_DIR)/$(PKG_SOURCE):
15 $(SCRIPT_DIR)/download.pl $(DL_DIR) $(PKG_SOURCE) $(PKG_MD5SUM) $(PKG_SOURCE_URL)
16
17 $(PKG_DIR)/.source: $(DL_DIR)/$(PKG_SOURCE)
18 zcat $(DL_DIR)/$(PKG_SOURCE) | tar -C $(BUILD_DIR) $(TAR_OPTIONS) -
19 touch $(PKG_DIR)/.source
20
21 $(PKG_DIR)/.configured: $(PKG_DIR)/.source
22 (cd $(PKG_DIR); \
23 $(TARGET_CONFIGURE_OPTS) \
24 CFLAGS="$(TARGET_CFLAGS)" \
25 ./configure \
26 --target=$(GNU_TARGET_NAME) \
27 --host=$(GNU_TARGET_NAME) \
28 --build=$(GNU_HOST_NAME) \
29 --prefix=/usr \
30 --sysconfdir=/etc \
31 );
32 touch $(PKG_DIR)/.configured;
33
34 $(PKG_DIR)/$(PKG_NAME) $(PKG_DIR)/.configured
35 $(MAKE) CC=$(TARGET_CC) -C $(PKG_DIR)
36
37 $(PKG_IPK): $(PKG_DIR)/$(PKG_NAME)
38 $(SCRIPT_DIR)/make-ipkg-dir.sh $(PKG_IPK_DIR) $(PKG_NAME).control $(PKG_VERSION)-$(PKG_RELEASE) $(ARCH)
39 $(MAKE) prefix=$(PKG_IPK_DIR)/usr -C $(PKG_DIR) install
40 rm -Rf $(PKG_IPK_DIR)/usr/man
41 $(IPKG_BUILD) $(PKG_IPK_DIR) $(PACKAGE_DIR)
42
43 $(IPKG_STATE_DIR)/info/$(PKG_NAME).list: $(PKG_IPK)
44 $(IPKG) install $(PKG_IPK)
45
46 prepare: $(PKG_DIR)/.source
47 compile: $(PKG_IPK)
48 install: $(IPKG_STATE_DIR)/info/$(PKG_NAME).list
49 clean:
50 rm -rf $(PKG_DIR)
51 rm -f $(PKG_IPK)
</pre>
<p>First of all, this <i>Makefile</i> example works for a single
binary software. For other software such as libraries or more
complex stuff with multiple binaries, it should be adapted. Look at
the other <code>Makefile</code> files in the <code>package</code>
directory.</p>
<p>At lines 3-12, a couple of useful variables are defined :</p>
<ul>
<li><code>PKG_NAME</code> : The package name, e.g. <i>foo</i>.</li>
<li><code>PKG_VERSION</code> : The version of the package that
should be downloaded.</li>
<li><code>PKG_RELEASE</code> : The release number that will be
appended to the version number of your <i>ipkg</i> package.
<li><code>PKG_MD5SUM</code> : The md5sum of the software archive.
<li><code>PKG_SOURCE</code> : The name of the tarball of
your package on the download website of FTP site. As you can see
<code>PKG_NAME</code> and <code>PKG_VERSION</code> are used.</li>
<li><code>PKG_SOURCE_URL</code> : Space separated list of the HTTP
or FTP sites from which the archive is downloaded. It must include the complete
path to the directory where <code>FOO_SOURCE</code> can be
found.</li>
<li><code>PKG_DIR</code> : The directory into which the software
will be configured and compiled. Basically, it's a subdirectory
of <code>BUILD_DIR</code> which is created upon decompression of
the tarball.</li>
<li><code>PKG_IPK</code> : The resulting <i>ipkg</i> package
</ul>
<p>Lines 14-15 defines a target that downloads the tarball from
the remote site to the download directory
(<code>DL_DIR</code>).</p>
<p>Lines 17-19 defines a target and associated rules that
uncompress the downloaded tarball. As you can see, this target
depends on the tarball file, so that the previous target (line
14-15) is called before executing the rules of the current
target. Uncompressing is followed by <i>touching</i> a hidden file
to mark the software has having been uncompressed. This trick is
used everywhere in Buildroot <i>Makefile</i> to split steps
(download, uncompress, configure, compile, install) while still
having correct dependencies.</p>
<p>Lines 21-32 defines a target and associated rules that
configures the software. It depends on the previous target (the
hidden <code>.source</code> file) so that we are sure the software has
been uncompressed. In order to configure it, it basically runs the
well-known <code>./configure</code>script. As we may be doing
cross-compilation, <code>target</code>, <code>host</code> and
<code>build</code> arguments are given. The prefix is also set to
<code>/usr</code>, not because the software will be installed in
<code>/usr</code> on your host system, but in the target
filesystem. Finally it creates a <code>.configured</code> file to
mark the software as configured.</p>
<p>Lines 34-35 defines a target and a rule that compiles the
software. This target will create the binary file in the
compilation directory, and depends on the software being already
configured (hence the reference to the <code>.configured</code>
file). It basically runs <code>make</code> inside the source
directory.</p>
<p>Lines 37-41 defines a target and associated rules that create
the <i>ipkg</i> package which can optionally be embedded into
the resulting firmware image. It depends on the binary file in
the source directory, to make sure the software has been compiled.
It uses the make-ipkg-dir.sh script, which will create the ipkg
build directory for your package, copy your control file into
that directory and add version and architecture information.
Then it calls the <code>install</code> target of the
software <code>Makefile</code> by passing a <code>prefix</code>
argument, so that the <code>Makefile</code> doesn't try to install
the software inside host <code>/usr</code> but inside target
<code>/usr</code>. After the installation, the
<code>/usr/man</code> directory inside the target filesystem is
removed to save space.
Finally <code>IPKG_BUILD</code> is called to create the package.</p>
<p>Line 43 and 44 define the installation target of your package,
which will embed the software into the target filesystem.</p>
<p>Lines 46-51 define the main targets that the Makefile in the
<code>package</code> dir calls.
<ul>
<li><code>prepare</code> : Download and unpack the source</li>
<li><code>compile</code> : Compile the source and create the package</li>
<li><code>install</code> : Embed the package into the target filesystem</li>
<li><code>clean</code> : Remove all the files created by the build process</li>
</ul></p>
<h3>Conclusion</h3>
<p>As you can see, adding a software to buildroot is simply a
matter of writing a <i>Makefile</i> using an already existing
example and to modify it according to the compilation process of
the software.</p>
<p>If you package software that might be useful for other persons,
don't forget to send a patch to OpenWrt developers !</p>
<h2><a name="links" id="links"></a>Resources</h2>
<p>To learn more about OpenWrt Buildroot you can visit this
website: <a href="http://openwrt.org/">http://openwrt.org/</a></p>
</div>
</body>
</html>