specific.html revision 1.1 1 1.1 mrg <html lang="en">
2 1.1 mrg <head>
3 1.1 mrg <title>Host/Target specific installation notes for GCC</title>
4 1.1 mrg <meta http-equiv="Content-Type" content="text/html">
5 1.1 mrg <meta name="description" content="Host/Target specific installation notes for GCC">
6 1.1 mrg <meta name="generator" content="makeinfo 4.12">
7 1.1 mrg <link title="Top" rel="top" href="#Top">
8 1.1 mrg <link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage">
9 1.1 mrg <!--
10 1.1 mrg Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
11 1.1 mrg 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008,
12 1.1 mrg 2009, 2010, 2011 Free Software Foundation, Inc.
13 1.1 mrg
14 1.1 mrg Permission is granted to copy, distribute and/or modify this document
15 1.1 mrg under the terms of the GNU Free Documentation License, Version 1.2 or
16 1.1 mrg any later version published by the Free Software Foundation; with no
17 1.1 mrg Invariant Sections, the Front-Cover texts being (a) (see below), and
18 1.1 mrg with the Back-Cover Texts being (b) (see below). A copy of the
19 1.1 mrg license is included in the section entitled "GNU Free Documentation License".
20 1.1 mrg
21 1.1 mrg (a) The FSF's Front-Cover Text is:
22 1.1 mrg
23 1.1 mrg A GNU Manual
24 1.1 mrg
25 1.1 mrg (b) The FSF's Back-Cover Text is:
26 1.1 mrg
27 1.1 mrg You have freedom to copy and modify this GNU Manual, like GNU
28 1.1 mrg software. Copies published by the Free Software Foundation raise
29 1.1 mrg funds for GNU development.-->
30 1.1 mrg <meta http-equiv="Content-Style-Type" content="text/css">
31 1.1 mrg <style type="text/css"><!--
32 1.1 mrg pre.display { font-family:inherit }
33 1.1 mrg pre.format { font-family:inherit }
34 1.1 mrg pre.smalldisplay { font-family:inherit; font-size:smaller }
35 1.1 mrg pre.smallformat { font-family:inherit; font-size:smaller }
36 1.1 mrg pre.smallexample { font-size:smaller }
37 1.1 mrg pre.smalllisp { font-size:smaller }
38 1.1 mrg span.sc { font-variant:small-caps }
39 1.1 mrg span.roman { font-family:serif; font-weight:normal; }
40 1.1 mrg span.sansserif { font-family:sans-serif; font-weight:normal; }
41 1.1 mrg --></style>
42 1.1 mrg </head>
43 1.1 mrg <body>
44 1.1 mrg <h1 class="settitle">Host/Target specific installation notes for GCC</h1>
45 1.1 mrg <a name="index-Specific-1"></a><a name="index-Specific-installation-notes-2"></a><a name="index-Target-specific-installation-3"></a><a name="index-Host-specific-installation-4"></a><a name="index-Target-specific-installation-notes-5"></a>
46 1.1 mrg Please read this document carefully <em>before</em> installing the
47 1.1 mrg GNU Compiler Collection on your machine.
48 1.1 mrg
49 1.1 mrg <p>Note that this list of install notes is <em>not</em> a list of supported
50 1.1 mrg hosts or targets. Not all supported hosts and targets are listed
51 1.1 mrg here, only the ones that require host-specific or target-specific
52 1.1 mrg information are.
53 1.1 mrg
54 1.1 mrg <ul>
55 1.1 mrg <li><a href="#alpha-x-x">alpha*-*-*</a>
56 1.1 mrg <li><a href="#alpha-dec-osf">alpha*-dec-osf*</a>
57 1.1 mrg <li><a href="#arc-x-elf">arc-*-elf</a>
58 1.1 mrg <li><a href="#arm-x-elf">arm-*-elf</a>
59 1.1 mrg <li><a href="#avr">avr</a>
60 1.1 mrg <li><a href="#bfin">Blackfin</a>
61 1.1 mrg <li><a href="#dos">DOS</a>
62 1.1 mrg <li><a href="#x-x-freebsd">*-*-freebsd*</a>
63 1.1 mrg <li><a href="#h8300-hms">h8300-hms</a>
64 1.1 mrg <li><a href="#hppa-hp-hpux">hppa*-hp-hpux*</a>
65 1.1 mrg <li><a href="#hppa-hp-hpux10">hppa*-hp-hpux10</a>
66 1.1 mrg <li><a href="#hppa-hp-hpux11">hppa*-hp-hpux11</a>
67 1.1 mrg <li><a href="#x-x-linux-gnu">*-*-linux-gnu</a>
68 1.1 mrg <li><a href="#ix86-x-linux">i?86-*-linux*</a>
69 1.1 mrg <li><a href="#ix86-x-solaris289">i?86-*-solaris2.[89]</a>
70 1.1 mrg <li><a href="#ix86-x-solaris210">i?86-*-solaris2.10</a>
71 1.1 mrg <li><a href="#ia64-x-linux">ia64-*-linux</a>
72 1.1 mrg <li><a href="#ia64-x-hpux">ia64-*-hpux*</a>
73 1.1 mrg <li><a href="#x-ibm-aix">*-ibm-aix*</a>
74 1.1 mrg <li><a href="#iq2000-x-elf">iq2000-*-elf</a>
75 1.1 mrg <li><a href="#lm32-x-elf">lm32-*-elf</a>
76 1.1 mrg <li><a href="#lm32-x-uclinux">lm32-*-uclinux</a>
77 1.1 mrg <li><a href="#m32c-x-elf">m32c-*-elf</a>
78 1.1 mrg <li><a href="#m32r-x-elf">m32r-*-elf</a>
79 1.1 mrg <li><a href="#m6811-elf">m6811-elf</a>
80 1.1 mrg <li><a href="#m6812-elf">m6812-elf</a>
81 1.1 mrg <li><a href="#m68k-x-x">m68k-*-*</a>
82 1.1 mrg <li><a href="#m68k-uclinux">m68k-uclinux</a>
83 1.1 mrg <li><a href="#mep-x-elf">mep-*-elf</a>
84 1.1 mrg <li><a href="#mips-x-x">mips-*-*</a>
85 1.1 mrg <li><a href="#mips-sgi-irix5">mips-sgi-irix5</a>
86 1.1 mrg <li><a href="#mips-sgi-irix6">mips-sgi-irix6</a>
87 1.1 mrg <li><a href="#powerpc-x-x">powerpc*-*-*</a>
88 1.1 mrg <li><a href="#powerpc-x-darwin">powerpc-*-darwin*</a>
89 1.1 mrg <li><a href="#powerpc-x-elf">powerpc-*-elf</a>
90 1.1 mrg <li><a href="#powerpc-x-linux-gnu">powerpc*-*-linux-gnu*</a>
91 1.1 mrg <li><a href="#powerpc-x-netbsd">powerpc-*-netbsd*</a>
92 1.1 mrg <li><a href="#powerpc-x-eabisim">powerpc-*-eabisim</a>
93 1.1 mrg <li><a href="#powerpc-x-eabi">powerpc-*-eabi</a>
94 1.1 mrg <li><a href="#powerpcle-x-elf">powerpcle-*-elf</a>
95 1.1 mrg <li><a href="#powerpcle-x-eabisim">powerpcle-*-eabisim</a>
96 1.1 mrg <li><a href="#powerpcle-x-eabi">powerpcle-*-eabi</a>
97 1.1 mrg <li><a href="#s390-x-linux">s390-*-linux*</a>
98 1.1 mrg <li><a href="#s390x-x-linux">s390x-*-linux*</a>
99 1.1 mrg <li><a href="#s390x-ibm-tpf">s390x-ibm-tpf*</a>
100 1.1 mrg <li><a href="#x-x-solaris2">*-*-solaris2*</a>
101 1.1 mrg <li><a href="#sparc-sun-solaris2">sparc-sun-solaris2*</a>
102 1.1 mrg <li><a href="#sparc-sun-solaris27">sparc-sun-solaris2.7</a>
103 1.1 mrg <li><a href="#sparc-sun-solaris210">sparc-sun-solaris2.10</a>
104 1.1 mrg <li><a href="#sparc-x-linux">sparc-*-linux*</a>
105 1.1 mrg <li><a href="#sparc64-x-solaris2">sparc64-*-solaris2*</a>
106 1.1 mrg <li><a href="#sparcv9-x-solaris2">sparcv9-*-solaris2*</a>
107 1.1 mrg <li><a href="#x-x-vxworks">*-*-vxworks*</a>
108 1.1 mrg <li><a href="#x86-64-x-x">x86_64-*-*</a> amd64-*-*
109 1.1 mrg <li><a href="#xtensa-x-elf">xtensa*-*-elf</a>
110 1.1 mrg <li><a href="#xtensa-x-linux">xtensa*-*-linux*</a>
111 1.1 mrg <li><a href="#windows">Microsoft Windows</a>
112 1.1 mrg <li><a href="#x-x-cygwin">*-*-cygwin</a>
113 1.1 mrg <li><a href="#x-x-interix">*-*-interix</a>
114 1.1 mrg <li><a href="#x-x-mingw32">*-*-mingw32</a>
115 1.1 mrg <li><a href="#os2">OS/2</a>
116 1.1 mrg <li><a href="#older">Older systems</a>
117 1.1 mrg </ul>
118 1.1 mrg
119 1.1 mrg <ul>
120 1.1 mrg <li><a href="#elf">all ELF targets</a> (SVR4, Solaris 2, etc.)
121 1.1 mrg </ul>
122 1.1 mrg
123 1.1 mrg <p><!-- -------- host/target specific issues start here ---------------- -->
124 1.1 mrg <hr />
125 1.1 mrg
126 1.1 mrg <h3 class="heading"><a name="TOC0"></a><a name="alpha_002dx_002dx"></a>alpha*-*-*</h3>
127 1.1 mrg
128 1.1 mrg <p>This section contains general configuration information for all
129 1.1 mrg alpha-based platforms using ELF (in particular, ignore this section for
130 1.1 mrg DEC OSF/1, Digital UNIX and Tru64 UNIX). In addition to reading this
131 1.1 mrg section, please read all other sections that match your target.
132 1.1 mrg
133 1.1 mrg <p>We require binutils 2.11.2 or newer.
134 1.1 mrg Previous binutils releases had a number of problems with DWARF 2
135 1.1 mrg debugging information, not the least of which is incorrect linking of
136 1.1 mrg shared libraries.
137 1.1 mrg
138 1.1 mrg <p><hr />
139 1.1 mrg
140 1.1 mrg <h3 class="heading"><a name="TOC1"></a><a name="alpha_002ddec_002dosf"></a>alpha*-dec-osf*</h3>
141 1.1 mrg
142 1.1 mrg <p>Systems using processors that implement the DEC Alpha architecture and
143 1.1 mrg are running the DEC/Compaq/HP Unix (DEC OSF/1, Digital UNIX, or Compaq/HP
144 1.1 mrg Tru64 UNIX) operating system, for example the DEC Alpha AXP systems.
145 1.1 mrg
146 1.1 mrg <p>As of GCC 3.2, versions before <code>alpha*-dec-osf4</code> are no longer
147 1.1 mrg supported. (These are the versions which identify themselves as DEC
148 1.1 mrg OSF/1.) As of GCC 4.5, support for Tru64 UNIX V4.0 and V5.0 has been
149 1.1 mrg obsoleted, but can still be enabled by configuring with
150 1.1 mrg <samp><span class="option">--enable-obsolete</span></samp>. Support will be removed in GCC 4.6.
151 1.1 mrg
152 1.1 mrg <p>On Tru64 UNIX, virtual memory exhausted bootstrap failures
153 1.1 mrg may be fixed by reconfiguring Kernel Virtual Memory and Swap parameters
154 1.1 mrg per the <samp><span class="command">/usr/sbin/sys_check</span></samp> Tuning Suggestions,
155 1.1 mrg or applying the patch in
156 1.1 mrg <a href="http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html">http://gcc.gnu.org/ml/gcc/2002-08/msg00822.html</a>. Depending on
157 1.1 mrg the OS version used, you need a data segment size between 512 MB and
158 1.1 mrg 1 GB, so simply use <samp><span class="command">ulimit -Sd unlimited</span></samp>.
159 1.1 mrg
160 1.1 mrg <p>As of GNU binutils 2.20.1, neither GNU <samp><span class="command">as</span></samp> nor GNU <samp><span class="command">ld</span></samp>
161 1.1 mrg are supported on Tru64 UNIX, so you must not configure GCC with
162 1.1 mrg <samp><span class="option">--with-gnu-as</span></samp> or <samp><span class="option">--with-gnu-ld</span></samp>.
163 1.1 mrg
164 1.1 mrg <p>GCC writes a ‘<samp><span class="samp">.verstamp</span></samp>’ directive to the assembler output file
165 1.1 mrg unless it is built as a cross-compiler. It gets the version to use from
166 1.1 mrg the system header file <samp><span class="file">/usr/include/stamp.h</span></samp>. If you install a
167 1.1 mrg new version of Tru64 UNIX, you should rebuild GCC to pick up the new version
168 1.1 mrg stamp.
169 1.1 mrg
170 1.1 mrg <p>GCC now supports both the native (ECOFF) debugging format used by DBX
171 1.1 mrg and GDB and an encapsulated STABS format for use only with GDB. See the
172 1.1 mrg discussion of the <samp><span class="option">--with-stabs</span></samp> option of <samp><span class="file">configure</span></samp> above
173 1.1 mrg for more information on these formats and how to select them.
174 1.1 mrg <!-- FIXME: does this work at all? If so, perhaps make default. -->
175 1.1 mrg
176 1.1 mrg <p>There is a bug in DEC's assembler that produces incorrect line numbers
177 1.1 mrg for ECOFF format when the ‘<samp><span class="samp">.align</span></samp>’ directive is used. To work
178 1.1 mrg around this problem, GCC will not emit such alignment directives
179 1.1 mrg while writing ECOFF format debugging information even if optimization is
180 1.1 mrg being performed. Unfortunately, this has the very undesirable
181 1.1 mrg side-effect that code addresses when <samp><span class="option">-O</span></samp> is specified are
182 1.1 mrg different depending on whether or not <samp><span class="option">-g</span></samp> is also specified.
183 1.1 mrg
184 1.1 mrg <p>To avoid this behavior, specify <samp><span class="option">-gstabs+</span></samp> and use GDB instead of
185 1.1 mrg DBX. DEC is now aware of this problem with the assembler and hopes to
186 1.1 mrg provide a fix shortly.
187 1.1 mrg
188 1.1 mrg <!-- FIXME: still applicable? -->
189 1.1 mrg <p><hr />
190 1.1 mrg
191 1.1 mrg <h3 class="heading"><a name="TOC2"></a><a name="arc_002dx_002delf"></a>arc-*-elf</h3>
192 1.1 mrg
193 1.1 mrg <p>Argonaut ARC processor.
194 1.1 mrg This configuration is intended for embedded systems.
195 1.1 mrg
196 1.1 mrg <p><hr />
197 1.1 mrg
198 1.1 mrg <h3 class="heading"><a name="TOC3"></a><a name="arm_002dx_002delf"></a>arm-*-elf</h3>
199 1.1 mrg
200 1.1 mrg <p>ARM-family processors. Subtargets that use the ELF object format
201 1.1 mrg require GNU binutils 2.13 or newer. Such subtargets include:
202 1.1 mrg <code>arm-*-freebsd</code>, <code>arm-*-netbsdelf</code>, <code>arm-*-*linux</code>
203 1.1 mrg and <code>arm-*-rtems</code>.
204 1.1 mrg
205 1.1 mrg <p><hr />
206 1.1 mrg
207 1.1 mrg <h3 class="heading"><a name="TOC4"></a><a name="avr"></a>avr</h3>
208 1.1 mrg
209 1.1 mrg <p>ATMEL AVR-family micro controllers. These are used in embedded
210 1.1 mrg applications. There are no standard Unix configurations.
211 1.1 mrg See “AVR Options” in the main manual
212 1.1 mrg for the list of supported MCU types.
213 1.1 mrg
214 1.1 mrg <p>Use ‘<samp><span class="samp">configure --target=avr --enable-languages="c"</span></samp>’ to configure GCC.
215 1.1 mrg
216 1.1 mrg <p>Further installation notes and other useful information about AVR tools
217 1.1 mrg can also be obtained from:
218 1.1 mrg
219 1.1 mrg <ul>
220 1.1 mrg <li><a href="http://www.nongnu.org/avr/">http://www.nongnu.org/avr/</a>
221 1.1 mrg <li><a href="http://www.amelek.gda.pl/avr/">http://www.amelek.gda.pl/avr/</a>
222 1.1 mrg </ul>
223 1.1 mrg
224 1.1 mrg <p>We <em>strongly</em> recommend using binutils 2.13 or newer.
225 1.1 mrg
226 1.1 mrg <p>The following error:
227 1.1 mrg <pre class="smallexample"> Error: register required
228 1.1 mrg </pre>
229 1.1 mrg <p>indicates that you should upgrade to a newer version of the binutils.
230 1.1 mrg
231 1.1 mrg <p><hr />
232 1.1 mrg
233 1.1 mrg <h3 class="heading"><a name="TOC5"></a><a name="bfin"></a>Blackfin</h3>
234 1.1 mrg
235 1.1 mrg <p>The Blackfin processor, an Analog Devices DSP.
236 1.1 mrg See “Blackfin Options” in the main manual
237 1.1 mrg
238 1.1 mrg <p>More information, and a version of binutils with support for this processor,
239 1.1 mrg is available at <a href="http://blackfin.uclinux.org">http://blackfin.uclinux.org</a>
240 1.1 mrg
241 1.1 mrg <p><hr />
242 1.1 mrg
243 1.1 mrg <h3 class="heading"><a name="TOC6"></a><a name="cris"></a>CRIS</h3>
244 1.1 mrg
245 1.1 mrg <p>CRIS is the CPU architecture in Axis Communications ETRAX system-on-a-chip
246 1.1 mrg series. These are used in embedded applications.
247 1.1 mrg
248 1.1 mrg <p>See “CRIS Options” in the main manual
249 1.1 mrg for a list of CRIS-specific options.
250 1.1 mrg
251 1.1 mrg <p>There are a few different CRIS targets:
252 1.1 mrg <dl>
253 1.1 mrg <dt><code>cris-axis-elf</code><dd>Mainly for monolithic embedded systems. Includes a multilib for the
254 1.1 mrg ‘<samp><span class="samp">v10</span></samp>’ core used in ‘<samp><span class="samp">ETRAX 100 LX</span></samp>’.
255 1.1 mrg <br><dt><code>cris-axis-linux-gnu</code><dd>A GNU/Linux port for the CRIS architecture, currently targeting
256 1.1 mrg ‘<samp><span class="samp">ETRAX 100 LX</span></samp>’ by default.
257 1.1 mrg </dl>
258 1.1 mrg
259 1.1 mrg <p>For <code>cris-axis-elf</code> you need binutils 2.11
260 1.1 mrg or newer. For <code>cris-axis-linux-gnu</code> you need binutils 2.12 or newer.
261 1.1 mrg
262 1.1 mrg <p>Pre-packaged tools can be obtained from
263 1.1 mrg <a href="ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/">ftp://ftp.axis.com/pub/axis/tools/cris/compiler-kit/</a>. More
264 1.1 mrg information about this platform is available at
265 1.1 mrg <a href="http://developer.axis.com/">http://developer.axis.com/</a>.
266 1.1 mrg
267 1.1 mrg <p><hr />
268 1.1 mrg
269 1.1 mrg <h3 class="heading"><a name="TOC7"></a><a name="crx"></a>CRX</h3>
270 1.1 mrg
271 1.1 mrg <p>The CRX CompactRISC architecture is a low-power 32-bit architecture with
272 1.1 mrg fast context switching and architectural extensibility features.
273 1.1 mrg
274 1.1 mrg <p>See “CRX Options” in the main manual for a list of CRX-specific options.
275 1.1 mrg
276 1.1 mrg <p>Use ‘<samp><span class="samp">configure --target=crx-elf --enable-languages=c,c++</span></samp>’ to configure
277 1.1 mrg GCC for building a CRX cross-compiler. The option ‘<samp><span class="samp">--target=crx-elf</span></samp>’
278 1.1 mrg is also used to build the ‘<samp><span class="samp">newlib</span></samp>’ C library for CRX.
279 1.1 mrg
280 1.1 mrg <p>It is also possible to build libstdc++-v3 for the CRX architecture. This
281 1.1 mrg needs to be done in a separate step with the following configure settings:
282 1.1 mrg ‘<samp><span class="samp">gcc/libstdc++-v3/configure --host=crx-elf --with-newlib
283 1.1 mrg --enable-sjlj-exceptions --enable-cxx-flags='-fexceptions -frtti'</span></samp>’
284 1.1 mrg
285 1.1 mrg <p><hr />
286 1.1 mrg
287 1.1 mrg <h3 class="heading"><a name="TOC8"></a><a name="dos"></a>DOS</h3>
288 1.1 mrg
289 1.1 mrg <p>Please have a look at the <a href="binaries.html">binaries page</a>.
290 1.1 mrg
291 1.1 mrg <p>You cannot install GCC by itself on MSDOS; it will not compile under
292 1.1 mrg any MSDOS compiler except itself. You need to get the complete
293 1.1 mrg compilation package DJGPP, which includes binaries as well as sources,
294 1.1 mrg and includes all the necessary compilation tools and libraries.
295 1.1 mrg
296 1.1 mrg <p><hr />
297 1.1 mrg
298 1.1 mrg <h3 class="heading"><a name="TOC9"></a><a name="x_002dx_002dfreebsd"></a>*-*-freebsd*</h3>
299 1.1 mrg
300 1.1 mrg <p>Support for FreeBSD 1 was discontinued in GCC 3.2. Support for
301 1.1 mrg FreeBSD 2 (and any mutant a.out variants of FreeBSD 3) was
302 1.1 mrg discontinued in GCC 4.0.
303 1.1 mrg
304 1.1 mrg <p>In GCC 4.5, we enabled the use of <code>dl_iterate_phdr</code> inside boehm-gc on
305 1.1 mrg FreeBSD 7 or later. In order to better match the configuration of the
306 1.1 mrg FreeBSD system compiler: We also enabled the check to see if libc
307 1.1 mrg provides SSP support (which it does on FreeBSD 7), the use of
308 1.1 mrg <code>dl_iterate_phdr</code> inside <samp><span class="file">libgcc_s.so.1</span></samp> (on FreeBSD 7 or later)
309 1.1 mrg and the use of <code>__cxa_atexit</code> by default (on FreeBSD 6 or later).
310 1.1 mrg
311 1.1 mrg <p>We support FreeBSD using the ELF file format with DWARF 2 debugging
312 1.1 mrg for all CPU architectures. You may use <samp><span class="option">-gstabs</span></samp> instead of
313 1.1 mrg <samp><span class="option">-g</span></samp>, if you really want the old debugging format. There are
314 1.1 mrg no known issues with mixing object files and libraries with different
315 1.1 mrg debugging formats. Otherwise, this release of GCC should now match
316 1.1 mrg more of the configuration used in the stock FreeBSD configuration of
317 1.1 mrg GCC. In particular, <samp><span class="option">--enable-threads</span></samp> is now configured by
318 1.1 mrg default. However, as a general user, do not attempt to replace the
319 1.1 mrg system compiler with this release. Known to bootstrap and check with
320 1.1 mrg good results on FreeBSD 7.2-STABLE. In the past, known to bootstrap
321 1.1 mrg and check with good results on FreeBSD 3.0, 3.4, 4.0, 4.2, 4.3, 4.4,
322 1.1 mrg 4.5, 4.8, 4.9 and 5-CURRENT.
323 1.1 mrg
324 1.1 mrg <p>The version of binutils installed in <samp><span class="file">/usr/bin</span></samp> probably works
325 1.1 mrg with this release of GCC. Bootstrapping against the latest GNU
326 1.1 mrg binutils and/or the version found in <samp><span class="file">/usr/ports/devel/binutils</span></samp> has
327 1.1 mrg been known to enable additional features and improve overall testsuite
328 1.1 mrg results. However, it is currently known that boehm-gc (which itself
329 1.1 mrg is required for java) may not configure properly on FreeBSD prior to
330 1.1 mrg the FreeBSD 7.0 release with GNU binutils after 2.16.1.
331 1.1 mrg
332 1.1 mrg <p><hr />
333 1.1 mrg
334 1.1 mrg <h3 class="heading"><a name="TOC10"></a><a name="h8300_002dhms"></a>h8300-hms</h3>
335 1.1 mrg
336 1.1 mrg <p>Renesas H8/300 series of processors.
337 1.1 mrg
338 1.1 mrg <p>Please have a look at the <a href="binaries.html">binaries page</a>.
339 1.1 mrg
340 1.1 mrg <p>The calling convention and structure layout has changed in release 2.6.
341 1.1 mrg All code must be recompiled. The calling convention now passes the
342 1.1 mrg first three arguments in function calls in registers. Structures are no
343 1.1 mrg longer a multiple of 2 bytes.
344 1.1 mrg
345 1.1 mrg <p><hr />
346 1.1 mrg
347 1.1 mrg <h3 class="heading"><a name="TOC11"></a><a name="hppa_002dhp_002dhpux"></a>hppa*-hp-hpux*</h3>
348 1.1 mrg
349 1.1 mrg <p>Support for HP-UX version 9 and older was discontinued in GCC 3.4.
350 1.1 mrg
351 1.1 mrg <p>We require using gas/binutils on all hppa platforms. Version 2.19 or
352 1.1 mrg later is recommended.
353 1.1 mrg
354 1.1 mrg <p>It may be helpful to configure GCC with the
355 1.1 mrg <a href="./configure.html#with-gnu-as"><samp><span class="option">--with-gnu-as</span></samp></a> and
356 1.1 mrg <samp><span class="option">--with-as=...</span></samp> options to ensure that GCC can find GAS.
357 1.1 mrg
358 1.1 mrg <p>The HP assembler should not be used with GCC. It is rarely tested and may
359 1.1 mrg not work. It shouldn't be used with any languages other than C due to its
360 1.1 mrg many limitations.
361 1.1 mrg
362 1.1 mrg <p>Specifically, <samp><span class="option">-g</span></samp> does not work (HP-UX uses a peculiar debugging
363 1.1 mrg format which GCC does not know about). It also inserts timestamps
364 1.1 mrg into each object file it creates, causing the 3-stage comparison test to
365 1.1 mrg fail during a bootstrap. You should be able to continue by saying
366 1.1 mrg ‘<samp><span class="samp">make all-host all-target</span></samp>’ after getting the failure from ‘<samp><span class="samp">make</span></samp>’.
367 1.1 mrg
368 1.1 mrg <p>Various GCC features are not supported. For example, it does not support weak
369 1.1 mrg symbols or alias definitions. As a result, explicit template instantiations
370 1.1 mrg are required when using C++. This makes it difficult if not impossible to
371 1.1 mrg build many C++ applications.
372 1.1 mrg
373 1.1 mrg <p>There are two default scheduling models for instructions. These are
374 1.1 mrg PROCESSOR_7100LC and PROCESSOR_8000. They are selected from the pa-risc
375 1.1 mrg architecture specified for the target machine when configuring.
376 1.1 mrg PROCESSOR_8000 is the default. PROCESSOR_7100LC is selected when
377 1.1 mrg the target is a ‘<samp><span class="samp">hppa1*</span></samp>’ machine.
378 1.1 mrg
379 1.1 mrg <p>The PROCESSOR_8000 model is not well suited to older processors. Thus,
380 1.1 mrg it is important to completely specify the machine architecture when
381 1.1 mrg configuring if you want a model other than PROCESSOR_8000. The macro
382 1.1 mrg TARGET_SCHED_DEFAULT can be defined in BOOT_CFLAGS if a different
383 1.1 mrg default scheduling model is desired.
384 1.1 mrg
385 1.1 mrg <p>As of GCC 4.0, GCC uses the UNIX 95 namespace for HP-UX 10.10
386 1.1 mrg through 11.00, and the UNIX 98 namespace for HP-UX 11.11 and later.
387 1.1 mrg This namespace change might cause problems when bootstrapping with
388 1.1 mrg an earlier version of GCC or the HP compiler as essentially the same
389 1.1 mrg namespace is required for an entire build. This problem can be avoided
390 1.1 mrg in a number of ways. With HP cc, <samp><span class="env">UNIX_STD</span></samp> can be set to ‘<samp><span class="samp">95</span></samp>’
391 1.1 mrg or ‘<samp><span class="samp">98</span></samp>’. Another way is to add an appropriate set of predefines
392 1.1 mrg to <samp><span class="env">CC</span></samp>. The description for the <samp><span class="option">munix=</span></samp> option contains
393 1.1 mrg a list of the predefines used with each standard.
394 1.1 mrg
395 1.1 mrg <p>More specific information to ‘<samp><span class="samp">hppa*-hp-hpux*</span></samp>’ targets follows.
396 1.1 mrg
397 1.1 mrg <p><hr />
398 1.1 mrg
399 1.1 mrg <h3 class="heading"><a name="TOC12"></a><a name="hppa_002dhp_002dhpux10"></a>hppa*-hp-hpux10</h3>
400 1.1 mrg
401 1.1 mrg <p>For hpux10.20, we <em>highly</em> recommend you pick up the latest sed patch
402 1.1 mrg <code>PHCO_19798</code> from HP.
403 1.1 mrg
404 1.1 mrg <p>The C++ ABI has changed incompatibly in GCC 4.0. COMDAT subspaces are
405 1.1 mrg used for one-only code and data. This resolves many of the previous
406 1.1 mrg problems in using C++ on this target. However, the ABI is not compatible
407 1.1 mrg with the one implemented under HP-UX 11 using secondary definitions.
408 1.1 mrg
409 1.1 mrg <p><hr />
410 1.1 mrg
411 1.1 mrg <h3 class="heading"><a name="TOC13"></a><a name="hppa_002dhp_002dhpux11"></a>hppa*-hp-hpux11</h3>
412 1.1 mrg
413 1.1 mrg <p>GCC 3.0 and up support HP-UX 11. GCC 2.95.x is not supported and cannot
414 1.1 mrg be used to compile GCC 3.0 and up.
415 1.1 mrg
416 1.1 mrg <p>The libffi and libjava libraries haven't been ported to 64-bit HP-UX and don't build.
417 1.1 mrg
418 1.1 mrg <p>Refer to <a href="binaries.html">binaries</a> for information about obtaining
419 1.1 mrg precompiled GCC binaries for HP-UX. Precompiled binaries must be obtained
420 1.1 mrg to build the Ada language as it can't be bootstrapped using C. Ada is
421 1.1 mrg only available for the 32-bit PA-RISC runtime.
422 1.1 mrg
423 1.1 mrg <p>Starting with GCC 3.4 an ISO C compiler is required to bootstrap. The
424 1.1 mrg bundled compiler supports only traditional C; you will need either HP's
425 1.1 mrg unbundled compiler, or a binary distribution of GCC.
426 1.1 mrg
427 1.1 mrg <p>It is possible to build GCC 3.3 starting with the bundled HP compiler,
428 1.1 mrg but the process requires several steps. GCC 3.3 can then be used to
429 1.1 mrg build later versions. The fastjar program contains ISO C code and
430 1.1 mrg can't be built with the HP bundled compiler. This problem can be
431 1.1 mrg avoided by not building the Java language. For example, use the
432 1.1 mrg <samp><span class="option">--enable-languages="c,c++,f77,objc"</span></samp> option in your configure
433 1.1 mrg command.
434 1.1 mrg
435 1.1 mrg <p>There are several possible approaches to building the distribution.
436 1.1 mrg Binutils can be built first using the HP tools. Then, the GCC
437 1.1 mrg distribution can be built. The second approach is to build GCC
438 1.1 mrg first using the HP tools, then build binutils, then rebuild GCC.
439 1.1 mrg There have been problems with various binary distributions, so it
440 1.1 mrg is best not to start from a binary distribution.
441 1.1 mrg
442 1.1 mrg <p>On 64-bit capable systems, there are two distinct targets. Different
443 1.1 mrg installation prefixes must be used if both are to be installed on
444 1.1 mrg the same system. The ‘<samp><span class="samp">hppa[1-2]*-hp-hpux11*</span></samp>’ target generates code
445 1.1 mrg for the 32-bit PA-RISC runtime architecture and uses the HP linker.
446 1.1 mrg The ‘<samp><span class="samp">hppa64-hp-hpux11*</span></samp>’ target generates 64-bit code for the
447 1.1 mrg PA-RISC 2.0 architecture.
448 1.1 mrg
449 1.1 mrg <p>The script config.guess now selects the target type based on the compiler
450 1.1 mrg detected during configuration. You must define <samp><span class="env">PATH</span></samp> or <samp><span class="env">CC</span></samp> so
451 1.1 mrg that configure finds an appropriate compiler for the initial bootstrap.
452 1.1 mrg When <samp><span class="env">CC</span></samp> is used, the definition should contain the options that are
453 1.1 mrg needed whenever <samp><span class="env">CC</span></samp> is used.
454 1.1 mrg
455 1.1 mrg <p>Specifically, options that determine the runtime architecture must be
456 1.1 mrg in <samp><span class="env">CC</span></samp> to correctly select the target for the build. It is also
457 1.1 mrg convenient to place many other compiler options in <samp><span class="env">CC</span></samp>. For example,
458 1.1 mrg <samp><span class="env">CC="cc -Ac +DA2.0W -Wp,-H16376 -D_CLASSIC_TYPES -D_HPUX_SOURCE"</span></samp>
459 1.1 mrg can be used to bootstrap the GCC 3.3 branch with the HP compiler in
460 1.1 mrg 64-bit K&R/bundled mode. The <samp><span class="option">+DA2.0W</span></samp> option will result in
461 1.1 mrg the automatic selection of the ‘<samp><span class="samp">hppa64-hp-hpux11*</span></samp>’ target. The
462 1.1 mrg macro definition table of cpp needs to be increased for a successful
463 1.1 mrg build with the HP compiler. _CLASSIC_TYPES and _HPUX_SOURCE need to
464 1.1 mrg be defined when building with the bundled compiler, or when using the
465 1.1 mrg <samp><span class="option">-Ac</span></samp> option. These defines aren't necessary with <samp><span class="option">-Ae</span></samp>.
466 1.1 mrg
467 1.1 mrg <p>It is best to explicitly configure the ‘<samp><span class="samp">hppa64-hp-hpux11*</span></samp>’ target
468 1.1 mrg with the <samp><span class="option">--with-ld=...</span></samp> option. This overrides the standard
469 1.1 mrg search for ld. The two linkers supported on this target require different
470 1.1 mrg commands. The default linker is determined during configuration. As a
471 1.1 mrg result, it's not possible to switch linkers in the middle of a GCC build.
472 1.1 mrg This has been reported to sometimes occur in unified builds of binutils
473 1.1 mrg and GCC.
474 1.1 mrg
475 1.1 mrg <p>A recent linker patch must be installed for the correct operation of
476 1.1 mrg GCC 3.3 and later. <code>PHSS_26559</code> and <code>PHSS_24304</code> are the
477 1.1 mrg oldest linker patches that are known to work. They are for HP-UX
478 1.1 mrg 11.00 and 11.11, respectively. <code>PHSS_24303</code>, the companion to
479 1.1 mrg <code>PHSS_24304</code>, might be usable but it hasn't been tested. These
480 1.1 mrg patches have been superseded. Consult the HP patch database to obtain
481 1.1 mrg the currently recommended linker patch for your system.
482 1.1 mrg
483 1.1 mrg <p>The patches are necessary for the support of weak symbols on the
484 1.1 mrg 32-bit port, and for the running of initializers and finalizers. Weak
485 1.1 mrg symbols are implemented using SOM secondary definition symbols. Prior
486 1.1 mrg to HP-UX 11, there are bugs in the linker support for secondary symbols.
487 1.1 mrg The patches correct a problem of linker core dumps creating shared
488 1.1 mrg libraries containing secondary symbols, as well as various other
489 1.1 mrg linking issues involving secondary symbols.
490 1.1 mrg
491 1.1 mrg <p>GCC 3.3 uses the ELF DT_INIT_ARRAY and DT_FINI_ARRAY capabilities to
492 1.1 mrg run initializers and finalizers on the 64-bit port. The 32-bit port
493 1.1 mrg uses the linker <samp><span class="option">+init</span></samp> and <samp><span class="option">+fini</span></samp> options for the same
494 1.1 mrg purpose. The patches correct various problems with the +init/+fini
495 1.1 mrg options, including program core dumps. Binutils 2.14 corrects a
496 1.1 mrg problem on the 64-bit port resulting from HP's non-standard use of
497 1.1 mrg the .init and .fini sections for array initializers and finalizers.
498 1.1 mrg
499 1.1 mrg <p>Although the HP and GNU linkers are both supported for the
500 1.1 mrg ‘<samp><span class="samp">hppa64-hp-hpux11*</span></samp>’ target, it is strongly recommended that the
501 1.1 mrg HP linker be used for link editing on this target.
502 1.1 mrg
503 1.1 mrg <p>At this time, the GNU linker does not support the creation of long
504 1.1 mrg branch stubs. As a result, it can't successfully link binaries
505 1.1 mrg containing branch offsets larger than 8 megabytes. In addition,
506 1.1 mrg there are problems linking shared libraries, linking executables
507 1.1 mrg with <samp><span class="option">-static</span></samp>, and with dwarf2 unwind and exception support.
508 1.1 mrg It also doesn't provide stubs for internal calls to global functions
509 1.1 mrg in shared libraries, so these calls can't be overloaded.
510 1.1 mrg
511 1.1 mrg <p>The HP dynamic loader does not support GNU symbol versioning, so symbol
512 1.1 mrg versioning is not supported. It may be necessary to disable symbol
513 1.1 mrg versioning with <samp><span class="option">--disable-symvers</span></samp> when using GNU ld.
514 1.1 mrg
515 1.1 mrg <p>POSIX threads are the default. The optional DCE thread library is not
516 1.1 mrg supported, so <samp><span class="option">--enable-threads=dce</span></samp> does not work.
517 1.1 mrg
518 1.1 mrg <p><hr />
519 1.1 mrg
520 1.1 mrg <h3 class="heading"><a name="TOC14"></a><a name="x_002dx_002dlinux_002dgnu"></a>*-*-linux-gnu</h3>
521 1.1 mrg
522 1.1 mrg <p>Versions of libstdc++-v3 starting with 3.2.1 require bug fixes present
523 1.1 mrg in glibc 2.2.5 and later. More information is available in the
524 1.1 mrg libstdc++-v3 documentation.
525 1.1 mrg
526 1.1 mrg <p><hr />
527 1.1 mrg
528 1.1 mrg <h3 class="heading"><a name="TOC15"></a><a name="ix86_002dx_002dlinux"></a>i?86-*-linux*</h3>
529 1.1 mrg
530 1.1 mrg <p>As of GCC 3.3, binutils 2.13.1 or later is required for this platform.
531 1.1 mrg See <a href="http://gcc.gnu.org/PR10877">bug 10877</a> for more information.
532 1.1 mrg
533 1.1 mrg <p>If you receive Signal 11 errors when building on GNU/Linux, then it is
534 1.1 mrg possible you have a hardware problem. Further information on this can be
535 1.1 mrg found on <a href="http://www.bitwizard.nl/sig11/">www.bitwizard.nl</a>.
536 1.1 mrg
537 1.1 mrg <p><hr />
538 1.1 mrg
539 1.1 mrg <h3 class="heading"><a name="TOC16"></a><a name="ix86_002dx_002dsolaris289"></a>i?86-*-solaris2.[89]</h3>
540 1.1 mrg
541 1.1 mrg <p>The Sun assembler in Solaris 8 and 9 has several bugs and limitations.
542 1.1 mrg While GCC works around them, several features are missing, so it is
543 1.1 mrg <!-- FIXME: which ones? -->
544 1.1 mrg recommended to use the GNU assembler instead. There is no bundled
545 1.1 mrg version, but the current version, from GNU binutils 2.20.1, is known to
546 1.1 mrg work.
547 1.1 mrg
548 1.1 mrg <p>Solaris~2/x86 doesn't support the execution of SSE/SSE2 instructions
549 1.1 mrg before Solaris~9 4/04, even if the CPU supports them. Programs will
550 1.1 mrg receive <code>SIGILL</code> if they try. The fix is available both in
551 1.1 mrg Solaris~9 Update~6 and kernel patch 112234-12 or newer. There is no
552 1.1 mrg corresponding patch for Solaris 8. To avoid this problem,
553 1.1 mrg <samp><span class="option">-march</span></samp> defaults to ‘<samp><span class="samp">pentiumpro</span></samp>’ on Solaris 8 and 9. If
554 1.1 mrg you have the patch installed, you can configure GCC with an appropriate
555 1.1 mrg <samp><span class="option">--with-arch</span></samp> option, but need GNU <samp><span class="command">as</span></samp> for SSE2 support.
556 1.1 mrg
557 1.1 mrg <p><hr />
558 1.1 mrg
559 1.1 mrg <h3 class="heading"><a name="TOC17"></a><a name="ix86_002dx_002dsolaris210"></a>i?86-*-solaris2.10</h3>
560 1.1 mrg
561 1.1 mrg <p>Use this for Solaris 10 or later on x86 and x86-64 systems. This
562 1.1 mrg configuration is supported by GCC 4.0 and later versions only. Unlike
563 1.1 mrg ‘<samp><span class="samp">sparcv9-sun-solaris2*</span></samp>’, there is no corresponding 64-bit
564 1.1 mrg configuration like ‘<samp><span class="samp">amd64-*-solaris2*</span></samp>’ or ‘<samp><span class="samp">x86_64-*-solaris2*</span></samp>’.
565 1.1 mrg <!-- FIXME: will there ever be? -->
566 1.1 mrg
567 1.1 mrg <p>It is recommended that you configure GCC to use the GNU assembler, in
568 1.1 mrg <samp><span class="file">/usr/sfw/bin/gas</span></samp>. The versions included in Solaris 10, from GNU
569 1.1 mrg binutils 2.15, and Solaris 11, from GNU binutils 2.19, work fine,
570 1.1 mrg although the current version, from GNU binutils
571 1.1 mrg 2.20.1, is known to work, too. Recent versions of the Sun assembler in
572 1.1 mrg <samp><span class="file">/usr/ccs/bin/as</span></samp> work almost as well, though.
573 1.1 mrg <!-- FIXME: as patch requirements? -->
574 1.1 mrg
575 1.1 mrg <p>For linking, the Sun linker, is preferred. If you want to use the GNU
576 1.1 mrg linker instead, which is available in <samp><span class="file">/usr/sfw/bin/gld</span></samp>, note that
577 1.1 mrg due to a packaging bug the version in Solaris 10, from GNU binutils
578 1.1 mrg 2.15, cannot be used, while the version in Solaris 11, from GNU binutils
579 1.1 mrg 2.19, works, as does the latest version, from GNU binutils 2.20.1.
580 1.1 mrg
581 1.1 mrg <p>To use GNU <samp><span class="command">as</span></samp>, configure with the options
582 1.1 mrg <samp><span class="option">--with-gnu-as --with-as=/usr/sfw/bin/gas</span></samp>. It may be necessary
583 1.1 mrg to configure with <samp><span class="option">--without-gnu-ld --with-ld=/usr/ccs/bin/ld</span></samp> to
584 1.1 mrg guarantee use of Sun <samp><span class="command">ld</span></samp>.
585 1.1 mrg <!-- FIXME: why -without-gnu-ld -with-ld? -->
586 1.1 mrg
587 1.1 mrg <p><hr />
588 1.1 mrg
589 1.1 mrg <h3 class="heading"><a name="TOC18"></a><a name="ia64_002dx_002dlinux"></a>ia64-*-linux</h3>
590 1.1 mrg
591 1.1 mrg <p>IA-64 processor (also known as IPF, or Itanium Processor Family)
592 1.1 mrg running GNU/Linux.
593 1.1 mrg
594 1.1 mrg <p>If you are using the installed system libunwind library with
595 1.1 mrg <samp><span class="option">--with-system-libunwind</span></samp>, then you must use libunwind 0.98 or
596 1.1 mrg later.
597 1.1 mrg
598 1.1 mrg <p>None of the following versions of GCC has an ABI that is compatible
599 1.1 mrg with any of the other versions in this list, with the exception that
600 1.1 mrg Red Hat 2.96 and Trillian 000171 are compatible with each other:
601 1.1 mrg 3.1, 3.0.2, 3.0.1, 3.0, Red Hat 2.96, and Trillian 000717.
602 1.1 mrg This primarily affects C++ programs and programs that create shared libraries.
603 1.1 mrg GCC 3.1 or later is recommended for compiling linux, the kernel.
604 1.1 mrg As of version 3.1 GCC is believed to be fully ABI compliant, and hence no
605 1.1 mrg more major ABI changes are expected.
606 1.1 mrg
607 1.1 mrg <p><hr />
608 1.1 mrg
609 1.1 mrg <h3 class="heading"><a name="TOC19"></a><a name="ia64_002dx_002dhpux"></a>ia64-*-hpux*</h3>
610 1.1 mrg
611 1.1 mrg <p>Building GCC on this target requires the GNU Assembler. The bundled HP
612 1.1 mrg assembler will not work. To prevent GCC from using the wrong assembler,
613 1.1 mrg the option <samp><span class="option">--with-gnu-as</span></samp> may be necessary.
614 1.1 mrg
615 1.1 mrg <p>The GCC libunwind library has not been ported to HPUX. This means that for
616 1.1 mrg GCC versions 3.2.3 and earlier, <samp><span class="option">--enable-libunwind-exceptions</span></samp>
617 1.1 mrg is required to build GCC. For GCC 3.3 and later, this is the default.
618 1.1 mrg For gcc 3.4.3 and later, <samp><span class="option">--enable-libunwind-exceptions</span></samp> is
619 1.1 mrg removed and the system libunwind library will always be used.
620 1.1 mrg
621 1.1 mrg <p><hr />
622 1.1 mrg <!-- rs6000-ibm-aix*, powerpc-ibm-aix* -->
623 1.1 mrg
624 1.1 mrg <h3 class="heading"><a name="TOC20"></a><a name="x_002dibm_002daix"></a>*-ibm-aix*</h3>
625 1.1 mrg
626 1.1 mrg <p>Support for AIX version 3 and older was discontinued in GCC 3.4.
627 1.1 mrg Support for AIX version 4.2 and older was discontinued in GCC 4.5.
628 1.1 mrg
629 1.1 mrg <p>“out of memory” bootstrap failures may indicate a problem with
630 1.1 mrg process resource limits (ulimit). Hard limits are configured in the
631 1.1 mrg <samp><span class="file">/etc/security/limits</span></samp> system configuration file.
632 1.1 mrg
633 1.1 mrg <p>GCC can bootstrap with recent versions of IBM XLC, but bootstrapping
634 1.1 mrg with an earlier release of GCC is recommended. Bootstrapping with XLC
635 1.1 mrg requires a larger data segment, which can be enabled through the
636 1.1 mrg <var>LDR_CNTRL</var> environment variable, e.g.,
637 1.1 mrg
638 1.1 mrg <pre class="smallexample"> % LDR_CNTRL=MAXDATA=0x50000000
639 1.1 mrg % export LDR_CNTRL
640 1.1 mrg </pre>
641 1.1 mrg <p>One can start with a pre-compiled version of GCC to build from
642 1.1 mrg sources. One may delete GCC's “fixed” header files when starting
643 1.1 mrg with a version of GCC built for an earlier release of AIX.
644 1.1 mrg
645 1.1 mrg <p>To speed up the configuration phases of bootstrapping and installing GCC,
646 1.1 mrg one may use GNU Bash instead of AIX <samp><span class="command">/bin/sh</span></samp>, e.g.,
647 1.1 mrg
648 1.1 mrg <pre class="smallexample"> % CONFIG_SHELL=/opt/freeware/bin/bash
649 1.1 mrg % export CONFIG_SHELL
650 1.1 mrg </pre>
651 1.1 mrg <p>and then proceed as described in <a href="build.html">the build instructions</a>, where we strongly recommend specifying an absolute path
652 1.1 mrg to invoke <var>srcdir</var>/configure.
653 1.1 mrg
654 1.1 mrg <p>Because GCC on AIX is built as a 32-bit executable by default,
655 1.1 mrg (although it can generate 64-bit programs) the GMP and MPFR libraries
656 1.1 mrg required by gfortran must be 32-bit libraries. Building GMP and MPFR
657 1.1 mrg as static archive libraries works better than shared libraries.
658 1.1 mrg
659 1.1 mrg <p>Errors involving <code>alloca</code> when building GCC generally are due
660 1.1 mrg to an incorrect definition of <code>CC</code> in the Makefile or mixing files
661 1.1 mrg compiled with the native C compiler and GCC. During the stage1 phase of
662 1.1 mrg the build, the native AIX compiler <strong>must</strong> be invoked as <samp><span class="command">cc</span></samp>
663 1.1 mrg (not <samp><span class="command">xlc</span></samp>). Once <samp><span class="command">configure</span></samp> has been informed of
664 1.1 mrg <samp><span class="command">xlc</span></samp>, one needs to use ‘<samp><span class="samp">make distclean</span></samp>’ to remove the
665 1.1 mrg configure cache files and ensure that <samp><span class="env">CC</span></samp> environment variable
666 1.1 mrg does not provide a definition that will confuse <samp><span class="command">configure</span></samp>.
667 1.1 mrg If this error occurs during stage2 or later, then the problem most likely
668 1.1 mrg is the version of Make (see above).
669 1.1 mrg
670 1.1 mrg <p>The native <samp><span class="command">as</span></samp> and <samp><span class="command">ld</span></samp> are recommended for bootstrapping
671 1.1 mrg on AIX. The GNU Assembler, GNU Linker, and GNU Binutils version 2.20
672 1.1 mrg is required to bootstrap on AIX 5. The native AIX tools do
673 1.1 mrg interoperate with GCC.
674 1.1 mrg
675 1.1 mrg <p>Building <samp><span class="file">libstdc++.a</span></samp> requires a fix for an AIX Assembler bug
676 1.1 mrg APAR IY26685 (AIX 4.3) or APAR IY25528 (AIX 5.1). It also requires a
677 1.1 mrg fix for another AIX Assembler bug and a co-dependent AIX Archiver fix
678 1.1 mrg referenced as APAR IY53606 (AIX 5.2) or as APAR IY54774 (AIX 5.1)
679 1.1 mrg
680 1.1 mrg <p>‘<samp><span class="samp">libstdc++</span></samp>’ in GCC 3.4 increments the major version number of the
681 1.1 mrg shared object and GCC installation places the <samp><span class="file">libstdc++.a</span></samp>
682 1.1 mrg shared library in a common location which will overwrite the and GCC
683 1.1 mrg 3.3 version of the shared library. Applications either need to be
684 1.1 mrg re-linked against the new shared library or the GCC 3.1 and GCC 3.3
685 1.1 mrg versions of the ‘<samp><span class="samp">libstdc++</span></samp>’ shared object needs to be available
686 1.1 mrg to the AIX runtime loader. The GCC 3.1 ‘<samp><span class="samp">libstdc++.so.4</span></samp>’, if
687 1.1 mrg present, and GCC 3.3 ‘<samp><span class="samp">libstdc++.so.5</span></samp>’ shared objects can be
688 1.1 mrg installed for runtime dynamic loading using the following steps to set
689 1.1 mrg the ‘<samp><span class="samp">F_LOADONLY</span></samp>’ flag in the shared object for <em>each</em>
690 1.1 mrg multilib <samp><span class="file">libstdc++.a</span></samp> installed:
691 1.1 mrg
692 1.1 mrg <p>Extract the shared objects from the currently installed
693 1.1 mrg <samp><span class="file">libstdc++.a</span></samp> archive:
694 1.1 mrg <pre class="smallexample"> % ar -x libstdc++.a libstdc++.so.4 libstdc++.so.5
695 1.1 mrg </pre>
696 1.1 mrg <p>Enable the ‘<samp><span class="samp">F_LOADONLY</span></samp>’ flag so that the shared object will be
697 1.1 mrg available for runtime dynamic loading, but not linking:
698 1.1 mrg <pre class="smallexample"> % strip -e libstdc++.so.4 libstdc++.so.5
699 1.1 mrg </pre>
700 1.1 mrg <p>Archive the runtime-only shared object in the GCC 3.4
701 1.1 mrg <samp><span class="file">libstdc++.a</span></samp> archive:
702 1.1 mrg <pre class="smallexample"> % ar -q libstdc++.a libstdc++.so.4 libstdc++.so.5
703 1.1 mrg </pre>
704 1.1 mrg <p>Linking executables and shared libraries may produce warnings of
705 1.1 mrg duplicate symbols. The assembly files generated by GCC for AIX always
706 1.1 mrg have included multiple symbol definitions for certain global variable
707 1.1 mrg and function declarations in the original program. The warnings should
708 1.1 mrg not prevent the linker from producing a correct library or runnable
709 1.1 mrg executable.
710 1.1 mrg
711 1.1 mrg <p>AIX 4.3 utilizes a “large format” archive to support both 32-bit and
712 1.1 mrg 64-bit object modules. The routines provided in AIX 4.3.0 and AIX 4.3.1
713 1.1 mrg to parse archive libraries did not handle the new format correctly.
714 1.1 mrg These routines are used by GCC and result in error messages during
715 1.1 mrg linking such as “not a COFF file”. The version of the routines shipped
716 1.1 mrg with AIX 4.3.1 should work for a 32-bit environment. The <samp><span class="option">-g</span></samp>
717 1.1 mrg option of the archive command may be used to create archives of 32-bit
718 1.1 mrg objects using the original “small format”. A correct version of the
719 1.1 mrg routines is shipped with AIX 4.3.2 and above.
720 1.1 mrg
721 1.1 mrg <p>Some versions of the AIX binder (linker) can fail with a relocation
722 1.1 mrg overflow severe error when the <samp><span class="option">-bbigtoc</span></samp> option is used to link
723 1.1 mrg GCC-produced object files into an executable that overflows the TOC. A fix
724 1.1 mrg for APAR IX75823 (OVERFLOW DURING LINK WHEN USING GCC AND -BBIGTOC) is
725 1.1 mrg available from IBM Customer Support and from its
726 1.1 mrg <a href="http://techsupport.services.ibm.com/">techsupport.services.ibm.com</a>
727 1.1 mrg website as PTF U455193.
728 1.1 mrg
729 1.1 mrg <p>The AIX 4.3.2.1 linker (bos.rte.bind_cmds Level 4.3.2.1) will dump core
730 1.1 mrg with a segmentation fault when invoked by any version of GCC. A fix for
731 1.1 mrg APAR IX87327 is available from IBM Customer Support and from its
732 1.1 mrg <a href="http://techsupport.services.ibm.com/">techsupport.services.ibm.com</a>
733 1.1 mrg website as PTF U461879. This fix is incorporated in AIX 4.3.3 and above.
734 1.1 mrg
735 1.1 mrg <p>The initial assembler shipped with AIX 4.3.0 generates incorrect object
736 1.1 mrg files. A fix for APAR IX74254 (64BIT DISASSEMBLED OUTPUT FROM COMPILER FAILS
737 1.1 mrg TO ASSEMBLE/BIND) is available from IBM Customer Support and from its
738 1.1 mrg <a href="http://techsupport.services.ibm.com/">techsupport.services.ibm.com</a>
739 1.1 mrg website as PTF U453956. This fix is incorporated in AIX 4.3.1 and above.
740 1.1 mrg
741 1.1 mrg <p>AIX provides National Language Support (NLS). Compilers and assemblers
742 1.1 mrg use NLS to support locale-specific representations of various data
743 1.1 mrg formats including floating-point numbers (e.g., ‘<samp><span class="samp">.</span></samp>’ vs ‘<samp><span class="samp">,</span></samp>’ for
744 1.1 mrg separating decimal fractions). There have been problems reported where
745 1.1 mrg GCC does not produce the same floating-point formats that the assembler
746 1.1 mrg expects. If one encounters this problem, set the <samp><span class="env">LANG</span></samp>
747 1.1 mrg environment variable to ‘<samp><span class="samp">C</span></samp>’ or ‘<samp><span class="samp">En_US</span></samp>’.
748 1.1 mrg
749 1.1 mrg <p>A default can be specified with the <samp><span class="option">-mcpu=</span><var>cpu_type</var></samp>
750 1.1 mrg switch and using the configure option <samp><span class="option">--with-cpu-</span><var>cpu_type</var></samp>.
751 1.1 mrg
752 1.1 mrg <p><hr />
753 1.1 mrg
754 1.1 mrg <h3 class="heading"><a name="TOC21"></a><a name="iq2000_002dx_002delf"></a>iq2000-*-elf</h3>
755 1.1 mrg
756 1.1 mrg <p>Vitesse IQ2000 processors. These are used in embedded
757 1.1 mrg applications. There are no standard Unix configurations.
758 1.1 mrg
759 1.1 mrg <p><hr />
760 1.1 mrg
761 1.1 mrg <h3 class="heading"><a name="TOC22"></a><a name="lm32_002dx_002delf"></a>lm32-*-elf</h3>
762 1.1 mrg
763 1.1 mrg <p>Lattice Mico32 processor.
764 1.1 mrg This configuration is intended for embedded systems.
765 1.1 mrg
766 1.1 mrg <p><hr />
767 1.1 mrg
768 1.1 mrg <h3 class="heading"><a name="TOC23"></a><a name="lm32_002dx_002duclinux"></a>lm32-*-uclinux</h3>
769 1.1 mrg
770 1.1 mrg <p>Lattice Mico32 processor.
771 1.1 mrg This configuration is intended for embedded systems running uClinux.
772 1.1 mrg
773 1.1 mrg <p><hr />
774 1.1 mrg
775 1.1 mrg <h3 class="heading"><a name="TOC24"></a><a name="m32c_002dx_002delf"></a>m32c-*-elf</h3>
776 1.1 mrg
777 1.1 mrg <p>Renesas M32C processor.
778 1.1 mrg This configuration is intended for embedded systems.
779 1.1 mrg
780 1.1 mrg <p><hr />
781 1.1 mrg
782 1.1 mrg <h3 class="heading"><a name="TOC25"></a><a name="m32r_002dx_002delf"></a>m32r-*-elf</h3>
783 1.1 mrg
784 1.1 mrg <p>Renesas M32R processor.
785 1.1 mrg This configuration is intended for embedded systems.
786 1.1 mrg
787 1.1 mrg <p><hr />
788 1.1 mrg
789 1.1 mrg <h3 class="heading"><a name="TOC26"></a><a name="m6811_002delf"></a>m6811-elf</h3>
790 1.1 mrg
791 1.1 mrg <p>Motorola 68HC11 family micro controllers. These are used in embedded
792 1.1 mrg applications. There are no standard Unix configurations.
793 1.1 mrg
794 1.1 mrg <p><hr />
795 1.1 mrg
796 1.1 mrg <h3 class="heading"><a name="TOC27"></a><a name="m6812_002delf"></a>m6812-elf</h3>
797 1.1 mrg
798 1.1 mrg <p>Motorola 68HC12 family micro controllers. These are used in embedded
799 1.1 mrg applications. There are no standard Unix configurations.
800 1.1 mrg
801 1.1 mrg <p><hr />
802 1.1 mrg
803 1.1 mrg <h3 class="heading"><a name="TOC28"></a><a name="m68k_002dx_002dx"></a>m68k-*-*</h3>
804 1.1 mrg
805 1.1 mrg <p>By default,
806 1.1 mrg ‘<samp><span class="samp">m68k-*-elf*</span></samp>’, ‘<samp><span class="samp">m68k-*-rtems</span></samp>’, ‘<samp><span class="samp">m68k-*-uclinux</span></samp>’ and
807 1.1 mrg ‘<samp><span class="samp">m68k-*-linux</span></samp>’
808 1.1 mrg build libraries for both M680x0 and ColdFire processors. If you only
809 1.1 mrg need the M680x0 libraries, you can omit the ColdFire ones by passing
810 1.1 mrg <samp><span class="option">--with-arch=m68k</span></samp> to <samp><span class="command">configure</span></samp>. Alternatively, you
811 1.1 mrg can omit the M680x0 libraries by passing <samp><span class="option">--with-arch=cf</span></samp> to
812 1.1 mrg <samp><span class="command">configure</span></samp>. These targets default to 5206 or 5475 code as
813 1.1 mrg appropriate for the target system when
814 1.1 mrg configured with <samp><span class="option">--with-arch=cf</span></samp> and 68020 code otherwise.
815 1.1 mrg
816 1.1 mrg <p>The ‘<samp><span class="samp">m68k-*-netbsd</span></samp>’ and
817 1.1 mrg ‘<samp><span class="samp">m68k-*-openbsd</span></samp>’ targets also support the <samp><span class="option">--with-arch</span></samp>
818 1.1 mrg option. They will generate ColdFire CFV4e code when configured with
819 1.1 mrg <samp><span class="option">--with-arch=cf</span></samp> and 68020 code otherwise.
820 1.1 mrg
821 1.1 mrg <p>You can override the default processors listed above by configuring
822 1.1 mrg with <samp><span class="option">--with-cpu=</span><var>target</var></samp>. This <var>target</var> can either
823 1.1 mrg be a <samp><span class="option">-mcpu</span></samp> argument or one of the following values:
824 1.1 mrg ‘<samp><span class="samp">m68000</span></samp>’, ‘<samp><span class="samp">m68010</span></samp>’, ‘<samp><span class="samp">m68020</span></samp>’, ‘<samp><span class="samp">m68030</span></samp>’,
825 1.1 mrg ‘<samp><span class="samp">m68040</span></samp>’, ‘<samp><span class="samp">m68060</span></samp>’, ‘<samp><span class="samp">m68020-40</span></samp>’ and ‘<samp><span class="samp">m68020-60</span></samp>’.
826 1.1 mrg
827 1.1 mrg <p><hr />
828 1.1 mrg
829 1.1 mrg <h3 class="heading"><a name="TOC29"></a><a name="m68k_002dx_002duclinux"></a>m68k-*-uclinux</h3>
830 1.1 mrg
831 1.1 mrg <p>GCC 4.3 changed the uClinux configuration so that it uses the
832 1.1 mrg ‘<samp><span class="samp">m68k-linux-gnu</span></samp>’ ABI rather than the ‘<samp><span class="samp">m68k-elf</span></samp>’ ABI.
833 1.1 mrg It also added improved support for C++ and flat shared libraries,
834 1.1 mrg both of which were ABI changes. However, you can still use the
835 1.1 mrg original ABI by configuring for ‘<samp><span class="samp">m68k-uclinuxoldabi</span></samp>’ or
836 1.1 mrg ‘<samp><span class="samp">m68k-</span><var>vendor</var><span class="samp">-uclinuxoldabi</span></samp>’.
837 1.1 mrg
838 1.1 mrg <p><hr />
839 1.1 mrg
840 1.1 mrg <h3 class="heading"><a name="TOC30"></a><a name="mep_002dx_002delf"></a>mep-*-elf</h3>
841 1.1 mrg
842 1.1 mrg <p>Toshiba Media embedded Processor.
843 1.1 mrg This configuration is intended for embedded systems.
844 1.1 mrg
845 1.1 mrg <p><hr />
846 1.1 mrg
847 1.1 mrg <h3 class="heading"><a name="TOC31"></a><a name="mips_002dx_002dx"></a>mips-*-*</h3>
848 1.1 mrg
849 1.1 mrg <p>If on a MIPS system you get an error message saying “does not have gp
850 1.1 mrg sections for all it's [sic] sectons [sic]”, don't worry about it. This
851 1.1 mrg happens whenever you use GAS with the MIPS linker, but there is not
852 1.1 mrg really anything wrong, and it is okay to use the output file. You can
853 1.1 mrg stop such warnings by installing the GNU linker.
854 1.1 mrg
855 1.1 mrg <p>It would be nice to extend GAS to produce the gp tables, but they are
856 1.1 mrg optional, and there should not be a warning about their absence.
857 1.1 mrg
858 1.1 mrg <p>The libstdc++ atomic locking routines for MIPS targets requires MIPS II
859 1.1 mrg and later. A patch went in just after the GCC 3.3 release to
860 1.1 mrg make ‘<samp><span class="samp">mips*-*-*</span></samp>’ use the generic implementation instead. You can also
861 1.1 mrg configure for ‘<samp><span class="samp">mipsel-elf</span></samp>’ as a workaround. The
862 1.1 mrg ‘<samp><span class="samp">mips*-*-linux*</span></samp>’ target continues to use the MIPS II routines. More
863 1.1 mrg work on this is expected in future releases.
864 1.1 mrg
865 1.1 mrg <!-- If you make -with-llsc the default for another target, please also -->
866 1.1 mrg <!-- update the description of the -with-llsc option. -->
867 1.1 mrg <p>The built-in <code>__sync_*</code> functions are available on MIPS II and
868 1.1 mrg later systems and others that support the ‘<samp><span class="samp">ll</span></samp>’, ‘<samp><span class="samp">sc</span></samp>’ and
869 1.1 mrg ‘<samp><span class="samp">sync</span></samp>’ instructions. This can be overridden by passing
870 1.1 mrg <samp><span class="option">--with-llsc</span></samp> or <samp><span class="option">--without-llsc</span></samp> when configuring GCC.
871 1.1 mrg Since the Linux kernel emulates these instructions if they are
872 1.1 mrg missing, the default for ‘<samp><span class="samp">mips*-*-linux*</span></samp>’ targets is
873 1.1 mrg <samp><span class="option">--with-llsc</span></samp>. The <samp><span class="option">--with-llsc</span></samp> and
874 1.1 mrg <samp><span class="option">--without-llsc</span></samp> configure options may be overridden at compile
875 1.1 mrg time by passing the <samp><span class="option">-mllsc</span></samp> or <samp><span class="option">-mno-llsc</span></samp> options to
876 1.1 mrg the compiler.
877 1.1 mrg
878 1.1 mrg <p>MIPS systems check for division by zero (unless
879 1.1 mrg <samp><span class="option">-mno-check-zero-division</span></samp> is passed to the compiler) by
880 1.1 mrg generating either a conditional trap or a break instruction. Using
881 1.1 mrg trap results in smaller code, but is only supported on MIPS II and
882 1.1 mrg later. Also, some versions of the Linux kernel have a bug that
883 1.1 mrg prevents trap from generating the proper signal (<code>SIGFPE</code>). To enable
884 1.1 mrg the use of break, use the <samp><span class="option">--with-divide=breaks</span></samp>
885 1.1 mrg <samp><span class="command">configure</span></samp> option when configuring GCC. The default is to
886 1.1 mrg use traps on systems that support them.
887 1.1 mrg
888 1.1 mrg <p>Cross-compilers for the MIPS as target using the MIPS assembler
889 1.1 mrg currently do not work, because the auxiliary programs
890 1.1 mrg <samp><span class="file">mips-tdump.c</span></samp> and <samp><span class="file">mips-tfile.c</span></samp> can't be compiled on
891 1.1 mrg anything but a MIPS. It does work to cross compile for a MIPS
892 1.1 mrg if you use the GNU assembler and linker.
893 1.1 mrg
894 1.1 mrg <p>The assembler from GNU binutils 2.17 and earlier has a bug in the way
895 1.1 mrg it sorts relocations for REL targets (o32, o64, EABI). This can cause
896 1.1 mrg bad code to be generated for simple C++ programs. Also the linker
897 1.1 mrg from GNU binutils versions prior to 2.17 has a bug which causes the
898 1.1 mrg runtime linker stubs in very large programs, like <samp><span class="file">libgcj.so</span></samp>, to
899 1.1 mrg be incorrectly generated. GNU Binutils 2.18 and later (and snapshots
900 1.1 mrg made after Nov. 9, 2006) should be free from both of these problems.
901 1.1 mrg
902 1.1 mrg <p><hr />
903 1.1 mrg
904 1.1 mrg <h3 class="heading"><a name="TOC32"></a><a name="mips_002dsgi_002dirix5"></a>mips-sgi-irix5</h3>
905 1.1 mrg
906 1.1 mrg <p>Support for IRIX 5 has been obsoleted in GCC 4.5, but can still be
907 1.1 mrg enabled by configuring with <samp><span class="option">--enable-obsolete</span></samp>. Support will be
908 1.1 mrg removed in GCC 4.6.
909 1.1 mrg
910 1.1 mrg <p>In order to compile GCC on an SGI running IRIX 5, the ‘<samp><span class="samp">compiler_dev.hdr</span></samp>’
911 1.1 mrg subsystem must be installed from the IDO CD-ROM supplied by SGI.
912 1.1 mrg It is also available for download from
913 1.1 mrg <a href="http://freeware.sgi.com/ido.html">http://freeware.sgi.com/ido.html</a>.
914 1.1 mrg
915 1.1 mrg <p>If you use the MIPS C compiler to bootstrap, it may be necessary
916 1.1 mrg to increase its table size for switch statements with the
917 1.1 mrg <samp><span class="option">-Wf,-XNg1500</span></samp> option. If you use the <samp><span class="option">-O2</span></samp>
918 1.1 mrg optimization option, you also need to use <samp><span class="option">-Olimit 3000</span></samp>.
919 1.1 mrg <!-- FIXME: verify. -->
920 1.1 mrg
921 1.1 mrg <p>GCC must be configured to use GNU <samp><span class="command">as</span></samp>. The latest version, from GNU
922 1.1 mrg binutils 2.20.1, is known to work.
923 1.1 mrg
924 1.1 mrg <p>To enable debugging under IRIX 5, you must use GNU binutils 2.15 or
925 1.1 mrg later, and use the <samp><span class="option">--with-gnu-ld</span></samp> <samp><span class="command">configure</span></samp> option
926 1.1 mrg when configuring GCC.
927 1.1 mrg You need to use GNU <samp><span class="command">ar</span></samp> and <samp><span class="command">nm</span></samp>,
928 1.1 mrg also distributed with GNU binutils.
929 1.1 mrg <!-- FIXME: which parts of this are still true? -->
930 1.1 mrg
931 1.1 mrg <p>Configuring GCC with <samp><span class="command">/bin/sh</span></samp> is <em>extremely</em> slow and may
932 1.1 mrg even hang. This problem can be avoided by running <samp><span class="command">configure</span></samp>
933 1.1 mrg like this:
934 1.1 mrg
935 1.1 mrg <pre class="smallexample"> % CONFIG_SHELL=/usr/local/bin/bash
936 1.1 mrg % export CONFIG_SHELL
937 1.1 mrg % $CONFIG_SHELL <var>srcdir</var>/configure [<var>options</var>]
938 1.1 mrg </pre>
939 1.1 mrg <p class="noindent"><samp><span class="command">/bin/ksh</span></samp> doesn't work properly either.
940 1.1 mrg
941 1.1 mrg <p><hr />
942 1.1 mrg
943 1.1 mrg <h3 class="heading"><a name="TOC33"></a><a name="mips_002dsgi_002dirix6"></a>mips-sgi-irix6</h3>
944 1.1 mrg
945 1.1 mrg <p>Support for IRIX 6 releases before 6.5 has been obsoleted in GCC 4.5,
946 1.1 mrg but can still be enabled by configuring with <samp><span class="option">--enable-obsolete</span></samp>.
947 1.1 mrg Support will be removed in GCC 4.6, which will also disable support for
948 1.1 mrg the O32 ABI. It is <em>strongly</em> recommended to upgrade to at least
949 1.1 mrg IRIX 6.5.18. This release introduced full ISO C99 support, though for
950 1.1 mrg the N32 and N64 ABIs only.
951 1.1 mrg
952 1.1 mrg <p>To build and use GCC on IRIX 6, you need the IRIX Development Foundation
953 1.1 mrg (IDF) and IRIX Development Libraries (IDL). They are included with the
954 1.1 mrg IRIX 6.5 media and can be downloaded from
955 1.1 mrg <a href="http://freeware.sgi.com/idf_idl.html">http://freeware.sgi.com/idf_idl.html</a> for older IRIX 6 releases.
956 1.1 mrg
957 1.1 mrg <p>If you are using SGI's MIPSpro <samp><span class="command">cc</span></samp> as your bootstrap compiler, you must
958 1.1 mrg ensure that the N32 ABI is in use. To test this, compile a simple C
959 1.1 mrg file with <samp><span class="command">cc</span></samp> and then run <samp><span class="command">file</span></samp> on the
960 1.1 mrg resulting object file. The output should look like:
961 1.1 mrg
962 1.1 mrg <pre class="smallexample"> test.o: ELF N32 MSB ...
963 1.1 mrg </pre>
964 1.1 mrg <p class="noindent">If you see:
965 1.1 mrg
966 1.1 mrg <pre class="smallexample"> test.o: ELF 32-bit MSB ...
967 1.1 mrg </pre>
968 1.1 mrg <p class="noindent">or
969 1.1 mrg
970 1.1 mrg <pre class="smallexample"> test.o: ELF 64-bit MSB ...
971 1.1 mrg </pre>
972 1.1 mrg <p class="noindent">then your version of <samp><span class="command">cc</span></samp> uses the O32 or N64 ABI by default. You
973 1.1 mrg should set the environment variable <samp><span class="env">CC</span></samp> to ‘<samp><span class="samp">cc -n32</span></samp>’
974 1.1 mrg before configuring GCC.
975 1.1 mrg
976 1.1 mrg <p>If you want the resulting <samp><span class="command">gcc</span></samp> to run on old 32-bit systems
977 1.1 mrg with the MIPS R4400 CPU, you need to ensure that only code for the ‘<samp><span class="samp">mips3</span></samp>’
978 1.1 mrg instruction set architecture (ISA) is generated. While GCC 3.x does
979 1.1 mrg this correctly, both GCC 2.95 and SGI's MIPSpro <samp><span class="command">cc</span></samp> may change
980 1.1 mrg the ISA depending on the machine where GCC is built. Using one of them
981 1.1 mrg as the bootstrap compiler may result in ‘<samp><span class="samp">mips4</span></samp>’ code, which won't run at
982 1.1 mrg all on ‘<samp><span class="samp">mips3</span></samp>’-only systems. For the test program above, you should see:
983 1.1 mrg
984 1.1 mrg <pre class="smallexample"> test.o: ELF N32 MSB mips-3 ...
985 1.1 mrg </pre>
986 1.1 mrg <p class="noindent">If you get:
987 1.1 mrg
988 1.1 mrg <pre class="smallexample"> test.o: ELF N32 MSB mips-4 ...
989 1.1 mrg </pre>
990 1.1 mrg <p class="noindent">instead, you should set the environment variable <samp><span class="env">CC</span></samp> to ‘<samp><span class="samp">cc
991 1.1 mrg -n32 -mips3</span></samp>’ or ‘<samp><span class="samp">gcc -mips3</span></samp>’ respectively before configuring GCC.
992 1.1 mrg
993 1.1 mrg <p>MIPSpro C 7.4 may cause bootstrap failures, due to a bug when inlining
994 1.1 mrg <code>memcmp</code>. Either add <code>-U__INLINE_INTRINSICS</code> to the <samp><span class="env">CC</span></samp>
995 1.1 mrg environment variable as a workaround or upgrade to MIPSpro C 7.4.1m.
996 1.1 mrg
997 1.1 mrg <p>GCC on IRIX 6 is usually built to support the N32, O32 and N64 ABIs. If
998 1.1 mrg you build GCC on a system that doesn't have the N64 libraries installed
999 1.1 mrg or cannot run 64-bit binaries,
1000 1.1 mrg you need to configure with <samp><span class="option">--disable-multilib</span></samp> so GCC doesn't
1001 1.1 mrg try to use them. This will disable building the O32 libraries, too.
1002 1.1 mrg Look for <samp><span class="file">/usr/lib64/libc.so.1</span></samp> to see if you
1003 1.1 mrg have the 64-bit libraries installed.
1004 1.1 mrg
1005 1.1 mrg <p>GCC must be configured with GNU <samp><span class="command">as</span></samp>. The latest version, from GNU
1006 1.1 mrg binutils 2.20.1, is known to work. On the other hand, bootstrap fails
1007 1.1 mrg with GNU <samp><span class="command">ld</span></samp> at least since GNU binutils 2.17.
1008 1.1 mrg
1009 1.1 mrg <p>The <samp><span class="option">--enable-libgcj</span></samp>
1010 1.1 mrg option is disabled by default: IRIX 6 uses a very low default limit
1011 1.1 mrg (20480) for the command line length. Although <samp><span class="command">libtool</span></samp> contains a
1012 1.1 mrg workaround for this problem, at least the N64 ‘<samp><span class="samp">libgcj</span></samp>’ is known not
1013 1.1 mrg to build despite this, running into an internal error of the native
1014 1.1 mrg <samp><span class="command">ld</span></samp>. A sure fix is to increase this limit (‘<samp><span class="samp">ncargs</span></samp>’) to
1015 1.1 mrg its maximum of 262144 bytes. If you have root access, you can use the
1016 1.1 mrg <samp><span class="command">systune</span></samp> command to do this.
1017 1.1 mrg <!-- FIXME: does this work with current libtool? -->
1018 1.1 mrg
1019 1.1 mrg <p><code>wchar_t</code> support in ‘<samp><span class="samp">libstdc++</span></samp>’ is not available for old
1020 1.1 mrg IRIX 6.5.x releases, x < 19. The problem cannot be autodetected
1021 1.1 mrg and in order to build GCC for such targets you need to configure with
1022 1.1 mrg <samp><span class="option">--disable-wchar_t</span></samp>.
1023 1.1 mrg
1024 1.1 mrg <p><hr />
1025 1.1 mrg
1026 1.1 mrg <h3 class="heading"><a name="TOC34"></a><a name="moxie_002dx_002delf"></a>moxie-*-elf</h3>
1027 1.1 mrg
1028 1.1 mrg <p>The moxie processor. See <a href="http://moxielogic.org/">http://moxielogic.org/</a> for more
1029 1.1 mrg information about this processor.
1030 1.1 mrg
1031 1.1 mrg <p><hr />
1032 1.1 mrg
1033 1.1 mrg <h3 class="heading"><a name="TOC35"></a><a name="powerpc_002dx_002dx"></a>powerpc-*-*</h3>
1034 1.1 mrg
1035 1.1 mrg <p>You can specify a default version for the <samp><span class="option">-mcpu=</span><var>cpu_type</var></samp>
1036 1.1 mrg switch by using the configure option <samp><span class="option">--with-cpu-</span><var>cpu_type</var></samp>.
1037 1.1 mrg
1038 1.1 mrg <p>You will need
1039 1.1 mrg <a href="ftp://ftp.kernel.org/pub/linux/devel/binutils">binutils 2.15</a>
1040 1.1 mrg or newer for a working GCC.
1041 1.1 mrg
1042 1.1 mrg <p><hr />
1043 1.1 mrg
1044 1.1 mrg <h3 class="heading"><a name="TOC36"></a><a name="powerpc_002dx_002ddarwin"></a>powerpc-*-darwin*</h3>
1045 1.1 mrg
1046 1.1 mrg <p>PowerPC running Darwin (Mac OS X kernel).
1047 1.1 mrg
1048 1.1 mrg <p>Pre-installed versions of Mac OS X may not include any developer tools,
1049 1.1 mrg meaning that you will not be able to build GCC from source. Tool
1050 1.1 mrg binaries are available at
1051 1.1 mrg <a href="http://developer.apple.com/darwin/projects/compiler/">http://developer.apple.com/darwin/projects/compiler/</a> (free
1052 1.1 mrg registration required).
1053 1.1 mrg
1054 1.1 mrg <p>This version of GCC requires at least cctools-590.36. The
1055 1.1 mrg cctools-590.36 package referenced from
1056 1.1 mrg <a href="http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html">http://gcc.gnu.org/ml/gcc/2006-03/msg00507.html</a> will not work
1057 1.1 mrg on systems older than 10.3.9 (aka darwin7.9.0).
1058 1.1 mrg
1059 1.1 mrg <p><hr />
1060 1.1 mrg
1061 1.1 mrg <h3 class="heading"><a name="TOC37"></a><a name="powerpc_002dx_002delf"></a>powerpc-*-elf</h3>
1062 1.1 mrg
1063 1.1 mrg <p>PowerPC system in big endian mode, running System V.4.
1064 1.1 mrg
1065 1.1 mrg <p><hr />
1066 1.1 mrg
1067 1.1 mrg <h3 class="heading"><a name="TOC38"></a><a name="powerpc_002dx_002dlinux_002dgnu"></a>powerpc*-*-linux-gnu*</h3>
1068 1.1 mrg
1069 1.1 mrg <p>PowerPC system in big endian mode running Linux.
1070 1.1 mrg
1071 1.1 mrg <p><hr />
1072 1.1 mrg
1073 1.1 mrg <h3 class="heading"><a name="TOC39"></a><a name="powerpc_002dx_002dnetbsd"></a>powerpc-*-netbsd*</h3>
1074 1.1 mrg
1075 1.1 mrg <p>PowerPC system in big endian mode running NetBSD.
1076 1.1 mrg
1077 1.1 mrg <p><hr />
1078 1.1 mrg
1079 1.1 mrg <h3 class="heading"><a name="TOC40"></a><a name="powerpc_002dx_002deabisim"></a>powerpc-*-eabisim</h3>
1080 1.1 mrg
1081 1.1 mrg <p>Embedded PowerPC system in big endian mode for use in running under the
1082 1.1 mrg PSIM simulator.
1083 1.1 mrg
1084 1.1 mrg <p><hr />
1085 1.1 mrg
1086 1.1 mrg <h3 class="heading"><a name="TOC41"></a><a name="powerpc_002dx_002deabi"></a>powerpc-*-eabi</h3>
1087 1.1 mrg
1088 1.1 mrg <p>Embedded PowerPC system in big endian mode.
1089 1.1 mrg
1090 1.1 mrg <p><hr />
1091 1.1 mrg
1092 1.1 mrg <h3 class="heading"><a name="TOC42"></a><a name="powerpcle_002dx_002delf"></a>powerpcle-*-elf</h3>
1093 1.1 mrg
1094 1.1 mrg <p>PowerPC system in little endian mode, running System V.4.
1095 1.1 mrg
1096 1.1 mrg <p><hr />
1097 1.1 mrg
1098 1.1 mrg <h3 class="heading"><a name="TOC43"></a><a name="powerpcle_002dx_002deabisim"></a>powerpcle-*-eabisim</h3>
1099 1.1 mrg
1100 1.1 mrg <p>Embedded PowerPC system in little endian mode for use in running under
1101 1.1 mrg the PSIM simulator.
1102 1.1 mrg
1103 1.1 mrg <p><hr />
1104 1.1 mrg
1105 1.1 mrg <h3 class="heading"><a name="TOC44"></a><a name="powerpcle_002dx_002deabi"></a>powerpcle-*-eabi</h3>
1106 1.1 mrg
1107 1.1 mrg <p>Embedded PowerPC system in little endian mode.
1108 1.1 mrg
1109 1.1 mrg <p><hr />
1110 1.1 mrg
1111 1.1 mrg <h3 class="heading"><a name="TOC45"></a><a name="rx_002dx_002delf"></a>rx-*-elf</h3>
1112 1.1 mrg
1113 1.1 mrg <p>The Renesas RX processor. See
1114 1.1 mrg <a href="http://eu.renesas.com/fmwk.jsp?cnt=rx600_series_landing.jsp&fp=/products/mpumcu/rx_family/rx600_series">http://eu.renesas.com/fmwk.jsp?cnt=rx600_series_landing.jsp&fp=/products/mpumcu/rx_family/rx600_series</a>
1115 1.1 mrg for more information about this processor.
1116 1.1 mrg
1117 1.1 mrg <p><hr />
1118 1.1 mrg
1119 1.1 mrg <h3 class="heading"><a name="TOC46"></a><a name="s390_002dx_002dlinux"></a>s390-*-linux*</h3>
1120 1.1 mrg
1121 1.1 mrg <p>S/390 system running GNU/Linux for S/390.
1122 1.1 mrg
1123 1.1 mrg <p><hr />
1124 1.1 mrg
1125 1.1 mrg <h3 class="heading"><a name="TOC47"></a><a name="s390x_002dx_002dlinux"></a>s390x-*-linux*</h3>
1126 1.1 mrg
1127 1.1 mrg <p>zSeries system (64-bit) running GNU/Linux for zSeries.
1128 1.1 mrg
1129 1.1 mrg <p><hr />
1130 1.1 mrg
1131 1.1 mrg <h3 class="heading"><a name="TOC48"></a><a name="s390x_002dibm_002dtpf"></a>s390x-ibm-tpf*</h3>
1132 1.1 mrg
1133 1.1 mrg <p>zSeries system (64-bit) running TPF. This platform is
1134 1.1 mrg supported as cross-compilation target only.
1135 1.1 mrg
1136 1.1 mrg <p><hr /><!-- Please use Solaris 2 to refer to all release of Solaris, starting -->
1137 1.1 mrg <!-- with 2.0 until 2.6, 7, 8, etc. Solaris 1 was a marketing name for -->
1138 1.1 mrg <!-- SunOS 4 releases which we don't use to avoid confusion. Solaris -->
1139 1.1 mrg <!-- alone is too unspecific and must be avoided. -->
1140 1.1 mrg
1141 1.1 mrg <h3 class="heading"><a name="TOC49"></a><a name="x_002dx_002dsolaris2"></a>*-*-solaris2*</h3>
1142 1.1 mrg
1143 1.1 mrg <p>Support for Solaris 7 has been obsoleted in GCC 4.5, but can still be
1144 1.1 mrg enabled by configuring with <samp><span class="option">--enable-obsolete</span></samp>. Support will be
1145 1.1 mrg removed in GCC 4.6.
1146 1.1 mrg
1147 1.1 mrg <p>Sun does not ship a C compiler with Solaris 2, though you can download
1148 1.1 mrg the Sun Studio compilers for free from
1149 1.1 mrg <a href="http://developers.sun.com/sunstudio/downloads/">http://developers.sun.com/sunstudio/downloads/</a>. Alternatively,
1150 1.1 mrg you can install a pre-built GCC to bootstrap and install GCC. See the
1151 1.1 mrg <a href="binaries.html">binaries page</a> for details.
1152 1.1 mrg
1153 1.1 mrg <p>The Solaris 2 <samp><span class="command">/bin/sh</span></samp> will often fail to configure
1154 1.1 mrg ‘<samp><span class="samp">libstdc++-v3</span></samp>’, ‘<samp><span class="samp">boehm-gc</span></samp>’ or ‘<samp><span class="samp">libjava</span></samp>’. We therefore
1155 1.1 mrg recommend using the following initial sequence of commands
1156 1.1 mrg
1157 1.1 mrg <pre class="smallexample"> % CONFIG_SHELL=/bin/ksh
1158 1.1 mrg % export CONFIG_SHELL
1159 1.1 mrg </pre>
1160 1.1 mrg <p class="noindent">and proceed as described in <a href="configure.html">the configure instructions</a>.
1161 1.1 mrg In addition we strongly recommend specifying an absolute path to invoke
1162 1.1 mrg <samp><var>srcdir</var><span class="command">/configure</span></samp>.
1163 1.1 mrg
1164 1.1 mrg <p>Solaris 2 comes with a number of optional OS packages. Some of these
1165 1.1 mrg are needed to use GCC fully, namely <code>SUNWarc</code>,
1166 1.1 mrg <code>SUNWbtool</code>, <code>SUNWesu</code>, <code>SUNWhea</code>, <code>SUNWlibm</code>,
1167 1.1 mrg <code>SUNWsprot</code>, and <code>SUNWtoo</code>. If you did not install all
1168 1.1 mrg optional packages when installing Solaris 2, you will need to verify that
1169 1.1 mrg the packages that GCC needs are installed.
1170 1.1 mrg
1171 1.1 mrg <p>To check whether an optional package is installed, use
1172 1.1 mrg the <samp><span class="command">pkginfo</span></samp> command. To add an optional package, use the
1173 1.1 mrg <samp><span class="command">pkgadd</span></samp> command. For further details, see the Solaris 2
1174 1.1 mrg documentation.
1175 1.1 mrg
1176 1.1 mrg <p>Trying to use the linker and other tools in
1177 1.1 mrg <samp><span class="file">/usr/ucb</span></samp> to install GCC has been observed to cause trouble.
1178 1.1 mrg For example, the linker may hang indefinitely. The fix is to remove
1179 1.1 mrg <samp><span class="file">/usr/ucb</span></samp> from your <samp><span class="env">PATH</span></samp>.
1180 1.1 mrg
1181 1.1 mrg <p>The build process works more smoothly with the legacy Sun tools so, if you
1182 1.1 mrg have <samp><span class="file">/usr/xpg4/bin</span></samp> in your <samp><span class="env">PATH</span></samp>, we recommend that you place
1183 1.1 mrg <samp><span class="file">/usr/bin</span></samp> before <samp><span class="file">/usr/xpg4/bin</span></samp> for the duration of the build.
1184 1.1 mrg
1185 1.1 mrg <p>We recommend the use of the Sun assembler or the GNU assembler, in
1186 1.1 mrg conjunction with the Sun linker. The GNU <samp><span class="command">as</span></samp>
1187 1.1 mrg versions included in Solaris 10, from GNU binutils 2.15, and Solaris 11,
1188 1.1 mrg from GNU binutils 2.19, are known to work. They can be found in
1189 1.1 mrg <samp><span class="file">/usr/sfw/bin/gas</span></samp>. Current versions of GNU binutils (2.20.1)
1190 1.1 mrg are known to work as well. Note that your mileage may vary
1191 1.1 mrg if you use a combination of the GNU tools and the Sun tools: while the
1192 1.1 mrg combination GNU <samp><span class="command">as</span></samp> + Sun <samp><span class="command">ld</span></samp> should reasonably work,
1193 1.1 mrg the reverse combination Sun <samp><span class="command">as</span></samp> + GNU <samp><span class="command">ld</span></samp> is known to
1194 1.1 mrg cause memory corruption at runtime in some cases for C++ programs.
1195 1.1 mrg <!-- FIXME: still? -->
1196 1.1 mrg GNU <samp><span class="command">ld</span></samp> usually works as well, although the version included in
1197 1.1 mrg Solaris 10 cannot be used due to several bugs. Again, the current
1198 1.1 mrg version (2.20.1) is known to work, but generally lacks platform specific
1199 1.1 mrg features, so better stay with Sun <samp><span class="command">ld</span></samp>.
1200 1.1 mrg
1201 1.1 mrg <p>Sun bug 4296832 turns up when compiling X11 headers with GCC 2.95 or
1202 1.1 mrg newer: <samp><span class="command">g++</span></samp> will complain that types are missing. These headers
1203 1.1 mrg assume that omitting the type means <code>int</code>; this assumption worked for
1204 1.1 mrg C90 but is wrong for C++, and is now wrong for C99 also.
1205 1.1 mrg
1206 1.1 mrg <p><samp><span class="command">g++</span></samp> accepts such (invalid) constructs with the option
1207 1.1 mrg <samp><span class="option">-fpermissive</span></samp>; it will assume that any missing type is <code>int</code>
1208 1.1 mrg (as defined by C90).
1209 1.1 mrg
1210 1.1 mrg <p>There are patches for Solaris 7 (108376-21 or newer for SPARC,
1211 1.1 mrg 108377-20 for Intel), and Solaris 8 (108652-24 or newer for SPARC,
1212 1.1 mrg 108653-22 for Intel) that fix this bug.
1213 1.1 mrg
1214 1.1 mrg <p>Sun bug 4927647 sometimes causes random spurious testsuite failures
1215 1.1 mrg related to missing diagnostic output. This bug doesn't affect GCC
1216 1.1 mrg itself, rather it is a kernel bug triggered by the <samp><span class="command">expect</span></samp>
1217 1.1 mrg program which is used only by the GCC testsuite driver. When the bug
1218 1.1 mrg causes the <samp><span class="command">expect</span></samp> program to miss anticipated output, extra
1219 1.1 mrg testsuite failures appear.
1220 1.1 mrg
1221 1.1 mrg <p>There are patches for Solaris 8 (117350-12 or newer for SPARC,
1222 1.1 mrg 117351-12 or newer for Intel) and Solaris 9 (117171-11 or newer for
1223 1.1 mrg SPARC, 117172-11 or newer for Intel) that address this problem.
1224 1.1 mrg
1225 1.1 mrg <p><hr />
1226 1.1 mrg
1227 1.1 mrg <h3 class="heading"><a name="TOC50"></a><a name="sparc_002dsun_002dsolaris2"></a>sparc-sun-solaris2*</h3>
1228 1.1 mrg
1229 1.1 mrg <p>When GCC is configured to use GNU binutils 2.14 or later, the binaries
1230 1.1 mrg produced are smaller than the ones produced using Sun's native tools;
1231 1.1 mrg this difference is quite significant for binaries containing debugging
1232 1.1 mrg information.
1233 1.1 mrg
1234 1.1 mrg <p>Starting with Solaris 7, the operating system is capable of executing
1235 1.1 mrg 64-bit SPARC V9 binaries. GCC 3.1 and later properly supports
1236 1.1 mrg this; the <samp><span class="option">-m64</span></samp> option enables 64-bit code generation.
1237 1.1 mrg However, if all you want is code tuned for the UltraSPARC CPU, you
1238 1.1 mrg should try the <samp><span class="option">-mtune=ultrasparc</span></samp> option instead, which produces
1239 1.1 mrg code that, unlike full 64-bit code, can still run on non-UltraSPARC
1240 1.1 mrg machines.
1241 1.1 mrg
1242 1.1 mrg <p>When configuring on a Solaris 7 or later system that is running a kernel
1243 1.1 mrg that supports only 32-bit binaries, one must configure with
1244 1.1 mrg <samp><span class="option">--disable-multilib</span></samp>, since we will not be able to build the
1245 1.1 mrg 64-bit target libraries.
1246 1.1 mrg
1247 1.1 mrg <p>GCC 3.3 and GCC 3.4 trigger code generation bugs in earlier versions of
1248 1.1 mrg the GNU compiler (especially GCC 3.0.x versions), which lead to the
1249 1.1 mrg miscompilation of the stage1 compiler and the subsequent failure of the
1250 1.1 mrg bootstrap process. A workaround is to use GCC 3.2.3 as an intermediary
1251 1.1 mrg stage, i.e. to bootstrap that compiler with the base compiler and then
1252 1.1 mrg use it to bootstrap the final compiler.
1253 1.1 mrg
1254 1.1 mrg <p>GCC 3.4 triggers a code generation bug in versions 5.4 (Sun ONE Studio 7)
1255 1.1 mrg and 5.5 (Sun ONE Studio 8) of the Sun compiler, which causes a bootstrap
1256 1.1 mrg failure in form of a miscompilation of the stage1 compiler by the Sun
1257 1.1 mrg compiler. This is Sun bug 4974440. This is fixed with patch 112760-07.
1258 1.1 mrg
1259 1.1 mrg <p>GCC 3.4 changed the default debugging format from Stabs to DWARF-2 for
1260 1.1 mrg 32-bit code on Solaris 7 and later. If you use the Sun assembler, this
1261 1.1 mrg change apparently runs afoul of Sun bug 4910101 (which is referenced as
1262 1.1 mrg an x86-only problem by Sun, probably because they do not use DWARF-2).
1263 1.1 mrg A symptom of the problem is that you cannot compile C++ programs like
1264 1.1 mrg <samp><span class="command">groff</span></samp> 1.19.1 without getting messages similar to the following:
1265 1.1 mrg
1266 1.1 mrg <pre class="smallexample"> ld: warning: relocation error: R_SPARC_UA32: ...
1267 1.1 mrg external symbolic relocation against non-allocatable section
1268 1.1 mrg .debug_info cannot be processed at runtime: relocation ignored.
1269 1.1 mrg </pre>
1270 1.1 mrg <p class="noindent">To work around this problem, compile with <samp><span class="option">-gstabs+</span></samp> instead of
1271 1.1 mrg plain <samp><span class="option">-g</span></samp>.
1272 1.1 mrg
1273 1.1 mrg <p>When configuring the GNU Multiple Precision Library (GMP) or the MPFR
1274 1.1 mrg library on a Solaris 7 or later system, the canonical target triplet
1275 1.1 mrg must be specified as the <samp><span class="command">build</span></samp> parameter on the configure
1276 1.1 mrg line. This triplet can be obtained by invoking <samp><span class="command">./config.guess</span></samp> in
1277 1.1 mrg the toplevel source directory of GCC (and not that of GMP or MPFR).
1278 1.1 mrg For example on a Solaris 7 system:
1279 1.1 mrg
1280 1.1 mrg <pre class="smallexample"> % ./configure --build=sparc-sun-solaris2.7 --prefix=xxx
1281 1.1 mrg </pre>
1282 1.1 mrg <p><hr />
1283 1.1 mrg
1284 1.1 mrg <h3 class="heading"><a name="TOC51"></a><a name="sparc_002dsun_002dsolaris27"></a>sparc-sun-solaris2.7</h3>
1285 1.1 mrg
1286 1.1 mrg <p><em>Note</em> that this configuration has been obsoleted in GCC 4.5, and will be
1287 1.1 mrg removed in GCC 4.6.
1288 1.1 mrg
1289 1.1 mrg <p>Sun patch 107058-01 (1999-01-13) for Solaris 7/SPARC triggers a bug in
1290 1.1 mrg the dynamic linker. This problem (Sun bug 4210064) affects GCC 2.8
1291 1.1 mrg and later, including all EGCS releases. Sun formerly recommended
1292 1.1 mrg 107058-01 for all Solaris 7 users, but around 1999-09-01 it started to
1293 1.1 mrg recommend it only for people who use Sun's compilers.
1294 1.1 mrg
1295 1.1 mrg <p>Here are some workarounds to this problem:
1296 1.1 mrg <ul>
1297 1.1 mrg <li>Do not install Sun patch 107058-01 until after Sun releases a
1298 1.1 mrg complete patch for bug 4210064. This is the simplest course to take,
1299 1.1 mrg unless you must also use Sun's C compiler. Unfortunately 107058-01
1300 1.1 mrg is preinstalled on some new Solaris 7-based hosts, so you may have to
1301 1.1 mrg back it out.
1302 1.1 mrg
1303 1.1 mrg <li>Copy the original, unpatched Solaris 7
1304 1.1 mrg <samp><span class="command">/usr/ccs/bin/as</span></samp> into
1305 1.1 mrg <samp><span class="command">/usr/local/libexec/gcc/sparc-sun-solaris2.7/3.4/as</span></samp>,
1306 1.1 mrg adjusting the latter name to fit your local conventions and software
1307 1.1 mrg version numbers.
1308 1.1 mrg
1309 1.1 mrg <li>Install Sun patch 106950-03 (1999-05-25) or later. Nobody with
1310 1.1 mrg both 107058-01 and 106950-03 installed has reported the bug with GCC
1311 1.1 mrg and Sun's dynamic linker. This last course of action is riskiest,
1312 1.1 mrg for two reasons. First, you must install 106950 on all hosts that
1313 1.1 mrg run code generated by GCC; it doesn't suffice to install it only on
1314 1.1 mrg the hosts that run GCC itself. Second, Sun says that 106950-03 is
1315 1.1 mrg only a partial fix for bug 4210064, but Sun doesn't know whether the
1316 1.1 mrg partial fix is adequate for GCC. Revision -08 or later should fix
1317 1.1 mrg the bug. The current (as of 2004-05-23) revision is -24, and is included in
1318 1.1 mrg the Solaris 7 Recommended Patch Cluster.
1319 1.1 mrg </ul>
1320 1.1 mrg
1321 1.1 mrg <p>GCC 3.3 triggers a bug in version 5.0 Alpha 03/27/98 of the Sun assembler,
1322 1.1 mrg which causes a bootstrap failure when linking the 64-bit shared version of
1323 1.1 mrg ‘<samp><span class="samp">libgcc</span></samp>’. A typical error message is:
1324 1.1 mrg
1325 1.1 mrg <pre class="smallexample"> ld: fatal: relocation error: R_SPARC_32: file libgcc/sparcv9/_muldi3.o:
1326 1.1 mrg symbol <unknown>: offset 0xffffffff7ec133e7 is non-aligned.
1327 1.1 mrg </pre>
1328 1.1 mrg <p class="noindent">This bug has been fixed in the final 5.0 version of the assembler.
1329 1.1 mrg
1330 1.1 mrg <p>A similar problem was reported for version Sun WorkShop 6 99/08/18 of the
1331 1.1 mrg Sun assembler, which causes a bootstrap failure with GCC 4.0.0:
1332 1.1 mrg
1333 1.1 mrg <pre class="smallexample"> ld: fatal: relocation error: R_SPARC_DISP32:
1334 1.1 mrg file .libs/libstdc++.lax/libsupc++convenience.a/vterminate.o:
1335 1.1 mrg symbol <unknown>: offset 0xfccd33ad is non-aligned
1336 1.1 mrg </pre>
1337 1.1 mrg <p class="noindent">This bug has been fixed in more recent revisions of the assembler.
1338 1.1 mrg
1339 1.1 mrg <p><hr />
1340 1.1 mrg
1341 1.1 mrg <h3 class="heading"><a name="TOC52"></a><a name="sparc_002dsun_002dsolaris210"></a>sparc-sun-solaris2.10</h3>
1342 1.1 mrg
1343 1.1 mrg <p>There is a bug in older versions of the Sun assembler which breaks
1344 1.1 mrg thread-local storage (TLS). A typical error message is
1345 1.1 mrg
1346 1.1 mrg <pre class="smallexample"> ld: fatal: relocation error: R_SPARC_TLS_LE_HIX22: file /var/tmp//ccamPA1v.o:
1347 1.1 mrg symbol <unknown>: bad symbol type SECT: symbol type must be TLS
1348 1.1 mrg </pre>
1349 1.1 mrg <p class="noindent">This bug is fixed in Sun patch 118683-03 or later.
1350 1.1 mrg
1351 1.1 mrg <p><hr />
1352 1.1 mrg
1353 1.1 mrg <h3 class="heading"><a name="TOC53"></a><a name="sparc_002dx_002dlinux"></a>sparc-*-linux*</h3>
1354 1.1 mrg
1355 1.1 mrg <p>GCC versions 3.0 and higher require binutils 2.11.2 and glibc 2.2.4
1356 1.1 mrg or newer on this platform. All earlier binutils and glibc
1357 1.1 mrg releases mishandled unaligned relocations on <code>sparc-*-*</code> targets.
1358 1.1 mrg
1359 1.1 mrg <p><hr />
1360 1.1 mrg
1361 1.1 mrg <h3 class="heading"><a name="TOC54"></a><a name="sparc64_002dx_002dsolaris2"></a>sparc64-*-solaris2*</h3>
1362 1.1 mrg
1363 1.1 mrg <p>When configuring the GNU Multiple Precision Library (GMP) or the
1364 1.1 mrg MPFR library, the canonical target triplet must be specified as
1365 1.1 mrg the <samp><span class="command">build</span></samp> parameter on the configure line. For example
1366 1.1 mrg on a Solaris 7 system:
1367 1.1 mrg
1368 1.1 mrg <pre class="smallexample"> % ./configure --build=sparc64-sun-solaris2.7 --prefix=xxx
1369 1.1 mrg </pre>
1370 1.1 mrg <p>The following compiler flags must be specified in the configure
1371 1.1 mrg step in order to bootstrap this target with the Sun compiler:
1372 1.1 mrg
1373 1.1 mrg <pre class="smallexample"> % CC="cc -xarch=v9 -xildoff" <var>srcdir</var>/configure [<var>options</var>] [<var>target</var>]
1374 1.1 mrg </pre>
1375 1.1 mrg <p class="noindent"><samp><span class="option">-xarch=v9</span></samp> specifies the SPARC-V9 architecture to the Sun toolchain
1376 1.1 mrg and <samp><span class="option">-xildoff</span></samp> turns off the incremental linker.
1377 1.1 mrg
1378 1.1 mrg <p><hr />
1379 1.1 mrg
1380 1.1 mrg <h3 class="heading"><a name="TOC55"></a><a name="sparcv9_002dx_002dsolaris2"></a>sparcv9-*-solaris2*</h3>
1381 1.1 mrg
1382 1.1 mrg <p>This is a synonym for ‘<samp><span class="samp">sparc64-*-solaris2*</span></samp>’.
1383 1.1 mrg
1384 1.1 mrg <p><hr />
1385 1.1 mrg
1386 1.1 mrg <h3 class="heading"><a name="TOC56"></a><a name="x_002dx_002dvxworks"></a>*-*-vxworks*</h3>
1387 1.1 mrg
1388 1.1 mrg <p>Support for VxWorks is in flux. At present GCC supports <em>only</em> the
1389 1.1 mrg very recent VxWorks 5.5 (aka Tornado 2.2) release, and only on PowerPC.
1390 1.1 mrg We welcome patches for other architectures supported by VxWorks 5.5.
1391 1.1 mrg Support for VxWorks AE would also be welcome; we believe this is merely
1392 1.1 mrg a matter of writing an appropriate “configlette” (see below). We are
1393 1.1 mrg not interested in supporting older, a.out or COFF-based, versions of
1394 1.1 mrg VxWorks in GCC 3.
1395 1.1 mrg
1396 1.1 mrg <p>VxWorks comes with an older version of GCC installed in
1397 1.1 mrg <samp><var>$WIND_BASE</var><span class="file">/host</span></samp>; we recommend you do not overwrite it.
1398 1.1 mrg Choose an installation <var>prefix</var> entirely outside <var>$WIND_BASE</var>.
1399 1.1 mrg Before running <samp><span class="command">configure</span></samp>, create the directories <samp><var>prefix</var></samp>
1400 1.1 mrg and <samp><var>prefix</var><span class="file">/bin</span></samp>. Link or copy the appropriate assembler,
1401 1.1 mrg linker, etc. into <samp><var>prefix</var><span class="file">/bin</span></samp>, and set your <var>PATH</var> to
1402 1.1 mrg include that directory while running both <samp><span class="command">configure</span></samp> and
1403 1.1 mrg <samp><span class="command">make</span></samp>.
1404 1.1 mrg
1405 1.1 mrg <p>You must give <samp><span class="command">configure</span></samp> the
1406 1.1 mrg <samp><span class="option">--with-headers=</span><var>$WIND_BASE</var><span class="option">/target/h</span></samp> switch so that it can
1407 1.1 mrg find the VxWorks system headers. Since VxWorks is a cross compilation
1408 1.1 mrg target only, you must also specify <samp><span class="option">--target=</span><var>target</var></samp>.
1409 1.1 mrg <samp><span class="command">configure</span></samp> will attempt to create the directory
1410 1.1 mrg <samp><var>prefix</var><span class="file">/</span><var>target</var><span class="file">/sys-include</span></samp> and copy files into it;
1411 1.1 mrg make sure the user running <samp><span class="command">configure</span></samp> has sufficient privilege
1412 1.1 mrg to do so.
1413 1.1 mrg
1414 1.1 mrg <p>GCC's exception handling runtime requires a special “configlette”
1415 1.1 mrg module, <samp><span class="file">contrib/gthr_supp_vxw_5x.c</span></samp>. Follow the instructions in
1416 1.1 mrg that file to add the module to your kernel build. (Future versions of
1417 1.1 mrg VxWorks will incorporate this module.)
1418 1.1 mrg
1419 1.1 mrg <p><hr />
1420 1.1 mrg
1421 1.1 mrg <h3 class="heading"><a name="TOC57"></a><a name="x86_002d64_002dx_002dx"></a>x86_64-*-*, amd64-*-*</h3>
1422 1.1 mrg
1423 1.1 mrg <p>GCC supports the x86-64 architecture implemented by the AMD64 processor
1424 1.1 mrg (amd64-*-* is an alias for x86_64-*-*) on GNU/Linux, FreeBSD and NetBSD.
1425 1.1 mrg On GNU/Linux the default is a bi-arch compiler which is able to generate
1426 1.1 mrg both 64-bit x86-64 and 32-bit x86 code (via the <samp><span class="option">-m32</span></samp> switch).
1427 1.1 mrg
1428 1.1 mrg <p><hr />
1429 1.1 mrg
1430 1.1 mrg <h3 class="heading"><a name="TOC58"></a><a name="xtensa_002dx_002delf"></a>xtensa*-*-elf</h3>
1431 1.1 mrg
1432 1.1 mrg <p>This target is intended for embedded Xtensa systems using the
1433 1.1 mrg ‘<samp><span class="samp">newlib</span></samp>’ C library. It uses ELF but does not support shared
1434 1.1 mrg objects. Designed-defined instructions specified via the
1435 1.1 mrg Tensilica Instruction Extension (TIE) language are only supported
1436 1.1 mrg through inline assembly.
1437 1.1 mrg
1438 1.1 mrg <p>The Xtensa configuration information must be specified prior to
1439 1.1 mrg building GCC. The <samp><span class="file">include/xtensa-config.h</span></samp> header
1440 1.1 mrg file contains the configuration information. If you created your
1441 1.1 mrg own Xtensa configuration with the Xtensa Processor Generator, the
1442 1.1 mrg downloaded files include a customized copy of this header file,
1443 1.1 mrg which you can use to replace the default header file.
1444 1.1 mrg
1445 1.1 mrg <p><hr />
1446 1.1 mrg
1447 1.1 mrg <h3 class="heading"><a name="TOC59"></a><a name="xtensa_002dx_002dlinux"></a>xtensa*-*-linux*</h3>
1448 1.1 mrg
1449 1.1 mrg <p>This target is for Xtensa systems running GNU/Linux. It supports ELF
1450 1.1 mrg shared objects and the GNU C library (glibc). It also generates
1451 1.1 mrg position-independent code (PIC) regardless of whether the
1452 1.1 mrg <samp><span class="option">-fpic</span></samp> or <samp><span class="option">-fPIC</span></samp> options are used. In other
1453 1.1 mrg respects, this target is the same as the
1454 1.1 mrg <a href="#xtensa*-*-elf">‘<samp><span class="samp">xtensa*-*-elf</span></samp>’</a> target.
1455 1.1 mrg
1456 1.1 mrg <p><hr />
1457 1.1 mrg
1458 1.1 mrg <h3 class="heading"><a name="TOC60"></a><a name="windows"></a>Microsoft Windows</h3>
1459 1.1 mrg
1460 1.1 mrg <h4 class="subheading"><a name="TOC61"></a>Intel 16-bit versions</h4>
1461 1.1 mrg
1462 1.1 mrg <p>The 16-bit versions of Microsoft Windows, such as Windows 3.1, are not
1463 1.1 mrg supported.
1464 1.1 mrg
1465 1.1 mrg <p>However, the 32-bit port has limited support for Microsoft
1466 1.1 mrg Windows 3.11 in the Win32s environment, as a target only. See below.
1467 1.1 mrg
1468 1.1 mrg <h4 class="subheading"><a name="TOC62"></a>Intel 32-bit versions</h4>
1469 1.1 mrg
1470 1.1 mrg <p>The 32-bit versions of Windows, including Windows 95, Windows NT, Windows
1471 1.1 mrg XP, and Windows Vista, are supported by several different target
1472 1.1 mrg platforms. These targets differ in which Windows subsystem they target
1473 1.1 mrg and which C libraries are used.
1474 1.1 mrg
1475 1.1 mrg <ul>
1476 1.1 mrg <li>Cygwin <a href="#x-x-cygwin">*-*-cygwin</a>: Cygwin provides a user-space
1477 1.1 mrg Linux API emulation layer in the Win32 subsystem.
1478 1.1 mrg <li>Interix <a href="#x-x-interix">*-*-interix</a>: The Interix subsystem
1479 1.1 mrg provides native support for POSIX.
1480 1.1 mrg <li>MinGW <a href="#x-x-mingw32">*-*-mingw32</a>: MinGW is a native GCC port for
1481 1.1 mrg the Win32 subsystem that provides a subset of POSIX.
1482 1.1 mrg <li>MKS i386-pc-mks: NuTCracker from MKS. See
1483 1.1 mrg <a href="http://www.mkssoftware.com/">http://www.mkssoftware.com/</a> for more information.
1484 1.1 mrg </ul>
1485 1.1 mrg
1486 1.1 mrg <h4 class="subheading"><a name="TOC63"></a>Intel 64-bit versions</h4>
1487 1.1 mrg
1488 1.1 mrg <p>GCC contains support for x86-64 using the mingw-w64
1489 1.1 mrg runtime library, available from <a href="http://mingw-w64.sourceforge.net/">http://mingw-w64.sourceforge.net/</a>.
1490 1.1 mrg This library should be used with the target triple x86_64-pc-mingw32.
1491 1.1 mrg
1492 1.1 mrg <p>Presently Windows for Itanium is not supported.
1493 1.1 mrg
1494 1.1 mrg <h4 class="subheading"><a name="TOC64"></a>Windows CE</h4>
1495 1.1 mrg
1496 1.1 mrg <p>Windows CE is supported as a target only on ARM (arm-wince-pe), Hitachi
1497 1.1 mrg SuperH (sh-wince-pe), and MIPS (mips-wince-pe).
1498 1.1 mrg
1499 1.1 mrg <h4 class="subheading"><a name="TOC65"></a>Other Windows Platforms</h4>
1500 1.1 mrg
1501 1.1 mrg <p>GCC no longer supports Windows NT on the Alpha or PowerPC.
1502 1.1 mrg
1503 1.1 mrg <p>GCC no longer supports the Windows POSIX subsystem. However, it does
1504 1.1 mrg support the Interix subsystem. See above.
1505 1.1 mrg
1506 1.1 mrg <p>Old target names including *-*-winnt and *-*-windowsnt are no longer used.
1507 1.1 mrg
1508 1.1 mrg <p>PW32 (i386-pc-pw32) support was never completed, and the project seems to
1509 1.1 mrg be inactive. See <a href="http://pw32.sourceforge.net/">http://pw32.sourceforge.net/</a> for more information.
1510 1.1 mrg
1511 1.1 mrg <p>UWIN support has been removed due to a lack of maintenance.
1512 1.1 mrg
1513 1.1 mrg <p><hr />
1514 1.1 mrg
1515 1.1 mrg <h3 class="heading"><a name="TOC66"></a><a name="x_002dx_002dcygwin"></a>*-*-cygwin</h3>
1516 1.1 mrg
1517 1.1 mrg <p>Ports of GCC are included with the
1518 1.1 mrg <a href="http://www.cygwin.com/">Cygwin environment</a>.
1519 1.1 mrg
1520 1.1 mrg <p>GCC will build under Cygwin without modification; it does not build
1521 1.1 mrg with Microsoft's C++ compiler and there are no plans to make it do so.
1522 1.1 mrg
1523 1.1 mrg <p>The Cygwin native compiler can be configured to target any 32-bit x86
1524 1.1 mrg cpu architecture desired; the default is i686-pc-cygwin. It should be
1525 1.1 mrg used with as up-to-date a version of binutils as possible; use either
1526 1.1 mrg the latest official GNU binutils release in the Cygwin distribution,
1527 1.1 mrg or version 2.20 or above if building your own.
1528 1.1 mrg
1529 1.1 mrg <p><hr />
1530 1.1 mrg
1531 1.1 mrg <h3 class="heading"><a name="TOC67"></a><a name="x_002dx_002dinterix"></a>*-*-interix</h3>
1532 1.1 mrg
1533 1.1 mrg <p>The Interix target is used by OpenNT, Interix, Services For UNIX (SFU),
1534 1.1 mrg and Subsystem for UNIX-based Applications (SUA). Applications compiled
1535 1.1 mrg with this target run in the Interix subsystem, which is separate from
1536 1.1 mrg the Win32 subsystem. This target was last known to work in GCC 3.3.
1537 1.1 mrg
1538 1.1 mrg <p>For more information, see <a href="http://www.interix.com/">http://www.interix.com/</a>.
1539 1.1 mrg
1540 1.1 mrg <p><hr />
1541 1.1 mrg
1542 1.1 mrg <h3 class="heading"><a name="TOC68"></a><a name="x_002dx_002dmingw32"></a>*-*-mingw32</h3>
1543 1.1 mrg
1544 1.1 mrg <p>GCC will build with and support only MinGW runtime 3.12 and later.
1545 1.1 mrg Earlier versions of headers are incompatible with the new default semantics
1546 1.1 mrg of <code>extern inline</code> in <code>-std=c99</code> and <code>-std=gnu99</code> modes.
1547 1.1 mrg
1548 1.1 mrg <p><hr />
1549 1.1 mrg
1550 1.1 mrg <h3 class="heading"><a name="TOC69"></a><a name="older"></a>Older systems</h3>
1551 1.1 mrg
1552 1.1 mrg <p>GCC contains support files for many older (1980s and early
1553 1.1 mrg 1990s) Unix variants. For the most part, support for these systems
1554 1.1 mrg has not been deliberately removed, but it has not been maintained for
1555 1.1 mrg several years and may suffer from bitrot.
1556 1.1 mrg
1557 1.1 mrg <p>Starting with GCC 3.1, each release has a list of “obsoleted” systems.
1558 1.1 mrg Support for these systems is still present in that release, but
1559 1.1 mrg <samp><span class="command">configure</span></samp> will fail unless the <samp><span class="option">--enable-obsolete</span></samp>
1560 1.1 mrg option is given. Unless a maintainer steps forward, support for these
1561 1.1 mrg systems will be removed from the next release of GCC.
1562 1.1 mrg
1563 1.1 mrg <p>Support for old systems as hosts for GCC can cause problems if the
1564 1.1 mrg workarounds for compiler, library and operating system bugs affect the
1565 1.1 mrg cleanliness or maintainability of the rest of GCC. In some cases, to
1566 1.1 mrg bring GCC up on such a system, if still possible with current GCC, may
1567 1.1 mrg require first installing an old version of GCC which did work on that
1568 1.1 mrg system, and using it to compile a more recent GCC, to avoid bugs in the
1569 1.1 mrg vendor compiler. Old releases of GCC 1 and GCC 2 are available in the
1570 1.1 mrg <samp><span class="file">old-releases</span></samp> directory on the <a href="../mirrors.html">GCC mirror sites</a>. Header bugs may generally be avoided using
1571 1.1 mrg <samp><span class="command">fixincludes</span></samp>, but bugs or deficiencies in libraries and the
1572 1.1 mrg operating system may still cause problems.
1573 1.1 mrg
1574 1.1 mrg <p>Support for older systems as targets for cross-compilation is less
1575 1.1 mrg problematic than support for them as hosts for GCC; if an enthusiast
1576 1.1 mrg wishes to make such a target work again (including resurrecting any of
1577 1.1 mrg the targets that never worked with GCC 2, starting from the last
1578 1.1 mrg version before they were removed), patches
1579 1.1 mrg <a href="../contribute.html">following the usual requirements</a> would be
1580 1.1 mrg likely to be accepted, since they should not affect the support for more
1581 1.1 mrg modern targets.
1582 1.1 mrg
1583 1.1 mrg <p>For some systems, old versions of GNU binutils may also be useful,
1584 1.1 mrg and are available from <samp><span class="file">pub/binutils/old-releases</span></samp> on
1585 1.1 mrg <a href="http://sourceware.org/mirrors.html">sourceware.org mirror sites</a>.
1586 1.1 mrg
1587 1.1 mrg <p>Some of the information on specific systems above relates to
1588 1.1 mrg such older systems, but much of the information
1589 1.1 mrg about GCC on such systems (which may no longer be applicable to
1590 1.1 mrg current GCC) is to be found in the GCC texinfo manual.
1591 1.1 mrg
1592 1.1 mrg <p><hr />
1593 1.1 mrg
1594 1.1 mrg <h3 class="heading"><a name="TOC70"></a><a name="elf"></a>all ELF targets (SVR4, Solaris 2, etc.)</h3>
1595 1.1 mrg
1596 1.1 mrg <p>C++ support is significantly better on ELF targets if you use the
1597 1.1 mrg <a href="./configure.html#with-gnu-ld">GNU linker</a>; duplicate copies of
1598 1.1 mrg inlines, vtables and template instantiations will be discarded
1599 1.1 mrg automatically.
1600 1.1 mrg
1601 1.1 mrg <p><hr />
1602 1.1 mrg <p><a href="./index.html">Return to the GCC Installation page</a>
1603 1.1 mrg
1604 1.1 mrg <!-- ***Old documentation****************************************************** -->
1605 1.1 mrg <!-- ***GFDL******************************************************************** -->
1606 1.1 mrg <!-- *************************************************************************** -->
1607 1.1 mrg <!-- Part 6 The End of the Document -->
1608 1.1 mrg </body></html>
1609 1.1 mrg
1610