toolchaintechnotes.xml 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337
  1. <?xml version="1.0" encoding="ISO-8859-1"?>
  2. <!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
  3. "http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [
  4. <!ENTITY % general-entities SYSTEM "../general.ent">
  5. %general-entities;
  6. ]>
  7. <sect1 id="ch-tools-toolchaintechnotes">
  8. <?dbhtml filename="toolchaintechnotes.html"?>
  9. <title>Toolchain Technical Notes</title>
  10. <para>This section explains some of the rationale and technical details
  11. behind the overall build method. It is not essential to immediately
  12. understand everything in this section. Most of this information will be
  13. clearer after performing an actual build. This section can be referred
  14. to at any time during the process.</para>
  15. <para>The overall goal of this chapter and <xref
  16. linkend="chapter-temporary-tools"/> is to produce a temporary area that
  17. contains a known-good set of tools that can be isolated from the host system.
  18. By using <command>chroot</command>, the commands in the remaining chapters
  19. will be contained within that environment, ensuring a clean, trouble-free
  20. build of the target LFS system. The build process has been designed to
  21. minimize the risks for new readers and to provide the most educational value
  22. at the same time.</para>
  23. <para>The build process is based on the process of
  24. <emphasis>cross-compilation</emphasis>. Cross-compilation is normally used
  25. for building a compiler and its toolchain for a machine different from
  26. the one that is used for the build. This is not strictly needed for LFS,
  27. since the machine where the new system will run is the same as the one
  28. used for the build. But cross-compilation has the great advantage that
  29. anything that is cross-compiled cannot depend on the host environment.</para>
  30. <sect2 id="cross-compile" xreflabel="About Cross-Compilation">
  31. <title>About Cross-Compilation</title>
  32. <para>Cross-compilation involves some concepts that deserve a section on
  33. their own. Although this section may be omitted in a first reading, it
  34. is strongly suggested to come back to it later in order to get a full
  35. grasp of the build process.</para>
  36. <para>Let us first define some terms used in this context:</para>
  37. <variablelist>
  38. <varlistentry><term>build</term><listitem>
  39. <para>is the machine where we build programs. Note that this machine
  40. is referred to as the <quote>host</quote> in other
  41. sections.</para></listitem>
  42. </varlistentry>
  43. <varlistentry><term>host</term><listitem>
  44. <para>is the machine/system where the built programs will run. Note
  45. that this use of <quote>host</quote> is not the same as in other
  46. sections.</para></listitem>
  47. </varlistentry>
  48. <varlistentry><term>target</term><listitem>
  49. <para>is only used for compilers. It is the machine the compiler
  50. produces code for. It may be different from both build and
  51. host.</para></listitem>
  52. </varlistentry>
  53. </variablelist>
  54. <para>As an example, let us imagine the following scenario: we may have a
  55. compiler on a slow machine only, let's call the machine A, and the compiler
  56. ccA. We may have also a fast machine (B), but with no compiler, and we may
  57. want to produce code for a another slow machine (C). Then, to build a
  58. compiler for machine C, we would have three stages:</para>
  59. <informaltable align="center">
  60. <tgroup cols="5">
  61. <colspec colnum="1" align="center"/>
  62. <colspec colnum="2" align="center"/>
  63. <colspec colnum="3" align="center"/>
  64. <colspec colnum="4" align="center"/>
  65. <colspec colnum="5" align="left"/>
  66. <thead>
  67. <row><entry>Stage</entry><entry>Build</entry><entry>Host</entry>
  68. <entry>Target</entry><entry>Action</entry></row>
  69. </thead>
  70. <tbody>
  71. <row>
  72. <entry>1</entry><entry>A</entry><entry>A</entry><entry>B</entry>
  73. <entry>build cross-compiler cc1 using ccA on machine A</entry>
  74. </row>
  75. <row>
  76. <entry>2</entry><entry>A</entry><entry>B</entry><entry>B</entry>
  77. <entry>build cross-compiler cc2 using cc1 on machine A</entry>
  78. </row>
  79. <row>
  80. <entry>3</entry><entry>B</entry><entry>C</entry><entry>C</entry>
  81. <entry>build compiler ccC using cc2 on machine B</entry>
  82. </row>
  83. </tbody>
  84. </tgroup>
  85. </informaltable>
  86. <para>Then, all the other programs needed by machine C can be compiled
  87. using cc2 on the fast machine B. Note that unless B can run programs
  88. produced for C, there is no way to test the built programs until machine
  89. C itself is running. For example, for testing ccC, we may want to add a
  90. fourth stage:</para>
  91. <informaltable align="center">
  92. <tgroup cols="5">
  93. <colspec colnum="1" align="center"/>
  94. <colspec colnum="2" align="center"/>
  95. <colspec colnum="3" align="center"/>
  96. <colspec colnum="4" align="center"/>
  97. <colspec colnum="5" align="left"/>
  98. <thead>
  99. <row><entry>Stage</entry><entry>Build</entry><entry>Host</entry>
  100. <entry>Target</entry><entry>Action</entry></row>
  101. </thead>
  102. <tbody>
  103. <row>
  104. <entry>4</entry><entry>C</entry><entry>C</entry><entry>C</entry>
  105. <entry>rebuild and test ccC using itself on machine C</entry>
  106. </row>
  107. </tbody>
  108. </tgroup>
  109. </informaltable>
  110. <para>In the example above, only cc1 and cc2 are cross-compilers, that is,
  111. they produce code for a machine different from the one they are run on.
  112. The other compilers ccA and ccC produce code for the machine they are run
  113. on. Such compilers are called <emphasis>native</emphasis> compilers.</para>
  114. </sect2>
  115. <sect2 id="lfs-cross">
  116. <title>Implementation of Cross-Compilation for LFS</title>
  117. <note>
  118. <para>Almost all the build systems use names of the form
  119. cpu-vendor-kernel-os referred to as the machine triplet. An astute
  120. reader may wonder why a <quote>triplet</quote> refers to a four component
  121. name. The reason is history: initially, three component names were enough
  122. to designate unambiguously a machine, but with new machines and systems
  123. appearing, that proved insufficient. The word <quote>triplet</quote>
  124. remained. A simple way to determine your machine triplet is to run
  125. the <command>config.guess</command>
  126. script that comes with the source for many packages. Unpack the binutils
  127. sources and run the script: <userinput>./config.guess</userinput> and note
  128. the output. For example, for a 32-bit Intel processor the
  129. output will be <emphasis>i686-pc-linux-gnu</emphasis>. On a 64-bit
  130. system it will be <emphasis>x86_64-pc-linux-gnu</emphasis>.</para>
  131. <para>Also be aware of the name of the platform's dynamic linker, often
  132. referred to as the dynamic loader (not to be confused with the standard
  133. linker <command>ld</command> that is part of binutils). The dynamic linker
  134. provided by Glibc finds and loads the shared libraries needed by a
  135. program, prepares the program to run, and then runs it. The name of the
  136. dynamic linker for a 32-bit Intel machine will be <filename
  137. class="libraryfile">ld-linux.so.2</filename> (<filename
  138. class="libraryfile">ld-linux-x86-64.so.2</filename> for 64-bit systems). A
  139. sure-fire way to determine the name of the dynamic linker is to inspect a
  140. random binary from the host system by running: <userinput>readelf -l
  141. &lt;name of binary&gt; | grep interpreter</userinput> and noting the
  142. output. The authoritative reference covering all platforms is in the
  143. <filename>shlib-versions</filename> file in the root of the Glibc source
  144. tree.</para>
  145. </note>
  146. <para>In order to fake a cross compilation, the name of the host triplet
  147. is slightly adjusted by changing the &quot;vendor&quot; field in the
  148. <envar>LFS_TGT</envar> variable. We also use the
  149. <parameter>--with-sysroot</parameter> option when building the cross linker and
  150. cross compiler to tell them where to find the needed host files. This
  151. ensures that none of the other programs built in <xref
  152. linkend="chapter-temporary-tools"/> can link to libraries on the build
  153. machine. Only two stages are mandatory, and one more for tests:</para>
  154. <informaltable align="center">
  155. <tgroup cols="5">
  156. <colspec colnum="1" align="center"/>
  157. <colspec colnum="2" align="center"/>
  158. <colspec colnum="3" align="center"/>
  159. <colspec colnum="4" align="center"/>
  160. <colspec colnum="5" align="left"/>
  161. <thead>
  162. <row><entry>Stage</entry><entry>Build</entry><entry>Host</entry>
  163. <entry>Target</entry><entry>Action</entry></row>
  164. </thead>
  165. <tbody>
  166. <row>
  167. <entry>1</entry><entry>pc</entry><entry>pc</entry><entry>lfs</entry>
  168. <entry>build cross-compiler cc1 using cc-pc on pc</entry>
  169. </row>
  170. <row>
  171. <entry>2</entry><entry>pc</entry><entry>lfs</entry><entry>lfs</entry>
  172. <entry>build compiler cc-lfs using cc1 on pc</entry>
  173. </row>
  174. <row>
  175. <entry>3</entry><entry>lfs</entry><entry>lfs</entry><entry>lfs</entry>
  176. <entry>rebuild and test cc-lfs using itself on lfs</entry>
  177. </row>
  178. </tbody>
  179. </tgroup>
  180. </informaltable>
  181. <para>In the above table, <quote>on pc</quote> means the commands are run
  182. on a machine using the already installed distribution. <quote>On
  183. lfs</quote> means the commands are run in a chrooted environment.</para>
  184. <para>Now, there is more about cross-compiling: the C language is not
  185. just a compiler, but also defines a standard library. In this book, the
  186. GNU C library, named glibc, is used. This library must
  187. be compiled for the lfs machine, that is, using the cross compiler cc1.
  188. But the compiler itself uses an internal library implementing complex
  189. instructions not available in the assembler instruction set. This
  190. internal library is named libgcc, and must be linked to the glibc
  191. library to be fully functional! Furthermore, the standard library for
  192. C++ (libstdc++) also needs being linked to glibc. The solution
  193. to this chicken and egg problem is to first build a degraded cc1 based libgcc,
  194. lacking some fuctionalities such as threads and exception handling, then
  195. build glibc using this degraded compiler (glibc itself is not
  196. degraded), then build libstdc++. But this last library will lack the
  197. same functionalities as libgcc.</para>
  198. <para>This is not the end of the story: the conclusion of the preceding
  199. paragraph is that cc1 is unable to build a fully functional libstdc++, but
  200. this is the only compiler available for building the C/C++ libraries
  201. during stage 2! Of course, the compiler built during stage 2, cc-lfs,
  202. would be able to build those libraries, but (1) the build system of
  203. gcc does not know that it is usable on pc, and (2) using it on pc
  204. would be at risk of linking to the pc libraries, since cc-lfs is a native
  205. compiler. So we have to build libstdc++ later, in chroot.</para>
  206. </sect2>
  207. <sect2 id="other-details">
  208. <title>Other procedural details</title>
  209. <para>The cross-compiler will be installed in a separate <filename
  210. class="directory">$LFS/tools</filename> directory, since it will not
  211. be part of the final system.</para>
  212. <para>Binutils is installed first because the <command>configure</command>
  213. runs of both GCC and Glibc perform various feature tests on the assembler
  214. and linker to determine which software features to enable or disable. This
  215. is more important than one might first realize. An incorrectly configured
  216. GCC or Glibc can result in a subtly broken toolchain, where the impact of
  217. such breakage might not show up until near the end of the build of an
  218. entire distribution. A test suite failure will usually highlight this error
  219. before too much additional work is performed.</para>
  220. <para>Binutils installs its assembler and linker in two locations,
  221. <filename class="directory">$LFS/tools/bin</filename> and <filename
  222. class="directory">$LFS/tools/$LFS_TGT/bin</filename>. The tools in one
  223. location are hard linked to the other. An important facet of the linker is
  224. its library search order. Detailed information can be obtained from
  225. <command>ld</command> by passing it the <parameter>--verbose</parameter>
  226. flag. For example, <command>$LFS_TGT-ld --verbose | grep SEARCH</command>
  227. will illustrate the current search paths and their order. It shows which
  228. files are linked by <command>ld</command> by compiling a dummy program and
  229. passing the <parameter>--verbose</parameter> switch to the linker. For
  230. example,
  231. <command>$LFS_TGT-gcc dummy.c -Wl,--verbose 2&gt;&amp;1 | grep succeeded</command>
  232. will show all the files successfully opened during the linking.</para>
  233. <para>The next package installed is GCC. An example of what can be
  234. seen during its run of <command>configure</command> is:</para>
  235. <screen><computeroutput>checking what assembler to use... /mnt/lfs/tools/i686-lfs-linux-gnu/bin/as
  236. checking what linker to use... /mnt/lfs/tools/i686-lfs-linux-gnu/bin/ld</computeroutput></screen>
  237. <para>This is important for the reasons mentioned above. It also
  238. demonstrates that GCC's configure script does not search the PATH
  239. directories to find which tools to use. However, during the actual
  240. operation of <command>gcc</command> itself, the same search paths are not
  241. necessarily used. To find out which standard linker <command>gcc</command>
  242. will use, run: <command>$LFS_TGT-gcc -print-prog-name=ld</command>.</para>
  243. <para>Detailed information can be obtained from <command>gcc</command> by
  244. passing it the <parameter>-v</parameter> command line option while compiling
  245. a dummy program. For example, <command>gcc -v dummy.c</command> will show
  246. detailed information about the preprocessor, compilation, and assembly
  247. stages, including <command>gcc</command>'s included search paths and their
  248. order.</para>
  249. <para>Next installed are sanitized Linux API headers. These allow the
  250. standard C library (Glibc) to interface with features that the Linux
  251. kernel will provide.</para>
  252. <para>The next package installed is Glibc. The most important
  253. considerations for building Glibc are the compiler, binary tools, and
  254. kernel headers. The compiler is generally not an issue since Glibc will
  255. always use the compiler relating to the <parameter>--host</parameter>
  256. parameter passed to its configure script; e.g. in our case, the compiler
  257. will be <command>$LFS_TGT-gcc</command>. The binary tools and kernel
  258. headers can be a bit more complicated. Therefore, take no risks and use
  259. the available configure switches to enforce the correct selections. After
  260. the run of <command>configure</command>, check the contents of the
  261. <filename>config.make</filename> file in the <filename
  262. class="directory">build</filename> directory for all important details.
  263. Note the use of <parameter>CC="$LFS_TGT-gcc"</parameter> (with
  264. <envar>$LFS_TGT</envar> expanded) to control which binary tools are used
  265. and the use of the <parameter>-nostdinc</parameter> and
  266. <parameter>-isystem</parameter> flags to control the compiler's include
  267. search path. These items highlight an important aspect of the Glibc
  268. package&mdash;it is very self-sufficient in terms of its build machinery
  269. and generally does not rely on toolchain defaults.</para>
  270. <para>As said above, the standard C++ library is compiled next, followed in
  271. Chapter 6 by all the programs that need themselves to be built. The install
  272. step of libstdc++ uses the <envar>DESTDIR</envar> variable to have the
  273. programs land into the LFS filesystem.</para>
  274. <para>In Chapter 7 the native lfs compiler is built. First binutils-pass2,
  275. with the same <envar>DESTDIR</envar> install as the other programs is
  276. built, and then the second pass of GCC is constructed, omitting libstdc++
  277. and other non-important libraries. Due to some weird logic in GCC's
  278. configure script, <envar>CC_FOR_TARGET</envar> ends up as
  279. <command>cc</command> when the host is the same as the target, but is
  280. different from the build system. This is why
  281. <parameter>CC_FOR_TARGET=$LFS_TGT-gcc</parameter> is put explicitely into
  282. the configure options.</para>
  283. <para>Upon entering the chroot environment in <xref
  284. linkend="chapter-chroot-temporary-tools"/>, the first task is to install
  285. libstdc++. Then temporary installations of programs needed for the proper
  286. operation of the toolchain are performed. Programs needed for testing
  287. other programs are also built. From this point onwards, the
  288. core toolchain is self-contained and self-hosted. In
  289. <xref linkend="chapter-building-system"/>, final versions of all the
  290. packages needed for a fully functional system are built, tested and
  291. installed.</para>
  292. </sect2>
  293. </sect1>