xref: /netbsd-src/external/gpl3/gcc.old/dist/INSTALL/configure.html (revision afab4e300d3a9fb07dd8c80daf53d0feb3345706)
1<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
2<html>
3<!-- Created by GNU Texinfo 6.8, https://www.gnu.org/software/texinfo/ -->
4<head>
5<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
6<!-- Copyright (C) 1988-2020 Free Software Foundation, Inc.
7
8Permission is granted to copy, distribute and/or modify this document
9under the terms of the GNU Free Documentation License, Version 1.3 or
10any later version published by the Free Software Foundation; with no
11Invariant Sections, the Front-Cover texts being (a) (see below), and
12with the Back-Cover Texts being (b) (see below).  A copy of the
13license is included in the section entitled "GNU
14Free Documentation License".
15
16(a) The FSF's Front-Cover Text is:
17
18A GNU Manual
19
20(b) The FSF's Back-Cover Text is:
21
22You have freedom to copy and modify this GNU Manual, like GNU
23     software.  Copies published by the Free Software Foundation raise
24     funds for GNU development. -->
25<title>Installing GCC: Configuration</title>
26
27<meta name="description" content="Installing GCC: Configuration">
28<meta name="keywords" content="Installing GCC: Configuration">
29<meta name="resource-type" content="document">
30<meta name="distribution" content="global">
31<meta name="Generator" content="makeinfo">
32<meta name="viewport" content="width=device-width,initial-scale=1">
33
34<style type="text/css">
35<!--
36a.copiable-anchor {visibility: hidden; text-decoration: none; line-height: 0em}
37a.summary-letter {text-decoration: none}
38blockquote.indentedblock {margin-right: 0em}
39div.display {margin-left: 3.2em}
40div.example {margin-left: 3.2em}
41kbd {font-style: oblique}
42pre.display {font-family: inherit}
43pre.format {font-family: inherit}
44pre.menu-comment {font-family: serif}
45pre.menu-preformatted {font-family: serif}
46span.nolinebreak {white-space: nowrap}
47span.roman {font-family: initial; font-weight: normal}
48span.sansserif {font-family: sans-serif; font-weight: normal}
49span:hover a.copiable-anchor {visibility: visible}
50ul.no-bullet {list-style: none}
51-->
52</style>
53
54
55</head>
56
57<body lang="en">
58<h1 class="settitle" align="center">Installing GCC: Configuration</h1>
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78<span id="index-Configuration"></span>
79<span id="index-Installing-GCC_003a-Configuration"></span>
80
81<p>Like most GNU software, GCC must be configured before it can be built.
82This document describes the recommended configuration procedure
83for both native and cross targets.
84</p>
85<p>We use <var>srcdir</var> to refer to the toplevel source directory for
86GCC; we use <var>objdir</var> to refer to the toplevel build/object directory.
87</p>
88<p>If you obtained the sources by cloning the repository, <var>srcdir</var>
89must refer to the top <samp>gcc</samp> directory, the one where the
90<samp>MAINTAINERS</samp> file can be found, and not its <samp>gcc</samp>
91subdirectory, otherwise the build will fail.
92</p>
93<p>If either <var>srcdir</var> or <var>objdir</var> is located on an automounted NFS
94file system, the shell&rsquo;s built-in <code>pwd</code> command will return
95temporary pathnames.  Using these can lead to various sorts of build
96problems.  To avoid this issue, set the <code>PWDCMD</code> environment
97variable to an automounter-aware <code>pwd</code> command, e.g.,
98<code>pawd</code> or &lsquo;<samp>amq -w</samp>&rsquo;, during the configuration and build
99phases.
100</p>
101<p>First, we <strong>highly</strong> recommend that GCC be built into a
102separate directory from the sources which does <strong>not</strong> reside
103within the source tree.  This is how we generally build GCC; building
104where <var>srcdir</var> == <var>objdir</var> should still work, but doesn&rsquo;t
105get extensive testing; building where <var>objdir</var> is a subdirectory
106of <var>srcdir</var> is unsupported.
107</p>
108<p>If you have previously built GCC in the same directory for a
109different target machine, do &lsquo;<samp>make distclean</samp>&rsquo; to delete all files
110that might be invalid.  One of the files this deletes is <samp>Makefile</samp>;
111if &lsquo;<samp>make distclean</samp>&rsquo; complains that <samp>Makefile</samp> does not exist
112or issues a message like &ldquo;don&rsquo;t know how to make distclean&rdquo; it probably
113means that the directory is already suitably clean.  However, with the
114recommended method of building in a separate <var>objdir</var>, you should
115simply use a different <var>objdir</var> for each target.
116</p>
117<p>Second, when configuring a native system, either <code>cc</code> or
118<code>gcc</code> must be in your path or you must set <code>CC</code> in
119your environment before running configure.  Otherwise the configuration
120scripts may fail.
121</p>
122
123<p>To configure GCC:
124</p>
125<div class="example">
126<pre class="example">% mkdir <var>objdir</var>
127% cd <var>objdir</var>
128% <var>srcdir</var>/configure [<var>options</var>] [<var>target</var>]
129</pre></div>
130
131<span id="Distributor-options"></span><h3 class="heading">Distributor options</h3>
132
133<p>If you will be distributing binary versions of GCC, with modifications
134to the source code, you should use the options described in this
135section to make clear that your version contains modifications.
136</p>
137<dl compact="compact">
138<dt><span><code>--with-pkgversion=<var>version</var></code></span></dt>
139<dd><p>Specify a string that identifies your package.  You may wish
140to include a build number or build date.  This version string will be
141included in the output of <code>gcc --version</code>.  This suffix does
142not replace the default version string, only the &lsquo;<samp>GCC</samp>&rsquo; part.
143</p>
144<p>The default value is &lsquo;<samp>GCC</samp>&rsquo;.
145</p>
146</dd>
147<dt><span><code>--with-bugurl=<var>url</var></code></span></dt>
148<dd><p>Specify the URL that users should visit if they wish to report a bug.
149You are of course welcome to forward bugs reported to you to the FSF,
150if you determine that they are not bugs in your modifications.
151</p>
152<p>The default value refers to the FSF&rsquo;s GCC bug tracker.
153</p>
154</dd>
155<dt><span><code>--with-documentation-root-url=<var>url</var></code></span></dt>
156<dd><p>Specify the URL root that contains GCC option documentation.  The <var>url</var>
157should end with a <code>/</code> character.
158</p>
159<p>The default value is <a href="https://gcc.gnu.org/onlinedocs/">https://gcc.gnu.org/onlinedocs/</a>.
160</p>
161</dd>
162<dt><span><code>--with-changes-root-url=<var>url</var></code></span></dt>
163<dd><p>Specify the URL root that contains information about changes in GCC
164releases like <code>gcc-<var>version</var>/changes.html</code>.
165The <var>url</var> should end with a <code>/</code> character.
166</p>
167<p>The default value is <a href="https://gcc.gnu.org/">https://gcc.gnu.org/</a>.
168</p>
169</dd>
170</dl>
171
172<span id="Target-specification"></span><h3 class="heading">Target specification</h3>
173<ul>
174<li> GCC has code to correctly determine the correct value for <var>target</var>
175for nearly all native systems.  Therefore, we highly recommend you do
176not provide a configure target when configuring a native compiler.
177
178</li><li> <var>target</var> must be specified as <samp>--target=<var>target</var></samp>
179when configuring a cross compiler; examples of valid targets would be
180m68k-elf, sh-elf, etc.
181
182</li><li> Specifying just <var>target</var> instead of <samp>--target=<var>target</var></samp>
183implies that the host defaults to <var>target</var>.
184</li></ul>
185
186
187<span id="Options-specification"></span><h3 class="heading">Options specification</h3>
188
189<p>Use <var>options</var> to override several configure time options for
190GCC.  A list of supported <var>options</var> follows; &lsquo;<samp>configure
191--help</samp>&rsquo; may list other options, but those not listed below may not
192work and should not normally be used.
193</p>
194<p>Note that each <samp>--enable</samp> option has a corresponding
195<samp>--disable</samp> option and that each <samp>--with</samp> option has a
196corresponding <samp>--without</samp> option.
197</p>
198<dl compact="compact">
199<dt><span><code>--prefix=<var>dirname</var></code></span></dt>
200<dd><p>Specify the toplevel installation
201directory.  This is the recommended way to install the tools into a directory
202other than the default.  The toplevel installation directory defaults to
203<samp>/usr/local</samp>.
204</p>
205<p>We <strong>highly</strong> recommend against <var>dirname</var> being the same or a
206subdirectory of <var>objdir</var> or vice versa.  If specifying a directory
207beneath a user&rsquo;s home directory tree, some shells will not expand
208<var>dirname</var> correctly if it contains the &lsquo;<samp>~</samp>&rsquo; metacharacter; use
209<code>$HOME</code> instead.
210</p>
211<p>The following standard <code>autoconf</code> options are supported.  Normally you
212should not need to use these options.
213</p><dl compact="compact">
214<dt><span><code>--exec-prefix=<var>dirname</var></code></span></dt>
215<dd><p>Specify the toplevel installation directory for architecture-dependent
216files.  The default is <samp><var>prefix</var></samp>.
217</p>
218</dd>
219<dt><span><code>--bindir=<var>dirname</var></code></span></dt>
220<dd><p>Specify the installation directory for the executables called by users
221(such as <code>gcc</code> and <code>g++</code>).  The default is
222<samp><var>exec-prefix</var>/bin</samp>.
223</p>
224</dd>
225<dt><span><code>--libdir=<var>dirname</var></code></span></dt>
226<dd><p>Specify the installation directory for object code libraries and
227internal data files of GCC.  The default is <samp><var>exec-prefix</var>/lib</samp>.
228</p>
229</dd>
230<dt><span><code>--libexecdir=<var>dirname</var></code></span></dt>
231<dd><p>Specify the installation directory for internal executables of GCC.
232The default is <samp><var>exec-prefix</var>/libexec</samp>.
233</p>
234</dd>
235<dt><span><code>--with-slibdir=<var>dirname</var></code></span></dt>
236<dd><p>Specify the installation directory for the shared libgcc library.  The
237default is <samp><var>libdir</var></samp>.
238</p>
239</dd>
240<dt><span><code>--datarootdir=<var>dirname</var></code></span></dt>
241<dd><p>Specify the root of the directory tree for read-only architecture-independent
242data files referenced by GCC.  The default is <samp><var>prefix</var>/share</samp>.
243</p>
244</dd>
245<dt><span><code>--infodir=<var>dirname</var></code></span></dt>
246<dd><p>Specify the installation directory for documentation in info format.
247The default is <samp><var>datarootdir</var>/info</samp>.
248</p>
249</dd>
250<dt><span><code>--datadir=<var>dirname</var></code></span></dt>
251<dd><p>Specify the installation directory for some architecture-independent
252data files referenced by GCC.  The default is <samp><var>datarootdir</var></samp>.
253</p>
254</dd>
255<dt><span><code>--docdir=<var>dirname</var></code></span></dt>
256<dd><p>Specify the installation directory for documentation files (other
257than Info) for GCC.  The default is <samp><var>datarootdir</var>/doc</samp>.
258</p>
259</dd>
260<dt><span><code>--htmldir=<var>dirname</var></code></span></dt>
261<dd><p>Specify the installation directory for HTML documentation files.
262The default is <samp><var>docdir</var></samp>.
263</p>
264</dd>
265<dt><span><code>--pdfdir=<var>dirname</var></code></span></dt>
266<dd><p>Specify the installation directory for PDF documentation files.
267The default is <samp><var>docdir</var></samp>.
268</p>
269</dd>
270<dt><span><code>--mandir=<var>dirname</var></code></span></dt>
271<dd><p>Specify the installation directory for manual pages.  The default is
272<samp><var>datarootdir</var>/man</samp>.  (Note that the manual pages are only extracts
273from the full GCC manuals, which are provided in Texinfo format.  The manpages
274are derived by an automatic conversion process from parts of the full
275manual.)
276</p>
277</dd>
278<dt><span><code>--with-gxx-include-dir=<var>dirname</var></code></span></dt>
279<dd><p>Specify
280the installation directory for G++ header files.  The default depends
281on other configuration options, and differs between cross and native
282configurations.
283</p>
284</dd>
285<dt><span><code>--with-specs=<var>specs</var></code></span></dt>
286<dd><p>Specify additional command line driver SPECS.
287This can be useful if you need to turn on a non-standard feature by
288default without modifying the compiler&rsquo;s source code, for instance
289<samp>--with-specs=%{!fcommon:%{!fno-common:-fno-common}}</samp>.
290See &ldquo;Spec Files&rdquo; in the main manual
291</p>
292</dd>
293</dl>
294
295</dd>
296<dt><span><code>--program-prefix=<var>prefix</var></code></span></dt>
297<dd><p>GCC supports some transformations of the names of its programs when
298installing them.  This option prepends <var>prefix</var> to the names of
299programs to install in <var>bindir</var> (see above).  For example, specifying
300<samp>--program-prefix=foo-</samp> would result in &lsquo;<samp>gcc</samp>&rsquo;
301being installed as <samp>/usr/local/bin/foo-gcc</samp>.
302</p>
303</dd>
304<dt><span><code>--program-suffix=<var>suffix</var></code></span></dt>
305<dd><p>Appends <var>suffix</var> to the names of programs to install in <var>bindir</var>
306(see above).  For example, specifying <samp>--program-suffix=-3.1</samp>
307would result in &lsquo;<samp>gcc</samp>&rsquo; being installed as
308<samp>/usr/local/bin/gcc-3.1</samp>.
309</p>
310</dd>
311<dt><span><code>--program-transform-name=<var>pattern</var></code></span></dt>
312<dd><p>Applies the &lsquo;<samp>sed</samp>&rsquo; script <var>pattern</var> to be applied to the names
313of programs to install in <var>bindir</var> (see above).  <var>pattern</var> has to
314consist of one or more basic &lsquo;<samp>sed</samp>&rsquo; editing commands, separated by
315semicolons.  For example, if you want the &lsquo;<samp>gcc</samp>&rsquo; program name to be
316transformed to the installed program <samp>/usr/local/bin/myowngcc</samp> and
317the &lsquo;<samp>g++</samp>&rsquo; program name to be transformed to
318<samp>/usr/local/bin/gspecial++</samp> without changing other program names,
319you could use the pattern
320<samp>--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'</samp>
321to achieve this effect.
322</p>
323<p>All three options can be combined and used together, resulting in more
324complex conversion patterns.  As a basic rule, <var>prefix</var> (and
325<var>suffix</var>) are prepended (appended) before further transformations
326can happen with a special transformation script <var>pattern</var>.
327</p>
328<p>As currently implemented, this option only takes effect for native
329builds; cross compiler binaries&rsquo; names are not transformed even when a
330transformation is explicitly asked for by one of these options.
331</p>
332<p>For native builds, some of the installed programs are also installed
333with the target alias in front of their name, as in
334&lsquo;<samp>i686-pc-linux-gnu-gcc</samp>&rsquo;.  All of the above transformations happen
335before the target alias is prepended to the name&mdash;so, specifying
336<samp>--program-prefix=foo-</samp> and <samp>program-suffix=-3.1</samp>, the
337resulting binary would be installed as
338<samp>/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1</samp>.
339</p>
340<p>As a last shortcoming, none of the installed Ada programs are
341transformed yet, which will be fixed in some time.
342</p>
343</dd>
344<dt><span><code>--with-local-prefix=<var>dirname</var></code></span></dt>
345<dd><p>Specify the
346installation directory for local include files.  The default is
347<samp>/usr/local</samp>.  Specify this option if you want the compiler to
348search directory <samp><var>dirname</var>/include</samp> for locally installed
349header files <em>instead</em> of <samp>/usr/local/include</samp>.
350</p>
351<p>You should specify <samp>--with-local-prefix</samp> <strong>only</strong> if your
352site has a different convention (not <samp>/usr/local</samp>) for where to put
353site-specific files.
354</p>
355<p>The default value for <samp>--with-local-prefix</samp> is <samp>/usr/local</samp>
356regardless of the value of <samp>--prefix</samp>.  Specifying
357<samp>--prefix</samp> has no effect on which directory GCC searches for
358local header files.  This may seem counterintuitive, but actually it is
359logical.
360</p>
361<p>The purpose of <samp>--prefix</samp> is to specify where to <em>install
362GCC</em>.  The local header files in <samp>/usr/local/include</samp>&mdash;if you put
363any in that directory&mdash;are not part of GCC.  They are part of other
364programs&mdash;perhaps many others.  (GCC installs its own header files in
365another directory which is based on the <samp>--prefix</samp> value.)
366</p>
367<p>Both the local-prefix include directory and the GCC-prefix include
368directory are part of GCC&rsquo;s &ldquo;system include&rdquo; directories.  Although these
369two directories are not fixed, they need to be searched in the proper
370order for the correct processing of the include_next directive.  The
371local-prefix include directory is searched before the GCC-prefix
372include directory.  Another characteristic of system include directories
373is that pedantic warnings are turned off for headers in these directories.
374</p>
375<p>Some autoconf macros add <samp>-I <var>directory</var></samp> options to the
376compiler command line, to ensure that directories containing installed
377packages&rsquo; headers are searched.  When <var>directory</var> is one of GCC&rsquo;s
378system include directories, GCC will ignore the option so that system
379directories continue to be processed in the correct order.  This
380may result in a search order different from what was specified but the
381directory will still be searched.
382</p>
383<p>GCC automatically searches for ordinary libraries using
384<code>GCC_EXEC_PREFIX</code>.  Thus, when the same installation prefix is
385used for both GCC and packages, GCC will automatically search for
386both headers and libraries.  This provides a configuration that is
387easy to use.  GCC behaves in a manner similar to that when it is
388installed as a system compiler in <samp>/usr</samp>.
389</p>
390<p>Sites that need to install multiple versions of GCC may not want to
391use the above simple configuration.  It is possible to use the
392<samp>--program-prefix</samp>, <samp>--program-suffix</samp> and
393<samp>--program-transform-name</samp> options to install multiple versions
394into a single directory, but it may be simpler to use different prefixes
395and the <samp>--with-local-prefix</samp> option to specify the location of the
396site-specific files for each version.  It will then be necessary for
397users to specify explicitly the location of local site libraries
398(e.g., with <code>LIBRARY_PATH</code>).
399</p>
400<p>The same value can be used for both <samp>--with-local-prefix</samp> and
401<samp>--prefix</samp> provided it is not <samp>/usr</samp>.  This can be used
402to avoid the default search of <samp>/usr/local/include</samp>.
403</p>
404<p><strong>Do not</strong> specify <samp>/usr</samp> as the <samp>--with-local-prefix</samp>!
405The directory you use for <samp>--with-local-prefix</samp> <strong>must not</strong>
406contain any of the system&rsquo;s standard header files.  If it did contain
407them, certain programs would be miscompiled (including GNU Emacs, on
408certain targets), because this would override and nullify the header
409file corrections made by the <code>fixincludes</code> script.
410</p>
411<p>Indications are that people who use this option use it based on mistaken
412ideas of what it is for.  People use it as if it specified where to
413install part of GCC.  Perhaps they make this assumption because
414installing GCC creates the directory.
415</p>
416</dd>
417<dt><span><code>--with-gcc-major-version-only</code></span></dt>
418<dd><p>Specifies that GCC should use only the major number rather than
419<var>major</var>.<var>minor</var>.<var>patchlevel</var> in filesystem paths.
420</p>
421</dd>
422<dt><span><code>--with-native-system-header-dir=<var>dirname</var></code></span></dt>
423<dd><p>Specifies that <var>dirname</var> is the directory that contains native system
424header files, rather than <samp>/usr/include</samp>.  This option is most useful
425if you are creating a compiler that should be isolated from the system
426as much as possible.  It is most commonly used with the
427<samp>--with-sysroot</samp> option and will cause GCC to search
428<var>dirname</var> inside the system root specified by that option.
429</p>
430</dd>
431<dt><span><code>--enable-shared[=<var>package</var>[,&hellip;]]</code></span></dt>
432<dd><p>Build shared versions of libraries, if shared libraries are supported on
433the target platform.  Unlike GCC 2.95.x and earlier, shared libraries
434are enabled by default on all platforms that support shared libraries.
435</p>
436<p>If a list of packages is given as an argument, build shared libraries
437only for the listed packages.  For other packages, only static libraries
438will be built.  Package names currently recognized in the GCC tree are
439&lsquo;<samp>libgcc</samp>&rsquo; (also known as &lsquo;<samp>gcc</samp>&rsquo;), &lsquo;<samp>libstdc++</samp>&rsquo; (not
440&lsquo;<samp>libstdc++-v3</samp>&rsquo;), &lsquo;<samp>libffi</samp>&rsquo;, &lsquo;<samp>zlib</samp>&rsquo;, &lsquo;<samp>boehm-gc</samp>&rsquo;,
441&lsquo;<samp>ada</samp>&rsquo;, &lsquo;<samp>libada</samp>&rsquo;, &lsquo;<samp>libgo</samp>&rsquo;, &lsquo;<samp>libobjc</samp>&rsquo;, and &lsquo;<samp>libphobos</samp>&rsquo;.
442Note &lsquo;<samp>libiberty</samp>&rsquo; does not support shared libraries at all.
443</p>
444<p>Use <samp>--disable-shared</samp> to build only static libraries.  Note that
445<samp>--disable-shared</samp> does not accept a list of package names as
446argument, only <samp>--enable-shared</samp> does.
447</p>
448<p>Contrast with <samp>--enable-host-shared</samp>, which affects <em>host</em>
449code.
450</p>
451</dd>
452<dt><span><code>--enable-host-shared</code></span></dt>
453<dd><p>Specify that the <em>host</em> code should be built into position-independent
454machine code (with -fPIC), allowing it to be used within shared libraries,
455but yielding a slightly slower compiler.
456</p>
457<p>This option is required when building the libgccjit.so library.
458</p>
459<p>Contrast with <samp>--enable-shared</samp>, which affects <em>target</em>
460libraries.
461</p>
462</dd>
463<dt><span><code><span id="with-gnu-as"></span>--with-gnu-as</code></span></dt>
464<dd><p>Specify that the compiler should assume that the
465assembler it finds is the GNU assembler.  However, this does not modify
466the rules to find an assembler and will result in confusion if the
467assembler found is not actually the GNU assembler.  (Confusion may also
468result if the compiler finds the GNU assembler but has not been
469configured with <samp>--with-gnu-as</samp>.)  If you have more than one
470assembler installed on your system, you may want to use this option in
471connection with <samp>--with-as=<var>pathname</var></samp> or
472<samp>--with-build-time-tools=<var>pathname</var></samp>.
473</p>
474<p>The following systems are the only ones where it makes a difference
475whether you use the GNU assembler.  On any other system,
476<samp>--with-gnu-as</samp> has no effect.
477</p>
478<ul>
479<li> &lsquo;<samp>hppa1.0-<var>any</var>-<var>any</var></samp>&rsquo;
480</li><li> &lsquo;<samp>hppa1.1-<var>any</var>-<var>any</var></samp>&rsquo;
481</li><li> &lsquo;<samp>sparc-sun-solaris2.<var>any</var></samp>&rsquo;
482</li><li> &lsquo;<samp>sparc64-<var>any</var>-solaris2.<var>any</var></samp>&rsquo;
483</li></ul>
484
485</dd>
486<dt><span><code><span id="with-as"></span>--with-as=<var>pathname</var></code></span></dt>
487<dd><p>Specify that the compiler should use the assembler pointed to by
488<var>pathname</var>, rather than the one found by the standard rules to find
489an assembler, which are:
490</p><ul>
491<li> Unless GCC is being built with a cross compiler, check the
492<samp><var>libexec</var>/gcc/<var>target</var>/<var>version</var></samp> directory.
493<var>libexec</var> defaults to <samp><var>exec-prefix</var>/libexec</samp>;
494<var>exec-prefix</var> defaults to <var>prefix</var>, which
495defaults to <samp>/usr/local</samp> unless overridden by the
496<samp>--prefix=<var>pathname</var></samp> switch described above.  <var>target</var>
497is the target system triple, such as &lsquo;<samp>sparc-sun-solaris2.7</samp>&rsquo;, and
498<var>version</var> denotes the GCC version, such as 3.0.
499
500</li><li> If the target system is the same that you are building on, check
501operating system specific directories (e.g. <samp>/usr/ccs/bin</samp> on
502Solaris 2).
503
504</li><li> Check in the <code>PATH</code> for a tool whose name is prefixed by the
505target system triple.
506
507</li><li> Check in the <code>PATH</code> for a tool whose name is not prefixed by the
508target system triple, if the host and target system triple are
509the same (in other words, we use a host tool if it can be used for
510the target as well).
511</li></ul>
512
513<p>You may want to use <samp>--with-as</samp> if no assembler
514is installed in the directories listed above, or if you have multiple
515assemblers installed and want to choose one that is not found by the
516above rules.
517</p>
518</dd>
519<dt><span><code><span id="with-gnu-ld"></span>--with-gnu-ld</code></span></dt>
520<dd><p>Same as <a href="#with-gnu-as"><samp>--with-gnu-as</samp></a>
521but for the linker.
522</p>
523</dd>
524<dt><span><code>--with-ld=<var>pathname</var></code></span></dt>
525<dd><p>Same as <a href="#with-as"><samp>--with-as</samp></a>
526but for the linker.
527</p>
528</dd>
529<dt><span><code>--with-dsymutil=<var>pathname</var></code></span></dt>
530<dd><p>Same as <a href="#with-as"><samp>--with-as</samp></a>
531but for the debug linker (only used on Darwin platforms so far).
532</p>
533</dd>
534<dt><span><code>--with-stabs</code></span></dt>
535<dd><p>Specify that stabs debugging
536information should be used instead of whatever format the host normally
537uses.  Normally GCC uses the same debug format as the host system.
538</p>
539</dd>
540<dt><span><code>--with-tls=<var>dialect</var></code></span></dt>
541<dd><p>Specify the default TLS dialect, for systems were there is a choice.
542For ARM targets, possible values for <var>dialect</var> are <code>gnu</code> or
543<code>gnu2</code>, which select between the original GNU dialect and the GNU TLS
544descriptor-based dialect.
545</p>
546</dd>
547<dt><span><code>--enable-multiarch</code></span></dt>
548<dd><p>Specify whether to enable or disable multiarch support.  The default is
549to check for glibc start files in a multiarch location, and enable it
550if the files are found.  The auto detection is enabled for native builds,
551and for cross builds configured with <samp>--with-sysroot</samp>, and without
552<samp>--with-native-system-header-dir</samp>.
553More documentation about multiarch can be found at
554<a href="https://wiki.debian.org/Multiarch">https://wiki.debian.org/Multiarch</a>.
555</p>
556</dd>
557<dt><span><code>--enable-sjlj-exceptions</code></span></dt>
558<dd><p>Force use of the <code>setjmp</code>/<code>longjmp</code>-based scheme for exceptions.
559&lsquo;<samp>configure</samp>&rsquo; ordinarily picks the correct value based on the platform.
560Only use this option if you are sure you need a different setting.
561</p>
562</dd>
563<dt><span><code>--enable-vtable-verify</code></span></dt>
564<dd><p>Specify whether to enable or disable the vtable verification feature.
565Enabling this feature causes libstdc++ to be built with its virtual calls
566in verifiable mode.  This means that, when linked with libvtv, every
567virtual call in libstdc++ will verify the vtable pointer through which the
568call will be made before actually making the call.  If not linked with libvtv,
569the verifier will call stub functions (in libstdc++ itself) and do nothing.
570If vtable verification is disabled, then libstdc++ is not built with its
571virtual calls in verifiable mode at all.  However the libvtv library will
572still be built (see <samp>--disable-libvtv</samp> to turn off building libvtv).
573<samp>--disable-vtable-verify</samp> is the default.
574</p>
575</dd>
576<dt><span><code>--disable-gcov</code></span></dt>
577<dd><p>Specify that the run-time library used for coverage analysis
578and associated host tools should not be built.
579</p>
580</dd>
581<dt><span><code>--disable-multilib</code></span></dt>
582<dd><p>Specify that multiple target
583libraries to support different target variants, calling
584conventions, etc. should not be built.  The default is to build a
585predefined set of them.
586</p>
587<p>Some targets provide finer-grained control over which multilibs are built
588(e.g., <samp>--disable-softfloat</samp>):
589</p><dl compact="compact">
590<dt><span><code>arm-*-*</code></span></dt>
591<dd><p>fpu, 26bit, underscore, interwork, biendian, nofmult.
592</p>
593</dd>
594<dt><span><code>m68*-*-*</code></span></dt>
595<dd><p>softfloat, m68881, m68000, m68020.
596</p>
597</dd>
598<dt><span><code>mips*-*-*</code></span></dt>
599<dd><p>single-float, biendian, softfloat.
600</p>
601</dd>
602<dt><span><code>msp430-*-*</code></span></dt>
603<dd><p>no-exceptions
604</p>
605</dd>
606<dt><span><code>powerpc*-*-*, rs6000*-*-*</code></span></dt>
607<dd><p>aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian,
608sysv, aix.
609</p>
610</dd>
611</dl>
612
613</dd>
614<dt><span><code>--with-multilib-list=<var>list</var></code></span></dt>
615<dt><span><code>--without-multilib-list</code></span></dt>
616<dd><p>Specify what multilibs to build.  <var>list</var> is a comma separated list of
617values, possibly consisting of a single value.  Currently only implemented
618for aarch64*-*-*, arm*-*-*, riscv*-*-*, sh*-*-* and x86-64-*-linux*.  The
619accepted values and meaning for each target is given below.
620</p>
621<dl compact="compact">
622<dt><span><code>aarch64*-*-*</code></span></dt>
623<dd><p><var>list</var> is a comma separated list of <code>ilp32</code>, and <code>lp64</code>
624to enable ILP32 and LP64 run-time libraries, respectively.  If
625<var>list</var> is empty, then there will be no multilibs and only the
626default run-time library will be built.  If <var>list</var> is
627<code>default</code> or &ndash;with-multilib-list= is not specified, then the
628default set of libraries is selected based on the value of
629<samp>--target</samp>.
630</p>
631</dd>
632<dt><span><code>arm*-*-*</code></span></dt>
633<dd><p><var>list</var> is a comma separated list of <code>aprofile</code> and
634<code>rmprofile</code> to build multilibs for A or R and M architecture
635profiles respectively.  Note that, due to some limitation of the current
636multilib framework, using the combined <code>aprofile,rmprofile</code>
637multilibs selects in some cases a less optimal multilib than when using
638the multilib profile for the architecture targetted.  The special value
639<code>default</code> is also accepted and is equivalent to omitting the
640option, i.e., only the default run-time library will be enabled.
641</p>
642<p><var>list</var> may instead contain <code>@name</code>, to use the multilib
643configuration Makefile fragment <samp>name</samp> in <samp>gcc/config/arm</samp> in
644the source tree (it is part of the corresponding sources, after all).
645It is recommended, but not required, that files used for this purpose to
646be named starting with <samp>t-ml-</samp>, to make their intended purpose
647self-evident, in line with GCC conventions.  Such files enable custom,
648user-chosen multilib lists to be configured.  Whether multiple such
649files can be used together depends on the contents of the supplied
650files.  See <samp>gcc/config/arm/t-multilib</samp> and its supplementary
651<samp>gcc/config/arm/t-*profile</samp> files for an example of what such
652Makefile fragments might look like for this version of GCC.  The macros
653expected to be defined in these fragments are not stable across GCC
654releases, so make sure they define the <code>MULTILIB</code>-related macros
655expected by the version of GCC you are building.
656See &ldquo;Target Makefile Fragments&rdquo; in the internals manual.
657</p>
658<p>The table below gives the combination of ISAs, architectures, FPUs and
659floating-point ABIs for which multilibs are built for each predefined
660profile.  The union of these options is considered when specifying both
661<code>aprofile</code> and <code>rmprofile</code>.
662</p>
663<table>
664<tr><td width="15%">Option</td><td width="28%">aprofile</td><td width="30%">rmprofile</td></tr>
665<tr><td width="15%">ISAs</td><td width="28%"><code>-marm</code> and <code>-mthumb</code></td><td width="30%"><code>-mthumb</code></td></tr>
666<tr><td width="15%">Architectures<br><br><br><br><br><br></td><td width="28%">default architecture<br>
667<code>-march=armv7-a</code><br>
668<code>-march=armv7ve</code><br>
669<code>-march=armv8-a</code><br><br><br></td><td width="30%">default architecture<br>
670<code>-march=armv6s-m</code><br>
671<code>-march=armv7-m</code><br>
672<code>-march=armv7e-m</code><br>
673<code>-march=armv8-m.base</code><br>
674<code>-march=armv8-m.main</code><br>
675<code>-march=armv7</code></td></tr>
676<tr><td width="15%">FPUs<br><br><br><br><br></td><td width="28%">none<br>
677<code>-mfpu=vfpv3-d16</code><br>
678<code>-mfpu=neon</code><br>
679<code>-mfpu=vfpv4-d16</code><br>
680<code>-mfpu=neon-vfpv4</code><br>
681<code>-mfpu=neon-fp-armv8</code></td><td width="30%">none<br>
682<code>-mfpu=vfpv3-d16</code><br>
683<code>-mfpu=fpv4-sp-d16</code><br>
684<code>-mfpu=fpv5-sp-d16</code><br>
685<code>-mfpu=fpv5-d16</code><br></td></tr>
686<tr><td width="15%">floating-point ABIs<br><br></td><td width="28%"><code>-mfloat-abi=soft</code><br>
687<code>-mfloat-abi=softfp</code><br>
688<code>-mfloat-abi=hard</code></td><td width="30%"><code>-mfloat-abi=soft</code><br>
689<code>-mfloat-abi=softfp</code><br>
690<code>-mfloat-abi=hard</code></td></tr>
691</table>
692
693</dd>
694<dt><span><code>riscv*-*-*</code></span></dt>
695<dd><p><var>list</var> is a single ABI name.  The target architecture must be either
696<code>rv32gc</code> or <code>rv64gc</code>.  This will build a single multilib for the
697specified architecture and ABI pair.  If <code>--with-multilib-list</code> is not
698given, then a default set of multilibs is selected based on the value of
699<samp>--target</samp>.  This is usually a large set of multilibs.
700</p>
701</dd>
702<dt><span><code>sh*-*-*</code></span></dt>
703<dd><p><var>list</var> is a comma separated list of CPU names.  These must be of the
704form <code>sh*</code> or <code>m*</code> (in which case they match the compiler option
705for that processor).  The list should not contain any endian options -
706these are handled by <samp>--with-endian</samp>.
707</p>
708<p>If <var>list</var> is empty, then there will be no multilibs for extra
709processors.  The multilib for the secondary endian remains enabled.
710</p>
711<p>As a special case, if an entry in the list starts with a <code>!</code>
712(exclamation point), then it is added to the list of excluded multilibs.
713Entries of this sort should be compatible with &lsquo;<samp>MULTILIB_EXCLUDES</samp>&rsquo;
714(once the leading <code>!</code> has been stripped).
715</p>
716<p>If <samp>--with-multilib-list</samp> is not given, then a default set of
717multilibs is selected based on the value of <samp>--target</samp>.  This is
718usually the complete set of libraries, but some targets imply a more
719specialized subset.
720</p>
721<p>Example 1: to configure a compiler for SH4A only, but supporting both
722endians, with little endian being the default:
723</p><div class="example">
724<pre class="example">--with-cpu=sh4a --with-endian=little,big --with-multilib-list=
725</pre></div>
726
727<p>Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with
728only little endian SH4AL:
729</p><div class="example">
730<pre class="example">--with-cpu=sh4a --with-endian=little,big \
731--with-multilib-list=sh4al,!mb/m4al
732</pre></div>
733
734</dd>
735<dt><span><code>x86-64-*-linux*</code></span></dt>
736<dd><p><var>list</var> is a comma separated list of <code>m32</code>, <code>m64</code> and
737<code>mx32</code> to enable 32-bit, 64-bit and x32 run-time libraries,
738respectively.  If <var>list</var> is empty, then there will be no multilibs
739and only the default run-time library will be enabled.
740</p>
741<p>If <samp>--with-multilib-list</samp> is not given, then only 32-bit and
74264-bit run-time libraries will be enabled.
743</p></dd>
744</dl>
745
746</dd>
747<dt><span><code>--with-endian=<var>endians</var></code></span></dt>
748<dd><p>Specify what endians to use.
749Currently only implemented for sh*-*-*.
750</p>
751<p><var>endians</var> may be one of the following:
752</p><dl compact="compact">
753<dt><span><code>big</code></span></dt>
754<dd><p>Use big endian exclusively.
755</p></dd>
756<dt><span><code>little</code></span></dt>
757<dd><p>Use little endian exclusively.
758</p></dd>
759<dt><span><code>big,little</code></span></dt>
760<dd><p>Use big endian by default.  Provide a multilib for little endian.
761</p></dd>
762<dt><span><code>little,big</code></span></dt>
763<dd><p>Use little endian by default.  Provide a multilib for big endian.
764</p></dd>
765</dl>
766
767</dd>
768<dt><span><code>--enable-threads</code></span></dt>
769<dd><p>Specify that the target
770supports threads.  This affects the Objective-C compiler and runtime
771library, and exception handling for other languages like C++.
772On some systems, this is the default.
773</p>
774<p>In general, the best (and, in many cases, the only known) threading
775model available will be configured for use.  Beware that on some
776systems, GCC has not been taught what threading models are generally
777available for the system.  In this case, <samp>--enable-threads</samp> is an
778alias for <samp>--enable-threads=single</samp>.
779</p>
780</dd>
781<dt><span><code>--disable-threads</code></span></dt>
782<dd><p>Specify that threading support should be disabled for the system.
783This is an alias for <samp>--enable-threads=single</samp>.
784</p>
785</dd>
786<dt><span><code>--enable-threads=<var>lib</var></code></span></dt>
787<dd><p>Specify that
788<var>lib</var> is the thread support library.  This affects the Objective-C
789compiler and runtime library, and exception handling for other languages
790like C++.  The possibilities for <var>lib</var> are:
791</p>
792<dl compact="compact">
793<dt><span><code>aix</code></span></dt>
794<dd><p>AIX thread support.
795</p></dd>
796<dt><span><code>dce</code></span></dt>
797<dd><p>DCE thread support.
798</p></dd>
799<dt><span><code>lynx</code></span></dt>
800<dd><p>LynxOS thread support.
801</p></dd>
802<dt><span><code>mipssde</code></span></dt>
803<dd><p>MIPS SDE thread support.
804</p></dd>
805<dt><span><code>no</code></span></dt>
806<dd><p>This is an alias for &lsquo;<samp>single</samp>&rsquo;.
807</p></dd>
808<dt><span><code>posix</code></span></dt>
809<dd><p>Generic POSIX/Unix98 thread support.
810</p></dd>
811<dt><span><code>rtems</code></span></dt>
812<dd><p>RTEMS thread support.
813</p></dd>
814<dt><span><code>single</code></span></dt>
815<dd><p>Disable thread support, should work for all platforms.
816</p></dd>
817<dt><span><code>tpf</code></span></dt>
818<dd><p>TPF thread support.
819</p></dd>
820<dt><span><code>vxworks</code></span></dt>
821<dd><p>VxWorks thread support.
822</p></dd>
823<dt><span><code>win32</code></span></dt>
824<dd><p>Microsoft Win32 API thread support.
825</p></dd>
826</dl>
827
828</dd>
829<dt><span><code>--enable-tls</code></span></dt>
830<dd><p>Specify that the target supports TLS (Thread Local Storage).  Usually
831configure can correctly determine if TLS is supported.  In cases where
832it guesses incorrectly, TLS can be explicitly enabled or disabled with
833<samp>--enable-tls</samp> or <samp>--disable-tls</samp>.  This can happen if
834the assembler supports TLS but the C library does not, or if the
835assumptions made by the configure test are incorrect.
836</p>
837</dd>
838<dt><span><code>--disable-tls</code></span></dt>
839<dd><p>Specify that the target does not support TLS.
840This is an alias for <samp>--enable-tls=no</samp>.
841</p>
842</dd>
843<dt><span><code>--disable-tm-clone-registry</code></span></dt>
844<dd><p>Disable TM clone registry in libgcc. It is enabled in libgcc by default.
845This option helps to reduce code size for embedded targets which do
846not use transactional memory.
847</p>
848</dd>
849<dt><span><code>--with-cpu=<var>cpu</var></code></span></dt>
850<dt><span><code>--with-cpu-32=<var>cpu</var></code></span></dt>
851<dt><span><code>--with-cpu-64=<var>cpu</var></code></span></dt>
852<dd><p>Specify which cpu variant the compiler should generate code for by default.
853<var>cpu</var> will be used as the default value of the <samp>-mcpu=</samp> switch.
854This option is only supported on some targets, including ARC, ARM, i386, M68k,
855PowerPC, and SPARC.  It is mandatory for ARC.  The <samp>--with-cpu-32</samp> and
856<samp>--with-cpu-64</samp> options specify separate default CPUs for
85732-bit and 64-bit modes; these options are only supported for aarch64, i386,
858x86-64, PowerPC, and SPARC.
859</p>
860</dd>
861<dt><span><code>--with-schedule=<var>cpu</var></code></span></dt>
862<dt><span><code>--with-arch=<var>cpu</var></code></span></dt>
863<dt><span><code>--with-arch-32=<var>cpu</var></code></span></dt>
864<dt><span><code>--with-arch-64=<var>cpu</var></code></span></dt>
865<dt><span><code>--with-tune=<var>cpu</var></code></span></dt>
866<dt><span><code>--with-tune-32=<var>cpu</var></code></span></dt>
867<dt><span><code>--with-tune-64=<var>cpu</var></code></span></dt>
868<dt><span><code>--with-abi=<var>abi</var></code></span></dt>
869<dt><span><code>--with-fpu=<var>type</var></code></span></dt>
870<dt><span><code>--with-float=<var>type</var></code></span></dt>
871<dd><p>These configure options provide default values for the <samp>-mschedule=</samp>,
872<samp>-march=</samp>, <samp>-mtune=</samp>, <samp>-mabi=</samp>, and <samp>-mfpu=</samp>
873options and for <samp>-mhard-float</samp> or <samp>-msoft-float</samp>.  As with
874<samp>--with-cpu</samp>, which switches will be accepted and acceptable values
875of the arguments depend on the target.
876</p>
877</dd>
878<dt><span><code>--with-mode=<var>mode</var></code></span></dt>
879<dd><p>Specify if the compiler should default to <samp>-marm</samp> or <samp>-mthumb</samp>.
880This option is only supported on ARM targets.
881</p>
882</dd>
883<dt><span><code>--with-stack-offset=<var>num</var></code></span></dt>
884<dd><p>This option sets the default for the -mstack-offset=<var>num</var> option,
885and will thus generally also control the setting of this option for
886libraries.  This option is only supported on Epiphany targets.
887</p>
888</dd>
889<dt><span><code>--with-fpmath=<var>isa</var></code></span></dt>
890<dd><p>This options sets <samp>-mfpmath=sse</samp> by default and specifies the default
891ISA for floating-point arithmetics.  You can select either &lsquo;<samp>sse</samp>&rsquo; which
892enables <samp>-msse2</samp> or &lsquo;<samp>avx</samp>&rsquo; which enables <samp>-mavx</samp> by default.
893This option is only supported on i386 and x86-64 targets.
894</p>
895</dd>
896<dt><span><code>--with-fp-32=<var>mode</var></code></span></dt>
897<dd><p>On MIPS targets, set the default value for the <samp>-mfp</samp> option when using
898the o32 ABI.  The possibilities for <var>mode</var> are:
899</p><dl compact="compact">
900<dt><span><code>32</code></span></dt>
901<dd><p>Use the o32 FP32 ABI extension, as with the <samp>-mfp32</samp> command-line
902option.
903</p></dd>
904<dt><span><code>xx</code></span></dt>
905<dd><p>Use the o32 FPXX ABI extension, as with the <samp>-mfpxx</samp> command-line
906option.
907</p></dd>
908<dt><span><code>64</code></span></dt>
909<dd><p>Use the o32 FP64 ABI extension, as with the <samp>-mfp64</samp> command-line
910option.
911</p></dd>
912</dl>
913<p>In the absence of this configuration option the default is to use the o32
914FP32 ABI extension.
915</p>
916</dd>
917<dt><span><code>--with-odd-spreg-32</code></span></dt>
918<dd><p>On MIPS targets, set the <samp>-modd-spreg</samp> option by default when using
919the o32 ABI.
920</p>
921</dd>
922<dt><span><code>--without-odd-spreg-32</code></span></dt>
923<dd><p>On MIPS targets, set the <samp>-mno-odd-spreg</samp> option by default when using
924the o32 ABI.  This is normally used in conjunction with
925<samp>--with-fp-32=64</samp> in order to target the o32 FP64A ABI extension.
926</p>
927</dd>
928<dt><span><code>--with-nan=<var>encoding</var></code></span></dt>
929<dd><p>On MIPS targets, set the default encoding convention to use for the
930special not-a-number (NaN) IEEE 754 floating-point data.  The
931possibilities for <var>encoding</var> are:
932</p><dl compact="compact">
933<dt><span><code>legacy</code></span></dt>
934<dd><p>Use the legacy encoding, as with the <samp>-mnan=legacy</samp> command-line
935option.
936</p></dd>
937<dt><span><code>2008</code></span></dt>
938<dd><p>Use the 754-2008 encoding, as with the <samp>-mnan=2008</samp> command-line
939option.
940</p></dd>
941</dl>
942<p>To use this configuration option you must have an assembler version
943installed that supports the <samp>-mnan=</samp> command-line option too.
944In the absence of this configuration option the default convention is
945the legacy encoding, as when neither of the <samp>-mnan=2008</samp> and
946<samp>-mnan=legacy</samp> command-line options has been used.
947</p>
948</dd>
949<dt><span><code>--with-divide=<var>type</var></code></span></dt>
950<dd><p>Specify how the compiler should generate code for checking for
951division by zero.  This option is only supported on the MIPS target.
952The possibilities for <var>type</var> are:
953</p><dl compact="compact">
954<dt><span><code>traps</code></span></dt>
955<dd><p>Division by zero checks use conditional traps (this is the default on
956systems that support conditional traps).
957</p></dd>
958<dt><span><code>breaks</code></span></dt>
959<dd><p>Division by zero checks use the break instruction.
960</p></dd>
961</dl>
962
963
964</dd>
965<dt><span><code>--with-llsc</code></span></dt>
966<dd><p>On MIPS targets, make <samp>-mllsc</samp> the default when no
967<samp>-mno-llsc</samp> option is passed.  This is the default for
968Linux-based targets, as the kernel will emulate them if the ISA does
969not provide them.
970</p>
971</dd>
972<dt><span><code>--without-llsc</code></span></dt>
973<dd><p>On MIPS targets, make <samp>-mno-llsc</samp> the default when no
974<samp>-mllsc</samp> option is passed.
975</p>
976</dd>
977<dt><span><code>--with-synci</code></span></dt>
978<dd><p>On MIPS targets, make <samp>-msynci</samp> the default when no
979<samp>-mno-synci</samp> option is passed.
980</p>
981</dd>
982<dt><span><code>--without-synci</code></span></dt>
983<dd><p>On MIPS targets, make <samp>-mno-synci</samp> the default when no
984<samp>-msynci</samp> option is passed.  This is the default.
985</p>
986</dd>
987<dt><span><code>--with-lxc1-sxc1</code></span></dt>
988<dd><p>On MIPS targets, make <samp>-mlxc1-sxc1</samp> the default when no
989<samp>-mno-lxc1-sxc1</samp> option is passed.  This is the default.
990</p>
991</dd>
992<dt><span><code>--without-lxc1-sxc1</code></span></dt>
993<dd><p>On MIPS targets, make <samp>-mno-lxc1-sxc1</samp> the default when no
994<samp>-mlxc1-sxc1</samp> option is passed.  The indexed load/store
995instructions are not directly a problem but can lead to unexpected
996behaviour when deployed in an application intended for a 32-bit address
997space but run on a 64-bit processor.  The issue is seen because all
998known MIPS 64-bit Linux kernels execute o32 and n32 applications
999with 64-bit addressing enabled which affects the overflow behaviour
1000of the indexed addressing mode.  GCC will assume that ordinary
100132-bit arithmetic overflow behaviour is the same whether performed
1002as an <code>addu</code> instruction or as part of the address calculation
1003in <code>lwxc1</code> type instructions.  This assumption holds true in a
1004pure 32-bit environment and can hold true in a 64-bit environment if
1005the address space is accurately set to be 32-bit for o32 and n32.
1006</p>
1007</dd>
1008<dt><span><code>--with-madd4</code></span></dt>
1009<dd><p>On MIPS targets, make <samp>-mmadd4</samp> the default when no
1010<samp>-mno-madd4</samp> option is passed.  This is the default.
1011</p>
1012</dd>
1013<dt><span><code>--without-madd4</code></span></dt>
1014<dd><p>On MIPS targets, make <samp>-mno-madd4</samp> the default when no
1015<samp>-mmadd4</samp> option is passed.  The <code>madd4</code> instruction
1016family can be problematic when targeting a combination of cores that
1017implement these instructions differently.  There are two known cores
1018that implement these as fused operations instead of unfused (where
1019unfused is normally expected).  Disabling these instructions is the
1020only way to ensure compatible code is generated; this will incur
1021a performance penalty.
1022</p>
1023</dd>
1024<dt><span><code>--with-mips-plt</code></span></dt>
1025<dd><p>On MIPS targets, make use of copy relocations and PLTs.
1026These features are extensions to the traditional
1027SVR4-based MIPS ABIs and require support from GNU binutils
1028and the runtime C library.
1029</p>
1030</dd>
1031<dt><span><code>--with-stack-clash-protection-guard-size=<var>size</var></code></span></dt>
1032<dd><p>On certain targets this option sets the default stack clash protection guard
1033size as a power of two in bytes.  On AArch64 <var>size</var> is required to be either
103412 (4KB) or 16 (64KB).
1035</p>
1036</dd>
1037<dt><span><code>--enable-__cxa_atexit</code></span></dt>
1038<dd><p>Define if you want to use __cxa_atexit, rather than atexit, to
1039register C++ destructors for local statics and global objects.
1040This is essential for fully standards-compliant handling of
1041destructors, but requires __cxa_atexit in libc.  This option is currently
1042only available on systems with GNU libc.  When enabled, this will cause
1043<samp>-fuse-cxa-atexit</samp> to be passed by default.
1044</p>
1045</dd>
1046<dt><span><code>--enable-gnu-indirect-function</code></span></dt>
1047<dd><p>Define if you want to enable the <code>ifunc</code> attribute.  This option is
1048currently only available on systems with GNU libc on certain targets.
1049</p>
1050</dd>
1051<dt><span><code>--enable-target-optspace</code></span></dt>
1052<dd><p>Specify that target
1053libraries should be optimized for code space instead of code speed.
1054This is the default for the m32r platform.
1055</p>
1056</dd>
1057<dt><span><code>--with-cpp-install-dir=<var>dirname</var></code></span></dt>
1058<dd><p>Specify that the user visible <code>cpp</code> program should be installed
1059in <samp><var>prefix</var>/<var>dirname</var>/cpp</samp>, in addition to <var>bindir</var>.
1060</p>
1061</dd>
1062<dt><span><code>--enable-comdat</code></span></dt>
1063<dd><p>Enable COMDAT group support.  This is primarily used to override the
1064automatically detected value.
1065</p>
1066</dd>
1067<dt><span><code>--enable-initfini-array</code></span></dt>
1068<dd><p>Force the use of sections <code>.init_array</code> and <code>.fini_array</code>
1069(instead of <code>.init</code> and <code>.fini</code>) for constructors and
1070destructors.  Option <samp>--disable-initfini-array</samp> has the
1071opposite effect.  If neither option is specified, the configure script
1072will try to guess whether the <code>.init_array</code> and
1073<code>.fini_array</code> sections are supported and, if they are, use them.
1074</p>
1075</dd>
1076<dt><span><code>--enable-link-mutex</code></span></dt>
1077<dd><p>When building GCC, use a mutex to avoid linking the compilers for
1078multiple languages at the same time, to avoid thrashing on build
1079systems with limited free memory.  The default is not to use such a mutex.
1080</p>
1081</dd>
1082<dt><span><code>--enable-maintainer-mode</code></span></dt>
1083<dd><p>The build rules that regenerate the Autoconf and Automake output files as
1084well as the GCC master message catalog <samp>gcc.pot</samp> are normally
1085disabled.  This is because it can only be rebuilt if the complete source
1086tree is present.  If you have changed the sources and want to rebuild the
1087catalog, configuring with <samp>--enable-maintainer-mode</samp> will enable
1088this.  Note that you need a recent version of the <code>gettext</code> tools
1089to do so.
1090</p>
1091</dd>
1092<dt><span><code>--disable-bootstrap</code></span></dt>
1093<dd><p>For a native build, the default configuration is to perform
1094a 3-stage bootstrap of the compiler when &lsquo;<samp>make</samp>&rsquo; is invoked,
1095testing that GCC can compile itself correctly.  If you want to disable
1096this process, you can configure with <samp>--disable-bootstrap</samp>.
1097</p>
1098</dd>
1099<dt><span><code>--enable-bootstrap</code></span></dt>
1100<dd><p>In special cases, you may want to perform a 3-stage build
1101even if the target and host triplets are different.
1102This is possible when the host can run code compiled for
1103the target (e.g. host is i686-linux, target is i486-linux).
1104Starting from GCC 4.2, to do this you have to configure explicitly
1105with <samp>--enable-bootstrap</samp>.
1106</p>
1107</dd>
1108<dt><span><code>--enable-generated-files-in-srcdir</code></span></dt>
1109<dd><p>Neither the .c and .h files that are generated from Bison and flex nor the
1110info manuals and man pages that are built from the .texi files are present
1111in the repository development tree.  When building GCC from that development tree,
1112or from one of our snapshots, those generated files are placed in your
1113build directory, which allows for the source to be in a readonly
1114directory.
1115</p>
1116<p>If you configure with <samp>--enable-generated-files-in-srcdir</samp> then those
1117generated files will go into the source directory.  This is mainly intended
1118for generating release or prerelease tarballs of the GCC sources, since it
1119is not a requirement that the users of source releases to have flex, Bison,
1120or makeinfo.
1121</p>
1122</dd>
1123<dt><span><code>--enable-version-specific-runtime-libs</code></span></dt>
1124<dd><p>Specify
1125that runtime libraries should be installed in the compiler specific
1126subdirectory (<samp><var>libdir</var>/gcc</samp>) rather than the usual places.  In
1127addition, &lsquo;<samp>libstdc++</samp>&rsquo;&rsquo;s include files will be installed into
1128<samp><var>libdir</var></samp> unless you overruled it by using
1129<samp>--with-gxx-include-dir=<var>dirname</var></samp>.  Using this option is
1130particularly useful if you intend to use several versions of GCC in
1131parallel.  The default is &lsquo;<samp>yes</samp>&rsquo; for &lsquo;<samp>libada</samp>&rsquo;, and &lsquo;<samp>no</samp>&rsquo; for
1132the remaining libraries.
1133</p>
1134</dd>
1135<dt><span><code><span id="WithAixSoname"></span>--with-aix-soname=&lsquo;<samp>aix</samp>&rsquo;, &lsquo;<samp>svr4</samp>&rsquo; or &lsquo;<samp>both</samp>&rsquo;</code></span></dt>
1136<dd><p>Traditional AIX shared library versioning (versioned <code>Shared Object</code>
1137files as members of unversioned <code>Archive Library</code> files named
1138&lsquo;<samp>lib.a</samp>&rsquo;) causes numerous headaches for package managers. However,
1139<code>Import Files</code> as members of <code>Archive Library</code> files allow for
1140<strong>filename-based versioning</strong> of shared libraries as seen on Linux/SVR4,
1141where this is called the &quot;SONAME&quot;. But as they prevent static linking,
1142<code>Import Files</code> may be used with <code>Runtime Linking</code> only, where the
1143linker does search for &lsquo;<samp>libNAME.so</samp>&rsquo; before &lsquo;<samp>libNAME.a</samp>&rsquo; library
1144filenames with the &lsquo;<samp>-lNAME</samp>&rsquo; linker flag.
1145</p>
1146<span id="AixLdCommand"></span><p>For detailed information please refer to the AIX
1147<a href="https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22">ld
1148Command</a> reference.
1149</p>
1150<p>As long as shared library creation is enabled, upon:
1151</p><dl compact="compact">
1152<dt><span><code>--with-aix-soname=aix</code></span></dt>
1153<dt><span><code>--with-aix-soname=both</code></span></dt>
1154<dd><p>A (traditional AIX) <code>Shared Archive Library</code> file is created:
1155 </p><ul>
1156<li> using the &lsquo;<samp>libNAME.a</samp>&rsquo; filename scheme
1157  </li><li> with the <code>Shared Object</code> file as archive member named
1158  &lsquo;<samp>libNAME.so.V</samp>&rsquo; (except for &lsquo;<samp>libgcc_s</samp>&rsquo;, where the <code>Shared
1159  Object</code> file is named &lsquo;<samp>shr.o</samp>&rsquo; for backwards compatibility), which
1160  <ul class="no-bullet">
1161<li>- is used for runtime loading from inside the &lsquo;<samp>libNAME.a</samp>&rsquo; file
1162   </li><li>- is used for dynamic loading via
1163   <code>dlopen(&quot;libNAME.a(libNAME.so.V)&quot;, RTLD_MEMBER)</code>
1164   </li><li>- is used for shared linking
1165   </li><li>- is used for static linking, so no separate <code>Static Archive
1166   Library</code> file is needed
1167  </li></ul>
1168</li></ul>
1169</dd>
1170<dt><span><code>--with-aix-soname=both</code></span></dt>
1171<dt><span><code>--with-aix-soname=svr4</code></span></dt>
1172<dd><p>A (second) <code>Shared Archive Library</code> file is created:
1173 </p><ul>
1174<li> using the &lsquo;<samp>libNAME.so.V</samp>&rsquo; filename scheme
1175 </li><li> with the <code>Shared Object</code> file as archive member named
1176 &lsquo;<samp>shr.o</samp>&rsquo;, which
1177  <ul class="no-bullet">
1178<li>- is created with the <code>-G linker flag</code>
1179   </li><li>- has the <code>F_LOADONLY</code> flag set
1180   </li><li>- is used for runtime loading from inside the &lsquo;<samp>libNAME.so.V</samp>&rsquo; file
1181   </li><li>- is used for dynamic loading via <code>dlopen(&quot;libNAME.so.V(shr.o)&quot;,
1182   RTLD_MEMBER)</code>
1183  </li></ul>
1184</li><li> with the <code>Import File</code> as archive member named &lsquo;<samp>shr.imp</samp>&rsquo;,
1185 which
1186  <ul class="no-bullet">
1187<li>- refers to &lsquo;<samp>libNAME.so.V(shr.o)</samp>&rsquo; as the &quot;SONAME&quot;, to be recorded
1188   in the <code>Loader Section</code> of subsequent binaries
1189   </li><li>- indicates whether &lsquo;<samp>libNAME.so.V(shr.o)</samp>&rsquo; is 32 or 64 bit
1190   </li><li>- lists all the public symbols exported by &lsquo;<samp>lib.so.V(shr.o)</samp>&rsquo;,
1191   eventually decorated with the <code>&lsquo;<samp>weak</samp>&rsquo; Keyword</code>
1192   </li><li>- is necessary for shared linking against &lsquo;<samp>lib.so.V(shr.o)</samp>&rsquo;
1193   </li></ul>
1194</li></ul>
1195<p>A symbolic link using the &lsquo;<samp>libNAME.so</samp>&rsquo; filename scheme is created:
1196  </p><ul>
1197<li> pointing to the &lsquo;<samp>libNAME.so.V</samp>&rsquo; <code>Shared Archive Library</code> file
1198  </li><li> to permit the <code>ld Command</code> to find &lsquo;<samp>lib.so.V(shr.imp)</samp>&rsquo; via
1199  the &lsquo;<samp>-lNAME</samp>&rsquo; argument (requires <code>Runtime Linking</code> to be enabled)
1200  </li><li> to permit dynamic loading of &lsquo;<samp>lib.so.V(shr.o)</samp>&rsquo; without the need
1201  to specify the version number via <code>dlopen(&quot;libNAME.so(shr.o)&quot;,
1202  RTLD_MEMBER)</code>
1203  </li></ul>
1204</dd>
1205</dl>
1206
1207<p>As long as static library creation is enabled, upon:
1208</p><dl compact="compact">
1209<dt><span><code>--with-aix-soname=svr4</code></span></dt>
1210<dd><p>A <code>Static Archive Library</code> is created:
1211 </p><ul>
1212<li> using the &lsquo;<samp>libNAME.a</samp>&rsquo; filename scheme
1213 </li><li> with all the <code>Static Object</code> files as archive members, which
1214  <ul class="no-bullet">
1215<li>- are used for static linking
1216  </li></ul>
1217</li></ul>
1218</dd>
1219</dl>
1220
1221<p>While the aix-soname=&lsquo;<samp>svr4</samp>&rsquo; option does not create <code>Shared Object</code>
1222files as members of unversioned <code>Archive Library</code> files any more, package
1223managers still are responsible to
1224<a href="./specific.html#TransferAixShobj">transfer</a> <code>Shared Object</code> files
1225found as member of a previously installed unversioned <code>Archive Library</code>
1226file into the newly installed <code>Archive Library</code> file with the same
1227filename.
1228</p>
1229<p><em>WARNING:</em> Creating <code>Shared Object</code> files with <code>Runtime Linking</code>
1230enabled may bloat the TOC, eventually leading to <code>TOC overflow</code> errors,
1231requiring the use of either the <samp>-Wl,-bbigtoc</samp> linker flag (seen to
1232break with the <code>GDB</code> debugger) or some of the TOC-related compiler flags,
1233see &ldquo;RS/6000 and PowerPC Options&rdquo; in the main manual.
1234</p>
1235<p><samp>--with-aix-soname</samp> is currently supported by &lsquo;<samp>libgcc_s</samp>&rsquo; only, so
1236this option is still experimental and not for normal use yet.
1237</p>
1238<p>Default is the traditional behavior <samp>--with-aix-soname=&lsquo;<samp>aix</samp>&rsquo;</samp>.
1239</p>
1240</dd>
1241<dt><span><code>--enable-languages=<var>lang1</var>,<var>lang2</var>,&hellip;</code></span></dt>
1242<dd><p>Specify that only a particular subset of compilers and
1243their runtime libraries should be built.  For a list of valid values for
1244<var>langN</var> you can issue the following command in the
1245<samp>gcc</samp> directory of your GCC source tree:<br>
1246</p><div class="example">
1247<pre class="example">grep ^language= */config-lang.in
1248</pre></div>
1249<p>Currently, you can use any of the following:
1250<code>all</code>, <code>default</code>, <code>ada</code>, <code>c</code>, <code>c++</code>, <code>d</code>,
1251<code>fortran</code>, <code>go</code>, <code>jit</code>, <code>lto</code>, <code>objc</code>, <code>obj-c++</code>.
1252Building the Ada compiler has special requirements, see below.
1253If you do not pass this flag, or specify the option <code>default</code>, then the
1254default languages available in the <samp>gcc</samp> sub-tree will be configured.
1255Ada, D, Go, Jit, and Objective-C++ are not default languages.  LTO is not a
1256default language, but is built by default because <samp>--enable-lto</samp> is
1257enabled by default.  The other languages are default languages.  If
1258<code>all</code> is specified, then all available languages are built.  An
1259exception is <code>jit</code> language, which requires
1260<samp>--enable-host-shared</samp> to be included with <code>all</code>.
1261</p>
1262</dd>
1263<dt><span><code>--enable-stage1-languages=<var>lang1</var>,<var>lang2</var>,&hellip;</code></span></dt>
1264<dd><p>Specify that a particular subset of compilers and their runtime
1265libraries should be built with the system C compiler during stage 1 of
1266the bootstrap process, rather than only in later stages with the
1267bootstrapped C compiler.  The list of valid values is the same as for
1268<samp>--enable-languages</samp>, and the option <code>all</code> will select all
1269of the languages enabled by <samp>--enable-languages</samp>.  This option is
1270primarily useful for GCC development; for instance, when a development
1271version of the compiler cannot bootstrap due to compiler bugs, or when
1272one is debugging front ends other than the C front end.  When this
1273option is used, one can then build the target libraries for the
1274specified languages with the stage-1 compiler by using <code>make
1275stage1-bubble all-target</code>, or run the testsuite on the stage-1 compiler
1276for the specified languages using <code>make stage1-start check-gcc</code>.
1277</p>
1278</dd>
1279<dt><span><code>--disable-libada</code></span></dt>
1280<dd><p>Specify that the run-time libraries and tools used by GNAT should not
1281be built.  This can be useful for debugging, or for compatibility with
1282previous Ada build procedures, when it was required to explicitly
1283do a &lsquo;<samp>make -C gcc gnatlib_and_tools</samp>&rsquo;.
1284</p>
1285</dd>
1286<dt><span><code>--disable-libsanitizer</code></span></dt>
1287<dd><p>Specify that the run-time libraries for the various sanitizers should
1288not be built.
1289</p>
1290</dd>
1291<dt><span><code>--disable-libssp</code></span></dt>
1292<dd><p>Specify that the run-time libraries for stack smashing protection
1293should not be built or linked against.  On many targets library support
1294is provided by the C library instead.
1295</p>
1296</dd>
1297<dt><span><code>--disable-libquadmath</code></span></dt>
1298<dd><p>Specify that the GCC quad-precision math library should not be built.
1299On some systems, the library is required to be linkable when building
1300the Fortran front end, unless <samp>--disable-libquadmath-support</samp>
1301is used.
1302</p>
1303</dd>
1304<dt><span><code>--disable-libquadmath-support</code></span></dt>
1305<dd><p>Specify that the Fortran front end and <code>libgfortran</code> do not add
1306support for <code>libquadmath</code> on systems supporting it.
1307</p>
1308</dd>
1309<dt><span><code>--disable-libgomp</code></span></dt>
1310<dd><p>Specify that the GNU Offloading and Multi Processing Runtime Library
1311should not be built.
1312</p>
1313</dd>
1314<dt><span><code>--disable-libvtv</code></span></dt>
1315<dd><p>Specify that the run-time libraries used by vtable verification
1316should not be built.
1317</p>
1318</dd>
1319<dt><span><code>--with-dwarf2</code></span></dt>
1320<dd><p>Specify that the compiler should
1321use DWARF 2 debugging information as the default.
1322</p>
1323</dd>
1324<dt><span><code>--with-advance-toolchain=<var>at</var></code></span></dt>
1325<dd><p>On 64-bit PowerPC Linux systems, configure the compiler to use the
1326header files, library files, and the dynamic linker from the Advance
1327Toolchain release <var>at</var> instead of the default versions that are
1328provided by the Linux distribution.  In general, this option is
1329intended for the developers of GCC, and it is not intended for general
1330use.
1331</p>
1332</dd>
1333<dt><span><code>--enable-targets=all</code></span></dt>
1334<dt><span><code>--enable-targets=<var>target_list</var></code></span></dt>
1335<dd><p>Some GCC targets, e.g. powerpc64-linux, build bi-arch compilers.
1336These are compilers that are able to generate either 64-bit or 32-bit
1337code.  Typically, the corresponding 32-bit target, e.g.
1338powerpc-linux for powerpc64-linux, only generates 32-bit code.  This
1339option enables the 32-bit target to be a bi-arch compiler, which is
1340useful when you want a bi-arch compiler that defaults to 32-bit, and
1341you are building a bi-arch or multi-arch binutils in a combined tree.
1342On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64),
1343defaulted to o32.
1344Currently, this option only affects sparc-linux, powerpc-linux, x86-linux,
1345mips-linux and s390-linux.
1346</p>
1347</dd>
1348<dt><span><code>--enable-default-pie</code></span></dt>
1349<dd><p>Turn on <samp>-fPIE</samp> and <samp>-pie</samp> by default.
1350</p>
1351</dd>
1352<dt><span><code>--enable-secureplt</code></span></dt>
1353<dd><p>This option enables <samp>-msecure-plt</samp> by default for powerpc-linux.
1354See &ldquo;RS/6000 and PowerPC Options&rdquo; in the main manual
1355</p>
1356</dd>
1357<dt><span><code>--enable-default-ssp</code></span></dt>
1358<dd><p>Turn on <samp>-fstack-protector-strong</samp> by default.
1359</p>
1360</dd>
1361<dt><span><code>--enable-cld</code></span></dt>
1362<dd><p>This option enables <samp>-mcld</samp> by default for 32-bit x86 targets.
1363See &ldquo;i386 and x86-64 Options&rdquo; in the main manual
1364</p>
1365</dd>
1366<dt><span><code>--enable-large-address-aware</code></span></dt>
1367<dd><p>The <samp>--enable-large-address-aware</samp> option arranges for MinGW
1368executables to be linked using the <samp>--large-address-aware</samp>
1369option, that enables the use of more than 2GB of memory.  If GCC is
1370configured with this option, its effects can be reversed by passing the
1371<samp>-Wl,--disable-large-address-aware</samp> option to the so-configured
1372compiler driver.
1373</p>
1374</dd>
1375<dt><span><code>--enable-win32-registry</code></span></dt>
1376<dt><span><code>--enable-win32-registry=<var>key</var></code></span></dt>
1377<dt><span><code>--disable-win32-registry</code></span></dt>
1378<dd><p>The <samp>--enable-win32-registry</samp> option enables Microsoft Windows-hosted GCC
1379to look up installations paths in the registry using the following key:
1380</p>
1381<div class="example">
1382<pre class="example"><code>HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\<var>key</var></code>
1383</pre></div>
1384
1385<p><var>key</var> defaults to GCC version number, and can be overridden by the
1386<samp>--enable-win32-registry=<var>key</var></samp> option.  Vendors and distributors
1387who use custom installers are encouraged to provide a different key,
1388perhaps one comprised of vendor name and GCC version number, to
1389avoid conflict with existing installations.  This feature is enabled
1390by default, and can be disabled by <samp>--disable-win32-registry</samp>
1391option.  This option has no effect on the other hosts.
1392</p>
1393</dd>
1394<dt><span><code>--nfp</code></span></dt>
1395<dd><p>Specify that the machine does not have a floating point unit.  This
1396option only applies to &lsquo;<samp>m68k-sun-sunos<var>n</var></samp>&rsquo;.  On any other
1397system, <samp>--nfp</samp> has no effect.
1398</p>
1399</dd>
1400<dt><span><code>--enable-werror</code></span></dt>
1401<dt><span><code>--disable-werror</code></span></dt>
1402<dt><span><code>--enable-werror=yes</code></span></dt>
1403<dt><span><code>--enable-werror=no</code></span></dt>
1404<dd><p>When you specify this option, it controls whether certain files in the
1405compiler are built with <samp>-Werror</samp> in bootstrap stage2 and later.
1406If you don&rsquo;t specify it, <samp>-Werror</samp> is turned on for the main
1407development trunk.  However it defaults to off for release branches and
1408final releases.  The specific files which get <samp>-Werror</samp> are
1409controlled by the Makefiles.
1410</p>
1411</dd>
1412<dt><span><code>--enable-checking</code></span></dt>
1413<dt><span><code>--disable-checking</code></span></dt>
1414<dt><span><code>--enable-checking=<var>list</var></code></span></dt>
1415<dd><p>This option controls performing internal consistency checks in the compiler.
1416It does not change the generated code, but adds error checking of the
1417requested complexity.  This slows down the compiler and may only work
1418properly if you are building the compiler with GCC.
1419</p>
1420<p>When the option is not specified, the active set of checks depends on context.
1421Namely, bootstrap stage 1 defaults to &lsquo;<samp>--enable-checking=yes</samp>&rsquo;, builds
1422from release branches or release archives default to
1423&lsquo;<samp>--enable-checking=release</samp>&rsquo;, and otherwise
1424&lsquo;<samp>--enable-checking=yes,extra</samp>&rsquo; is used.  When the option is
1425specified without a <var>list</var>, the result is the same as
1426&lsquo;<samp>--enable-checking=yes</samp>&rsquo;.  Likewise, &lsquo;<samp>--disable-checking</samp>&rsquo; is
1427equivalent to &lsquo;<samp>--enable-checking=no</samp>&rsquo;.
1428</p>
1429<p>The categories of checks available in <var>list</var> are &lsquo;<samp>yes</samp>&rsquo; (most common
1430checks &lsquo;<samp>assert,misc,gc,gimple,rtlflag,runtime,tree,types</samp>&rsquo;), &lsquo;<samp>no</samp>&rsquo;
1431(no checks at all), &lsquo;<samp>all</samp>&rsquo; (all but &lsquo;<samp>valgrind</samp>&rsquo;), &lsquo;<samp>release</samp>&rsquo;
1432(cheapest checks &lsquo;<samp>assert,runtime</samp>&rsquo;) or &lsquo;<samp>none</samp>&rsquo; (same as &lsquo;<samp>no</samp>&rsquo;).
1433&lsquo;<samp>release</samp>&rsquo; checks are always on and to disable them
1434&lsquo;<samp>--disable-checking</samp>&rsquo; or &lsquo;<samp>--enable-checking=no[,&lt;other checks&gt;]</samp>&rsquo;
1435must be explicitly requested.  Disabling assertions makes the compiler and
1436runtime slightly faster but increases the risk of undetected internal errors
1437causing wrong code to be generated.
1438</p>
1439<p>Individual checks can be enabled with these flags: &lsquo;<samp>assert</samp>&rsquo;, &lsquo;<samp>df</samp>&rsquo;,
1440&lsquo;<samp>extra</samp>&rsquo;, &lsquo;<samp>fold</samp>&rsquo;, &lsquo;<samp>gc</samp>&rsquo;, &lsquo;<samp>gcac</samp>&rsquo;, &lsquo;<samp>gimple</samp>&rsquo;,
1441&lsquo;<samp>misc</samp>&rsquo;, &lsquo;<samp>rtl</samp>&rsquo;, &lsquo;<samp>rtlflag</samp>&rsquo;, &lsquo;<samp>runtime</samp>&rsquo;, &lsquo;<samp>tree</samp>&rsquo;,
1442&lsquo;<samp>types</samp>&rsquo; and &lsquo;<samp>valgrind</samp>&rsquo;.  &lsquo;<samp>extra</samp>&rsquo; extends &lsquo;<samp>misc</samp>&rsquo;
1443checking with extra checks that might affect code generation and should
1444therefore not differ between stage1 and later stages in bootstrap.
1445</p>
1446<p>The &lsquo;<samp>valgrind</samp>&rsquo; check requires the external <code>valgrind</code> simulator,
1447available from <a href="http://valgrind.org/">http://valgrind.org/</a>.  The &lsquo;<samp>rtl</samp>&rsquo; checks are
1448expensive and the &lsquo;<samp>df</samp>&rsquo;, &lsquo;<samp>gcac</samp>&rsquo; and &lsquo;<samp>valgrind</samp>&rsquo; checks are very
1449expensive.
1450</p>
1451</dd>
1452<dt><span><code>--disable-stage1-checking</code></span></dt>
1453<dt><span><code>--enable-stage1-checking</code></span></dt>
1454<dt><span><code>--enable-stage1-checking=<var>list</var></code></span></dt>
1455<dd><p>This option affects only bootstrap build.  If no <samp>--enable-checking</samp>
1456option is specified the stage1 compiler is built with &lsquo;<samp>yes</samp>&rsquo; checking
1457enabled, otherwise the stage1 checking flags are the same as specified by
1458<samp>--enable-checking</samp>.  To build the stage1 compiler with
1459different checking options use <samp>--enable-stage1-checking</samp>.
1460The list of checking options is the same as for <samp>--enable-checking</samp>.
1461If your system is too slow or too small to bootstrap a released compiler
1462with checking for stage1 enabled, you can use &lsquo;<samp>--disable-stage1-checking</samp>&rsquo;
1463to disable checking for the stage1 compiler.
1464</p>
1465</dd>
1466<dt><span><code>--enable-coverage</code></span></dt>
1467<dt><span><code>--enable-coverage=<var>level</var></code></span></dt>
1468<dd><p>With this option, the compiler is built to collect self coverage
1469information, every time it is run.  This is for internal development
1470purposes, and only works when the compiler is being built with gcc.  The
1471<var>level</var> argument controls whether the compiler is built optimized or
1472not, values are &lsquo;<samp>opt</samp>&rsquo; and &lsquo;<samp>noopt</samp>&rsquo;.  For coverage analysis you
1473want to disable optimization, for performance analysis you want to
1474enable optimization.  When coverage is enabled, the default level is
1475without optimization.
1476</p>
1477</dd>
1478<dt><span><code>--enable-gather-detailed-mem-stats</code></span></dt>
1479<dd><p>When this option is specified more detailed information on memory
1480allocation is gathered.  This information is printed when using
1481<samp>-fmem-report</samp>.
1482</p>
1483</dd>
1484<dt><span><code>--enable-valgrind-annotations</code></span></dt>
1485<dd><p>Mark selected memory related operations in the compiler when run under
1486valgrind to suppress false positives.
1487</p>
1488</dd>
1489<dt><span><code>--enable-nls</code></span></dt>
1490<dt><span><code>--disable-nls</code></span></dt>
1491<dd><p>The <samp>--enable-nls</samp> option enables Native Language Support (NLS),
1492which lets GCC output diagnostics in languages other than American
1493English.  Native Language Support is enabled by default if not doing a
1494canadian cross build.  The <samp>--disable-nls</samp> option disables NLS.
1495</p>
1496</dd>
1497<dt><span><code>--with-included-gettext</code></span></dt>
1498<dd><p>If NLS is enabled, the <samp>--with-included-gettext</samp> option causes the build
1499procedure to prefer its copy of GNU <code>gettext</code>.
1500</p>
1501</dd>
1502<dt><span><code>--with-catgets</code></span></dt>
1503<dd><p>If NLS is enabled, and if the host lacks <code>gettext</code> but has the
1504inferior <code>catgets</code> interface, the GCC build procedure normally
1505ignores <code>catgets</code> and instead uses GCC&rsquo;s copy of the GNU
1506<code>gettext</code> library.  The <samp>--with-catgets</samp> option causes the
1507build procedure to use the host&rsquo;s <code>catgets</code> in this situation.
1508</p>
1509</dd>
1510<dt><span><code>--with-libiconv-prefix=<var>dir</var></code></span></dt>
1511<dd><p>Search for libiconv header files in <samp><var>dir</var>/include</samp> and
1512libiconv library files in <samp><var>dir</var>/lib</samp>.
1513</p>
1514</dd>
1515<dt><span><code>--enable-obsolete</code></span></dt>
1516<dd><p>Enable configuration for an obsoleted system.  If you attempt to
1517configure GCC for a system (build, host, or target) which has been
1518obsoleted, and you do not specify this flag, configure will halt with an
1519error message.
1520</p>
1521<p>All support for systems which have been obsoleted in one release of GCC
1522is removed entirely in the next major release, unless someone steps
1523forward to maintain the port.
1524</p>
1525</dd>
1526<dt><span><code>--enable-decimal-float</code></span></dt>
1527<dt><span><code>--enable-decimal-float=yes</code></span></dt>
1528<dt><span><code>--enable-decimal-float=no</code></span></dt>
1529<dt><span><code>--enable-decimal-float=bid</code></span></dt>
1530<dt><span><code>--enable-decimal-float=dpd</code></span></dt>
1531<dt><span><code>--disable-decimal-float</code></span></dt>
1532<dd><p>Enable (or disable) support for the C decimal floating point extension
1533that is in the IEEE 754-2008 standard.  This is enabled by default only
1534on PowerPC, i386, and x86_64 GNU/Linux systems.  Other systems may also
1535support it, but require the user to specifically enable it.  You can
1536optionally control which decimal floating point format is used (either
1537&lsquo;<samp>bid</samp>&rsquo; or &lsquo;<samp>dpd</samp>&rsquo;).  The &lsquo;<samp>bid</samp>&rsquo; (binary integer decimal)
1538format is default on i386 and x86_64 systems, and the &lsquo;<samp>dpd</samp>&rsquo;
1539(densely packed decimal) format is default on PowerPC systems.
1540</p>
1541</dd>
1542<dt><span><code>--enable-fixed-point</code></span></dt>
1543<dt><span><code>--disable-fixed-point</code></span></dt>
1544<dd><p>Enable (or disable) support for C fixed-point arithmetic.
1545This option is enabled by default for some targets (such as MIPS) which
1546have hardware-support for fixed-point operations.  On other targets, you
1547may enable this option manually.
1548</p>
1549</dd>
1550<dt><span><code>--with-long-double-128</code></span></dt>
1551<dd><p>Specify if <code>long double</code> type should be 128-bit by default on selected
1552GNU/Linux architectures.  If using <code>--without-long-double-128</code>,
1553<code>long double</code> will be by default 64-bit, the same as <code>double</code> type.
1554When neither of these configure options are used, the default will be
1555128-bit <code>long double</code> when built against GNU C Library 2.4 and later,
155664-bit <code>long double</code> otherwise.
1557</p>
1558</dd>
1559<dt><span><code>--with-long-double-format=ibm</code></span></dt>
1560<dt><span><code>--with-long-double-format=ieee</code></span></dt>
1561<dd><p>Specify whether <code>long double</code> uses the IBM extended double format
1562or the IEEE 128-bit floating point format on PowerPC Linux systems.
1563This configuration switch will only work on little endian PowerPC
1564Linux systems and on big endian 64-bit systems where the default cpu
1565is at least power7 (i.e. <samp>--with-cpu=power7</samp>,
1566<samp>--with-cpu=power8</samp>, or <samp>--with-cpu=power9</samp> is used).
1567</p>
1568<p>If you use the <samp>--with-long-double-64</samp> configuration option,
1569the <samp>--with-long-double-format=ibm</samp> and
1570<samp>--with-long-double-format=ieee</samp> options are ignored.
1571</p>
1572<p>The default <code>long double</code> format is to use IBM extended double.
1573Until all of the libraries are converted to use IEEE 128-bit floating
1574point, it is not recommended to use
1575<samp>--with-long-double-format=ieee</samp>.
1576</p>
1577<p>On little endian PowerPC Linux systems, if you explicitly set the
1578<code>long double</code> type, it will build multilibs to allow you to
1579select either <code>long double</code> format, unless you disable multilibs
1580with the <code>--disable-multilib</code> option.  At present,
1581<code>long double</code> multilibs are not built on big endian PowerPC Linux
1582systems.  If you are building multilibs, you will need to configure
1583the compiler using the <samp>--with-system-zlib</samp> option.
1584</p>
1585<p>If you do not set the <code>long double</code> type explicitly, no multilibs
1586will be generated.
1587</p>
1588</dd>
1589<dt><span><code>--enable-fdpic</code></span></dt>
1590<dd><p>On SH Linux systems, generate ELF FDPIC code.
1591</p>
1592</dd>
1593<dt><span><code>--with-gmp=<var>pathname</var></code></span></dt>
1594<dt><span><code>--with-gmp-include=<var>pathname</var></code></span></dt>
1595<dt><span><code>--with-gmp-lib=<var>pathname</var></code></span></dt>
1596<dt><span><code>--with-mpfr=<var>pathname</var></code></span></dt>
1597<dt><span><code>--with-mpfr-include=<var>pathname</var></code></span></dt>
1598<dt><span><code>--with-mpfr-lib=<var>pathname</var></code></span></dt>
1599<dt><span><code>--with-mpc=<var>pathname</var></code></span></dt>
1600<dt><span><code>--with-mpc-include=<var>pathname</var></code></span></dt>
1601<dt><span><code>--with-mpc-lib=<var>pathname</var></code></span></dt>
1602<dd><p>If you want to build GCC but do not have the GMP library, the MPFR
1603library and/or the MPC library installed in a standard location and
1604do not have their sources present in the GCC source tree then you
1605can explicitly specify the directory where they are installed
1606(&lsquo;<samp>--with-gmp=<var>gmpinstalldir</var></samp>&rsquo;,
1607&lsquo;<samp>--with-mpfr=<var>mpfrinstalldir</var></samp>&rsquo;,
1608&lsquo;<samp>--with-mpc=<var>mpcinstalldir</var></samp>&rsquo;).  The
1609<samp>--with-gmp=<var>gmpinstalldir</var></samp> option is shorthand for
1610<samp>--with-gmp-lib=<var>gmpinstalldir</var>/lib</samp> and
1611<samp>--with-gmp-include=<var>gmpinstalldir</var>/include</samp>.  Likewise the
1612<samp>--with-mpfr=<var>mpfrinstalldir</var></samp> option is shorthand for
1613<samp>--with-mpfr-lib=<var>mpfrinstalldir</var>/lib</samp> and
1614<samp>--with-mpfr-include=<var>mpfrinstalldir</var>/include</samp>, also the
1615<samp>--with-mpc=<var>mpcinstalldir</var></samp> option is shorthand for
1616<samp>--with-mpc-lib=<var>mpcinstalldir</var>/lib</samp> and
1617<samp>--with-mpc-include=<var>mpcinstalldir</var>/include</samp>.  If these
1618shorthand assumptions are not correct, you can use the explicit
1619include and lib options directly.  You might also need to ensure the
1620shared libraries can be found by the dynamic linker when building and
1621using GCC, for example by setting the runtime shared library path
1622variable (<code>LD_LIBRARY_PATH</code> on GNU/Linux and Solaris systems).
1623</p>
1624<p>These flags are applicable to the host platform only.  When building
1625a cross compiler, they will not be used to configure target libraries.
1626</p>
1627</dd>
1628<dt><span><code>--with-isl=<var>pathname</var></code></span></dt>
1629<dt><span><code>--with-isl-include=<var>pathname</var></code></span></dt>
1630<dt><span><code>--with-isl-lib=<var>pathname</var></code></span></dt>
1631<dd><p>If you do not have the isl library installed in a standard location and you
1632want to build GCC, you can explicitly specify the directory where it is
1633installed (&lsquo;<samp>--with-isl=<var>islinstalldir</var></samp>&rsquo;). The
1634<samp>--with-isl=<var>islinstalldir</var></samp> option is shorthand for
1635<samp>--with-isl-lib=<var>islinstalldir</var>/lib</samp> and
1636<samp>--with-isl-include=<var>islinstalldir</var>/include</samp>. If this
1637shorthand assumption is not correct, you can use the explicit
1638include and lib options directly.
1639</p>
1640<p>These flags are applicable to the host platform only.  When building
1641a cross compiler, they will not be used to configure target libraries.
1642</p>
1643</dd>
1644<dt><span><code>--with-stage1-ldflags=<var>flags</var></code></span></dt>
1645<dd><p>This option may be used to set linker flags to be used when linking
1646stage 1 of GCC.  These are also used when linking GCC if configured with
1647<samp>--disable-bootstrap</samp>.  If <samp>--with-stage1-libs</samp> is not set to a
1648value, then the default is &lsquo;<samp>-static-libstdc++ -static-libgcc</samp>&rsquo;, if
1649supported.
1650</p>
1651</dd>
1652<dt><span><code>--with-stage1-libs=<var>libs</var></code></span></dt>
1653<dd><p>This option may be used to set libraries to be used when linking stage 1
1654of GCC.  These are also used when linking GCC if configured with
1655<samp>--disable-bootstrap</samp>.
1656</p>
1657</dd>
1658<dt><span><code>--with-boot-ldflags=<var>flags</var></code></span></dt>
1659<dd><p>This option may be used to set linker flags to be used when linking
1660stage 2 and later when bootstrapping GCC.  If &ndash;with-boot-libs
1661is not is set to a value, then the default is
1662&lsquo;<samp>-static-libstdc++ -static-libgcc</samp>&rsquo;.
1663</p>
1664</dd>
1665<dt><span><code>--with-boot-libs=<var>libs</var></code></span></dt>
1666<dd><p>This option may be used to set libraries to be used when linking stage 2
1667and later when bootstrapping GCC.
1668</p>
1669</dd>
1670<dt><span><code>--with-debug-prefix-map=<var>map</var></code></span></dt>
1671<dd><p>Convert source directory names using <samp>-fdebug-prefix-map</samp> when
1672building runtime libraries.  &lsquo;<samp><var>map</var></samp>&rsquo; is a space-separated
1673list of maps of the form &lsquo;<samp><var>old</var>=<var>new</var></samp>&rsquo;.
1674</p>
1675</dd>
1676<dt><span><code>--enable-linker-build-id</code></span></dt>
1677<dd><p>Tells GCC to pass <samp>--build-id</samp> option to the linker for all final
1678links (links performed without the <samp>-r</samp> or <samp>--relocatable</samp>
1679option), if the linker supports it.  If you specify
1680<samp>--enable-linker-build-id</samp>, but your linker does not
1681support <samp>--build-id</samp> option, a warning is issued and the
1682<samp>--enable-linker-build-id</samp> option is ignored.  The default is off.
1683</p>
1684</dd>
1685<dt><span><code>--with-linker-hash-style=<var>choice</var></code></span></dt>
1686<dd><p>Tells GCC to pass <samp>--hash-style=<var>choice</var></samp> option to the
1687linker for all final links. <var>choice</var> can be one of
1688&lsquo;<samp>sysv</samp>&rsquo;, &lsquo;<samp>gnu</samp>&rsquo;, and &lsquo;<samp>both</samp>&rsquo; where &lsquo;<samp>sysv</samp>&rsquo; is the default.
1689</p>
1690</dd>
1691<dt><span><code>--enable-gnu-unique-object</code></span></dt>
1692<dt><span><code>--disable-gnu-unique-object</code></span></dt>
1693<dd><p>Tells GCC to use the gnu_unique_object relocation for C++ template
1694static data members and inline function local statics.  Enabled by
1695default for a toolchain with an assembler that accepts it and
1696GLIBC 2.11 or above, otherwise disabled.
1697</p>
1698</dd>
1699<dt><span><code>--with-diagnostics-color=<var>choice</var></code></span></dt>
1700<dd><p>Tells GCC to use <var>choice</var> as the default for <samp>-fdiagnostics-color=</samp>
1701option (if not used explicitly on the command line).  <var>choice</var>
1702can be one of &lsquo;<samp>never</samp>&rsquo;, &lsquo;<samp>auto</samp>&rsquo;, &lsquo;<samp>always</samp>&rsquo;, and &lsquo;<samp>auto-if-env</samp>&rsquo;
1703where &lsquo;<samp>auto</samp>&rsquo; is the default.  &lsquo;<samp>auto-if-env</samp>&rsquo; makes
1704<samp>-fdiagnostics-color=auto</samp> the default if <code>GCC_COLORS</code>
1705is present and non-empty in the environment of the compiler, and
1706<samp>-fdiagnostics-color=never</samp> otherwise.
1707</p>
1708</dd>
1709<dt><span><code>--with-diagnostics-urls=<var>choice</var></code></span></dt>
1710<dd><p>Tells GCC to use <var>choice</var> as the default for <samp>-fdiagnostics-urls=</samp>
1711option (if not used explicitly on the command line).  <var>choice</var>
1712can be one of &lsquo;<samp>never</samp>&rsquo;, &lsquo;<samp>auto</samp>&rsquo;, &lsquo;<samp>always</samp>&rsquo;, and &lsquo;<samp>auto-if-env</samp>&rsquo;
1713where &lsquo;<samp>auto</samp>&rsquo; is the default.  &lsquo;<samp>auto-if-env</samp>&rsquo; makes
1714<samp>-fdiagnostics-urls=auto</samp> the default if <code>GCC_URLS</code>
1715or <code>TERM_URLS</code> is present and non-empty in the environment of the
1716compiler, and <samp>-fdiagnostics-urls=never</samp> otherwise.
1717</p>
1718</dd>
1719<dt><span><code>--enable-lto</code></span></dt>
1720<dt><span><code>--disable-lto</code></span></dt>
1721<dd><p>Enable support for link-time optimization (LTO).  This is enabled by
1722default, and may be disabled using <samp>--disable-lto</samp>.
1723</p>
1724</dd>
1725<dt><span><code>--enable-linker-plugin-configure-flags=FLAGS</code></span></dt>
1726<dt><span><code>--enable-linker-plugin-flags=FLAGS</code></span></dt>
1727<dd><p>By default, linker plugins (such as the LTO plugin) are built for the
1728host system architecture.  For the case that the linker has a
1729different (but run-time compatible) architecture, these flags can be
1730specified to build plugins that are compatible to the linker.  For
1731example, if you are building GCC for a 64-bit x86_64
1732(&lsquo;<samp>x86_64-pc-linux-gnu</samp>&rsquo;) host system, but have a 32-bit x86
1733GNU/Linux (&lsquo;<samp>i686-pc-linux-gnu</samp>&rsquo;) linker executable (which is
1734executable on the former system), you can configure GCC as follows for
1735getting compatible linker plugins:
1736</p>
1737<div class="example">
1738<pre class="example">% <var>srcdir</var>/configure \
1739    --host=x86_64-pc-linux-gnu \
1740    --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \
1741    --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib'
1742</pre></div>
1743
1744</dd>
1745<dt><span><code>--with-plugin-ld=<var>pathname</var></code></span></dt>
1746<dd><p>Enable an alternate linker to be used at link-time optimization (LTO)
1747link time when <samp>-fuse-linker-plugin</samp> is enabled.
1748This linker should have plugin support such as gold starting with
1749version 2.20 or GNU ld starting with version 2.21.
1750See <samp>-fuse-linker-plugin</samp> for details.
1751</p>
1752</dd>
1753<dt><span><code>--enable-canonical-system-headers</code></span></dt>
1754<dt><span><code>--disable-canonical-system-headers</code></span></dt>
1755<dd><p>Enable system header path canonicalization for <samp>libcpp</samp>.  This can
1756produce shorter header file paths in diagnostics and dependency output
1757files, but these changed header paths may conflict with some compilation
1758environments.  Enabled by default, and may be disabled using
1759<samp>--disable-canonical-system-headers</samp>.
1760</p>
1761</dd>
1762<dt><span><code>--with-glibc-version=<var>major</var>.<var>minor</var></code></span></dt>
1763<dd><p>Tell GCC that when the GNU C Library (glibc) is used on the target it
1764will be version <var>major</var>.<var>minor</var> or later.  Normally this can
1765be detected from the C library&rsquo;s header files, but this option may be
1766needed when bootstrapping a cross toolchain without the header files
1767available for building the initial bootstrap compiler.
1768</p>
1769<p>If GCC is configured with some multilibs that use glibc and some that
1770do not, this option applies only to the multilibs that use glibc.
1771However, such configurations may not work well as not all the relevant
1772configuration in GCC is on a per-multilib basis.
1773</p>
1774</dd>
1775<dt><span><code>--enable-as-accelerator-for=<var>target</var></code></span></dt>
1776<dd><p>Build as offload target compiler. Specify offload host triple by <var>target</var>.
1777</p>
1778</dd>
1779<dt><span><code>--enable-offload-targets=<var>target1</var>[=<var>path1</var>],&hellip;,<var>targetN</var>[=<var>pathN</var>]</code></span></dt>
1780<dd><p>Enable offloading to targets <var>target1</var>, &hellip;, <var>targetN</var>.
1781Offload compilers are expected to be already installed.  Default search
1782path for them is <samp><var>exec-prefix</var></samp>, but it can be changed by
1783specifying paths <var>path1</var>, &hellip;, <var>pathN</var>.
1784</p>
1785<div class="example">
1786<pre class="example">% <var>srcdir</var>/configure \
1787    --enable-offload-targets=x86_64-intelmicemul-linux-gnu=/path/to/x86_64/compiler,nvptx-none,hsa
1788</pre></div>
1789
1790<p>If &lsquo;<samp>hsa</samp>&rsquo; is specified as one of the targets, the compiler will be
1791built with support for HSA GPU accelerators.  Because the same
1792compiler will emit the accelerator code, no path should be specified.
1793</p>
1794</dd>
1795<dt><span><code>--with-hsa-runtime=<var>pathname</var></code></span></dt>
1796<dt><span><code>--with-hsa-runtime-include=<var>pathname</var></code></span></dt>
1797<dt><span><code>--with-hsa-runtime-lib=<var>pathname</var></code></span></dt>
1798<dd>
1799<p>If you configure GCC with HSA offloading but do not have the HSA
1800run-time library installed in a standard location then you can
1801explicitly specify the directory where they are installed.  The
1802<samp>--with-hsa-runtime=<var>hsainstalldir</var></samp> option is a
1803shorthand for
1804<samp>--with-hsa-runtime-lib=<var>hsainstalldir</var>/lib</samp> and
1805<samp>--with-hsa-runtime-include=<var>hsainstalldir</var>/include</samp>.
1806</p>
1807</dd>
1808<dt><span><code>--enable-cet</code></span></dt>
1809<dt><span><code>--disable-cet</code></span></dt>
1810<dd><p>Enable building target run-time libraries with control-flow
1811instrumentation, see <samp>-fcf-protection</samp> option.  When
1812<code>--enable-cet</code> is specified target libraries are configured
1813to add <samp>-fcf-protection</samp> and, if needed, other target
1814specific options to a set of building options.
1815</p>
1816<p>The option is disabled by default.  When <code>--enable-cet=auto</code>
1817is used, it is enabled on Linux/x86 if target binutils
1818supports <code>Intel CET</code> instructions and disabled otherwise.
1819In this case the target libraries are configured to get additional
1820<samp>-fcf-protection</samp> option.
1821</p>
1822</dd>
1823<dt><span><code>--with-riscv-attribute=&lsquo;<samp>yes</samp>&rsquo;, &lsquo;<samp>no</samp>&rsquo; or &lsquo;<samp>default</samp>&rsquo;</code></span></dt>
1824<dd><p>Generate RISC-V attribute by default, in order to record extra build
1825information in object.
1826</p>
1827<p>The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal)
1828target if target binutils supported.
1829</p></dd>
1830</dl>
1831
1832<span id="Cross-Compiler-Specific-Options"></span><h4 class="subheading">Cross-Compiler-Specific Options</h4>
1833<p>The following options only apply to building cross compilers.
1834</p>
1835<dl compact="compact">
1836<dt><span><code>--with-toolexeclibdir=<var>dir</var></code></span></dt>
1837<dd><p>Specify the installation directory for libraries built with a cross compiler.
1838The default is <samp>${gcc_tooldir}/lib</samp>.
1839</p>
1840</dd>
1841<dt><span><code>--with-sysroot</code></span></dt>
1842<dt><span><code>--with-sysroot=<var>dir</var></code></span></dt>
1843<dd><p>Tells GCC to consider <var>dir</var> as the root of a tree that contains
1844(a subset of) the root filesystem of the target operating system.
1845Target system headers, libraries and run-time object files will be
1846searched for in there.  More specifically, this acts as if
1847<samp>--sysroot=<var>dir</var></samp> was added to the default options of the built
1848compiler.  The specified directory is not copied into the
1849install tree, unlike the options <samp>--with-headers</samp> and
1850<samp>--with-libs</samp> that this option obsoletes.  The default value,
1851in case <samp>--with-sysroot</samp> is not given an argument, is
1852<samp>${gcc_tooldir}/sys-root</samp>.  If the specified directory is a
1853subdirectory of <samp>${exec_prefix}</samp>, then it will be found relative to
1854the GCC binaries if the installation tree is moved.
1855</p>
1856<p>This option affects the system root for the compiler used to build
1857target libraries (which runs on the build system) and the compiler newly
1858installed with <code>make install</code>; it does not affect the compiler which is
1859used to build GCC itself.
1860</p>
1861<p>If you specify the <samp>--with-native-system-header-dir=<var>dirname</var></samp>
1862option then the compiler will search that directory within <var>dirname</var> for
1863native system headers rather than the default <samp>/usr/include</samp>.
1864</p>
1865</dd>
1866<dt><span><code>--with-build-sysroot</code></span></dt>
1867<dt><span><code>--with-build-sysroot=<var>dir</var></code></span></dt>
1868<dd><p>Tells GCC to consider <var>dir</var> as the system root (see
1869<samp>--with-sysroot</samp>) while building target libraries, instead of
1870the directory specified with <samp>--with-sysroot</samp>.  This option is
1871only useful when you are already using <samp>--with-sysroot</samp>.  You
1872can use <samp>--with-build-sysroot</samp> when you are configuring with
1873<samp>--prefix</samp> set to a directory that is different from the one in
1874which you are installing GCC and your target libraries.
1875</p>
1876<p>This option affects the system root for the compiler used to build
1877target libraries (which runs on the build system); it does not affect
1878the compiler which is used to build GCC itself.
1879</p>
1880<p>If you specify the <samp>--with-native-system-header-dir=<var>dirname</var></samp>
1881option then the compiler will search that directory within <var>dirname</var> for
1882native system headers rather than the default <samp>/usr/include</samp>.
1883</p>
1884</dd>
1885<dt><span><code>--with-headers</code></span></dt>
1886<dt><span><code>--with-headers=<var>dir</var></code></span></dt>
1887<dd><p>Deprecated in favor of <samp>--with-sysroot</samp>.
1888Specifies that target headers are available when building a cross compiler.
1889The <var>dir</var> argument specifies a directory which has the target include
1890files.  These include files will be copied into the <samp>gcc</samp> install
1891directory.  <em>This option with the <var>dir</var> argument is required</em> when
1892building a cross compiler, if <samp><var>prefix</var>/<var>target</var>/sys-include</samp>
1893doesn&rsquo;t pre-exist.  If <samp><var>prefix</var>/<var>target</var>/sys-include</samp> does
1894pre-exist, the <var>dir</var> argument may be omitted.  <code>fixincludes</code>
1895will be run on these files to make them compatible with GCC.
1896</p>
1897</dd>
1898<dt><span><code>--without-headers</code></span></dt>
1899<dd><p>Tells GCC not use any target headers from a libc when building a cross
1900compiler.  When crossing to GNU/Linux, you need the headers so GCC
1901can build the exception handling for libgcc.
1902</p>
1903</dd>
1904<dt><span><code>--with-libs</code></span></dt>
1905<dt><span><code>--with-libs=&quot;<var>dir1</var> <var>dir2</var> &hellip; <var>dirN</var>&quot;</code></span></dt>
1906<dd><p>Deprecated in favor of <samp>--with-sysroot</samp>.
1907Specifies a list of directories which contain the target runtime
1908libraries.  These libraries will be copied into the <samp>gcc</samp> install
1909directory.  If the directory list is omitted, this option has no
1910effect.
1911</p>
1912</dd>
1913<dt><span><code>--with-newlib</code></span></dt>
1914<dd><p>Specifies that &lsquo;<samp>newlib</samp>&rsquo; is
1915being used as the target C library.  This causes <code>__eprintf</code> to be
1916omitted from <samp>libgcc.a</samp> on the assumption that it will be provided by
1917&lsquo;<samp>newlib</samp>&rsquo;.
1918</p>
1919<a name="avr"></a>
1920</dd>
1921<dt><span><code>--with-avrlibc</code></span></dt>
1922<dd><p>Only supported for the AVR target. Specifies that &lsquo;<samp>AVR-Libc</samp>&rsquo; is
1923being used as the target C&nbsp; library.  This causes float support
1924functions like <code>__addsf3</code> to be omitted from <samp>libgcc.a</samp> on
1925the assumption that it will be provided by <samp>libm.a</samp>.  For more
1926technical details, cf. <a href="http://gcc.gnu.org/PR54461">PR54461</a>.
1927It is not supported for
1928RTEMS configurations, which currently use newlib.  The option is
1929supported since version 4.7.2 and is the default in 4.8.0 and newer.
1930</p>
1931</dd>
1932<dt><span><code>--with-double={32|64|32,64|64,32}</code></span></dt>
1933<dt><span><code>--with-long-double={32|64|32,64|64,32|double}</code></span></dt>
1934<dd><p>Only supported for the AVR target since version&nbsp;10.
1935Specify the default layout available for the C/C++ &lsquo;<samp>double</samp>&rsquo;
1936and &lsquo;<samp>long double</samp>&rsquo; type, respectively. The following rules apply:
1937</p><ul>
1938<li> The first value after the &lsquo;<samp>=</samp>&rsquo; specifies the default layout (in bits)
1939of the type and also the default for the <samp>-mdouble=</samp> resp.
1940<samp>-mlong-double=</samp> compiler option.
1941</li><li> If more than one value is specified, respective multilib variants are
1942available, and  <samp>-mdouble=</samp> resp. <samp>-mlong-double=</samp> acts
1943as a multilib option.
1944</li><li> If <samp>--with-long-double=double</samp> is specified, &lsquo;<samp>double</samp>&rsquo; and
1945&lsquo;<samp>long double</samp>&rsquo; will have the same layout.
1946</li><li> The defaults are <samp>--with-long-double=64,32</samp> and
1947<samp>--with-double=32,64</samp>.  The default &lsquo;<samp>double</samp>&rsquo; layout imposed by
1948the latter is compatible with older versions of the compiler that implement
1949&lsquo;<samp>double</samp>&rsquo; as a 32-bit type, which does not comply to the language standard.
1950</li></ul>
1951<p>Not all combinations of <samp>--with-double=</samp> and
1952<samp>--with-long-double=</samp> are valid.  For example, the combination
1953<samp>--with-double=32,64</samp> <samp>--with-long-double=32</samp> will be
1954rejected because the first option specifies the availability of
1955multilibs for &lsquo;<samp>double</samp>&rsquo;, whereas the second option implies
1956that &lsquo;<samp>long double</samp>&rsquo; &mdash; and hence also &lsquo;<samp>double</samp>&rsquo; &mdash; is always
195732&nbsp;bits wide.
1958</p>
1959</dd>
1960<dt><span><code>--with-double-comparison={tristate|bool|libf7}</code></span></dt>
1961<dd><p>Only supported for the AVR target since version&nbsp;10.
1962Specify what result format is returned by library functions that
1963compare 64-bit floating point values (<code>DFmode</code>).
1964The GCC default is &lsquo;<samp>tristate</samp>&rsquo;.  If the floating point
1965implementation returns a boolean instead, set it to &lsquo;<samp>bool</samp>&rsquo;.
1966</p>
1967</dd>
1968<dt><span><code>--with-libf7={libgcc|math|math-symbols|no}</code></span></dt>
1969<dd><p>Only supported for the AVR target since version&nbsp;10.
1970Specify to which degree code from LibF7 is included in libgcc.
1971LibF7 is an ad-hoc, AVR-specific, 64-bit floating point emulation
1972written in C and (inline) assembly. &lsquo;<samp>libgcc</samp>&rsquo; adds support
1973for functions that one would usually expect in libgcc like double addition,
1974double comparisons and double conversions. &lsquo;<samp>math</samp>&rsquo; also adds routines
1975that one would expect in <samp>libm.a</samp>, but with <code>__</code> (two underscores)
1976prepended to the symbol names as specified by <samp>math.h</samp>.
1977&lsquo;<samp>math-symbols</samp>&rsquo; also defines weak aliases for the functions
1978declared in <samp>math.h</samp>.  However, <code>--with-libf7</code> won&rsquo;t
1979install no <samp>math.h</samp> header file whatsoever, this file must come
1980from elsewhere.  This option sets <samp>--with-double-comparison</samp>
1981to &lsquo;<samp>bool</samp>&rsquo;.
1982</p>
1983</dd>
1984<dt><span><code>--with-nds32-lib=<var>library</var></code></span></dt>
1985<dd><p>Specifies that <var>library</var> setting is used for building <samp>libgcc.a</samp>.
1986Currently, the valid <var>library</var> is &lsquo;<samp>newlib</samp>&rsquo; or &lsquo;<samp>mculib</samp>&rsquo;.
1987This option is only supported for the NDS32 target.
1988</p>
1989</dd>
1990<dt><span><code>--with-build-time-tools=<var>dir</var></code></span></dt>
1991<dd><p>Specifies where to find the set of target tools (assembler, linker, etc.)
1992that will be used while building GCC itself.  This option can be useful
1993if the directory layouts are different between the system you are building
1994GCC on, and the system where you will deploy it.
1995</p>
1996<p>For example, on an &lsquo;<samp>ia64-hp-hpux</samp>&rsquo; system, you may have the GNU
1997assembler and linker in <samp>/usr/bin</samp>, and the native tools in a
1998different path, and build a toolchain that expects to find the
1999native tools in <samp>/usr/bin</samp>.
2000</p>
2001<p>When you use this option, you should ensure that <var>dir</var> includes
2002<code>ar</code>, <code>as</code>, <code>ld</code>, <code>nm</code>,
2003<code>ranlib</code> and <code>strip</code> if necessary, and possibly
2004<code>objdump</code>.  Otherwise, GCC may use an inconsistent set of
2005tools.
2006</p></dd>
2007</dl>
2008
2009<span id="Overriding-configure-test-results"></span><h4 class="subsubheading">Overriding <code>configure</code> test results</h4>
2010
2011<p>Sometimes, it might be necessary to override the result of some
2012<code>configure</code> test, for example in order to ease porting to a new
2013system or work around a bug in a test.  The toplevel <code>configure</code>
2014script provides three variables for this:
2015</p>
2016<dl compact="compact">
2017<dt id='index-build_005fconfigargs'><span><code>build_configargs</code><a href='#index-build_005fconfigargs' class='copiable-anchor'> &para;</a></span></dt>
2018<dd><p>The contents of this variable is passed to all build <code>configure</code>
2019scripts.
2020</p>
2021</dd>
2022<dt id='index-host_005fconfigargs'><span><code>host_configargs</code><a href='#index-host_005fconfigargs' class='copiable-anchor'> &para;</a></span></dt>
2023<dd><p>The contents of this variable is passed to all host <code>configure</code>
2024scripts.
2025</p>
2026</dd>
2027<dt id='index-target_005fconfigargs'><span><code>target_configargs</code><a href='#index-target_005fconfigargs' class='copiable-anchor'> &para;</a></span></dt>
2028<dd><p>The contents of this variable is passed to all target <code>configure</code>
2029scripts.
2030</p>
2031</dd>
2032</dl>
2033
2034<p>In order to avoid shell and <code>make</code> quoting issues for complex
2035overrides, you can pass a setting for <code>CONFIG_SITE</code> and set
2036variables in the site file.
2037</p>
2038<span id="Objective-C-Specific-Options"></span><h4 class="subheading">Objective-C-Specific Options</h4>
2039
2040<p>The following options apply to the build of the Objective-C runtime library.
2041</p>
2042<dl compact="compact">
2043<dt><span><code>--enable-objc-gc</code></span></dt>
2044<dd><p>Specify that an additional variant of the GNU Objective-C runtime library
2045is built, using an external build of the Boehm-Demers-Weiser garbage
2046collector (<a href="https://www.hboehm.info/gc/">https://www.hboehm.info/gc/</a>).  This library needs to be
2047available for each multilib variant, unless configured with
2048<samp>--enable-objc-gc=&lsquo;<samp>auto</samp>&rsquo;</samp> in which case the build of the
2049additional runtime library is skipped when not available and the build
2050continues.
2051</p>
2052</dd>
2053<dt><span><code>--with-target-bdw-gc=<var>list</var></code></span></dt>
2054<dt><span><code>--with-target-bdw-gc-include=<var>list</var></code></span></dt>
2055<dt><span><code>--with-target-bdw-gc-lib=<var>list</var></code></span></dt>
2056<dd><p>Specify search directories for the garbage collector header files and
2057libraries. <var>list</var> is a comma separated list of key value pairs of the
2058form &lsquo;<samp><var>multilibdir</var>=<var>path</var></samp>&rsquo;, where the default multilib key
2059is named as &lsquo;<samp>.</samp>&rsquo; (dot), or is omitted (e.g.
2060&lsquo;<samp>--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32</samp>&rsquo;).
2061</p>
2062<p>The options <samp>--with-target-bdw-gc-include</samp> and
2063<samp>--with-target-bdw-gc-lib</samp> must always be specified together
2064for each multilib variant and they take precedence over
2065<samp>--with-target-bdw-gc</samp>.  If <samp>--with-target-bdw-gc-include</samp>
2066is missing values for a multilib, then the value for the default
2067multilib is used (e.g. &lsquo;<samp>--with-target-bdw-gc-include=/opt/bdw-gc/include</samp>&rsquo;
2068&lsquo;<samp>--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32</samp>&rsquo;).
2069If none of these options are specified, the library is assumed in
2070default locations.
2071</p></dd>
2072</dl>
2073
2074<span id="D-Specific-Options"></span><h4 class="subheading">D-Specific Options</h4>
2075
2076<p>The following options apply to the build of the D runtime library.
2077</p>
2078<dl compact="compact">
2079<dt><span><code>--enable-libphobos-checking</code></span></dt>
2080<dt><span><code>--disable-libphobos-checking</code></span></dt>
2081<dt><span><code>--enable-libphobos-checking=<var>list</var></code></span></dt>
2082<dd><p>This option controls whether run-time checks and contracts are compiled into
2083the D runtime library.  When the option is not specified, the library is built
2084with &lsquo;<samp>release</samp>&rsquo; checking.  When the option is specified without a
2085<var>list</var>, the result is the same as &lsquo;<samp>--enable-libphobos-checking=yes</samp>&rsquo;.
2086Likewise, &lsquo;<samp>--disable-libphobos-checking</samp>&rsquo; is equivalent to
2087&lsquo;<samp>--enable-libphobos-checking=no</samp>&rsquo;.
2088</p>
2089<p>The categories of checks available in <var>list</var> are &lsquo;<samp>yes</samp>&rsquo; (compiles
2090libphobos with <samp>-fno-release</samp>), &lsquo;<samp>no</samp>&rsquo; (compiles libphobos with
2091<samp>-frelease</samp>), &lsquo;<samp>all</samp>&rsquo; (same as &lsquo;<samp>yes</samp>&rsquo;), &lsquo;<samp>none</samp>&rsquo; or
2092&lsquo;<samp>release</samp>&rsquo; (same as &lsquo;<samp>no</samp>&rsquo;).
2093</p>
2094<p>Individual checks available in <var>list</var> are &lsquo;<samp>assert</samp>&rsquo; (compiles libphobos
2095with an extra option <samp>-fassert</samp>).
2096</p>
2097</dd>
2098<dt><span><code>--with-libphobos-druntime-only</code></span></dt>
2099<dt><span><code>--with-libphobos-druntime-only=<var>choice</var></code></span></dt>
2100<dd><p>Specify whether to build only the core D runtime library (druntime), or both
2101the core and standard library (phobos) into libphobos.  This is useful for
2102targets that have full support in druntime, but no or incomplete support
2103in phobos.  <var>choice</var> can be one of &lsquo;<samp>auto</samp>&rsquo;, &lsquo;<samp>yes</samp>&rsquo;, and &lsquo;<samp>no</samp>&rsquo;
2104where &lsquo;<samp>auto</samp>&rsquo; is the default.
2105</p>
2106<p>When the option is not specified, the default choice &lsquo;<samp>auto</samp>&rsquo; means that it
2107is inferred whether the target has support for the phobos standard library.
2108When the option is specified without a <var>choice</var>,  the result is the same as
2109&lsquo;<samp>--with-libphobos-druntime-only=yes</samp>&rsquo;.
2110</p>
2111</dd>
2112<dt><span><code>--with-target-system-zlib</code></span></dt>
2113<dd><p>Use installed &lsquo;<samp>zlib</samp>&rsquo; rather than that included with GCC.  This needs
2114to be available for each multilib variant, unless configured with
2115<samp>--with-target-system-zlib=&lsquo;<samp>auto</samp>&rsquo;</samp> in which case the GCC&nbsp;included
2116&lsquo;<samp>zlib</samp>&rsquo; is only used when the system installed library is not available.
2117</p></dd>
2118</dl>
2119
2120<hr />
2121<p>
2122<p><a href="./index.html">Return to the GCC Installation page</a>
2123</p>
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135</body>
2136</html>
2137