[98c95f5] | 1 | <sect1 id="ch05-toolchaintechnotes">
|
---|
| 2 | <title>Toolchain technical notes</title>
|
---|
| 3 | <?dbhtml filename="toolchaintechnotes.html" dir="chapter05"?>
|
---|
| 4 |
|
---|
| 5 | <para>This section attempts to explain some of the rationale and technical
|
---|
| 6 | details behind the overall build method. It's not essential that you understand
|
---|
| 7 | everything here immediately. Most of it will make sense once you have performed
|
---|
| 8 | an actual build. Feel free to refer back here at any time.</para>
|
---|
| 9 |
|
---|
| 10 | <para>The overall goal of Chapter 5 is to provide a sane, temporary environment
|
---|
| 11 | that we can chroot into, and from which we can produce a clean, trouble-free
|
---|
| 12 | build of the target LFS system in Chapter 6. Along the way, we attempt to
|
---|
| 13 | divorce ourselves from the host system as much as possible, and in so doing
|
---|
| 14 | build a self-contained and self-hosted toolchain. It should be noted that the
|
---|
| 15 | build process has been designed in such a way so as to minimize the risks for
|
---|
[1668d8e] | 16 | new readers and provide maximum educational value at the same time. In other
|
---|
| 17 | words, more advanced techniques could be used to build the system.</para>
|
---|
[98c95f5] | 18 |
|
---|
| 19 | <important>
|
---|
| 20 | <para>Before continuing, you really should be aware of the name of your working
|
---|
| 21 | platform, often also referred to as the <emphasis>target triplet</emphasis>. For
|
---|
| 22 | many folks the target triplet will be, for example:
|
---|
| 23 | <emphasis>i686-pc-linux-gnu</emphasis>. A simple way to determine your target
|
---|
| 24 | triplet is to run the <filename>config.guess</filename> script that comes with
|
---|
| 25 | the source for many packages. Unpack the Binutils sources and run the script:
|
---|
| 26 | <userinput>./config.guess</userinput> and note the output.</para>
|
---|
| 27 |
|
---|
| 28 | <para>You'll also need to be aware of the name of your platform's
|
---|
| 29 | <emphasis>dynamic linker</emphasis>, often also referred to as the
|
---|
| 30 | <emphasis>dynamic loader</emphasis>, not to be confused with the standard linker
|
---|
| 31 | <emphasis>ld</emphasis> that is part of Binutils. The dynamic linker is provided
|
---|
| 32 | by Glibc and has the job of finding and loading the shared libraries needed by a
|
---|
| 33 | program, preparing the program to run and then running it. For most folks, the
|
---|
| 34 | name of the dynamic linker will be <emphasis>ld-linux.so.2</emphasis>. On
|
---|
| 35 | platforms that are less prevalent, the name might be
|
---|
| 36 | <emphasis>ld.so.1</emphasis> and newer 64 bit platforms might even have
|
---|
| 37 | something completely different. You should be able to determine the name
|
---|
| 38 | of your platform's dynamic linker by looking in the
|
---|
| 39 | <filename class="directory">/lib</filename> directory on your host system. A
|
---|
| 40 | surefire way is to inspect a random binary from your host system by running:
|
---|
[a3f6e124] | 41 | <userinput>'readelf -l <name of binary> | grep interpreter'</userinput>
|
---|
[98c95f5] | 42 | and noting the output. The authoritative reference covering all platforms is in
|
---|
| 43 | the <filename>shlib-versions</filename> file in the root of the Glibc source
|
---|
| 44 | tree.</para>
|
---|
| 45 | </important>
|
---|
| 46 |
|
---|
| 47 | <para>Some key technical points of how the Chapter 5 build method works:</para>
|
---|
| 48 |
|
---|
| 49 | <itemizedlist>
|
---|
| 50 | <listitem><para>Similar in principle to cross compiling whereby tools installed
|
---|
| 51 | into the same prefix work in cooperation and thus utilize a little GNU
|
---|
| 52 | "magic".</para></listitem>
|
---|
| 53 |
|
---|
| 54 | <listitem><para>Careful manipulation of the standard linker's library search
|
---|
| 55 | path to ensure programs are linked only against libraries we
|
---|
| 56 | choose.</para></listitem>
|
---|
| 57 |
|
---|
[1668d8e] | 58 | <listitem><para>Careful manipulation of <userinput>gcc</userinput>'s
|
---|
| 59 | <emphasis>specs</emphasis> file to tell the compiler which target dynamic
|
---|
| 60 | linker will be used.</para></listitem>
|
---|
[98c95f5] | 61 | </itemizedlist>
|
---|
| 62 |
|
---|
| 63 | <para>Binutils is installed first because both GCC and Glibc perform various
|
---|
| 64 | feature tests on the assembler and linker during their respective runs of
|
---|
[1668d8e] | 65 | <userinput>./configure</userinput> to determine which software features to enable
|
---|
[98c95f5] | 66 | or disable. This is more important than one might first realize. An incorrectly
|
---|
| 67 | configured GCC or Glibc can result in a subtly broken toolchain where the impact
|
---|
[1668d8e] | 68 | of such breakage might not show up until near the end of the build of a whole
|
---|
[98c95f5] | 69 | distribution. Thankfully, a test suite failure will usually alert us before too
|
---|
[1668d8e] | 70 | much time is wasted.</para>
|
---|
[98c95f5] | 71 |
|
---|
| 72 | <para>Binutils installs its assembler and linker into two locations,
|
---|
| 73 | <filename class="directory">/tools/bin</filename> and
|
---|
| 74 | <filename class="directory">/tools/$TARGET_TRIPLET/bin</filename>. In reality,
|
---|
[1668d8e] | 75 | the tools in one location are hard linked to the other. An important facet of
|
---|
| 76 | the linker is its library search order. Detailed information can be obtained
|
---|
| 77 | from <userinput>ld</userinput> by passing it the <emphasis>--verbose</emphasis>
|
---|
[a3f6e124] | 78 | flag. For example: <userinput>'ld --verbose | grep SEARCH'</userinput> will
|
---|
[1668d8e] | 79 | show you the current search paths and their order. You can see what files are
|
---|
| 80 | actually linked by <userinput>ld</userinput> by compiling a dummy program and
|
---|
| 81 | passing the <emphasis>--verbose</emphasis> switch. For example:
|
---|
[a3f6e124] | 82 | <userinput>'gcc dummy.c -Wl,--verbose 2>&1 | grep succeeded'</userinput>
|
---|
[98c95f5] | 83 | will show you all the files successfully opened during the link.</para>
|
---|
| 84 |
|
---|
| 85 | <para>The next package installed is GCC and during its run of
|
---|
[1668d8e] | 86 | <userinput>./configure</userinput> you'll see, for example:</para>
|
---|
[98c95f5] | 87 |
|
---|
| 88 | <blockquote><screen>checking what assembler to use... /tools/i686-pc-linux-gnu/bin/as
|
---|
| 89 | checking what linker to use... /tools/i686-pc-linux-gnu/bin/ld</screen></blockquote>
|
---|
| 90 |
|
---|
| 91 | <para>This is important for the reasons mentioned above. It also demonstrates
|
---|
| 92 | that GCC's configure script does not search the $PATH directories to find which
|
---|
[1668d8e] | 93 | tools to use. However, during the actual operation of <userinput>gcc</userinput>
|
---|
| 94 | itself, the same search paths are not necessarily used. You can find out which
|
---|
| 95 | standard linker <userinput>gcc</userinput> will use by running:
|
---|
[a3f6e124] | 96 | <userinput>'gcc -print-prog-name=ld'</userinput>.
|
---|
[1668d8e] | 97 | Detailed information can be obtained from <userinput>gcc</userinput> by passing
|
---|
| 98 | it the <emphasis>-v</emphasis> flag while compiling a dummy program. For
|
---|
[a3f6e124] | 99 | example: <userinput>'gcc -v dummy.c'</userinput> will show you detailed
|
---|
[1668d8e] | 100 | information about the preprocessor, compilation and assembly stages, including
|
---|
| 101 | <userinput>gcc</userinput>'s include search paths and their order.</para>
|
---|
[98c95f5] | 102 |
|
---|
| 103 | <para>The next package installed is Glibc. The most important considerations for
|
---|
| 104 | building Glibc are the compiler, binary tools and kernel headers. The compiler
|
---|
[1668d8e] | 105 | is generally no problem as Glibc will always use the <userinput>gcc</userinput>
|
---|
| 106 | found in a $PATH directory. The binary tools and kernel headers can be a little
|
---|
| 107 | more troublesome. Therefore we take no risks and use the available configure
|
---|
| 108 | switches to enforce the correct selections. After the run of
|
---|
| 109 | <userinput>./configure</userinput> you can check the contents of the
|
---|
[98c95f5] | 110 | <filename>config.make</filename> file in the
|
---|
| 111 | <filename class="directory">glibc-build</filename> directory for all the
|
---|
| 112 | important details. You'll note some interesting items like the use of
|
---|
| 113 | <userinput>CC="gcc -B/tools/bin/"</userinput> to control which binary tools are
|
---|
[1668d8e] | 114 | used, and also the use of the <emphasis>-nostdinc</emphasis> and
|
---|
| 115 | <emphasis>-isystem</emphasis> flags to control the compiler's include search
|
---|
| 116 | path. These items help to highlight an important aspect of the Glibc package:
|
---|
| 117 | it is very self-sufficient in terms of its build machinery and generally does
|
---|
| 118 | not rely on toolchain defaults.</para>
|
---|
[98c95f5] | 119 |
|
---|
| 120 | <para>After the Glibc installation, we make some adjustments to ensure that
|
---|
[1668d8e] | 121 | searching and linking take place only within our <filename>/tools</filename>
|
---|
| 122 | prefix. We install an adjusted <userinput>ld</userinput>, which has a hard-wired
|
---|
| 123 | search path limited to <filename class="directory">/tools/lib</filename>. Then
|
---|
| 124 | we amend <userinput>gcc</userinput>'s specs file to point to our new dynamic
|
---|
| 125 | linker in <filename class="directory">/tools/lib</filename>. This last step is
|
---|
[98c95f5] | 126 | <emphasis>vital</emphasis> to the whole process. As mentioned above, a
|
---|
[f57e3d1] | 127 | hard-wired path to a dynamic linker is embedded into every ELF shared
|
---|
| 128 | executable. You can inspect this by running:
|
---|
[a3f6e124] | 129 | <userinput>'readelf -l <name of binary> | grep interpreter'</userinput>.
|
---|
[1668d8e] | 130 | By amending <userinput>gcc</userinput>'s specs file, we are ensuring that every
|
---|
| 131 | program compiled from here through the end of Chapter 5 will use our new
|
---|
| 132 | dynamic linker in <filename class="directory">/tools/lib</filename>.</para>
|
---|
[98c95f5] | 133 |
|
---|
| 134 | <para>The need to use the new dynamic linker is also the reason why we apply the
|
---|
[1668d8e] | 135 | Specs patch for the second pass of GCC. Failure to do so will result in the GCC
|
---|
| 136 | programs themselves having the name of the dynamic linker from the host system's
|
---|
[98c95f5] | 137 | <filename class="directory">/lib</filename> directory embedded into them, which
|
---|
[1668d8e] | 138 | would defeat our goal of getting away from the host.</para>
|
---|
[98c95f5] | 139 |
|
---|
| 140 | <para>During the second pass of Binutils, we are able to utilize the
|
---|
[1668d8e] | 141 | <emphasis>--with-lib-path</emphasis> configure switch to control
|
---|
| 142 | <userinput>ld</userinput>'s library search path. From this point onwards, the
|
---|
| 143 | core toolchain is self-contained and self-hosted. The remainder of the
|
---|
| 144 | Chapter 5 packages all build against the new Glibc in
|
---|
| 145 | <filename class="directory">/tools</filename> and all is well.</para>
|
---|
[98c95f5] | 146 |
|
---|
| 147 | <para>Upon entering the chroot environment in Chapter 6, the first major package
|
---|
[1668d8e] | 148 | we install is Glibc, due to its self-sufficient nature that we mentioned above.
|
---|
[98c95f5] | 149 | Once this Glibc is installed into <filename class="directory">/usr</filename>,
|
---|
| 150 | we perform a quick changeover of the toolchain defaults, then proceed for real
|
---|
| 151 | in building the rest of the target Chapter 6 LFS system.</para>
|
---|
| 152 |
|
---|
| 153 | <sect2>
|
---|
| 154 | <title>Notes on static linking</title>
|
---|
| 155 |
|
---|
| 156 | <para>Most programs have to perform, beside their specific task, many rather
|
---|
[8b437193] | 157 | common and sometimes trivial operations. These include allocating memory,
|
---|
[98c95f5] | 158 | searching directories, reading and writing files, string handling, pattern
|
---|
[8b437193] | 159 | matching, arithmetic and many other tasks. Instead of obliging each program to
|
---|
[98c95f5] | 160 | reinvent the wheel, the GNU system provides all these basic functions in
|
---|
| 161 | ready-made libraries. The major library on any Linux system is
|
---|
| 162 | <emphasis>Glibc</emphasis>.</para>
|
---|
| 163 |
|
---|
| 164 | <para>There are two primary ways of linking the functions from a library to a
|
---|
| 165 | program that uses them: statically or dynamically. When a program is linked
|
---|
| 166 | statically, the code of the used functions is included in the executable,
|
---|
| 167 | resulting in a rather bulky program. When a program is dynamically linked, what
|
---|
| 168 | is included is a reference to the dynamic linker, the name of the library, and
|
---|
[1668d8e] | 169 | the name of the function, resulting in a much smaller executable. (A third way
|
---|
| 170 | is to use the programming interface of the dynamic linker. See the
|
---|
| 171 | <emphasis>dlopen</emphasis> man page for more information.)</para>
|
---|
[98c95f5] | 172 |
|
---|
| 173 | <para>Dynamic linking is the default on Linux and has three major advantages
|
---|
| 174 | over static linking. First, you need only one copy of the executable library
|
---|
| 175 | code on your hard disk, instead of having many copies of the same code included
|
---|
| 176 | into a whole bunch of programs -- thus saving disk space. Second, when several
|
---|
| 177 | programs use the same library function at the same time, only one copy of the
|
---|
| 178 | function's code is required in core -- thus saving memory space. Third, when a
|
---|
| 179 | library function gets a bug fixed or is otherwise improved, you only need to
|
---|
| 180 | recompile this one library, instead of having to recompile all the programs that
|
---|
| 181 | make use of the improved function.</para>
|
---|
| 182 |
|
---|
| 183 | <para>Why do we statically link the first two packages in Chapter 5? The reasons
|
---|
| 184 | are threefold: historical, educational and technical. Historical because earlier
|
---|
| 185 | versions of LFS statically linked every program in Chapter 5. Educational
|
---|
| 186 | because knowing the difference is useful. Technical because we gain an element
|
---|
| 187 | of independence from the host in doing so, i.e. those programs can be used
|
---|
| 188 | independently of the host system. However, it's worth noting that an overall
|
---|
| 189 | successful LFS build can still be achieved when the first two packages are built
|
---|
| 190 | dynamically.</para>
|
---|
| 191 |
|
---|
| 192 | </sect2>
|
---|
| 193 |
|
---|
| 194 | </sect1>
|
---|
| 195 |
|
---|