]>
Commit | Line | Data |
---|---|---|
e9a25f70 | 1 | @c Copyright (C) 1988,89,92,93,94,95,96,97 Free Software Foundation, Inc. |
2284f91b DE |
2 | @c This is part of the GCC manual. |
3 | @c For copying conditions, see the file gcc.texi. | |
4 | ||
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. | |
7 | ||
8 | @ifclear INSTALLONLY | |
9 | @node Installation | |
10 | @chapter Installing GNU CC | |
11 | @end ifclear | |
12 | @cindex installing GNU CC | |
13 | ||
14 | @menu | |
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. | |
22 | @end menu | |
23 | ||
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 | |
28 | systems. | |
29 | ||
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. | |
34 | ||
35 | @enumerate | |
36 | @item | |
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 | |
42 | clean. | |
43 | ||
44 | @item | |
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. | |
48 | ||
49 | @item | |
50 | Specify the host, build and target machine configurations. You do this | |
51 | by running the file @file{configure}. | |
52 | ||
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. | |
57 | ||
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 | |
64 | wrong. | |
65 | ||
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}.) | |
70 | ||
71 | Here is an example: | |
72 | ||
73 | @smallexample | |
74 | ./configure --build=sparc-sun-sunos4.1 | |
75 | @end smallexample | |
76 | ||
77 | A configuration name may be canonical or it may be more or less | |
78 | abbreviated. | |
79 | ||
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. | |
85 | ||
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. | |
91 | ||
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. | |
95 | ||
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. | |
99 | ||
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 | |
103 | @samp{--nfp}. | |
104 | ||
105 | @table @samp | |
106 | @item --with-gnu-as | |
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 | |
109 | @file{configure}. | |
110 | ||
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. | |
113 | ||
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. | |
120 | ||
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. | |
130 | ||
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}). | |
134 | ||
135 | @item --with-gnu-ld | |
136 | Specify the option @samp{--with-gnu-ld} if you plan to use the GNU | |
137 | linker with GNU CC. | |
138 | ||
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 | |
143 | configurations. | |
144 | ||
145 | @item --with-stabs | |
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. | |
151 | ||
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 | |
154 | CC. | |
155 | ||
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. | |
159 | ||
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. | |
164 | ||
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. | |
171 | ||
172 | @item --nfp | |
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 | |
177 | make a difference. | |
178 | ||
9101297d DL |
179 | @cindex Haifa scheduler |
180 | @cindex scheduler, experimental | |
181 | @item --enable-haifa | |
182 | @item --disable-haifa | |
183 | Use @samp{--enable-haifa} to enable use of an experimental instruction | |
184 | scheduler (from IBM Haifa). This may or may not produce better code. | |
185 | Some targets on which it is known to be a win enable it by default; use | |
186 | @samp{--disable-haifa} to disable it in these cases. @code{configure} | |
187 | will print out whether the Haifa scheduler is enabled when it is run. | |
188 | ||
2284f91b DE |
189 | @cindex Objective C threads |
190 | @cindex threads, Objective C | |
191 | @item --enable-objcthreads=@var{type} | |
861bb6c1 JL |
192 | Certain systems, notably Linux-based GNU systems, can't be relied on to |
193 | supply a threads facility for the Objective C runtime and so will | |
194 | default to single-threaded runtime. They may, however, have a library | |
195 | threads implementation available, in which case threads can be enabled | |
196 | with this option by supplying a suitable @var{type}, probably | |
197 | @samp{posix}. The possibilities for @var{type} are @samp{single}, | |
198 | @samp{posix}, @samp{win32}, @samp{solaris}, @samp{irix} and @samp{mach}. | |
2284f91b DE |
199 | @end table |
200 | ||
201 | The @file{configure} script searches subdirectories of the source | |
202 | directory for other compilers that are to be integrated into GNU CC. | |
203 | The GNU compiler for C++, called G++ is in a subdirectory named | |
204 | @file{cp}. @file{configure} inserts rules into @file{Makefile} to build | |
205 | all of those compilers. | |
206 | ||
207 | Here we spell out what files will be set up by @code{configure}. Normally | |
208 | you need not be concerned with these files. | |
209 | ||
210 | @itemize @bullet | |
211 | @item | |
212 | @ifset INTERNALS | |
213 | A file named @file{config.h} is created that contains a @samp{#include} | |
214 | of the top-level config file for the machine you will run the compiler | |
215 | on (@pxref{Config}). This file is responsible for defining information | |
216 | about the host machine. It includes @file{tm.h}. | |
217 | @end ifset | |
218 | @ifclear INTERNALS | |
219 | A file named @file{config.h} is created that contains a @samp{#include} | |
220 | of the top-level config file for the machine you will run the compiler | |
221 | on (@pxref{Config,,The Configuration File, gcc.info, Using and Porting | |
222 | GCC}). This file is responsible for defining information about the host | |
223 | machine. It includes @file{tm.h}. | |
224 | @end ifclear | |
225 | ||
226 | The top-level config file is located in the subdirectory @file{config}. | |
227 | Its name is always @file{xm-@var{something}.h}; usually | |
228 | @file{xm-@var{machine}.h}, but there are some exceptions. | |
229 | ||
230 | If your system does not support symbolic links, you might want to | |
231 | set up @file{config.h} to contain a @samp{#include} command which | |
232 | refers to the appropriate file. | |
233 | ||
234 | @item | |
235 | A file named @file{tconfig.h} is created which includes the top-level config | |
236 | file for your target machine. This is used for compiling certain | |
237 | programs to run on that machine. | |
238 | ||
239 | @item | |
240 | A file named @file{tm.h} is created which includes the | |
241 | machine-description macro file for your target machine. It should be in | |
242 | the subdirectory @file{config} and its name is often | |
243 | @file{@var{machine}.h}. | |
244 | ||
245 | @item | |
246 | The command file @file{configure} also constructs the file | |
247 | @file{Makefile} by adding some text to the template file | |
248 | @file{Makefile.in}. The additional text comes from files in the | |
249 | @file{config} directory, named @file{t-@var{target}} and | |
250 | @file{x-@var{host}}. If these files do not exist, it means nothing | |
251 | needs to be added for a given target or host. | |
252 | @end itemize | |
253 | ||
254 | @item | |
255 | The standard directory for installing GNU CC is @file{/usr/local/lib}. | |
256 | If you want to install its files somewhere else, specify | |
257 | @samp{--prefix=@var{dir}} when you run @file{configure}. Here @var{dir} | |
258 | is a directory name to use instead of @file{/usr/local} for all purposes | |
259 | with one exception: the directory @file{/usr/local/include} is searched | |
260 | for header files no matter where you install the compiler. To override | |
261 | this name, use the @code{--local-prefix} option below. | |
262 | ||
263 | @item | |
264 | Specify @samp{--local-prefix=@var{dir}} if you want the compiler to | |
265 | search directory @file{@var{dir}/include} for locally installed header | |
266 | files @emph{instead} of @file{/usr/local/include}. | |
267 | ||
268 | You should specify @samp{--local-prefix} @strong{only} if your site has | |
269 | a different convention (not @file{/usr/local}) for where to put | |
270 | site-specific files. | |
271 | ||
272 | The default value for @samp{--local-prefix} is @file{/usr/local} | |
273 | regardless of the value of @samp{--prefix}. Specifying @samp{--prefix} | |
274 | has no effect on which directory GNU CC searches for local header files. | |
275 | This may seem counterintuitive, but actually it is logical. | |
276 | ||
277 | The purpose of @samp{--prefix} is to specify where to @emph{install GNU | |
278 | CC}. The local header files in @file{/usr/local/include}---if you put | |
279 | any in that directory---are not part of GNU CC. They are part of other | |
280 | programs---perhaps many others. (GNU CC installs its own header files | |
281 | in another directory which is based on the @samp{--prefix} value.) | |
282 | ||
283 | @strong{Do not} specify @file{/usr} as the @samp{--local-prefix}! The | |
284 | directory you use for @samp{--local-prefix} @strong{must not} contain | |
285 | any of the system's standard header files. If it did contain them, | |
286 | certain programs would be miscompiled (including GNU Emacs, on certain | |
287 | targets), because this would override and nullify the header file | |
288 | corrections made by the @code{fixincludes} script. | |
289 | ||
290 | Indications are that people who use this option use it based on | |
291 | mistaken ideas of what it is for. People use it as if it specified | |
292 | where to install part of GNU CC. Perhaps they make this assumption | |
293 | because installing GNU CC creates the directory. | |
294 | ||
295 | @cindex Bison parser generator | |
296 | @cindex parser generator, Bison | |
297 | @item | |
298 | Make sure the Bison parser generator is installed. (This is | |
299 | unnecessary if the Bison output files @file{c-parse.c} and | |
300 | @file{cexp.c} are more recent than @file{c-parse.y} and @file{cexp.y} | |
301 | and you do not plan to change the @samp{.y} files.) | |
302 | ||
303 | Bison versions older than Sept 8, 1988 will produce incorrect output | |
304 | for @file{c-parse.c}. | |
305 | ||
306 | @item | |
307 | If you have chosen a configuration for GNU CC which requires other GNU | |
308 | tools (such as GAS or the GNU linker) instead of the standard system | |
309 | tools, install the required tools in the build directory under the names | |
310 | @file{as}, @file{ld} or whatever is appropriate. This will enable the | |
311 | compiler to find the proper tools for compilation of the program | |
312 | @file{enquire}. | |
313 | ||
314 | Alternatively, you can do subsequent compilation using a value of the | |
315 | @code{PATH} environment variable such that the necessary GNU tools come | |
316 | before the standard system tools. | |
317 | ||
318 | @item | |
319 | Build the compiler. Just type @samp{make LANGUAGES=c} in the compiler | |
320 | directory. | |
321 | ||
322 | @samp{LANGUAGES=c} specifies that only the C compiler should be | |
323 | compiled. The makefile normally builds compilers for all the supported | |
324 | languages; currently, C, C++ and Objective C. However, C is the only | |
325 | language that is sure to work when you build with other non-GNU C | |
326 | compilers. In addition, building anything but C at this stage is a | |
327 | waste of time. | |
328 | ||
329 | In general, you can specify the languages to build by typing the | |
330 | argument @samp{LANGUAGES="@var{list}"}, where @var{list} is one or more | |
331 | words from the list @samp{c}, @samp{c++}, and @samp{objective-c}. If | |
332 | you have any additional GNU compilers as subdirectories of the GNU CC | |
333 | source directory, you may also specify their names in this list. | |
334 | ||
335 | Ignore any warnings you may see about ``statement not reached'' in | |
336 | @file{insn-emit.c}; they are normal. Also, warnings about ``unknown | |
337 | escape sequence'' are normal in @file{genopinit.c} and perhaps some | |
338 | other files. Likewise, you should ignore warnings about ``constant is | |
339 | so large that it is unsigned'' in @file{insn-emit.c} and | |
340 | @file{insn-recog.c} and a warning about a comparison always being zero | |
341 | in @file{enquire.o}. Any other compilation errors may represent bugs in | |
342 | the port to your machine or operating system, and | |
343 | @ifclear INSTALLONLY | |
344 | should be investigated and reported (@pxref{Bugs}). | |
345 | @end ifclear | |
346 | @ifset INSTALLONLY | |
347 | should be investigated and reported. | |
348 | @end ifset | |
349 | ||
350 | Some commercial compilers fail to compile GNU CC because they have bugs | |
351 | or limitations. For example, the Microsoft compiler is said to run out | |
352 | of macro space. Some Ultrix compilers run out of expression space; then | |
353 | you need to break up the statement where the problem happens. | |
354 | ||
355 | @item | |
356 | If you are building a cross-compiler, stop here. @xref{Cross-Compiler}. | |
357 | ||
358 | @cindex stage1 | |
359 | @item | |
360 | Move the first-stage object files and executables into a subdirectory | |
361 | with this command: | |
362 | ||
363 | @smallexample | |
364 | make stage1 | |
365 | @end smallexample | |
366 | ||
367 | The files are moved into a subdirectory named @file{stage1}. | |
368 | Once installation is complete, you may wish to delete these files | |
369 | with @code{rm -r stage1}. | |
370 | ||
371 | @item | |
372 | If you have chosen a configuration for GNU CC which requires other GNU | |
373 | tools (such as GAS or the GNU linker) instead of the standard system | |
374 | tools, install the required tools in the @file{stage1} subdirectory | |
375 | under the names @file{as}, @file{ld} or whatever is appropriate. This | |
376 | will enable the stage 1 compiler to find the proper tools in the | |
377 | following stage. | |
378 | ||
379 | Alternatively, you can do subsequent compilation using a value of the | |
380 | @code{PATH} environment variable such that the necessary GNU tools come | |
381 | before the standard system tools. | |
382 | ||
383 | @item | |
384 | Recompile the compiler with itself, with this command: | |
385 | ||
386 | @smallexample | |
387 | make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2" | |
388 | @end smallexample | |
389 | ||
390 | This is called making the stage 2 compiler. | |
391 | ||
392 | The command shown above builds compilers for all the supported | |
393 | languages. If you don't want them all, you can specify the languages to | |
394 | build by typing the argument @samp{LANGUAGES="@var{list}"}. @var{list} | |
395 | should contain one or more words from the list @samp{c}, @samp{c++}, | |
396 | @samp{objective-c}, and @samp{proto}. Separate the words with spaces. | |
397 | @samp{proto} stands for the programs @code{protoize} and | |
398 | @code{unprotoize}; they are not a separate language, but you use | |
399 | @code{LANGUAGES} to enable or disable their installation. | |
400 | ||
401 | If you are going to build the stage 3 compiler, then you might want to | |
402 | build only the C language in stage 2. | |
403 | ||
404 | Once you have built the stage 2 compiler, if you are short of disk | |
405 | space, you can delete the subdirectory @file{stage1}. | |
406 | ||
407 | On a 68000 or 68020 system lacking floating point hardware, | |
408 | unless you have selected a @file{tm.h} file that expects by default | |
409 | that there is no such hardware, do this instead: | |
410 | ||
411 | @smallexample | |
412 | make CC="stage1/xgcc -Bstage1/" CFLAGS="-g -O2 -msoft-float" | |
413 | @end smallexample | |
414 | ||
415 | @item | |
416 | If you wish to test the compiler by compiling it with itself one more | |
417 | time, install any other necessary GNU tools (such as GAS or the GNU | |
418 | linker) in the @file{stage2} subdirectory as you did in the | |
419 | @file{stage1} subdirectory, then do this: | |
420 | ||
421 | @smallexample | |
422 | make stage2 | |
423 | make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" | |
424 | @end smallexample | |
425 | ||
426 | @noindent | |
427 | This is called making the stage 3 compiler. Aside from the @samp{-B} | |
428 | option, the compiler options should be the same as when you made the | |
429 | stage 2 compiler. But the @code{LANGUAGES} option need not be the | |
430 | same. The command shown above builds compilers for all the supported | |
431 | languages; if you don't want them all, you can specify the languages to | |
432 | build by typing the argument @samp{LANGUAGES="@var{list}"}, as described | |
433 | above. | |
434 | ||
435 | If you do not have to install any additional GNU tools, you may use the | |
436 | command | |
437 | ||
438 | @smallexample | |
439 | make bootstrap LANGUAGES=@var{language-list} BOOT_CFLAGS=@var{option-list} | |
440 | @end smallexample | |
441 | ||
442 | @noindent | |
443 | instead of making @file{stage1}, @file{stage2}, and performing | |
444 | the two compiler builds. | |
445 | ||
446 | @item | |
447 | Then compare the latest object files with the stage 2 object | |
448 | files---they ought to be identical, aside from time stamps (if any). | |
449 | ||
450 | On some systems, meaningful comparison of object files is impossible; | |
451 | they always appear ``different.'' This is currently true on Solaris and | |
452 | some systems that use ELF object file format. On some versions of Irix | |
453 | on SGI machines and DEC Unix (OSF/1) on Alpha systems, you will not be | |
454 | able to compare the files without specifying @file{-save-temps}; see the | |
455 | description of individual systems above to see if you get comparison | |
456 | failures. You may have similar problems on other systems. | |
457 | ||
458 | Use this command to compare the files: | |
459 | ||
460 | @smallexample | |
461 | make compare | |
462 | @end smallexample | |
463 | ||
464 | This will mention any object files that differ between stage 2 and stage | |
465 | 3. Any difference, no matter how innocuous, indicates that the stage 2 | |
466 | compiler has compiled GNU CC incorrectly, and is therefore a potentially | |
467 | @ifclear INSTALLONLY | |
468 | serious bug which you should investigate and report (@pxref{Bugs}). | |
469 | @end ifclear | |
470 | @ifset INSTALLONLY | |
471 | serious bug which you should investigate and report. | |
472 | @end ifset | |
473 | ||
474 | If your system does not put time stamps in the object files, then this | |
475 | is a faster way to compare them (using the Bourne shell): | |
476 | ||
477 | @smallexample | |
478 | for file in *.o; do | |
479 | cmp $file stage2/$file | |
480 | done | |
481 | @end smallexample | |
482 | ||
483 | If you have built the compiler with the @samp{-mno-mips-tfile} option on | |
484 | MIPS machines, you will not be able to compare the files. | |
485 | ||
486 | @item | |
487 | Install the compiler driver, the compiler's passes and run-time support | |
488 | with @samp{make install}. Use the same value for @code{CC}, | |
489 | @code{CFLAGS} and @code{LANGUAGES} that you used when compiling the | |
490 | files that are being installed. One reason this is necessary is that | |
491 | some versions of Make have bugs and recompile files gratuitously when | |
492 | you do this step. If you use the same variable values, those files will | |
493 | be recompiled properly. | |
494 | ||
495 | For example, if you have built the stage 2 compiler, you can use the | |
496 | following command: | |
497 | ||
498 | @smallexample | |
499 | make install CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O" LANGUAGES="@var{list}" | |
500 | @end smallexample | |
501 | ||
502 | @noindent | |
503 | This copies the files @file{cc1}, @file{cpp} and @file{libgcc.a} to | |
504 | files @file{cc1}, @file{cpp} and @file{libgcc.a} in the directory | |
505 | @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}, which is where | |
506 | the compiler driver program looks for them. Here @var{target} is the | |
507 | target machine type specified when you ran @file{configure}, and | |
508 | @var{version} is the version number of GNU CC. This naming scheme | |
509 | permits various versions and/or cross-compilers to coexist. | |
510 | It also copies the executables for compilers for other languages | |
511 | (e.g., @file{cc1plus} for C++) to the same directory. | |
512 | ||
513 | This also copies the driver program @file{xgcc} into | |
514 | @file{/usr/local/bin/gcc}, so that it appears in typical execution | |
515 | search paths. It also copies @file{gcc.1} into | |
516 | @file{/usr/local/man/man1} and info pages into @file{/usr/local/info}. | |
517 | ||
518 | On some systems, this command causes recompilation of some files. This | |
519 | is usually due to bugs in @code{make}. You should either ignore this | |
520 | problem, or use GNU Make. | |
521 | ||
522 | @cindex @code{alloca} and SunOS | |
523 | @strong{Warning: there is a bug in @code{alloca} in the Sun library. To | |
524 | avoid this bug, be sure to install the executables of GNU CC that were | |
525 | compiled by GNU CC. (That is, the executables from stage 2 or 3, not | |
526 | stage 1.) They use @code{alloca} as a built-in function and never the | |
527 | one in the library.} | |
528 | ||
529 | (It is usually better to install GNU CC executables from stage 2 or 3, | |
530 | since they usually run faster than the ones compiled with some other | |
531 | compiler.) | |
532 | ||
533 | @item | |
534 | If you're going to use C++, it's likely that you need to also install | |
535 | the libg++ distribution. It should be available from the same | |
536 | place where you got the GNU C distribution. Just as GNU C does not | |
537 | distribute a C runtime library, it also does not include a C++ run-time | |
538 | library. All I/O functionality, special class libraries, etc., are | |
539 | available in the libg++ distribution. | |
540 | ||
541 | @item | |
542 | GNU CC includes a runtime library for Objective-C because it is an | |
543 | integral part of the language. You can find the files associated with | |
544 | the library in the subdirectory @file{objc}. The GNU Objective-C | |
545 | Runtime Library requires header files for the target's C library in | |
546 | order to be compiled,and also requires the header files for the target's | |
547 | thread library if you want thread support. @xref{Cross Headers, | |
548 | Cross-Compilers and Header Files, Cross-Compilers and Header Files}, for | |
549 | discussion about header files issues for cross-compilation. | |
550 | ||
551 | When you run @file{configure}, it picks the appropriate Objective-C | |
552 | thread implementation file for the target platform. In some situations, | |
553 | you may wish to choose a different back-end as some platforms support | |
554 | multiple thread implementations or you may wish to disable thread | |
555 | support completely. You do this by specifying a value for the | |
556 | @var{OBJC_THREAD_FILE} makefile variable on the command line when you | |
557 | run make, for example: | |
558 | ||
559 | @smallexample | |
560 | make CC="stage2/xgcc -Bstage2/" CFLAGS="-g -O2" OBJC_THREAD_FILE=thr-single | |
561 | @end smallexample | |
562 | ||
563 | @noindent | |
564 | Below is a list of the currently available back-ends. | |
565 | ||
566 | @itemize @bullet | |
567 | @item thr-single | |
568 | Disable thread support, should work for all platforms. | |
569 | @item thr-decosf1 | |
570 | DEC OSF/1 thread support. | |
571 | @item thr-irix | |
572 | SGI IRIX thread support. | |
573 | @item thr-mach | |
574 | Generic MACH thread support, known to work on NEXTSTEP. | |
575 | @item thr-os2 | |
576 | IBM OS/2 thread support. | |
577 | @item thr-posix | |
578 | Generix POSIX thread support. | |
579 | @item thr-pthreads | |
580 | PCThreads on Linux-based GNU systems. | |
581 | @item thr-solaris | |
582 | SUN Solaris thread support. | |
583 | @item thr-win32 | |
584 | Microsoft Win32 API thread support. | |
585 | @end itemize | |
586 | @end enumerate | |
587 | ||
588 | @node Configurations | |
589 | @section Configurations Supported by GNU CC | |
590 | @cindex configurations supported by GNU CC | |
591 | ||
592 | Here are the possible CPU types: | |
593 | ||
594 | @quotation | |
595 | @c gmicro, alliant, spur and tahoe omitted since they don't work. | |
596 | 1750a, a29k, alpha, arm, c@var{n}, clipper, dsp16xx, elxsi, h8300, | |
65376d28 | 597 | hppa1.0, hppa1.1, i370, i386, i486, i586, i860, i960, m32r, m68000, m68k, |
2284f91b DE |
598 | m88k, mips, mipsel, mips64, mips64el, ns32k, powerpc, powerpcle, |
599 | pyramid, romp, rs6000, sh, sparc, sparclite, sparc64, vax, we32k. | |
600 | @end quotation | |
601 | ||
602 | Here are the recognized company names. As you can see, customary | |
603 | abbreviations are used rather than the longer official names. | |
604 | ||
605 | @c What should be done about merlin, tek*, dolphin? | |
606 | @quotation | |
607 | acorn, alliant, altos, apollo, apple, att, bull, | |
608 | cbm, convergent, convex, crds, dec, dg, dolphin, | |
609 | elxsi, encore, harris, hitachi, hp, ibm, intergraph, isi, | |
610 | mips, motorola, ncr, next, ns, omron, plexus, | |
611 | sequent, sgi, sony, sun, tti, unicom, wrs. | |
612 | @end quotation | |
613 | ||
614 | The company name is meaningful only to disambiguate when the rest of | |
615 | the information supplied is insufficient. You can omit it, writing | |
616 | just @samp{@var{cpu}-@var{system}}, if it is not needed. For example, | |
617 | @samp{vax-ultrix4.2} is equivalent to @samp{vax-dec-ultrix4.2}. | |
618 | ||
619 | Here is a list of system types: | |
620 | ||
621 | @quotation | |
0c82f6bf | 622 | 386bsd, aix, acis, amigaos, aos, aout, aux, bosx, bsd, clix, coff, ctix, cxux, |
861bb6c1 | 623 | dgux, dynix, ebmon, ecoff, elf, esix, freebsd, hms, genix, gnu, linux-gnu, |
2284f91b DE |
624 | hiux, hpux, iris, irix, isc, luna, lynxos, mach, minix, msdos, mvs, |
625 | netbsd, newsos, nindy, ns, osf, osfrose, ptx, riscix, riscos, rtu, sco, sim, | |
626 | solaris, sunos, sym, sysv, udi, ultrix, unicos, uniplus, unos, vms, vsta, | |
627 | vxworks, winnt, xenix. | |
628 | @end quotation | |
629 | ||
630 | @noindent | |
631 | You can omit the system type; then @file{configure} guesses the | |
632 | operating system from the CPU and company. | |
633 | ||
634 | You can add a version number to the system type; this may or may not | |
635 | make a difference. For example, you can write @samp{bsd4.3} or | |
636 | @samp{bsd4.4} to distinguish versions of BSD. In practice, the version | |
637 | number is most needed for @samp{sysv3} and @samp{sysv4}, which are often | |
638 | treated differently. | |
639 | ||
640 | If you specify an impossible combination such as @samp{i860-dg-vms}, | |
641 | then you may get an error message from @file{configure}, or it may | |
642 | ignore part of the information and do the best it can with the rest. | |
643 | @file{configure} always prints the canonical name for the alternative | |
644 | that it used. GNU CC does not support all possible alternatives. | |
645 | ||
646 | Often a particular model of machine has a name. Many machine names are | |
647 | recognized as aliases for CPU/company combinations. Thus, the machine | |
648 | name @samp{sun3}, mentioned above, is an alias for @samp{m68k-sun}. | |
649 | Sometimes we accept a company name as a machine name, when the name is | |
650 | popularly used for a particular machine. Here is a table of the known | |
651 | machine names: | |
652 | ||
653 | @quotation | |
654 | 3300, 3b1, 3b@var{n}, 7300, altos3068, altos, | |
655 | apollo68, att-7300, balance, | |
656 | convex-c@var{n}, crds, decstation-3100, | |
657 | decstation, delta, encore, | |
658 | fx2800, gmicro, hp7@var{nn}, hp8@var{nn}, | |
659 | hp9k2@var{nn}, hp9k3@var{nn}, hp9k7@var{nn}, | |
660 | hp9k8@var{nn}, iris4d, iris, isi68, | |
661 | m3230, magnum, merlin, miniframe, | |
662 | mmax, news-3600, news800, news, next, | |
663 | pbd, pc532, pmax, powerpc, powerpcle, ps2, risc-news, | |
664 | rtpc, sun2, sun386i, sun386, sun3, | |
665 | sun4, symmetry, tower-32, tower. | |
666 | @end quotation | |
667 | ||
668 | @noindent | |
669 | Remember that a machine name specifies both the cpu type and the company | |
670 | name. | |
671 | If you want to install your own homemade configuration files, you can | |
672 | use @samp{local} as the company name to access them. If you use | |
673 | configuration @samp{@var{cpu}-local}, the configuration name | |
674 | without the cpu prefix | |
675 | is used to form the configuration file names. | |
676 | ||
677 | Thus, if you specify @samp{m68k-local}, configuration uses | |
678 | files @file{m68k.md}, @file{local.h}, @file{m68k.c}, | |
679 | @file{xm-local.h}, @file{t-local}, and @file{x-local}, all in the | |
680 | directory @file{config/m68k}. | |
681 | ||
682 | Here is a list of configurations that have special treatment or special | |
683 | things you must know: | |
684 | ||
685 | @table @samp | |
686 | @item 1750a-*-* | |
687 | MIL-STD-1750A processors. | |
688 | ||
689 | The MIL-STD-1750A cross configuration produces output for | |
690 | @code{as1750}, an assembler/linker available under the GNU Public | |
691 | License for the 1750A. @code{as1750} can be obtained at | |
692 | @emph{ftp://ftp.fta-berlin.de/pub/crossgcc/1750gals/}. | |
693 | A similarly licensed simulator for | |
694 | the 1750A is available from same address. | |
695 | ||
696 | You should ignore a fatal error during the building of libgcc (libgcc is | |
697 | not yet implemented for the 1750A.) | |
698 | ||
699 | The @code{as1750} assembler requires the file @file{ms1750.inc}, which is | |
700 | found in the directory @file{config/1750a}. | |
701 | ||
702 | GNU CC produced the same sections as the Fairchild F9450 C Compiler, | |
703 | namely: | |
704 | ||
705 | @table @code | |
706 | @item Normal | |
707 | The program code section. | |
708 | ||
709 | @item Static | |
710 | The read/write (RAM) data section. | |
711 | ||
712 | @item Konst | |
713 | The read-only (ROM) constants section. | |
714 | ||
715 | @item Init | |
716 | Initialization section (code to copy KREL to SREL). | |
717 | @end table | |
718 | ||
719 | The smallest addressable unit is 16 bits (BITS_PER_UNIT is 16). This | |
720 | means that type `char' is represented with a 16-bit word per character. | |
721 | The 1750A's "Load/Store Upper/Lower Byte" instructions are not used by | |
722 | GNU CC. | |
723 | ||
724 | @item alpha-*-osf1 | |
725 | Systems using processors that implement the DEC Alpha architecture and | |
726 | are running the DEC Unix (OSF/1) operating system, for example the DEC | |
727 | Alpha AXP systems. (VMS on the Alpha is not currently supported by GNU | |
728 | CC.) | |
729 | ||
730 | GNU CC writes a @samp{.verstamp} directive to the assembler output file | |
731 | unless it is built as a cross-compiler. It gets the version to use from | |
732 | the system header file @file{/usr/include/stamp.h}. If you install a | |
733 | new version of DEC Unix, you should rebuild GCC to pick up the new version | |
734 | stamp. | |
735 | ||
736 | Note that since the Alpha is a 64-bit architecture, cross-compilers from | |
737 | 32-bit machines will not generate code as efficient as that generated | |
738 | when the compiler is running on a 64-bit machine because many | |
739 | optimizations that depend on being able to represent a word on the | |
740 | target in an integral value on the host cannot be performed. Building | |
741 | cross-compilers on the Alpha for 32-bit machines has only been tested in | |
742 | a few cases and may not work properly. | |
743 | ||
744 | @code{make compare} may fail on old versions of DEC Unix unless you add | |
745 | @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the | |
746 | assembler input file is stored in the object file, and that makes | |
747 | comparison fail if it differs between the @code{stage1} and | |
748 | @code{stage2} compilations. The option @samp{-save-temps} forces a | |
749 | fixed name to be used for the assembler input file, instead of a | |
750 | randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps} | |
751 | unless the comparisons fail without that option. If you add | |
752 | @samp{-save-temps}, you will have to manually delete the @samp{.i} and | |
753 | @samp{.s} files after each series of compilations. | |
754 | ||
755 | GNU CC now supports both the native (ECOFF) debugging format used by DBX | |
756 | and GDB and an encapsulated STABS format for use only with GDB. See the | |
757 | discussion of the @samp{--with-stabs} option of @file{configure} above | |
758 | for more information on these formats and how to select them. | |
759 | ||
760 | There is a bug in DEC's assembler that produces incorrect line numbers | |
761 | for ECOFF format when the @samp{.align} directive is used. To work | |
762 | around this problem, GNU CC will not emit such alignment directives | |
763 | while writing ECOFF format debugging information even if optimization is | |
764 | being performed. Unfortunately, this has the very undesirable | |
765 | side-effect that code addresses when @samp{-O} is specified are | |
766 | different depending on whether or not @samp{-g} is also specified. | |
767 | ||
768 | To avoid this behavior, specify @samp{-gstabs+} and use GDB instead of | |
769 | DBX. DEC is now aware of this problem with the assembler and hopes to | |
770 | provide a fix shortly. | |
771 | ||
e9a25f70 JL |
772 | @item arc-*-elf |
773 | Argonaut ARC processor. | |
774 | This configuration is intended for embedded systems. | |
775 | ||
861bb6c1 | 776 | @item arm-*-aout |
2284f91b DE |
777 | Advanced RISC Machines ARM-family processors. These are often used in |
778 | embedded applications. There are no standard Unix configurations. | |
779 | This configuration corresponds to the basic instruction sequences and will | |
861bb6c1 | 780 | produce @file{a.out} format object modules. |
2284f91b DE |
781 | |
782 | You may need to make a variant of the file @file{arm.h} for your particular | |
783 | configuration. | |
784 | ||
861bb6c1 JL |
785 | @item arm-*-linuxaout |
786 | Any of the ARM family processors running the Linux-based GNU system with | |
787 | the @file{a.out} binary format (ELF is not yet supported). You must use | |
788 | version 2.8.1.0.7 or later of the Linux binutils, which you can download | |
789 | from @file{sunsite.unc.edu:/pub/Linux/GCC} and other mirror sites for | |
790 | Linux-based GNU systems. | |
791 | ||
2284f91b DE |
792 | @item arm-*-riscix |
793 | The ARM2 or ARM3 processor running RISC iX, Acorn's port of BSD Unix. If | |
794 | you are running a version of RISC iX prior to 1.2 then you must specify | |
795 | the version number during configuration. Note that the assembler | |
796 | shipped with RISC iX does not support stabs debugging information; a | |
797 | new version of the assembler, with stabs support included, is now | |
798 | available from Acorn. | |
799 | ||
800 | @item a29k | |
801 | AMD Am29k-family processors. These are normally used in embedded | |
802 | applications. There are no standard Unix configurations. | |
803 | This configuration | |
804 | corresponds to AMD's standard calling sequence and binary interface | |
805 | and is compatible with other 29k tools. | |
806 | ||
807 | You may need to make a variant of the file @file{a29k.h} for your | |
808 | particular configuration. | |
809 | ||
810 | @item a29k-*-bsd | |
811 | AMD Am29050 used in a system running a variant of BSD Unix. | |
812 | ||
813 | @item decstation-* | |
814 | DECstations can support three different personalities: Ultrix, | |
815 | DEC OSF/1, and OSF/rose. To configure GCC for these platforms | |
816 | use the following configurations: | |
817 | ||
818 | @table @samp | |
819 | @item decstation-ultrix | |
820 | Ultrix configuration. | |
821 | ||
822 | @item decstation-osf1 | |
823 | Dec's version of OSF/1. | |
824 | ||
825 | @item decstation-osfrose | |
826 | Open Software Foundation reference port of OSF/1 which uses the | |
827 | OSF/rose object file format instead of ECOFF. Normally, you | |
828 | would not select this configuration. | |
829 | @end table | |
830 | ||
831 | The MIPS C compiler needs to be told to increase its table size | |
832 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
833 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
834 | optimization option, you also need to use @samp{-Olimit 3000}. | |
835 | Both of these options are automatically generated in the | |
836 | @file{Makefile} that the shell script @file{configure} builds. | |
837 | If you override the @code{CC} make variable and use the MIPS | |
838 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
839 | ||
840 | @item elxsi-elxsi-bsd | |
841 | The Elxsi's C compiler has known limitations that prevent it from | |
842 | compiling GNU C. Please contact @code{mrs@@cygnus.com} for more details. | |
843 | ||
844 | @item dsp16xx | |
845 | A port to the AT&T DSP1610 family of processors. | |
846 | ||
847 | @ignore | |
848 | @item fx80 | |
849 | Alliant FX/8 computer. Note that the standard installed C compiler in | |
850 | Concentrix 5.0 has a bug which prevent it from compiling GNU CC | |
851 | correctly. You can patch the compiler bug as follows: | |
852 | ||
853 | @smallexample | |
854 | cp /bin/pcc ./pcc | |
855 | adb -w ./pcc - << EOF | |
856 | 15f6?w 6610 | |
857 | EOF | |
858 | @end smallexample | |
859 | ||
860 | Then you must use the @samp{-ip12} option when compiling GNU CC | |
861 | with the patched compiler, as shown here: | |
862 | ||
863 | @smallexample | |
864 | make CC="./pcc -ip12" CFLAGS=-w | |
865 | @end smallexample | |
866 | ||
867 | Note also that Alliant's version of DBX does not manage to work with the | |
868 | output from GNU CC. | |
869 | @end ignore | |
870 | ||
871 | @item h8300-*-* | |
e9a25f70 JL |
872 | Hitachi H8/300 series of processors. |
873 | ||
2284f91b DE |
874 | The calling convention and structure layout has changed in release 2.6. |
875 | All code must be recompiled. The calling convention now passes the | |
876 | first three arguments in function calls in registers. Structures are no | |
877 | longer a multiple of 2 bytes. | |
878 | ||
879 | @item hppa*-*-* | |
880 | There are several variants of the HP-PA processor which run a variety | |
881 | of operating systems. GNU CC must be configured to use the correct | |
882 | processor type and operating system, or GNU CC will not function correctly. | |
883 | The easiest way to handle this problem is to @emph{not} specify a target | |
884 | when configuring GNU CC, the @file{configure} script will try to automatically | |
885 | determine the right processor type and operating system. | |
886 | ||
887 | @samp{-g} does not work on HP-UX, since that system uses a peculiar | |
888 | debugging format which GNU CC does not know about. However, @samp{-g} | |
889 | will work if you also use GAS and GDB in conjunction with GCC. We | |
890 | highly recommend using GAS for all HP-PA configurations. | |
891 | ||
892 | You should be using GAS-2.6 (or later) along with GDB-4.16 (or later). These | |
893 | can be retrieved from all the traditional GNU ftp archive sites. | |
894 | ||
895 | GAS will need to be installed into a directory before @code{/bin}, | |
896 | @code{/usr/bin}, and @code{/usr/ccs/bin} in your search path. You | |
897 | should install GAS before you build GNU CC. | |
898 | ||
899 | To enable debugging, you must configure GNU CC with the @samp{--with-gnu-as} | |
900 | option before building. | |
901 | ||
902 | @item i370-*-* | |
903 | This port is very preliminary and has many known bugs. We hope to | |
904 | have a higher-quality port for this machine soon. | |
905 | ||
861bb6c1 JL |
906 | @item i386-*-linux-gnuoldld |
907 | Use this configuration to generate @file{a.out} binaries on Linux-based | |
908 | GNU systems if you do not have gas/binutils version 2.5.2 or later | |
909 | installed. This is an obsolete configuration. | |
2284f91b | 910 | |
861bb6c1 JL |
911 | @item i386-*-linux-gnuaout |
912 | Use this configuration to generate @file{a.out} binaries on Linux-based | |
913 | GNU systems. This configuration is being superseded. You must use | |
914 | gas/binutils version 2.5.2 or later. | |
2284f91b | 915 | |
861bb6c1 JL |
916 | @item i386-*-linux-gnu |
917 | Use this configuration to generate ELF binaries on Linux-based GNU | |
918 | systems. You must use gas/binutils version 2.5.2 or later. | |
2284f91b DE |
919 | |
920 | @item i386-*-sco | |
921 | Compilation with RCC is recommended. Also, it may be a good idea to | |
922 | link with GNU malloc instead of the malloc that comes with the system. | |
923 | ||
924 | @item i386-*-sco3.2v4 | |
925 | Use this configuration for SCO release 3.2 version 4. | |
926 | ||
927 | @item i386-*-sco3.2v5* | |
e9a25f70 JL |
928 | Use this for the SCO OpenServer Release family including 5.0.0, 5.0.2, |
929 | 5.0.4, Internet FastStart 1.0, and Internet FastStart 1.1. | |
930 | ||
931 | GNU CC can generate ELF binaries (if you specify @samp{-melf}) or COFF | |
932 | binaries (the default). If you are going to build your compiler in ELF | |
933 | mode (once you have bootstrapped the first stage compiler) you | |
934 | @strong{must} specify @samp{-melf} as part of @code{CC}, | |
935 | @emph{not} @code{CFLAGS}, for example as | |
936 | @samp{CC="stage1/xgcc -melf -Bstage1/" }. If you do not do this, the | |
937 | bootstrap will generate incorrect versions of @file{libgcc.a}. | |
2284f91b DE |
938 | |
939 | You must have TLS597 (from ftp.sco.com/TLS) installed for ELF | |
940 | binaries to work correctly. Note that Open Server 5.0.2 @emph{does} | |
941 | need TLS597 installed. | |
942 | ||
e9a25f70 JL |
943 | @emph{NOTE:} You must follow the instructions about invoking |
944 | @samp{make bootstrap} because the native OpenServer compiler builds | |
945 | a @file{cc1plus} that will not correctly parse many valid C++ programs. | |
946 | You must do a @samp{make bootstrap} if you are building with the native | |
947 | compiler. | |
948 | ||
2284f91b DE |
949 | @item i386-*-isc |
950 | It may be a good idea to link with GNU malloc instead of the malloc that | |
951 | comes with the system. | |
952 | ||
953 | In ISC version 4.1, @file{sed} core dumps when building | |
954 | @file{deduced.h}. Use the version of @file{sed} from version 4.0. | |
955 | ||
956 | @item i386-*-esix | |
957 | It may be good idea to link with GNU malloc instead of the malloc that | |
958 | comes with the system. | |
959 | ||
960 | @item i386-ibm-aix | |
961 | You need to use GAS version 2.1 or later, and LD from | |
962 | GNU binutils version 2.2 or later. | |
963 | ||
964 | @item i386-sequent-bsd | |
965 | Go to the Berkeley universe before compiling. In addition, you probably | |
966 | need to create a file named @file{string.h} containing just one line: | |
967 | @samp{#include <strings.h>}. | |
968 | ||
969 | @item i386-sequent-ptx1* | |
970 | Sequent DYNIX/ptx 1.x. | |
971 | ||
972 | @item i386-sequent-ptx2* | |
973 | Sequent DYNIX/ptx 2.x. | |
974 | ||
975 | @item i386-sun-sunos4 | |
976 | You may find that you need another version of GNU CC to begin | |
977 | bootstrapping with, since the current version when built with the | |
978 | system's own compiler seems to get an infinite loop compiling part of | |
979 | @file{libgcc2.c}. GNU CC version 2 compiled with GNU CC (any version) | |
980 | seems not to have this problem. | |
981 | ||
982 | See @ref{Sun Install}, for information on installing GNU CC on Sun | |
983 | systems. | |
984 | ||
985 | @item i[345]86-*-winnt3.5 | |
e9a25f70 | 986 | This version requires a GAS that has not yet been released. Until it |
2284f91b DE |
987 | is, you can get a prebuilt binary version via anonymous ftp from |
988 | @file{cs.washington.edu:pub/gnat} or @file{cs.nyu.edu:pub/gnat}. You | |
989 | must also use the Microsoft header files from the Windows NT 3.5 SDK. | |
990 | Find these on the CDROM in the @file{/mstools/h} directory dated 9/4/94. You | |
991 | must use a fixed version of Microsoft linker made especially for NT 3.5, | |
992 | which is also is available on the NT 3.5 SDK CDROM. If you do not have | |
993 | this linker, can you also use the linker from Visual C/C++ 1.0 or 2.0. | |
994 | ||
995 | Installing GNU CC for NT builds a wrapper linker, called @file{ld.exe}, | |
996 | which mimics the behaviour of Unix @file{ld} in the specification of | |
997 | libraries (@samp{-L} and @samp{-l}). @file{ld.exe} looks for both Unix | |
998 | and Microsoft named libraries. For example, if you specify | |
999 | @samp{-lfoo}, @file{ld.exe} will look first for @file{libfoo.a} | |
1000 | and then for @file{foo.lib}. | |
1001 | ||
1002 | You may install GNU CC for Windows NT in one of two ways, depending on | |
1003 | whether or not you have a Unix-like shell and various Unix-like | |
1004 | utilities. | |
1005 | ||
1006 | @enumerate | |
1007 | @item | |
1008 | If you do not have a Unix-like shell and few Unix-like utilities, you | |
1009 | will use a DOS style batch script called @file{configure.bat}. Invoke | |
1010 | it as @code{configure winnt} from an MSDOS console window or from the | |
1011 | program manager dialog box. @file{configure.bat} assumes you have | |
1012 | already installed and have in your path a Unix-like @file{sed} program | |
1013 | which is used to create a working @file{Makefile} from @file{Makefile.in}. | |
1014 | ||
1015 | @file{Makefile} uses the Microsoft Nmake program maintenance utility and | |
1016 | the Visual C/C++ V8.00 compiler to build GNU CC. You need only have the | |
1017 | utilities @file{sed} and @file{touch} to use this installation method, | |
1018 | which only automatically builds the compiler itself. You must then | |
1019 | examine what @file{fixinc.winnt} does, edit the header files by hand and | |
1020 | build @file{libgcc.a} manually. | |
1021 | ||
1022 | @item | |
1023 | The second type of installation assumes you are running a Unix-like | |
1024 | shell, have a complete suite of Unix-like utilities in your path, and | |
1025 | have a previous version of GNU CC already installed, either through | |
1026 | building it via the above installation method or acquiring a pre-built | |
1027 | binary. In this case, use the @file{configure} script in the normal | |
1028 | fashion. | |
1029 | @end enumerate | |
1030 | ||
1031 | @item i860-intel-osf1 | |
1032 | This is the Paragon. | |
1033 | @ifset INSTALLONLY | |
1034 | If you have version 1.0 of the operating system, you need to take | |
1035 | special steps to build GNU CC due to peculiarities of the system. Newer | |
1036 | system versions have no problem. See the section `Installation Problems' | |
1037 | in the GNU CC Manual. | |
1038 | @end ifset | |
1039 | @ifclear INSTALLONLY | |
1040 | If you have version 1.0 of the operating system, | |
1041 | see @ref{Installation Problems}, for special things you need to do to | |
1042 | compensate for peculiarities in the system. | |
1043 | @end ifclear | |
1044 | ||
1045 | @item *-lynx-lynxos | |
1046 | LynxOS 2.2 and earlier comes with GNU CC 1.x already installed as | |
1047 | @file{/bin/gcc}. You should compile with this instead of @file{/bin/cc}. | |
1048 | You can tell GNU CC to use the GNU assembler and linker, by specifying | |
1049 | @samp{--with-gnu-as --with-gnu-ld} when configuring. These will produce | |
1050 | COFF format object files and executables; otherwise GNU CC will use the | |
861bb6c1 | 1051 | installed tools, which produce @file{a.out} format executables. |
2284f91b | 1052 | |
65376d28 | 1053 | @item m32r-*-elf |
e9a25f70 JL |
1054 | Mitsubishi M32R processor. |
1055 | This configuration is intended for embedded systems. | |
65376d28 | 1056 | |
2284f91b DE |
1057 | @item m68000-hp-bsd |
1058 | HP 9000 series 200 running BSD. Note that the C compiler that comes | |
1059 | with this system cannot compile GNU CC; contact @code{law@@cs.utah.edu} | |
1060 | to get binaries of GNU CC for bootstrapping. | |
1061 | ||
1062 | @item m68k-altos | |
1063 | Altos 3068. You must use the GNU assembler, linker and debugger. | |
1064 | Also, you must fix a kernel bug. Details in the file @file{README.ALTOS}. | |
1065 | ||
1066 | @item m68k-apple-aux | |
1067 | Apple Macintosh running A/UX. | |
1068 | You may configure GCC to use either the system assembler and | |
1069 | linker or the GNU assembler and linker. You should use the GNU configuration | |
1070 | if you can, especially if you also want to use GNU C++. You enabled | |
1071 | that configuration with + the @samp{--with-gnu-as} and @samp{--with-gnu-ld} | |
1072 | options to @code{configure}. | |
1073 | ||
1074 | Note the C compiler that comes | |
1075 | with this system cannot compile GNU CC. You can fine binaries of GNU CC | |
1076 | for bootstrapping on @code{jagubox.gsfc.nasa.gov}. | |
1077 | You will also a patched version of @file{/bin/ld} there that | |
1078 | raises some of the arbitrary limits found in the original. | |
1079 | ||
1080 | @item m68k-att-sysv | |
1081 | AT&T 3b1, a.k.a. 7300 PC. Special procedures are needed to compile GNU | |
1082 | CC with this machine's standard C compiler, due to bugs in that | |
1083 | compiler. You can bootstrap it more easily with | |
1084 | previous versions of GNU CC if you have them. | |
1085 | ||
1086 | Installing GNU CC on the 3b1 is difficult if you do not already have | |
1087 | GNU CC running, due to bugs in the installed C compiler. However, | |
1088 | the following procedure might work. We are unable to test it. | |
1089 | ||
1090 | @enumerate | |
1091 | @item | |
e9a25f70 | 1092 | Comment out the @samp{#include "config.h"} line near the start of |
2284f91b DE |
1093 | @file{cccp.c} and do @samp{make cpp}. This makes a preliminary version |
1094 | of GNU cpp. | |
1095 | ||
1096 | @item | |
1097 | Save the old @file{/lib/cpp} and copy the preliminary GNU cpp to that | |
1098 | file name. | |
1099 | ||
1100 | @item | |
1101 | Undo your change in @file{cccp.c}, or reinstall the original version, | |
1102 | and do @samp{make cpp} again. | |
1103 | ||
1104 | @item | |
1105 | Copy this final version of GNU cpp into @file{/lib/cpp}. | |
1106 | ||
1107 | @findex obstack_free | |
1108 | @item | |
1109 | Replace every occurrence of @code{obstack_free} in the file | |
1110 | @file{tree.c} with @code{_obstack_free}. | |
1111 | ||
1112 | @item | |
1113 | Run @code{make} to get the first-stage GNU CC. | |
1114 | ||
1115 | @item | |
1116 | Reinstall the original version of @file{/lib/cpp}. | |
1117 | ||
1118 | @item | |
1119 | Now you can compile GNU CC with itself and install it in the normal | |
1120 | fashion. | |
1121 | @end enumerate | |
1122 | ||
1123 | @item m68k-bull-sysv | |
1124 | Bull DPX/2 series 200 and 300 with BOS-2.00.45 up to BOS-2.01. GNU CC works | |
1125 | either with native assembler or GNU assembler. You can use | |
1126 | GNU assembler with native coff generation by providing @samp{--with-gnu-as} to | |
1127 | the configure script or use GNU assembler with dbx-in-coff encapsulation | |
1128 | by providing @samp{--with-gnu-as --stabs}. For any problem with native | |
1129 | assembler or for availability of the DPX/2 port of GAS, contact | |
1130 | @code{F.Pierresteguy@@frcl.bull.fr}. | |
1131 | ||
1132 | @item m68k-crds-unox | |
1133 | Use @samp{configure unos} for building on Unos. | |
1134 | ||
1135 | The Unos assembler is named @code{casm} instead of @code{as}. For some | |
1136 | strange reason linking @file{/bin/as} to @file{/bin/casm} changes the | |
1137 | behavior, and does not work. So, when installing GNU CC, you should | |
1138 | install the following script as @file{as} in the subdirectory where | |
1139 | the passes of GCC are installed: | |
1140 | ||
1141 | @example | |
1142 | #!/bin/sh | |
1143 | casm $* | |
1144 | @end example | |
1145 | ||
1146 | The default Unos library is named @file{libunos.a} instead of | |
1147 | @file{libc.a}. To allow GNU CC to function, either change all | |
1148 | references to @samp{-lc} in @file{gcc.c} to @samp{-lunos} or link | |
1149 | @file{/lib/libc.a} to @file{/lib/libunos.a}. | |
1150 | ||
1151 | @cindex @code{alloca}, for Unos | |
1152 | When compiling GNU CC with the standard compiler, to overcome bugs in | |
1153 | the support of @code{alloca}, do not use @samp{-O} when making stage 2. | |
1154 | Then use the stage 2 compiler with @samp{-O} to make the stage 3 | |
1155 | compiler. This compiler will have the same characteristics as the usual | |
1156 | stage 2 compiler on other systems. Use it to make a stage 4 compiler | |
1157 | and compare that with stage 3 to verify proper compilation. | |
1158 | ||
1159 | (Perhaps simply defining @code{ALLOCA} in @file{x-crds} as described in | |
1160 | the comments there will make the above paragraph superfluous. Please | |
1161 | inform us of whether this works.) | |
1162 | ||
1163 | Unos uses memory segmentation instead of demand paging, so you will need | |
1164 | a lot of memory. 5 Mb is barely enough if no other tasks are running. | |
1165 | If linking @file{cc1} fails, try putting the object files into a library | |
1166 | and linking from that library. | |
1167 | ||
1168 | @item m68k-hp-hpux | |
1169 | HP 9000 series 300 or 400 running HP-UX. HP-UX version 8.0 has a bug in | |
1170 | the assembler that prevents compilation of GNU CC. To fix it, get patch | |
1171 | PHCO_4484 from HP. | |
1172 | ||
1173 | In addition, if you wish to use gas @samp{--with-gnu-as} you must use | |
1174 | gas version 2.1 or later, and you must use the GNU linker version 2.1 or | |
1175 | later. Earlier versions of gas relied upon a program which converted the | |
1176 | gas output into the native HP/UX format, but that program has not been | |
1177 | kept up to date. gdb does not understand that native HP/UX format, so | |
1178 | you must use gas if you wish to use gdb. | |
1179 | ||
1180 | @item m68k-sun | |
1181 | Sun 3. We do not provide a configuration file to use the Sun FPA by | |
1182 | default, because programs that establish signal handlers for floating | |
1183 | point traps inherently cannot work with the FPA. | |
1184 | ||
1185 | See @ref{Sun Install}, for information on installing GNU CC on Sun | |
1186 | systems. | |
1187 | ||
1188 | @item m88k-*-svr3 | |
1189 | Motorola m88k running the AT&T/Unisoft/Motorola V.3 reference port. | |
1190 | These systems tend to use the Green Hills C, revision 1.8.5, as the | |
1191 | standard C compiler. There are apparently bugs in this compiler that | |
1192 | result in object files differences between stage 2 and stage 3. If this | |
1193 | happens, make the stage 4 compiler and compare it to the stage 3 | |
1194 | compiler. If the stage 3 and stage 4 object files are identical, this | |
1195 | suggests you encountered a problem with the standard C compiler; the | |
1196 | stage 3 and 4 compilers may be usable. | |
1197 | ||
1198 | It is best, however, to use an older version of GNU CC for bootstrapping | |
1199 | if you have one. | |
1200 | ||
1201 | @item m88k-*-dgux | |
1202 | Motorola m88k running DG/UX. To build 88open BCS native or cross | |
1203 | compilers on DG/UX, specify the configuration name as | |
1204 | @samp{m88k-*-dguxbcs} and build in the 88open BCS software development | |
1205 | environment. To build ELF native or cross compilers on DG/UX, specify | |
1206 | @samp{m88k-*-dgux} and build in the DG/UX ELF development environment. | |
1207 | You set the software development environment by issuing | |
1208 | @samp{sde-target} command and specifying either @samp{m88kbcs} or | |
1209 | @samp{m88kdguxelf} as the operand. | |
1210 | ||
1211 | If you do not specify a configuration name, @file{configure} guesses the | |
1212 | configuration based on the current software development environment. | |
1213 | ||
1214 | @item m88k-tektronix-sysv3 | |
1215 | Tektronix XD88 running UTekV 3.2e. Do not turn on | |
1216 | optimization while building stage1 if you bootstrap with | |
1217 | the buggy Green Hills compiler. Also, The bundled LAI | |
1218 | System V NFS is buggy so if you build in an NFS mounted | |
1219 | directory, start from a fresh reboot, or avoid NFS all together. | |
1220 | Otherwise you may have trouble getting clean comparisons | |
1221 | between stages. | |
1222 | ||
1223 | @item mips-mips-bsd | |
1224 | MIPS machines running the MIPS operating system in BSD mode. It's | |
1225 | possible that some old versions of the system lack the functions | |
1226 | @code{memcpy}, @code{memcmp}, and @code{memset}. If your system lacks | |
1227 | these, you must remove or undo the definition of | |
1228 | @code{TARGET_MEM_FUNCTIONS} in @file{mips-bsd.h}. | |
1229 | ||
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}. | |
1238 | ||
1239 | @item mips-mips-riscos* | |
1240 | The MIPS C compiler needs to be told to increase its table size | |
1241 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
1242 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
1243 | optimization option, you also need to use @samp{-Olimit 3000}. | |
1244 | Both of these options are automatically generated in the | |
1245 | @file{Makefile} that the shell script @file{configure} builds. | |
1246 | If you override the @code{CC} make variable and use the MIPS | |
1247 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
1248 | ||
1249 | MIPS computers running RISC-OS can support four different | |
1250 | personalities: default, BSD 4.3, System V.3, and System V.4 | |
1251 | (older versions of RISC-OS don't support V.4). To configure GCC | |
1252 | for these platforms use the following configurations: | |
1253 | ||
1254 | @table @samp | |
1255 | @item mips-mips-riscos@code{rev} | |
1256 | Default configuration for RISC-OS, revision @code{rev}. | |
1257 | ||
1258 | @item mips-mips-riscos@code{rev}bsd | |
1259 | BSD 4.3 configuration for RISC-OS, revision @code{rev}. | |
1260 | ||
1261 | @item mips-mips-riscos@code{rev}sysv4 | |
1262 | System V.4 configuration for RISC-OS, revision @code{rev}. | |
1263 | ||
1264 | @item mips-mips-riscos@code{rev}sysv | |
1265 | System V.3 configuration for RISC-OS, revision @code{rev}. | |
1266 | @end table | |
1267 | ||
1268 | The revision @code{rev} mentioned above is the revision of | |
1269 | RISC-OS to use. You must reconfigure GCC when going from a | |
1270 | RISC-OS revision 4 to RISC-OS revision 5. This has the effect of | |
1271 | avoiding a linker | |
1272 | @ifclear INSTALLONLY | |
1273 | bug (see @ref{Installation Problems}, for more details). | |
1274 | @end ifclear | |
1275 | @ifset INSTALLONLY | |
1276 | bug. | |
1277 | @end ifset | |
1278 | ||
1279 | @item mips-sgi-* | |
1280 | In order to compile GCC on an SGI running IRIX 4, the "c.hdr.lib" | |
1281 | option must be installed from the CD-ROM supplied from Silicon Graphics. | |
1282 | This is found on the 2nd CD in release 4.0.1. | |
1283 | ||
1284 | In order to compile GCC on an SGI running IRIX 5, the "compiler_dev.hdr" | |
1285 | subsystem must be installed from the IDO CD-ROM supplied by Silicon | |
1286 | Graphics. | |
1287 | ||
1288 | @code{make compare} may fail on version 5 of IRIX unless you add | |
1289 | @samp{-save-temps} to @code{CFLAGS}. On these systems, the name of the | |
1290 | assembler input file is stored in the object file, and that makes | |
1291 | comparison fail if it differs between the @code{stage1} and | |
1292 | @code{stage2} compilations. The option @samp{-save-temps} forces a | |
1293 | fixed name to be used for the assembler input file, instead of a | |
1294 | randomly chosen name in @file{/tmp}. Do not add @samp{-save-temps} | |
1295 | unless the comparisons fail without that option. If you do you | |
1296 | @samp{-save-temps}, you will have to manually delete the @samp{.i} and | |
1297 | @samp{.s} files after each series of compilations. | |
1298 | ||
1299 | The MIPS C compiler needs to be told to increase its table size | |
1300 | for switch statements with the @samp{-Wf,-XNg1500} option in | |
1301 | order to compile @file{cp/parse.c}. If you use the @samp{-O2} | |
1302 | optimization option, you also need to use @samp{-Olimit 3000}. | |
1303 | Both of these options are automatically generated in the | |
1304 | @file{Makefile} that the shell script @file{configure} builds. | |
1305 | If you override the @code{CC} make variable and use the MIPS | |
1306 | compilers, you may need to add @samp{-Wf,-XNg1500 -Olimit 3000}. | |
1307 | ||
1308 | On Irix version 4.0.5F, and perhaps on some other versions as well, | |
1309 | there is an assembler bug that reorders instructions incorrectly. To | |
1310 | work around it, specify the target configuration | |
1311 | @samp{mips-sgi-irix4loser}. This configuration inhibits assembler | |
1312 | optimization. | |
1313 | ||
1314 | In a compiler configured with target @samp{mips-sgi-irix4}, you can turn | |
1315 | off assembler optimization by using the @samp{-noasmopt} option. This | |
1316 | compiler option passes the option @samp{-O0} to the assembler, to | |
1317 | inhibit reordering. | |
1318 | ||
1319 | The @samp{-noasmopt} option can be useful for testing whether a problem | |
1320 | is due to erroneous assembler reordering. Even if a problem does not go | |
1321 | away with @samp{-noasmopt}, it may still be due to assembler | |
1322 | reordering---perhaps GNU CC itself was miscompiled as a result. | |
1323 | ||
1324 | To enable debugging under Irix 5, you must use GNU as 2.5 or later, | |
1325 | and use the @samp{--with-gnu-as} configure option when configuring gcc. | |
1326 | GNU as is distributed as part of the binutils package. | |
1327 | ||
1328 | @item mips-sony-sysv | |
1329 | Sony MIPS NEWS. This works in NEWSOS 5.0.1, but not in 5.0.2 (which | |
1330 | uses ELF instead of COFF). Support for 5.0.2 will probably be provided | |
1331 | soon by volunteers. In particular, the linker does not like the | |
1332 | code generated by GCC when shared libraries are linked in. | |
1333 | ||
1334 | @item ns32k-encore | |
1335 | Encore ns32000 system. Encore systems are supported only under BSD. | |
1336 | ||
1337 | @item ns32k-*-genix | |
1338 | National Semiconductor ns32000 system. Genix has bugs in @code{alloca} | |
1339 | and @code{malloc}; you must get the compiled versions of these from GNU | |
1340 | Emacs. | |
1341 | ||
1342 | @item ns32k-sequent | |
1343 | Go to the Berkeley universe before compiling. In addition, you probably | |
1344 | need to create a file named @file{string.h} containing just one line: | |
1345 | @samp{#include <strings.h>}. | |
1346 | ||
1347 | @item ns32k-utek | |
1348 | UTEK ns32000 system (``merlin''). The C compiler that comes with this | |
1349 | system cannot compile GNU CC; contact @samp{tektronix!reed!mason} to get | |
1350 | binaries of GNU CC for bootstrapping. | |
1351 | ||
1352 | @item romp-*-aos | |
1353 | @itemx romp-*-mach | |
1354 | The only operating systems supported for the IBM RT PC are AOS and | |
1355 | MACH. GNU CC does not support AIX running on the RT. We recommend you | |
1356 | compile GNU CC with an earlier version of itself; if you compile GNU CC | |
1357 | with @code{hc}, the Metaware compiler, it will work, but you will get | |
1358 | mismatches between the stage 2 and stage 3 compilers in various files. | |
1359 | These errors are minor differences in some floating-point constants and | |
1360 | can be safely ignored; the stage 3 compiler is correct. | |
1361 | ||
1362 | @item rs6000-*-aix | |
1363 | @itemx powerpc-*-aix | |
1364 | Various early versions of each release of the IBM XLC compiler will not | |
1365 | bootstrap GNU CC. Symptoms include differences between the stage2 and | |
1366 | stage3 object files, and errors when compiling @file{libgcc.a} or | |
1367 | @file{enquire}. Known problematic releases include: xlc-1.2.1.8, | |
1368 | xlc-1.3.0.0 (distributed with AIX 3.2.5), and xlc-1.3.0.19. Both | |
1369 | xlc-1.2.1.28 and xlc-1.3.0.24 (PTF 432238) are known to produce working | |
1370 | versions of GNU CC, but most other recent releases correctly bootstrap | |
1371 | GNU CC. Also, releases of AIX prior to AIX 3.2.4 include a version of | |
1372 | the IBM assembler which does not accept debugging directives: assembler | |
1373 | updates are available as PTFs. Also, if you are using AIX 3.2.5 or | |
1374 | greater and the GNU assembler, you must have a version modified after | |
1375 | October 16th, 1995 in order for the GNU C compiler to build. See the | |
1376 | file @file{README.RS6000} for more details on of these problems. | |
1377 | ||
1378 | GNU CC does not yet support the 64-bit PowerPC instructions. | |
1379 | ||
1380 | Objective C does not work on this architecture because it makes assumptions | |
1381 | that are incompatible with the calling conventions. | |
1382 | ||
1383 | AIX on the RS/6000 provides support (NLS) for environments outside of | |
1384 | the United States. Compilers and assemblers use NLS to support | |
1385 | locale-specific representations of various objects including | |
1386 | floating-point numbers ("." vs "," for separating decimal fractions). | |
1387 | There have been problems reported where the library linked with GNU CC | |
1388 | does not produce the same floating-point formats that the assembler | |
1389 | accepts. If you have this problem, set the LANG environment variable to | |
1390 | "C" or "En_US". | |
1391 | ||
1392 | Due to changes in the way that GNU CC invokes the binder (linker) for AIX | |
1393 | 4.1, you may now receive warnings of duplicate symbols from the link step | |
1394 | that were not reported before. The assembly files generated by GNU CC for | |
1395 | AIX have always included multiple symbol definitions for certain global | |
1396 | variable and function declarations in the original program. The warnings | |
1397 | should not prevent the linker from producing a correct library or runnable | |
1398 | executable. | |
1399 | ||
1400 | By default, AIX 4.1 produces code that can be used on either Power or | |
1401 | PowerPC processors. | |
1402 | ||
1403 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1404 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1405 | ||
1406 | @item powerpc-*-elf | |
1407 | @itemx powerpc-*-sysv4 | |
1408 | PowerPC system in big endian mode, running System V.4. | |
1409 | ||
1410 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1411 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1412 | ||
861bb6c1 JL |
1413 | @item powerpc-*-linux-gnu |
1414 | PowerPC system in big endian mode, running the Linux-based GNU system. | |
2284f91b DE |
1415 | |
1416 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1417 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1418 | ||
1419 | @item powerpc-*-eabiaix | |
1420 | Embedded PowerPC system in big endian mode with -mcall-aix selected as | |
1421 | the default. | |
1422 | ||
1423 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1424 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1425 | ||
1426 | @item powerpc-*-eabisim | |
1427 | Embedded PowerPC system in big endian mode for use in running under the | |
1428 | PSIM simulator. | |
1429 | ||
1430 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1431 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1432 | ||
1433 | @item powerpc-*-eabi | |
1434 | Embedded PowerPC system in big endian mode. | |
1435 | ||
1436 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1437 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1438 | ||
1439 | @item powerpcle-*-elf | |
1440 | @itemx powerpcle-*-sysv4 | |
1441 | PowerPC system in little endian mode, running System V.4. | |
1442 | ||
1443 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1444 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1445 | ||
1446 | @item powerpcle-*-solaris2* | |
1447 | PowerPC system in little endian mode, running Solaris 2.5.1 or higher. | |
1448 | ||
1449 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1450 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1451 | Beta versions of the Sun 4.0 compiler do not seem to be able to build | |
1452 | GNU CC correctly. There are also problems with the host assembler and | |
1453 | linker that are fixed by using the GNU versions of these tools. | |
1454 | ||
1455 | @item powerpcle-*-eabisim | |
1456 | Embedded PowerPC system in little endian mode for use in running under | |
1457 | the PSIM simulator. | |
1458 | ||
1459 | @itemx powerpcle-*-eabi | |
1460 | Embedded PowerPC system in little endian mode. | |
1461 | ||
1462 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1463 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1464 | ||
1465 | @item powerpcle-*-winnt | |
1466 | @itemx powerpcle-*-pe | |
1467 | PowerPC system in little endian mode running Windows NT. | |
1468 | ||
1469 | You can specify a default version for the @samp{-mcpu=}@var{cpu_type} | |
1470 | switch by using the configure option @samp{--with-cpu-}@var{cpu_type}. | |
1471 | ||
1472 | @item vax-dec-ultrix | |
1473 | Don't try compiling with Vax C (@code{vcc}). It produces incorrect code | |
1474 | in some cases (for example, when @code{alloca} is used). | |
1475 | ||
1476 | Meanwhile, compiling @file{cp/parse.c} with pcc does not work because of | |
1477 | an internal table size limitation in that compiler. To avoid this | |
1478 | problem, compile just the GNU C compiler first, and use it to recompile | |
1479 | building all the languages that you want to run. | |
1480 | ||
1481 | @item sparc-sun-* | |
2284f91b DE |
1482 | See @ref{Sun Install}, for information on installing GNU CC on Sun |
1483 | systems. | |
1484 | ||
1485 | @item vax-dec-vms | |
1486 | See @ref{VMS Install}, for details on how to install GNU CC on VMS. | |
1487 | ||
1488 | @item we32k-*-* | |
1489 | These computers are also known as the 3b2, 3b5, 3b20 and other similar | |
1490 | names. (However, the 3b1 is actually a 68000; see | |
1491 | @ref{Configurations}.) | |
1492 | ||
1493 | Don't use @samp{-g} when compiling with the system's compiler. The | |
1494 | system's linker seems to be unable to handle such a large program with | |
1495 | debugging information. | |
1496 | ||
1497 | The system's compiler runs out of capacity when compiling @file{stmt.c} | |
1498 | in GNU CC. You can work around this by building @file{cpp} in GNU CC | |
1499 | first, then use that instead of the system's preprocessor with the | |
1500 | system's C compiler to compile @file{stmt.c}. Here is how: | |
1501 | ||
1502 | @smallexample | |
1503 | mv /lib/cpp /lib/cpp.att | |
1504 | cp cpp /lib/cpp.gnu | |
1505 | echo '/lib/cpp.gnu -traditional $@{1+"$@@"@}' > /lib/cpp | |
1506 | chmod +x /lib/cpp | |
1507 | @end smallexample | |
1508 | ||
1509 | The system's compiler produces bad code for some of the GNU CC | |
1510 | optimization files. So you must build the stage 2 compiler without | |
1511 | optimization. Then build a stage 3 compiler with optimization. | |
1512 | That executable should work. Here are the necessary commands: | |
1513 | ||
1514 | @smallexample | |
1515 | make LANGUAGES=c CC=stage1/xgcc CFLAGS="-Bstage1/ -g" | |
1516 | make stage2 | |
1517 | make CC=stage2/xgcc CFLAGS="-Bstage2/ -g -O" | |
1518 | @end smallexample | |
1519 | ||
1520 | You may need to raise the ULIMIT setting to build a C++ compiler, | |
1521 | as the file @file{cc1plus} is larger than one megabyte. | |
1522 | @end table | |
1523 | ||
1524 | @node Other Dir | |
1525 | @section Compilation in a Separate Directory | |
1526 | @cindex other directory, compilation in | |
1527 | @cindex compilation in a separate directory | |
1528 | @cindex separate directory, compilation in | |
1529 | ||
1530 | If you wish to build the object files and executables in a directory | |
1531 | other than the one containing the source files, here is what you must | |
1532 | do differently: | |
1533 | ||
1534 | @enumerate | |
1535 | @item | |
1536 | Make sure you have a version of Make that supports the @code{VPATH} | |
1537 | feature. (GNU Make supports it, as do Make versions on most BSD | |
1538 | systems.) | |
1539 | ||
1540 | @item | |
1541 | If you have ever run @file{configure} in the source directory, you must undo | |
1542 | the configuration. Do this by running: | |
1543 | ||
1544 | @example | |
1545 | make distclean | |
1546 | @end example | |
1547 | ||
1548 | @item | |
1549 | Go to the directory in which you want to build the compiler before | |
1550 | running @file{configure}: | |
1551 | ||
1552 | @example | |
1553 | mkdir gcc-sun3 | |
1554 | cd gcc-sun3 | |
1555 | @end example | |
1556 | ||
1557 | On systems that do not support symbolic links, this directory must be | |
1558 | on the same file system as the source code directory. | |
1559 | ||
1560 | @item | |
1561 | Specify where to find @file{configure} when you run it: | |
1562 | ||
1563 | @example | |
1564 | ../gcc/configure @dots{} | |
1565 | @end example | |
1566 | ||
1567 | This also tells @code{configure} where to find the compiler sources; | |
1568 | @code{configure} takes the directory from the file name that was used to | |
1569 | invoke it. But if you want to be sure, you can specify the source | |
1570 | directory with the @samp{--srcdir} option, like this: | |
1571 | ||
1572 | @example | |
1573 | ../gcc/configure --srcdir=../gcc @var{other options} | |
1574 | @end example | |
1575 | ||
1576 | The directory you specify with @samp{--srcdir} need not be the same | |
1577 | as the one that @code{configure} is found in. | |
1578 | @end enumerate | |
1579 | ||
1580 | Now, you can run @code{make} in that directory. You need not repeat the | |
1581 | configuration steps shown above, when ordinary source files change. You | |
1582 | must, however, run @code{configure} again when the configuration files | |
1583 | change, if your system does not support symbolic links. | |
1584 | ||
1585 | @node Cross-Compiler | |
1586 | @section Building and Installing a Cross-Compiler | |
1587 | @cindex cross-compiler, installation | |
1588 | ||
1589 | GNU CC can function as a cross-compiler for many machines, but not all. | |
1590 | ||
1591 | @itemize @bullet | |
1592 | @item | |
1593 | Cross-compilers for the Mips as target using the Mips assembler | |
1594 | currently do not work, because the auxiliary programs | |
1595 | @file{mips-tdump.c} and @file{mips-tfile.c} can't be compiled on | |
1596 | anything but a Mips. It does work to cross compile for a Mips | |
1597 | if you use the GNU assembler and linker. | |
1598 | ||
1599 | @item | |
1600 | Cross-compilers between machines with different floating point formats | |
1601 | have not all been made to work. GNU CC now has a floating point | |
1602 | emulator with which these can work, but each target machine description | |
1603 | needs to be updated to take advantage of it. | |
1604 | ||
1605 | @item | |
1606 | Cross-compilation between machines of different word sizes is | |
1607 | somewhat problematic and sometimes does not work. | |
1608 | @end itemize | |
1609 | ||
1610 | Since GNU CC generates assembler code, you probably need a | |
1611 | cross-assembler that GNU CC can run, in order to produce object files. | |
1612 | If you want to link on other than the target machine, you need a | |
1613 | cross-linker as well. You also need header files and libraries suitable | |
1614 | for the target machine that you can install on the host machine. | |
1615 | ||
1616 | @menu | |
1617 | * Steps of Cross:: Using a cross-compiler involves several steps | |
1618 | that may be carried out on different machines. | |
1619 | * Configure Cross:: Configuring a cross-compiler. | |
1620 | * Tools and Libraries:: Where to put the linker and assembler, and the C library. | |
1621 | * Cross Headers:: Finding and installing header files | |
1622 | for a cross-compiler. | |
1623 | * Cross Runtime:: Supplying arithmetic runtime routines (@file{libgcc1.a}). | |
1624 | * Build Cross:: Actually compiling the cross-compiler. | |
1625 | @end menu | |
1626 | ||
1627 | @node Steps of Cross | |
1628 | @subsection Steps of Cross-Compilation | |
1629 | ||
1630 | To compile and run a program using a cross-compiler involves several | |
1631 | steps: | |
1632 | ||
1633 | @itemize @bullet | |
1634 | @item | |
1635 | Run the cross-compiler on the host machine to produce assembler files | |
1636 | for the target machine. This requires header files for the target | |
1637 | machine. | |
1638 | ||
1639 | @item | |
1640 | Assemble the files produced by the cross-compiler. You can do this | |
1641 | either with an assembler on the target machine, or with a | |
1642 | cross-assembler on the host machine. | |
1643 | ||
1644 | @item | |
1645 | Link those files to make an executable. You can do this either with a | |
1646 | linker on the target machine, or with a cross-linker on the host | |
1647 | machine. Whichever machine you use, you need libraries and certain | |
1648 | startup files (typically @file{crt@dots{}.o}) for the target machine. | |
1649 | @end itemize | |
1650 | ||
1651 | It is most convenient to do all of these steps on the same host machine, | |
1652 | since then you can do it all with a single invocation of GNU CC. This | |
1653 | requires a suitable cross-assembler and cross-linker. For some targets, | |
1654 | the GNU assembler and linker are available. | |
1655 | ||
1656 | @node Configure Cross | |
1657 | @subsection Configuring a Cross-Compiler | |
1658 | ||
1659 | To build GNU CC as a cross-compiler, you start out by running | |
1660 | @file{configure}. Use the @samp{--target=@var{target}} to specify the | |
1661 | target type. If @file{configure} was unable to correctly identify the | |
1662 | system you are running on, also specify the @samp{--build=@var{build}} | |
1663 | option. For example, here is how to configure for a cross-compiler that | |
1664 | produces code for an HP 68030 system running BSD on a system that | |
1665 | @file{configure} can correctly identify: | |
1666 | ||
1667 | @smallexample | |
1668 | ./configure --target=m68k-hp-bsd4.3 | |
1669 | @end smallexample | |
1670 | ||
1671 | @node Tools and Libraries | |
1672 | @subsection Tools and Libraries for a Cross-Compiler | |
1673 | ||
1674 | If you have a cross-assembler and cross-linker available, you should | |
1675 | install them now. Put them in the directory | |
1676 | @file{/usr/local/@var{target}/bin}. Here is a table of the tools | |
1677 | you should put in this directory: | |
1678 | ||
1679 | @table @file | |
1680 | @item as | |
1681 | This should be the cross-assembler. | |
1682 | ||
1683 | @item ld | |
1684 | This should be the cross-linker. | |
1685 | ||
1686 | @item ar | |
1687 | This should be the cross-archiver: a program which can manipulate | |
1688 | archive files (linker libraries) in the target machine's format. | |
1689 | ||
1690 | @item ranlib | |
1691 | This should be a program to construct a symbol table in an archive file. | |
1692 | @end table | |
1693 | ||
1694 | The installation of GNU CC will find these programs in that directory, | |
1695 | and copy or link them to the proper place to for the cross-compiler to | |
1696 | find them when run later. | |
1697 | ||
1698 | The easiest way to provide these files is to build the Binutils package | |
1699 | and GAS. Configure them with the same @samp{--host} and @samp{--target} | |
1700 | options that you use for configuring GNU CC, then build and install | |
1701 | them. They install their executables automatically into the proper | |
1702 | directory. Alas, they do not support all the targets that GNU CC | |
1703 | supports. | |
1704 | ||
1705 | If you want to install libraries to use with the cross-compiler, such as | |
1706 | a standard C library, put them in the directory | |
1707 | @file{/usr/local/@var{target}/lib}; installation of GNU CC copies | |
1708 | all the files in that subdirectory into the proper place for GNU CC to | |
1709 | find them and link with them. Here's an example of copying some | |
1710 | libraries from a target machine: | |
1711 | ||
1712 | @example | |
1713 | ftp @var{target-machine} | |
1714 | lcd /usr/local/@var{target}/lib | |
1715 | cd /lib | |
1716 | get libc.a | |
1717 | cd /usr/lib | |
1718 | get libg.a | |
1719 | get libm.a | |
1720 | quit | |
1721 | @end example | |
1722 | ||
1723 | @noindent | |
1724 | The precise set of libraries you'll need, and their locations on | |
1725 | the target machine, vary depending on its operating system. | |
1726 | ||
1727 | @cindex start files | |
1728 | Many targets require ``start files'' such as @file{crt0.o} and | |
1729 | @file{crtn.o} which are linked into each executable; these too should be | |
1730 | placed in @file{/usr/local/@var{target}/lib}. There may be several | |
1731 | alternatives for @file{crt0.o}, for use with profiling or other | |
1732 | compilation options. Check your target's definition of | |
1733 | @code{STARTFILE_SPEC} to find out what start files it uses. | |
1734 | Here's an example of copying these files from a target machine: | |
1735 | ||
1736 | @example | |
1737 | ftp @var{target-machine} | |
1738 | lcd /usr/local/@var{target}/lib | |
1739 | prompt | |
1740 | cd /lib | |
1741 | mget *crt*.o | |
1742 | cd /usr/lib | |
1743 | mget *crt*.o | |
1744 | quit | |
1745 | @end example | |
1746 | ||
1747 | @node Cross Runtime | |
1748 | @subsection @file{libgcc.a} and Cross-Compilers | |
1749 | ||
1750 | Code compiled by GNU CC uses certain runtime support functions | |
1751 | implicitly. Some of these functions can be compiled successfully with | |
1752 | GNU CC itself, but a few cannot be. These problem functions are in the | |
1753 | source file @file{libgcc1.c}; the library made from them is called | |
1754 | @file{libgcc1.a}. | |
1755 | ||
1756 | When you build a native compiler, these functions are compiled with some | |
1757 | other compiler--the one that you use for bootstrapping GNU CC. | |
1758 | Presumably it knows how to open code these operations, or else knows how | |
1759 | to call the run-time emulation facilities that the machine comes with. | |
1760 | But this approach doesn't work for building a cross-compiler. The | |
1761 | compiler that you use for building knows about the host system, not the | |
1762 | target system. | |
1763 | ||
1764 | So, when you build a cross-compiler you have to supply a suitable | |
1765 | library @file{libgcc1.a} that does the job it is expected to do. | |
1766 | ||
1767 | To compile @file{libgcc1.c} with the cross-compiler itself does not | |
1768 | work. The functions in this file are supposed to implement arithmetic | |
1769 | operations that GNU CC does not know how to open code for your target | |
1770 | machine. If these functions are compiled with GNU CC itself, they | |
1771 | will compile into infinite recursion. | |
1772 | ||
1773 | On any given target, most of these functions are not needed. If GNU CC | |
1774 | can open code an arithmetic operation, it will not call these functions | |
1775 | to perform the operation. It is possible that on your target machine, | |
1776 | none of these functions is needed. If so, you can supply an empty | |
1777 | library as @file{libgcc1.a}. | |
1778 | ||
1779 | Many targets need library support only for multiplication and division. | |
1780 | If you are linking with a library that contains functions for | |
1781 | multiplication and division, you can tell GNU CC to call them directly | |
1782 | by defining the macros @code{MULSI3_LIBCALL}, and the like. These | |
1783 | macros need to be defined in the target description macro file. For | |
1784 | some targets, they are defined already. This may be sufficient to | |
1785 | avoid the need for libgcc1.a; if so, you can supply an empty library. | |
1786 | ||
1787 | Some targets do not have floating point instructions; they need other | |
1788 | functions in @file{libgcc1.a}, which do floating arithmetic. | |
1789 | Recent versions of GNU CC have a file which emulates floating point. | |
1790 | With a certain amount of work, you should be able to construct a | |
1791 | floating point emulator that can be used as @file{libgcc1.a}. Perhaps | |
1792 | future versions will contain code to do this automatically and | |
1793 | conveniently. That depends on whether someone wants to implement it. | |
1794 | ||
1795 | Some embedded targets come with all the necessary @file{libgcc1.a} | |
1796 | routines written in C or assembler. These targets build | |
1797 | @file{libgcc1.a} automatically and you do not need to do anything | |
1798 | special for them. Other embedded targets do not need any | |
1799 | @file{libgcc1.a} routines since all the necessary operations are | |
1800 | supported by the hardware. | |
1801 | ||
1802 | If your target system has another C compiler, you can configure GNU CC | |
1803 | as a native compiler on that machine, build just @file{libgcc1.a} with | |
1804 | @samp{make libgcc1.a} on that machine, and use the resulting file with | |
1805 | the cross-compiler. To do this, execute the following on the target | |
1806 | machine: | |
1807 | ||
1808 | @example | |
1809 | cd @var{target-build-dir} | |
1810 | ./configure --host=sparc --target=sun3 | |
1811 | make libgcc1.a | |
1812 | @end example | |
1813 | ||
1814 | @noindent | |
1815 | And then this on the host machine: | |
1816 | ||
1817 | @example | |
1818 | ftp @var{target-machine} | |
1819 | binary | |
1820 | cd @var{target-build-dir} | |
1821 | get libgcc1.a | |
1822 | quit | |
1823 | @end example | |
1824 | ||
1825 | Another way to provide the functions you need in @file{libgcc1.a} is to | |
1826 | define the appropriate @code{perform_@dots{}} macros for those | |
1827 | functions. If these definitions do not use the C arithmetic operators | |
1828 | that they are meant to implement, you should be able to compile them | |
1829 | with the cross-compiler you are building. (If these definitions already | |
1830 | exist for your target file, then you are all set.) | |
1831 | ||
1832 | To build @file{libgcc1.a} using the perform macros, use | |
1833 | @samp{LIBGCC1=libgcc1.a OLDCC=./xgcc} when building the compiler. | |
1834 | Otherwise, you should place your replacement library under the name | |
1835 | @file{libgcc1.a} in the directory in which you will build the | |
1836 | cross-compiler, before you run @code{make}. | |
1837 | ||
1838 | @node Cross Headers | |
1839 | @subsection Cross-Compilers and Header Files | |
1840 | ||
1841 | If you are cross-compiling a standalone program or a program for an | |
1842 | embedded system, then you may not need any header files except the few | |
1843 | that are part of GNU CC (and those of your program). However, if you | |
1844 | intend to link your program with a standard C library such as | |
1845 | @file{libc.a}, then you probably need to compile with the header files | |
1846 | that go with the library you use. | |
1847 | ||
1848 | The GNU C compiler does not come with these files, because (1) they are | |
1849 | system-specific, and (2) they belong in a C library, not in a compiler. | |
1850 | ||
1851 | If the GNU C library supports your target machine, then you can get the | |
1852 | header files from there (assuming you actually use the GNU library when | |
1853 | you link your program). | |
1854 | ||
1855 | If your target machine comes with a C compiler, it probably comes with | |
1856 | suitable header files also. If you make these files accessible from the host | |
1857 | machine, the cross-compiler can use them also. | |
1858 | ||
1859 | Otherwise, you're on your own in finding header files to use when | |
1860 | cross-compiling. | |
1861 | ||
1862 | When you have found suitable header files, put them in the directory | |
1863 | @file{/usr/local/@var{target}/include}, before building the cross | |
1864 | compiler. Then installation will run fixincludes properly and install | |
1865 | the corrected versions of the header files where the compiler will use | |
1866 | them. | |
1867 | ||
1868 | Provide the header files before you build the cross-compiler, because | |
1869 | the build stage actually runs the cross-compiler to produce parts of | |
1870 | @file{libgcc.a}. (These are the parts that @emph{can} be compiled with | |
1871 | GNU CC.) Some of them need suitable header files. | |
1872 | ||
1873 | Here's an example showing how to copy the header files from a target | |
1874 | machine. On the target machine, do this: | |
1875 | ||
1876 | @example | |
1877 | (cd /usr/include; tar cf - .) > tarfile | |
1878 | @end example | |
1879 | ||
1880 | Then, on the host machine, do this: | |
1881 | ||
1882 | @example | |
1883 | ftp @var{target-machine} | |
1884 | lcd /usr/local/@var{target}/include | |
1885 | get tarfile | |
1886 | quit | |
1887 | tar xf tarfile | |
1888 | @end example | |
1889 | ||
1890 | @node Build Cross | |
1891 | @subsection Actually Building the Cross-Compiler | |
1892 | ||
1893 | Now you can proceed just as for compiling a single-machine compiler | |
1894 | through the step of building stage 1. If you have not provided some | |
1895 | sort of @file{libgcc1.a}, then compilation will give up at the point | |
1896 | where it needs that file, printing a suitable error message. If you | |
1897 | do provide @file{libgcc1.a}, then building the compiler will automatically | |
1898 | compile and link a test program called @file{libgcc1-test}; if you get | |
1899 | errors in the linking, it means that not all of the necessary routines | |
1900 | in @file{libgcc1.a} are available. | |
1901 | ||
1902 | You must provide the header file @file{float.h}. One way to do this is | |
1903 | to compile @file{enquire} and run it on your target machine. The job of | |
1904 | @file{enquire} is to run on the target machine and figure out by | |
1905 | experiment the nature of its floating point representation. | |
1906 | @file{enquire} records its findings in the header file @file{float.h}. | |
1907 | If you can't produce this file by running @file{enquire} on the target | |
1908 | machine, then you will need to come up with a suitable @file{float.h} in | |
1909 | some other way (or else, avoid using it in your programs). | |
1910 | ||
1911 | Do not try to build stage 2 for a cross-compiler. It doesn't work to | |
1912 | rebuild GNU CC as a cross-compiler using the cross-compiler, because | |
1913 | that would produce a program that runs on the target machine, not on the | |
1914 | host. For example, if you compile a 386-to-68030 cross-compiler with | |
1915 | itself, the result will not be right either for the 386 (because it was | |
1916 | compiled into 68030 code) or for the 68030 (because it was configured | |
1917 | for a 386 as the host). If you want to compile GNU CC into 68030 code, | |
1918 | whether you compile it on a 68030 or with a cross-compiler on a 386, you | |
1919 | must specify a 68030 as the host when you configure it. | |
1920 | ||
1921 | To install the cross-compiler, use @samp{make install}, as usual. | |
1922 | ||
1923 | @node Sun Install | |
1924 | @section Installing GNU CC on the Sun | |
1925 | @cindex Sun installation | |
1926 | @cindex installing GNU CC on the Sun | |
1927 | ||
1928 | On Solaris (version 2.1), do not use the linker or other tools in | |
1929 | @file{/usr/ucb} to build GNU CC. Use @code{/usr/ccs/bin}. | |
1930 | ||
1931 | Make sure the environment variable @code{FLOAT_OPTION} is not set when | |
1932 | you compile @file{libgcc.a}. If this option were set to @code{f68881} | |
1933 | when @file{libgcc.a} is compiled, the resulting code would demand to be | |
1934 | linked with a special startup file and would not link properly without | |
1935 | special pains. | |
1936 | ||
1937 | @cindex @code{alloca}, for SunOS | |
1938 | There is a bug in @code{alloca} in certain versions of the Sun library. | |
1939 | To avoid this bug, install the binaries of GNU CC that were compiled by | |
1940 | GNU CC. They use @code{alloca} as a built-in function and never the one | |
1941 | in the library. | |
1942 | ||
1943 | Some versions of the Sun compiler crash when compiling GNU CC. The | |
1944 | problem is a segmentation fault in cpp. This problem seems to be due to | |
1945 | the bulk of data in the environment variables. You may be able to avoid | |
1946 | it by using the following command to compile GNU CC with Sun CC: | |
1947 | ||
1948 | @example | |
1949 | make CC="TERMCAP=x OBJS=x LIBFUNCS=x STAGESTUFF=x cc" | |
1950 | @end example | |
1951 | ||
1952 | SunOS 4.1.3 and 4.1.3_U1 have bugs that can cause intermittent core | |
1953 | dumps when compiling GNU CC. A common symptom is an | |
1954 | internal compiler error which does not recur if you run it again. | |
1955 | To fix the problem, install Sun recommended patch 100726 (for SunOS 4.1.3) | |
1956 | or 101508 (for SunOS 4.1.3_U1), or upgrade to a later SunOS release. | |
1957 | ||
1958 | @node VMS Install | |
1959 | @section Installing GNU CC on VMS | |
1960 | @cindex VMS installation | |
1961 | @cindex installing GNU CC on VMS | |
1962 | ||
1963 | The VMS version of GNU CC is distributed in a backup saveset containing | |
1964 | both source code and precompiled binaries. | |
1965 | ||
1966 | To install the @file{gcc} command so you can use the compiler easily, in | |
1967 | the same manner as you use the VMS C compiler, you must install the VMS CLD | |
1968 | file for GNU CC as follows: | |
1969 | ||
1970 | @enumerate | |
1971 | @item | |
1972 | Define the VMS logical names @samp{GNU_CC} and @samp{GNU_CC_INCLUDE} | |
1973 | to point to the directories where the GNU CC executables | |
1974 | (@file{gcc-cpp.exe}, @file{gcc-cc1.exe}, etc.) and the C include files are | |
1975 | kept respectively. This should be done with the commands:@refill | |
1976 | ||
1977 | @smallexample | |
1978 | $ assign /system /translation=concealed - | |
1979 | disk:[gcc.] gnu_cc | |
1980 | $ assign /system /translation=concealed - | |
1981 | disk:[gcc.include.] gnu_cc_include | |
1982 | @end smallexample | |
1983 | ||
1984 | @noindent | |
1985 | with the appropriate disk and directory names. These commands can be | |
1986 | placed in your system startup file so they will be executed whenever | |
1987 | the machine is rebooted. You may, if you choose, do this via the | |
1988 | @file{GCC_INSTALL.COM} script in the @file{[GCC]} directory. | |
1989 | ||
1990 | @item | |
1991 | Install the @file{GCC} command with the command line: | |
1992 | ||
1993 | @smallexample | |
1994 | $ set command /table=sys$common:[syslib]dcltables - | |
1995 | /output=sys$common:[syslib]dcltables gnu_cc:[000000]gcc | |
1996 | $ install replace sys$common:[syslib]dcltables | |
1997 | @end smallexample | |
1998 | ||
1999 | @item | |
2000 | To install the help file, do the following: | |
2001 | ||
2002 | @smallexample | |
2003 | $ library/help sys$library:helplib.hlb gcc.hlp | |
2004 | @end smallexample | |
2005 | ||
2006 | @noindent | |
2007 | Now you can invoke the compiler with a command like @samp{gcc /verbose | |
2008 | file.c}, which is equivalent to the command @samp{gcc -v -c file.c} in | |
2009 | Unix. | |
2010 | @end enumerate | |
2011 | ||
2012 | If you wish to use GNU C++ you must first install GNU CC, and then | |
2013 | perform the following steps: | |
2014 | ||
2015 | @enumerate | |
2016 | @item | |
2017 | Define the VMS logical name @samp{GNU_GXX_INCLUDE} to point to the | |
2018 | directory where the preprocessor will search for the C++ header files. | |
2019 | This can be done with the command:@refill | |
2020 | ||
2021 | @smallexample | |
2022 | $ assign /system /translation=concealed - | |
2023 | disk:[gcc.gxx_include.] gnu_gxx_include | |
2024 | @end smallexample | |
2025 | ||
2026 | @noindent | |
2027 | with the appropriate disk and directory name. If you are going to be | |
2028 | using libg++, this is where the libg++ install procedure will install | |
2029 | the libg++ header files. | |
2030 | ||
2031 | @item | |
2032 | Obtain the file @file{gcc-cc1plus.exe}, and place this in the same | |
2033 | directory that @file{gcc-cc1.exe} is kept. | |
2034 | ||
2035 | The GNU C++ compiler can be invoked with a command like @samp{gcc /plus | |
2036 | /verbose file.cc}, which is equivalent to the command @samp{g++ -v -c | |
2037 | file.cc} in Unix. | |
2038 | @end enumerate | |
2039 | ||
2040 | We try to put corresponding binaries and sources on the VMS distribution | |
2041 | tape. But sometimes the binaries will be from an older version than the | |
2042 | sources, because we don't always have time to update them. (Use the | |
2043 | @samp{/version} option to determine the version number of the binaries and | |
2044 | compare it with the source file @file{version.c} to tell whether this is | |
2045 | so.) In this case, you should use the binaries you get to recompile the | |
2046 | sources. If you must recompile, here is how: | |
2047 | ||
2048 | @enumerate | |
2049 | @item | |
2050 | Execute the command procedure @file{vmsconfig.com} to set up the files | |
2051 | @file{tm.h}, @file{config.h}, @file{aux-output.c}, and @file{md.}, and | |
2052 | to create files @file{tconfig.h} and @file{hconfig.h}. This procedure | |
2053 | also creates several linker option files used by @file{make-cc1.com} and | |
2054 | a data file used by @file{make-l2.com}.@refill | |
2055 | ||
2056 | @smallexample | |
2057 | $ @@vmsconfig.com | |
2058 | @end smallexample | |
2059 | ||
2060 | @item | |
2061 | Setup the logical names and command tables as defined above. In | |
2062 | addition, define the VMS logical name @samp{GNU_BISON} to point at the | |
2063 | to the directories where the Bison executable is kept. This should be | |
2064 | done with the command:@refill | |
2065 | ||
2066 | @smallexample | |
2067 | $ assign /system /translation=concealed - | |
2068 | disk:[bison.] gnu_bison | |
2069 | @end smallexample | |
2070 | ||
2071 | You may, if you choose, use the @file{INSTALL_BISON.COM} script in the | |
2072 | @file{[BISON]} directory. | |
2073 | ||
2074 | @item | |
2075 | Install the @samp{BISON} command with the command line:@refill | |
2076 | ||
2077 | @smallexample | |
2078 | $ set command /table=sys$common:[syslib]dcltables - | |
2079 | /output=sys$common:[syslib]dcltables - | |
2080 | gnu_bison:[000000]bison | |
2081 | $ install replace sys$common:[syslib]dcltables | |
2082 | @end smallexample | |
2083 | ||
2084 | @item | |
2085 | Type @samp{@@make-gcc} to recompile everything (alternatively, submit | |
2086 | the file @file{make-gcc.com} to a batch queue). If you wish to build | |
2087 | the GNU C++ compiler as well as the GNU CC compiler, you must first edit | |
2088 | @file{make-gcc.com} and follow the instructions that appear in the | |
2089 | comments.@refill | |
2090 | ||
2091 | @item | |
2092 | In order to use GCC, you need a library of functions which GCC compiled code | |
2093 | will call to perform certain tasks, and these functions are defined in the | |
2094 | file @file{libgcc2.c}. To compile this you should use the command procedure | |
2095 | @file{make-l2.com}, which will generate the library @file{libgcc2.olb}. | |
2096 | @file{libgcc2.olb} should be built using the compiler built from | |
2097 | the same distribution that @file{libgcc2.c} came from, and | |
2098 | @file{make-gcc.com} will automatically do all of this for you. | |
2099 | ||
2100 | To install the library, use the following commands:@refill | |
2101 | ||
2102 | @smallexample | |
2103 | $ library gnu_cc:[000000]gcclib/delete=(new,eprintf) | |
2104 | $ library gnu_cc:[000000]gcclib/delete=L_* | |
2105 | $ library libgcc2/extract=*/output=libgcc2.obj | |
2106 | $ library gnu_cc:[000000]gcclib libgcc2.obj | |
2107 | @end smallexample | |
2108 | ||
2109 | The first command simply removes old modules that will be replaced with | |
2110 | modules from @file{libgcc2} under different module names. The modules | |
2111 | @code{new} and @code{eprintf} may not actually be present in your | |
2112 | @file{gcclib.olb}---if the VMS librarian complains about those modules | |
2113 | not being present, simply ignore the message and continue on with the | |
2114 | next command. The second command removes the modules that came from the | |
2115 | previous version of the library @file{libgcc2.c}. | |
2116 | ||
2117 | Whenever you update the compiler on your system, you should also update the | |
2118 | library with the above procedure. | |
2119 | ||
2120 | @item | |
2121 | You may wish to build GCC in such a way that no files are written to the | |
2122 | directory where the source files reside. An example would be the when | |
2123 | the source files are on a read-only disk. In these cases, execute the | |
2124 | following DCL commands (substituting your actual path names): | |
2125 | ||
2126 | @smallexample | |
2127 | $ assign dua0:[gcc.build_dir.]/translation=concealed, - | |
2128 | dua1:[gcc.source_dir.]/translation=concealed gcc_build | |
2129 | $ set default gcc_build:[000000] | |
2130 | @end smallexample | |
2131 | ||
2132 | @noindent | |
2133 | where the directory @file{dua1:[gcc.source_dir]} contains the source | |
2134 | code, and the directory @file{dua0:[gcc.build_dir]} is meant to contain | |
2135 | all of the generated object files and executables. Once you have done | |
2136 | this, you can proceed building GCC as described above. (Keep in mind | |
2137 | that @file{gcc_build} is a rooted logical name, and thus the device | |
2138 | names in each element of the search list must be an actual physical | |
2139 | device name rather than another rooted logical name). | |
2140 | ||
2141 | @item | |
2142 | @strong{If you are building GNU CC with a previous version of GNU CC, | |
2143 | you also should check to see that you have the newest version of the | |
2144 | assembler}. In particular, GNU CC version 2 treats global constant | |
2145 | variables slightly differently from GNU CC version 1, and GAS version | |
2146 | 1.38.1 does not have the patches required to work with GCC version 2. | |
2147 | If you use GAS 1.38.1, then @code{extern const} variables will not have | |
2148 | the read-only bit set, and the linker will generate warning messages | |
2149 | about mismatched psect attributes for these variables. These warning | |
2150 | messages are merely a nuisance, and can safely be ignored. | |
2151 | ||
2152 | If you are compiling with a version of GNU CC older than 1.33, specify | |
2153 | @samp{/DEFINE=("inline=")} as an option in all the compilations. This | |
2154 | requires editing all the @code{gcc} commands in @file{make-cc1.com}. | |
2155 | (The older versions had problems supporting @code{inline}.) Once you | |
2156 | have a working 1.33 or newer GNU CC, you can change this file back. | |
2157 | ||
2158 | @item | |
2159 | If you want to build GNU CC with the VAX C compiler, you will need to | |
2160 | make minor changes in @file{make-cccp.com} and @file{make-cc1.com} | |
2161 | to choose alternate definitions of @code{CC}, @code{CFLAGS}, and | |
2162 | @code{LIBS}. See comments in those files. However, you must | |
2163 | also have a working version of the GNU assembler (GNU as, aka GAS) as | |
2164 | it is used as the back-end for GNU CC to produce binary object modules | |
2165 | and is not included in the GNU CC sources. GAS is also needed to | |
2166 | compile @file{libgcc2} in order to build @file{gcclib} (see above); | |
2167 | @file{make-l2.com} expects to be able to find it operational in | |
2168 | @file{gnu_cc:[000000]gnu-as.exe}. | |
2169 | ||
2170 | To use GNU CC on VMS, you need the VMS driver programs | |
2171 | @file{gcc.exe}, @file{gcc.com}, and @file{gcc.cld}. They are | |
2172 | distributed with the VMS binaries (@file{gcc-vms}) rather than the | |
2173 | GNU CC sources. GAS is also included in @file{gcc-vms}, as is Bison. | |
2174 | ||
2175 | Once you have successfully built GNU CC with VAX C, you should use the | |
2176 | resulting compiler to rebuild itself. Before doing this, be sure to | |
2177 | restore the @code{CC}, @code{CFLAGS}, and @code{LIBS} definitions in | |
2178 | @file{make-cccp.com} and @file{make-cc1.com}. The second generation | |
2179 | compiler will be able to take advantage of many optimizations that must | |
2180 | be suppressed when building with other compilers. | |
2181 | @end enumerate | |
2182 | ||
2183 | Under previous versions of GNU CC, the generated code would occasionally | |
2184 | give strange results when linked with the sharable @file{VAXCRTL} library. | |
2185 | Now this should work. | |
2186 | ||
2187 | Even with this version, however, GNU CC itself should not be linked with | |
2188 | the sharable @file{VAXCRTL}. The version of @code{qsort} in | |
2189 | @file{VAXCRTL} has a bug (known to be present in VMS versions V4.6 | |
2190 | through V5.5) which causes the compiler to fail. | |
2191 | ||
2192 | The executables are generated by @file{make-cc1.com} and | |
2193 | @file{make-cccp.com} use the object library version of @file{VAXCRTL} in | |
2194 | order to make use of the @code{qsort} routine in @file{gcclib.olb}. If | |
2195 | you wish to link the compiler executables with the shareable image | |
2196 | version of @file{VAXCRTL}, you should edit the file @file{tm.h} (created | |
2197 | by @file{vmsconfig.com}) to define the macro @code{QSORT_WORKAROUND}. | |
2198 | ||
2199 | @code{QSORT_WORKAROUND} is always defined when GNU CC is compiled with | |
2200 | VAX C, to avoid a problem in case @file{gcclib.olb} is not yet | |
2201 | available. | |
2202 | ||
2203 | @node Collect2 | |
2204 | @section @code{collect2} | |
2205 | ||
2206 | Many target systems do not have support in the assembler and linker for | |
2207 | ``constructors''---initialization functions to be called before the | |
2208 | official ``start'' of @code{main}. On such systems, GNU CC uses a | |
2209 | utility called @code{collect2} to arrange to call these functions at | |
2210 | start time. | |
2211 | ||
2212 | The program @code{collect2} works by linking the program once and | |
2213 | looking through the linker output file for symbols with particular names | |
2214 | indicating they are constructor functions. If it finds any, it | |
2215 | creates a new temporary @samp{.c} file containing a table of them, | |
2216 | compiles it, and links the program a second time including that file. | |
2217 | ||
2218 | @findex __main | |
2219 | @cindex constructors, automatic calls | |
2220 | The actual calls to the constructors are carried out by a subroutine | |
2221 | called @code{__main}, which is called (automatically) at the beginning | |
2222 | of the body of @code{main} (provided @code{main} was compiled with GNU | |
2223 | CC). Calling @code{__main} is necessary, even when compiling C code, to | |
2224 | allow linking C and C++ object code together. (If you use | |
2225 | @samp{-nostdlib}, you get an unresolved reference to @code{__main}, | |
2226 | since it's defined in the standard GCC library. Include @samp{-lgcc} at | |
2227 | the end of your compiler command line to resolve this reference.) | |
2228 | ||
2229 | The program @code{collect2} is installed as @code{ld} in the directory | |
2230 | where the passes of the compiler are installed. When @code{collect2} | |
2231 | needs to find the @emph{real} @code{ld}, it tries the following file | |
2232 | names: | |
2233 | ||
2234 | @itemize @bullet | |
2235 | @item | |
2236 | @file{real-ld} in the directories listed in the compiler's search | |
2237 | directories. | |
2238 | ||
2239 | @item | |
2240 | @file{real-ld} in the directories listed in the environment variable | |
2241 | @code{PATH}. | |
2242 | ||
2243 | @item | |
2244 | The file specified in the @code{REAL_LD_FILE_NAME} configuration macro, | |
2245 | if specified. | |
2246 | ||
2247 | @item | |
2248 | @file{ld} in the compiler's search directories, except that | |
2249 | @code{collect2} will not execute itself recursively. | |
2250 | ||
2251 | @item | |
2252 | @file{ld} in @code{PATH}. | |
2253 | @end itemize | |
2254 | ||
2255 | ``The compiler's search directories'' means all the directories where | |
2256 | @code{gcc} searches for passes of the compiler. This includes | |
2257 | directories that you specify with @samp{-B}. | |
2258 | ||
2259 | Cross-compilers search a little differently: | |
2260 | ||
2261 | @itemize @bullet | |
2262 | @item | |
2263 | @file{real-ld} in the compiler's search directories. | |
2264 | ||
2265 | @item | |
2266 | @file{@var{target}-real-ld} in @code{PATH}. | |
2267 | ||
2268 | @item | |
2269 | The file specified in the @code{REAL_LD_FILE_NAME} configuration macro, | |
2270 | if specified. | |
2271 | ||
2272 | @item | |
2273 | @file{ld} in the compiler's search directories. | |
2274 | ||
2275 | @item | |
2276 | @file{@var{target}-ld} in @code{PATH}. | |
2277 | @end itemize | |
2278 | ||
2279 | @code{collect2} explicitly avoids running @code{ld} using the file name | |
2280 | under which @code{collect2} itself was invoked. In fact, it remembers | |
2281 | up a list of such names---in case one copy of @code{collect2} finds | |
2282 | another copy (or version) of @code{collect2} installed as @code{ld} in a | |
2283 | second place in the search path. | |
2284 | ||
2285 | @code{collect2} searches for the utilities @code{nm} and @code{strip} | |
2286 | using the same algorithm as above for @code{ld}. | |
2287 | ||
2288 | @node Header Dirs | |
2289 | @section Standard Header File Directories | |
2290 | ||
2291 | @code{GCC_INCLUDE_DIR} means the same thing for native and cross. It is | |
2292 | where GNU CC stores its private include files, and also where GNU CC | |
2293 | stores the fixed include files. A cross compiled GNU CC runs | |
2294 | @code{fixincludes} on the header files in @file{$(tooldir)/include}. | |
2295 | (If the cross compilation header files need to be fixed, they must be | |
2296 | installed before GNU CC is built. If the cross compilation header files | |
2297 | are already suitable for ANSI C and GNU CC, nothing special need be | |
2298 | done). | |
2299 | ||
2300 | @code{GPLUS_INCLUDE_DIR} means the same thing for native and cross. It | |
2301 | is where @code{g++} looks first for header files. @code{libg++} | |
2302 | installs only target independent header files in that directory. | |
2303 | ||
2304 | @code{LOCAL_INCLUDE_DIR} is used only for a native compiler. It is | |
2305 | normally @file{/usr/local/include}. GNU CC searches this directory so | |
2306 | that users can install header files in @file{/usr/local/include}. | |
2307 | ||
2308 | @code{CROSS_INCLUDE_DIR} is used only for a cross compiler. GNU CC | |
2309 | doesn't install anything there. | |
2310 | ||
2311 | @code{TOOL_INCLUDE_DIR} is used for both native and cross compilers. It | |
2312 | is the place for other packages to install header files that GNU CC will | |
2313 | use. For a cross-compiler, this is the equivalent of | |
2314 | @file{/usr/include}. When you build a cross-compiler, | |
2315 | @code{fixincludes} processes any header files in this directory. |