1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995 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. @samp{sun3-bsd} also works, since
91 @file{configure} knows that the only BSD variant on a Sun 3 is SunOS.
93 You can specify a version number after any of the system types, and some
94 of the CPU types. In most cases, the version is irrelevant, and will be
95 ignored. So you might as well specify the version if you know it.
97 See @ref{Configurations}, for a list of supported configuration names and
98 notes on many of the configurations. You should check the notes in that
99 section before proceeding any further with the installation of GNU CC.
101 There are four additional options you can specify independently to
102 describe variant hardware and software configurations. These are
103 @samp{--with-gnu-as}, @samp{--with-gnu-ld}, @samp{--with-stabs} and
108 If you will use GNU CC with the GNU assembler (GAS), you should declare
109 this by using the @samp{--with-gnu-as} option when you run
112 Using this option does not install GAS. It only modifies the output of
113 GNU CC to work with GAS. Building and installing GAS is up to you.
115 Conversely, if you @emph{do not} wish to use GAS and do not specify
116 @samp{--with-gnu-as} when building GNU CC, it is up to you to make sure
117 that GAS is not installed. GNU CC searches for a program named
118 @code{as} in various directories; if the program it finds is GAS, then
119 it runs GAS. If you are not sure where GNU CC finds the assembler it is
120 using, try specifying @samp{-v} when you run it.
122 The systems where it makes a difference whether you use GAS are@*
123 @samp{hppa1.0-@var{any}-@var{any}}, @samp{hppa1.1-@var{any}-@var{any}},
124 @samp{i386-@var{any}-sysv}, @samp{i386-@var{any}-isc},@*
125 @samp{i860-@var{any}-bsd}, @samp{m68k-bull-sysv}, @samp{m68k-hp-hpux},
126 @samp{m68k-sony-bsd},@*
127 @samp{m68k-altos-sysv}, @samp{m68000-hp-hpux}, @samp{m68000-att-sysv},
128 @samp{@var{any}-lynx-lynxos}, 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
180 The @file{configure} script searches subdirectories of the source
181 directory for other compilers that are to be integrated into GNU CC.
182 The GNU compiler for C++, called G++ is in a subdirectory named
183 @file{cp}. @file{configure} inserts rules into @file{Makefile} to build
184 all of those compilers.
186 Here we spell out what files will be set up by @code{configure}. Normally
187 you need not be concerned with these files.
192 A file named @file{config.h} is created that contains a @samp{#include}
193 of the top-level config file for the machine you will run the compiler
194 on (@pxref{Config}). This file is responsible for defining information
195 about the host machine. It includes @file{tm.h}.
198 A file named @file{config.h} is created that contains a @samp{#include}
199 of the top-level config file for the machine you will run the compiler
200 on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting
201 GCC}). This file is responsible for defining information about the host
202 machine. It includes @file{tm.h}.
205 The top-level config file is located in the subdirectory @file{config}.
206 Its name is always @file{xm-@var{something}.h}; usually
207 @file{xm-@var{machine}.h}, but there are some exceptions.
209 If your system does not support symbolic links, you might want to
210 set up @file{config.h} to contain a @samp{#include} command which
211 refers to the appropriate file.
214 A file named @file{tconfig.h} is created which includes the top-level config
215 file for your target machine. This is used for compiling certain
216 programs to run on that machine.
219 A file named @file{tm.h} is created which includes the
220 machine-description macro file for your target machine. It should be in
221 the subdirectory @file{config} and its name is often
222 @file{@var{machine}.h}.
225 The command file @file{configure} also constructs the file
226 @file{Makefile} by adding some text to the template file
227 @file{Makefile.in}. The additional text comes from files in the
228 @file{config} directory, named @file{t-@var{target}} and
229 @file{x-@var{host}}. If these files do not exist, it means nothing
230 needs to be added for a given target or host.
234 The standard directory for installing GNU CC is @file{/usr/local/lib}.
235 If you want to install its files somewhere else, specify
236 @samp{--prefix=@var{dir}} when you run @file{configure}. Here @var{dir}
237 is a directory name to use instead of @file{/usr/local} for all purposes
238 with one exception: the directory @file{/usr/local/include} is searched
239 for header files no matter where you install the compiler. To override
240 this name, use the @code{--local-prefix} option below.
243 Specify @samp{--local-prefix=@var{dir}} if you want the compiler to
244 search directory @file{@var{dir}/include} for locally installed header
245 files @emph{instead} of @file{/usr/local/include}.
247 You should specify @samp{--local-prefix} @strong{only} if your site has
248 a different convention (not @file{/usr/local}) for where to put
251 @strong{Do not} specify @file{/usr} as the @samp{--local-prefix}! The
252 directory you use for @samp{--local-prefix} @strong{must not} contain
253 any of the system's standard header files. If it did contain them,
254 certain programs would be miscompiled (including GNU Emacs, on certain
255 targets), because this would override and nullify the header file
256 corrections made by the @code{fixincludes} script.
258 @cindex Bison parser generator
259 @cindex parser generator, Bison
261 Make sure the Bison parser generator is installed. (This is
262 unnecessary if the Bison output files @file{c-parse.c} and
263 @file{cexp.c} are more recent than @file{c-parse.y} and @file{cexp.y}
264 and you do not plan to change the @samp{.y} files.)
266 Bison versions older than Sept 8, 1988 will produce incorrect output
267 for @file{c-parse.c}.
270 If you have chosen a configuration for GNU CC which requires other GNU
271 tools (such as GAS or the GNU linker) instead of the standard system
272 tools, install the required tools in the build directory under the names
273 @file{as}, @file{ld} or whatever is appropriate. This will enable the
274 compiler to find the proper tools for compilation of the program
277 Alternatively, you can do subsequent compilation using a value of the
278 @code{PATH} environment variable such that the necessary GNU tools come
279 before the standard system tools.
282 Build the compiler. Just type @samp{make LANGUAGES=c} in the compiler
285 @samp{LANGUAGES=c} specifies that only the C compiler should be
286 compiled. The makefile normally builds compilers for all the supported
287 languages; currently, C, C++ and Objective C. However, C is the only
288 language that is sure to work when you build with other non-GNU C
289 compilers. In addition, building anything but C at this stage is a
292 In general, you can specify the languages to build by typing the
293 argument @samp{LANGUAGES="@var{list}"}, where @var{list} is one or more
294 words from the list @samp{c}, @samp{c++}, and @samp{objective-c}. If
295 you have any additional GNU compilers as subdirectories of the GNU CC
296 source directory, you may also specify their names in this list.
298 Ignore any warnings you may see about ``statement not reached'' in
299 @file{insn-emit.c}; they are normal. Also, warnings about ``unknown
300 escape sequence'' are normal in @file{genopinit.c} and perhaps some
301 other files. Likewise, you should ignore warnings about ``constant is
302 so large that it is unsigned'' in @file{insn-emit.c} and
303 @file{insn-recog.c} and a warning about a comparison always being zero
304 in @file{enquire.o}. Any other compilation errors may represent bugs in
305 the port to your machine or operating system, and
307 should be investigated and reported (@pxref{Bugs}).
310 should be investigated and reported.
313 Some commercial compilers fail to compile GNU CC because they have bugs
314 or limitations. For example, the Microsoft compiler is said to run out
315 of macro space. Some Ultrix compilers run out of expression space; then
316 you need to break up the statement where the problem happens.
319 If you are building a cross-compiler, stop here. @xref{Cross-Compiler}.
323 Move the first-stage object files and executables into a subdirectory
330 The files are moved into a subdirectory named @file{stage1}.
331 Once installation is complete, you may wish to delete these files
332 with @code{rm -r stage1}.
335 If you have chosen a configuration for GNU CC which requires other GNU
336 tools (such as GAS or the GNU linker) instead of the standard system
337 tools, install the required tools in the @file{stage1} subdirectory
338 under the names @file{as}, @file{ld} or whatever is appropriate. This
339 will enable the stage 1 compiler to find the proper tools in the
342 Alternatively, you can do subsequent compilation using a value of the
343 @code{PATH} environment variable such that the necessary GNU tools come
344 before the standard system tools.
347 Recompile the compiler with itself, with this command:
350 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2"
353 This is called making the stage 2 compiler.
355 The command shown above builds compilers for all the supported
356 languages. If you don't want them all, you can specify the languages to
357 build by typing the argument @samp{LANGUAGES="@var{list}"}. @var{list}
358 should contain one or more words from the list @samp{c}, @samp{c++},
359 @samp{objective-c}, and @samp{proto}. Separate the words with spaces.
360 @samp{proto} stands for the programs @code{protoize} and
361 @code{unprotoize}; they are not a separate language, but you use
362 @code{LANGUAGES} to enable or disable their installation.
364 If you are going to build the stage 3 compiler, then you might want to
365 build only the C language in stage 2.
367 Once you have built the stage 2 compiler, if you are short of disk
368 space, you can delete the subdirectory @file{stage1}.
370 On a 68000 or 68020 system lacking floating point hardware,
371 unless you have selected a @file{tm.h} file that expects by default
372 that there is no such hardware, do this instead:
375 make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2 -msoft-float"
379 If you wish to test the compiler by compiling it with itself one more
380 time, install any other necessary GNU tools (such as GAS or the GNU
381 linker) in the @file{stage2} subdirectory as you did in the
382 @file{stage1} subdirectory, then do this:
386 make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2"
390 This is called making the stage 3 compiler. Aside from the @samp{-B}
391 option, the compiler options should be the same as when you made the
392 stage 2 compiler. But the @code{LANGUAGES} option need not be the
393 same. The command shown above builds compilers for all the supported
394 languages; if you don't want them all, you can specify the languages to
395 build by typing the argument @samp{LANGUAGES="@var{list}"}, as described
398 If you do not have to install any additional GNU tools, you may use the
402 make bootstrap LANGUAGES=@var{language-list} BOOT_CFLAGS=@var{option-list}
406 instead of making @file{stage1}, @file{stage2}, and performing
407 the two compiler builds.
410 Then compare the latest object files with the stage 2 object
411 files---they ought to be identical, aside from time stamps (if any).
413 On some systems, meaningful comparison of object files is impossible;
414 they always appear ``different.'' This is currently true on Solaris and
415 some systems that use ELF object file format. On some versions of Irix
416 on SGI machines and DEC Unix (OSF/1) on Alpha systems, you will not be
417 able to compare the files without specifying @file{-save-temps}; see the
418 description of individual systems above to see if you get comparison
419 failures. You may have similar problems on other systems.
421 Use this command to compare the files:
427 This will mention any object files that differ between stage 2 and stage
428 3. Any difference, no matter how innocuous, indicates that the stage 2
429 compiler has compiled GNU CC incorrectly, and is therefore a potentially
431 serious bug which you should investigate and report (@pxref{Bugs}).
434 serious bug which you should investigate and report.
437 If your system does not put time stamps in the object files, then this
438 is a faster way to compare them (using the Bourne shell):
442 cmp $file stage2/$file
446 If you have built the compiler with the @samp{-mno-mips-tfile} option on
447 MIPS machines, you will not be able to compare the files.
450 Install the compiler driver, the compiler's passes and run-time support
451 with @samp{make install}. Use the same value for @code{CC},
452 @code{CFLAGS} and @code{LANGUAGES} that you used when compiling the
453 files that are being installed. One reason this is necessary is that
454 some versions of Make have bugs and recompile files gratuitously when
455 you do this step. If you use the same variable values, those files will
456 be recompiled properly.
458 For example, if you have built the stage 2 compiler, you can use the
462 make install CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O" LANGUAGES="@var{list}"
466 This copies the files @file{cc1}, @file{cpp} and @file{libgcc.a} to
467 files @file{cc1}, @file{cpp} and @file{libgcc.a} in the directory
468 @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}, which is where
469 the compiler driver program looks for them. Here @var{target} is the
470 target machine type specified when you ran @file{configure}, and
471 @var{version} is the version number of GNU CC. This naming scheme
472 permits various versions and/or cross-compilers to coexist.
474 This also copies the driver program @file{xgcc} into
475 @file{/usr/local/bin/gcc}, so that it appears in typical execution
478 On some systems, this command causes recompilation of some files. This
479 is usually due to bugs in @code{make}. You should either ignore this
480 problem, or use GNU Make.
482 @cindex @code{alloca} and SunOs
483 @strong{Warning: there is a bug in @code{alloca} in the Sun library. To
484 avoid this bug, be sure to install the executables of GNU CC that were
485 compiled by GNU CC. (That is, the executables from stage 2 or 3, not
486 stage 1.) They use @code{alloca} as a built-in function and never the
489 (It is usually better to install GNU CC executables from stage 2 or 3,
490 since they usually run faster than the ones compiled with some other
494 If you're going to use C++, it's likely that you need to also install
495 the libg++ distribution. It should be available from the same
496 place where you got the GNU C distribution. Just as GNU C does not
497 distribute a C runtime library, it also does not include a C++ run-time
498 library. All I/O functionality, special class libraries, etc., are
499 available in the libg++ distribution.
503 @section Configurations Supported by GNU CC
504 @cindex configurations supported by GNU CC
506 Here are the possible CPU types:
509 @c gmicro, alliant, spur and tahoe omitted since they don't work.
510 1750a, a29k, alpha, arm, c@var{n}, clipper, dsp16xx, elxsi, h8300,
511 hppa1.0, hppa1.1, i370, i386, i486, i586, i860, i960, m68000, m68k,
512 m88k, mips, mipsel, mips64, mips64el, ns32k, powerpc, powerpcle,
513 pyramid, romp, rs6000, sh, sparc, sparclite, sparc64, vax, we32k.
516 Here are the recognized company names. As you can see, customary
517 abbreviations are used rather than the longer official names.
519 @c What should be done about merlin, tek*, dolphin?
521 acorn, alliant, altos, apollo, att, bull,
522 cbm, convergent, convex, crds, dec, dg, dolphin,
523 elxsi, encore, harris, hitachi, hp, ibm, intergraph, isi,
524 mips, motorola, ncr, next, ns, omron, plexus,
525 sequent, sgi, sony, sun, tti, unicom, wrs.
528 The company name is meaningful only to disambiguate when the rest of
529 the information supplied is insufficient. You can omit it, writing
530 just @samp{@var{cpu}-@var{system}}, if it is not needed. For example,
531 @samp{vax-ultrix4.2} is equivalent to @samp{vax-dec-ultrix4.2}.
533 Here is a list of system types:
536 386bsd, aix, acis, amigados, aos, aout, bosx, bsd, clix, coff, ctix, cxux,
537 dgux, dynix, ebmon, ecoff, elf, esix, freebsd, hms, genix, gnu, gnu/linux,
538 hiux, hpux, iris, irix, isc, luna, lynxos, mach, minix, msdos, mvs,
539 netbsd, newsos, nindy, ns, osf, osfrose, ptx, riscix, riscos, rtu, sco, sim,
540 solaris, sunos, sym, sysv, udi, ultrix, unicos, uniplus, unos, vms, vsta,
541 vxworks, winnt, xenix.
545 You can omit the system type; then @file{configure} guesses the
546 operating system from the CPU and company.
548 You can add a version number to the system type; this may or may not
549 make a difference. For example, you can write @samp{bsd4.3} or
550 @samp{bsd4.4} to distinguish versions of BSD. In practice, the version
551 number is most needed for @samp{sysv3} and @samp{sysv4}, which are often
554 If you specify an impossible combination such as @samp{i860-dg-vms},
555 then you may get an error message from @file{configure}, or it may
556 ignore part of the information and do the best it can with the rest.
557 @file{configure} always prints the canonical name for the alternative
558 that it used. GNU CC does not support all possible alternatives.
560 Often a particular model of machine has a name. Many machine names are
561 recognized as aliases for CPU/company combinations. Thus, the machine
562 name @samp{sun3}, mentioned above, is an alias for @samp{m68k-sun}.
563 Sometimes we accept a company name as a machine name, when the name is
564 popularly used for a particular machine. Here is a table of the known
568 3300, 3b1, 3b@var{n}, 7300, altos3068, altos,
569 apollo68, att-7300, balance,
570 convex-c@var{n}, crds, decstation-3100,
571 decstation, delta, encore,
572 fx2800, gmicro, hp7@var{nn}, hp8@var{nn},
573 hp9k2@var{nn}, hp9k3@var{nn}, hp9k7@var{nn},
574 hp9k8@var{nn}, iris4d, iris, isi68,
575 m3230, magnum, merlin, miniframe,
576 mmax, news-3600, news800, news, next,
577 pbd, pc532, pmax, powerpc, powerpcle, ps2, risc-news,
578 rtpc, sun2, sun386i, sun386, sun3,
579 sun4, symmetry, tower-32, tower.
583 Remember that a machine name specifies both the cpu type and the company
585 If you want to install your own homemade configuration files, you can
586 use @samp{local} as the company name to access them. If you use
587 configuration @samp{@var{cpu}-local}, the configuration name
588 without the cpu prefix
589 is used to form the configuration file names.
591 Thus, if you specify @samp{m68k-local}, configuration uses
592 files @file{m68k.md}, @file{local.h}, @file{m68k.c},
593 @file{xm-local.h}, @file{t-local}, and @file{x-local}, all in the
594 directory @file{config/m68k}.
596 Here is a list of configurations that have special treatment or special
597 things you must know:
601 MIL-STD-1750A processors.
603 Starting with GCC 2.6.1, the MIL-STD-1750A cross configuration no longer
604 supports the Tektronix Assembler, but instead produces output for
605 @code{as1750}, an assembler/linker available under the GNU Public
606 License for the 1750A. Contact @emph{kellogg@@space.otn.dasa.de} for more
607 details on obtaining @samp{as1750}. A similarly licensed simulator for
608 the 1750A is available from same address.
610 You should ignore a fatal error during the building of libgcc (libgcc is
611 not yet implemented for the 1750A.)
613 The @code{as1750} assembler requires the file @file{ms1750.inc}, which is
614 found in the directory @file{config/1750a}.
616 GNU CC produced the same sections as the Fairchild F9450 C Compiler,
621 The program code section.
624 The read/write (RAM) data section.
627 The read-only (ROM) constants section.
630 Initialization section (code to copy KREL to SREL).
633 The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This
634 means that type `char' is represented with a 16-bit word per character.
635 The 1750A's "Load/Store Upper/Lower Byte" instructions are not used by
639 Systems using processors that implement the DEC Alpha architecture and
640 are running the DEC Unix (OSF/1) operating system, for example the DEC
641 Alpha AXP systems. (VMS on the Alpha is not currently supported by GNU
644 GNU CC writes a @samp{.verstamp} directive to the assembler output file
645 unless it is built as a cross-compiler. It gets the version to use from
646 the system header file @file{/usr/include/stamp.h}. If you install a
647 new version of DEC Unix, you should rebuild GCC to pick up the new version
650 Note that since the Alpha is a 64-bit architecture, cross-compilers from
651 32-bit machines will not generate code as efficient as that generated
652 when the compiler is running on a 64-bit machine because many
653 optimizations that depend on being able to represent a word on the
654 target in an integral value on the host cannot be performed. Building
655 cross-compilers on the Alpha for 32-bit machines has only been tested in
656 a few cases and may not work properly.
658 @code{make compare} may fail on old versions of DEC Unix unless you add
659 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
660 assembler input file is stored in the object file, and that makes
661 comparison fail if it differs between the @code{stage1} and
662 @code{stage2} compilations. The option @samp{-save-temps} forces a
663 fixed name to be used for the assembler input file, instead of a
664 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
665 unless the comparisons fail without that option. If you add
666 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
667 @samp{.s} files after each series of compilations.
669 GNU CC now supports both the native (ECOFF) debugging format used by DBX
670 and GDB and an encapsulated STABS format for use only with GDB. See the
671 discussion of the @samp{--with-stabs} option of @file{configure} above
672 for more information on these formats and how to select them.
674 There is a bug in DEC's assembler that produces incorrect line numbers
675 for ECOFF format when the @samp{.align} directive is used. To work
676 around this problem, GNU CC will not emit such alignment directives
677 while writing ECOFF format debugging information even if optimization is
678 being performed. Unfortunately, this has the very undesirable
679 side-effect that code addresses when @samp{-O} is specified are
680 different depending on whether or not @samp{-g} is also specified.
682 To avoid this behavior, specify @samp{-gstabs+} and use GDB instead of
683 DBX. DEC is now aware of this problem with the assembler and hopes to
684 provide a fix shortly.
687 Advanced RISC Machines ARM-family processors. These are often used in
688 embedded applications. There are no standard Unix configurations.
689 This configuration corresponds to the basic instruction sequences and will
690 produce a.out format object modules.
692 You may need to make a variant of the file @file{arm.h} for your particular
696 The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix. If
697 you are running a version of RISC iX prior to 1.2 then you must specify
698 the version number during configuration. Note that the assembler
699 shipped with RISC iX does not support stabs debugging information; a
700 new version of the assembler, with stabs support included, is now
701 available from Acorn.
704 AMD Am29k-family processors. These are normally used in embedded
705 applications. There are no standard Unix configurations.
707 corresponds to AMD's standard calling sequence and binary interface
708 and is compatible with other 29k tools.
710 You may need to make a variant of the file @file{a29k.h} for your
711 particular configuration.
714 AMD Am29050 used in a system running a variant of BSD Unix.
717 DECstations can support three different personalities: Ultrix,
718 DEC OSF/1, and OSF/rose. To configure GCC for these platforms
719 use the following configurations:
722 @item decstation-ultrix
723 Ultrix configuration.
725 @item decstation-osf1
726 Dec's version of OSF/1.
728 @item decstation-osfrose
729 Open Software Foundation reference port of OSF/1 which uses the
730 OSF/rose object file format instead of ECOFF. Normally, you
731 would not select this configuration.
734 The MIPS C compiler needs to be told to increase its table size
735 for switch statements with the @samp{-Wf,-XNg1500} option in
736 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
737 optimization option, you also need to use @samp{-Olimit 3000}.
738 Both of these options are automatically generated in the
739 @file{Makefile} that the shell script @file{configure} builds.
740 If you override the @code{CC} make variable and use the MIPS
741 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
743 @item elxsi-elxsi-bsd
744 The Elxsi's C compiler has known limitations that prevent it from
745 compiling GNU C. Please contact @code{mrs@@cygnus.com} for more details.
748 A port to the AT&T DSP1610 family of processors.
752 Alliant FX/8 computer. Note that the standard installed C compiler in
753 Concentrix 5.0 has a bug which prevent it from compiling GNU CC
754 correctly. You can patch the compiler bug as follows:
758 adb -w ./pcc - << EOF
763 Then you must use the @samp{-ip12} option when compiling GNU CC
764 with the patched compiler, as shown here:
767 make CC="./pcc -ip12" CFLAGS=-w
770 Note also that Alliant's version of DBX does not manage to work with the
775 The calling convention and structure layout has changed in release 2.6.
776 All code must be recompiled. The calling convention now passes the
777 first three arguments in function calls in registers. Structures are no
778 longer a multiple of 2 bytes.
781 There are several variants of the HP-PA processor which run a variety
782 of operating systems. GNU CC must be configured to use the correct
783 processor type and operating system, or GNU CC will not function correctly.
784 The easiest way to handle this problem is to @emph{not} specify a target
785 when configuring GNU CC, the @file{configure} script will try to automatically
786 determine the right processor type and operating system.
788 @samp{-g} does not work on HP-UX, since that system uses a peculiar
789 debugging format which GNU CC does not know about. However, @samp{-g}
790 will work if you also use GAS and GDB in conjunction with GCC. We
791 highly recommend using GAS for all HP-PA configurations.
793 You should be using GAS-2.6 (or later) along with GDB-4.16 (or later). These
794 can be retrieved from all the traditional GNU ftp archive sites.
796 GAS will need to be installed into a directory before @code{/bin},
797 @code{/usr/bin}, and @code{/usr/ccs/bin} in your search path. You
798 should install GAS before you build GNU CC.
800 To enable debugging, you must configure GNU CC with the @samp{--with-gnu-as}
801 option before building.
804 This port is very preliminary and has many known bugs. We hope to
805 have a higher-quality port for this machine soon.
807 @item i386-*-linuxoldld
808 Use this configuration to generate a.out binaries on Linux-based GNU
809 systems, if you do not have gas/binutils version 2.5.2 or later
810 installed. This is an obsolete configuration.
812 @item i386-*-linuxaout
813 Use this configuration to generate a.out binaries on Linux-based GNU
814 systems. This configuration is being superseded. You must use
815 gas/binutils version 2.5.2 or later.
818 Use this configuration to generate ELF binaries on Linux-based GNU
819 systems. You must use gas/binutils version 2.5.2 or later.
822 Compilation with RCC is recommended. Also, it may be a good idea to
823 link with GNU malloc instead of the malloc that comes with the system.
825 @item i386-*-sco3.2v4
826 Use this configuration for SCO release 3.2 version 4.
829 It may be a good idea to link with GNU malloc instead of the malloc that
830 comes with the system.
832 In ISC version 4.1, @file{sed} core dumps when building
833 @file{deduced.h}. Use the version of @file{sed} from version 4.0.
836 It may be good idea to link with GNU malloc instead of the malloc that
837 comes with the system.
840 You need to use GAS version 2.1 or later, and and LD from
841 GNU binutils version 2.2 or later.
843 @item i386-sequent-bsd
844 Go to the Berkeley universe before compiling. In addition, you probably
845 need to create a file named @file{string.h} containing just one line:
846 @samp{#include <strings.h>}.
848 @item i386-sequent-ptx1*
849 Sequent DYNIX/ptx 1.x.
851 @item i386-sequent-ptx2*
852 Sequent DYNIX/ptx 2.x.
854 @item i386-sun-sunos4
855 You may find that you need another version of GNU CC to begin
856 bootstrapping with, since the current version when built with the
857 system's own compiler seems to get an infinite loop compiling part of
858 @file{libgcc2.c}. GNU CC version 2 compiled with GNU CC (any version)
859 seems not to have this problem.
861 See @ref{Sun Install}, for information on installing GNU CC on Sun
864 @item i[345]86-*-winnt3.5
865 This version requires a GAS that has not let been released. Until it
866 is, you can get a prebuilt binary version via anonymous ftp from
867 @file{cs.washington.edu:pub/gnat} or @file{cs.nyu.edu:pub/gnat}. You
868 must also use the Microsoft header files from the Windows NT 3.5 SDK.
869 Find these on the CDROM in the @file{/mstools/h} directory dated 9/4/94. You
870 must use a fixed version of Microsoft linker made especially for NT 3.5,
871 which is also is available on the NT 3.5 SDK CDROM. If you do not have
872 this linker, can you also use the linker from Visual C/C++ 1.0 or 2.0.
874 Installing GNU CC for NT builds a wrapper linker, called @file{ld.exe},
875 which mimics the behaviour of Unix @file{ld} in the specification of
876 libraries (@samp{-L} and @samp{-l}). @file{ld.exe} looks for both Unix
877 and Microsoft named libraries. For example, if you specify
878 @samp{-lfoo}, @file{ld.exe} will look first for @file{libfoo.a}
879 and then for @file{foo.lib}.
881 You may install GNU CC for Windows NT in one of two ways, depending on
882 whether or not you have a Unix-like shell and various Unix-like
887 If you do not have a Unix-like shell and few Unix-like utilities, you
888 will use a DOS style batch script called @file{configure.bat}. Invoke
889 it as @code{configure winnt} from an MSDOS console window or from the
890 program manager dialog box. @file{configure.bat} assumes you have
891 already installed and have in your path a Unix-like @file{sed} program
892 which is used to create a working @file{Makefile} from @file{Makefile.in}.
894 @file{Makefile} uses the Microsoft Nmake program maintenance utility and
895 the Visual C/C++ V8.00 compiler to build GNU CC. You need only have the
896 utilities @file{sed} and @file{touch} to use this installation method,
897 which only automatically builds the compiler itself. You must then
898 examine what @file{fixinc.winnt} does, edit the header files by hand and
899 build @file{libgcc.a} manually.
902 The second type of installation assumes you are running a Unix-like
903 shell, have a complete suite of Unix-like utilities in your path, and
904 have a previous version of GNU CC already installed, either through
905 building it via the above installation method or acquiring a pre-built
906 binary. In this case, use the @file{configure} script in the normal
910 @item i860-intel-osf1
913 If you have version 1.0 of the operating system, you need to take
914 special steps to build GNU CC due to peculiarities of the system. Newer
915 system versions have no problem. See the section `Installation Problems'
916 in the GNU CC Manual.
919 If you have version 1.0 of the operating system,
920 see @ref{Installation Problems}, for special things you need to do to
921 compensate for peculiarities in the system.
925 LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as
926 @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}.
927 You can tell GNU CC to use the GNU assembler and linker, by specifying
928 @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce
929 COFF format object files and executables; otherwise GNU CC will use the
930 installed tools, which produce a.out format executables.
933 HP 9000 series 200 running BSD. Note that the C compiler that comes
934 with this system cannot compile GNU CC; contact @code{law@@cs.utah.edu}
935 to get binaries of GNU CC for bootstrapping.
938 Altos 3068. You must use the GNU assembler, linker and debugger.
939 Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}.
942 AT&T 3b1, a.k.a. 7300 PC. Special procedures are needed to compile GNU
943 CC with this machine's standard C compiler, due to bugs in that
944 compiler. You can bootstrap it more easily with
945 previous versions of GNU CC if you have them.
947 Installing GNU CC on the 3b1 is difficult if you do not already have
948 GNU CC running, due to bugs in the installed C compiler. However,
949 the following procedure might work. We are unable to test it.
953 Comment out the @samp{#include "config.h"} line on line 37 of
954 @file{cccp.c} and do @samp{make cpp}. This makes a preliminary version
958 Save the old @file{/lib/cpp} and copy the preliminary GNU cpp to that
962 Undo your change in @file{cccp.c}, or reinstall the original version,
963 and do @samp{make cpp} again.
966 Copy this final version of GNU cpp into @file{/lib/cpp}.
970 Replace every occurrence of @code{obstack_free} in the file
971 @file{tree.c} with @code{_obstack_free}.
974 Run @code{make} to get the first-stage GNU CC.
977 Reinstall the original version of @file{/lib/cpp}.
980 Now you can compile GNU CC with itself and install it in the normal
985 Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works
986 either with native assembler or GNU assembler. You can use
987 GNU assembler with native coff generation by providing @samp{--with-gnu-as} to
988 the configure script or use GNU assembler with dbx-in-coff encapsulation
989 by providing @samp{--with-gnu-as --stabs}. For any problem with native
990 assembler or for availability of the DPX/2 port of GAS, contact
991 @code{F.Pierresteguy@@frcl.bull.fr}.
994 Use @samp{configure unos} for building on Unos.
996 The Unos assembler is named @code{casm} instead of @code{as}. For some
997 strange reason linking @file{/bin/as} to @file{/bin/casm} changes the
998 behavior, and does not work. So, when installing GNU CC, you should
999 install the following script as @file{as} in the subdirectory where
1000 the passes of GCC are installed:
1007 The default Unos library is named @file{libunos.a} instead of
1008 @file{libc.a}. To allow GNU CC to function, either change all
1009 references to @samp{-lc} in @file{gcc.c} to @samp{-lunos} or link
1010 @file{/lib/libc.a} to @file{/lib/libunos.a}.
1012 @cindex @code{alloca}, for Unos
1013 When compiling GNU CC with the standard compiler, to overcome bugs in
1014 the support of @code{alloca}, do not use @samp{-O} when making stage 2.
1015 Then use the stage 2 compiler with @samp{-O} to make the stage 3
1016 compiler. This compiler will have the same characteristics as the usual
1017 stage 2 compiler on other systems. Use it to make a stage 4 compiler
1018 and compare that with stage 3 to verify proper compilation.
1020 (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in
1021 the comments there will make the above paragraph superfluous. Please
1022 inform us of whether this works.)
1024 Unos uses memory segmentation instead of demand paging, so you will need
1025 a lot of memory. 5 Mb is barely enough if no other tasks are running.
1026 If linking @file{cc1} fails, try putting the object files into a library
1027 and linking from that library.
1030 HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in
1031 the assembler that prevents compilation of GNU CC. To fix it, get patch
1034 In addition, if you wish to use gas @samp{--with-gnu-as} you must use
1035 gas version 2.1 or later, and you must use the GNU linker version 2.1 or
1036 later. Earlier versions of gas relied upon a program which converted the
1037 gas output into the native HP/UX format, but that program has not been
1038 kept up to date. gdb does not understand that native HP/UX format, so
1039 you must use gas if you wish to use gdb.
1042 Sun 3. We do not provide a configuration file to use the Sun FPA by
1043 default, because programs that establish signal handlers for floating
1044 point traps inherently cannot work with the FPA.
1046 See @ref{Sun Install}, for information on installing GNU CC on Sun
1050 Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port.
1051 These systems tend to use the Green Hills C, revision 1.8.5, as the
1052 standard C compiler. There are apparently bugs in this compiler that
1053 result in object files differences between stage 2 and stage 3. If this
1054 happens, make the stage 4 compiler and compare it to the stage 3
1055 compiler. If the stage 3 and stage 4 object files are identical, this
1056 suggests you encountered a problem with the standard C compiler; the
1057 stage 3 and 4 compilers may be usable.
1059 It is best, however, to use an older version of GNU CC for bootstrapping
1063 Motorola m88k running DG/UX. To build 88open BCS native or cross
1064 compilers on DG/UX, specify the configuration name as
1065 @samp{m88k-*-dguxbcs} and build in the 88open BCS software development
1066 environment. To build ELF native or cross compilers on DG/UX, specify
1067 @samp{m88k-*-dgux} and build in the DG/UX ELF development environment.
1068 You set the software development environment by issuing
1069 @samp{sde-target} command and specifying either @samp{m88kbcs} or
1070 @samp{m88kdguxelf} as the operand.
1072 If you do not specify a configuration name, @file{configure} guesses the
1073 configuration based on the current software development environment.
1075 @item m88k-tektronix-sysv3
1076 Tektronix XD88 running UTekV 3.2e. Do not turn on
1077 optimization while building stage1 if you bootstrap with
1078 the buggy Green Hills compiler. Also, The bundled LAI
1079 System V NFS is buggy so if you build in an NFS mounted
1080 directory, start from a fresh reboot, or avoid NFS all together.
1081 Otherwise you may have trouble getting clean comparisons
1085 MIPS machines running the MIPS operating system in BSD mode. It's
1086 possible that some old versions of the system lack the functions
1087 @code{memcpy}, @code{memcmp}, and @code{memset}. If your system lacks
1088 these, you must remove or undo the definition of
1089 @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}.
1091 The MIPS C compiler needs to be told to increase its table size
1092 for switch statements with the @samp{-Wf,-XNg1500} option in
1093 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1094 optimization option, you also need to use @samp{-Olimit 3000}.
1095 Both of these options are automatically generated in the
1096 @file{Makefile} that the shell script @file{configure} builds.
1097 If you override the @code{CC} make variable and use the MIPS
1098 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1100 @item mips-mips-riscos*
1101 The MIPS C compiler needs to be told to increase its table size
1102 for switch statements with the @samp{-Wf,-XNg1500} option in
1103 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1104 optimization option, you also need to use @samp{-Olimit 3000}.
1105 Both of these options are automatically generated in the
1106 @file{Makefile} that the shell script @file{configure} builds.
1107 If you override the @code{CC} make variable and use the MIPS
1108 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1110 MIPS computers running RISC-OS can support four different
1111 personalities: default, BSD 4.3, System V.3, and System V.4
1112 (older versions of RISC-OS don't support V.4). To configure GCC
1113 for these platforms use the following configurations:
1116 @item mips-mips-riscos@code{rev}
1117 Default configuration for RISC-OS, revision @code{rev}.
1119 @item mips-mips-riscos@code{rev}bsd
1120 BSD 4.3 configuration for RISC-OS, revision @code{rev}.
1122 @item mips-mips-riscos@code{rev}sysv4
1123 System V.4 configuration for RISC-OS, revision @code{rev}.
1125 @item mips-mips-riscos@code{rev}sysv
1126 System V.3 configuration for RISC-OS, revision @code{rev}.
1129 The revision @code{rev} mentioned above is the revision of
1130 RISC-OS to use. You must reconfigure GCC when going from a
1131 RISC-OS revision 4 to RISC-OS revision 5. This has the effect of
1133 @ifclear INSTALLONLY
1134 bug (see @ref{Installation Problems}, for more details).
1141 In order to compile GCC on an SGI running IRIX 4, the "c.hdr.lib"
1142 option must be installed from the CD-ROM supplied from Silicon Graphics.
1143 This is found on the 2nd CD in release 4.0.1.
1145 In order to compile GCC on an SGI running IRIX 5, the "compiler_dev.hdr"
1146 subsystem must be installed from the IDO CD-ROM supplied by Silicon
1149 @code{make compare} may fail on version 5 of IRIX unless you add
1150 @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the
1151 assembler input file is stored in the object file, and that makes
1152 comparison fail if it differs between the @code{stage1} and
1153 @code{stage2} compilations. The option @samp{-save-temps} forces a
1154 fixed name to be used for the assembler input file, instead of a
1155 randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps}
1156 unless the comparisons fail without that option. If you do you
1157 @samp{-save-temps}, you will have to manually delete the @samp{.i} and
1158 @samp{.s} files after each series of compilations.
1160 The MIPS C compiler needs to be told to increase its table size
1161 for switch statements with the @samp{-Wf,-XNg1500} option in
1162 order to compile @file{cp/parse.c}. If you use the @samp{-O2}
1163 optimization option, you also need to use @samp{-Olimit 3000}.
1164 Both of these options are automatically generated in the
1165 @file{Makefile} that the shell script @file{configure} builds.
1166 If you override the @code{CC} make variable and use the MIPS
1167 compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}.
1169 On Irix version 4.0.5F, and perhaps on some other versions as well,
1170 there is an assembler bug that reorders instructions incorrectly. To
1171 work around it, specify the target configuration
1172 @samp{mips-sgi-irix4loser}. This configuration inhibits assembler
1175 In a compiler configured with target @samp{mips-sgi-irix4}, you can turn
1176 off assembler optimization by using the @samp{-noasmopt} option. This
1177 compiler option passes the option @samp{-O0} to the assembler, to
1180 The @samp{-noasmopt} option can be useful for testing whether a problem
1181 is due to erroneous assembler reordering. Even if a problem does not go
1182 away with @samp{-noasmopt}, it may still be due to assembler
1183 reordering---perhaps GNU CC itself was miscompiled as a result.
1185 To enable debugging under Irix 5, you must use GNU as 2.5 or later,
1186 and use the @samp{--with-gnu-as} configure option when configuring gcc.
1187 GNU as is distributed as part of the binutils package.
1189 @item mips-sony-sysv
1190 Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which
1191 uses ELF instead of COFF). Support for 5.0.2 will probably be provided
1192 soon by volunteers. In particular, the linker does not like the
1193 code generated by GCC when shared libraries are linked in.
1196 Encore ns32000 system. Encore systems are supported only under BSD.
1199 National Semiconductor ns32000 system. Genix has bugs in @code{alloca}
1200 and @code{malloc}; you must get the compiled versions of these from GNU
1204 Go to the Berkeley universe before compiling. In addition, you probably
1205 need to create a file named @file{string.h} containing just one line:
1206 @samp{#include <strings.h>}.
1209 UTEK ns32000 system (``merlin''). The C compiler that comes with this
1210 system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get
1211 binaries of GNU CC for bootstrapping.
1215 The only operating systems supported for the IBM RT PC are AOS and
1216 MACH. GNU CC does not support AIX running on the RT. We recommend you
1217 compile GNU CC with an earlier version of itself; if you compile GNU CC
1218 with @code{hc}, the Metaware compiler, it will work, but you will get
1219 mismatches between the stage 2 and stage 3 compilers in various files.
1220 These errors are minor differences in some floating-point constants and
1221 can be safely ignored; the stage 3 compiler is correct.
1224 @itemx powerpc-*-aix
1225 Various early versions of each release of the IBM XLC compiler will not
1226 bootstrap GNU CC. Symptoms include differences between the stage2 and
1227 stage3 object files, and errors when compiling @file{libgcc.a} or
1228 @file{enquire}. Known problematic releases include: xlc-1.2.1.8,
1229 xlc-1.3.0.0 (distributed with AIX 3.2.5), and xlc-1.3.0.19. Both
1230 xlc-1.2.1.28 and xlc-1.3.0.24 (PTF 432238) are known to produce working
1231 versions of GNU CC, but most other recent releases correctly bootstrap
1232 GNU CC. Also, releases of AIX prior to AIX 3.2.4 include a version of
1233 the IBM assembler which does not accept debugging directives: assembler
1234 updates are available as PTFs. Also, if you are using AIX 3.2.5 or
1235 greater and the GNU assembler, you must have a version modified after
1236 October 16th, 1995 in order for the GNU C compiler to build. See the
1237 file @file{README.RS6000} for more details on of these problems.
1239 GNU CC does not yet support the 64-bit PowerPC instructions.
1241 Objective C does not work on this architecture because it makes assumptions
1242 that are incompatible with the calling conventions.
1244 AIX on the RS/6000 provides support (NLS) for environments outside of
1245 the United States. Compilers and assemblers use NLS to support
1246 locale-specific representations of various objects including
1247 floating-point numbers ("." vs "," for separating decimal fractions).
1248 There have been problems reported where the library linked with GNU CC
1249 does not produce the same floating-point formats that the assembler
1250 accepts. If you have this problem, set the LANG environment variable to
1253 Due to changes in the way that GNU CC invokes the binder (linker) for AIX
1254 4.1, you may now receive warnings of duplicate symbols from the link step
1255 that were not reported before. The assembly files generated by GNU CC for
1256 AIX have always included multiple symbol definitions for certain global
1257 variable and function declarations in the original program. The warnings
1258 should not prevent the linker from producing a correct library or runnable
1262 @itemx powerpc-*-sysv4
1263 PowerPC system in big endian mode, running System V.4.
1265 This configuration is currently under development.
1267 @item powerpc-*-eabiaix
1268 Embedded PowerPC system in big endian mode with -mcall-aix selected as
1269 the default. This system is currently under development.
1271 @item powerpc-*-eabisim
1272 Embedded PowerPC system in big endian mode for use in running under the
1273 PSIM simulator. This system is currently under development.
1275 @item powerpc-*-eabi
1276 Embedded PowerPC system in big endian mode.
1278 This configuration is currently under development.
1280 @item powerpcle-*-elf
1281 @itemx powerpcle-*-sysv4
1282 PowerPC system in little endian mode, running System V.4.
1284 This configuration is currently under development.
1286 @itemx powerpcle-*-sysv4
1287 Embedded PowerPC system in little endian mode.
1289 This system is currently under development.
1291 @item powerpcle-*-eabisim
1292 Embedded PowerPC system in little endian mode for use in running under
1295 This system is currently under development.
1297 @itemx powerpcle-*-eabi
1298 Embedded PowerPC system in little endian mode.
1300 This configuration is currently under development.
1302 @item vax-dec-ultrix
1303 Don't try compiling with Vax C (@code{vcc}). It produces incorrect code
1304 in some cases (for example, when @code{alloca} is used).
1306 Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of
1307 an internal table size limitation in that compiler. To avoid this
1308 problem, compile just the GNU C compiler first, and use it to recompile
1309 building all the languages that you want to run.
1312 See @ref{Sun Install}, for information on installing GNU CC on Sun
1316 See @ref{VMS Install}, for details on how to install GNU CC on VMS.
1319 These computers are also known as the 3b2, 3b5, 3b20 and other similar
1320 names. (However, the 3b1 is actually a 68000; see
1321 @ref{Configurations}.)
1323 Don't use @samp{-g} when compiling with the system's compiler. The
1324 system's linker seems to be unable to handle such a large program with
1325 debugging information.
1327 The system's compiler runs out of capacity when compiling @file{stmt.c}
1328 in GNU CC. You can work around this by building @file{cpp} in GNU CC
1329 first, then use that instead of the system's preprocessor with the
1330 system's C compiler to compile @file{stmt.c}. Here is how:
1333 mv /lib/cpp /lib/cpp.att
1335 echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp
1339 The system's compiler produces bad code for some of the GNU CC
1340 optimization files. So you must build the stage 2 compiler without
1341 optimization. Then build a stage 3 compiler with optimization.
1342 That executable should work. Here are the necessary commands:
1345 make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g"
1347 make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O"
1350 You may need to raise the ULIMIT setting to build a C++ compiler,
1351 as the file @file{cc1plus} is larger than one megabyte.
1355 @section Compilation in a Separate Directory
1356 @cindex other directory, compilation in
1357 @cindex compilation in a separate directory
1358 @cindex separate directory, compilation in
1360 If you wish to build the object files and executables in a directory
1361 other than the one containing the source files, here is what you must
1366 Make sure you have a version of Make that supports the @code{VPATH}
1367 feature. (GNU Make supports it, as do Make versions on most BSD
1371 If you have ever run @file{configure} in the source directory, you must undo
1372 the configuration. Do this by running:
1379 Go to the directory in which you want to build the compiler before
1380 running @file{configure}:
1387 On systems that do not support symbolic links, this directory must be
1388 on the same file system as the source code directory.
1391 Specify where to find @file{configure} when you run it:
1394 ../gcc/configure @dots{}
1397 This also tells @code{configure} where to find the compiler sources;
1398 @code{configure} takes the directory from the file name that was used to
1399 invoke it. But if you want to be sure, you can specify the source
1400 directory with the @samp{--srcdir} option, like this:
1403 ../gcc/configure --srcdir=../gcc @var{other options}
1406 The directory you specify with @samp{--srcdir} need not be the same
1407 as the one that @code{configure} is found in.
1410 Now, you can run @code{make} in that directory. You need not repeat the
1411 configuration steps shown above, when ordinary source files change. You
1412 must, however, run @code{configure} again when the configuration files
1413 change, if your system does not support symbolic links.
1415 @node Cross-Compiler
1416 @section Building and Installing a Cross-Compiler
1417 @cindex cross-compiler, installation
1419 GNU CC can function as a cross-compiler for many machines, but not all.
1423 Cross-compilers for the Mips as target using the Mips assembler
1424 currently do not work, because the auxiliary programs
1425 @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on
1426 anything but a Mips. It does work to cross compile for a Mips
1427 if you use the GNU assembler and linker.
1430 Cross-compilers between machines with different floating point formats
1431 have not all been made to work. GNU CC now has a floating point
1432 emulator with which these can work, but each target machine description
1433 needs to be updated to take advantage of it.
1436 Cross-compilation between machines of different word sizes is
1437 somewhat problematic and sometimes does not work.
1440 Since GNU CC generates assembler code, you probably need a
1441 cross-assembler that GNU CC can run, in order to produce object files.
1442 If you want to link on other than the target machine, you need a
1443 cross-linker as well. You also need header files and libraries suitable
1444 for the target machine that you can install on the host machine.
1447 * Steps of Cross:: Using a cross-compiler involves several steps
1448 that may be carried out on different machines.
1449 * Configure Cross:: Configuring a cross-compiler.
1450 * Tools and Libraries:: Where to put the linker and assembler, and the C library.
1451 * Cross Headers:: Finding and installing header files
1452 for a cross-compiler.
1453 * Cross Runtime:: Supplying arithmetic runtime routines (@file{libgcc1.a}).
1454 * Build Cross:: Actually compiling the cross-compiler.
1457 @node Steps of Cross
1458 @subsection Steps of Cross-Compilation
1460 To compile and run a program using a cross-compiler involves several
1465 Run the cross-compiler on the host machine to produce assembler files
1466 for the target machine. This requires header files for the target
1470 Assemble the files produced by the cross-compiler. You can do this
1471 either with an assembler on the target machine, or with a
1472 cross-assembler on the host machine.
1475 Link those files to make an executable. You can do this either with a
1476 linker on the target machine, or with a cross-linker on the host
1477 machine. Whichever machine you use, you need libraries and certain
1478 startup files (typically @file{crt@dots{}.o}) for the target machine.
1481 It is most convenient to do all of these steps on the same host machine,
1482 since then you can do it all with a single invocation of GNU CC. This
1483 requires a suitable cross-assembler and cross-linker. For some targets,
1484 the GNU assembler and linker are available.
1486 @node Configure Cross
1487 @subsection Configuring a Cross-Compiler
1489 To build GNU CC as a cross-compiler, you start out by running
1490 @file{configure}. Use the @samp{--target=@var{target}} to specify the
1491 target type. If @file{configure} was unable to correctly identify the
1492 system you are running on, also specify the @samp{--build=@var{build}}
1493 option. For example, here is how to configure for a cross-compiler that
1494 produces code for an HP 68030 system running BSD on a system that
1495 @file{configure} can correctly identify:
1498 ./configure --target=m68k-hp-bsd4.3
1501 @node Tools and Libraries
1502 @subsection Tools and Libraries for a Cross-Compiler
1504 If you have a cross-assembler and cross-linker available, you should
1505 install them now. Put them in the directory
1506 @file{/usr/local/@var{target}/bin}. Here is a table of the tools
1507 you should put in this directory:
1511 This should be the cross-assembler.
1514 This should be the cross-linker.
1517 This should be the cross-archiver: a program which can manipulate
1518 archive files (linker libraries) in the target machine's format.
1521 This should be a program to construct a symbol table in an archive file.
1524 The installation of GNU CC will find these programs in that directory,
1525 and copy or link them to the proper place to for the cross-compiler to
1526 find them when run later.
1528 The easiest way to provide these files is to build the Binutils package
1529 and GAS. Configure them with the same @samp{--host} and @samp{--target}
1530 options that you use for configuring GNU CC, then build and install
1531 them. They install their executables automatically into the proper
1532 directory. Alas, they do not support all the targets that GNU CC
1535 If you want to install libraries to use with the cross-compiler, such as
1536 a standard C library, put them in the directory
1537 @file{/usr/local/@var{target}/lib}; installation of GNU CC copies all
1538 all the files in that subdirectory into the proper place for GNU CC to
1539 find them and link with them. Here's an example of copying some
1540 libraries from a target machine:
1543 ftp @var{target-machine}
1544 lcd /usr/local/@var{target}/lib
1554 The precise set of libraries you'll need, and their locations on
1555 the target machine, vary depending on its operating system.
1558 Many targets require ``start files'' such as @file{crt0.o} and
1559 @file{crtn.o} which are linked into each executable; these too should be
1560 placed in @file{/usr/local/@var{target}/lib}. There may be several
1561 alternatives for @file{crt0.o}, for use with profiling or other
1562 compilation options. Check your target's definition of
1563 @code{STARTFILE_SPEC} to find out what start files it uses.
1564 Here's an example of copying these files from a target machine:
1567 ftp @var{target-machine}
1568 lcd /usr/local/@var{target}/lib
1578 @subsection @file{libgcc.a} and Cross-Compilers
1580 Code compiled by GNU CC uses certain runtime support functions
1581 implicitly. Some of these functions can be compiled successfully with
1582 GNU CC itself, but a few cannot be. These problem functions are in the
1583 source file @file{libgcc1.c}; the library made from them is called
1586 When you build a native compiler, these functions are compiled with some
1587 other compiler--the one that you use for bootstrapping GNU CC.
1588 Presumably it knows how to open code these operations, or else knows how
1589 to call the run-time emulation facilities that the machine comes with.
1590 But this approach doesn't work for building a cross-compiler. The
1591 compiler that you use for building knows about the host system, not the
1594 So, when you build a cross-compiler you have to supply a suitable
1595 library @file{libgcc1.a} that does the job it is expected to do.
1597 To compile @file{libgcc1.c} with the cross-compiler itself does not
1598 work. The functions in this file are supposed to implement arithmetic
1599 operations that GNU CC does not know how to open code for your target
1600 machine. If these functions are compiled with GNU CC itself, they
1601 will compile into infinite recursion.
1603 On any given target, most of these functions are not needed. If GNU CC
1604 can open code an arithmetic operation, it will not call these functions
1605 to perform the operation. It is possible that on your target machine,
1606 none of these functions is needed. If so, you can supply an empty
1607 library as @file{libgcc1.a}.
1609 Many targets need library support only for multiplication and division.
1610 If you are linking with a library that contains functions for
1611 multiplication and division, you can tell GNU CC to call them directly
1612 by defining the macros @code{MULSI3_LIBCALL}, and the like. These
1613 macros need to be defined in the target description macro file. For
1614 some targets, they are defined already. This may be sufficient to
1615 avoid the need for libgcc1.a; if so, you can supply an empty library.
1617 Some targets do not have floating point instructions; they need other
1618 functions in @file{libgcc1.a}, which do floating arithmetic.
1619 Recent versions of GNU CC have a file which emulates floating point.
1620 With a certain amount of work, you should be able to construct a
1621 floating point emulator that can be used as @file{libgcc1.a}. Perhaps
1622 future versions will contain code to do this automatically and
1623 conveniently. That depends on whether someone wants to implement it.
1625 Some embedded targets come with all the necessary @file{libgcc1.a}
1626 routines written in C or assembler. These targets build
1627 @file{libgcc1.a} automatically and you do not need to do anything
1628 special for them. Other embedded targets do not need any
1629 @file{libgcc1.a} routines since all the necessary operations are
1630 supported by the hardware.
1632 If your target system has another C compiler, you can configure GNU CC
1633 as a native compiler on that machine, build just @file{libgcc1.a} with
1634 @samp{make libgcc1.a} on that machine, and use the resulting file with
1635 the cross-compiler. To do this, execute the following on the target
1639 cd @var{target-build-dir}
1640 ./configure --host=sparc --target=sun3
1645 And then this on the host machine:
1648 ftp @var{target-machine}
1650 cd @var{target-build-dir}
1655 Another way to provide the functions you need in @file{libgcc1.a} is to
1656 define the appropriate @code{perform_@dots{}} macros for those
1657 functions. If these definitions do not use the C arithmetic operators
1658 that they are meant to implement, you should be able to compile them
1659 with the cross-compiler you are building. (If these definitions already
1660 exist for your target file, then you are all set.)
1662 To build @file{libgcc1.a} using the perform macros, use
1663 @samp{LIBGCC1=libgcc1.a OLDCC=./xgcc} when building the compiler.
1664 Otherwise, you should place your replacement library under the name
1665 @file{libgcc1.a} in the directory in which you will build the
1666 cross-compiler, before you run @code{make}.
1669 @subsection Cross-Compilers and Header Files
1671 If you are cross-compiling a standalone program or a program for an
1672 embedded system, then you may not need any header files except the few
1673 that are part of GNU CC (and those of your program). However, if you
1674 intend to link your program with a standard C library such as
1675 @file{libc.a}, then you probably need to compile with the header files
1676 that go with the library you use.
1678 The GNU C compiler does not come with these files, because (1) they are
1679 system-specific, and (2) they belong in a C library, not in a compiler.
1681 If the GNU C library supports your target machine, then you can get the
1682 header files from there (assuming you actually use the GNU library when
1683 you link your program).
1685 If your target machine comes with a C compiler, it probably comes with
1686 suitable header files also. If you make these files accessible from the host
1687 machine, the cross-compiler can use them also.
1689 Otherwise, you're on your own in finding header files to use when
1692 When you have found suitable header files, put them in
1693 @file{/usr/local/@var{target}/include}, before building the cross
1694 compiler. Then installation will run fixincludes properly and install
1695 the corrected versions of the header files where the compiler will use
1698 Provide the header files before you build the cross-compiler, because
1699 the build stage actually runs the cross-compiler to produce parts of
1700 @file{libgcc.a}. (These are the parts that @emph{can} be compiled with
1701 GNU CC.) Some of them need suitable header files.
1703 Here's an example showing how to copy the header files from a target
1704 machine. On the target machine, do this:
1707 (cd /usr/include; tar cf - .) > tarfile
1710 Then, on the host machine, do this:
1713 ftp @var{target-machine}
1714 lcd /usr/local/@var{target}/include
1721 @subsection Actually Building the Cross-Compiler
1723 Now you can proceed just as for compiling a single-machine compiler
1724 through the step of building stage 1. If you have not provided some
1725 sort of @file{libgcc1.a}, then compilation will give up at the point
1726 where it needs that file, printing a suitable error message. If you
1727 do provide @file{libgcc1.a}, then building the compiler will automatically
1728 compile and link a test program called @file{libgcc1-test}; if you get
1729 errors in the linking, it means that not all of the necessary routines
1730 in @file{libgcc1.a} are available.
1732 You must provide the header file @file{float.h}. One way to do this is
1733 to compile @file{enquire} and run it on your target machine. The job of
1734 @file{enquire} is to run on the target machine and figure out by
1735 experiment the nature of its floating point representation.
1736 @file{enquire} records its findings in the header file @file{float.h}.
1737 If you can't produce this file by running @file{enquire} on the target
1738 machine, then you will need to come up with a suitable @file{float.h} in
1739 some other way (or else, avoid using it in your programs).
1741 Do not try to build stage 2 for a cross-compiler. It doesn't work to
1742 rebuild GNU CC as a cross-compiler using the cross-compiler, because
1743 that would produce a program that runs on the target machine, not on the
1744 host. For example, if you compile a 386-to-68030 cross-compiler with
1745 itself, the result will not be right either for the 386 (because it was
1746 compiled into 68030 code) or for the 68030 (because it was configured
1747 for a 386 as the host). If you want to compile GNU CC into 68030 code,
1748 whether you compile it on a 68030 or with a cross-compiler on a 386, you
1749 must specify a 68030 as the host when you configure it.
1751 To install the cross-compiler, use @samp{make install}, as usual.
1754 @section Installing GNU CC on the Sun
1755 @cindex Sun installation
1756 @cindex installing GNU CC on the Sun
1758 On Solaris (version 2.1), do not use the linker or other tools in
1759 @file{/usr/ucb} to build GNU CC. Use @code{/usr/ccs/bin}.
1761 Make sure the environment variable @code{FLOAT_OPTION} is not set when
1762 you compile @file{libgcc.a}. If this option were set to @code{f68881}
1763 when @file{libgcc.a} is compiled, the resulting code would demand to be
1764 linked with a special startup file and would not link properly without
1767 @cindex @code{alloca}, for SunOs
1768 There is a bug in @code{alloca} in certain versions of the Sun library.
1769 To avoid this bug, install the binaries of GNU CC that were compiled by
1770 GNU CC. They use @code{alloca} as a built-in function and never the one
1773 Some versions of the Sun compiler crash when compiling GNU CC. The
1774 problem is a segmentation fault in cpp. This problem seems to be due to
1775 the bulk of data in the environment variables. You may be able to avoid
1776 it by using the following command to compile GNU CC with Sun CC:
1779 make CC="TERMCAP=x OBJS=x LIBFUNCS=x STAGESTUFF=x cc"
1783 @section Installing GNU CC on VMS
1784 @cindex VMS installation
1785 @cindex installing GNU CC on VMS
1787 The VMS version of GNU CC is distributed in a backup saveset containing
1788 both source code and precompiled binaries.
1790 To install the @file{gcc} command so you can use the compiler easily, in
1791 the same manner as you use the VMS C compiler, you must install the VMS CLD
1792 file for GNU CC as follows:
1796 Define the VMS logical names @samp{GNU_CC} and @samp{GNU_CC_INCLUDE}
1797 to point to the directories where the GNU CC executables
1798 (@file{gcc-cpp.exe}, @file{gcc-cc1.exe}, etc.) and the C include files are
1799 kept respectively. This should be done with the commands:@refill
1802 $ assign /system /translation=concealed -
1804 $ assign /system /translation=concealed -
1805 disk:[gcc.include.] gnu_cc_include
1809 with the appropriate disk and directory names. These commands can be
1810 placed in your system startup file so they will be executed whenever
1811 the machine is rebooted. You may, if you choose, do this via the
1812 @file{GCC_INSTALL.COM} script in the @file{[GCC]} directory.
1815 Install the @file{GCC} command with the command line:
1818 $ set command /table=sys$common:[syslib]dcltables -
1819 /output=sys$common:[syslib]dcltables gnu_cc:[000000]gcc
1820 $ install replace sys$common:[syslib]dcltables
1824 To install the help file, do the following:
1827 $ library/help sys$library:helplib.hlb gcc.hlp
1831 Now you can invoke the compiler with a command like @samp{gcc /verbose
1832 file.c}, which is equivalent to the command @samp{gcc -v -c file.c} in
1836 If you wish to use GNU C++ you must first install GNU CC, and then
1837 perform the following steps:
1841 Define the VMS logical name @samp{GNU_GXX_INCLUDE} to point to the
1842 directory where the preprocessor will search for the C++ header files.
1843 This can be done with the command:@refill
1846 $ assign /system /translation=concealed -
1847 disk:[gcc.gxx_include.] gnu_gxx_include
1851 with the appropriate disk and directory name. If you are going to be
1852 using libg++, this is where the libg++ install procedure will install
1853 the libg++ header files.
1856 Obtain the file @file{gcc-cc1plus.exe}, and place this in the same
1857 directory that @file{gcc-cc1.exe} is kept.
1859 The GNU C++ compiler can be invoked with a command like @samp{gcc /plus
1860 /verbose file.cc}, which is equivalent to the command @samp{g++ -v -c
1864 We try to put corresponding binaries and sources on the VMS distribution
1865 tape. But sometimes the binaries will be from an older version than the
1866 sources, because we don't always have time to update them. (Use the
1867 @samp{/version} option to determine the version number of the binaries and
1868 compare it with the source file @file{version.c} to tell whether this is
1869 so.) In this case, you should use the binaries you get to recompile the
1870 sources. If you must recompile, here is how:
1874 Execute the command procedure @file{vmsconfig.com} to set up the files
1875 @file{tm.h}, @file{config.h}, @file{aux-output.c}, and @file{md.}, and
1876 to create files @file{tconfig.h} and @file{hconfig.h}. This procedure
1877 also creates several linker option files used by @file{make-cc1.com} and
1878 a data file used by @file{make-l2.com}.@refill
1885 Setup the logical names and command tables as defined above. In
1886 addition, define the VMS logical name @samp{GNU_BISON} to point at the
1887 to the directories where the Bison executable is kept. This should be
1888 done with the command:@refill
1891 $ assign /system /translation=concealed -
1892 disk:[bison.] gnu_bison
1895 You may, if you choose, use the @file{INSTALL_BISON.COM} script in the
1896 @file{[BISON]} directory.
1899 Install the @samp{BISON} command with the command line:@refill
1902 $ set command /table=sys$common:[syslib]dcltables -
1903 /output=sys$common:[syslib]dcltables -
1904 gnu_bison:[000000]bison
1905 $ install replace sys$common:[syslib]dcltables
1909 Type @samp{@@make-gcc} to recompile everything (alternatively, submit
1910 the file @file{make-gcc.com} to a batch queue). If you wish to build
1911 the GNU C++ compiler as well as the GNU CC compiler, you must first edit
1912 @file{make-gcc.com} and follow the instructions that appear in the
1916 In order to use GCC, you need a library of functions which GCC compiled code
1917 will call to perform certain tasks, and these functions are defined in the
1918 file @file{libgcc2.c}. To compile this you should use the command procedure
1919 @file{make-l2.com}, which will generate the library @file{libgcc2.olb}.
1920 @file{libgcc2.olb} should be built using the compiler built from
1921 the same distribution that @file{libgcc2.c} came from, and
1922 @file{make-gcc.com} will automatically do all of this for you.
1924 To install the library, use the following commands:@refill
1927 $ library gnu_cc:[000000]gcclib/delete=(new,eprintf)
1928 $ library gnu_cc:[000000]gcclib/delete=L_*
1929 $ library libgcc2/extract=*/output=libgcc2.obj
1930 $ library gnu_cc:[000000]gcclib libgcc2.obj
1933 The first command simply removes old modules that will be replaced with
1934 modules from @file{libgcc2} under different module names. The modules
1935 @code{new} and @code{eprintf} may not actually be present in your
1936 @file{gcclib.olb}---if the VMS librarian complains about those modules
1937 not being present, simply ignore the message and continue on with the
1938 next command. The second command removes the modules that came from the
1939 previous version of the library @file{libgcc2.c}.
1941 Whenever you update the compiler on your system, you should also update the
1942 library with the above procedure.
1945 You may wish to build GCC in such a way that no files are written to the
1946 directory where the source files reside. An example would be the when
1947 the source files are on a read-only disk. In these cases, execute the
1948 following DCL commands (substituting your actual path names):
1951 $ assign dua0:[gcc.build_dir.]/translation=concealed, -
1952 dua1:[gcc.source_dir.]/translation=concealed gcc_build
1953 $ set default gcc_build:[000000]
1957 where the directory @file{dua1:[gcc.source_dir]} contains the source
1958 code, and the directory @file{dua0:[gcc.build_dir]} is meant to contain
1959 all of the generated object files and executables. Once you have done
1960 this, you can proceed building GCC as described above. (Keep in mind
1961 that @file{gcc_build} is a rooted logical name, and thus the device
1962 names in each element of the search list must be an actual physical
1963 device name rather than another rooted logical name).
1966 @strong{If you are building GNU CC with a previous version of GNU CC,
1967 you also should check to see that you have the newest version of the
1968 assembler}. In particular, GNU CC version 2 treats global constant
1969 variables slightly differently from GNU CC version 1, and GAS version
1970 1.38.1 does not have the patches required to work with GCC version 2.
1971 If you use GAS 1.38.1, then @code{extern const} variables will not have
1972 the read-only bit set, and the linker will generate warning messages
1973 about mismatched psect attributes for these variables. These warning
1974 messages are merely a nuisance, and can safely be ignored.
1976 If you are compiling with a version of GNU CC older than 1.33, specify
1977 @samp{/DEFINE=("inline=")} as an option in all the compilations. This
1978 requires editing all the @code{gcc} commands in @file{make-cc1.com}.
1979 (The older versions had problems supporting @code{inline}.) Once you
1980 have a working 1.33 or newer GNU CC, you can change this file back.
1983 If you want to build GNU CC with the VAX C compiler, you will need to
1984 make minor changes in @file{make-cccp.com} and @file{make-cc1.com}
1985 to choose alternate definitions of @code{CC}, @code{CFLAGS}, and
1986 @code{LIBS}. See comments in those files. However, you must
1987 also have a working version of the GNU assembler (GNU as, aka GAS) as
1988 it is used as the back-end for GNU CC to produce binary object modules
1989 and is not included in the GNU CC sources. GAS is also needed to
1990 compile @file{libgcc2} in order to build @file{gcclib} (see above);
1991 @file{make-l2.com} expects to be able to find it operational in
1992 @file{gnu_cc:[000000]gnu-as.exe}.
1994 To use GNU CC on VMS, you need the VMS driver programs
1995 @file{gcc.exe}, @file{gcc.com}, and @file{gcc.cld}. They are
1996 distributed with the VMS binaries (@file{gcc-vms}) rather than the
1997 GNU CC sources. GAS is also included in @file{gcc-vms}, as is Bison.
1999 Once you have successfully built GNU CC with VAX C, you should use the
2000 resulting compiler to rebuild itself. Before doing this, be sure to
2001 restore the @code{CC}, @code{CFLAGS}, and @code{LIBS} definitions in
2002 @file{make-cccp.com} and @file{make-cc1.com}. The second generation
2003 compiler will be able to take advantage of many optimizations that must
2004 be suppressed when building with other compilers.
2007 Under previous versions of GNU CC, the generated code would occasionally
2008 give strange results when linked with the sharable @file{VAXCRTL} library.
2009 Now this should work.
2011 Even with this version, however, GNU CC itself should not be linked with
2012 the sharable @file{VAXCRTL}. The version of @code{qsort} in
2013 @file{VAXCRTL} has a bug (known to be present in VMS versions V4.6
2014 through V5.5) which causes the compiler to fail.
2016 The executables are generated by @file{make-cc1.com} and
2017 @file{make-cccp.com} use the object library version of @file{VAXCRTL} in
2018 order to make use of the @code{qsort} routine in @file{gcclib.olb}. If
2019 you wish to link the compiler executables with the shareable image
2020 version of @file{VAXCRTL}, you should edit the file @file{tm.h} (created
2021 by @file{vmsconfig.com}) to define the macro @code{QSORT_WORKAROUND}.
2023 @code{QSORT_WORKAROUND} is always defined when GNU CC is compiled with
2024 VAX C, to avoid a problem in case @file{gcclib.olb} is not yet
2028 @section @code{collect2}
2030 Many target systems do not have support in the assembler and linker for
2031 ``constructors''---initialization functions to be called before the
2032 official ``start'' of @code{main}. On such systems, GNU CC uses a
2033 utility called @code{collect2} to arrange to call these functions at
2036 The program @code{collect2} works by linking the program once and
2037 looking through the linker output file for symbols with particular names
2038 indicating they are constructor functions. If it finds any, it
2039 creates a new temporary @samp{.c} file containing a table of them,
2040 compiles it, and links the program a second time including that file.
2043 @cindex constructors, automatic calls
2044 The actual calls to the constructors are carried out by a subroutine
2045 called @code{__main}, which is called (automatically) at the beginning
2046 of the body of @code{main} (provided @code{main} was compiled with GNU
2047 CC). Calling @code{__main} is necessary, even when compiling C code, to
2048 allow linking C and C++ object code together. (If you use
2049 @samp{-nostdlib}, you get an unresolved reference to @code{__main},
2050 since it's defined in the standard GCC library. Include @samp{-lgcc} at
2051 the end of your compiler command line to resolve this reference.)
2053 The program @code{collect2} is installed as @code{ld} in the directory
2054 where the passes of the compiler are installed. When @code{collect2}
2055 needs to find the @emph{real} @code{ld}, it tries the following file
2060 @file{real-ld} in the directories listed in the compiler's search
2064 @file{real-ld} in the directories listed in the environment variable
2068 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2072 @file{ld} in the compiler's search directories, except that
2073 @code{collect2} will not execute itself recursively.
2076 @file{ld} in @code{PATH}.
2079 ``The compiler's search directories'' means all the directories where
2080 @code{gcc} searches for passes of the compiler. This includes
2081 directories that you specify with @samp{-B}.
2083 Cross-compilers search a little differently:
2087 @file{real-ld} in the compiler's search directories.
2090 @file{@var{target}-real-ld} in @code{PATH}.
2093 The file specified in the @code{REAL_LD_FILE_NAME} configuration macro,
2097 @file{ld} in the compiler's search directories.
2100 @file{@var{target}-ld} in @code{PATH}.
2103 @code{collect2} explicitly avoids running @code{ld} using the file name
2104 under which @code{collect2} itself was invoked. In fact, it remembers
2105 up a list of such names---in case one copy of @code{collect2} finds
2106 another copy (or version) of @code{collect2} installed as @code{ld} in a
2107 second place in the search path.
2109 @code{collect2} searches for the utilities @code{nm} and @code{strip}
2110 using the same algorithm as above for @code{ld}.
2113 @section Standard Header File Directories
2115 @code{GCC_INCLUDE_DIR} means the same thing for native and cross. It is
2116 where GNU CC stores its private include files, and also where GNU CC
2117 stores the fixed include files. A cross compiled GNU CC runs
2118 @code{fixincludes} on the header files in @file{$(tooldir)/include}.
2119 (If the cross compilation header files need to be fixed, they must be
2120 installed before GNU CC is built. If the cross compilation header files
2121 are already suitable for ANSI C and GNU CC, nothing special need be
2124 @code{GPLUS_INCLUDE_DIR} means the same thing for native and cross. It
2125 is where @code{g++} looks first for header files. @code{libg++}
2126 installs only target independent header files in that directory.
2128 @code{LOCAL_INCLUDE_DIR} is used only for a native compiler. It is
2129 normally @file{/usr/local/include}. GNU CC searches this directory so
2130 that users can install header files in @file{/usr/local/include}.
2132 @code{CROSS_INCLUDE_DIR} is used only for a cross compiler. GNU CC
2133 doesn't install anything there.
2135 @code{TOOL_INCLUDE_DIR} is used for both native and cross compilers. It
2136 is the place for other packages to install header files that GNU CC will
2137 use. For a cross-compiler, this is the equivalent of
2138 @file{/usr/include}. When you build a cross-compiler,
2139 @code{fixincludes} processes any header files in this directory.