1 @c Copyright (C) 1988,89,92,93,94,95,96,97 Free Software Foundation, Inc.
2 @c This is part of the GCC manual.
3 @c For copying conditions, see the file gcc.texi.
5 @c The text of this file appears in the file INSTALL
6 @c in the GCC distribution, as well as in the GCC manual.
10 @chapter Installing GNU CC
12 @cindex installing GNU CC
15 * Configurations:: Configurations Supported by GNU CC.
16 * Other Dir:: Compiling in a separate directory (not where the source is).
17 * Cross-Compiler:: Building and installing a cross-compiler.
18 * Sun Install:: See below for installation on the Sun.
19 * VMS Install:: See below for installation on VMS.
20 * Collect2:: How @code{collect2} works; how it finds @code{ld}.
21 * Header Dirs:: Understanding the standard header file directories.
24 Here is the procedure for installing GNU CC on a Unix system. See
25 @ref{VMS Install}, for VMS systems. In this section we assume you
26 compile in the same directory that contains the source files; see
27 @ref{Other Dir}, to find out how to compile in a separate directory on Unix
30 You cannot install GNU C by itself on MSDOS; it will not compile under
31 any MSDOS compiler except itself. You need to get the complete
32 compilation package DJGPP, which includes binaries as well as sources,
33 and includes all the necessary compilation tools and libraries.
37 If you have built GNU CC previously in the same directory for a
38 different target machine, do @samp{make distclean} to delete all files
39 that might be invalid. One of the files this deletes is
40 @file{Makefile}; if @samp{make distclean} complains that @file{Makefile}
41 does not exist, it probably means that the directory is already suitably
45 On a System V release 4 system, make sure @file{/usr/bin} precedes
46 @file{/usr/ucb} in @code{PATH}. The @code{cc} command in
47 @file{/usr/ucb} uses libraries which have bugs.
50 Specify the host, build and target machine configurations. You do this
51 by running the file @file{configure}.
53 The @dfn{build} machine is the system which you are using, the
54 @dfn{host} machine is the system where you want to run the resulting
55 compiler (normally the build machine), and the @dfn{target} machine is
56 the system for which you want the compiler to generate code.
58 If you are building a compiler to produce code for the machine it runs
59 on (a native compiler), you normally do not need to specify any operands
60 to @file{configure}; it will try to guess the type of machine you are on
61 and use that as the build, host and target machines. So you don't need
62 to specify a configuration when building a native compiler unless
63 @file{configure} cannot figure out what your configuration is or guesses
66 In those cases, specify the build machine's @dfn{configuration name}
67 with the @samp{--build} option; the host and target will default to be
68 the same as the build machine. (If you are building a cross-compiler,
69 see @ref{Cross-Compiler}.)
74 ./configure --build=sparc-sun-sunos4.1
77 A configuration name may be canonical or it may be more or less
80 A canonical configuration name has three parts, separated by dashes.
81 It looks like this: @samp{@var{cpu}-@var{company}-@var{system}}.
82 (The three parts may themselves contain dashes; @file{configure}
83 can figure out which dashes serve which purpose.) For example,
84 @samp{m68k-sun-sunos4.1} specifies a Sun 3.
86 You can also replace parts of the configuration by nicknames or aliases.
87 For example, @samp{sun3} stands for @samp{m68k-sun}, so
88 @samp{sun3-sunos4.1} is another way to specify a Sun 3. You can also
89 use simply @samp{sun3-sunos}, since the version of SunOS is assumed by
90 default to be version 4.
92 You can specify a version number after any of the system types, and some
93 of the CPU types. In most cases, the version is irrelevant, and will be
94 ignored. So you might as well specify the version if you know it.
96 See @ref{Configurations}, for a list of supported configuration names and
97 notes on many of the configurations. You should check the notes in that
98 section before proceeding any further with the installation of GNU CC.
100 There are four additional options you can specify independently to
101 describe variant hardware and software configurations. These are
102 @samp{--with-gnu-as}, @samp{--with-gnu-ld}, @samp{--with-stabs} and
107 If you will use GNU CC with the GNU assembler (GAS), you should declare
108 this by using the @samp{--with-gnu-as} option when you run
111 Using this option does not install GAS. It only modifies the output of
112 GNU CC to work with GAS. Building and installing GAS is up to you.
114 Conversely, if you @emph{do not} wish to use GAS and do not specify
115 @samp{--with-gnu-as} when building GNU CC, it is up to you to make sure
116 that GAS is not installed. GNU CC searches for a program named
117 @code{as} in various directories; if the program it finds is GAS, then
118 it runs GAS. If you are not sure where GNU CC finds the assembler it is
119 using, try specifying @samp{-v} when you run it.
121 The systems where it makes a difference whether you use GAS are@*
122 @samp{hppa1.0-@var{any}-@var{any}}, @samp{hppa1.1-@var{any}-@var{any}},
123 @samp{i386-@var{any}-sysv}, @samp{i386-@var{any}-isc},@*
124 @samp{i860-@var{any}-bsd}, @samp{m68k-bull-sysv},@*
125 @samp{m68k-hp-hpux}, @samp{m68k-sony-bsd},@*
126 @samp{m68k-altos-sysv}, @samp{m68000-hp-hpux},@*
127 @samp{m68000-att-sysv}, @samp{@var{any}-lynx-lynxos},
128 and @samp{mips-@var{any}}).
129 On any other system, @samp{--with-gnu-as} has no effect.
131 On the systems listed above (except for the HP-PA, for ISC on the
132 386, and for @samp{mips-sgi-irix5.*}), if you use GAS, you should also
133 use the GNU linker (and specify @samp{--with-gnu-ld}).
136 Specify the option @samp{--with-gnu-ld} if you plan to use the GNU
139 This option does not cause the GNU linker to be installed; it just
140 modifies the behavior of GNU CC to work with the GNU linker.
141 Specifically, it inhibits the installation of @code{collect2}, a program
142 which otherwise serves as a front-end for the system's linker on most
146 On MIPS based systems and on Alphas, you must specify whether you want
147 GNU CC to create the normal ECOFF debugging format, or to use BSD-style
148 stabs passed through the ECOFF symbol table. The normal ECOFF debug
149 format cannot fully handle languages other than C. BSD stabs format can
150 handle other languages, but it only works with the GNU debugger GDB.
152 Normally, GNU CC uses the ECOFF debugging format by default; if you
153 prefer BSD stabs, specify @samp{--with-stabs} when you configure GNU
156 No matter which default you choose when you configure GNU CC, the user
157 can use the @samp{-gcoff} and @samp{-gstabs+} options to specify explicitly
158 the debug format for a particular compilation.
160 @samp{--with-stabs} is meaningful on the ISC system on the 386, also, if
161 @samp{--with-gas} is used. It selects use of stabs debugging
162 information embedded in COFF output. This kind of debugging information
163 supports C++ well; ordinary COFF debugging information does not.
165 @samp{--with-stabs} is also meaningful on 386 systems running SVR4. It
166 selects use of stabs debugging information embedded in ELF output. The
167 C++ compiler currently (2.6.0) does not support the DWARF debugging
168 information normally used on 386 SVR4 platforms; stabs provide a
169 workable alternative. This requires gas and gdb, as the normal SVR4
170 tools can not generate or interpret stabs.
173 On certain systems, you must specify whether the machine has a floating
174 point unit. These systems include @samp{m68k-sun-sunos@var{n}} and
175 @samp{m68k-isi-bsd}. On any other system, @samp{--nfp} currently has no
176 effect, though perhaps there are other systems where it could usefully
179 @cindex Objective C threads
180 @cindex threads, Objective C
181 @item --enable-objcthreads=@var{type}
182 Certain systems, notably Linux-based GNU systems, can't be relied on to
183 supply a threads facility for the Objective C runtime and so will
184 default to single-threaded runtime. They may, however, have a library
185 threads implementation available, in which case threads can be enabled
186 with this option by supplying a suitable @var{type}, probably
187 @samp{posix}. The possibilities for @var{type} are @samp{single},
188 @samp{posix}, @samp{win32}, @samp{solaris}, @samp{irix} and @samp{mach}.
191 The @file{configure} script searches subdirectories of the source
192 directory for other compilers that are to be integrated into GNU CC.
193 The GNU compiler for C++, called G++ is in a subdirectory named
194 @file{cp}. @file{configure} inserts rules into @file{Makefile} to build
195 all of those compilers.
197 Here we spell out what files will be set up by @code{configure}. Normally
198 you need not be concerned with these files.
203 A file named @file{config.h} is created that contains a @samp{#include}
204 of the top-level config file for the machine you will run the compiler
205 on (@pxref{Config}). This file is responsible for defining information
206 about the host machine. It includes @file{tm.h}.
209 A file named @file{config.h} is created that contains a @samp{#include}
210 of the top-level config file for the machine you will run the compiler
211 on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting
212 GCC}). This file is responsible for defining information about the host
213 machine. It includes @file{tm.h}.
216 The top-level config file is located in the subdirectory @file{config}.
217 Its name is always @file{xm-@var{something}.h}; usually
218 @file{xm-@var{machine}.h}, but there are some exceptions.
220 If your system does not support symbolic links, you might want to
221 set up @file{config.h} to contain a @samp{#include} command which
222 refers to the appropriate file.
225 A file named @file{tconfig.h} is created which includes the top-level config
226 file for your target machine. This is used for compiling certain
227 programs to run on that machine.
230 A file named @file{tm.h} is created which includes the
231 machine-description macro file for your target machine. It should be in
232 the subdirectory @file{config} and its name is often
233 @file{@var{machine}.h}.
236 The command file @file{configure} also constructs the file
237 @file{Makefile} by adding some text to the template file
238 @file{Makefile.in}. The additional text comes from files in the
239 @file{config} directory, named @file{t-@var{target}} and
240 @file{x-@var{host}}. If these files do not exist, it means nothing
241 needs to be added for a given target or host.
245 The standard directory for installing GNU CC is @file{/usr/local/lib}.
246 If you want to install its files somewhere else, specify
247 @samp{--prefix=@var{dir}} when you run @file{configure}. Here @var{dir}
248 is a directory name to use instead of @file{/usr/local} for all purposes
249 with one exception: the directory @file{/usr/local/include} is searched
250 for header files no matter where you install the compiler. To override
251 this name, use the @code{--local-prefix} option below.
254 Specify @samp{--local-prefix=@var{dir}} if you want the compiler to
255 search directory @file{@var{dir}/include} for locally installed header
256 files @emph{instead} of @file{/usr/local/include}.
258 You should specify @samp{--local-prefix} @strong{only} if your site has
259 a different convention (not @file{/usr/local}) for where to put
262 The default value for @samp{--local-prefix} is @file{/usr/local}
263 regardless of the value of @samp{--prefix}. Specifying @samp{--prefix}
264 has no effect on which directory GNU CC searches for local header files.
265 This may seem counterintuitive, but actually it is logical.
267 The purpose of @samp{--prefix} is to specify where to @emph{install GNU
268 CC}. The local header files in @file{/usr/local/include}---if you put
269 any in that directory---are not part of GNU CC. They are part of other
270 programs---perhaps many others. (GNU CC installs its own header files
271 in another directory which is based on the @samp{--prefix} value.)
273 @strong{Do not} specify @file{/usr} as the @samp{--local-prefix}! The
274 directory you use for @samp{--local-prefix} @strong{must not} contain
275 any of the system's standard header files. If it did contain them,
276 certain programs would be miscompiled (including GNU Emacs, on certain
277 targets), because this would override and nullify the header file
278 corrections made by the @code{fixincludes} script.
280 Indications are that people who use this option use it based on
281 mistaken ideas of what it is for. People use it as if it specified
282 where to install part of GNU CC. Perhaps they make this assumption
283 because installing GNU CC creates the directory.
285 @cindex Bison parser generator
286 @cindex parser generator, Bison
288 Make sure the Bison parser generator is installed. (This is
289 unnecessary if the Bison output files @file{c-parse.c} and
290 @file{cexp.c} are more recent than @file{c-parse.y} and @file{cexp.y}
291 and you do not plan to change the @samp{.y} files.)
293 Bison versions older than Sept 8, 1988 will produce incorrect output
294 for @file{c-parse.c}.
297 If you have chosen a configuration for GNU CC which requires other GNU
298 tools (such as GAS or the GNU linker) instead of the standard system
299 tools, install the required tools in the build directory under the names
300 @file{as}, @file{ld} or whatever is appropriate. This will enable the
301 compiler to find the proper tools for compilation of the program
304 Alternatively, you can do subsequent compilation using a value of the
305 @code{PATH} environment variable such that the necessary GNU tools come
306 before the standard system tools.
309 Build the compiler. Just type @samp{make LANGUAGES=c} in the compiler
312 @samp{LANGUAGES=c} specifies that only the C compiler should be
313 compiled. The makefile normally builds compilers for all the supported
314 languages; currently, C, C++ and Objective C. However, C is the only
315 language that is sure to work when you build with other non-GNU C
316 compilers. In addition, building anything but C at this stage is a
319 In general, you can specify the languages to build by typing the
320 argument @samp{LANGUAGES="@var{list}"}, where @var{list} is one or more
321 words from the list @samp{c}, @samp{c++}, and @samp{objective-c}. If
322 you have any additional GNU compilers as subdirectories of the GNU CC
323 source directory, you may also specify their names in this list.
325 Ignore any warnings you may see about ``statement not reached'' in
326 @file{insn-emit.c}; they are normal. Also, warnings about ``unknown
327 escape sequence'' are normal in @file{genopinit.c} and perhaps some
328 other files. Likewise, you should ignore warnings about ``constant is
329 so large that it is unsigned'' in @file{insn-emit.c} and
330 @file{insn-recog.c} and a warning about a comparison always being zero
331 in @file{enquire.o}. Any other compilation errors may represent bugs in
332 the port to your machine or operating system, and
334 should be investigated and reported (@pxref{Bugs}).
337 should be investigated and reported.
340 Some commercial compilers fail to compile GNU CC because they have bugs
341 or limitations. For example, the Microsoft compiler is said to run out
342 of macro space. Some Ultrix compilers run out of expression space; then
343 you need to break up the statement where the problem happens.
346 If you are building a cross-compiler, stop here. @xref{Cross-Compiler}.
350 Move the first-stage object files and executables into a subdirectory
357 The files are moved into a subdirectory named @file{stage1}.
358 Once installation is complete, you may wish to delete these files
359 with @code{rm -r stage1}.
362 If you have chosen a configuration for GNU CC which requires other GNU
363 tools (such as GAS or the GNU linker) instead of the standard system
364 tools, install the required tools in the @file{stage1} subdirectory
365 under the names @file{as}, @file{ld} or whatever is appropriate. This
366 will enable the stage 1 compiler to find the proper tools in the
369 Alternatively, you can do subsequent compilation using a value of the
370 @code{PATH} environment variable such that the necessary GNU tools come
371 before the standard system tools.
374 Recompile the compiler with itself, with this command:
377 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2"
380 This is called making the stage 2 compiler.
382 The command shown above builds compilers for all the supported
383 languages. If you don't want them all, you can specify the languages to
384 build by typing the argument @samp{LANGUAGES="@var{list}"}. @var{list}
385 should contain one or more words from the list @samp{c}, @samp{c++},
386 @samp{objective-c}, and @samp{proto}. Separate the words with spaces.
387 @samp{proto} stands for the programs @code{protoize} and
388 @code{unprotoize}; they are not a separate language, but you use
389 @code{LANGUAGES} to enable or disable their installation.
391 If you are going to build the stage 3 compiler, then you might want to
392 build only the C language in stage 2.
394 Once you have built the stage 2 compiler, if you are short of disk
395 space, you can delete the subdirectory @file{stage1}.
397 On a 68000 or 68020 system lacking floating point hardware,
398 unless you have selected a @file{tm.h} file that expects by default
399 that there is no such hardware, do this instead:
402 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2 -msoft-float"
406 If you wish to test the compiler by compiling it with itself one more
407 time, install any other necessary GNU tools (such as GAS or the GNU
408 linker) in the @file{stage2} subdirectory as you did in the
409 @file{stage1} subdirectory, then do this:
413 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2"
417 This is called making the stage 3 compiler. Aside from the @samp{-B}
418 option, the compiler options should be the same as when you made the
419 stage 2 compiler. But the @code{LANGUAGES} option need not be the
420 same. The command shown above builds compilers for all the supported
421 languages; if you don't want them all, you can specify the languages to
422 build by typing the argument @samp{LANGUAGES="@var{list}"}, as described
425 If you do not have to install any additional GNU tools, you may use the
429 make bootstrap LANGUAGES=@var{language-list} BOOT_CFLAGS=@var{option-list}
433 instead of making @file{stage1}, @file{stage2}, and performing
434 the two compiler builds.
437 Then compare the latest object files with the stage 2 object
438 files---they ought to be identical, aside from time stamps (if any).
440 On some systems, meaningful comparison of object files is impossible;
441 they always appear ``different.'' This is currently true on Solaris and
442 some systems that use ELF object file format. On some versions of Irix
443 on SGI machines and DEC Unix (OSF/1) on Alpha systems, you will not be
444 able to compare the files without specifying @file{-save-temps}; see the
445 description of individual systems above to see if you get comparison
446 failures. You may have similar problems on other systems.
448 Use this command to compare the files:
454 This will mention any object files that differ between stage 2 and stage
455 3. Any difference, no matter how innocuous, indicates that the stage 2
456 compiler has compiled GNU CC incorrectly, and is therefore a potentially
458 serious bug which you should investigate and report (@pxref{Bugs}).
461 serious bug which you should investigate and report.
464 If your system does not put time stamps in the object files, then this
465 is a faster way to compare them (using the Bourne shell):
469 cmp $file stage2/$file
473 If you have built the compiler with the @samp{-mno-mips-tfile} option on
474 MIPS machines, you will not be able to compare the files.
477 Install the compiler driver, the compiler's passes and run-time support
478 with @samp{make install}. Use the same value for @code{CC},
479 @code{CFLAGS} and @code{LANGUAGES} that you used when compiling the
480 files that are being installed. One reason this is necessary is that
481 some versions of Make have bugs and recompile files gratuitously when
482 you do this step. If you use the same variable values, those files will
483 be recompiled properly.
485 For example, if you have built the stage 2 compiler, you can use the
489 make install CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O" LANGUAGES="@var{list}"
493 This copies the files @file{cc1}, @file{cpp} and @file{libgcc.a} to
494 files @file{cc1}, @file{cpp} and @file{libgcc.a} in the directory
495 @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}, which is where
496 the compiler driver program looks for them. Here @var{target} is the
497 target machine type specified when you ran @file{configure}, and
498 @var{version} is the version number of GNU CC. This naming scheme
499 permits various versions and/or cross-compilers to coexist.
500 It also copies the executables for compilers for other languages
501 (e.g., @file{cc1plus} for C++) to the same directory.
503 This also copies the driver program @file{xgcc} into
504 @file{/usr/local/bin/gcc}, so that it appears in typical execution
505 search paths. It also copies @file{gcc.1} into
506 @file{/usr/local/man/man1} and info pages into @file{/usr/local/info}.
508 On some systems, this command causes recompilation of some files. This
509 is usually due to bugs in @code{make}. You should either ignore this
510 problem, or use GNU Make.
512 @cindex @code{alloca} and SunOS
513 @strong{Warning: there is a bug in @code{alloca} in the Sun library. To
514 avoid this bug, be sure to install the executables of GNU CC that were
515 compiled by GNU CC. (That is, the executables from stage 2 or 3, not
516 stage 1.) They use @code{alloca} as a built-in function and never the
519 (It is usually better to install GNU CC executables from stage 2 or 3,
520 since they usually run faster than the ones compiled with some other
524 If you're going to use C++, it's likely that you need to also install
525 the libg++ distribution. It should be available from the same
526 place where you got the GNU C distribution. Just as GNU C does not
527 distribute a C runtime library, it also does not include a C++ run-time
528 library. All I/O functionality, special class libraries, etc., are
529 available in the libg++ distribution.
532 GNU CC includes a runtime library for Objective-C because it is an
533 integral part of the language. You can find the files associated with
534 the library in the subdirectory @file{objc}. The GNU Objective-C
535 Runtime Library requires header files for the target's C library in
536 order to be compiled,and also requires the header files for the target's
537 thread library if you want thread support. @xref{Cross Headers,
538 Cross-Compilers and Header Files, Cross-Compilers and Header Files}, for
539 discussion about header files issues for cross-compilation.
541 When you run @file{configure}, it picks the appropriate Objective-C
542 thread implementation file for the target platform. In some situations,
543 you may wish to choose a different back-end as some platforms support
544 multiple thread implementations or you may wish to disable thread
545 support completely. You do this by specifying a value for the
546 @var{OBJC_THREAD_FILE} makefile variable on the command line when you
547 run make, for example:
550 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" OBJC_THREAD_FILE=thr-single
554 Below is a list of the currently available back-ends.
558 Disable thread support, should work for all platforms.
560 DEC OSF/1 thread support.
562 SGI IRIX thread support.
564 Generic MACH thread support, known to work on NEXTSTEP.
566 IBM OS/2 thread support.
568 Generix POSIX thread support.
570 PCThreads on Linux-based GNU systems.
572 SUN Solaris thread support.
574 Microsoft Win32 API thread support.
579 @section Configurations Supported by GNU CC
580 @cindex configurations supported by GNU CC
582 Here are the possible CPU types:
585 @c gmicro, alliant, spur and tahoe omitted since they don't work.
586 1750a, a29k, alpha, arm, c@var{n}, clipper, dsp16xx, elxsi, h8300,
587 hppa1.0, hppa1.1, i370, i386, i486, i586, i860, i960, m32r, m68000, m68k,
588 m88k, mips, mipsel, mips64, mips64el, ns32k, powerpc, powerpcle,
589 pyramid, romp, rs6000, sh, sparc, sparclite, sparc64, vax, we32k.
592 Here are the recognized company names. As you can see, customary
593 abbreviations are used rather than the longer official names.
595 @c What should be done about merlin, tek*, dolphin?
597 acorn, alliant, altos, apollo, apple, att, bull,
598 cbm, convergent, convex, crds, dec, dg, dolphin,
599 elxsi, encore, harris, hitachi, hp, ibm, intergraph, isi,
600 mips, motorola, ncr, next, ns, omron, plexus,
601 sequent, sgi, sony, sun, tti, unicom, wrs.
604 The company name is meaningful only to disambiguate when the rest of
605 the information supplied is insufficient. You can omit it, writing
606 just @samp{@var{cpu}-@var{system}}, if it is not needed. For example,
607 @samp{vax-ultrix4.2} is equivalent to @samp{vax-dec-ultrix4.2}.
609 Here is a list of system types:
612 386bsd, aix, acis, amigaos, aos, aout, aux, bosx, bsd, clix, coff, ctix, cxux,
613 dgux, dynix, ebmon, ecoff, elf, esix, freebsd, hms, genix, gnu, linux-gnu,
614 hiux, hpux, iris, irix, isc, luna, lynxos, mach, minix, msdos, mvs,
615 netbsd, newsos, nindy, ns, osf, osfrose, ptx, riscix, riscos, rtu, sco, sim,
616 solaris, sunos, sym, sysv, udi, ultrix, unicos, uniplus, unos, vms, vsta,
617 vxworks, winnt, xenix.
621 You can omit the system type; then @file{configure} guesses the
622 operating system from the CPU and company.
624 You can add a version number to the system type; this may or may not
625 make a difference. For example, you can write @samp{bsd4.3} or
626 @samp{bsd4.4} to distinguish versions of BSD. In practice, the version
627 number is most needed for @samp{sysv3} and @samp{sysv4}, which are often
630 If you specify an impossible combination such as @samp{i860-dg-vms},
631 then you may get an error message from @file{configure}, or it may
632 ignore part of the information and do the best it can with the rest.
633 @file{configure} always prints the canonical name for the alternative
634 that it used. GNU CC does not support all possible alternatives.
636 Often a particular model of machine has a name. Many machine names are
637 recognized as aliases for CPU/company combinations. Thus, the machine
638 name @samp{sun3}, mentioned above, is an alias for @samp{m68k-sun}.
639 Sometimes we accept a company name as a machine name, when the name is
640 popularly used for a particular machine. Here is a table of the known
644 3300, 3b1, 3b@var{n}, 7300, altos3068, altos,
645 apollo68, att-7300, balance,
646 convex-c@var{n}, crds, decstation-3100,
647 decstation, delta, encore,
648 fx2800, gmicro, hp7@var{nn}, hp8@var{nn},
649 hp9k2@var{nn}, hp9k3@var{nn}, hp9k7@var{nn},
650 hp9k8@var{nn}, iris4d, iris, isi68,
651 m3230, magnum, merlin, miniframe,
652 mmax, news-3600, news800, news, next,
653 pbd, pc532, pmax, powerpc, powerpcle, ps2, risc-news,
654 rtpc, sun2, sun386i, sun386, sun3,
655 sun4, symmetry, tower-32, tower.
659 Remember that a machine name specifies both the cpu type and the company
661 If you want to install your own homemade configuration files, you can
662 use @samp{local} as the company name to access them. If you use
663 configuration @samp{@var{cpu}-local}, the configuration name
664 without the cpu prefix
665 is used to form the configuration file names.
667 Thus, if you specify @samp{m68k-local}, configuration uses
668 files @file{m68k.md}, @file{local.h}, @file{m68k.c},
669 @file{xm-local.h}, @file{t-local}, and @file{x-local}, all in the
670 directory @file{config/m68k}.
672 Here is a list of configurations that have special treatment or special
673 things you must know:
677 MIL-STD-1750A processors.
679 The MIL-STD-1750A cross configuration produces output for
680 @code{as1750}, an assembler/linker available under the GNU Public
681 License for the 1750A. @code{as1750} can be obtained at
682 @emph{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}.
683 A similarly licensed simulator for
684 the 1750A is available from same address.
686 You should ignore a fatal error during the building of libgcc (libgcc is
687 not yet implemented for the 1750A.)
689 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
690 found in the directory @file{config/1750a}.
692 GNU CC produced the same sections as the Fairchild F9450 C Compiler,
697 The program code section.
700 The read/write (RAM) data section.
703 The read-only (ROM) constants section.
706 Initialization section (code to copy KREL to SREL).
709 The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This
710 means that type `char' is represented with a 16-bit word per character.
711 The 1750A's "Load/Store Upper/Lower Byte" instructions are not used by
715 Systems using processors that implement the DEC Alpha architecture and
716 are running the DEC Unix (OSF/1) operating system, for example the DEC
717 Alpha AXP systems. (VMS on the Alpha is not currently supported by GNU
720 GNU CC writes a @samp{.verstamp} directive to the assembler output file
721 unless it is built as a cross-compiler. It gets the version to use from
722 the system header file @file{/usr/include/stamp.h}. If you install a
723 new version of DEC Unix, you should rebuild GCC to pick up the new version
726 Note that since the Alpha is a 64-bit architecture, cross-compilers from
727 32-bit machines will not generate code as efficient as that generated
728 when the compiler is running on a 64-bit machine because many
729 optimizations that depend on being able to represent a word on the
730 target in an integral value on the host cannot be performed. Building
731 cross-compilers on the Alpha for 32-bit machines has only been tested in
732 a few cases and may not work properly.
734 @code{make compare} may fail on old versions of DEC Unix unless you add
735 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
736 assembler input file is stored in the object file, and that makes
737 comparison fail if it differs between the @code{stage1} and
738 @code{stage2} compilations. The option @samp{-save-temps} forces a
739 fixed name to be used for the assembler input file, instead of a
740 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
741 unless the comparisons fail without that option. If you add
742 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
743 @samp{.s} files after each series of compilations.
745 GNU CC now supports both the native (ECOFF) debugging format used by DBX
746 and GDB and an encapsulated STABS format for use only with GDB. See the
747 discussion of the @samp{--with-stabs} option of @file{configure} above
748 for more information on these formats and how to select them.
750 There is a bug in DEC's assembler that produces incorrect line numbers
751 for ECOFF format when the @samp{.align} directive is used. To work
752 around this problem, GNU CC will not emit such alignment directives
753 while writing ECOFF format debugging information even if optimization is
754 being performed. Unfortunately, this has the very undesirable
755 side-effect that code addresses when @samp{-O} is specified are
756 different depending on whether or not @samp{-g} is also specified.
758 To avoid this behavior, specify @samp{-gstabs+} and use GDB instead of
759 DBX. DEC is now aware of this problem with the assembler and hopes to
760 provide a fix shortly.
763 Argonaut ARC processor.
764 This configuration is intended for embedded systems.
767 Advanced RISC Machines ARM-family processors. These are often used in
768 embedded applications. There are no standard Unix configurations.
769 This configuration corresponds to the basic instruction sequences and will
770 produce @file{a.out} format object modules.
772 You may need to make a variant of the file @file{arm.h} for your particular
775 @item arm-*-linuxaout
776 Any of the ARM family processors running the Linux-based GNU system with
777 the @file{a.out} binary format (ELF is not yet supported). You must use
778 version 2.8.1.0.7 or later of the Linux binutils, which you can download
779 from @file{sunsite.unc.edu:/pub/Linux/GCC} and other mirror sites for
780 Linux-based GNU systems.
783 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix. If
784 you are running a version of RISC iX prior to 1.2 then you must specify
785 the version number during configuration. Note that the assembler
786 shipped with RISC iX does not support stabs debugging information; a
787 new version of the assembler, with stabs support included, is now
788 available from Acorn.
791 AMD Am29k-family processors. These are normally used in embedded
792 applications. There are no standard Unix configurations.
794 corresponds to AMD's standard calling sequence and binary interface
795 and is compatible with other 29k tools.
797 You may need to make a variant of the file @file{a29k.h} for your
798 particular configuration.
801 AMD Am29050 used in a system running a variant of BSD Unix.
804 DECstations can support three different personalities: Ultrix,
805 DEC OSF/1, and OSF/rose. To configure GCC for these platforms
806 use the following configurations:
809 @item decstation-ultrix
810 Ultrix configuration.
812 @item decstation-osf1
813 Dec's version of OSF/1.
815 @item decstation-osfrose
816 Open Software Foundation reference port of OSF/1 which uses the
817 OSF/rose object file format instead of ECOFF. Normally, you
818 would not select this configuration.
821 The MIPS C compiler needs to be told to increase its table size
822 for switch statements with the @samp{-Wf,-XNg1500} option in
823 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
824 optimization option, you also need to use @samp{-Olimit 3000}.
825 Both of these options are automatically generated in the
826 @file{Makefile} that the shell script @file{configure} builds.
827 If you override the @code{CC} make variable and use the MIPS
828 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
830 @item elxsi-elxsi-bsd
831 The Elxsi's C compiler has known limitations that prevent it from
832 compiling GNU C. Please contact @code{mrs@@cygnus.com} for more details.
835 A port to the AT&T DSP1610 family of processors.
839 Alliant FX/8 computer. Note that the standard installed C compiler in
840 Concentrix 5.0 has a bug which prevent it from compiling GNU CC
841 correctly. You can patch the compiler bug as follows:
845 adb -w ./pcc - << EOF
850 Then you must use the @samp{-ip12} option when compiling GNU CC
851 with the patched compiler, as shown here:
854 make CC="./pcc -ip12" CFLAGS=-w
857 Note also that Alliant's version of DBX does not manage to work with the
862 Hitachi H8/300 series of processors.
864 The calling convention and structure layout has changed in release 2.6.
865 All code must be recompiled. The calling convention now passes the
866 first three arguments in function calls in registers. Structures are no
867 longer a multiple of 2 bytes.
870 There are several variants of the HP-PA processor which run a variety
871 of operating systems. GNU CC must be configured to use the correct
872 processor type and operating system, or GNU CC will not function correctly.
873 The easiest way to handle this problem is to @emph{not} specify a target
874 when configuring GNU CC, the @file{configure} script will try to automatically
875 determine the right processor type and operating system.
877 @samp{-g} does not work on HP-UX, since that system uses a peculiar
878 debugging format which GNU CC does not know about. However, @samp{-g}
879 will work if you also use GAS and GDB in conjunction with GCC. We
880 highly recommend using GAS for all HP-PA configurations.
882 You should be using GAS-2.6 (or later) along with GDB-4.16 (or later). These
883 can be retrieved from all the traditional GNU ftp archive sites.
885 GAS will need to be installed into a directory before @code{/bin},
886 @code{/usr/bin}, and @code{/usr/ccs/bin} in your search path. You
887 should install GAS before you build GNU CC.
889 To enable debugging, you must configure GNU CC with the @samp{--with-gnu-as}
890 option before building.
893 This port is very preliminary and has many known bugs. We hope to
894 have a higher-quality port for this machine soon.
896 @item i386-*-linux-gnuoldld
897 Use this configuration to generate @file{a.out} binaries on Linux-based
898 GNU systems if you do not have gas/binutils version 2.5.2 or later
899 installed. This is an obsolete configuration.
901 @item i386-*-linux-gnuaout
902 Use this configuration to generate @file{a.out} binaries on Linux-based
903 GNU systems. This configuration is being superseded. You must use
904 gas/binutils version 2.5.2 or later.
906 @item i386-*-linux-gnu
907 Use this configuration to generate ELF binaries on Linux-based GNU
908 systems. You must use gas/binutils version 2.5.2 or later.
911 Compilation with RCC is recommended. Also, it may be a good idea to
912 link with GNU malloc instead of the malloc that comes with the system.
914 @item i386-*-sco3.2v4
915 Use this configuration for SCO release 3.2 version 4.
917 @item i386-*-sco3.2v5*
918 Use this for the SCO OpenServer Release family including 5.0.0, 5.0.2,
919 5.0.4, Internet FastStart 1.0, and Internet FastStart 1.1.
921 GNU CC can generate ELF binaries (if you specify @samp{-melf}) or COFF
922 binaries (the default). If you are going to build your compiler in ELF
923 mode (once you have bootstrapped the first stage compiler) you
924 @strong{must} specify @samp{-melf} as part of @code{CC},
925 @emph{not} @code{CFLAGS}, for example as
926 @samp{CC="stage1/xgcc -melf -Bstage1/" }. If you do not do this, the
927 bootstrap will generate incorrect versions of @file{libgcc.a}.
929 You must have TLS597 (from ftp.sco.com/TLS) installed for ELF
930 binaries to work correctly. Note that Open Server 5.0.2 @emph{does}
931 need TLS597 installed.
933 @emph{NOTE:} You must follow the instructions about invoking
934 @samp{make bootstrap} because the native OpenServer compiler builds
935 a @file{cc1plus} that will not correctly parse many valid C++ programs.
936 You must do a @samp{make bootstrap} if you are building with the native
940 It may be a good idea to link with GNU malloc instead of the malloc that
941 comes with the system.
943 In ISC version 4.1, @file{sed} core dumps when building
944 @file{deduced.h}. Use the version of @file{sed} from version 4.0.
947 It may be good idea to link with GNU malloc instead of the malloc that
948 comes with the system.
951 You need to use GAS version 2.1 or later, and LD from
952 GNU binutils version 2.2 or later.
954 @item i386-sequent-bsd
955 Go to the Berkeley universe before compiling. In addition, you probably
956 need to create a file named @file{string.h} containing just one line:
957 @samp{#include <strings.h>}.
959 @item i386-sequent-ptx1*
960 Sequent DYNIX/ptx 1.x.
962 @item i386-sequent-ptx2*
963 Sequent DYNIX/ptx 2.x.
965 @item i386-sun-sunos4
966 You may find that you need another version of GNU CC to begin
967 bootstrapping with, since the current version when built with the
968 system's own compiler seems to get an infinite loop compiling part of
969 @file{libgcc2.c}. GNU CC version 2 compiled with GNU CC (any version)
970 seems not to have this problem.
972 See @ref{Sun Install}, for information on installing GNU CC on Sun
975 @item i[345]86-*-winnt3.5
976 This version requires a GAS that has not yet been released. Until it
977 is, you can get a prebuilt binary version via anonymous ftp from
978 @file{cs.washington.edu:pub/gnat} or @file{cs.nyu.edu:pub/gnat}. You
979 must also use the Microsoft header files from the Windows NT 3.5 SDK.
980 Find these on the CDROM in the @file{/mstools/h} directory dated 9/4/94. You
981 must use a fixed version of Microsoft linker made especially for NT 3.5,
982 which is also is available on the NT 3.5 SDK CDROM. If you do not have
983 this linker, can you also use the linker from Visual C/C++ 1.0 or 2.0.
985 Installing GNU CC for NT builds a wrapper linker, called @file{ld.exe},
986 which mimics the behaviour of Unix @file{ld} in the specification of
987 libraries (@samp{-L} and @samp{-l}). @file{ld.exe} looks for both Unix
988 and Microsoft named libraries. For example, if you specify
989 @samp{-lfoo}, @file{ld.exe} will look first for @file{libfoo.a}
990 and then for @file{foo.lib}.
992 You may install GNU CC for Windows NT in one of two ways, depending on
993 whether or not you have a Unix-like shell and various Unix-like
998 If you do not have a Unix-like shell and few Unix-like utilities, you
999 will use a DOS style batch script called @file{configure.bat}. Invoke
1000 it as @code{configure winnt} from an MSDOS console window or from the
1001 program manager dialog box. @file{configure.bat} assumes you have
1002 already installed and have in your path a Unix-like @file{sed} program
1003 which is used to create a working @file{Makefile} from @file{Makefile.in}.
1005 @file{Makefile} uses the Microsoft Nmake program maintenance utility and
1006 the Visual C/C++ V8.00 compiler to build GNU CC. You need only have the
1007 utilities @file{sed} and @file{touch} to use this installation method,
1008 which only automatically builds the compiler itself. You must then
1009 examine what @file{fixinc.winnt} does, edit the header files by hand and
1010 build @file{libgcc.a} manually.
1013 The second type of installation assumes you are running a Unix-like
1014 shell, have a complete suite of Unix-like utilities in your path, and
1015 have a previous version of GNU CC already installed, either through
1016 building it via the above installation method or acquiring a pre-built
1017 binary. In this case, use the @file{configure} script in the normal
1021 @item i860-intel-osf1
1022 This is the Paragon.
1024 If you have version 1.0 of the operating system, you need to take
1025 special steps to build GNU CC due to peculiarities of the system. Newer
1026 system versions have no problem. See the section `Installation Problems'
1027 in the GNU CC Manual.
1029 @ifclear INSTALLONLY
1030 If you have version 1.0 of the operating system,
1031 see @ref{Installation Problems}, for special things you need to do to
1032 compensate for peculiarities in the system.
1036 LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as
1037 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
1038 You can tell GNU CC to use the GNU assembler and linker, by specifying
1039 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
1040 COFF format object files and executables; otherwise GNU CC will use the
1041 installed tools, which produce @file{a.out} format executables.
1044 Mitsubishi M32R processor.
1045 This configuration is intended for embedded systems.
1048 HP 9000 series 200 running BSD. Note that the C compiler that comes
1049 with this system cannot compile GNU CC; contact @code{law@@cs.utah.edu}
1050 to get binaries of GNU CC for bootstrapping.
1053 Altos 3068. You must use the GNU assembler, linker and debugger.
1054 Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}.
1056 @item m68k-apple-aux
1057 Apple Macintosh running A/UX.
1058 You may configure GCC to use either the system assembler and
1059 linker or the GNU assembler and linker. You should use the GNU configuration
1060 if you can, especially if you also want to use GNU C++. You enabled
1061 that configuration with + the @samp{--with-gnu-as} and @samp{--with-gnu-ld}
1062 options to @code{configure}.
1064 Note the C compiler that comes
1065 with this system cannot compile GNU CC. You can fine binaries of GNU CC
1066 for bootstrapping on @code{jagubox.gsfc.nasa.gov}.
1067 You will also a patched version of @file{/bin/ld} there that
1068 raises some of the arbitrary limits found in the original.
1071 AT&T 3b1, a.k.a. 7300 PC. Special procedures are needed to compile GNU
1072 CC with this machine's standard C compiler, due to bugs in that
1073 compiler. You can bootstrap it more easily with
1074 previous versions of GNU CC if you have them.
1076 Installing GNU CC on the 3b1 is difficult if you do not already have
1077 GNU CC running, due to bugs in the installed C compiler. However,
1078 the following procedure might work. We are unable to test it.
1082 Comment out the @samp{#include "config.h"} line near the start of
1083 @file{cccp.c} and do @samp{make cpp}. This makes a preliminary version
1087 Save the old @file{/lib/cpp} and copy the preliminary GNU cpp to that
1091 Undo your change in @file{cccp.c}, or reinstall the original version,
1092 and do @samp{make cpp} again.
1095 Copy this final version of GNU cpp into @file{/lib/cpp}.
1097 @findex obstack_free
1099 Replace every occurrence of @code{obstack_free} in the file
1100 @file{tree.c} with @code{_obstack_free}.
1103 Run @code{make} to get the first-stage GNU CC.
1106 Reinstall the original version of @file{/lib/cpp}.
1109 Now you can compile GNU CC with itself and install it in the normal
1113 @item m68k-bull-sysv
1114 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works
1115 either with native assembler or GNU assembler. You can use
1116 GNU assembler with native coff generation by providing @samp{--with-gnu-as} to
1117 the configure script or use GNU assembler with dbx-in-coff encapsulation
1118 by providing @samp{--with-gnu-as --stabs}. For any problem with native
1119 assembler or for availability of the DPX/2 port of GAS, contact
1120 @code{F.Pierresteguy@@frcl.bull.fr}.
1122 @item m68k-crds-unox
1123 Use @samp{configure unos} for building on Unos.
1125 The Unos assembler is named @code{casm} instead of @code{as}. For some
1126 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
1127 behavior, and does not work. So, when installing GNU CC, you should
1128 install the following script as @file{as} in the subdirectory where
1129 the passes of GCC are installed:
1136 The default Unos library is named @file{libunos.a} instead of
1137 @file{libc.a}. To allow GNU CC to function, either change all
1138 references to @samp{-lc} in @file{gcc.c} to @samp{-lunos} or link
1139 @file{/lib/libc.a} to @file{/lib/libunos.a}.
1141 @cindex @code{alloca}, for Unos
1142 When compiling GNU CC with the standard compiler, to overcome bugs in
1143 the support of @code{alloca}, do not use @samp{-O} when making stage 2.
1144 Then use the stage 2 compiler with @samp{-O} to make the stage 3
1145 compiler. This compiler will have the same characteristics as the usual
1146 stage 2 compiler on other systems. Use it to make a stage 4 compiler
1147 and compare that with stage 3 to verify proper compilation.
1149 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
1150 the comments there will make the above paragraph superfluous. Please
1151 inform us of whether this works.)
1153 Unos uses memory segmentation instead of demand paging, so you will need
1154 a lot of memory. 5 Mb is barely enough if no other tasks are running.
1155 If linking @file{cc1} fails, try putting the object files into a library
1156 and linking from that library.
1159 HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in
1160 the assembler that prevents compilation of GNU CC. To fix it, get patch
1163 In addition, if you wish to use gas @samp{--with-gnu-as} you must use
1164 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
1165 later. Earlier versions of gas relied upon a program which converted the
1166 gas output into the native HP/UX format, but that program has not been
1167 kept up to date. gdb does not understand that native HP/UX format, so
1168 you must use gas if you wish to use gdb.
1171 Sun 3. We do not provide a configuration file to use the Sun FPA by
1172 default, because programs that establish signal handlers for floating
1173 point traps inherently cannot work with the FPA.
1175 See @ref{Sun Install}, for information on installing GNU CC on Sun
1179 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
1180 These systems tend to use the Green Hills C, revision 1.8.5, as the
1181 standard C compiler. There are apparently bugs in this compiler that
1182 result in object files differences between stage 2 and stage 3. If this
1183 happens, make the stage 4 compiler and compare it to the stage 3
1184 compiler. If the stage 3 and stage 4 object files are identical, this
1185 suggests you encountered a problem with the standard C compiler; the
1186 stage 3 and 4 compilers may be usable.
1188 It is best, however, to use an older version of GNU CC for bootstrapping
1192 Motorola m88k running DG/UX. To build 88open BCS native or cross
1193 compilers on DG/UX, specify the configuration name as
1194 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
1195 environment. To build ELF native or cross compilers on DG/UX, specify
1196 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
1197 You set the software development environment by issuing
1198 @samp{sde-target} command and specifying either @samp{m88kbcs} or
1199 @samp{m88kdguxelf} as the operand.
1201 If you do not specify a configuration name, @file{configure} guesses the
1202 configuration based on the current software development environment.
1204 @item m88k-tektronix-sysv3
1205 Tektronix XD88 running UTekV 3.2e. Do not turn on
1206 optimization while building stage1 if you bootstrap with
1207 the buggy Green Hills compiler. Also, The bundled LAI
1208 System V NFS is buggy so if you build in an NFS mounted
1209 directory, start from a fresh reboot, or avoid NFS all together.
1210 Otherwise you may have trouble getting clean comparisons
1214 MIPS machines running the MIPS operating system in BSD mode. It's
1215 possible that some old versions of the system lack the functions
1216 @code{memcpy}, @code{memcmp}, and @code{memset}. If your system lacks
1217 these, you must remove or undo the definition of
1218 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
1220 The MIPS C compiler needs to be told to increase its table size
1221 for switch statements with the @samp{-Wf,-XNg1500} option in
1222 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1223 optimization option, you also need to use @samp{-Olimit 3000}.
1224 Both of these options are automatically generated in the
1225 @file{Makefile} that the shell script @file{configure} builds.
1226 If you override the @code{CC} make variable and use the MIPS
1227 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1229 @item mips-mips-riscos*
1230 The MIPS C compiler needs to be told to increase its table size
1231 for switch statements with the @samp{-Wf,-XNg1500} option in
1232 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1233 optimization option, you also need to use @samp{-Olimit 3000}.
1234 Both of these options are automatically generated in the
1235 @file{Makefile} that the shell script @file{configure} builds.
1236 If you override the @code{CC} make variable and use the MIPS
1237 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1239 MIPS computers running RISC-OS can support four different
1240 personalities: default, BSD 4.3, System V.3, and System V.4
1241 (older versions of RISC-OS don't support V.4). To configure GCC
1242 for these platforms use the following configurations:
1245 @item mips-mips-riscos@code{rev}
1246 Default configuration for RISC-OS, revision @code{rev}.
1248 @item mips-mips-riscos@code{rev}bsd
1249 BSD 4.3 configuration for RISC-OS, revision @code{rev}.
1251 @item mips-mips-riscos@code{rev}sysv4
1252 System V.4 configuration for RISC-OS, revision @code{rev}.
1254 @item mips-mips-riscos@code{rev}sysv
1255 System V.3 configuration for RISC-OS, revision @code{rev}.
1258 The revision @code{rev} mentioned above is the revision of
1259 RISC-OS to use. You must reconfigure GCC when going from a
1260 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
1262 @ifclear INSTALLONLY
1263 bug (see @ref{Installation Problems}, for more details).
1270 In order to compile GCC on an SGI running IRIX 4, the "c.hdr.lib"
1271 option must be installed from the CD-ROM supplied from Silicon Graphics.
1272 This is found on the 2nd CD in release 4.0.1.
1274 In order to compile GCC on an SGI running IRIX 5, the "compiler_dev.hdr"
1275 subsystem must be installed from the IDO CD-ROM supplied by Silicon
1278 @code{make compare} may fail on version 5 of IRIX unless you add
1279 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1280 assembler input file is stored in the object file, and that makes
1281 comparison fail if it differs between the @code{stage1} and
1282 @code{stage2} compilations. The option @samp{-save-temps} forces a
1283 fixed name to be used for the assembler input file, instead of a
1284 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
1285 unless the comparisons fail without that option. If you do you
1286 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
1287 @samp{.s} files after each series of compilations.
1289 The MIPS C compiler needs to be told to increase its table size
1290 for switch statements with the @samp{-Wf,-XNg1500} option in
1291 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1292 optimization option, you also need to use @samp{-Olimit 3000}.
1293 Both of these options are automatically generated in the
1294 @file{Makefile} that the shell script @file{configure} builds.
1295 If you override the @code{CC} make variable and use the MIPS
1296 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1298 On Irix version 4.0.5F, and perhaps on some other versions as well,
1299 there is an assembler bug that reorders instructions incorrectly. To
1300 work around it, specify the target configuration
1301 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
1304 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
1305 off assembler optimization by using the @samp{-noasmopt} option. This
1306 compiler option passes the option @samp{-O0} to the assembler, to
1309 The @samp{-noasmopt} option can be useful for testing whether a problem
1310 is due to erroneous assembler reordering. Even if a problem does not go
1311 away with @samp{-noasmopt}, it may still be due to assembler
1312 reordering---perhaps GNU CC itself was miscompiled as a result.
1314 To enable debugging under Irix 5, you must use GNU as 2.5 or later,
1315 and use the @samp{--with-gnu-as} configure option when configuring gcc.
1316 GNU as is distributed as part of the binutils package.
1318 @item mips-sony-sysv
1319 Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
1320 uses ELF instead of COFF). Support for 5.0.2 will probably be provided
1321 soon by volunteers. In particular, the linker does not like the
1322 code generated by GCC when shared libraries are linked in.
1325 Encore ns32000 system. Encore systems are supported only under BSD.
1328 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
1329 and @code{malloc}; you must get the compiled versions of these from GNU
1333 Go to the Berkeley universe before compiling. In addition, you probably
1334 need to create a file named @file{string.h} containing just one line:
1335 @samp{#include <strings.h>}.
1338 UTEK ns32000 system (``merlin''). The C compiler that comes with this
1339 system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get
1340 binaries of GNU CC for bootstrapping.
1344 The only operating systems supported for the IBM RT PC are AOS and
1345 MACH. GNU CC does not support AIX running on the RT. We recommend you
1346 compile GNU CC with an earlier version of itself; if you compile GNU CC
1347 with @code{hc}, the Metaware compiler, it will work, but you will get
1348 mismatches between the stage 2 and stage 3 compilers in various files.
1349 These errors are minor differences in some floating-point constants and
1350 can be safely ignored; the stage 3 compiler is correct.
1353 @itemx powerpc-*-aix
1354 Various early versions of each release of the IBM XLC compiler will not
1355 bootstrap GNU CC. Symptoms include differences between the stage2 and
1356 stage3 object files, and errors when compiling @file{libgcc.a} or
1357 @file{enquire}. Known problematic releases include: xlc-1.2.1.8,
1358 xlc-1.3.0.0 (distributed with AIX 3.2.5), and xlc-1.3.0.19. Both
1359 xlc-1.2.1.28 and xlc-1.3.0.24 (PTF 432238) are known to produce working
1360 versions of GNU CC, but most other recent releases correctly bootstrap
1361 GNU CC. Also, releases of AIX prior to AIX 3.2.4 include a version of
1362 the IBM assembler which does not accept debugging directives: assembler
1363 updates are available as PTFs. Also, if you are using AIX 3.2.5 or
1364 greater and the GNU assembler, you must have a version modified after
1365 October 16th, 1995 in order for the GNU C compiler to build. See the
1366 file @file{README.RS6000} for more details on of these problems.
1368 GNU CC does not yet support the 64-bit PowerPC instructions.
1370 Objective C does not work on this architecture because it makes assumptions
1371 that are incompatible with the calling conventions.
1373 AIX on the RS/6000 provides support (NLS) for environments outside of
1374 the United States. Compilers and assemblers use NLS to support
1375 locale-specific representations of various objects including
1376 floating-point numbers ("." vs "," for separating decimal fractions).
1377 There have been problems reported where the library linked with GNU CC
1378 does not produce the same floating-point formats that the assembler
1379 accepts. If you have this problem, set the LANG environment variable to
1382 Due to changes in the way that GNU CC invokes the binder (linker) for AIX
1383 4.1, you may now receive warnings of duplicate symbols from the link step
1384 that were not reported before. The assembly files generated by GNU CC for
1385 AIX have always included multiple symbol definitions for certain global
1386 variable and function declarations in the original program. The warnings
1387 should not prevent the linker from producing a correct library or runnable
1390 By default, AIX 4.1 produces code that can be used on either Power or
1393 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1394 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1397 @itemx powerpc-*-sysv4
1398 PowerPC system in big endian mode, running System V.4.
1400 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1401 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1403 @item powerpc-*-linux-gnu
1404 PowerPC system in big endian mode, running the Linux-based GNU system.
1406 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1407 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1409 @item powerpc-*-eabiaix
1410 Embedded PowerPC system in big endian mode with -mcall-aix selected as
1413 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1414 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1416 @item powerpc-*-eabisim
1417 Embedded PowerPC system in big endian mode for use in running under the
1420 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1421 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1423 @item powerpc-*-eabi
1424 Embedded PowerPC system in big endian mode.
1426 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1427 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1429 @item powerpcle-*-elf
1430 @itemx powerpcle-*-sysv4
1431 PowerPC system in little endian mode, running System V.4.
1433 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1434 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1436 @item powerpcle-*-solaris2*
1437 PowerPC system in little endian mode, running Solaris 2.5.1 or higher.
1439 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1440 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1441 Beta versions of the Sun 4.0 compiler do not seem to be able to build
1442 GNU CC correctly. There are also problems with the host assembler and
1443 linker that are fixed by using the GNU versions of these tools.
1445 @item powerpcle-*-eabisim
1446 Embedded PowerPC system in little endian mode for use in running under
1449 @itemx powerpcle-*-eabi
1450 Embedded PowerPC system in little endian mode.
1452 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1453 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1455 @item powerpcle-*-winnt
1456 @itemx powerpcle-*-pe
1457 PowerPC system in little endian mode running Windows NT.
1459 You can specify a default version for the @samp{-mcpu=}@var{cpu_type}
1460 switch by using the configure option @samp{--with-cpu-}@var{cpu_type}.
1462 @item vax-dec-ultrix
1463 Don't try compiling with Vax C (@code{vcc}). It produces incorrect code
1464 in some cases (for example, when @code{alloca} is used).
1466 Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of
1467 an internal table size limitation in that compiler. To avoid this
1468 problem, compile just the GNU C compiler first, and use it to recompile
1469 building all the languages that you want to run.
1472 See @ref{Sun Install}, for information on installing GNU CC on Sun
1476 See @ref{VMS Install}, for details on how to install GNU CC on VMS.
1479 These computers are also known as the 3b2, 3b5, 3b20 and other similar
1480 names. (However, the 3b1 is actually a 68000; see
1481 @ref{Configurations}.)
1483 Don't use @samp{-g} when compiling with the system's compiler. The
1484 system's linker seems to be unable to handle such a large program with
1485 debugging information.
1487 The system's compiler runs out of capacity when compiling @file{stmt.c}
1488 in GNU CC. You can work around this by building @file{cpp} in GNU CC
1489 first, then use that instead of the system's preprocessor with the
1490 system's C compiler to compile @file{stmt.c}. Here is how:
1493 mv /lib/cpp /lib/cpp.att
1495 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
1499 The system's compiler produces bad code for some of the GNU CC
1500 optimization files. So you must build the stage 2 compiler without
1501 optimization. Then build a stage 3 compiler with optimization.
1502 That executable should work. Here are the necessary commands:
1505 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
1507 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
1510 You may need to raise the ULIMIT setting to build a C++ compiler,
1511 as the file @file{cc1plus} is larger than one megabyte.
1515 @section Compilation in a Separate Directory
1516 @cindex other directory, compilation in
1517 @cindex compilation in a separate directory
1518 @cindex separate directory, compilation in
1520 If you wish to build the object files and executables in a directory
1521 other than the one containing the source files, here is what you must
1526 Make sure you have a version of Make that supports the @code{VPATH}
1527 feature. (GNU Make supports it, as do Make versions on most BSD
1531 If you have ever run @file{configure} in the source directory, you must undo
1532 the configuration. Do this by running:
1539 Go to the directory in which you want to build the compiler before
1540 running @file{configure}:
1547 On systems that do not support symbolic links, this directory must be
1548 on the same file system as the source code directory.
1551 Specify where to find @file{configure} when you run it:
1554 ../gcc/configure @dots{}
1557 This also tells @code{configure} where to find the compiler sources;
1558 @code{configure} takes the directory from the file name that was used to
1559 invoke it. But if you want to be sure, you can specify the source
1560 directory with the @samp{--srcdir} option, like this:
1563 ../gcc/configure --srcdir=../gcc @var{other options}
1566 The directory you specify with @samp{--srcdir} need not be the same
1567 as the one that @code{configure} is found in.
1570 Now, you can run @code{make} in that directory. You need not repeat the
1571 configuration steps shown above, when ordinary source files change. You
1572 must, however, run @code{configure} again when the configuration files
1573 change, if your system does not support symbolic links.
1575 @node Cross-Compiler
1576 @section Building and Installing a Cross-Compiler
1577 @cindex cross-compiler, installation
1579 GNU CC can function as a cross-compiler for many machines, but not all.
1583 Cross-compilers for the Mips as target using the Mips assembler
1584 currently do not work, because the auxiliary programs
1585 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
1586 anything but a Mips. It does work to cross compile for a Mips
1587 if you use the GNU assembler and linker.
1590 Cross-compilers between machines with different floating point formats
1591 have not all been made to work. GNU CC now has a floating point
1592 emulator with which these can work, but each target machine description
1593 needs to be updated to take advantage of it.
1596 Cross-compilation between machines of different word sizes is
1597 somewhat problematic and sometimes does not work.
1600 Since GNU CC generates assembler code, you probably need a
1601 cross-assembler that GNU CC can run, in order to produce object files.
1602 If you want to link on other than the target machine, you need a
1603 cross-linker as well. You also need header files and libraries suitable
1604 for the target machine that you can install on the host machine.
1607 * Steps of Cross:: Using a cross-compiler involves several steps
1608 that may be carried out on different machines.
1609 * Configure Cross:: Configuring a cross-compiler.
1610 * Tools and Libraries:: Where to put the linker and assembler, and the C library.
1611 * Cross Headers:: Finding and installing header files
1612 for a cross-compiler.
1613 * Cross Runtime:: Supplying arithmetic runtime routines (@file{libgcc1.a}).
1614 * Build Cross:: Actually compiling the cross-compiler.
1617 @node Steps of Cross
1618 @subsection Steps of Cross-Compilation
1620 To compile and run a program using a cross-compiler involves several
1625 Run the cross-compiler on the host machine to produce assembler files
1626 for the target machine. This requires header files for the target
1630 Assemble the files produced by the cross-compiler. You can do this
1631 either with an assembler on the target machine, or with a
1632 cross-assembler on the host machine.
1635 Link those files to make an executable. You can do this either with a
1636 linker on the target machine, or with a cross-linker on the host
1637 machine. Whichever machine you use, you need libraries and certain
1638 startup files (typically @file{crt@dots{}.o}) for the target machine.
1641 It is most convenient to do all of these steps on the same host machine,
1642 since then you can do it all with a single invocation of GNU CC. This
1643 requires a suitable cross-assembler and cross-linker. For some targets,
1644 the GNU assembler and linker are available.
1646 @node Configure Cross
1647 @subsection Configuring a Cross-Compiler
1649 To build GNU CC as a cross-compiler, you start out by running
1650 @file{configure}. Use the @samp{--target=@var{target}} to specify the
1651 target type. If @file{configure} was unable to correctly identify the
1652 system you are running on, also specify the @samp{--build=@var{build}}
1653 option. For example, here is how to configure for a cross-compiler that
1654 produces code for an HP 68030 system running BSD on a system that
1655 @file{configure} can correctly identify:
1658 ./configure --target=m68k-hp-bsd4.3
1661 @node Tools and Libraries
1662 @subsection Tools and Libraries for a Cross-Compiler
1664 If you have a cross-assembler and cross-linker available, you should
1665 install them now. Put them in the directory
1666 @file{/usr/local/@var{target}/bin}. Here is a table of the tools
1667 you should put in this directory:
1671 This should be the cross-assembler.
1674 This should be the cross-linker.
1677 This should be the cross-archiver: a program which can manipulate
1678 archive files (linker libraries) in the target machine's format.
1681 This should be a program to construct a symbol table in an archive file.
1684 The installation of GNU CC will find these programs in that directory,
1685 and copy or link them to the proper place to for the cross-compiler to
1686 find them when run later.
1688 The easiest way to provide these files is to build the Binutils package
1689 and GAS. Configure them with the same @samp{--host} and @samp{--target}
1690 options that you use for configuring GNU CC, then build and install
1691 them. They install their executables automatically into the proper
1692 directory. Alas, they do not support all the targets that GNU CC
1695 If you want to install libraries to use with the cross-compiler, such as
1696 a standard C library, put them in the directory
1697 @file{/usr/local/@var{target}/lib}; installation of GNU CC copies
1698 all the files in that subdirectory into the proper place for GNU CC to
1699 find them and link with them. Here's an example of copying some
1700 libraries from a target machine:
1703 ftp @var{target-machine}
1704 lcd /usr/local/@var{target}/lib
1714 The precise set of libraries you'll need, and their locations on
1715 the target machine, vary depending on its operating system.
1718 Many targets require ``start files'' such as @file{crt0.o} and
1719 @file{crtn.o} which are linked into each executable; these too should be
1720 placed in @file{/usr/local/@var{target}/lib}. There may be several
1721 alternatives for @file{crt0.o}, for use with profiling or other
1722 compilation options. Check your target's definition of
1723 @code{STARTFILE_SPEC} to find out what start files it uses.
1724 Here's an example of copying these files from a target machine:
1727 ftp @var{target-machine}
1728 lcd /usr/local/@var{target}/lib
1738 @subsection @file{libgcc.a} and Cross-Compilers
1740 Code compiled by GNU CC uses certain runtime support functions
1741 implicitly. Some of these functions can be compiled successfully with
1742 GNU CC itself, but a few cannot be. These problem functions are in the
1743 source file @file{libgcc1.c}; the library made from them is called
1746 When you build a native compiler, these functions are compiled with some
1747 other compiler--the one that you use for bootstrapping GNU CC.
1748 Presumably it knows how to open code these operations, or else knows how
1749 to call the run-time emulation facilities that the machine comes with.
1750 But this approach doesn't work for building a cross-compiler. The
1751 compiler that you use for building knows about the host system, not the
1754 So, when you build a cross-compiler you have to supply a suitable
1755 library @file{libgcc1.a} that does the job it is expected to do.
1757 To compile @file{libgcc1.c} with the cross-compiler itself does not
1758 work. The functions in this file are supposed to implement arithmetic
1759 operations that GNU CC does not know how to open code for your target
1760 machine. If these functions are compiled with GNU CC itself, they
1761 will compile into infinite recursion.
1763 On any given target, most of these functions are not needed. If GNU CC
1764 can open code an arithmetic operation, it will not call these functions
1765 to perform the operation. It is possible that on your target machine,
1766 none of these functions is needed. If so, you can supply an empty
1767 library as @file{libgcc1.a}.
1769 Many targets need library support only for multiplication and division.
1770 If you are linking with a library that contains functions for
1771 multiplication and division, you can tell GNU CC to call them directly
1772 by defining the macros @code{MULSI3_LIBCALL}, and the like. These
1773 macros need to be defined in the target description macro file. For
1774 some targets, they are defined already. This may be sufficient to
1775 avoid the need for libgcc1.a; if so, you can supply an empty library.
1777 Some targets do not have floating point instructions; they need other
1778 functions in @file{libgcc1.a}, which do floating arithmetic.
1779 Recent versions of GNU CC have a file which emulates floating point.
1780 With a certain amount of work, you should be able to construct a
1781 floating point emulator that can be used as @file{libgcc1.a}. Perhaps
1782 future versions will contain code to do this automatically and
1783 conveniently. That depends on whether someone wants to implement it.
1785 Some embedded targets come with all the necessary @file{libgcc1.a}
1786 routines written in C or assembler. These targets build
1787 @file{libgcc1.a} automatically and you do not need to do anything
1788 special for them. Other embedded targets do not need any
1789 @file{libgcc1.a} routines since all the necessary operations are
1790 supported by the hardware.
1792 If your target system has another C compiler, you can configure GNU CC
1793 as a native compiler on that machine, build just @file{libgcc1.a} with
1794 @samp{make libgcc1.a} on that machine, and use the resulting file with
1795 the cross-compiler. To do this, execute the following on the target
1799 cd @var{target-build-dir}
1800 ./configure --host=sparc --target=sun3
1805 And then this on the host machine:
1808 ftp @var{target-machine}
1810 cd @var{target-build-dir}
1815 Another way to provide the functions you need in @file{libgcc1.a} is to
1816 define the appropriate @code{perform_@dots{}} macros for those
1817 functions. If these definitions do not use the C arithmetic operators
1818 that they are meant to implement, you should be able to compile them
1819 with the cross-compiler you are building. (If these definitions already
1820 exist for your target file, then you are all set.)
1822 To build @file{libgcc1.a} using the perform macros, use
1823 @samp{LIBGCC1=libgcc1.a OLDCC=./xgcc} when building the compiler.
1824 Otherwise, you should place your replacement library under the name
1825 @file{libgcc1.a} in the directory in which you will build the
1826 cross-compiler, before you run @code{make}.
1829 @subsection Cross-Compilers and Header Files
1831 If you are cross-compiling a standalone program or a program for an
1832 embedded system, then you may not need any header files except the few
1833 that are part of GNU CC (and those of your program). However, if you
1834 intend to link your program with a standard C library such as
1835 @file{libc.a}, then you probably need to compile with the header files
1836 that go with the library you use.
1838 The GNU C compiler does not come with these files, because (1) they are
1839 system-specific, and (2) they belong in a C library, not in a compiler.
1841 If the GNU C library supports your target machine, then you can get the
1842 header files from there (assuming you actually use the GNU library when
1843 you link your program).
1845 If your target machine comes with a C compiler, it probably comes with
1846 suitable header files also. If you make these files accessible from the host
1847 machine, the cross-compiler can use them also.
1849 Otherwise, you're on your own in finding header files to use when
1852 When you have found suitable header files, put them in the directory
1853 @file{/usr/local/@var{target}/include}, before building the cross
1854 compiler. Then installation will run fixincludes properly and install
1855 the corrected versions of the header files where the compiler will use
1858 Provide the header files before you build the cross-compiler, because
1859 the build stage actually runs the cross-compiler to produce parts of
1860 @file{libgcc.a}. (These are the parts that @emph{can} be compiled with
1861 GNU CC.) Some of them need suitable header files.
1863 Here's an example showing how to copy the header files from a target
1864 machine. On the target machine, do this:
1867 (cd /usr/include; tar cf - .) > tarfile
1870 Then, on the host machine, do this:
1873 ftp @var{target-machine}
1874 lcd /usr/local/@var{target}/include
1881 @subsection Actually Building the Cross-Compiler
1883 Now you can proceed just as for compiling a single-machine compiler
1884 through the step of building stage 1. If you have not provided some
1885 sort of @file{libgcc1.a}, then compilation will give up at the point
1886 where it needs that file, printing a suitable error message. If you
1887 do provide @file{libgcc1.a}, then building the compiler will automatically
1888 compile and link a test program called @file{libgcc1-test}; if you get
1889 errors in the linking, it means that not all of the necessary routines
1890 in @file{libgcc1.a} are available.
1892 You must provide the header file @file{float.h}. One way to do this is
1893 to compile @file{enquire} and run it on your target machine. The job of
1894 @file{enquire} is to run on the target machine and figure out by
1895 experiment the nature of its floating point representation.
1896 @file{enquire} records its findings in the header file @file{float.h}.
1897 If you can't produce this file by running @file{enquire} on the target
1898 machine, then you will need to come up with a suitable @file{float.h} in
1899 some other way (or else, avoid using it in your programs).
1901 Do not try to build stage 2 for a cross-compiler. It doesn't work to
1902 rebuild GNU CC as a cross-compiler using the cross-compiler, because
1903 that would produce a program that runs on the target machine, not on the
1904 host. For example, if you compile a 386-to-68030 cross-compiler with
1905 itself, the result will not be right either for the 386 (because it was
1906 compiled into 68030 code) or for the 68030 (because it was configured
1907 for a 386 as the host). If you want to compile GNU CC into 68030 code,
1908 whether you compile it on a 68030 or with a cross-compiler on a 386, you
1909 must specify a 68030 as the host when you configure it.
1911 To install the cross-compiler, use @samp{make install}, as usual.
1914 @section Installing GNU CC on the Sun
1915 @cindex Sun installation
1916 @cindex installing GNU CC on the Sun
1918 On Solaris (version 2.1), do not use the linker or other tools in
1919 @file{/usr/ucb} to build GNU CC. Use @code{/usr/ccs/bin}.
1921 Make sure the environment variable @code{FLOAT_OPTION} is not set when
1922 you compile @file{libgcc.a}. If this option were set to @code{f68881}
1923 when @file{libgcc.a} is compiled, the resulting code would demand to be
1924 linked with a special startup file and would not link properly without
1927 @cindex @code{alloca}, for SunOS
1928 There is a bug in @code{alloca} in certain versions of the Sun library.
1929 To avoid this bug, install the binaries of GNU CC that were compiled by
1930 GNU CC. They use @code{alloca} as a built-in function and never the one
1933 Some versions of the Sun compiler crash when compiling GNU CC. The
1934 problem is a segmentation fault in cpp. This problem seems to be due to
1935 the bulk of data in the environment variables. You may be able to avoid
1936 it by using the following command to compile GNU CC with Sun CC:
1939 make CC="TERMCAP=x OBJS=x LIBFUNCS=x STAGESTUFF=x cc"
1942 SunOS 4.1.3 and 4.1.3_U1 have bugs that can cause intermittent core
1943 dumps when compiling GNU CC. A common symptom is an
1944 internal compiler error which does not recur if you run it again.
1945 To fix the problem, install Sun recommended patch 100726 (for SunOS 4.1.3)
1946 or 101508 (for SunOS 4.1.3_U1), or upgrade to a later SunOS release.
1949 @section Installing GNU CC on VMS
1950 @cindex VMS installation
1951 @cindex installing GNU CC on VMS
1953 The VMS version of GNU CC is distributed in a backup saveset containing
1954 both source code and precompiled binaries.
1956 To install the @file{gcc} command so you can use the compiler easily, in
1957 the same manner as you use the VMS C compiler, you must install the VMS CLD
1958 file for GNU CC as follows:
1962 Define the VMS logical names @samp{GNU_CC} and @samp{GNU_CC_INCLUDE}
1963 to point to the directories where the GNU CC executables
1964 (@file{gcc-cpp.exe}, @file{gcc-cc1.exe}, etc.) and the C include files are
1965 kept respectively. This should be done with the commands:@refill
1968 $ assign /system /translation=concealed -
1970 $ assign /system /translation=concealed -
1971 disk:[gcc.include.] gnu_cc_include
1975 with the appropriate disk and directory names. These commands can be
1976 placed in your system startup file so they will be executed whenever
1977 the machine is rebooted. You may, if you choose, do this via the
1978 @file{GCC_INSTALL.COM} script in the @file{[GCC]} directory.
1981 Install the @file{GCC} command with the command line:
1984 $ set command /table=sys$common:[syslib]dcltables -
1985 /output=sys$common:[syslib]dcltables gnu_cc:[000000]gcc
1986 $ install replace sys$common:[syslib]dcltables
1990 To install the help file, do the following:
1993 $ library/help sys$library:helplib.hlb gcc.hlp
1997 Now you can invoke the compiler with a command like @samp{gcc /verbose
1998 file.c}, which is equivalent to the command @samp{gcc -v -c file.c} in
2002 If you wish to use GNU C++ you must first install GNU CC, and then
2003 perform the following steps:
2007 Define the VMS logical name @samp{GNU_GXX_INCLUDE} to point to the
2008 directory where the preprocessor will search for the C++ header files.
2009 This can be done with the command:@refill
2012 $ assign /system /translation=concealed -
2013 disk:[gcc.gxx_include.] gnu_gxx_include
2017 with the appropriate disk and directory name. If you are going to be
2018 using libg++, this is where the libg++ install procedure will install
2019 the libg++ header files.
2022 Obtain the file @file{gcc-cc1plus.exe}, and place this in the same
2023 directory that @file{gcc-cc1.exe} is kept.
2025 The GNU C++ compiler can be invoked with a command like @samp{gcc /plus
2026 /verbose file.cc}, which is equivalent to the command @samp{g++ -v -c
2030 We try to put corresponding binaries and sources on the VMS distribution
2031 tape. But sometimes the binaries will be from an older version than the
2032 sources, because we don't always have time to update them. (Use the
2033 @samp{/version} option to determine the version number of the binaries and
2034 compare it with the source file @file{version.c} to tell whether this is
2035 so.) In this case, you should use the binaries you get to recompile the
2036 sources. If you must recompile, here is how:
2040 Execute the command procedure @file{vmsconfig.com} to set up the files
2041 @file{tm.h}, @file{config.h}, @file{aux-output.c}, and @file{md.}, and
2042 to create files @file{tconfig.h} and @file{hconfig.h}. This procedure
2043 also creates several linker option files used by @file{make-cc1.com} and
2044 a data file used by @file{make-l2.com}.@refill
2051 Setup the logical names and command tables as defined above. In
2052 addition, define the VMS logical name @samp{GNU_BISON} to point at the
2053 to the directories where the Bison executable is kept. This should be
2054 done with the command:@refill
2057 $ assign /system /translation=concealed -
2058 disk:[bison.] gnu_bison
2061 You may, if you choose, use the @file{INSTALL_BISON.COM} script in the
2062 @file{[BISON]} directory.
2065 Install the @samp{BISON} command with the command line:@refill
2068 $ set command /table=sys$common:[syslib]dcltables -
2069 /output=sys$common:[syslib]dcltables -
2070 gnu_bison:[000000]bison
2071 $ install replace sys$common:[syslib]dcltables
2075 Type @samp{@@make-gcc} to recompile everything (alternatively, submit
2076 the file @file{make-gcc.com} to a batch queue). If you wish to build
2077 the GNU C++ compiler as well as the GNU CC compiler, you must first edit
2078 @file{make-gcc.com} and follow the instructions that appear in the
2082 In order to use GCC, you need a library of functions which GCC compiled code
2083 will call to perform certain tasks, and these functions are defined in the
2084 file @file{libgcc2.c}. To compile this you should use the command procedure
2085 @file{make-l2.com}, which will generate the library @file{libgcc2.olb}.
2086 @file{libgcc2.olb} should be built using the compiler built from
2087 the same distribution that @file{libgcc2.c} came from, and
2088 @file{make-gcc.com} will automatically do all of this for you.
2090 To install the library, use the following commands:@refill
2093 $ library gnu_cc:[000000]gcclib/delete=(new,eprintf)
2094 $ library gnu_cc:[000000]gcclib/delete=L_*
2095 $ library libgcc2/extract=*/output=libgcc2.obj
2096 $ library gnu_cc:[000000]gcclib libgcc2.obj
2099 The first command simply removes old modules that will be replaced with
2100 modules from @file{libgcc2} under different module names. The modules
2101 @code{new} and @code{eprintf} may not actually be present in your
2102 @file{gcclib.olb}---if the VMS librarian complains about those modules
2103 not being present, simply ignore the message and continue on with the
2104 next command. The second command removes the modules that came from the
2105 previous version of the library @file{libgcc2.c}.
2107 Whenever you update the compiler on your system, you should also update the
2108 library with the above procedure.
2111 You may wish to build GCC in such a way that no files are written to the
2112 directory where the source files reside. An example would be the when
2113 the source files are on a read-only disk. In these cases, execute the
2114 following DCL commands (substituting your actual path names):
2117 $ assign dua0:[gcc.build_dir.]/translation=concealed, -
2118 dua1:[gcc.source_dir.]/translation=concealed gcc_build
2119 $ set default gcc_build:[000000]
2123 where the directory @file{dua1:[gcc.source_dir]} contains the source
2124 code, and the directory @file{dua0:[gcc.build_dir]} is meant to contain
2125 all of the generated object files and executables. Once you have done
2126 this, you can proceed building GCC as described above. (Keep in mind
2127 that @file{gcc_build} is a rooted logical name, and thus the device
2128 names in each element of the search list must be an actual physical
2129 device name rather than another rooted logical name).
2132 @strong{If you are building GNU CC with a previous version of GNU CC,
2133 you also should check to see that you have the newest version of the
2134 assembler}. In particular, GNU CC version 2 treats global constant
2135 variables slightly differently from GNU CC version 1, and GAS version
2136 1.38.1 does not have the patches required to work with GCC version 2.
2137 If you use GAS 1.38.1, then @code{extern const} variables will not have
2138 the read-only bit set, and the linker will generate warning messages
2139 about mismatched psect attributes for these variables. These warning
2140 messages are merely a nuisance, and can safely be ignored.
2142 If you are compiling with a version of GNU CC older than 1.33, specify
2143 @samp{/DEFINE=("inline=")} as an option in all the compilations. This
2144 requires editing all the @code{gcc} commands in @file{make-cc1.com}.
2145 (The older versions had problems supporting @code{inline}.) Once you
2146 have a working 1.33 or newer GNU CC, you can change this file back.
2149 If you want to build GNU CC with the VAX C compiler, you will need to
2150 make minor changes in @file{make-cccp.com} and @file{make-cc1.com}
2151 to choose alternate definitions of @code{CC}, @code{CFLAGS}, and
2152 @code{LIBS}. See comments in those files. However, you must
2153 also have a working version of the GNU assembler (GNU as, aka GAS) as
2154 it is used as the back-end for GNU CC to produce binary object modules
2155 and is not included in the GNU CC sources. GAS is also needed to
2156 compile @file{libgcc2} in order to build @file{gcclib} (see above);
2157 @file{make-l2.com} expects to be able to find it operational in
2158 @file{gnu_cc:[000000]gnu-as.exe}.
2160 To use GNU CC on VMS, you need the VMS driver programs
2161 @file{gcc.exe}, @file{gcc.com}, and @file{gcc.cld}. They are
2162 distributed with the VMS binaries (@file{gcc-vms}) rather than the
2163 GNU CC sources. GAS is also included in @file{gcc-vms}, as is Bison.
2165 Once you have successfully built GNU CC with VAX C, you should use the
2166 resulting compiler to rebuild itself. Before doing this, be sure to
2167 restore the @code{CC}, @code{CFLAGS}, and @code{LIBS} definitions in
2168 @file{make-cccp.com} and @file{make-cc1.com}. The second generation
2169 compiler will be able to take advantage of many optimizations that must
2170 be suppressed when building with other compilers.
2173 Under previous versions of GNU CC, the generated code would occasionally
2174 give strange results when linked with the sharable @file{VAXCRTL} library.
2175 Now this should work.
2177 Even with this version, however, GNU CC itself should not be linked with
2178 the sharable @file{VAXCRTL}. The version of @code{qsort} in
2179 @file{VAXCRTL} has a bug (known to be present in VMS versions V4.6
2180 through V5.5) which causes the compiler to fail.
2182 The executables are generated by @file{make-cc1.com} and
2183 @file{make-cccp.com} use the object library version of @file{VAXCRTL} in
2184 order to make use of the @code{qsort} routine in @file{gcclib.olb}. If
2185 you wish to link the compiler executables with the shareable image
2186 version of @file{VAXCRTL}, you should edit the file @file{tm.h} (created
2187 by @file{vmsconfig.com}) to define the macro @code{QSORT_WORKAROUND}.
2189 @code{QSORT_WORKAROUND} is always defined when GNU CC is compiled with
2190 VAX C, to avoid a problem in case @file{gcclib.olb} is not yet
2194 @section @code{collect2}
2196 Many target systems do not have support in the assembler and linker for
2197 ``constructors''---initialization functions to be called before the
2198 official ``start'' of @code{main}. On such systems, GNU CC uses a
2199 utility called @code{collect2} to arrange to call these functions at
2202 The program @code{collect2} works by linking the program once and
2203 looking through the linker output file for symbols with particular names
2204 indicating they are constructor functions. If it finds any, it
2205 creates a new temporary @samp{.c} file containing a table of them,
2206 compiles it, and links the program a second time including that file.
2209 @cindex constructors, automatic calls
2210 The actual calls to the constructors are carried out by a subroutine
2211 called @code{__main}, which is called (automatically) at the beginning
2212 of the body of @code{main} (provided @code{main} was compiled with GNU
2213 CC). Calling @code{__main} is necessary, even when compiling C code, to
2214 allow linking C and C++ object code together. (If you use
2215 @samp{-nostdlib}, you get an unresolved reference to @code{__main},
2216 since it's defined in the standard GCC library. Include @samp{-lgcc} at
2217 the end of your compiler command line to resolve this reference.)
2219 The program @code{collect2} is installed as @code{ld} in the directory
2220 where the passes of the compiler are installed. When @code{collect2}
2221 needs to find the @emph{real} @code{ld}, it tries the following file
2226 @file{real-ld} in the directories listed in the compiler's search
2230 @file{real-ld} in the directories listed in the environment variable
2234 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2238 @file{ld} in the compiler's search directories, except that
2239 @code{collect2} will not execute itself recursively.
2242 @file{ld} in @code{PATH}.
2245 ``The compiler's search directories'' means all the directories where
2246 @code{gcc} searches for passes of the compiler. This includes
2247 directories that you specify with @samp{-B}.
2249 Cross-compilers search a little differently:
2253 @file{real-ld} in the compiler's search directories.
2256 @file{@var{target}-real-ld} in @code{PATH}.
2259 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2263 @file{ld} in the compiler's search directories.
2266 @file{@var{target}-ld} in @code{PATH}.
2269 @code{collect2} explicitly avoids running @code{ld} using the file name
2270 under which @code{collect2} itself was invoked. In fact, it remembers
2271 up a list of such names---in case one copy of @code{collect2} finds
2272 another copy (or version) of @code{collect2} installed as @code{ld} in a
2273 second place in the search path.
2275 @code{collect2} searches for the utilities @code{nm} and @code{strip}
2276 using the same algorithm as above for @code{ld}.
2279 @section Standard Header File Directories
2281 @code{GCC_INCLUDE_DIR} means the same thing for native and cross. It is
2282 where GNU CC stores its private include files, and also where GNU CC
2283 stores the fixed include files. A cross compiled GNU CC runs
2284 @code{fixincludes} on the header files in @file{$(tooldir)/include}.
2285 (If the cross compilation header files need to be fixed, they must be
2286 installed before GNU CC is built. If the cross compilation header files
2287 are already suitable for ANSI C and GNU CC, nothing special need be
2290 @code{GPLUS_INCLUDE_DIR} means the same thing for native and cross. It
2291 is where @code{g++} looks first for header files. @code{libg++}
2292 installs only target independent header files in that directory.
2294 @code{LOCAL_INCLUDE_DIR} is used only for a native compiler. It is
2295 normally @file{/usr/local/include}. GNU CC searches this directory so
2296 that users can install header files in @file{/usr/local/include}.
2298 @code{CROSS_INCLUDE_DIR} is used only for a cross compiler. GNU CC
2299 doesn't install anything there.
2301 @code{TOOL_INCLUDE_DIR} is used for both native and cross compilers. It
2302 is the place for other packages to install header files that GNU CC will
2303 use. For a cross-compiler, this is the equivalent of
2304 @file{/usr/include}. When you build a cross-compiler,
2305 @code{fixincludes} processes any header files in this directory.