configure.html revision 1.13
1<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> 2<html> 3<!-- Copyright (C) 1988-2020 Free Software Foundation, Inc. 4 5Permission is granted to copy, distribute and/or modify this document 6under the terms of the GNU Free Documentation License, Version 1.3 or 7any later version published by the Free Software Foundation; with no 8Invariant Sections, the Front-Cover texts being (a) (see below), and 9with the Back-Cover Texts being (b) (see below). A copy of the 10license is included in the section entitled "GNU 11Free Documentation License". 12 13(a) The FSF's Front-Cover Text is: 14 15A GNU Manual 16 17(b) The FSF's Back-Cover Text is: 18 19You have freedom to copy and modify this GNU Manual, like GNU 20 software. Copies published by the Free Software Foundation raise 21 funds for GNU development. --> 22<!-- Created by GNU Texinfo 6.5, http://www.gnu.org/software/texinfo/ --> 23<head> 24<meta http-equiv="Content-Type" content="text/html; charset=utf-8"> 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<style type="text/css"> 33<!-- 34a.summary-letter {text-decoration: none} 35blockquote.indentedblock {margin-right: 0em} 36blockquote.smallindentedblock {margin-right: 0em; font-size: smaller} 37blockquote.smallquotation {font-size: smaller} 38div.display {margin-left: 3.2em} 39div.example {margin-left: 3.2em} 40div.lisp {margin-left: 3.2em} 41div.smalldisplay {margin-left: 3.2em} 42div.smallexample {margin-left: 3.2em} 43div.smalllisp {margin-left: 3.2em} 44kbd {font-style: oblique} 45pre.display {font-family: inherit} 46pre.format {font-family: inherit} 47pre.menu-comment {font-family: serif} 48pre.menu-preformatted {font-family: serif} 49pre.smalldisplay {font-family: inherit; font-size: smaller} 50pre.smallexample {font-size: smaller} 51pre.smallformat {font-family: inherit; font-size: smaller} 52pre.smalllisp {font-size: smaller} 53span.nolinebreak {white-space: nowrap} 54span.roman {font-family: initial; font-weight: normal} 55span.sansserif {font-family: sans-serif; font-weight: normal} 56ul.no-bullet {list-style: none} 57--> 58</style> 59 60 61</head> 62 63<body lang="en"> 64<h1 class="settitle" align="center">Installing GCC: Configuration</h1> 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84<a name="index-Configuration"></a> 85<a name="index-Installing-GCC_003a-Configuration"></a> 86 87<p>Like most GNU software, GCC must be configured before it can be built. 88This document describes the recommended configuration procedure 89for both native and cross targets. 90</p> 91<p>We use <var>srcdir</var> to refer to the toplevel source directory for 92GCC; we use <var>objdir</var> to refer to the toplevel build/object directory. 93</p> 94<p>If you obtained the sources by cloning the repository, <var>srcdir</var> 95must refer to the top <samp>gcc</samp> directory, the one where the 96<samp>MAINTAINERS</samp> file can be found, and not its <samp>gcc</samp> 97subdirectory, otherwise the build will fail. 98</p> 99<p>If either <var>srcdir</var> or <var>objdir</var> is located on an automounted NFS 100file system, the shell’s built-in <code>pwd</code> command will return 101temporary pathnames. Using these can lead to various sorts of build 102problems. To avoid this issue, set the <code>PWDCMD</code> environment 103variable to an automounter-aware <code>pwd</code> command, e.g., 104<code>pawd</code> or ‘<samp>amq -w</samp>’, during the configuration and build 105phases. 106</p> 107<p>First, we <strong>highly</strong> recommend that GCC be built into a 108separate directory from the sources which does <strong>not</strong> reside 109within the source tree. This is how we generally build GCC; building 110where <var>srcdir</var> == <var>objdir</var> should still work, but doesn’t 111get extensive testing; building where <var>objdir</var> is a subdirectory 112of <var>srcdir</var> is unsupported. 113</p> 114<p>If you have previously built GCC in the same directory for a 115different target machine, do ‘<samp>make distclean</samp>’ to delete all files 116that might be invalid. One of the files this deletes is <samp>Makefile</samp>; 117if ‘<samp>make distclean</samp>’ complains that <samp>Makefile</samp> does not exist 118or issues a message like “don’t know how to make distclean” it probably 119means that the directory is already suitably clean. However, with the 120recommended method of building in a separate <var>objdir</var>, you should 121simply use a different <var>objdir</var> for each target. 122</p> 123<p>Second, when configuring a native system, either <code>cc</code> or 124<code>gcc</code> must be in your path or you must set <code>CC</code> in 125your environment before running configure. Otherwise the configuration 126scripts may fail. 127</p> 128 129<p>To configure GCC: 130</p> 131<div class="smallexample"> 132<pre class="smallexample">% mkdir <var>objdir</var> 133% cd <var>objdir</var> 134% <var>srcdir</var>/configure [<var>options</var>] [<var>target</var>] 135</pre></div> 136 137<a name="Distributor-options"></a> 138<h3 class="heading">Distributor options</h3> 139 140<p>If you will be distributing binary versions of GCC, with modifications 141to the source code, you should use the options described in this 142section to make clear that your version contains modifications. 143</p> 144<dl compact="compact"> 145<dt><code>--with-pkgversion=<var>version</var></code></dt> 146<dd><p>Specify a string that identifies your package. You may wish 147to include a build number or build date. This version string will be 148included in the output of <code>gcc --version</code>. This suffix does 149not replace the default version string, only the ‘<samp>GCC</samp>’ part. 150</p> 151<p>The default value is ‘<samp>GCC</samp>’. 152</p> 153</dd> 154<dt><code>--with-bugurl=<var>url</var></code></dt> 155<dd><p>Specify the URL that users should visit if they wish to report a bug. 156You are of course welcome to forward bugs reported to you to the FSF, 157if you determine that they are not bugs in your modifications. 158</p> 159<p>The default value refers to the FSF’s GCC bug tracker. 160</p> 161</dd> 162<dt><code>--with-documentation-root-url=<var>url</var></code></dt> 163<dd><p>Specify the URL root that contains GCC option documentation. The <var>url</var> 164should end with a <code>/</code> character. 165</p> 166<p>The default value is <a href="https://gcc.gnu.org/onlinedocs/">https://gcc.gnu.org/onlinedocs/</a>. 167</p> 168</dd> 169<dt><code>--with-changes-root-url=<var>url</var></code></dt> 170<dd><p>Specify the URL root that contains information about changes in GCC 171releases like <code>gcc-<var>version</var>/changes.html</code>. 172The <var>url</var> should end with a <code>/</code> character. 173</p> 174<p>The default value is <a href="https://gcc.gnu.org/">https://gcc.gnu.org/</a>. 175</p> 176</dd> 177</dl> 178 179<a name="Target-specification"></a> 180<h3 class="heading">Target specification</h3> 181<ul> 182<li> GCC has code to correctly determine the correct value for <var>target</var> 183for nearly all native systems. Therefore, we highly recommend you do 184not provide a configure target when configuring a native compiler. 185 186</li><li> <var>target</var> must be specified as <samp>--target=<var>target</var></samp> 187when configuring a cross compiler; examples of valid targets would be 188m68k-elf, sh-elf, etc. 189 190</li><li> Specifying just <var>target</var> instead of <samp>--target=<var>target</var></samp> 191implies that the host defaults to <var>target</var>. 192</li></ul> 193 194 195<a name="Options-specification"></a> 196<h3 class="heading">Options specification</h3> 197 198<p>Use <var>options</var> to override several configure time options for 199GCC. A list of supported <var>options</var> follows; ‘<samp>configure 200--help</samp>’ may list other options, but those not listed below may not 201work and should not normally be used. 202</p> 203<p>Note that each <samp>--enable</samp> option has a corresponding 204<samp>--disable</samp> option and that each <samp>--with</samp> option has a 205corresponding <samp>--without</samp> option. 206</p> 207<dl compact="compact"> 208<dt><code>--prefix=<var>dirname</var></code></dt> 209<dd><p>Specify the toplevel installation 210directory. This is the recommended way to install the tools into a directory 211other than the default. The toplevel installation directory defaults to 212<samp>/usr/local</samp>. 213</p> 214<p>We <strong>highly</strong> recommend against <var>dirname</var> being the same or a 215subdirectory of <var>objdir</var> or vice versa. If specifying a directory 216beneath a user’s home directory tree, some shells will not expand 217<var>dirname</var> correctly if it contains the ‘<samp>~</samp>’ metacharacter; use 218<code>$HOME</code> instead. 219</p> 220<p>The following standard <code>autoconf</code> options are supported. Normally you 221should not need to use these options. 222</p><dl compact="compact"> 223<dt><code>--exec-prefix=<var>dirname</var></code></dt> 224<dd><p>Specify the toplevel installation directory for architecture-dependent 225files. The default is <samp><var>prefix</var></samp>. 226</p> 227</dd> 228<dt><code>--bindir=<var>dirname</var></code></dt> 229<dd><p>Specify the installation directory for the executables called by users 230(such as <code>gcc</code> and <code>g++</code>). The default is 231<samp><var>exec-prefix</var>/bin</samp>. 232</p> 233</dd> 234<dt><code>--libdir=<var>dirname</var></code></dt> 235<dd><p>Specify the installation directory for object code libraries and 236internal data files of GCC. The default is <samp><var>exec-prefix</var>/lib</samp>. 237</p> 238</dd> 239<dt><code>--libexecdir=<var>dirname</var></code></dt> 240<dd><p>Specify the installation directory for internal executables of GCC. 241The default is <samp><var>exec-prefix</var>/libexec</samp>. 242</p> 243</dd> 244<dt><code>--with-slibdir=<var>dirname</var></code></dt> 245<dd><p>Specify the installation directory for the shared libgcc library. The 246default is <samp><var>libdir</var></samp>. 247</p> 248</dd> 249<dt><code>--datarootdir=<var>dirname</var></code></dt> 250<dd><p>Specify the root of the directory tree for read-only architecture-independent 251data files referenced by GCC. The default is <samp><var>prefix</var>/share</samp>. 252</p> 253</dd> 254<dt><code>--infodir=<var>dirname</var></code></dt> 255<dd><p>Specify the installation directory for documentation in info format. 256The default is <samp><var>datarootdir</var>/info</samp>. 257</p> 258</dd> 259<dt><code>--datadir=<var>dirname</var></code></dt> 260<dd><p>Specify the installation directory for some architecture-independent 261data files referenced by GCC. The default is <samp><var>datarootdir</var></samp>. 262</p> 263</dd> 264<dt><code>--docdir=<var>dirname</var></code></dt> 265<dd><p>Specify the installation directory for documentation files (other 266than Info) for GCC. The default is <samp><var>datarootdir</var>/doc</samp>. 267</p> 268</dd> 269<dt><code>--htmldir=<var>dirname</var></code></dt> 270<dd><p>Specify the installation directory for HTML documentation files. 271The default is <samp><var>docdir</var></samp>. 272</p> 273</dd> 274<dt><code>--pdfdir=<var>dirname</var></code></dt> 275<dd><p>Specify the installation directory for PDF documentation files. 276The default is <samp><var>docdir</var></samp>. 277</p> 278</dd> 279<dt><code>--mandir=<var>dirname</var></code></dt> 280<dd><p>Specify the installation directory for manual pages. The default is 281<samp><var>datarootdir</var>/man</samp>. (Note that the manual pages are only extracts 282from the full GCC manuals, which are provided in Texinfo format. The manpages 283are derived by an automatic conversion process from parts of the full 284manual.) 285</p> 286</dd> 287<dt><code>--with-gxx-include-dir=<var>dirname</var></code></dt> 288<dd><p>Specify 289the installation directory for G++ header files. The default depends 290on other configuration options, and differs between cross and native 291configurations. 292</p> 293</dd> 294<dt><code>--with-specs=<var>specs</var></code></dt> 295<dd><p>Specify additional command line driver SPECS. 296This can be useful if you need to turn on a non-standard feature by 297default without modifying the compiler’s source code, for instance 298<samp>--with-specs=%{!fcommon:%{!fno-common:-fno-common}}</samp>. 299See “Spec Files” in the main manual 300</p> 301</dd> 302</dl> 303 304</dd> 305<dt><code>--program-prefix=<var>prefix</var></code></dt> 306<dd><p>GCC supports some transformations of the names of its programs when 307installing them. This option prepends <var>prefix</var> to the names of 308programs to install in <var>bindir</var> (see above). For example, specifying 309<samp>--program-prefix=foo-</samp> would result in ‘<samp>gcc</samp>’ 310being installed as <samp>/usr/local/bin/foo-gcc</samp>. 311</p> 312</dd> 313<dt><code>--program-suffix=<var>suffix</var></code></dt> 314<dd><p>Appends <var>suffix</var> to the names of programs to install in <var>bindir</var> 315(see above). For example, specifying <samp>--program-suffix=-3.1</samp> 316would result in ‘<samp>gcc</samp>’ being installed as 317<samp>/usr/local/bin/gcc-3.1</samp>. 318</p> 319</dd> 320<dt><code>--program-transform-name=<var>pattern</var></code></dt> 321<dd><p>Applies the ‘<samp>sed</samp>’ script <var>pattern</var> to be applied to the names 322of programs to install in <var>bindir</var> (see above). <var>pattern</var> has to 323consist of one or more basic ‘<samp>sed</samp>’ editing commands, separated by 324semicolons. For example, if you want the ‘<samp>gcc</samp>’ program name to be 325transformed to the installed program <samp>/usr/local/bin/myowngcc</samp> and 326the ‘<samp>g++</samp>’ program name to be transformed to 327<samp>/usr/local/bin/gspecial++</samp> without changing other program names, 328you could use the pattern 329<samp>--program-transform-name='s/^gcc$/myowngcc/; s/^g++$/gspecial++/'</samp> 330to achieve this effect. 331</p> 332<p>All three options can be combined and used together, resulting in more 333complex conversion patterns. As a basic rule, <var>prefix</var> (and 334<var>suffix</var>) are prepended (appended) before further transformations 335can happen with a special transformation script <var>pattern</var>. 336</p> 337<p>As currently implemented, this option only takes effect for native 338builds; cross compiler binaries’ names are not transformed even when a 339transformation is explicitly asked for by one of these options. 340</p> 341<p>For native builds, some of the installed programs are also installed 342with the target alias in front of their name, as in 343‘<samp>i686-pc-linux-gnu-gcc</samp>’. All of the above transformations happen 344before the target alias is prepended to the name—so, specifying 345<samp>--program-prefix=foo-</samp> and <samp>program-suffix=-3.1</samp>, the 346resulting binary would be installed as 347<samp>/usr/local/bin/i686-pc-linux-gnu-foo-gcc-3.1</samp>. 348</p> 349<p>As a last shortcoming, none of the installed Ada programs are 350transformed yet, which will be fixed in some time. 351</p> 352</dd> 353<dt><code>--with-local-prefix=<var>dirname</var></code></dt> 354<dd><p>Specify the 355installation directory for local include files. The default is 356<samp>/usr/local</samp>. Specify this option if you want the compiler to 357search directory <samp><var>dirname</var>/include</samp> for locally installed 358header files <em>instead</em> of <samp>/usr/local/include</samp>. 359</p> 360<p>You should specify <samp>--with-local-prefix</samp> <strong>only</strong> if your 361site has a different convention (not <samp>/usr/local</samp>) for where to put 362site-specific files. 363</p> 364<p>The default value for <samp>--with-local-prefix</samp> is <samp>/usr/local</samp> 365regardless of the value of <samp>--prefix</samp>. Specifying 366<samp>--prefix</samp> has no effect on which directory GCC searches for 367local header files. This may seem counterintuitive, but actually it is 368logical. 369</p> 370<p>The purpose of <samp>--prefix</samp> is to specify where to <em>install 371GCC</em>. The local header files in <samp>/usr/local/include</samp>—if you put 372any in that directory—are not part of GCC. They are part of other 373programs—perhaps many others. (GCC installs its own header files in 374another directory which is based on the <samp>--prefix</samp> value.) 375</p> 376<p>Both the local-prefix include directory and the GCC-prefix include 377directory are part of GCC’s “system include” directories. Although these 378two directories are not fixed, they need to be searched in the proper 379order for the correct processing of the include_next directive. The 380local-prefix include directory is searched before the GCC-prefix 381include directory. Another characteristic of system include directories 382is that pedantic warnings are turned off for headers in these directories. 383</p> 384<p>Some autoconf macros add <samp>-I <var>directory</var></samp> options to the 385compiler command line, to ensure that directories containing installed 386packages’ headers are searched. When <var>directory</var> is one of GCC’s 387system include directories, GCC will ignore the option so that system 388directories continue to be processed in the correct order. This 389may result in a search order different from what was specified but the 390directory will still be searched. 391</p> 392<p>GCC automatically searches for ordinary libraries using 393<code>GCC_EXEC_PREFIX</code>. Thus, when the same installation prefix is 394used for both GCC and packages, GCC will automatically search for 395both headers and libraries. This provides a configuration that is 396easy to use. GCC behaves in a manner similar to that when it is 397installed as a system compiler in <samp>/usr</samp>. 398</p> 399<p>Sites that need to install multiple versions of GCC may not want to 400use the above simple configuration. It is possible to use the 401<samp>--program-prefix</samp>, <samp>--program-suffix</samp> and 402<samp>--program-transform-name</samp> options to install multiple versions 403into a single directory, but it may be simpler to use different prefixes 404and the <samp>--with-local-prefix</samp> option to specify the location of the 405site-specific files for each version. It will then be necessary for 406users to specify explicitly the location of local site libraries 407(e.g., with <code>LIBRARY_PATH</code>). 408</p> 409<p>The same value can be used for both <samp>--with-local-prefix</samp> and 410<samp>--prefix</samp> provided it is not <samp>/usr</samp>. This can be used 411to avoid the default search of <samp>/usr/local/include</samp>. 412</p> 413<p><strong>Do not</strong> specify <samp>/usr</samp> as the <samp>--with-local-prefix</samp>! 414The directory you use for <samp>--with-local-prefix</samp> <strong>must not</strong> 415contain any of the system’s standard header files. If it did contain 416them, certain programs would be miscompiled (including GNU Emacs, on 417certain targets), because this would override and nullify the header 418file corrections made by the <code>fixincludes</code> script. 419</p> 420<p>Indications are that people who use this option use it based on mistaken 421ideas of what it is for. People use it as if it specified where to 422install part of GCC. Perhaps they make this assumption because 423installing GCC creates the directory. 424</p> 425</dd> 426<dt><code>--with-gcc-major-version-only</code></dt> 427<dd><p>Specifies that GCC should use only the major number rather than 428<var>major</var>.<var>minor</var>.<var>patchlevel</var> in filesystem paths. 429</p> 430</dd> 431<dt><code>--with-native-system-header-dir=<var>dirname</var></code></dt> 432<dd><p>Specifies that <var>dirname</var> is the directory that contains native system 433header files, rather than <samp>/usr/include</samp>. This option is most useful 434if you are creating a compiler that should be isolated from the system 435as much as possible. It is most commonly used with the 436<samp>--with-sysroot</samp> option and will cause GCC to search 437<var>dirname</var> inside the system root specified by that option. 438</p> 439</dd> 440<dt><code>--enable-shared[=<var>package</var>[,…]]</code></dt> 441<dd><p>Build shared versions of libraries, if shared libraries are supported on 442the target platform. Unlike GCC 2.95.x and earlier, shared libraries 443are enabled by default on all platforms that support shared libraries. 444</p> 445<p>If a list of packages is given as an argument, build shared libraries 446only for the listed packages. For other packages, only static libraries 447will be built. Package names currently recognized in the GCC tree are 448‘<samp>libgcc</samp>’ (also known as ‘<samp>gcc</samp>’), ‘<samp>libstdc++</samp>’ (not 449‘<samp>libstdc++-v3</samp>’), ‘<samp>libffi</samp>’, ‘<samp>zlib</samp>’, ‘<samp>boehm-gc</samp>’, 450‘<samp>ada</samp>’, ‘<samp>libada</samp>’, ‘<samp>libgo</samp>’, ‘<samp>libobjc</samp>’, and ‘<samp>libphobos</samp>’. 451Note ‘<samp>libiberty</samp>’ does not support shared libraries at all. 452</p> 453<p>Use <samp>--disable-shared</samp> to build only static libraries. Note that 454<samp>--disable-shared</samp> does not accept a list of package names as 455argument, only <samp>--enable-shared</samp> does. 456</p> 457<p>Contrast with <samp>--enable-host-shared</samp>, which affects <em>host</em> 458code. 459</p> 460</dd> 461<dt><code>--enable-host-shared</code></dt> 462<dd><p>Specify that the <em>host</em> code should be built into position-independent 463machine code (with -fPIC), allowing it to be used within shared libraries, 464but yielding a slightly slower compiler. 465</p> 466<p>This option is required when building the libgccjit.so library. 467</p> 468<p>Contrast with <samp>--enable-shared</samp>, which affects <em>target</em> 469libraries. 470</p> 471</dd> 472<dt><code><a name="with-gnu-as"></a>--with-gnu-as</code></dt> 473<dd><p>Specify that the compiler should assume that the 474assembler it finds is the GNU assembler. However, this does not modify 475the rules to find an assembler and will result in confusion if the 476assembler found is not actually the GNU assembler. (Confusion may also 477result if the compiler finds the GNU assembler but has not been 478configured with <samp>--with-gnu-as</samp>.) If you have more than one 479assembler installed on your system, you may want to use this option in 480connection with <samp>--with-as=<var>pathname</var></samp> or 481<samp>--with-build-time-tools=<var>pathname</var></samp>. 482</p> 483<p>The following systems are the only ones where it makes a difference 484whether you use the GNU assembler. On any other system, 485<samp>--with-gnu-as</samp> has no effect. 486</p> 487<ul> 488<li> ‘<samp>hppa1.0-<var>any</var>-<var>any</var></samp>’ 489</li><li> ‘<samp>hppa1.1-<var>any</var>-<var>any</var></samp>’ 490</li><li> ‘<samp>sparc-sun-solaris2.<var>any</var></samp>’ 491</li><li> ‘<samp>sparc64-<var>any</var>-solaris2.<var>any</var></samp>’ 492</li></ul> 493 494</dd> 495<dt><code><a name="with-as"></a>--with-as=<var>pathname</var></code></dt> 496<dd><p>Specify that the compiler should use the assembler pointed to by 497<var>pathname</var>, rather than the one found by the standard rules to find 498an assembler, which are: 499</p><ul> 500<li> Unless GCC is being built with a cross compiler, check the 501<samp><var>libexec</var>/gcc/<var>target</var>/<var>version</var></samp> directory. 502<var>libexec</var> defaults to <samp><var>exec-prefix</var>/libexec</samp>; 503<var>exec-prefix</var> defaults to <var>prefix</var>, which 504defaults to <samp>/usr/local</samp> unless overridden by the 505<samp>--prefix=<var>pathname</var></samp> switch described above. <var>target</var> 506is the target system triple, such as ‘<samp>sparc-sun-solaris2.7</samp>’, and 507<var>version</var> denotes the GCC version, such as 3.0. 508 509</li><li> If the target system is the same that you are building on, check 510operating system specific directories (e.g. <samp>/usr/ccs/bin</samp> on 511Solaris 2). 512 513</li><li> Check in the <code>PATH</code> for a tool whose name is prefixed by the 514target system triple. 515 516</li><li> Check in the <code>PATH</code> for a tool whose name is not prefixed by the 517target system triple, if the host and target system triple are 518the same (in other words, we use a host tool if it can be used for 519the target as well). 520</li></ul> 521 522<p>You may want to use <samp>--with-as</samp> if no assembler 523is installed in the directories listed above, or if you have multiple 524assemblers installed and want to choose one that is not found by the 525above rules. 526</p> 527</dd> 528<dt><code><a name="with-gnu-ld"></a>--with-gnu-ld</code></dt> 529<dd><p>Same as <a href="#with-gnu-as"><samp>--with-gnu-as</samp></a> 530but for the linker. 531</p> 532</dd> 533<dt><code>--with-ld=<var>pathname</var></code></dt> 534<dd><p>Same as <a href="#with-as"><samp>--with-as</samp></a> 535but for the linker. 536</p> 537</dd> 538<dt><code>--with-stabs</code></dt> 539<dd><p>Specify that stabs debugging 540information should be used instead of whatever format the host normally 541uses. Normally GCC uses the same debug format as the host system. 542</p> 543</dd> 544<dt><code>--with-tls=<var>dialect</var></code></dt> 545<dd><p>Specify the default TLS dialect, for systems were there is a choice. 546For ARM targets, possible values for <var>dialect</var> are <code>gnu</code> or 547<code>gnu2</code>, which select between the original GNU dialect and the GNU TLS 548descriptor-based dialect. 549</p> 550</dd> 551<dt><code>--enable-multiarch</code></dt> 552<dd><p>Specify whether to enable or disable multiarch support. The default is 553to check for glibc start files in a multiarch location, and enable it 554if the files are found. The auto detection is enabled for native builds, 555and for cross builds configured with <samp>--with-sysroot</samp>, and without 556<samp>--with-native-system-header-dir</samp>. 557More documentation about multiarch can be found at 558<a href="https://wiki.debian.org/Multiarch">https://wiki.debian.org/Multiarch</a>. 559</p> 560</dd> 561<dt><code>--enable-sjlj-exceptions</code></dt> 562<dd><p>Force use of the <code>setjmp</code>/<code>longjmp</code>-based scheme for exceptions. 563‘<samp>configure</samp>’ ordinarily picks the correct value based on the platform. 564Only use this option if you are sure you need a different setting. 565</p> 566</dd> 567<dt><code>--enable-vtable-verify</code></dt> 568<dd><p>Specify whether to enable or disable the vtable verification feature. 569Enabling this feature causes libstdc++ to be built with its virtual calls 570in verifiable mode. This means that, when linked with libvtv, every 571virtual call in libstdc++ will verify the vtable pointer through which the 572call will be made before actually making the call. If not linked with libvtv, 573the verifier will call stub functions (in libstdc++ itself) and do nothing. 574If vtable verification is disabled, then libstdc++ is not built with its 575virtual calls in verifiable mode at all. However the libvtv library will 576still be built (see <samp>--disable-libvtv</samp> to turn off building libvtv). 577<samp>--disable-vtable-verify</samp> is the default. 578</p> 579</dd> 580<dt><code>--disable-gcov</code></dt> 581<dd><p>Specify that the run-time library used for coverage analysis 582and associated host tools should not be built. 583</p> 584</dd> 585<dt><code>--disable-multilib</code></dt> 586<dd><p>Specify that multiple target 587libraries to support different target variants, calling 588conventions, etc. should not be built. The default is to build a 589predefined set of them. 590</p> 591<p>Some targets provide finer-grained control over which multilibs are built 592(e.g., <samp>--disable-softfloat</samp>): 593</p><dl compact="compact"> 594<dt><code>arm-*-*</code></dt> 595<dd><p>fpu, 26bit, underscore, interwork, biendian, nofmult. 596</p> 597</dd> 598<dt><code>m68*-*-*</code></dt> 599<dd><p>softfloat, m68881, m68000, m68020. 600</p> 601</dd> 602<dt><code>mips*-*-*</code></dt> 603<dd><p>single-float, biendian, softfloat. 604</p> 605</dd> 606<dt><code>msp430-*-*</code></dt> 607<dd><p>no-exceptions 608</p> 609</dd> 610<dt><code>powerpc*-*-*, rs6000*-*-*</code></dt> 611<dd><p>aix64, pthread, softfloat, powercpu, powerpccpu, powerpcos, biendian, 612sysv, aix. 613</p> 614</dd> 615</dl> 616 617</dd> 618<dt><code>--with-multilib-list=<var>list</var></code></dt> 619<dt><code>--without-multilib-list</code></dt> 620<dd><p>Specify what multilibs to build. <var>list</var> is a comma separated list of 621values, possibly consisting of a single value. Currently only implemented 622for aarch64*-*-*, arm*-*-*, riscv*-*-*, sh*-*-* and x86-64-*-linux*. The 623accepted values and meaning for each target is given below. 624</p> 625<dl compact="compact"> 626<dt><code>aarch64*-*-*</code></dt> 627<dd><p><var>list</var> is a comma separated list of <code>ilp32</code>, and <code>lp64</code> 628to enable ILP32 and LP64 run-time libraries, respectively. If 629<var>list</var> is empty, then there will be no multilibs and only the 630default run-time library will be built. If <var>list</var> is 631<code>default</code> or –with-multilib-list= is not specified, then the 632default set of libraries is selected based on the value of 633<samp>--target</samp>. 634</p> 635</dd> 636<dt><code>arm*-*-*</code></dt> 637<dd><p><var>list</var> is a comma separated list of <code>aprofile</code> and 638<code>rmprofile</code> to build multilibs for A or R and M architecture 639profiles respectively. Note that, due to some limitation of the current 640multilib framework, using the combined <code>aprofile,rmprofile</code> 641multilibs selects in some cases a less optimal multilib than when using 642the multilib profile for the architecture targetted. The special value 643<code>default</code> is also accepted and is equivalent to omitting the 644option, i.e., only the default run-time library will be enabled. 645</p> 646<p><var>list</var> may instead contain <code>@name</code>, to use the multilib 647configuration Makefile fragment <samp>name</samp> in <samp>gcc/config/arm</samp> in 648the source tree (it is part of the corresponding sources, after all). 649It is recommended, but not required, that files used for this purpose to 650be named starting with <samp>t-ml-</samp>, to make their intended purpose 651self-evident, in line with GCC conventions. Such files enable custom, 652user-chosen multilib lists to be configured. Whether multiple such 653files can be used together depends on the contents of the supplied 654files. See <samp>gcc/config/arm/t-multilib</samp> and its supplementary 655<samp>gcc/config/arm/t-*profile</samp> files for an example of what such 656Makefile fragments might look like for this version of GCC. The macros 657expected to be defined in these fragments are not stable across GCC 658releases, so make sure they define the <code>MULTILIB</code>-related macros 659expected by the version of GCC you are building. 660See “Target Makefile Fragments” in the internals manual. 661</p> 662<p>The table below gives the combination of ISAs, architectures, FPUs and 663floating-point ABIs for which multilibs are built for each predefined 664profile. The union of these options is considered when specifying both 665<code>aprofile</code> and <code>rmprofile</code>. 666</p> 667<table> 668<tr><td width="15%">Option</td><td width="28%">aprofile</td><td width="30%">rmprofile</td></tr> 669<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> 670<tr><td width="15%">Architectures<br><br><br><br><br><br></td><td width="28%">default architecture<br> 671<code>-march=armv7-a</code><br> 672<code>-march=armv7ve</code><br> 673<code>-march=armv8-a</code><br><br><br></td><td width="30%">default architecture<br> 674<code>-march=armv6s-m</code><br> 675<code>-march=armv7-m</code><br> 676<code>-march=armv7e-m</code><br> 677<code>-march=armv8-m.base</code><br> 678<code>-march=armv8-m.main</code><br> 679<code>-march=armv7</code></td></tr> 680<tr><td width="15%">FPUs<br><br><br><br><br></td><td width="28%">none<br> 681<code>-mfpu=vfpv3-d16</code><br> 682<code>-mfpu=neon</code><br> 683<code>-mfpu=vfpv4-d16</code><br> 684<code>-mfpu=neon-vfpv4</code><br> 685<code>-mfpu=neon-fp-armv8</code></td><td width="30%">none<br> 686<code>-mfpu=vfpv3-d16</code><br> 687<code>-mfpu=fpv4-sp-d16</code><br> 688<code>-mfpu=fpv5-sp-d16</code><br> 689<code>-mfpu=fpv5-d16</code><br></td></tr> 690<tr><td width="15%">floating-point ABIs<br><br></td><td width="28%"><code>-mfloat-abi=soft</code><br> 691<code>-mfloat-abi=softfp</code><br> 692<code>-mfloat-abi=hard</code></td><td width="30%"><code>-mfloat-abi=soft</code><br> 693<code>-mfloat-abi=softfp</code><br> 694<code>-mfloat-abi=hard</code></td></tr> 695</table> 696 697</dd> 698<dt><code>riscv*-*-*</code></dt> 699<dd><p><var>list</var> is a single ABI name. The target architecture must be either 700<code>rv32gc</code> or <code>rv64gc</code>. This will build a single multilib for the 701specified architecture and ABI pair. If <code>--with-multilib-list</code> is not 702given, then a default set of multilibs is selected based on the value of 703<samp>--target</samp>. This is usually a large set of multilibs. 704</p> 705</dd> 706<dt><code>sh*-*-*</code></dt> 707<dd><p><var>list</var> is a comma separated list of CPU names. These must be of the 708form <code>sh*</code> or <code>m*</code> (in which case they match the compiler option 709for that processor). The list should not contain any endian options - 710these are handled by <samp>--with-endian</samp>. 711</p> 712<p>If <var>list</var> is empty, then there will be no multilibs for extra 713processors. The multilib for the secondary endian remains enabled. 714</p> 715<p>As a special case, if an entry in the list starts with a <code>!</code> 716(exclamation point), then it is added to the list of excluded multilibs. 717Entries of this sort should be compatible with ‘<samp>MULTILIB_EXCLUDES</samp>’ 718(once the leading <code>!</code> has been stripped). 719</p> 720<p>If <samp>--with-multilib-list</samp> is not given, then a default set of 721multilibs is selected based on the value of <samp>--target</samp>. This is 722usually the complete set of libraries, but some targets imply a more 723specialized subset. 724</p> 725<p>Example 1: to configure a compiler for SH4A only, but supporting both 726endians, with little endian being the default: 727</p><div class="smallexample"> 728<pre class="smallexample">--with-cpu=sh4a --with-endian=little,big --with-multilib-list= 729</pre></div> 730 731<p>Example 2: to configure a compiler for both SH4A and SH4AL-DSP, but with 732only little endian SH4AL: 733</p><div class="smallexample"> 734<pre class="smallexample">--with-cpu=sh4a --with-endian=little,big \ 735--with-multilib-list=sh4al,!mb/m4al 736</pre></div> 737 738</dd> 739<dt><code>x86-64-*-linux*</code></dt> 740<dd><p><var>list</var> is a comma separated list of <code>m32</code>, <code>m64</code> and 741<code>mx32</code> to enable 32-bit, 64-bit and x32 run-time libraries, 742respectively. If <var>list</var> is empty, then there will be no multilibs 743and only the default run-time library will be enabled. 744</p> 745<p>If <samp>--with-multilib-list</samp> is not given, then only 32-bit and 74664-bit run-time libraries will be enabled. 747</p></dd> 748</dl> 749 750</dd> 751<dt><code>--with-endian=<var>endians</var></code></dt> 752<dd><p>Specify what endians to use. 753Currently only implemented for sh*-*-*. 754</p> 755<p><var>endians</var> may be one of the following: 756</p><dl compact="compact"> 757<dt><code>big</code></dt> 758<dd><p>Use big endian exclusively. 759</p></dd> 760<dt><code>little</code></dt> 761<dd><p>Use little endian exclusively. 762</p></dd> 763<dt><code>big,little</code></dt> 764<dd><p>Use big endian by default. Provide a multilib for little endian. 765</p></dd> 766<dt><code>little,big</code></dt> 767<dd><p>Use little endian by default. Provide a multilib for big endian. 768</p></dd> 769</dl> 770 771</dd> 772<dt><code>--enable-threads</code></dt> 773<dd><p>Specify that the target 774supports threads. This affects the Objective-C compiler and runtime 775library, and exception handling for other languages like C++. 776On some systems, this is the default. 777</p> 778<p>In general, the best (and, in many cases, the only known) threading 779model available will be configured for use. Beware that on some 780systems, GCC has not been taught what threading models are generally 781available for the system. In this case, <samp>--enable-threads</samp> is an 782alias for <samp>--enable-threads=single</samp>. 783</p> 784</dd> 785<dt><code>--disable-threads</code></dt> 786<dd><p>Specify that threading support should be disabled for the system. 787This is an alias for <samp>--enable-threads=single</samp>. 788</p> 789</dd> 790<dt><code>--enable-threads=<var>lib</var></code></dt> 791<dd><p>Specify that 792<var>lib</var> is the thread support library. This affects the Objective-C 793compiler and runtime library, and exception handling for other languages 794like C++. The possibilities for <var>lib</var> are: 795</p> 796<dl compact="compact"> 797<dt><code>aix</code></dt> 798<dd><p>AIX thread support. 799</p></dd> 800<dt><code>dce</code></dt> 801<dd><p>DCE thread support. 802</p></dd> 803<dt><code>lynx</code></dt> 804<dd><p>LynxOS thread support. 805</p></dd> 806<dt><code>mipssde</code></dt> 807<dd><p>MIPS SDE thread support. 808</p></dd> 809<dt><code>no</code></dt> 810<dd><p>This is an alias for ‘<samp>single</samp>’. 811</p></dd> 812<dt><code>posix</code></dt> 813<dd><p>Generic POSIX/Unix98 thread support. 814</p></dd> 815<dt><code>rtems</code></dt> 816<dd><p>RTEMS thread support. 817</p></dd> 818<dt><code>single</code></dt> 819<dd><p>Disable thread support, should work for all platforms. 820</p></dd> 821<dt><code>tpf</code></dt> 822<dd><p>TPF thread support. 823</p></dd> 824<dt><code>vxworks</code></dt> 825<dd><p>VxWorks thread support. 826</p></dd> 827<dt><code>win32</code></dt> 828<dd><p>Microsoft Win32 API thread support. 829</p></dd> 830</dl> 831 832</dd> 833<dt><code>--enable-tls</code></dt> 834<dd><p>Specify that the target supports TLS (Thread Local Storage). Usually 835configure can correctly determine if TLS is supported. In cases where 836it guesses incorrectly, TLS can be explicitly enabled or disabled with 837<samp>--enable-tls</samp> or <samp>--disable-tls</samp>. This can happen if 838the assembler supports TLS but the C library does not, or if the 839assumptions made by the configure test are incorrect. 840</p> 841</dd> 842<dt><code>--disable-tls</code></dt> 843<dd><p>Specify that the target does not support TLS. 844This is an alias for <samp>--enable-tls=no</samp>. 845</p> 846</dd> 847<dt><code>--disable-tm-clone-registry</code></dt> 848<dd><p>Disable TM clone registry in libgcc. It is enabled in libgcc by default. 849This option helps to reduce code size for embedded targets which do 850not use transactional memory. 851</p> 852</dd> 853<dt><code>--with-cpu=<var>cpu</var></code></dt> 854<dt><code>--with-cpu-32=<var>cpu</var></code></dt> 855<dt><code>--with-cpu-64=<var>cpu</var></code></dt> 856<dd><p>Specify which cpu variant the compiler should generate code for by default. 857<var>cpu</var> will be used as the default value of the <samp>-mcpu=</samp> switch. 858This option is only supported on some targets, including ARC, ARM, i386, M68k, 859PowerPC, and SPARC. It is mandatory for ARC. The <samp>--with-cpu-32</samp> and 860<samp>--with-cpu-64</samp> options specify separate default CPUs for 86132-bit and 64-bit modes; these options are only supported for aarch64, i386, 862x86-64, PowerPC, and SPARC. 863</p> 864</dd> 865<dt><code>--with-schedule=<var>cpu</var></code></dt> 866<dt><code>--with-arch=<var>cpu</var></code></dt> 867<dt><code>--with-arch-32=<var>cpu</var></code></dt> 868<dt><code>--with-arch-64=<var>cpu</var></code></dt> 869<dt><code>--with-tune=<var>cpu</var></code></dt> 870<dt><code>--with-tune-32=<var>cpu</var></code></dt> 871<dt><code>--with-tune-64=<var>cpu</var></code></dt> 872<dt><code>--with-abi=<var>abi</var></code></dt> 873<dt><code>--with-fpu=<var>type</var></code></dt> 874<dt><code>--with-float=<var>type</var></code></dt> 875<dd><p>These configure options provide default values for the <samp>-mschedule=</samp>, 876<samp>-march=</samp>, <samp>-mtune=</samp>, <samp>-mabi=</samp>, and <samp>-mfpu=</samp> 877options and for <samp>-mhard-float</samp> or <samp>-msoft-float</samp>. As with 878<samp>--with-cpu</samp>, which switches will be accepted and acceptable values 879of the arguments depend on the target. 880</p> 881</dd> 882<dt><code>--with-mode=<var>mode</var></code></dt> 883<dd><p>Specify if the compiler should default to <samp>-marm</samp> or <samp>-mthumb</samp>. 884This option is only supported on ARM targets. 885</p> 886</dd> 887<dt><code>--with-stack-offset=<var>num</var></code></dt> 888<dd><p>This option sets the default for the -mstack-offset=<var>num</var> option, 889and will thus generally also control the setting of this option for 890libraries. This option is only supported on Epiphany targets. 891</p> 892</dd> 893<dt><code>--with-fpmath=<var>isa</var></code></dt> 894<dd><p>This options sets <samp>-mfpmath=sse</samp> by default and specifies the default 895ISA for floating-point arithmetics. You can select either ‘<samp>sse</samp>’ which 896enables <samp>-msse2</samp> or ‘<samp>avx</samp>’ which enables <samp>-mavx</samp> by default. 897This option is only supported on i386 and x86-64 targets. 898</p> 899</dd> 900<dt><code>--with-fp-32=<var>mode</var></code></dt> 901<dd><p>On MIPS targets, set the default value for the <samp>-mfp</samp> option when using 902the o32 ABI. The possibilities for <var>mode</var> are: 903</p><dl compact="compact"> 904<dt><code>32</code></dt> 905<dd><p>Use the o32 FP32 ABI extension, as with the <samp>-mfp32</samp> command-line 906option. 907</p></dd> 908<dt><code>xx</code></dt> 909<dd><p>Use the o32 FPXX ABI extension, as with the <samp>-mfpxx</samp> command-line 910option. 911</p></dd> 912<dt><code>64</code></dt> 913<dd><p>Use the o32 FP64 ABI extension, as with the <samp>-mfp64</samp> command-line 914option. 915</p></dd> 916</dl> 917<p>In the absence of this configuration option the default is to use the o32 918FP32 ABI extension. 919</p> 920</dd> 921<dt><code>--with-odd-spreg-32</code></dt> 922<dd><p>On MIPS targets, set the <samp>-modd-spreg</samp> option by default when using 923the o32 ABI. 924</p> 925</dd> 926<dt><code>--without-odd-spreg-32</code></dt> 927<dd><p>On MIPS targets, set the <samp>-mno-odd-spreg</samp> option by default when using 928the o32 ABI. This is normally used in conjunction with 929<samp>--with-fp-32=64</samp> in order to target the o32 FP64A ABI extension. 930</p> 931</dd> 932<dt><code>--with-nan=<var>encoding</var></code></dt> 933<dd><p>On MIPS targets, set the default encoding convention to use for the 934special not-a-number (NaN) IEEE 754 floating-point data. The 935possibilities for <var>encoding</var> are: 936</p><dl compact="compact"> 937<dt><code>legacy</code></dt> 938<dd><p>Use the legacy encoding, as with the <samp>-mnan=legacy</samp> command-line 939option. 940</p></dd> 941<dt><code>2008</code></dt> 942<dd><p>Use the 754-2008 encoding, as with the <samp>-mnan=2008</samp> command-line 943option. 944</p></dd> 945</dl> 946<p>To use this configuration option you must have an assembler version 947installed that supports the <samp>-mnan=</samp> command-line option too. 948In the absence of this configuration option the default convention is 949the legacy encoding, as when neither of the <samp>-mnan=2008</samp> and 950<samp>-mnan=legacy</samp> command-line options has been used. 951</p> 952</dd> 953<dt><code>--with-divide=<var>type</var></code></dt> 954<dd><p>Specify how the compiler should generate code for checking for 955division by zero. This option is only supported on the MIPS target. 956The possibilities for <var>type</var> are: 957</p><dl compact="compact"> 958<dt><code>traps</code></dt> 959<dd><p>Division by zero checks use conditional traps (this is the default on 960systems that support conditional traps). 961</p></dd> 962<dt><code>breaks</code></dt> 963<dd><p>Division by zero checks use the break instruction. 964</p></dd> 965</dl> 966 967 968</dd> 969<dt><code>--with-llsc</code></dt> 970<dd><p>On MIPS targets, make <samp>-mllsc</samp> the default when no 971<samp>-mno-llsc</samp> option is passed. This is the default for 972Linux-based targets, as the kernel will emulate them if the ISA does 973not provide them. 974</p> 975</dd> 976<dt><code>--without-llsc</code></dt> 977<dd><p>On MIPS targets, make <samp>-mno-llsc</samp> the default when no 978<samp>-mllsc</samp> option is passed. 979</p> 980</dd> 981<dt><code>--with-synci</code></dt> 982<dd><p>On MIPS targets, make <samp>-msynci</samp> the default when no 983<samp>-mno-synci</samp> option is passed. 984</p> 985</dd> 986<dt><code>--without-synci</code></dt> 987<dd><p>On MIPS targets, make <samp>-mno-synci</samp> the default when no 988<samp>-msynci</samp> option is passed. This is the default. 989</p> 990</dd> 991<dt><code>--with-lxc1-sxc1</code></dt> 992<dd><p>On MIPS targets, make <samp>-mlxc1-sxc1</samp> the default when no 993<samp>-mno-lxc1-sxc1</samp> option is passed. This is the default. 994</p> 995</dd> 996<dt><code>--without-lxc1-sxc1</code></dt> 997<dd><p>On MIPS targets, make <samp>-mno-lxc1-sxc1</samp> the default when no 998<samp>-mlxc1-sxc1</samp> option is passed. The indexed load/store 999instructions are not directly a problem but can lead to unexpected 1000behaviour when deployed in an application intended for a 32-bit address 1001space but run on a 64-bit processor. The issue is seen because all 1002known MIPS 64-bit Linux kernels execute o32 and n32 applications 1003with 64-bit addressing enabled which affects the overflow behaviour 1004of the indexed addressing mode. GCC will assume that ordinary 100532-bit arithmetic overflow behaviour is the same whether performed 1006as an <code>addu</code> instruction or as part of the address calculation 1007in <code>lwxc1</code> type instructions. This assumption holds true in a 1008pure 32-bit environment and can hold true in a 64-bit environment if 1009the address space is accurately set to be 32-bit for o32 and n32. 1010</p> 1011</dd> 1012<dt><code>--with-madd4</code></dt> 1013<dd><p>On MIPS targets, make <samp>-mmadd4</samp> the default when no 1014<samp>-mno-madd4</samp> option is passed. This is the default. 1015</p> 1016</dd> 1017<dt><code>--without-madd4</code></dt> 1018<dd><p>On MIPS targets, make <samp>-mno-madd4</samp> the default when no 1019<samp>-mmadd4</samp> option is passed. The <code>madd4</code> instruction 1020family can be problematic when targeting a combination of cores that 1021implement these instructions differently. There are two known cores 1022that implement these as fused operations instead of unfused (where 1023unfused is normally expected). Disabling these instructions is the 1024only way to ensure compatible code is generated; this will incur 1025a performance penalty. 1026</p> 1027</dd> 1028<dt><code>--with-mips-plt</code></dt> 1029<dd><p>On MIPS targets, make use of copy relocations and PLTs. 1030These features are extensions to the traditional 1031SVR4-based MIPS ABIs and require support from GNU binutils 1032and the runtime C library. 1033</p> 1034</dd> 1035<dt><code>--with-stack-clash-protection-guard-size=<var>size</var></code></dt> 1036<dd><p>On certain targets this option sets the default stack clash protection guard 1037size as a power of two in bytes. On AArch64 <var>size</var> is required to be either 103812 (4KB) or 16 (64KB). 1039</p> 1040</dd> 1041<dt><code>--enable-__cxa_atexit</code></dt> 1042<dd><p>Define if you want to use __cxa_atexit, rather than atexit, to 1043register C++ destructors for local statics and global objects. 1044This is essential for fully standards-compliant handling of 1045destructors, but requires __cxa_atexit in libc. This option is currently 1046only available on systems with GNU libc. When enabled, this will cause 1047<samp>-fuse-cxa-atexit</samp> to be passed by default. 1048</p> 1049</dd> 1050<dt><code>--enable-gnu-indirect-function</code></dt> 1051<dd><p>Define if you want to enable the <code>ifunc</code> attribute. This option is 1052currently only available on systems with GNU libc on certain targets. 1053</p> 1054</dd> 1055<dt><code>--enable-target-optspace</code></dt> 1056<dd><p>Specify that target 1057libraries should be optimized for code space instead of code speed. 1058This is the default for the m32r platform. 1059</p> 1060</dd> 1061<dt><code>--with-cpp-install-dir=<var>dirname</var></code></dt> 1062<dd><p>Specify that the user visible <code>cpp</code> program should be installed 1063in <samp><var>prefix</var>/<var>dirname</var>/cpp</samp>, in addition to <var>bindir</var>. 1064</p> 1065</dd> 1066<dt><code>--enable-comdat</code></dt> 1067<dd><p>Enable COMDAT group support. This is primarily used to override the 1068automatically detected value. 1069</p> 1070</dd> 1071<dt><code>--enable-initfini-array</code></dt> 1072<dd><p>Force the use of sections <code>.init_array</code> and <code>.fini_array</code> 1073(instead of <code>.init</code> and <code>.fini</code>) for constructors and 1074destructors. Option <samp>--disable-initfini-array</samp> has the 1075opposite effect. If neither option is specified, the configure script 1076will try to guess whether the <code>.init_array</code> and 1077<code>.fini_array</code> sections are supported and, if they are, use them. 1078</p> 1079</dd> 1080<dt><code>--enable-link-mutex</code></dt> 1081<dd><p>When building GCC, use a mutex to avoid linking the compilers for 1082multiple languages at the same time, to avoid thrashing on build 1083systems with limited free memory. The default is not to use such a mutex. 1084</p> 1085</dd> 1086<dt><code>--enable-maintainer-mode</code></dt> 1087<dd><p>The build rules that regenerate the Autoconf and Automake output files as 1088well as the GCC master message catalog <samp>gcc.pot</samp> are normally 1089disabled. This is because it can only be rebuilt if the complete source 1090tree is present. If you have changed the sources and want to rebuild the 1091catalog, configuring with <samp>--enable-maintainer-mode</samp> will enable 1092this. Note that you need a recent version of the <code>gettext</code> tools 1093to do so. 1094</p> 1095</dd> 1096<dt><code>--disable-bootstrap</code></dt> 1097<dd><p>For a native build, the default configuration is to perform 1098a 3-stage bootstrap of the compiler when ‘<samp>make</samp>’ is invoked, 1099testing that GCC can compile itself correctly. If you want to disable 1100this process, you can configure with <samp>--disable-bootstrap</samp>. 1101</p> 1102</dd> 1103<dt><code>--enable-bootstrap</code></dt> 1104<dd><p>In special cases, you may want to perform a 3-stage build 1105even if the target and host triplets are different. 1106This is possible when the host can run code compiled for 1107the target (e.g. host is i686-linux, target is i486-linux). 1108Starting from GCC 4.2, to do this you have to configure explicitly 1109with <samp>--enable-bootstrap</samp>. 1110</p> 1111</dd> 1112<dt><code>--enable-generated-files-in-srcdir</code></dt> 1113<dd><p>Neither the .c and .h files that are generated from Bison and flex nor the 1114info manuals and man pages that are built from the .texi files are present 1115in the repository development tree. When building GCC from that development tree, 1116or from one of our snapshots, those generated files are placed in your 1117build directory, which allows for the source to be in a readonly 1118directory. 1119</p> 1120<p>If you configure with <samp>--enable-generated-files-in-srcdir</samp> then those 1121generated files will go into the source directory. This is mainly intended 1122for generating release or prerelease tarballs of the GCC sources, since it 1123is not a requirement that the users of source releases to have flex, Bison, 1124or makeinfo. 1125</p> 1126</dd> 1127<dt><code>--enable-version-specific-runtime-libs</code></dt> 1128<dd><p>Specify 1129that runtime libraries should be installed in the compiler specific 1130subdirectory (<samp><var>libdir</var>/gcc</samp>) rather than the usual places. In 1131addition, ‘<samp>libstdc++</samp>’’s include files will be installed into 1132<samp><var>libdir</var></samp> unless you overruled it by using 1133<samp>--with-gxx-include-dir=<var>dirname</var></samp>. Using this option is 1134particularly useful if you intend to use several versions of GCC in 1135parallel. The default is ‘<samp>yes</samp>’ for ‘<samp>libada</samp>’, and ‘<samp>no</samp>’ for 1136the remaining libraries. 1137</p> 1138</dd> 1139<dt><code><a name="WithAixSoname"></a>--with-aix-soname=‘<samp>aix</samp>’, ‘<samp>svr4</samp>’ or ‘<samp>both</samp>’</code></dt> 1140<dd><p>Traditional AIX shared library versioning (versioned <code>Shared Object</code> 1141files as members of unversioned <code>Archive Library</code> files named 1142‘<samp>lib.a</samp>’) causes numerous headaches for package managers. However, 1143<code>Import Files</code> as members of <code>Archive Library</code> files allow for 1144<strong>filename-based versioning</strong> of shared libraries as seen on Linux/SVR4, 1145where this is called the "SONAME". But as they prevent static linking, 1146<code>Import Files</code> may be used with <code>Runtime Linking</code> only, where the 1147linker does search for ‘<samp>libNAME.so</samp>’ before ‘<samp>libNAME.a</samp>’ library 1148filenames with the ‘<samp>-lNAME</samp>’ linker flag. 1149</p> 1150<a name="AixLdCommand"></a><p>For detailed information please refer to the AIX 1151<a href="https://www.ibm.com/support/knowledgecenter/search/%22the%20ld%20command%2C%20also%20called%20the%20linkage%20editor%20or%20binder%22">ld 1152Command</a> reference. 1153</p> 1154<p>As long as shared library creation is enabled, upon: 1155</p><dl compact="compact"> 1156<dt><code>--with-aix-soname=aix</code></dt> 1157<dt><code>--with-aix-soname=both</code></dt> 1158<dd><p>A (traditional AIX) <code>Shared Archive Library</code> file is created: 1159 </p><ul> 1160<li> using the ‘<samp>libNAME.a</samp>’ filename scheme 1161 </li><li> with the <code>Shared Object</code> file as archive member named 1162 ‘<samp>libNAME.so.V</samp>’ (except for ‘<samp>libgcc_s</samp>’, where the <code>Shared 1163 Object</code> file is named ‘<samp>shr.o</samp>’ for backwards compatibility), which 1164 <ul class="no-bullet"> 1165<li>- is used for runtime loading from inside the ‘<samp>libNAME.a</samp>’ file 1166 </li><li>- is used for dynamic loading via 1167 <code>dlopen("libNAME.a(libNAME.so.V)", RTLD_MEMBER)</code> 1168 </li><li>- is used for shared linking 1169 </li><li>- is used for static linking, so no separate <code>Static Archive 1170 Library</code> file is needed 1171 </li></ul> 1172</li></ul> 1173</dd> 1174<dt><code>--with-aix-soname=both</code></dt> 1175<dt><code>--with-aix-soname=svr4</code></dt> 1176<dd><p>A (second) <code>Shared Archive Library</code> file is created: 1177 </p><ul> 1178<li> using the ‘<samp>libNAME.so.V</samp>’ filename scheme 1179 </li><li> with the <code>Shared Object</code> file as archive member named 1180 ‘<samp>shr.o</samp>’, which 1181 <ul class="no-bullet"> 1182<li>- is created with the <code>-G linker flag</code> 1183 </li><li>- has the <code>F_LOADONLY</code> flag set 1184 </li><li>- is used for runtime loading from inside the ‘<samp>libNAME.so.V</samp>’ file 1185 </li><li>- is used for dynamic loading via <code>dlopen("libNAME.so.V(shr.o)", 1186 RTLD_MEMBER)</code> 1187 </li></ul> 1188</li><li> with the <code>Import File</code> as archive member named ‘<samp>shr.imp</samp>’, 1189 which 1190 <ul class="no-bullet"> 1191<li>- refers to ‘<samp>libNAME.so.V(shr.o)</samp>’ as the "SONAME", to be recorded 1192 in the <code>Loader Section</code> of subsequent binaries 1193 </li><li>- indicates whether ‘<samp>libNAME.so.V(shr.o)</samp>’ is 32 or 64 bit 1194 </li><li>- lists all the public symbols exported by ‘<samp>lib.so.V(shr.o)</samp>’, 1195 eventually decorated with the <code>‘<samp>weak</samp>’ Keyword</code> 1196 </li><li>- is necessary for shared linking against ‘<samp>lib.so.V(shr.o)</samp>’ 1197 </li></ul> 1198</li></ul> 1199<p>A symbolic link using the ‘<samp>libNAME.so</samp>’ filename scheme is created: 1200 </p><ul> 1201<li> pointing to the ‘<samp>libNAME.so.V</samp>’ <code>Shared Archive Library</code> file 1202 </li><li> to permit the <code>ld Command</code> to find ‘<samp>lib.so.V(shr.imp)</samp>’ via 1203 the ‘<samp>-lNAME</samp>’ argument (requires <code>Runtime Linking</code> to be enabled) 1204 </li><li> to permit dynamic loading of ‘<samp>lib.so.V(shr.o)</samp>’ without the need 1205 to specify the version number via <code>dlopen("libNAME.so(shr.o)", 1206 RTLD_MEMBER)</code> 1207 </li></ul> 1208</dd> 1209</dl> 1210 1211<p>As long as static library creation is enabled, upon: 1212</p><dl compact="compact"> 1213<dt><code>--with-aix-soname=svr4</code></dt> 1214<dd><p>A <code>Static Archive Library</code> is created: 1215 </p><ul> 1216<li> using the ‘<samp>libNAME.a</samp>’ filename scheme 1217 </li><li> with all the <code>Static Object</code> files as archive members, which 1218 <ul class="no-bullet"> 1219<li>- are used for static linking 1220 </li></ul> 1221</li></ul> 1222</dd> 1223</dl> 1224 1225<p>While the aix-soname=‘<samp>svr4</samp>’ option does not create <code>Shared Object</code> 1226files as members of unversioned <code>Archive Library</code> files any more, package 1227managers still are responsible to 1228<a href="./specific.html#TransferAixShobj">transfer</a> <code>Shared Object</code> files 1229found as member of a previously installed unversioned <code>Archive Library</code> 1230file into the newly installed <code>Archive Library</code> file with the same 1231filename. 1232</p> 1233<p><em>WARNING:</em> Creating <code>Shared Object</code> files with <code>Runtime Linking</code> 1234enabled may bloat the TOC, eventually leading to <code>TOC overflow</code> errors, 1235requiring the use of either the <samp>-Wl,-bbigtoc</samp> linker flag (seen to 1236break with the <code>GDB</code> debugger) or some of the TOC-related compiler flags, 1237see “RS/6000 and PowerPC Options” in the main manual. 1238</p> 1239<p><samp>--with-aix-soname</samp> is currently supported by ‘<samp>libgcc_s</samp>’ only, so 1240this option is still experimental and not for normal use yet. 1241</p> 1242<p>Default is the traditional behavior <samp>--with-aix-soname=‘<samp>aix</samp>’</samp>. 1243</p> 1244</dd> 1245<dt><code>--enable-languages=<var>lang1</var>,<var>lang2</var>,…</code></dt> 1246<dd><p>Specify that only a particular subset of compilers and 1247their runtime libraries should be built. For a list of valid values for 1248<var>langN</var> you can issue the following command in the 1249<samp>gcc</samp> directory of your GCC source tree:<br> 1250</p><div class="smallexample"> 1251<pre class="smallexample">grep ^language= */config-lang.in 1252</pre></div> 1253<p>Currently, you can use any of the following: 1254<code>all</code>, <code>default</code>, <code>ada</code>, <code>c</code>, <code>c++</code>, <code>d</code>, 1255<code>fortran</code>, <code>go</code>, <code>jit</code>, <code>lto</code>, <code>objc</code>, <code>obj-c++</code>. 1256Building the Ada compiler has special requirements, see below. 1257If you do not pass this flag, or specify the option <code>default</code>, then the 1258default languages available in the <samp>gcc</samp> sub-tree will be configured. 1259Ada, D, Go, Jit, and Objective-C++ are not default languages. LTO is not a 1260default language, but is built by default because <samp>--enable-lto</samp> is 1261enabled by default. The other languages are default languages. If 1262<code>all</code> is specified, then all available languages are built. An 1263exception is <code>jit</code> language, which requires 1264<samp>--enable-host-shared</samp> to be included with <code>all</code>. 1265</p> 1266</dd> 1267<dt><code>--enable-stage1-languages=<var>lang1</var>,<var>lang2</var>,…</code></dt> 1268<dd><p>Specify that a particular subset of compilers and their runtime 1269libraries should be built with the system C compiler during stage 1 of 1270the bootstrap process, rather than only in later stages with the 1271bootstrapped C compiler. The list of valid values is the same as for 1272<samp>--enable-languages</samp>, and the option <code>all</code> will select all 1273of the languages enabled by <samp>--enable-languages</samp>. This option is 1274primarily useful for GCC development; for instance, when a development 1275version of the compiler cannot bootstrap due to compiler bugs, or when 1276one is debugging front ends other than the C front end. When this 1277option is used, one can then build the target libraries for the 1278specified languages with the stage-1 compiler by using <code>make 1279stage1-bubble all-target</code>, or run the testsuite on the stage-1 compiler 1280for the specified languages using <code>make stage1-start check-gcc</code>. 1281</p> 1282</dd> 1283<dt><code>--disable-libada</code></dt> 1284<dd><p>Specify that the run-time libraries and tools used by GNAT should not 1285be built. This can be useful for debugging, or for compatibility with 1286previous Ada build procedures, when it was required to explicitly 1287do a ‘<samp>make -C gcc gnatlib_and_tools</samp>’. 1288</p> 1289</dd> 1290<dt><code>--disable-libsanitizer</code></dt> 1291<dd><p>Specify that the run-time libraries for the various sanitizers should 1292not be built. 1293</p> 1294</dd> 1295<dt><code>--disable-libssp</code></dt> 1296<dd><p>Specify that the run-time libraries for stack smashing protection 1297should not be built or linked against. On many targets library support 1298is provided by the C library instead. 1299</p> 1300</dd> 1301<dt><code>--disable-libquadmath</code></dt> 1302<dd><p>Specify that the GCC quad-precision math library should not be built. 1303On some systems, the library is required to be linkable when building 1304the Fortran front end, unless <samp>--disable-libquadmath-support</samp> 1305is used. 1306</p> 1307</dd> 1308<dt><code>--disable-libquadmath-support</code></dt> 1309<dd><p>Specify that the Fortran front end and <code>libgfortran</code> do not add 1310support for <code>libquadmath</code> on systems supporting it. 1311</p> 1312</dd> 1313<dt><code>--disable-libgomp</code></dt> 1314<dd><p>Specify that the GNU Offloading and Multi Processing Runtime Library 1315should not be built. 1316</p> 1317</dd> 1318<dt><code>--disable-libvtv</code></dt> 1319<dd><p>Specify that the run-time libraries used by vtable verification 1320should not be built. 1321</p> 1322</dd> 1323<dt><code>--with-dwarf2</code></dt> 1324<dd><p>Specify that the compiler should 1325use DWARF 2 debugging information as the default. 1326</p> 1327</dd> 1328<dt><code>--with-advance-toolchain=<var>at</var></code></dt> 1329<dd><p>On 64-bit PowerPC Linux systems, configure the compiler to use the 1330header files, library files, and the dynamic linker from the Advance 1331Toolchain release <var>at</var> instead of the default versions that are 1332provided by the Linux distribution. In general, this option is 1333intended for the developers of GCC, and it is not intended for general 1334use. 1335</p> 1336</dd> 1337<dt><code>--enable-targets=all</code></dt> 1338<dt><code>--enable-targets=<var>target_list</var></code></dt> 1339<dd><p>Some GCC targets, e.g. powerpc64-linux, build bi-arch compilers. 1340These are compilers that are able to generate either 64-bit or 32-bit 1341code. Typically, the corresponding 32-bit target, e.g. 1342powerpc-linux for powerpc64-linux, only generates 32-bit code. This 1343option enables the 32-bit target to be a bi-arch compiler, which is 1344useful when you want a bi-arch compiler that defaults to 32-bit, and 1345you are building a bi-arch or multi-arch binutils in a combined tree. 1346On mips-linux, this will build a tri-arch compiler (ABI o32/n32/64), 1347defaulted to o32. 1348Currently, this option only affects sparc-linux, powerpc-linux, x86-linux, 1349mips-linux and s390-linux. 1350</p> 1351</dd> 1352<dt><code>--enable-default-pie</code></dt> 1353<dd><p>Turn on <samp>-fPIE</samp> and <samp>-pie</samp> by default. 1354</p> 1355</dd> 1356<dt><code>--enable-secureplt</code></dt> 1357<dd><p>This option enables <samp>-msecure-plt</samp> by default for powerpc-linux. 1358See “RS/6000 and PowerPC Options” in the main manual 1359</p> 1360</dd> 1361<dt><code>--enable-default-ssp</code></dt> 1362<dd><p>Turn on <samp>-fstack-protector-strong</samp> by default. 1363</p> 1364</dd> 1365<dt><code>--enable-cld</code></dt> 1366<dd><p>This option enables <samp>-mcld</samp> by default for 32-bit x86 targets. 1367See “i386 and x86-64 Options” in the main manual 1368</p> 1369</dd> 1370<dt><code>--enable-large-address-aware</code></dt> 1371<dd><p>The <samp>--enable-large-address-aware</samp> option arranges for MinGW 1372executables to be linked using the <samp>--large-address-aware</samp> 1373option, that enables the use of more than 2GB of memory. If GCC is 1374configured with this option, its effects can be reversed by passing the 1375<samp>-Wl,--disable-large-address-aware</samp> option to the so-configured 1376compiler driver. 1377</p> 1378</dd> 1379<dt><code>--enable-win32-registry</code></dt> 1380<dt><code>--enable-win32-registry=<var>key</var></code></dt> 1381<dt><code>--disable-win32-registry</code></dt> 1382<dd><p>The <samp>--enable-win32-registry</samp> option enables Microsoft Windows-hosted GCC 1383to look up installations paths in the registry using the following key: 1384</p> 1385<div class="smallexample"> 1386<pre class="smallexample"><code>HKEY_LOCAL_MACHINE\SOFTWARE\Free Software Foundation\<var>key</var></code> 1387</pre></div> 1388 1389<p><var>key</var> defaults to GCC version number, and can be overridden by the 1390<samp>--enable-win32-registry=<var>key</var></samp> option. Vendors and distributors 1391who use custom installers are encouraged to provide a different key, 1392perhaps one comprised of vendor name and GCC version number, to 1393avoid conflict with existing installations. This feature is enabled 1394by default, and can be disabled by <samp>--disable-win32-registry</samp> 1395option. This option has no effect on the other hosts. 1396</p> 1397</dd> 1398<dt><code>--nfp</code></dt> 1399<dd><p>Specify that the machine does not have a floating point unit. This 1400option only applies to ‘<samp>m68k-sun-sunos<var>n</var></samp>’. On any other 1401system, <samp>--nfp</samp> has no effect. 1402</p> 1403</dd> 1404<dt><code>--enable-werror</code></dt> 1405<dt><code>--disable-werror</code></dt> 1406<dt><code>--enable-werror=yes</code></dt> 1407<dt><code>--enable-werror=no</code></dt> 1408<dd><p>When you specify this option, it controls whether certain files in the 1409compiler are built with <samp>-Werror</samp> in bootstrap stage2 and later. 1410If you don’t specify it, <samp>-Werror</samp> is turned on for the main 1411development trunk. However it defaults to off for release branches and 1412final releases. The specific files which get <samp>-Werror</samp> are 1413controlled by the Makefiles. 1414</p> 1415</dd> 1416<dt><code>--enable-checking</code></dt> 1417<dt><code>--disable-checking</code></dt> 1418<dt><code>--enable-checking=<var>list</var></code></dt> 1419<dd><p>This option controls performing internal consistency checks in the compiler. 1420It does not change the generated code, but adds error checking of the 1421requested complexity. This slows down the compiler and may only work 1422properly if you are building the compiler with GCC. 1423</p> 1424<p>When the option is not specified, the active set of checks depends on context. 1425Namely, bootstrap stage 1 defaults to ‘<samp>--enable-checking=yes</samp>’, builds 1426from release branches or release archives default to 1427‘<samp>--enable-checking=release</samp>’, and otherwise 1428‘<samp>--enable-checking=yes,extra</samp>’ is used. When the option is 1429specified without a <var>list</var>, the result is the same as 1430‘<samp>--enable-checking=yes</samp>’. Likewise, ‘<samp>--disable-checking</samp>’ is 1431equivalent to ‘<samp>--enable-checking=no</samp>’. 1432</p> 1433<p>The categories of checks available in <var>list</var> are ‘<samp>yes</samp>’ (most common 1434checks ‘<samp>assert,misc,gc,gimple,rtlflag,runtime,tree,types</samp>’), ‘<samp>no</samp>’ 1435(no checks at all), ‘<samp>all</samp>’ (all but ‘<samp>valgrind</samp>’), ‘<samp>release</samp>’ 1436(cheapest checks ‘<samp>assert,runtime</samp>’) or ‘<samp>none</samp>’ (same as ‘<samp>no</samp>’). 1437‘<samp>release</samp>’ checks are always on and to disable them 1438‘<samp>--disable-checking</samp>’ or ‘<samp>--enable-checking=no[,<other checks>]</samp>’ 1439must be explicitly requested. Disabling assertions makes the compiler and 1440runtime slightly faster but increases the risk of undetected internal errors 1441causing wrong code to be generated. 1442</p> 1443<p>Individual checks can be enabled with these flags: ‘<samp>assert</samp>’, ‘<samp>df</samp>’, 1444‘<samp>extra</samp>’, ‘<samp>fold</samp>’, ‘<samp>gc</samp>’, ‘<samp>gcac</samp>’, ‘<samp>gimple</samp>’, 1445‘<samp>misc</samp>’, ‘<samp>rtl</samp>’, ‘<samp>rtlflag</samp>’, ‘<samp>runtime</samp>’, ‘<samp>tree</samp>’, 1446‘<samp>types</samp>’ and ‘<samp>valgrind</samp>’. ‘<samp>extra</samp>’ extends ‘<samp>misc</samp>’ 1447checking with extra checks that might affect code generation and should 1448therefore not differ between stage1 and later stages in bootstrap. 1449</p> 1450<p>The ‘<samp>valgrind</samp>’ check requires the external <code>valgrind</code> simulator, 1451available from <a href="http://valgrind.org/">http://valgrind.org/</a>. The ‘<samp>rtl</samp>’ checks are 1452expensive and the ‘<samp>df</samp>’, ‘<samp>gcac</samp>’ and ‘<samp>valgrind</samp>’ checks are very 1453expensive. 1454</p> 1455</dd> 1456<dt><code>--disable-stage1-checking</code></dt> 1457<dt><code>--enable-stage1-checking</code></dt> 1458<dt><code>--enable-stage1-checking=<var>list</var></code></dt> 1459<dd><p>This option affects only bootstrap build. If no <samp>--enable-checking</samp> 1460option is specified the stage1 compiler is built with ‘<samp>yes</samp>’ checking 1461enabled, otherwise the stage1 checking flags are the same as specified by 1462<samp>--enable-checking</samp>. To build the stage1 compiler with 1463different checking options use <samp>--enable-stage1-checking</samp>. 1464The list of checking options is the same as for <samp>--enable-checking</samp>. 1465If your system is too slow or too small to bootstrap a released compiler 1466with checking for stage1 enabled, you can use ‘<samp>--disable-stage1-checking</samp>’ 1467to disable checking for the stage1 compiler. 1468</p> 1469</dd> 1470<dt><code>--enable-coverage</code></dt> 1471<dt><code>--enable-coverage=<var>level</var></code></dt> 1472<dd><p>With this option, the compiler is built to collect self coverage 1473information, every time it is run. This is for internal development 1474purposes, and only works when the compiler is being built with gcc. The 1475<var>level</var> argument controls whether the compiler is built optimized or 1476not, values are ‘<samp>opt</samp>’ and ‘<samp>noopt</samp>’. For coverage analysis you 1477want to disable optimization, for performance analysis you want to 1478enable optimization. When coverage is enabled, the default level is 1479without optimization. 1480</p> 1481</dd> 1482<dt><code>--enable-gather-detailed-mem-stats</code></dt> 1483<dd><p>When this option is specified more detailed information on memory 1484allocation is gathered. This information is printed when using 1485<samp>-fmem-report</samp>. 1486</p> 1487</dd> 1488<dt><code>--enable-valgrind-annotations</code></dt> 1489<dd><p>Mark selected memory related operations in the compiler when run under 1490valgrind to suppress false positives. 1491</p> 1492</dd> 1493<dt><code>--enable-nls</code></dt> 1494<dt><code>--disable-nls</code></dt> 1495<dd><p>The <samp>--enable-nls</samp> option enables Native Language Support (NLS), 1496which lets GCC output diagnostics in languages other than American 1497English. Native Language Support is enabled by default if not doing a 1498canadian cross build. The <samp>--disable-nls</samp> option disables NLS. 1499</p> 1500</dd> 1501<dt><code>--with-included-gettext</code></dt> 1502<dd><p>If NLS is enabled, the <samp>--with-included-gettext</samp> option causes the build 1503procedure to prefer its copy of GNU <code>gettext</code>. 1504</p> 1505</dd> 1506<dt><code>--with-catgets</code></dt> 1507<dd><p>If NLS is enabled, and if the host lacks <code>gettext</code> but has the 1508inferior <code>catgets</code> interface, the GCC build procedure normally 1509ignores <code>catgets</code> and instead uses GCC’s copy of the GNU 1510<code>gettext</code> library. The <samp>--with-catgets</samp> option causes the 1511build procedure to use the host’s <code>catgets</code> in this situation. 1512</p> 1513</dd> 1514<dt><code>--with-libiconv-prefix=<var>dir</var></code></dt> 1515<dd><p>Search for libiconv header files in <samp><var>dir</var>/include</samp> and 1516libiconv library files in <samp><var>dir</var>/lib</samp>. 1517</p> 1518</dd> 1519<dt><code>--enable-obsolete</code></dt> 1520<dd><p>Enable configuration for an obsoleted system. If you attempt to 1521configure GCC for a system (build, host, or target) which has been 1522obsoleted, and you do not specify this flag, configure will halt with an 1523error message. 1524</p> 1525<p>All support for systems which have been obsoleted in one release of GCC 1526is removed entirely in the next major release, unless someone steps 1527forward to maintain the port. 1528</p> 1529</dd> 1530<dt><code>--enable-decimal-float</code></dt> 1531<dt><code>--enable-decimal-float=yes</code></dt> 1532<dt><code>--enable-decimal-float=no</code></dt> 1533<dt><code>--enable-decimal-float=bid</code></dt> 1534<dt><code>--enable-decimal-float=dpd</code></dt> 1535<dt><code>--disable-decimal-float</code></dt> 1536<dd><p>Enable (or disable) support for the C decimal floating point extension 1537that is in the IEEE 754-2008 standard. This is enabled by default only 1538on PowerPC, i386, and x86_64 GNU/Linux systems. Other systems may also 1539support it, but require the user to specifically enable it. You can 1540optionally control which decimal floating point format is used (either 1541‘<samp>bid</samp>’ or ‘<samp>dpd</samp>’). The ‘<samp>bid</samp>’ (binary integer decimal) 1542format is default on i386 and x86_64 systems, and the ‘<samp>dpd</samp>’ 1543(densely packed decimal) format is default on PowerPC systems. 1544</p> 1545</dd> 1546<dt><code>--enable-fixed-point</code></dt> 1547<dt><code>--disable-fixed-point</code></dt> 1548<dd><p>Enable (or disable) support for C fixed-point arithmetic. 1549This option is enabled by default for some targets (such as MIPS) which 1550have hardware-support for fixed-point operations. On other targets, you 1551may enable this option manually. 1552</p> 1553</dd> 1554<dt><code>--with-long-double-128</code></dt> 1555<dd><p>Specify if <code>long double</code> type should be 128-bit by default on selected 1556GNU/Linux architectures. If using <code>--without-long-double-128</code>, 1557<code>long double</code> will be by default 64-bit, the same as <code>double</code> type. 1558When neither of these configure options are used, the default will be 1559128-bit <code>long double</code> when built against GNU C Library 2.4 and later, 156064-bit <code>long double</code> otherwise. 1561</p> 1562</dd> 1563<dt><code>--with-long-double-format=ibm</code></dt> 1564<dt><code>--with-long-double-format=ieee</code></dt> 1565<dd><p>Specify whether <code>long double</code> uses the IBM extended double format 1566or the IEEE 128-bit floating point format on PowerPC Linux systems. 1567This configuration switch will only work on little endian PowerPC 1568Linux systems and on big endian 64-bit systems where the default cpu 1569is at least power7 (i.e. <samp>--with-cpu=power7</samp>, 1570<samp>--with-cpu=power8</samp>, or <samp>--with-cpu=power9</samp> is used). 1571</p> 1572<p>If you use the <samp>--with-long-double-64</samp> configuration option, 1573the <samp>--with-long-double-format=ibm</samp> and 1574<samp>--with-long-double-format=ieee</samp> options are ignored. 1575</p> 1576<p>The default <code>long double</code> format is to use IBM extended double. 1577Until all of the libraries are converted to use IEEE 128-bit floating 1578point, it is not recommended to use 1579<samp>--with-long-double-format=ieee</samp>. 1580</p> 1581<p>On little endian PowerPC Linux systems, if you explicitly set the 1582<code>long double</code> type, it will build multilibs to allow you to 1583select either <code>long double</code> format, unless you disable multilibs 1584with the <code>--disable-multilib</code> option. At present, 1585<code>long double</code> multilibs are not built on big endian PowerPC Linux 1586systems. If you are building multilibs, you will need to configure 1587the compiler using the <samp>--with-system-zlib</samp> option. 1588</p> 1589<p>If you do not set the <code>long double</code> type explicitly, no multilibs 1590will be generated. 1591</p> 1592</dd> 1593<dt><code>--enable-fdpic</code></dt> 1594<dd><p>On SH Linux systems, generate ELF FDPIC code. 1595</p> 1596</dd> 1597<dt><code>--with-gmp=<var>pathname</var></code></dt> 1598<dt><code>--with-gmp-include=<var>pathname</var></code></dt> 1599<dt><code>--with-gmp-lib=<var>pathname</var></code></dt> 1600<dt><code>--with-mpfr=<var>pathname</var></code></dt> 1601<dt><code>--with-mpfr-include=<var>pathname</var></code></dt> 1602<dt><code>--with-mpfr-lib=<var>pathname</var></code></dt> 1603<dt><code>--with-mpc=<var>pathname</var></code></dt> 1604<dt><code>--with-mpc-include=<var>pathname</var></code></dt> 1605<dt><code>--with-mpc-lib=<var>pathname</var></code></dt> 1606<dd><p>If you want to build GCC but do not have the GMP library, the MPFR 1607library and/or the MPC library installed in a standard location and 1608do not have their sources present in the GCC source tree then you 1609can explicitly specify the directory where they are installed 1610(‘<samp>--with-gmp=<var>gmpinstalldir</var></samp>’, 1611‘<samp>--with-mpfr=<var>mpfrinstalldir</var></samp>’, 1612‘<samp>--with-mpc=<var>mpcinstalldir</var></samp>’). The 1613<samp>--with-gmp=<var>gmpinstalldir</var></samp> option is shorthand for 1614<samp>--with-gmp-lib=<var>gmpinstalldir</var>/lib</samp> and 1615<samp>--with-gmp-include=<var>gmpinstalldir</var>/include</samp>. Likewise the 1616<samp>--with-mpfr=<var>mpfrinstalldir</var></samp> option is shorthand for 1617<samp>--with-mpfr-lib=<var>mpfrinstalldir</var>/lib</samp> and 1618<samp>--with-mpfr-include=<var>mpfrinstalldir</var>/include</samp>, also the 1619<samp>--with-mpc=<var>mpcinstalldir</var></samp> option is shorthand for 1620<samp>--with-mpc-lib=<var>mpcinstalldir</var>/lib</samp> and 1621<samp>--with-mpc-include=<var>mpcinstalldir</var>/include</samp>. If these 1622shorthand assumptions are not correct, you can use the explicit 1623include and lib options directly. You might also need to ensure the 1624shared libraries can be found by the dynamic linker when building and 1625using GCC, for example by setting the runtime shared library path 1626variable (<code>LD_LIBRARY_PATH</code> on GNU/Linux and Solaris systems). 1627</p> 1628<p>These flags are applicable to the host platform only. When building 1629a cross compiler, they will not be used to configure target libraries. 1630</p> 1631</dd> 1632<dt><code>--with-isl=<var>pathname</var></code></dt> 1633<dt><code>--with-isl-include=<var>pathname</var></code></dt> 1634<dt><code>--with-isl-lib=<var>pathname</var></code></dt> 1635<dd><p>If you do not have the isl library installed in a standard location and you 1636want to build GCC, you can explicitly specify the directory where it is 1637installed (‘<samp>--with-isl=<var>islinstalldir</var></samp>’). The 1638<samp>--with-isl=<var>islinstalldir</var></samp> option is shorthand for 1639<samp>--with-isl-lib=<var>islinstalldir</var>/lib</samp> and 1640<samp>--with-isl-include=<var>islinstalldir</var>/include</samp>. If this 1641shorthand assumption is not correct, you can use the explicit 1642include and lib options directly. 1643</p> 1644<p>These flags are applicable to the host platform only. When building 1645a cross compiler, they will not be used to configure target libraries. 1646</p> 1647</dd> 1648<dt><code>--with-stage1-ldflags=<var>flags</var></code></dt> 1649<dd><p>This option may be used to set linker flags to be used when linking 1650stage 1 of GCC. These are also used when linking GCC if configured with 1651<samp>--disable-bootstrap</samp>. If <samp>--with-stage1-libs</samp> is not set to a 1652value, then the default is ‘<samp>-static-libstdc++ -static-libgcc</samp>’, if 1653supported. 1654</p> 1655</dd> 1656<dt><code>--with-stage1-libs=<var>libs</var></code></dt> 1657<dd><p>This option may be used to set libraries to be used when linking stage 1 1658of GCC. These are also used when linking GCC if configured with 1659<samp>--disable-bootstrap</samp>. 1660</p> 1661</dd> 1662<dt><code>--with-boot-ldflags=<var>flags</var></code></dt> 1663<dd><p>This option may be used to set linker flags to be used when linking 1664stage 2 and later when bootstrapping GCC. If –with-boot-libs 1665is not is set to a value, then the default is 1666‘<samp>-static-libstdc++ -static-libgcc</samp>’. 1667</p> 1668</dd> 1669<dt><code>--with-boot-libs=<var>libs</var></code></dt> 1670<dd><p>This option may be used to set libraries to be used when linking stage 2 1671and later when bootstrapping GCC. 1672</p> 1673</dd> 1674<dt><code>--with-debug-prefix-map=<var>map</var></code></dt> 1675<dd><p>Convert source directory names using <samp>-fdebug-prefix-map</samp> when 1676building runtime libraries. ‘<samp><var>map</var></samp>’ is a space-separated 1677list of maps of the form ‘<samp><var>old</var>=<var>new</var></samp>’. 1678</p> 1679</dd> 1680<dt><code>--enable-linker-build-id</code></dt> 1681<dd><p>Tells GCC to pass <samp>--build-id</samp> option to the linker for all final 1682links (links performed without the <samp>-r</samp> or <samp>--relocatable</samp> 1683option), if the linker supports it. If you specify 1684<samp>--enable-linker-build-id</samp>, but your linker does not 1685support <samp>--build-id</samp> option, a warning is issued and the 1686<samp>--enable-linker-build-id</samp> option is ignored. The default is off. 1687</p> 1688</dd> 1689<dt><code>--with-linker-hash-style=<var>choice</var></code></dt> 1690<dd><p>Tells GCC to pass <samp>--hash-style=<var>choice</var></samp> option to the 1691linker for all final links. <var>choice</var> can be one of 1692‘<samp>sysv</samp>’, ‘<samp>gnu</samp>’, and ‘<samp>both</samp>’ where ‘<samp>sysv</samp>’ is the default. 1693</p> 1694</dd> 1695<dt><code>--enable-gnu-unique-object</code></dt> 1696<dt><code>--disable-gnu-unique-object</code></dt> 1697<dd><p>Tells GCC to use the gnu_unique_object relocation for C++ template 1698static data members and inline function local statics. Enabled by 1699default for a toolchain with an assembler that accepts it and 1700GLIBC 2.11 or above, otherwise disabled. 1701</p> 1702</dd> 1703<dt><code>--with-diagnostics-color=<var>choice</var></code></dt> 1704<dd><p>Tells GCC to use <var>choice</var> as the default for <samp>-fdiagnostics-color=</samp> 1705option (if not used explicitly on the command line). <var>choice</var> 1706can be one of ‘<samp>never</samp>’, ‘<samp>auto</samp>’, ‘<samp>always</samp>’, and ‘<samp>auto-if-env</samp>’ 1707where ‘<samp>auto</samp>’ is the default. ‘<samp>auto-if-env</samp>’ makes 1708<samp>-fdiagnostics-color=auto</samp> the default if <code>GCC_COLORS</code> 1709is present and non-empty in the environment of the compiler, and 1710<samp>-fdiagnostics-color=never</samp> otherwise. 1711</p> 1712</dd> 1713<dt><code>--with-diagnostics-urls=<var>choice</var></code></dt> 1714<dd><p>Tells GCC to use <var>choice</var> as the default for <samp>-fdiagnostics-urls=</samp> 1715option (if not used explicitly on the command line). <var>choice</var> 1716can be one of ‘<samp>never</samp>’, ‘<samp>auto</samp>’, ‘<samp>always</samp>’, and ‘<samp>auto-if-env</samp>’ 1717where ‘<samp>auto</samp>’ is the default. ‘<samp>auto-if-env</samp>’ makes 1718<samp>-fdiagnostics-urls=auto</samp> the default if <code>GCC_URLS</code> 1719or <code>TERM_URLS</code> is present and non-empty in the environment of the 1720compiler, and <samp>-fdiagnostics-urls=never</samp> otherwise. 1721</p> 1722</dd> 1723<dt><code>--enable-lto</code></dt> 1724<dt><code>--disable-lto</code></dt> 1725<dd><p>Enable support for link-time optimization (LTO). This is enabled by 1726default, and may be disabled using <samp>--disable-lto</samp>. 1727</p> 1728</dd> 1729<dt><code>--enable-linker-plugin-configure-flags=FLAGS</code></dt> 1730<dt><code>--enable-linker-plugin-flags=FLAGS</code></dt> 1731<dd><p>By default, linker plugins (such as the LTO plugin) are built for the 1732host system architecture. For the case that the linker has a 1733different (but run-time compatible) architecture, these flags can be 1734specified to build plugins that are compatible to the linker. For 1735example, if you are building GCC for a 64-bit x86_64 1736(‘<samp>x86_64-pc-linux-gnu</samp>’) host system, but have a 32-bit x86 1737GNU/Linux (‘<samp>i686-pc-linux-gnu</samp>’) linker executable (which is 1738executable on the former system), you can configure GCC as follows for 1739getting compatible linker plugins: 1740</p> 1741<div class="smallexample"> 1742<pre class="smallexample">% <var>srcdir</var>/configure \ 1743 --host=x86_64-pc-linux-gnu \ 1744 --enable-linker-plugin-configure-flags=--host=i686-pc-linux-gnu \ 1745 --enable-linker-plugin-flags='CC=gcc\ -m32\ -Wl,-rpath,[...]/i686-pc-linux-gnu/lib' 1746</pre></div> 1747 1748</dd> 1749<dt><code>--with-plugin-ld=<var>pathname</var></code></dt> 1750<dd><p>Enable an alternate linker to be used at link-time optimization (LTO) 1751link time when <samp>-fuse-linker-plugin</samp> is enabled. 1752This linker should have plugin support such as gold starting with 1753version 2.20 or GNU ld starting with version 2.21. 1754See <samp>-fuse-linker-plugin</samp> for details. 1755</p> 1756</dd> 1757<dt><code>--enable-canonical-system-headers</code></dt> 1758<dt><code>--disable-canonical-system-headers</code></dt> 1759<dd><p>Enable system header path canonicalization for <samp>libcpp</samp>. This can 1760produce shorter header file paths in diagnostics and dependency output 1761files, but these changed header paths may conflict with some compilation 1762environments. Enabled by default, and may be disabled using 1763<samp>--disable-canonical-system-headers</samp>. 1764</p> 1765</dd> 1766<dt><code>--with-glibc-version=<var>major</var>.<var>minor</var></code></dt> 1767<dd><p>Tell GCC that when the GNU C Library (glibc) is used on the target it 1768will be version <var>major</var>.<var>minor</var> or later. Normally this can 1769be detected from the C library’s header files, but this option may be 1770needed when bootstrapping a cross toolchain without the header files 1771available for building the initial bootstrap compiler. 1772</p> 1773<p>If GCC is configured with some multilibs that use glibc and some that 1774do not, this option applies only to the multilibs that use glibc. 1775However, such configurations may not work well as not all the relevant 1776configuration in GCC is on a per-multilib basis. 1777</p> 1778</dd> 1779<dt><code>--enable-as-accelerator-for=<var>target</var></code></dt> 1780<dd><p>Build as offload target compiler. Specify offload host triple by <var>target</var>. 1781</p> 1782</dd> 1783<dt><code>--enable-offload-targets=<var>target1</var>[=<var>path1</var>],…,<var>targetN</var>[=<var>pathN</var>]</code></dt> 1784<dd><p>Enable offloading to targets <var>target1</var>, …, <var>targetN</var>. 1785Offload compilers are expected to be already installed. Default search 1786path for them is <samp><var>exec-prefix</var></samp>, but it can be changed by 1787specifying paths <var>path1</var>, …, <var>pathN</var>. 1788</p> 1789<div class="smallexample"> 1790<pre class="smallexample">% <var>srcdir</var>/configure \ 1791 --enable-offload-targets=x86_64-intelmicemul-linux-gnu=/path/to/x86_64/compiler,nvptx-none,hsa 1792</pre></div> 1793 1794<p>If ‘<samp>hsa</samp>’ is specified as one of the targets, the compiler will be 1795built with support for HSA GPU accelerators. Because the same 1796compiler will emit the accelerator code, no path should be specified. 1797</p> 1798</dd> 1799<dt><code>--with-hsa-runtime=<var>pathname</var></code></dt> 1800<dt><code>--with-hsa-runtime-include=<var>pathname</var></code></dt> 1801<dt><code>--with-hsa-runtime-lib=<var>pathname</var></code></dt> 1802<dd> 1803<p>If you configure GCC with HSA offloading but do not have the HSA 1804run-time library installed in a standard location then you can 1805explicitly specify the directory where they are installed. The 1806<samp>--with-hsa-runtime=<var>hsainstalldir</var></samp> option is a 1807shorthand for 1808<samp>--with-hsa-runtime-lib=<var>hsainstalldir</var>/lib</samp> and 1809<samp>--with-hsa-runtime-include=<var>hsainstalldir</var>/include</samp>. 1810</p> 1811</dd> 1812<dt><code>--enable-cet</code></dt> 1813<dt><code>--disable-cet</code></dt> 1814<dd><p>Enable building target run-time libraries with control-flow 1815instrumentation, see <samp>-fcf-protection</samp> option. When 1816<code>--enable-cet</code> is specified target libraries are configured 1817to add <samp>-fcf-protection</samp> and, if needed, other target 1818specific options to a set of building options. 1819</p> 1820<p>The option is disabled by default. When <code>--enable-cet=auto</code> 1821is used, it is enabled on Linux/x86 if target binutils 1822supports <code>Intel CET</code> instructions and disabled otherwise. 1823In this case the target libraries are configured to get additional 1824<samp>-fcf-protection</samp> option. 1825</p> 1826</dd> 1827<dt><code>--with-riscv-attribute=‘<samp>yes</samp>’, ‘<samp>no</samp>’ or ‘<samp>default</samp>’</code></dt> 1828<dd><p>Generate RISC-V attribute by default, in order to record extra build 1829information in object. 1830</p> 1831<p>The option is disabled by default. It is enabled on RISC-V/ELF (bare-metal) 1832target if target binutils supported. 1833</p></dd> 1834</dl> 1835 1836<a name="Cross-Compiler-Specific-Options"></a> 1837<h4 class="subheading">Cross-Compiler-Specific Options</h4> 1838<p>The following options only apply to building cross compilers. 1839</p> 1840<dl compact="compact"> 1841<dt><code>--with-toolexeclibdir=<var>dir</var></code></dt> 1842<dd><p>Specify the installation directory for libraries built with a cross compiler. 1843The default is <samp>${gcc_tooldir}/lib</samp>. 1844</p> 1845</dd> 1846<dt><code>--with-sysroot</code></dt> 1847<dt><code>--with-sysroot=<var>dir</var></code></dt> 1848<dd><p>Tells GCC to consider <var>dir</var> as the root of a tree that contains 1849(a subset of) the root filesystem of the target operating system. 1850Target system headers, libraries and run-time object files will be 1851searched for in there. More specifically, this acts as if 1852<samp>--sysroot=<var>dir</var></samp> was added to the default options of the built 1853compiler. The specified directory is not copied into the 1854install tree, unlike the options <samp>--with-headers</samp> and 1855<samp>--with-libs</samp> that this option obsoletes. The default value, 1856in case <samp>--with-sysroot</samp> is not given an argument, is 1857<samp>${gcc_tooldir}/sys-root</samp>. If the specified directory is a 1858subdirectory of <samp>${exec_prefix}</samp>, then it will be found relative to 1859the GCC binaries if the installation tree is moved. 1860</p> 1861<p>This option affects the system root for the compiler used to build 1862target libraries (which runs on the build system) and the compiler newly 1863installed with <code>make install</code>; it does not affect the compiler which is 1864used to build GCC itself. 1865</p> 1866<p>If you specify the <samp>--with-native-system-header-dir=<var>dirname</var></samp> 1867option then the compiler will search that directory within <var>dirname</var> for 1868native system headers rather than the default <samp>/usr/include</samp>. 1869</p> 1870</dd> 1871<dt><code>--with-build-sysroot</code></dt> 1872<dt><code>--with-build-sysroot=<var>dir</var></code></dt> 1873<dd><p>Tells GCC to consider <var>dir</var> as the system root (see 1874<samp>--with-sysroot</samp>) while building target libraries, instead of 1875the directory specified with <samp>--with-sysroot</samp>. This option is 1876only useful when you are already using <samp>--with-sysroot</samp>. You 1877can use <samp>--with-build-sysroot</samp> when you are configuring with 1878<samp>--prefix</samp> set to a directory that is different from the one in 1879which you are installing GCC and your target libraries. 1880</p> 1881<p>This option affects the system root for the compiler used to build 1882target libraries (which runs on the build system); it does not affect 1883the compiler which is used to build GCC itself. 1884</p> 1885<p>If you specify the <samp>--with-native-system-header-dir=<var>dirname</var></samp> 1886option then the compiler will search that directory within <var>dirname</var> for 1887native system headers rather than the default <samp>/usr/include</samp>. 1888</p> 1889</dd> 1890<dt><code>--with-headers</code></dt> 1891<dt><code>--with-headers=<var>dir</var></code></dt> 1892<dd><p>Deprecated in favor of <samp>--with-sysroot</samp>. 1893Specifies that target headers are available when building a cross compiler. 1894The <var>dir</var> argument specifies a directory which has the target include 1895files. These include files will be copied into the <samp>gcc</samp> install 1896directory. <em>This option with the <var>dir</var> argument is required</em> when 1897building a cross compiler, if <samp><var>prefix</var>/<var>target</var>/sys-include</samp> 1898doesn’t pre-exist. If <samp><var>prefix</var>/<var>target</var>/sys-include</samp> does 1899pre-exist, the <var>dir</var> argument may be omitted. <code>fixincludes</code> 1900will be run on these files to make them compatible with GCC. 1901</p> 1902</dd> 1903<dt><code>--without-headers</code></dt> 1904<dd><p>Tells GCC not use any target headers from a libc when building a cross 1905compiler. When crossing to GNU/Linux, you need the headers so GCC 1906can build the exception handling for libgcc. 1907</p> 1908</dd> 1909<dt><code>--with-libs</code></dt> 1910<dt><code>--with-libs="<var>dir1</var> <var>dir2</var> … <var>dirN</var>"</code></dt> 1911<dd><p>Deprecated in favor of <samp>--with-sysroot</samp>. 1912Specifies a list of directories which contain the target runtime 1913libraries. These libraries will be copied into the <samp>gcc</samp> install 1914directory. If the directory list is omitted, this option has no 1915effect. 1916</p> 1917</dd> 1918<dt><code>--with-newlib</code></dt> 1919<dd><p>Specifies that ‘<samp>newlib</samp>’ is 1920being used as the target C library. This causes <code>__eprintf</code> to be 1921omitted from <samp>libgcc.a</samp> on the assumption that it will be provided by 1922‘<samp>newlib</samp>’. 1923</p> 1924<a name="avr"></a> 1925</dd> 1926<dt><code>--with-avrlibc</code></dt> 1927<dd><p>Only supported for the AVR target. Specifies that ‘<samp>AVR-Libc</samp>’ is 1928being used as the target C library. This causes float support 1929functions like <code>__addsf3</code> to be omitted from <samp>libgcc.a</samp> on 1930the assumption that it will be provided by <samp>libm.a</samp>. For more 1931technical details, cf. <a href="http://gcc.gnu.org/PR54461">PR54461</a>. 1932It is not supported for 1933RTEMS configurations, which currently use newlib. The option is 1934supported since version 4.7.2 and is the default in 4.8.0 and newer. 1935</p> 1936</dd> 1937<dt><code>--with-double={32|64|32,64|64,32}</code></dt> 1938<dt><code>--with-long-double={32|64|32,64|64,32|double}</code></dt> 1939<dd><p>Only supported for the AVR target since version 10. 1940Specify the default layout available for the C/C++ ‘<samp>double</samp>’ 1941and ‘<samp>long double</samp>’ type, respectively. The following rules apply: 1942</p><ul> 1943<li> The first value after the ‘<samp>=</samp>’ specifies the default layout (in bits) 1944of the type and also the default for the <samp>-mdouble=</samp> resp. 1945<samp>-mlong-double=</samp> compiler option. 1946</li><li> If more than one value is specified, respective multilib variants are 1947available, and <samp>-mdouble=</samp> resp. <samp>-mlong-double=</samp> acts 1948as a multilib option. 1949</li><li> If <samp>--with-long-double=double</samp> is specified, ‘<samp>double</samp>’ and 1950‘<samp>long double</samp>’ will have the same layout. 1951</li><li> The defaults are <samp>--with-long-double=64,32</samp> and 1952<samp>--with-double=32,64</samp>. The default ‘<samp>double</samp>’ layout imposed by 1953the latter is compatible with older versions of the compiler that implement 1954‘<samp>double</samp>’ as a 32-bit type, which does not comply to the language standard. 1955</li></ul> 1956<p>Not all combinations of <samp>--with-double=</samp> and 1957<samp>--with-long-double=</samp> are valid. For example, the combination 1958<samp>--with-double=32,64</samp> <samp>--with-long-double=32</samp> will be 1959rejected because the first option specifies the availability of 1960multilibs for ‘<samp>double</samp>’, whereas the second option implies 1961that ‘<samp>long double</samp>’ — and hence also ‘<samp>double</samp>’ — is always 196232 bits wide. 1963</p> 1964</dd> 1965<dt><code>--with-double-comparison={tristate|bool|libf7}</code></dt> 1966<dd><p>Only supported for the AVR target since version 10. 1967Specify what result format is returned by library functions that 1968compare 64-bit floating point values (<code>DFmode</code>). 1969The GCC default is ‘<samp>tristate</samp>’. If the floating point 1970implementation returns a boolean instead, set it to ‘<samp>bool</samp>’. 1971</p> 1972</dd> 1973<dt><code>--with-libf7={libgcc|math|math-symbols|no}</code></dt> 1974<dd><p>Only supported for the AVR target since version 10. 1975Specify to which degree code from LibF7 is included in libgcc. 1976LibF7 is an ad-hoc, AVR-specific, 64-bit floating point emulation 1977written in C and (inline) assembly. ‘<samp>libgcc</samp>’ adds support 1978for functions that one would usually expect in libgcc like double addition, 1979double comparisons and double conversions. ‘<samp>math</samp>’ also adds routines 1980that one would expect in <samp>libm.a</samp>, but with <code>__</code> (two underscores) 1981prepended to the symbol names as specified by <samp>math.h</samp>. 1982‘<samp>math-symbols</samp>’ also defines weak aliases for the functions 1983declared in <samp>math.h</samp>. However, <code>--with-libf7</code> won’t 1984install no <samp>math.h</samp> header file whatsoever, this file must come 1985from elsewhere. This option sets <samp>--with-double-comparison</samp> 1986to ‘<samp>bool</samp>’. 1987</p> 1988</dd> 1989<dt><code>--with-nds32-lib=<var>library</var></code></dt> 1990<dd><p>Specifies that <var>library</var> setting is used for building <samp>libgcc.a</samp>. 1991Currently, the valid <var>library</var> is ‘<samp>newlib</samp>’ or ‘<samp>mculib</samp>’. 1992This option is only supported for the NDS32 target. 1993</p> 1994</dd> 1995<dt><code>--with-build-time-tools=<var>dir</var></code></dt> 1996<dd><p>Specifies where to find the set of target tools (assembler, linker, etc.) 1997that will be used while building GCC itself. This option can be useful 1998if the directory layouts are different between the system you are building 1999GCC on, and the system where you will deploy it. 2000</p> 2001<p>For example, on an ‘<samp>ia64-hp-hpux</samp>’ system, you may have the GNU 2002assembler and linker in <samp>/usr/bin</samp>, and the native tools in a 2003different path, and build a toolchain that expects to find the 2004native tools in <samp>/usr/bin</samp>. 2005</p> 2006<p>When you use this option, you should ensure that <var>dir</var> includes 2007<code>ar</code>, <code>as</code>, <code>ld</code>, <code>nm</code>, 2008<code>ranlib</code> and <code>strip</code> if necessary, and possibly 2009<code>objdump</code>. Otherwise, GCC may use an inconsistent set of 2010tools. 2011</p></dd> 2012</dl> 2013 2014<a name="Overriding-configure-test-results"></a> 2015<h4 class="subsubheading">Overriding <code>configure</code> test results</h4> 2016 2017<p>Sometimes, it might be necessary to override the result of some 2018<code>configure</code> test, for example in order to ease porting to a new 2019system or work around a bug in a test. The toplevel <code>configure</code> 2020script provides three variables for this: 2021</p> 2022<dl compact="compact"> 2023<dt><code>build_configargs</code></dt> 2024<dd><a name="index-build_005fconfigargs"></a> 2025<p>The contents of this variable is passed to all build <code>configure</code> 2026scripts. 2027</p> 2028</dd> 2029<dt><code>host_configargs</code></dt> 2030<dd><a name="index-host_005fconfigargs"></a> 2031<p>The contents of this variable is passed to all host <code>configure</code> 2032scripts. 2033</p> 2034</dd> 2035<dt><code>target_configargs</code></dt> 2036<dd><a name="index-target_005fconfigargs"></a> 2037<p>The contents of this variable is passed to all target <code>configure</code> 2038scripts. 2039</p> 2040</dd> 2041</dl> 2042 2043<p>In order to avoid shell and <code>make</code> quoting issues for complex 2044overrides, you can pass a setting for <code>CONFIG_SITE</code> and set 2045variables in the site file. 2046</p> 2047<a name="Objective-C-Specific-Options"></a> 2048<h4 class="subheading">Objective-C-Specific Options</h4> 2049 2050<p>The following options apply to the build of the Objective-C runtime library. 2051</p> 2052<dl compact="compact"> 2053<dt><code>--enable-objc-gc</code></dt> 2054<dd><p>Specify that an additional variant of the GNU Objective-C runtime library 2055is built, using an external build of the Boehm-Demers-Weiser garbage 2056collector (<a href="https://www.hboehm.info/gc/">https://www.hboehm.info/gc/</a>). This library needs to be 2057available for each multilib variant, unless configured with 2058<samp>--enable-objc-gc=‘<samp>auto</samp>’</samp> in which case the build of the 2059additional runtime library is skipped when not available and the build 2060continues. 2061</p> 2062</dd> 2063<dt><code>--with-target-bdw-gc=<var>list</var></code></dt> 2064<dt><code>--with-target-bdw-gc-include=<var>list</var></code></dt> 2065<dt><code>--with-target-bdw-gc-lib=<var>list</var></code></dt> 2066<dd><p>Specify search directories for the garbage collector header files and 2067libraries. <var>list</var> is a comma separated list of key value pairs of the 2068form ‘<samp><var>multilibdir</var>=<var>path</var></samp>’, where the default multilib key 2069is named as ‘<samp>.</samp>’ (dot), or is omitted (e.g. 2070‘<samp>--with-target-bdw-gc=/opt/bdw-gc,32=/opt-bdw-gc32</samp>’). 2071</p> 2072<p>The options <samp>--with-target-bdw-gc-include</samp> and 2073<samp>--with-target-bdw-gc-lib</samp> must always be specified together 2074for each multilib variant and they take precedence over 2075<samp>--with-target-bdw-gc</samp>. If <samp>--with-target-bdw-gc-include</samp> 2076is missing values for a multilib, then the value for the default 2077multilib is used (e.g. ‘<samp>--with-target-bdw-gc-include=/opt/bdw-gc/include</samp>’ 2078‘<samp>--with-target-bdw-gc-lib=/opt/bdw-gc/lib64,32=/opt-bdw-gc/lib32</samp>’). 2079If none of these options are specified, the library is assumed in 2080default locations. 2081</p></dd> 2082</dl> 2083 2084<a name="D-Specific-Options"></a> 2085<h4 class="subheading">D-Specific Options</h4> 2086 2087<p>The following options apply to the build of the D runtime library. 2088</p> 2089<dl compact="compact"> 2090<dt><code>--enable-libphobos-checking</code></dt> 2091<dt><code>--disable-libphobos-checking</code></dt> 2092<dt><code>--enable-libphobos-checking=<var>list</var></code></dt> 2093<dd><p>This option controls whether run-time checks and contracts are compiled into 2094the D runtime library. When the option is not specified, the library is built 2095with ‘<samp>release</samp>’ checking. When the option is specified without a 2096<var>list</var>, the result is the same as ‘<samp>--enable-libphobos-checking=yes</samp>’. 2097Likewise, ‘<samp>--disable-libphobos-checking</samp>’ is equivalent to 2098‘<samp>--enable-libphobos-checking=no</samp>’. 2099</p> 2100<p>The categories of checks available in <var>list</var> are ‘<samp>yes</samp>’ (compiles 2101libphobos with <samp>-fno-release</samp>), ‘<samp>no</samp>’ (compiles libphobos with 2102<samp>-frelease</samp>), ‘<samp>all</samp>’ (same as ‘<samp>yes</samp>’), ‘<samp>none</samp>’ or 2103‘<samp>release</samp>’ (same as ‘<samp>no</samp>’). 2104</p> 2105<p>Individual checks available in <var>list</var> are ‘<samp>assert</samp>’ (compiles libphobos 2106with an extra option <samp>-fassert</samp>). 2107</p> 2108</dd> 2109<dt><code>--with-libphobos-druntime-only</code></dt> 2110<dt><code>--with-libphobos-druntime-only=<var>choice</var></code></dt> 2111<dd><p>Specify whether to build only the core D runtime library (druntime), or both 2112the core and standard library (phobos) into libphobos. This is useful for 2113targets that have full support in druntime, but no or incomplete support 2114in phobos. <var>choice</var> can be one of ‘<samp>auto</samp>’, ‘<samp>yes</samp>’, and ‘<samp>no</samp>’ 2115where ‘<samp>auto</samp>’ is the default. 2116</p> 2117<p>When the option is not specified, the default choice ‘<samp>auto</samp>’ means that it 2118is inferred whether the target has support for the phobos standard library. 2119When the option is specified without a <var>choice</var>, the result is the same as 2120‘<samp>--with-libphobos-druntime-only=yes</samp>’. 2121</p> 2122</dd> 2123<dt><code>--with-target-system-zlib</code></dt> 2124<dd><p>Use installed ‘<samp>zlib</samp>’ rather than that included with GCC. This needs 2125to be available for each multilib variant, unless configured with 2126<samp>--with-target-system-zlib=‘<samp>auto</samp>’</samp> in which case the GCC included 2127‘<samp>zlib</samp>’ is only used when the system installed library is not available. 2128</p></dd> 2129</dl> 2130 2131<hr /> 2132<p> 2133<p><a href="./index.html">Return to the GCC Installation page</a> 2134</p> 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146</body> 2147</html> 2148