]>
Commit | Line | Data |
---|---|---|
74291a4b MM |
1 | @c Copyright (C) 1988, 89, 92, 93, 94, 1995, 1996 Free Software Foundation, Inc. |
2 | @c This is part of the GCC manual. | |
3 | @c For copying conditions, see the file gcc.texi. | |
4 | ||
5 | @node Invoking GCC | |
6 | @chapter GNU CC Command Options | |
7 | @cindex GNU CC command options | |
8 | @cindex command options | |
9 | @cindex options, GNU CC command | |
10 | ||
11 | When you invoke GNU CC, it normally does preprocessing, compilation, | |
12 | assembly and linking. The ``overall options'' allow you to stop this | |
13 | process at an intermediate stage. For example, the @samp{-c} option | |
14 | says not to run the linker. Then the output consists of object files | |
15 | output by the assembler. | |
16 | ||
17 | Other options are passed on to one stage of processing. Some options | |
18 | control the preprocessor and others the compiler itself. Yet other | |
19 | options control the assembler and linker; most of these are not | |
20 | documented here, since you rarely need to use any of them. | |
21 | ||
22 | @cindex C compilation options | |
23 | Most of the command line options that you can use with GNU CC are useful | |
24 | for C programs; when an option is only useful with another language | |
25 | (usually C++), the explanation says so explicitly. If the description | |
26 | for a particular option does not mention a source language, you can use | |
27 | that option with all supported languages. | |
28 | ||
29 | @cindex C++ compilation options | |
30 | @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special | |
31 | options for compiling C++ programs. | |
32 | ||
33 | @cindex grouping options | |
34 | @cindex options, grouping | |
35 | The @code{gcc} program accepts options and file names as operands. Many | |
36 | options have multiletter names; therefore multiple single-letter options | |
37 | may @emph{not} be grouped: @samp{-dr} is very different from @w{@samp{-d | |
38 | -r}}. | |
39 | ||
40 | @cindex order of options | |
41 | @cindex options, order | |
42 | You can mix options and other arguments. For the most part, the order | |
43 | you use doesn't matter. Order does matter when you use several options | |
44 | of the same kind; for example, if you specify @samp{-L} more than once, | |
45 | the directories are searched in the order specified. | |
46 | ||
47 | Many options have long names starting with @samp{-f} or with | |
48 | @samp{-W}---for example, @samp{-fforce-mem}, | |
49 | @samp{-fstrength-reduce}, @samp{-Wformat} and so on. Most of | |
50 | these have both positive and negative forms; the negative form of | |
51 | @samp{-ffoo} would be @samp{-fno-foo}. This manual documents | |
52 | only one of these two forms, whichever one is not the default. | |
53 | ||
54 | @menu | |
55 | * Option Summary:: Brief list of all options, without explanations. | |
56 | * Overall Options:: Controlling the kind of output: | |
57 | an executable, object files, assembler files, | |
58 | or preprocessed source. | |
59 | * Invoking G++:: Compiling C++ programs. | |
60 | * C Dialect Options:: Controlling the variant of C language compiled. | |
61 | * C++ Dialect Options:: Variations on C++. | |
62 | * Warning Options:: How picky should the compiler be? | |
63 | * Debugging Options:: Symbol tables, measurements, and debugging dumps. | |
64 | * Optimize Options:: How much optimization? | |
65 | * Preprocessor Options:: Controlling header files and macro definitions. | |
66 | Also, getting dependency information for Make. | |
67 | * Assembler Options:: Passing options to the assembler. | |
68 | * Link Options:: Specifying libraries and so on. | |
69 | * Directory Options:: Where to find header files and libraries. | |
70 | Where to find the compiler executable files. | |
71 | * Target Options:: Running a cross-compiler, or an old version of GNU CC. | |
72 | * Submodel Options:: Specifying minor hardware or convention variations, | |
73 | such as 68010 vs 68020. | |
74 | * Code Gen Options:: Specifying conventions for function calls, data layout | |
75 | and register usage. | |
76 | * Environment Variables:: Env vars that affect GNU CC. | |
77 | * Running Protoize:: Automatically adding or removing function prototypes. | |
78 | @end menu | |
79 | ||
80 | @node Option Summary | |
81 | @section Option Summary | |
82 | ||
83 | Here is a summary of all the options, grouped by type. Explanations are | |
84 | in the following sections. | |
85 | ||
86 | @table @emph | |
87 | @item Overall Options | |
88 | @xref{Overall Options,,Options Controlling the Kind of Output}. | |
89 | @smallexample | |
90 | -c -S -E -o @var{file} -pipe -v -x @var{language} | |
91 | @end smallexample | |
92 | ||
93 | @item C Language Options | |
94 | @xref{C Dialect Options,,Options Controlling C Dialect}. | |
95 | @smallexample | |
96 | -ansi -fallow-single-precision -fcond-mismatch -fno-asm | |
97 | -fno-builtin -fsigned-bitfields -fsigned-char | |
98 | -funsigned-bitfields -funsigned-char -fwritable-strings | |
99 | -traditional -traditional-cpp -trigraphs | |
100 | @end smallexample | |
101 | ||
102 | @item C++ Language Options | |
103 | @xref{C++ Dialect Options,,Options Controlling C++ Dialect}. | |
104 | @smallexample | |
105 | -fall-virtual -fdollars-in-identifiers -felide-constructors | |
106 | -fenum-int-equiv -fexternal-templates -ffor-scope -fno-for-scope | |
107 | -fhandle-signatures -fmemoize-lookups -fno-default-inline -fno-gnu-keywords | |
108 | -fnonnull-objects -foperator-names -fstrict-prototype | |
109 | -fthis-is-variable -nostdinc++ -traditional +e@var{n} | |
110 | @end smallexample | |
111 | ||
112 | @item Warning Options | |
113 | @xref{Warning Options,,Options to Request or Suppress Warnings}. | |
114 | @smallexample | |
115 | -fsyntax-only -pedantic -pedantic-errors | |
116 | -w -W -Wall -Waggregate-return -Wbad-function-cast | |
117 | -Wcast-align -Wcast-qual -Wchar-subscript -Wcomment | |
118 | -Wconversion -Werror -Wformat | |
119 | -Wid-clash-@var{len} -Wimplicit -Wimport -Winline | |
120 | -Wlarger-than-@var{len} -Wmissing-declarations | |
121 | -Wmissing-prototypes -Wnested-externs | |
122 | -Wno-import -Woverloaded-virtual -Wparentheses | |
123 | -Wpointer-arith -Wredundant-decls -Wreorder -Wreturn-type -Wshadow | |
124 | -Wsign-compare -Wstrict-prototypes -Wswitch -Wsynth | |
125 | -Wtemplate-debugging -Wtraditional -Wtrigraphs | |
126 | -Wuninitialized -Wunused -Wwrite-strings | |
127 | @end smallexample | |
128 | ||
129 | @item Debugging Options | |
130 | @xref{Debugging Options,,Options for Debugging Your Program or GCC}. | |
131 | @smallexample | |
132 | -a -ax -d@var{letters} -fpretend-float | |
133 | -g -g@var{level} -gcoff -gdwarf -gdwarf+ | |
134 | -ggdb -gstabs -gstabs+ -gxcoff -gxcoff+ | |
135 | -p -pg -print-file-name=@var{library} -print-libgcc-file-name | |
136 | -print-prog-name=@var{program} -print-search-dirs -save-temps | |
137 | @end smallexample | |
138 | ||
139 | @item Optimization Options | |
140 | @xref{Optimize Options,,Options that Control Optimization}. | |
141 | @smallexample | |
142 | -fcaller-saves -fcse-follow-jumps -fcse-skip-blocks | |
143 | -fdelayed-branch -fexpensive-optimizations | |
144 | -ffast-math -ffloat-store -fforce-addr -fforce-mem | |
145 | -ffunction-sections -finline-functions | |
146 | -fkeep-inline-functions -fno-default-inline | |
147 | -fno-defer-pop -fno-function-cse | |
148 | -fno-inline -fno-peephole -fomit-frame-pointer | |
149 | -frerun-cse-after-loop -fschedule-insns | |
150 | -fschedule-insns2 -fstrength-reduce -fthread-jumps | |
151 | -funroll-all-loops -funroll-loops | |
152 | -O -O0 -O1 -O2 -O3 | |
153 | @end smallexample | |
154 | ||
155 | @item Preprocessor Options | |
156 | @xref{Preprocessor Options,,Options Controlling the Preprocessor}. | |
157 | @smallexample | |
158 | -A@var{question}(@var{answer}) -C -dD -dM -dN | |
159 | -D@var{macro}@r{[}=@var{defn}@r{]} -E -H | |
160 | -idirafter @var{dir} | |
161 | -include @var{file} -imacros @var{file} | |
162 | -iprefix @var{file} -iwithprefix @var{dir} | |
163 | -iwithprefixbefore @var{dir} -isystem @var{dir} | |
164 | -M -MD -MM -MMD -MG -nostdinc -P -trigraphs | |
165 | -undef -U@var{macro} -Wp,@var{option} | |
166 | @end smallexample | |
167 | ||
168 | @item Assembler Option | |
169 | @xref{Assembler Options,,Passing Options to the Assembler}. | |
170 | @smallexample | |
171 | -Wa,@var{option} | |
172 | @end smallexample | |
173 | ||
174 | @item Linker Options | |
175 | @xref{Link Options,,Options for Linking}. | |
176 | @smallexample | |
177 | @var{object-file-name} -l@var{library} | |
178 | -nostartfiles -nodefaultlibs -nostdlib | |
179 | -s -static -shared -symbolic | |
180 | -Wl,@var{option} -Xlinker @var{option} | |
181 | -u @var{symbol} | |
182 | @end smallexample | |
183 | ||
184 | @item Directory Options | |
185 | @xref{Directory Options,,Options for Directory Search}. | |
186 | @smallexample | |
187 | -B@var{prefix} -I@var{dir} -I- -L@var{dir} | |
188 | @end smallexample | |
189 | ||
190 | @item Target Options | |
191 | @c I wrote this xref this way to avoid overfull hbox. -- rms | |
192 | @xref{Target Options}. | |
193 | @smallexample | |
194 | -b @var{machine} -V @var{version} | |
195 | @end smallexample | |
196 | ||
197 | @item Machine Dependent Options | |
198 | @xref{Submodel Options,,Hardware Models and Configurations}. | |
199 | @smallexample | |
200 | @emph{M680x0 Options} | |
201 | -m68000 -m68020 -m68020-40 -m68030 -m68040 -m68060 | |
202 | -m5200 -m68881 -mbitfield -mc68000 -mc68020 -mfpa | |
203 | -mnobitfield -mrtd -mshort -msoft-float | |
204 | ||
205 | @emph{VAX Options} | |
206 | -mg -mgnu -munix | |
207 | ||
208 | @emph{SPARC Options} | |
209 | -mcpu=@var{cpu type} | |
210 | -mtune=@var{cpu type} | |
211 | -mapp-regs -mbroken-saverestore -mcypress -mepilogue | |
212 | -mflat -mfpu -mfullany -mhard-float -mhard-quad-float | |
213 | -mimpure-text -mint32 -mint64 -mlive-g0 | |
214 | -mlong32 -mlong64 -mmedlow -mmedany | |
215 | -mno-app-regs -mno-epilogue | |
216 | -mno-flat -mno-fpu -mno-impure-text | |
217 | -mno-stack-bias -mno-unaligned-doubles | |
218 | -msoft-float -msoft-quad-float -msparclite -mstack-bias | |
219 | -msupersparc -munaligned-doubles -mv8 | |
220 | ||
221 | @emph{Convex Options} | |
222 | -mc1 -mc2 -mc32 -mc34 -mc38 | |
223 | -margcount -mnoargcount | |
224 | -mlong32 -mlong64 | |
225 | -mvolatile-cache -mvolatile-nocache | |
226 | ||
227 | @emph{AMD29K Options} | |
228 | -m29000 -m29050 -mbw -mnbw -mdw -mndw | |
229 | -mlarge -mnormal -msmall | |
230 | -mkernel-registers -mno-reuse-arg-regs | |
231 | -mno-stack-check -mno-storem-bug | |
232 | -mreuse-arg-regs -msoft-float -mstack-check | |
233 | -mstorem-bug -muser-registers | |
234 | ||
235 | @emph{ARM Options} | |
236 | -mapcs-frame -mapcs-26 -mapcs-32 | |
237 | -mlittle-endian -mbig-endian -mwords-little-endian | |
238 | -mshort-load-bytes -mno-short-load-bytes | |
239 | -msoft-float -mhard-float | |
240 | -mbsd -mxopen -mno-symrename | |
241 | ||
242 | @emph{M88K Options} | |
243 | -m88000 -m88100 -m88110 -mbig-pic | |
244 | -mcheck-zero-division -mhandle-large-shift | |
245 | -midentify-revision -mno-check-zero-division | |
246 | -mno-ocs-debug-info -mno-ocs-frame-position | |
247 | -mno-optimize-arg-area -mno-serialize-volatile | |
248 | -mno-underscores -mocs-debug-info | |
249 | -mocs-frame-position -moptimize-arg-area | |
250 | -mserialize-volatile -mshort-data-@var{num} -msvr3 | |
251 | -msvr4 -mtrap-large-shift -muse-div-instruction | |
252 | -mversion-03.00 -mwarn-passed-structs | |
253 | ||
254 | @emph{RS/6000 and PowerPC Options} | |
255 | -mcpu=@var{cpu type} | |
256 | -mtune=@var{cpu type} | |
257 | -mpower -mno-power -mpower2 -mno-power2 | |
258 | -mpowerpc -mno-powerpc | |
259 | -mpowerpc-gpopt -mno-powerpc-gpopt | |
260 | -mpowerpc-gfxopt -mno-powerpc-gfxopt | |
261 | -mnew-mnemonics -mno-new-mnemonics | |
262 | -mfull-toc -mminimal-toc -mno-fop-in-toc -mno-sum-in-toc | |
263 | -mxl-call -mno-xl-call | |
264 | -msoft-float -mhard-float -mmultiple -mno-multiple | |
265 | -mstring -mno-string -mbit-align -mno-bit-align | |
266 | -mstrict-align -mno-strict-align -mrelocatable | |
267 | -mno-relocatable -mrelocatable-lib -mno-relocatable-lib | |
268 | -mtoc -mno-toc -mtraceback -mno-traceback | |
269 | -mlittle -mlittle-endian -mbig -mbig-endian | |
270 | -mcall-aix -mcall-sysv -mprototype -mno-prototype | |
271 | -msim -mmvme -memb -msdata -msdata=@var{opt} -G @var{num} | |
272 | ||
273 | @emph{RT Options} | |
274 | -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs | |
275 | -mfull-fp-blocks -mhc-struct-return -min-line-mul | |
276 | -mminimum-fp-blocks -mnohc-struct-return | |
277 | ||
278 | @emph{MIPS Options} | |
279 | -mabicalls -mcpu=@var{cpu type} -membedded-data | |
280 | -membedded-pic -mfp32 -mfp64 -mgas -mgp32 -mgp64 | |
281 | -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 | |
282 | -mips2 -mips3 -mlong64 -mlong-calls -mmemcpy | |
283 | -mmips-as -mmips-tfile -mno-abicalls | |
284 | -mno-embedded-data -mno-embedded-pic | |
285 | -mno-gpopt -mno-long-calls | |
286 | -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats | |
287 | -mrnames -msoft-float | |
288 | -m4650 -msingle-float -mmad | |
289 | -mstats -EL -EB -G @var{num} -nocpp | |
290 | ||
291 | @emph{i386 Options} | |
292 | -m486 -m386 -mieee-fp -mno-fancy-math-387 | |
293 | -mno-fp-ret-in-387 -msoft-float -msvr3-shlib | |
294 | -mno-wide-multiply -mrtd -malign-double | |
295 | -mreg-alloc=@var{list} -mregparm=@var{num} | |
296 | -malign-jumps=@var{num} -malign-loops=@var{num} | |
297 | -malign-functions=@var{num} | |
298 | ||
299 | @emph{HPPA Options} | |
300 | -mdisable-fpregs -mdisable-indexing -mfast-indirect-calls | |
301 | -mgas -mjump-in-delay -mlong-load-store -mno-disable-fpregs | |
302 | -mno-disable-indexing -mno-fast-indirect-calls -mno-gas | |
303 | -mno-jump-in-delay | |
304 | -mno-long-load-store | |
305 | -mno-portable-runtime -mno-soft-float -mno-space -mno-space-regs -msoft-float | |
306 | -mpa-risc-1-0 -mpa-risc-1-1 -mportable-runtime | |
307 | -mschedule=@var{list} -mspace -mspace-regs | |
308 | ||
309 | @emph{Intel 960 Options} | |
310 | -m@var{cpu type} -masm-compat -mclean-linkage | |
311 | -mcode-align -mcomplex-addr -mleaf-procedures | |
312 | -mic-compat -mic2.0-compat -mic3.0-compat | |
313 | -mintel-asm -mno-clean-linkage -mno-code-align | |
314 | -mno-complex-addr -mno-leaf-procedures | |
315 | -mno-old-align -mno-strict-align -mno-tail-call | |
316 | -mnumerics -mold-align -msoft-float -mstrict-align | |
317 | -mtail-call | |
318 | ||
319 | @emph{DEC Alpha Options} | |
320 | -mfp-regs -mno-fp-regs | |
321 | -mno-soft-float -msoft-float | |
322 | -mieee -mieee-with-inexact -mieee-conformant | |
323 | -mfp-trap-mode -mfp-rounding-mode -mtrap-precision | |
324 | -mbuild-constants | |
325 | ||
326 | @emph{Clipper Options} | |
327 | -mc300 -mc400 | |
328 | ||
329 | @emph{H8/300 Options} | |
330 | -mrelax -mh -mint32 -malign-300 | |
331 | ||
332 | @emph{SH Options} | |
333 | -m1 -m2 -m3 -m3e -mb -ml -mrelax | |
334 | ||
335 | @emph{System V Options} | |
336 | -Qy -Qn -YP,@var{paths} -Ym,@var{dir} | |
337 | @end smallexample | |
338 | ||
339 | @item Code Generation Options | |
340 | @xref{Code Gen Options,,Options for Code Generation Conventions}. | |
341 | @smallexample | |
342 | -fcall-saved-@var{reg} -fcall-used-@var{reg} | |
343 | -ffixed-@var{reg} -finhibit-size-directive | |
344 | -fno-common -fno-ident -fno-gnu-linker | |
345 | -fpcc-struct-return -fpic -fPIC | |
346 | -freg-struct-return -fshared-data -fshort-enums | |
347 | -fshort-double -fvolatile -fvolatile-global | |
348 | -fverbose-asm -fpack-struct +e0 +e1 | |
349 | @end smallexample | |
350 | @end table | |
351 | ||
352 | @menu | |
353 | * Overall Options:: Controlling the kind of output: | |
354 | an executable, object files, assembler files, | |
355 | or preprocessed source. | |
356 | * C Dialect Options:: Controlling the variant of C language compiled. | |
357 | * C++ Dialect Options:: Variations on C++. | |
358 | * Warning Options:: How picky should the compiler be? | |
359 | * Debugging Options:: Symbol tables, measurements, and debugging dumps. | |
360 | * Optimize Options:: How much optimization? | |
361 | * Preprocessor Options:: Controlling header files and macro definitions. | |
362 | Also, getting dependency information for Make. | |
363 | * Assembler Options:: Passing options to the assembler. | |
364 | * Link Options:: Specifying libraries and so on. | |
365 | * Directory Options:: Where to find header files and libraries. | |
366 | Where to find the compiler executable files. | |
367 | * Target Options:: Running a cross-compiler, or an old version of GNU CC. | |
368 | @end menu | |
369 | ||
370 | @node Overall Options | |
371 | @section Options Controlling the Kind of Output | |
372 | ||
373 | Compilation can involve up to four stages: preprocessing, compilation | |
374 | proper, assembly and linking, always in that order. The first three | |
375 | stages apply to an individual source file, and end by producing an | |
376 | object file; linking combines all the object files (those newly | |
377 | compiled, and those specified as input) into an executable file. | |
378 | ||
379 | @cindex file name suffix | |
380 | For any given input file, the file name suffix determines what kind of | |
381 | compilation is done: | |
382 | ||
383 | @table @code | |
384 | @item @var{file}.c | |
385 | C source code which must be preprocessed. | |
386 | ||
387 | @item @var{file}.i | |
388 | C source code which should not be preprocessed. | |
389 | ||
390 | @item @var{file}.ii | |
391 | C++ source code which should not be preprocessed. | |
392 | ||
393 | @item @var{file}.m | |
394 | Objective-C source code. Note that you must link with the library | |
395 | @file{libobjc.a} to make an Objective-C program work. | |
396 | ||
397 | @item @var{file}.h | |
398 | C header file (not to be compiled or linked). | |
399 | ||
400 | @item @var{file}.cc | |
401 | @itemx @var{file}.cxx | |
402 | @itemx @var{file}.cpp | |
403 | @itemx @var{file}.C | |
404 | C++ source code which must be preprocessed. Note that in @samp{.cxx}, | |
405 | the last two letters must both be literally @samp{x}. Likewise, | |
406 | @samp{.C} refers to a literal capital C. | |
407 | ||
408 | @item @var{file}.s | |
409 | Assembler code. | |
410 | ||
411 | @item @var{file}.S | |
412 | Assembler code which must be preprocessed. | |
413 | ||
414 | @item @var{other} | |
415 | An object file to be fed straight into linking. | |
416 | Any file name with no recognized suffix is treated this way. | |
417 | @end table | |
418 | ||
419 | You can specify the input language explicitly with the @samp{-x} option: | |
420 | ||
421 | @table @code | |
422 | @item -x @var{language} | |
423 | Specify explicitly the @var{language} for the following input files | |
424 | (rather than letting the compiler choose a default based on the file | |
425 | name suffix). This option applies to all following input files until | |
426 | the next @samp{-x} option. Possible values for @var{language} are: | |
427 | @example | |
428 | c objective-c c++ | |
429 | c-header cpp-output c++-cpp-output | |
430 | assembler assembler-with-cpp | |
431 | @end example | |
432 | ||
433 | @item -x none | |
434 | Turn off any specification of a language, so that subsequent files are | |
435 | handled according to their file name suffixes (as they are if @samp{-x} | |
436 | has not been used at all). | |
437 | @end table | |
438 | ||
439 | If you only want some of the stages of compilation, you can use | |
440 | @samp{-x} (or filename suffixes) to tell @code{gcc} where to start, and | |
441 | one of the options @samp{-c}, @samp{-S}, or @samp{-E} to say where | |
442 | @code{gcc} is to stop. Note that some combinations (for example, | |
443 | @samp{-x cpp-output -E} instruct @code{gcc} to do nothing at all. | |
444 | ||
445 | @table @code | |
446 | @item -c | |
447 | Compile or assemble the source files, but do not link. The linking | |
448 | stage simply is not done. The ultimate output is in the form of an | |
449 | object file for each source file. | |
450 | ||
451 | By default, the object file name for a source file is made by replacing | |
452 | the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}. | |
453 | ||
454 | Unrecognized input files, not requiring compilation or assembly, are | |
455 | ignored. | |
456 | ||
457 | @item -S | |
458 | Stop after the stage of compilation proper; do not assemble. The output | |
459 | is in the form of an assembler code file for each non-assembler input | |
460 | file specified. | |
461 | ||
462 | By default, the assembler file name for a source file is made by | |
463 | replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}. | |
464 | ||
465 | Input files that don't require compilation are ignored. | |
466 | ||
467 | @item -E | |
468 | Stop after the preprocessing stage; do not run the compiler proper. The | |
469 | output is in the form of preprocessed source code, which is sent to the | |
470 | standard output. | |
471 | ||
472 | Input files which don't require preprocessing are ignored. | |
473 | ||
474 | @cindex output file option | |
475 | @item -o @var{file} | |
476 | Place output in file @var{file}. This applies regardless to whatever | |
477 | sort of output is being produced, whether it be an executable file, | |
478 | an object file, an assembler file or preprocessed C code. | |
479 | ||
480 | Since only one output file can be specified, it does not make sense to | |
481 | use @samp{-o} when compiling more than one input file, unless you are | |
482 | producing an executable file as output. | |
483 | ||
484 | If @samp{-o} is not specified, the default is to put an executable file | |
485 | in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in | |
486 | @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and | |
487 | all preprocessed C source on standard output.@refill | |
488 | ||
489 | @item -v | |
490 | Print (on standard error output) the commands executed to run the stages | |
491 | of compilation. Also print the version number of the compiler driver | |
492 | program and of the preprocessor and the compiler proper. | |
493 | ||
494 | @item -pipe | |
495 | Use pipes rather than temporary files for communication between the | |
496 | various stages of compilation. This fails to work on some systems where | |
497 | the assembler is unable to read from a pipe; but the GNU assembler has | |
498 | no trouble. | |
499 | @end table | |
500 | ||
501 | @node Invoking G++ | |
502 | @section Compiling C++ Programs | |
503 | ||
504 | @cindex suffixes for C++ source | |
505 | @cindex C++ source file suffixes | |
506 | C++ source files conventionally use one of the suffixes @samp{.C}, | |
507 | @samp{.cc}, @samp{cpp}, or @samp{.cxx}; preprocessed C++ files use the | |
508 | suffix @samp{.ii}. GNU CC recognizes files with these names and | |
509 | compiles them as C++ programs even if you call the compiler the same way | |
510 | as for compiling C programs (usually with the name @code{gcc}). | |
511 | ||
512 | @findex g++ | |
513 | @findex c++ | |
514 | However, C++ programs often require class libraries as well as a | |
515 | compiler that understands the C++ language---and under some | |
516 | circumstances, you might want to compile programs from standard input, | |
517 | or otherwise without a suffix that flags them as C++ programs. | |
518 | @code{g++} is a program that calls GNU CC with the default language | |
519 | set to C++, and automatically specifies linking against the GNU class | |
520 | library libg++. | |
521 | @cindex @code{g++ 1.@var{xx}} | |
522 | @cindex @code{g++}, separate compiler | |
523 | @cindex @code{g++} older version | |
524 | @footnote{Prior to release 2 of the compiler, | |
525 | there was a separate @code{g++} compiler. That version was based on GNU | |
526 | CC, but not integrated with it. Versions of @code{g++} with a | |
527 | @samp{1.@var{xx}} version number---for example, @code{g++} version 1.37 | |
528 | or 1.42---are much less reliable than the versions integrated with GCC | |
529 | 2. Moreover, combining G++ @samp{1.@var{xx}} with a version 2 GCC will | |
530 | simply not work.} On many systems, the script @code{g++} is also | |
531 | installed with the name @code{c++}. | |
532 | ||
533 | @cindex invoking @code{g++} | |
534 | When you compile C++ programs, you may specify many of the same | |
535 | command-line options that you use for compiling programs in any | |
536 | language; or command-line options meaningful for C and related | |
537 | languages; or options that are meaningful only for C++ programs. | |
538 | @xref{C Dialect Options,,Options Controlling C Dialect}, for | |
539 | explanations of options for languages related to C. | |
540 | @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for | |
541 | explanations of options that are meaningful only for C++ programs. | |
542 | ||
543 | @node C Dialect Options | |
544 | @section Options Controlling C Dialect | |
545 | @cindex dialect options | |
546 | @cindex language dialect options | |
547 | @cindex options, dialect | |
548 | ||
549 | The following options control the dialect of C (or languages derived | |
550 | from C, such as C++ and Objective C) that the compiler accepts: | |
551 | ||
552 | @table @code | |
553 | @cindex ANSI support | |
554 | @item -ansi | |
555 | Support all ANSI standard C programs. | |
556 | ||
557 | This turns off certain features of GNU C that are incompatible with ANSI | |
558 | C, such as the @code{asm}, @code{inline} and @code{typeof} keywords, and | |
559 | predefined macros such as @code{unix} and @code{vax} that identify the | |
560 | type of system you are using. It also enables the undesirable and | |
561 | rarely used ANSI trigraph feature, disallows @samp{$} as part of | |
562 | identifiers, and disables recognition of C++ style @samp{//} comments. | |
563 | ||
564 | The alternate keywords @code{__asm__}, @code{__extension__}, | |
565 | @code{__inline__} and @code{__typeof__} continue to work despite | |
566 | @samp{-ansi}. You would not want to use them in an ANSI C program, of | |
567 | course, but it is useful to put them in header files that might be included | |
568 | in compilations done with @samp{-ansi}. Alternate predefined macros | |
569 | such as @code{__unix__} and @code{__vax__} are also available, with or | |
570 | without @samp{-ansi}. | |
571 | ||
572 | The @samp{-ansi} option does not cause non-ANSI programs to be | |
573 | rejected gratuitously. For that, @samp{-pedantic} is required in | |
574 | addition to @samp{-ansi}. @xref{Warning Options}. | |
575 | ||
576 | The macro @code{__STRICT_ANSI__} is predefined when the @samp{-ansi} | |
577 | option is used. Some header files may notice this macro and refrain | |
578 | from declaring certain functions or defining certain macros that the | |
579 | ANSI standard doesn't call for; this is to avoid interfering with any | |
580 | programs that might use these names for other things. | |
581 | ||
582 | The functions @code{alloca}, @code{abort}, @code{exit}, and | |
583 | @code{_exit} are not builtin functions when @samp{-ansi} is used. | |
584 | ||
585 | @item -fno-asm | |
586 | Do not recognize @code{asm}, @code{inline} or @code{typeof} as a | |
587 | keyword, so that code can use these words as identifiers. You can use | |
588 | the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__} | |
589 | instead. @samp{-ansi} implies @samp{-fno-asm}. | |
590 | ||
591 | In C++, this switch only affects the @code{typeof} keyword, since | |
592 | @code{asm} and @code{inline} are standard keywords. You may want to | |
593 | use the @samp{-fno-gnu-keywords} flag instead, as it also disables the | |
594 | other, C++-specific, extension keywords such as @code{headof}. | |
595 | ||
596 | @item -fno-builtin | |
597 | @cindex builtin functions | |
598 | @findex abort | |
599 | @findex abs | |
600 | @findex alloca | |
601 | @findex cos | |
602 | @findex exit | |
603 | @findex fabs | |
604 | @findex ffs | |
605 | @findex labs | |
606 | @findex memcmp | |
607 | @findex memcpy | |
608 | @findex sin | |
609 | @findex sqrt | |
610 | @findex strcmp | |
611 | @findex strcpy | |
612 | @findex strlen | |
613 | Don't recognize builtin functions that do not begin with two leading | |
614 | underscores. Currently, the functions affected include @code{abort}, | |
615 | @code{abs}, @code{alloca}, @code{cos}, @code{exit}, @code{fabs}, | |
616 | @code{ffs}, @code{labs}, @code{memcmp}, @code{memcpy}, @code{sin}, | |
617 | @code{sqrt}, @code{strcmp}, @code{strcpy}, and @code{strlen}. | |
618 | ||
619 | GCC normally generates special code to handle certain builtin functions | |
620 | more efficiently; for instance, calls to @code{alloca} may become single | |
621 | instructions that adjust the stack directly, and calls to @code{memcpy} | |
622 | may become inline copy loops. The resulting code is often both smaller | |
623 | and faster, but since the function calls no longer appear as such, you | |
624 | cannot set a breakpoint on those calls, nor can you change the behavior | |
625 | of the functions by linking with a different library. | |
626 | ||
627 | The @samp{-ansi} option prevents @code{alloca} and @code{ffs} from being | |
628 | builtin functions, since these functions do not have an ANSI standard | |
629 | meaning. | |
630 | ||
631 | @item -trigraphs | |
632 | Support ANSI C trigraphs. You don't want to know about this | |
633 | brain-damage. The @samp{-ansi} option implies @samp{-trigraphs}. | |
634 | ||
635 | @cindex traditional C language | |
636 | @cindex C language, traditional | |
637 | @item -traditional | |
638 | Attempt to support some aspects of traditional C compilers. | |
639 | Specifically: | |
640 | ||
641 | @itemize @bullet | |
642 | @item | |
643 | All @code{extern} declarations take effect globally even if they | |
644 | are written inside of a function definition. This includes implicit | |
645 | declarations of functions. | |
646 | ||
647 | @item | |
648 | The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const} | |
649 | and @code{volatile} are not recognized. (You can still use the | |
650 | alternative keywords such as @code{__typeof__}, @code{__inline__}, and | |
651 | so on.) | |
652 | ||
653 | @item | |
654 | Comparisons between pointers and integers are always allowed. | |
655 | ||
656 | @item | |
657 | Integer types @code{unsigned short} and @code{unsigned char} promote | |
658 | to @code{unsigned int}. | |
659 | ||
660 | @item | |
661 | Out-of-range floating point literals are not an error. | |
662 | ||
663 | @item | |
664 | Certain constructs which ANSI regards as a single invalid preprocessing | |
665 | number, such as @samp{0xe-0xd}, are treated as expressions instead. | |
666 | ||
667 | @item | |
668 | String ``constants'' are not necessarily constant; they are stored in | |
669 | writable space, and identical looking constants are allocated | |
670 | separately. (This is the same as the effect of | |
671 | @samp{-fwritable-strings}.) | |
672 | ||
673 | @cindex @code{longjmp} and automatic variables | |
674 | @item | |
675 | All automatic variables not declared @code{register} are preserved by | |
676 | @code{longjmp}. Ordinarily, GNU C follows ANSI C: automatic variables | |
677 | not declared @code{volatile} may be clobbered. | |
678 | ||
679 | @item | |
680 | @kindex \x | |
681 | @kindex \a | |
682 | @cindex escape sequences, traditional | |
683 | The character escape sequences @samp{\x} and @samp{\a} evaluate as the | |
684 | literal characters @samp{x} and @samp{a} respectively. Without | |
685 | @w{@samp{-traditional}}, @samp{\x} is a prefix for the hexadecimal | |
686 | representation of a character, and @samp{\a} produces a bell. | |
687 | ||
688 | @item | |
689 | In C++ programs, assignment to @code{this} is permitted with | |
690 | @samp{-traditional}. (The option @samp{-fthis-is-variable} also has | |
691 | this effect.) | |
692 | @end itemize | |
693 | ||
694 | You may wish to use @samp{-fno-builtin} as well as @samp{-traditional} | |
695 | if your program uses names that are normally GNU C builtin functions for | |
696 | other purposes of its own. | |
697 | ||
698 | You cannot use @samp{-traditional} if you include any header files that | |
699 | rely on ANSI C features. Some vendors are starting to ship systems with | |
700 | ANSI C header files and you cannot use @samp{-traditional} on such | |
701 | systems to compile files that include any system headers. | |
702 | ||
703 | The @samp{-traditional} option also enables the @samp{-traditional-cpp} | |
704 | option, which is described next. | |
705 | ||
706 | @item -traditional-cpp | |
707 | Attempt to support some aspects of traditional C preprocessors. | |
708 | Specifically: | |
709 | ||
710 | @itemize @bullet | |
711 | @item | |
712 | Comments convert to nothing at all, rather than to a space. This allows | |
713 | traditional token concatenation. | |
714 | ||
715 | @item | |
716 | In a preprocessing directive, the @samp{#} symbol must appear as the first | |
717 | character of a line. | |
718 | ||
719 | @item | |
720 | Macro arguments are recognized within string constants in a macro | |
721 | definition (and their values are stringified, though without additional | |
722 | quote marks, when they appear in such a context). The preprocessor | |
723 | always considers a string constant to end at a newline. | |
724 | ||
725 | @item | |
726 | @cindex detecting @w{@samp{-traditional}} | |
727 | The predefined macro @code{__STDC__} is not defined when you use | |
728 | @samp{-traditional}, but @code{__GNUC__} is (since the GNU extensions | |
729 | which @code{__GNUC__} indicates are not affected by | |
730 | @samp{-traditional}). If you need to write header files that work | |
731 | differently depending on whether @samp{-traditional} is in use, by | |
732 | testing both of these predefined macros you can distinguish four | |
733 | situations: GNU C, traditional GNU C, other ANSI C compilers, and other | |
734 | old C compilers. The predefined macro @code{__STDC_VERSION__} is also | |
735 | not defined when you use @samp{-traditional}. @xref{Standard | |
736 | Predefined,,Standard Predefined Macros,cpp.info,The C Preprocessor}, | |
737 | for more discussion of these and other predefined macros. | |
738 | ||
739 | @item | |
740 | @cindex string constants vs newline | |
741 | @cindex newline vs string constants | |
742 | The preprocessor considers a string constant to end at a newline (unless | |
743 | the newline is escaped with @samp{\}). (Without @w{@samp{-traditional}}, | |
744 | string constants can contain the newline character as typed.) | |
745 | @end itemize | |
746 | ||
747 | @item -fcond-mismatch | |
748 | Allow conditional expressions with mismatched types in the second and | |
749 | third arguments. The value of such an expression is void. | |
750 | ||
751 | @item -funsigned-char | |
752 | Let the type @code{char} be unsigned, like @code{unsigned char}. | |
753 | ||
754 | Each kind of machine has a default for what @code{char} should | |
755 | be. It is either like @code{unsigned char} by default or like | |
756 | @code{signed char} by default. | |
757 | ||
758 | Ideally, a portable program should always use @code{signed char} or | |
759 | @code{unsigned char} when it depends on the signedness of an object. | |
760 | But many programs have been written to use plain @code{char} and | |
761 | expect it to be signed, or expect it to be unsigned, depending on the | |
762 | machines they were written for. This option, and its inverse, let you | |
763 | make such a program work with the opposite default. | |
764 | ||
765 | The type @code{char} is always a distinct type from each of | |
766 | @code{signed char} or @code{unsigned char}, even though its behavior | |
767 | is always just like one of those two. | |
768 | ||
769 | @item -fsigned-char | |
770 | Let the type @code{char} be signed, like @code{signed char}. | |
771 | ||
772 | Note that this is equivalent to @samp{-fno-unsigned-char}, which is | |
773 | the negative form of @samp{-funsigned-char}. Likewise, the option | |
774 | @samp{-fno-signed-char} is equivalent to @samp{-funsigned-char}. | |
775 | ||
776 | You may wish to use @samp{-fno-builtin} as well as @samp{-traditional} | |
777 | if your program uses names that are normally GNU C builtin functions for | |
778 | other purposes of its own. | |
779 | ||
780 | You cannot use @samp{-traditional} if you include any header files that | |
781 | rely on ANSI C features. Some vendors are starting to ship systems with | |
782 | ANSI C header files and you cannot use @samp{-traditional} on such | |
783 | systems to compile files that include any system headers. | |
784 | ||
785 | @item -fsigned-bitfields | |
786 | @itemx -funsigned-bitfields | |
787 | @itemx -fno-signed-bitfields | |
788 | @itemx -fno-unsigned-bitfields | |
789 | These options control whether a bitfield is signed or unsigned, when the | |
790 | declaration does not use either @code{signed} or @code{unsigned}. By | |
791 | default, such a bitfield is signed, because this is consistent: the | |
792 | basic integer types such as @code{int} are signed types. | |
793 | ||
794 | However, when @samp{-traditional} is used, bitfields are all unsigned | |
795 | no matter what. | |
796 | ||
797 | @item -fwritable-strings | |
798 | Store string constants in the writable data segment and don't uniquize | |
799 | them. This is for compatibility with old programs which assume they can | |
800 | write into string constants. The option @samp{-traditional} also has | |
801 | this effect. | |
802 | ||
803 | Writing into string constants is a very bad idea; ``constants'' should | |
804 | be constant. | |
805 | ||
806 | @item -fallow-single-precision | |
807 | Do not promote single precision math operations to double precision, | |
808 | even when compiling with @samp{-traditional}. | |
809 | ||
810 | Traditional K&R C promotes all floating point operations to double | |
811 | precision, regardless of the sizes of the operands. On the | |
812 | architecture for which you are compiling, single precision may be faster | |
813 | than double precision. If you must use @samp{-traditional}, but want | |
814 | to use single precision operations when the operands are single | |
815 | precision, use this option. This option has no effect when compiling | |
816 | with ANSI or GNU C conventions (the default). | |
817 | ||
818 | @end table | |
819 | ||
820 | @node C++ Dialect Options | |
821 | @section Options Controlling C++ Dialect | |
822 | ||
823 | @cindex compiler options, C++ | |
824 | @cindex C++ options, command line | |
825 | @cindex options, C++ | |
826 | This section describes the command-line options that are only meaningful | |
827 | for C++ programs; but you can also use most of the GNU compiler options | |
828 | regardless of what language your program is in. For example, you | |
829 | might compile a file @code{firstClass.C} like this: | |
830 | ||
831 | @example | |
832 | g++ -g -felide-constructors -O -c firstClass.C | |
833 | @end example | |
834 | ||
835 | @noindent | |
836 | In this example, only @samp{-felide-constructors} is an option meant | |
837 | only for C++ programs; you can use the other options with any | |
838 | language supported by GNU CC. | |
839 | ||
840 | Here is a list of options that are @emph{only} for compiling C++ programs: | |
841 | ||
842 | @table @code | |
843 | @item -fno-access-control | |
844 | Turn off all access checking. This switch is mainly useful for working | |
845 | around bugs in the access control code. | |
846 | ||
847 | @item -fall-virtual | |
848 | Treat all possible member functions as virtual, implicitly. | |
849 | All member functions (except for constructor functions and @code{new} or | |
850 | @code{delete} member operators) are treated as virtual functions of the | |
851 | class where they appear. | |
852 | ||
853 | This does not mean that all calls to these member functions will be made | |
854 | through the internal table of virtual functions. Under some | |
855 | circumstances, the compiler can determine that a call to a given virtual | |
856 | function can be made directly; in these cases the calls are direct in | |
857 | any case. | |
858 | ||
859 | @item -fcheck-new | |
860 | Check that the pointer returned by @code{operator new} is non-null | |
861 | before attempting to modify the storage allocated. The current Working | |
862 | Paper requires that @code{operator new} never return a null pointer, so | |
863 | this check is normally unnecessary. | |
864 | ||
865 | @item -fconserve-space | |
866 | Put uninitialized or runtime-initialized global variables into the | |
867 | common segment, as C does. This saves space in the executable at the | |
868 | cost of not diagnosing duplicate definitions. If you compile with this | |
869 | flag and your program mysteriously crashes after @code{main()} has | |
870 | completed, you may have an object that is being destroyed twice because | |
871 | two definitions were merged. | |
872 | ||
873 | @item -fdollars-in-identifiers | |
874 | Accept @samp{$} in identifiers. You can also explicitly prohibit use of | |
875 | @samp{$} with the option @samp{-fno-dollars-in-identifiers}. (GNU C++ | |
876 | allows @samp{$} by default on some target systems but not others.) | |
877 | Traditional C allowed the character @samp{$} to form part of | |
878 | identifiers. However, ANSI C and C++ forbid @samp{$} in identifiers. | |
879 | ||
880 | @item -fenum-int-equiv | |
881 | Anachronistically permit implicit conversion of @code{int} to | |
882 | enumeration types. Current C++ allows conversion of @code{enum} to | |
883 | @code{int}, but not the other way around. | |
884 | ||
885 | @item -fexternal-templates | |
886 | Cause template instantiations to obey @samp{#pragma interface} and | |
887 | @samp{implementation}; template instances are emitted or not according | |
888 | to the location of the template definition. @xref{Template | |
889 | Instantiation}, for more information. | |
890 | ||
891 | @item -falt-external-templates | |
892 | Similar to -fexternal-templates, but template instances are emitted or | |
893 | not according to the place where they are first instantiated. | |
894 | @xref{Template Instantiation}, for more information. | |
895 | ||
896 | @item -ffor-scope | |
897 | @item -fno-for-scope | |
898 | If -ffor-scope is specified, the scope of variables declared in | |
899 | a @i{for-init-statement} is limited to the @samp{for} loop itself, | |
900 | as specified by the draft C++ standard. | |
901 | If -fno-for-scope is specified, the scope of variables declared in | |
902 | a @i{for-init-statement} extends to the end of the enclosing scope, | |
903 | as was the case in old versions of gcc, and other (traditional) | |
904 | implementations of C++. | |
905 | ||
906 | The default if neither flag is given to follow the standard, | |
907 | but to allow and give a warning for old-style code that would | |
908 | otherwise be invalid, or have different behavior. | |
909 | ||
910 | @item -fno-gnu-keywords | |
911 | Do not recognize @code{classof}, @code{headof}, @code{signature}, | |
912 | @code{sigof} or @code{typeof} as a keyword, so that code can use these | |
913 | words as identifiers. You can use the keywords @code{__classof__}, | |
914 | @code{__headof__}, @code{__signature__}, @code{__sigof__}, and | |
915 | @code{__typeof__} instead. @samp{-ansi} implies | |
916 | @samp{-fno-gnu-keywords}. | |
917 | ||
918 | @item -fno-implicit-templates | |
919 | Never emit code for templates which are instantiated implicitly (i.e. by | |
920 | use); only emit code for explicit instantiations. @xref{Template | |
921 | Instantiation}, for more information. | |
922 | ||
923 | @item -fhandle-signatures | |
924 | Recognize the @code{signature} and @code{sigof} keywords for specifying | |
925 | abstract types. The default (@samp{-fno-handle-signatures}) is not to | |
926 | recognize them. @xref{C++ Signatures, Type Abstraction using | |
927 | Signatures}. | |
928 | ||
929 | @item -fhuge-objects | |
930 | Support virtual function calls for objects that exceed the size | |
931 | representable by a @samp{short int}. Users should not use this flag by | |
932 | default; if you need to use it, the compiler will tell you so. If you | |
933 | compile any of your code with this flag, you must compile @emph{all} of | |
934 | your code with this flag (including libg++, if you use it). | |
935 | ||
936 | This flag is not useful when compiling with -fvtable-thunks. | |
937 | ||
938 | @item -fno-implement-inlines | |
939 | To save space, do not emit out-of-line copies of inline functions | |
940 | controlled by @samp{#pragma implementation}. This will cause linker | |
941 | errors if these functions are not inlined everywhere they are called. | |
942 | ||
943 | @item -fmemoize-lookups | |
944 | @itemx -fsave-memoized | |
945 | Use heuristics to compile faster. These heuristics are not enabled by | |
946 | default, since they are only effective for certain input files. Other | |
947 | input files compile more slowly. | |
948 | ||
949 | The first time the compiler must build a call to a member function (or | |
950 | reference to a data member), it must (1) determine whether the class | |
951 | implements member functions of that name; (2) resolve which member | |
952 | function to call (which involves figuring out what sorts of type | |
953 | conversions need to be made); and (3) check the visibility of the member | |
954 | function to the caller. All of this adds up to slower compilation. | |
955 | Normally, the second time a call is made to that member function (or | |
956 | reference to that data member), it must go through the same lengthy | |
957 | process again. This means that code like this: | |
958 | ||
959 | @smallexample | |
960 | cout << "This " << p << " has " << n << " legs.\n"; | |
961 | @end smallexample | |
962 | ||
963 | @noindent | |
964 | makes six passes through all three steps. By using a software cache, a | |
965 | ``hit'' significantly reduces this cost. Unfortunately, using the cache | |
966 | introduces another layer of mechanisms which must be implemented, and so | |
967 | incurs its own overhead. @samp{-fmemoize-lookups} enables the software | |
968 | cache. | |
969 | ||
970 | Because access privileges (visibility) to members and member functions | |
971 | may differ from one function context to the next, G++ may need to flush | |
972 | the cache. With the @samp{-fmemoize-lookups} flag, the cache is flushed | |
973 | after every function that is compiled. The @samp{-fsave-memoized} flag | |
974 | enables the same software cache, but when the compiler determines that | |
975 | the context of the last function compiled would yield the same access | |
976 | privileges of the next function to compile, it preserves the cache. | |
977 | This is most helpful when defining many member functions for the same | |
978 | class: with the exception of member functions which are friends of other | |
979 | classes, each member function has exactly the same access privileges as | |
980 | every other, and the cache need not be flushed. | |
981 | ||
982 | The code that implements these flags has rotted; you should probably | |
983 | avoid using them. | |
984 | ||
985 | @item -fstrict-prototype | |
986 | Within an @samp{extern "C"} linkage specification, treat a function | |
987 | declaration with no arguments, such as @samp{int foo ();}, as declaring | |
988 | the function to take no arguments. Normally, such a declaration means | |
989 | that the function @code{foo} can take any combination of arguments, as | |
990 | in C. @samp{-pedantic} implies @samp{-fstrict-prototype} unless | |
991 | overridden with @samp{-fno-strict-prototype}. | |
992 | ||
993 | This flag no longer affects declarations with C++ linkage. | |
994 | ||
995 | @item -fno-nonnull-objects | |
996 | Don't assume that a reference is initialized to refer to a valid object. | |
997 | Although the current C++ Working Paper prohibits null references, some | |
998 | old code may rely on them, and you can use @samp{-fno-nonnull-objects} | |
999 | to turn on checking. | |
1000 | ||
1001 | At the moment, the compiler only does this checking for conversions to | |
1002 | virtual base classes. | |
1003 | ||
1004 | @item -foperator-names | |
1005 | Recognize the operator name keywords @code{and}, @code{bitand}, | |
1006 | @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as | |
1007 | synonyms for the symbols they refer to. @samp{-ansi} implies | |
1008 | @samp{-foperator-names}. | |
1009 | ||
1010 | @item -fthis-is-variable | |
1011 | Permit assignment to @code{this}. The incorporation of user-defined | |
1012 | free store management into C++ has made assignment to @samp{this} an | |
1013 | anachronism. Therefore, by default it is invalid to assign to | |
1014 | @code{this} within a class member function; that is, GNU C++ treats | |
1015 | @samp{this} in a member function of class @code{X} as a non-lvalue of | |
1016 | type @samp{X *}. However, for backwards compatibility, you can make it | |
1017 | valid with @samp{-fthis-is-variable}. | |
1018 | ||
1019 | @item -fvtable-thunks | |
1020 | Use @samp{thunks} to implement the virtual function dispatch table | |
1021 | (@samp{vtable}). The traditional (cfront-style) approach to | |
1022 | implementing vtables was to store a pointer to the function and two | |
1023 | offsets for adjusting the @samp{this} pointer at the call site. Newer | |
1024 | implementations store a single pointer to a @samp{thunk} function which | |
1025 | does any necessary adjustment and then calls the target function. | |
1026 | ||
1027 | This option also enables a heuristic for controlling emission of | |
1028 | vtables; if a class has any non-inline virtual functions, the vtable | |
1029 | will be emitted in the translation unit containing the first one of | |
1030 | those. | |
1031 | ||
1032 | @item -nostdinc++ | |
1033 | Do not search for header files in the standard directories specific to | |
1034 | C++, but do still search the other standard directories. (This option | |
1035 | is used when building libg++.) | |
1036 | ||
1037 | @item -traditional | |
1038 | For C++ programs (in addition to the effects that apply to both C and | |
1039 | C++), this has the same effect as @samp{-fthis-is-variable}. | |
1040 | @xref{C Dialect Options,, Options Controlling C Dialect}. | |
1041 | @end table | |
1042 | ||
1043 | In addition, these optimization, warning, and code generation options | |
1044 | have meanings only for C++ programs: | |
1045 | ||
1046 | @table @code | |
1047 | @item -fno-default-inline | |
1048 | Do not assume @samp{inline} for functions defined inside a class scope. | |
1049 | @xref{Optimize Options,,Options That Control Optimization}. | |
1050 | ||
1051 | @item -Woverloaded-virtual | |
1052 | @itemx -Wtemplate-debugging | |
1053 | Warnings that apply only to C++ programs. @xref{Warning | |
1054 | Options,,Options to Request or Suppress Warnings}. | |
1055 | ||
1056 | @item +e@var{n} | |
1057 | Control how virtual function definitions are used, in a fashion | |
1058 | compatible with @code{cfront} 1.x. @xref{Code Gen Options,,Options for | |
1059 | Code Generation Conventions}. | |
1060 | @end table | |
1061 | ||
1062 | @node Warning Options | |
1063 | @section Options to Request or Suppress Warnings | |
1064 | @cindex options to control warnings | |
1065 | @cindex warning messages | |
1066 | @cindex messages, warning | |
1067 | @cindex suppressing warnings | |
1068 | ||
1069 | Warnings are diagnostic messages that report constructions which | |
1070 | are not inherently erroneous but which are risky or suggest there | |
1071 | may have been an error. | |
1072 | ||
1073 | You can request many specific warnings with options beginning @samp{-W}, | |
1074 | for example @samp{-Wimplicit} to request warnings on implicit | |
1075 | declarations. Each of these specific warning options also has a | |
1076 | negative form beginning @samp{-Wno-} to turn off warnings; | |
1077 | for example, @samp{-Wno-implicit}. This manual lists only one of the | |
1078 | two forms, whichever is not the default. | |
1079 | ||
1080 | These options control the amount and kinds of warnings produced by GNU | |
1081 | CC: | |
1082 | ||
1083 | @table @code | |
1084 | @cindex syntax checking | |
1085 | @item -fsyntax-only | |
1086 | Check the code for syntax errors, but don't do anything beyond that. | |
1087 | ||
1088 | @item -pedantic | |
1089 | Issue all the warnings demanded by strict ANSI standard C; reject | |
1090 | all programs that use forbidden extensions. | |
1091 | ||
1092 | Valid ANSI standard C programs should compile properly with or without | |
1093 | this option (though a rare few will require @samp{-ansi}). However, | |
1094 | without this option, certain GNU extensions and traditional C features | |
1095 | are supported as well. With this option, they are rejected. | |
1096 | ||
1097 | @samp{-pedantic} does not cause warning messages for use of the | |
1098 | alternate keywords whose names begin and end with @samp{__}. Pedantic | |
1099 | warnings are also disabled in the expression that follows | |
1100 | @code{__extension__}. However, only system header files should use | |
1101 | these escape routes; application programs should avoid them. | |
1102 | @xref{Alternate Keywords}. | |
1103 | ||
1104 | This option is not intended to be @i{useful}; it exists only to satisfy | |
1105 | pedants who would otherwise claim that GNU CC fails to support the ANSI | |
1106 | standard. | |
1107 | ||
1108 | Some users try to use @samp{-pedantic} to check programs for strict ANSI | |
1109 | C conformance. They soon find that it does not do quite what they want: | |
1110 | it finds some non-ANSI practices, but not all---only those for which | |
1111 | ANSI C @emph{requires} a diagnostic. | |
1112 | ||
1113 | A feature to report any failure to conform to ANSI C might be useful in | |
1114 | some instances, but would require considerable additional work and would | |
1115 | be quite different from @samp{-pedantic}. We recommend, rather, that | |
1116 | users take advantage of the extensions of GNU C and disregard the | |
1117 | limitations of other compilers. Aside from certain supercomputers and | |
1118 | obsolete small machines, there is less and less reason ever to use any | |
1119 | other C compiler other than for bootstrapping GNU CC. | |
1120 | ||
1121 | @item -pedantic-errors | |
1122 | Like @samp{-pedantic}, except that errors are produced rather than | |
1123 | warnings. | |
1124 | ||
1125 | @item -w | |
1126 | Inhibit all warning messages. | |
1127 | ||
1128 | @item -Wno-import | |
1129 | Inhibit warning messages about the use of @samp{#import}. | |
1130 | ||
1131 | @item -Wchar-subscripts | |
1132 | Warn if an array subscript has type @code{char}. This is a common cause | |
1133 | of error, as programmers often forget that this type is signed on some | |
1134 | machines. | |
1135 | ||
1136 | @item -Wcomment | |
1137 | Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*} | |
1138 | comment, or whenever a Backslash-Newline appears in a @samp{//} comment. | |
1139 | ||
1140 | @item -Wformat | |
1141 | Check calls to @code{printf} and @code{scanf}, etc., to make sure that | |
1142 | the arguments supplied have types appropriate to the format string | |
1143 | specified. | |
1144 | ||
1145 | @item -Wimplicit | |
1146 | Warn whenever a function or parameter is implicitly declared. | |
1147 | ||
1148 | @item -Wparentheses | |
1149 | Warn if parentheses are omitted in certain contexts, such | |
1150 | as when there is an assignment in a context where a truth value | |
1151 | is expected, or when operators are nested whose precedence people | |
1152 | often get confused about. | |
1153 | ||
1154 | @item -Wreturn-type | |
1155 | Warn whenever a function is defined with a return-type that defaults | |
1156 | to @code{int}. Also warn about any @code{return} statement with no | |
1157 | return-value in a function whose return-type is not @code{void}. | |
1158 | ||
1159 | @item -Wswitch | |
1160 | Warn whenever a @code{switch} statement has an index of enumeral type | |
1161 | and lacks a @code{case} for one or more of the named codes of that | |
1162 | enumeration. (The presence of a @code{default} label prevents this | |
1163 | warning.) @code{case} labels outside the enumeration range also | |
1164 | provoke warnings when this option is used. | |
1165 | ||
1166 | @item -Wtrigraphs | |
1167 | Warn if any trigraphs are encountered (assuming they are enabled). | |
1168 | ||
1169 | @item -Wunused | |
1170 | Warn whenever a variable is unused aside from its declaration, | |
1171 | whenever a function is declared static but never defined, whenever a | |
1172 | label is declared but not used, and whenever a statement computes a | |
1173 | result that is explicitly not used. | |
1174 | ||
1175 | To suppress this warning for an expression, simply cast it to void. For | |
1176 | unused variables and parameters, use the @samp{unused} attribute | |
1177 | (@pxref{Variable Attributes}). | |
1178 | ||
1179 | @item -Wuninitialized | |
1180 | An automatic variable is used without first being initialized. | |
1181 | ||
1182 | These warnings are possible only in optimizing compilation, | |
1183 | because they require data flow information that is computed only | |
1184 | when optimizing. If you don't specify @samp{-O}, you simply won't | |
1185 | get these warnings. | |
1186 | ||
1187 | These warnings occur only for variables that are candidates for | |
1188 | register allocation. Therefore, they do not occur for a variable that | |
1189 | is declared @code{volatile}, or whose address is taken, or whose size | |
1190 | is other than 1, 2, 4 or 8 bytes. Also, they do not occur for | |
1191 | structures, unions or arrays, even when they are in registers. | |
1192 | ||
1193 | Note that there may be no warning about a variable that is used only | |
1194 | to compute a value that itself is never used, because such | |
1195 | computations may be deleted by data flow analysis before the warnings | |
1196 | are printed. | |
1197 | ||
1198 | These warnings are made optional because GNU CC is not smart | |
1199 | enough to see all the reasons why the code might be correct | |
1200 | despite appearing to have an error. Here is one example of how | |
1201 | this can happen: | |
1202 | ||
1203 | @smallexample | |
1204 | @{ | |
1205 | int x; | |
1206 | switch (y) | |
1207 | @{ | |
1208 | case 1: x = 1; | |
1209 | break; | |
1210 | case 2: x = 4; | |
1211 | break; | |
1212 | case 3: x = 5; | |
1213 | @} | |
1214 | foo (x); | |
1215 | @} | |
1216 | @end smallexample | |
1217 | ||
1218 | @noindent | |
1219 | If the value of @code{y} is always 1, 2 or 3, then @code{x} is | |
1220 | always initialized, but GNU CC doesn't know this. Here is | |
1221 | another common case: | |
1222 | ||
1223 | @smallexample | |
1224 | @{ | |
1225 | int save_y; | |
1226 | if (change_y) save_y = y, y = new_y; | |
1227 | @dots{} | |
1228 | if (change_y) y = save_y; | |
1229 | @} | |
1230 | @end smallexample | |
1231 | ||
1232 | @noindent | |
1233 | This has no bug because @code{save_y} is used only if it is set. | |
1234 | ||
1235 | Some spurious warnings can be avoided if you declare all the functions | |
1236 | you use that never return as @code{noreturn}. @xref{Function | |
1237 | Attributes}. | |
1238 | ||
1239 | @item -Wreorder (C++ only) | |
1240 | @cindex reordering, warning | |
1241 | @cindex warning for reordering of member initializers | |
1242 | Warn when the order of member initializers given in the code does not | |
1243 | match the order in which they must be executed. For instance: | |
1244 | ||
1245 | @smallexample | |
1246 | struct A @{ | |
1247 | int i; | |
1248 | int j; | |
1249 | A(): j (0), i (1) @{ @} | |
1250 | @}; | |
1251 | @end smallexample | |
1252 | ||
1253 | Here the compiler will warn that the member initializers for @samp{i} | |
1254 | and @samp{j} will be rearranged to match the declaration order of the | |
1255 | members. | |
1256 | ||
1257 | @item -Wsign-compare | |
1258 | @cindex warning for comparison of signed and unsigned values | |
1259 | @cindex comparison of signed and unsigned values, warning | |
1260 | @cindex signed and unsigned values, comparison warning | |
1261 | Warn when a comparison between signed and unsigned values could produce | |
1262 | an incorrect result when the signed value is converted to unsigned. | |
1263 | ||
1264 | @item -Wtemplate-debugging | |
1265 | @cindex template debugging | |
1266 | When using templates in a C++ program, warn if debugging is not yet | |
1267 | fully available (C++ only). | |
1268 | ||
1269 | @item -Wall | |
1270 | All of the above @samp{-W} options combined. This enables all the | |
1271 | warnings about constructions that some users consider questionable, and | |
1272 | that are easy to avoid (or modify to prevent the warning), even in | |
1273 | conjunction with macros. | |
1274 | @end table | |
1275 | ||
1276 | The following @samp{-W@dots{}} options are not implied by @samp{-Wall}. | |
1277 | Some of them warn about constructions that users generally do not | |
1278 | consider questionable, but which occasionally you might wish to check | |
1279 | for; others warn about constructions that are necessary or hard to avoid | |
1280 | in some cases, and there is no simple way to modify the code to suppress | |
1281 | the warning. | |
1282 | ||
1283 | @table @code | |
1284 | @item -W | |
1285 | Print extra warning messages for these events: | |
1286 | ||
1287 | @itemize @bullet | |
1288 | @cindex @code{longjmp} warnings | |
1289 | @item | |
1290 | A nonvolatile automatic variable might be changed by a call to | |
1291 | @code{longjmp}. These warnings as well are possible only in | |
1292 | optimizing compilation. | |
1293 | ||
1294 | The compiler sees only the calls to @code{setjmp}. It cannot know | |
1295 | where @code{longjmp} will be called; in fact, a signal handler could | |
1296 | call it at any point in the code. As a result, you may get a warning | |
1297 | even when there is in fact no problem because @code{longjmp} cannot | |
1298 | in fact be called at the place which would cause a problem. | |
1299 | ||
1300 | @item | |
1301 | A function can return either with or without a value. (Falling | |
1302 | off the end of the function body is considered returning without | |
1303 | a value.) For example, this function would evoke such a | |
1304 | warning: | |
1305 | ||
1306 | @smallexample | |
1307 | @group | |
1308 | foo (a) | |
1309 | @{ | |
1310 | if (a > 0) | |
1311 | return a; | |
1312 | @} | |
1313 | @end group | |
1314 | @end smallexample | |
1315 | ||
1316 | @item | |
1317 | An expression-statement or the left-hand side of a comma expression | |
1318 | contains no side effects. | |
1319 | To suppress the warning, cast the unused expression to void. | |
1320 | For example, an expression such as @samp{x[i,j]} will cause a warning, | |
1321 | but @samp{x[(void)i,j]} will not. | |
1322 | ||
1323 | @item | |
1324 | An unsigned value is compared against zero with @samp{<} or @samp{<=}. | |
1325 | ||
1326 | @item | |
1327 | A comparison like @samp{x<=y<=z} appears; this is equivalent to | |
1328 | @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from | |
1329 | that of ordinary mathematical notation. | |
1330 | ||
1331 | @item | |
1332 | Storage-class specifiers like @code{static} are not the first things in | |
1333 | a declaration. According to the C Standard, this usage is obsolescent. | |
1334 | ||
1335 | @item | |
1336 | If @samp{-Wall} or @samp{-Wunused} is also specified, warn about unused | |
1337 | arguments. | |
1338 | ||
1339 | @item | |
1340 | An aggregate has a partly bracketed initializer. | |
1341 | For example, the following code would evoke such a warning, | |
1342 | because braces are missing around the initializer for @code{x.h}: | |
1343 | ||
1344 | @smallexample | |
1345 | struct s @{ int f, g; @}; | |
1346 | struct t @{ struct s h; int i; @}; | |
1347 | struct t x = @{ 1, 2, 3 @}; | |
1348 | @end smallexample | |
1349 | @end itemize | |
1350 | ||
1351 | @item -Wtraditional | |
1352 | Warn about certain constructs that behave differently in traditional and | |
1353 | ANSI C. | |
1354 | ||
1355 | @itemize @bullet | |
1356 | @item | |
1357 | Macro arguments occurring within string constants in the macro body. | |
1358 | These would substitute the argument in traditional C, but are part of | |
1359 | the constant in ANSI C. | |
1360 | ||
1361 | @item | |
1362 | A function declared external in one block and then used after the end of | |
1363 | the block. | |
1364 | ||
1365 | @item | |
1366 | A @code{switch} statement has an operand of type @code{long}. | |
1367 | @end itemize | |
1368 | ||
1369 | @item -Wshadow | |
1370 | Warn whenever a local variable shadows another local variable. | |
1371 | ||
1372 | @item -Wid-clash-@var{len} | |
1373 | Warn whenever two distinct identifiers match in the first @var{len} | |
1374 | characters. This may help you prepare a program that will compile | |
1375 | with certain obsolete, brain-damaged compilers. | |
1376 | ||
1377 | @item -Wlarger-than-@var{len} | |
1378 | Warn whenever an object of larger than @var{len} bytes is defined. | |
1379 | ||
1380 | @item -Wpointer-arith | |
1381 | Warn about anything that depends on the ``size of'' a function type or | |
1382 | of @code{void}. GNU C assigns these types a size of 1, for | |
1383 | convenience in calculations with @code{void *} pointers and pointers | |
1384 | to functions. | |
1385 | ||
1386 | @item -Wbad-function-cast | |
1387 | Warn whenever a function call is cast to a non-matching type. | |
1388 | For example, warn if @code{int malloc()} is cast to @code{anything *}. | |
1389 | ||
1390 | @item -Wcast-qual | |
1391 | Warn whenever a pointer is cast so as to remove a type qualifier from | |
1392 | the target type. For example, warn if a @code{const char *} is cast | |
1393 | to an ordinary @code{char *}. | |
1394 | ||
1395 | @item -Wcast-align | |
1396 | Warn whenever a pointer is cast such that the required alignment of the | |
1397 | target is increased. For example, warn if a @code{char *} is cast to | |
1398 | an @code{int *} on machines where integers can only be accessed at | |
1399 | two- or four-byte boundaries. | |
1400 | ||
1401 | @item -Wwrite-strings | |
1402 | Give string constants the type @code{const char[@var{length}]} so that | |
1403 | copying the address of one into a non-@code{const} @code{char *} | |
1404 | pointer will get a warning. These warnings will help you find at | |
1405 | compile time code that can try to write into a string constant, but | |
1406 | only if you have been very careful about using @code{const} in | |
1407 | declarations and prototypes. Otherwise, it will just be a nuisance; | |
1408 | this is why we did not make @samp{-Wall} request these warnings. | |
1409 | ||
1410 | @item -Wconversion | |
1411 | Warn if a prototype causes a type conversion that is different from what | |
1412 | would happen to the same argument in the absence of a prototype. This | |
1413 | includes conversions of fixed point to floating and vice versa, and | |
1414 | conversions changing the width or signedness of a fixed point argument | |
1415 | except when the same as the default promotion. | |
1416 | ||
1417 | Also, warn if a negative integer constant expression is implicitly | |
1418 | converted to an unsigned type. For example, warn about the assignment | |
1419 | @code{x = -1} if @code{x} is unsigned. But do not warn about explicit | |
1420 | casts like @code{(unsigned) -1}. | |
1421 | ||
1422 | @item -Waggregate-return | |
1423 | Warn if any functions that return structures or unions are defined or | |
1424 | called. (In languages where you can return an array, this also elicits | |
1425 | a warning.) | |
1426 | ||
1427 | @item -Wstrict-prototypes | |
1428 | Warn if a function is declared or defined without specifying the | |
1429 | argument types. (An old-style function definition is permitted without | |
1430 | a warning if preceded by a declaration which specifies the argument | |
1431 | types.) | |
1432 | ||
1433 | @item -Wmissing-prototypes | |
1434 | Warn if a global function is defined without a previous prototype | |
1435 | declaration. This warning is issued even if the definition itself | |
1436 | provides a prototype. The aim is to detect global functions that fail | |
1437 | to be declared in header files. | |
1438 | ||
1439 | @item -Wmissing-declarations | |
1440 | Warn if a global function is defined without a previous declaration. | |
1441 | Do so even if the definition itself provides a prototype. | |
1442 | Use this option to detect global functions that are not declared in | |
1443 | header files. | |
1444 | ||
1445 | @item -Wredundant-decls | |
1446 | Warn if anything is declared more than once in the same scope, even in | |
1447 | cases where multiple declaration is valid and changes nothing. | |
1448 | ||
1449 | @item -Wnested-externs | |
1450 | Warn if an @code{extern} declaration is encountered within an function. | |
1451 | ||
1452 | @item -Winline | |
1453 | Warn if a function can not be inlined, and either it was declared as inline, | |
1454 | or else the @samp{-finline-functions} option was given. | |
1455 | ||
1456 | @item -Woverloaded-virtual | |
1457 | @cindex overloaded virtual fn, warning | |
1458 | @cindex warning for overloaded virtual fn | |
1459 | Warn when a derived class function declaration may be an error in | |
1460 | defining a virtual function (C++ only). In a derived class, the | |
1461 | definitions of virtual functions must match the type signature of a | |
1462 | virtual function declared in the base class. With this option, the | |
1463 | compiler warns when you define a function with the same name as a | |
1464 | virtual function, but with a type signature that does not match any | |
1465 | declarations from the base class. | |
1466 | ||
1467 | @item -Wsynth (C++ only) | |
1468 | @cindex warning for synthesized methods | |
1469 | @cindex synthesized methods, warning | |
1470 | Warn when g++'s synthesis behavior does not match that of cfront. For | |
1471 | instance: | |
1472 | ||
1473 | @smallexample | |
1474 | struct A @{ | |
1475 | operator int (); | |
1476 | A& operator = (int); | |
1477 | @}; | |
1478 | ||
1479 | main () | |
1480 | @{ | |
1481 | A a,b; | |
1482 | a = b; | |
1483 | @} | |
1484 | @end smallexample | |
1485 | ||
1486 | In this example, g++ will synthesize a default @samp{A& operator = | |
1487 | (const A&);}, while cfront will use the user-defined @samp{operator =}. | |
1488 | ||
1489 | @item -Werror | |
1490 | Make all warnings into errors. | |
1491 | @end table | |
1492 | ||
1493 | @node Debugging Options | |
1494 | @section Options for Debugging Your Program or GNU CC | |
1495 | @cindex options, debugging | |
1496 | @cindex debugging information options | |
1497 | ||
1498 | GNU CC has various special options that are used for debugging | |
1499 | either your program or GCC: | |
1500 | ||
1501 | @table @code | |
1502 | @item -g | |
1503 | Produce debugging information in the operating system's native format | |
1504 | (stabs, COFF, XCOFF, or DWARF). GDB can work with this debugging | |
1505 | information. | |
1506 | ||
1507 | On most systems that use stabs format, @samp{-g} enables use of extra | |
1508 | debugging information that only GDB can use; this extra information | |
1509 | makes debugging work better in GDB but will probably make other debuggers | |
1510 | crash or | |
1511 | refuse to read the program. If you want to control for certain whether | |
1512 | to generate the extra information, use @samp{-gstabs+}, @samp{-gstabs}, | |
1513 | @samp{-gxcoff+}, @samp{-gxcoff}, @samp{-gdwarf+}, or @samp{-gdwarf} | |
1514 | (see below). | |
1515 | ||
1516 | Unlike most other C compilers, GNU CC allows you to use @samp{-g} with | |
1517 | @samp{-O}. The shortcuts taken by optimized code may occasionally | |
1518 | produce surprising results: some variables you declared may not exist | |
1519 | at all; flow of control may briefly move where you did not expect it; | |
1520 | some statements may not be executed because they compute constant | |
1521 | results or their values were already at hand; some statements may | |
1522 | execute in different places because they were moved out of loops. | |
1523 | ||
1524 | Nevertheless it proves possible to debug optimized output. This makes | |
1525 | it reasonable to use the optimizer for programs that might have bugs. | |
1526 | ||
1527 | The following options are useful when GNU CC is generated with the | |
1528 | capability for more than one debugging format. | |
1529 | ||
1530 | @item -ggdb | |
1531 | Produce debugging information in the native format (if that is supported), | |
1532 | including GDB extensions if at all possible. | |
1533 | ||
1534 | @item -gstabs | |
1535 | Produce debugging information in stabs format (if that is supported), | |
1536 | without GDB extensions. This is the format used by DBX on most BSD | |
1537 | systems. On MIPS, Alpha and System V Release 4 systems this option | |
1538 | produces stabs debugging output which is not understood by DBX or SDB. | |
1539 | On System V Release 4 systems this option requires the GNU assembler. | |
1540 | ||
1541 | @item -gstabs+ | |
1542 | Produce debugging information in stabs format (if that is supported), | |
1543 | using GNU extensions understood only by the GNU debugger (GDB). The | |
1544 | use of these extensions is likely to make other debuggers crash or | |
1545 | refuse to read the program. | |
1546 | ||
1547 | @item -gcoff | |
1548 | Produce debugging information in COFF format (if that is supported). | |
1549 | This is the format used by SDB on most System V systems prior to | |
1550 | System V Release 4. | |
1551 | ||
1552 | @item -gxcoff | |
1553 | Produce debugging information in XCOFF format (if that is supported). | |
1554 | This is the format used by the DBX debugger on IBM RS/6000 systems. | |
1555 | ||
1556 | @item -gxcoff+ | |
1557 | Produce debugging information in XCOFF format (if that is supported), | |
1558 | using GNU extensions understood only by the GNU debugger (GDB). The | |
1559 | use of these extensions is likely to make other debuggers crash or | |
1560 | refuse to read the program, and may cause assemblers other than the GNU | |
1561 | assembler (GAS) to fail with an error. | |
1562 | ||
1563 | @item -gdwarf | |
1564 | Produce debugging information in DWARF format (if that is supported). | |
1565 | This is the format used by SDB on most System V Release 4 systems. | |
1566 | ||
1567 | @item -gdwarf+ | |
1568 | Produce debugging information in DWARF format (if that is supported), | |
1569 | using GNU extensions understood only by the GNU debugger (GDB). The | |
1570 | use of these extensions is likely to make other debuggers crash or | |
1571 | refuse to read the program. | |
1572 | ||
1573 | @item -g@var{level} | |
1574 | @itemx -ggdb@var{level} | |
1575 | @itemx -gstabs@var{level} | |
1576 | @itemx -gcoff@var{level} | |
1577 | @itemx -gxcoff@var{level} | |
1578 | @itemx -gdwarf@var{level} | |
1579 | Request debugging information and also use @var{level} to specify how | |
1580 | much information. The default level is 2. | |
1581 | ||
1582 | Level 1 produces minimal information, enough for making backtraces in | |
1583 | parts of the program that you don't plan to debug. This includes | |
1584 | descriptions of functions and external variables, but no information | |
1585 | about local variables and no line numbers. | |
1586 | ||
1587 | Level 3 includes extra information, such as all the macro definitions | |
1588 | present in the program. Some debuggers support macro expansion when | |
1589 | you use @samp{-g3}. | |
1590 | ||
1591 | @cindex @code{prof} | |
1592 | @item -p | |
1593 | Generate extra code to write profile information suitable for the | |
1594 | analysis program @code{prof}. You must use this option when compiling | |
1595 | the source files you want data about, and you must also use it when | |
1596 | linking. | |
1597 | ||
1598 | @cindex @code{gprof} | |
1599 | @item -pg | |
1600 | Generate extra code to write profile information suitable for the | |
1601 | analysis program @code{gprof}. You must use this option when compiling | |
1602 | the source files you want data about, and you must also use it when | |
1603 | linking. | |
1604 | ||
1605 | @cindex @code{tcov} | |
1606 | @item -a | |
1607 | Generate extra code to write profile information for basic blocks, which will | |
1608 | record the number of times each basic block is executed, the basic block start | |
1609 | address, and the function name containing the basic block. If @samp{-g} is | |
1610 | used, the line number and filename of the start of the basic block will also be | |
1611 | recorded. If not overridden by the machine description, the default action is | |
1612 | to append to the text file @file{bb.out}. | |
1613 | ||
1614 | This data could be analyzed by a program like @code{tcov}. Note, | |
1615 | however, that the format of the data is not what @code{tcov} expects. | |
1616 | Eventually GNU @code{gprof} should be extended to process this data. | |
1617 | ||
1618 | @item -ax | |
1619 | Generate extra code to profile basic blocks. Your executable will | |
1620 | produce output that is a superset of that produced when @samp{-a} is | |
1621 | used. Additional output is the source and target address of the basic | |
1622 | blocks where a jump takes place, the number of times a jump is executed, | |
1623 | and (optionally) the complete sequence of basic blocks being executed. | |
1624 | The output is appended to file @file{bb.out}. | |
1625 | ||
1626 | You can examine different profiling aspects without recompilation. Your | |
1627 | execuable will read a list of function names from file @file{bb.in}. | |
1628 | Profiling starts when a function on the list is entered and stops when | |
1629 | that invocation is exited. To exclude a function from profiling, prefix | |
1630 | its name with `-'. If a function name is not unique, you can | |
1631 | disambiguate it by writing it in the form | |
1632 | @samp{/path/filename.d:functionname}. Your executable will write the | |
1633 | available paths and filenames in file @file{bb.out}. | |
1634 | ||
1635 | Several function names have a special meaning: | |
1636 | @table @code | |
1637 | @item __bb_jumps__ | |
1638 | Write source, target and frequency of jumps to file @file{bb.out}. | |
1639 | @item __bb_hidecall__ | |
1640 | Exclude function calls from frequency count. | |
1641 | @item __bb_showret__ | |
1642 | Include function returns in frequency count. | |
1643 | @item __bb_trace__ | |
1644 | Write the sequence of basic blocks executed to file @file{bbtrace.gz}. | |
1645 | The file will be compressed using the program @samp{gzip}, which must | |
1646 | exist in your @code{PATH}. On systems without the @samp{popen} | |
1647 | function, the file will be named @file{bbtrace} and will not be | |
1648 | compressed. @strong{Profiling for even a few seconds on these systems | |
1649 | will produce a very large file.} Note: @code{__bb_hidecall__} and | |
1650 | @code{__bb_showret__} will not affect the sequence written to | |
1651 | @file{bbtrace.gz}. | |
1652 | @end table | |
1653 | ||
1654 | Here's a short example using different profiling parameters | |
1655 | in file @file{bb.in}. Assume function @code{foo} consists of basic blocks | |
1656 | 1 and 2 and is called twice from block 3 of function @code{main}. After | |
1657 | the calls, block 3 transfers control to block 4 of @code{main}. | |
1658 | ||
1659 | With @code{__bb_trace__} and @code{main} contained in file @file{bb.in}, | |
1660 | the following sequence of blocks is written to file @file{bbtrace.gz}: | |
1661 | 0 3 1 2 1 2 4. The return from block 2 to block 3 is not shown, because | |
1662 | the return is to a point inside the block and not to the top. The | |
1663 | block address 0 always indicates, that control is transferred | |
1664 | to the trace from somewhere outside the observed functions. With | |
1665 | @samp{-foo} added to @file{bb.in}, the blocks of function | |
1666 | @code{foo} are removed from the trace, so only 0 3 4 remains. | |
1667 | ||
1668 | With @code{__bb_jumps__} and @code{main} contained in file @file{bb.in}, | |
1669 | jump frequencies will be written to file @file{bb.out}. The | |
1670 | frequencies are obtained by constructing a trace of blocks | |
1671 | and incrementing a counter for every neighbouring pair of blocks | |
1672 | in the trace. The trace 0 3 1 2 1 2 4 displays the following | |
1673 | frequencies: | |
1674 | ||
1675 | @example | |
1676 | Jump from block 0x0 to block 0x3 executed 1 time(s) | |
1677 | Jump from block 0x3 to block 0x1 executed 1 time(s) | |
1678 | Jump from block 0x1 to block 0x2 executed 2 time(s) | |
1679 | Jump from block 0x2 to block 0x1 executed 1 time(s) | |
1680 | Jump from block 0x2 to block 0x4 executed 1 time(s) | |
1681 | @end example | |
1682 | ||
1683 | With @code{__bb_hidecall__}, control transfer due to call instructions | |
1684 | is removed from the trace, that is the trace is cut into three parts: 0 | |
1685 | 3 4, 0 1 2 and 0 1 2. With @code{__bb_showret__}, control transfer due | |
1686 | to return instructions is added to the trace. The trace becomes: 0 3 1 | |
1687 | 2 3 1 2 3 4. Note, that this trace is not the same, as the sequence | |
1688 | written to @file{bbtrace.gz}. It is solely used for counting jump | |
1689 | frequencies. | |
1690 | ||
1691 | @item -d@var{letters} | |
1692 | Says to make debugging dumps during compilation at times specified by | |
1693 | @var{letters}. This is used for debugging the compiler. The file names | |
1694 | for most of the dumps are made by appending a word to the source file | |
1695 | name (e.g. @file{foo.c.rtl} or @file{foo.c.jump}). Here are the | |
1696 | possible letters for use in @var{letters}, and their meanings: | |
1697 | ||
1698 | @table @samp | |
1699 | @item M | |
1700 | Dump all macro definitions, at the end of preprocessing, and write no | |
1701 | output. | |
1702 | @item N | |
1703 | Dump all macro names, at the end of preprocessing. | |
1704 | @item D | |
1705 | Dump all macro definitions, at the end of preprocessing, in addition to | |
1706 | normal output. | |
1707 | @item y | |
1708 | Dump debugging information during parsing, to standard error. | |
1709 | @item r | |
1710 | Dump after RTL generation, to @file{@var{file}.rtl}. | |
1711 | @item x | |
1712 | Just generate RTL for a function instead of compiling it. Usually used | |
1713 | with @samp{r}. | |
1714 | @item j | |
1715 | Dump after first jump optimization, to @file{@var{file}.jump}. | |
1716 | @item s | |
1717 | Dump after CSE (including the jump optimization that sometimes | |
1718 | follows CSE), to @file{@var{file}.cse}. | |
1719 | @item L | |
1720 | Dump after loop optimization, to @file{@var{file}.loop}. | |
1721 | @item t | |
1722 | Dump after the second CSE pass (including the jump optimization that | |
1723 | sometimes follows CSE), to @file{@var{file}.cse2}. | |
1724 | @item f | |
1725 | Dump after flow analysis, to @file{@var{file}.flow}. | |
1726 | @item c | |
1727 | Dump after instruction combination, to the file | |
1728 | @file{@var{file}.combine}. | |
1729 | @item S | |
1730 | Dump after the first instruction scheduling pass, to | |
1731 | @file{@var{file}.sched}. | |
1732 | @item l | |
1733 | Dump after local register allocation, to | |
1734 | @file{@var{file}.lreg}. | |
1735 | @item g | |
1736 | Dump after global register allocation, to | |
1737 | @file{@var{file}.greg}. | |
1738 | @item R | |
1739 | Dump after the second instruction scheduling pass, to | |
1740 | @file{@var{file}.sched2}. | |
1741 | @item J | |
1742 | Dump after last jump optimization, to @file{@var{file}.jump2}. | |
1743 | @item d | |
1744 | Dump after delayed branch scheduling, to @file{@var{file}.dbr}. | |
1745 | @item k | |
1746 | Dump after conversion from registers to stack, to @file{@var{file}.stack}. | |
1747 | @item a | |
1748 | Produce all the dumps listed above. | |
1749 | @item m | |
1750 | Print statistics on memory usage, at the end of the run, to | |
1751 | standard error. | |
1752 | @item p | |
1753 | Annotate the assembler output with a comment indicating which | |
1754 | pattern and alternative was used. | |
1755 | @item A | |
1756 | Annotate the assembler output with miscellaneous debugging information. | |
1757 | @end table | |
1758 | ||
1759 | @item -fpretend-float | |
1760 | When running a cross-compiler, pretend that the target machine uses the | |
1761 | same floating point format as the host machine. This causes incorrect | |
1762 | output of the actual floating constants, but the actual instruction | |
1763 | sequence will probably be the same as GNU CC would make when running on | |
1764 | the target machine. | |
1765 | ||
1766 | @item -save-temps | |
1767 | Store the usual ``temporary'' intermediate files permanently; place them | |
1768 | in the current directory and name them based on the source file. Thus, | |
1769 | compiling @file{foo.c} with @samp{-c -save-temps} would produce files | |
1770 | @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. | |
1771 | ||
1772 | @item -print-file-name=@var{library} | |
1773 | Print the full absolute name of the library file @var{library} that | |
1774 | would be used when linking---and don't do anything else. With this | |
1775 | option, GNU CC does not compile or link anything; it just prints the | |
1776 | file name. | |
1777 | ||
1778 | @item -print-prog-name=@var{program} | |
1779 | Like @samp{-print-file-name}, but searches for a program such as @samp{cpp}. | |
1780 | ||
1781 | @item -print-libgcc-file-name | |
1782 | Same as @samp{-print-file-name=libgcc.a}. | |
1783 | ||
1784 | This is useful when you use @samp{-nostdlib} or @samp{-nodefaultlibs} | |
1785 | but you do want to link with @file{libgcc.a}. You can do | |
1786 | ||
1787 | @example | |
1788 | gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name` | |
1789 | @end example | |
1790 | ||
1791 | @item -print-search-dirs | |
1792 | Print the name of the configured installation directory and a list of | |
1793 | program and library directories gcc will search---and don't do anything else. | |
1794 | ||
1795 | This is useful when gcc prints the error message | |
1796 | @samp{installation problem, cannot exec cpp: No such file or directory}. | |
1797 | To resolve this you either need to put @file{cpp} and the other compiler | |
1798 | components where gcc expects to find them, or you can set the environment | |
1799 | variable @code{GCC_EXEC_PREFIX} to the directory where you installed them. | |
1800 | Don't forget the trailing '/'. | |
1801 | @xref{Environment Variables}. | |
1802 | @end table | |
1803 | ||
1804 | @node Optimize Options | |
1805 | @section Options That Control Optimization | |
1806 | @cindex optimize options | |
1807 | @cindex options, optimization | |
1808 | ||
1809 | These options control various sorts of optimizations: | |
1810 | ||
1811 | @table @code | |
1812 | @item -O | |
1813 | @itemx -O1 | |
1814 | Optimize. Optimizing compilation takes somewhat more time, and a lot | |
1815 | more memory for a large function. | |
1816 | ||
1817 | Without @samp{-O}, the compiler's goal is to reduce the cost of | |
1818 | compilation and to make debugging produce the expected results. | |
1819 | Statements are independent: if you stop the program with a breakpoint | |
1820 | between statements, you can then assign a new value to any variable or | |
1821 | change the program counter to any other statement in the function and | |
1822 | get exactly the results you would expect from the source code. | |
1823 | ||
1824 | Without @samp{-O}, the compiler only allocates variables declared | |
1825 | @code{register} in registers. The resulting compiled code is a little | |
1826 | worse than produced by PCC without @samp{-O}. | |
1827 | ||
1828 | With @samp{-O}, the compiler tries to reduce code size and execution | |
1829 | time. | |
1830 | ||
1831 | When you specify @samp{-O}, the compiler turns on @samp{-fthread-jumps} | |
1832 | and @samp{-fdefer-pop} on all machines. The compiler turns on | |
1833 | @samp{-fdelayed-branch} on machines that have delay slots, and | |
1834 | @samp{-fomit-frame-pointer} on machines that can support debugging even | |
1835 | without a frame pointer. On some machines the compiler also turns | |
1836 | on other flags.@refill | |
1837 | ||
1838 | @item -O2 | |
1839 | Optimize even more. GNU CC performs nearly all supported optimizations | |
1840 | that do not involve a space-speed tradeoff. The compiler does not | |
1841 | perform loop unrolling or function inlining when you specify @samp{-O2}. | |
1842 | As compared to @samp{-O}, this option increases both compilation time | |
1843 | and the performance of the generated code. | |
1844 | ||
1845 | @samp{-O2} turns on all optional optimizations except for loop unrolling | |
1846 | and function inlining. It also turns on the @samp{-fforce-mem} option | |
1847 | on all machines and frame pointer elimination on machines where doing so | |
1848 | does not interfere with debugging. | |
1849 | ||
1850 | @item -O3 | |
1851 | Optimize yet more. @samp{-O3} turns on all optimizations specified by | |
1852 | @samp{-O2} and also turns on the @samp{inline-functions} option. | |
1853 | ||
1854 | @item -O0 | |
1855 | Do not optimize. | |
1856 | ||
1857 | If you use multiple @samp{-O} options, with or without level numbers, | |
1858 | the last such option is the one that is effective. | |
1859 | @end table | |
1860 | ||
1861 | Options of the form @samp{-f@var{flag}} specify machine-independent | |
1862 | flags. Most flags have both positive and negative forms; the negative | |
1863 | form of @samp{-ffoo} would be @samp{-fno-foo}. In the table below, | |
1864 | only one of the forms is listed---the one which is not the default. | |
1865 | You can figure out the other form by either removing @samp{no-} or | |
1866 | adding it. | |
1867 | ||
1868 | @table @code | |
1869 | @item -ffloat-store | |
1870 | Do not store floating point variables in registers, and inhibit other | |
1871 | options that might change whether a floating point value is taken from a | |
1872 | register or memory. | |
1873 | ||
1874 | @cindex floating point precision | |
1875 | This option prevents undesirable excess precision on machines such as | |
1876 | the 68000 where the floating registers (of the 68881) keep more | |
1877 | precision than a @code{double} is supposed to have. Similarly for the | |
1878 | x86 architecture. For most programs, the excess precision does only | |
1879 | good, but a few programs rely on the precise definition of IEEE floating | |
1880 | point. Use @samp{-ffloat-store} for such programs. | |
1881 | ||
1882 | @item -fno-default-inline | |
1883 | Do not make member functions inline by default merely because they are | |
1884 | defined inside the class scope (C++ only). Otherwise, when you specify | |
1885 | @w{@samp{-O}}, member functions defined inside class scope are compiled | |
1886 | inline by default; i.e., you don't need to add @samp{inline} in front of | |
1887 | the member function name. | |
1888 | ||
1889 | @item -fno-defer-pop | |
1890 | Always pop the arguments to each function call as soon as that function | |
1891 | returns. For machines which must pop arguments after a function call, | |
1892 | the compiler normally lets arguments accumulate on the stack for several | |
1893 | function calls and pops them all at once. | |
1894 | ||
1895 | @item -fforce-mem | |
1896 | Force memory operands to be copied into registers before doing | |
1897 | arithmetic on them. This produces better code by making all memory | |
1898 | references potential common subexpressions. When they are not common | |
1899 | subexpressions, instruction combination should eliminate the separate | |
1900 | register-load. The @samp{-O2} option turns on this option. | |
1901 | ||
1902 | @item -fforce-addr | |
1903 | Force memory address constants to be copied into registers before | |
1904 | doing arithmetic on them. This may produce better code just as | |
1905 | @samp{-fforce-mem} may. | |
1906 | ||
1907 | @item -fomit-frame-pointer | |
1908 | Don't keep the frame pointer in a register for functions that | |
1909 | don't need one. This avoids the instructions to save, set up and | |
1910 | restore frame pointers; it also makes an extra register available | |
1911 | in many functions. @strong{It also makes debugging impossible on | |
1912 | some machines.} | |
1913 | ||
1914 | @ifset INTERNALS | |
1915 | On some machines, such as the Vax, this flag has no effect, because | |
1916 | the standard calling sequence automatically handles the frame pointer | |
1917 | and nothing is saved by pretending it doesn't exist. The | |
1918 | machine-description macro @code{FRAME_POINTER_REQUIRED} controls | |
1919 | whether a target machine supports this flag. @xref{Registers}.@refill | |
1920 | @end ifset | |
1921 | @ifclear INTERNALS | |
1922 | On some machines, such as the Vax, this flag has no effect, because | |
1923 | the standard calling sequence automatically handles the frame pointer | |
1924 | and nothing is saved by pretending it doesn't exist. The | |
1925 | machine-description macro @code{FRAME_POINTER_REQUIRED} controls | |
1926 | whether a target machine supports this flag. @xref{Registers,,Register | |
1927 | Usage, gcc.info, Using and Porting GCC}.@refill | |
1928 | @end ifclear | |
1929 | ||
1930 | @item -fno-inline | |
1931 | Don't pay attention to the @code{inline} keyword. Normally this option | |
1932 | is used to keep the compiler from expanding any functions inline. | |
1933 | Note that if you are not optimizing, no functions can be expanded inline. | |
1934 | ||
1935 | @item -finline-functions | |
1936 | Integrate all simple functions into their callers. The compiler | |
1937 | heuristically decides which functions are simple enough to be worth | |
1938 | integrating in this way. | |
1939 | ||
1940 | If all calls to a given function are integrated, and the function is | |
1941 | declared @code{static}, then the function is normally not output as | |
1942 | assembler code in its own right. | |
1943 | ||
1944 | @item -fkeep-inline-functions | |
1945 | Even if all calls to a given function are integrated, and the function | |
1946 | is declared @code{static}, nevertheless output a separate run-time | |
1947 | callable version of the function. This switch does not affect | |
1948 | @code{extern inline} functions. | |
1949 | ||
1950 | @item -fkeep-static-consts | |
1951 | Emit variables declared @code{static const} when optimization isn't turned | |
1952 | on, even if the variables aren't referenced. | |
1953 | ||
1954 | GNU CC enables this option by default. If you want to force the compiler to | |
1955 | check if the variable was referenced, regardless of whether or not | |
1956 | optimization is turned on, use the @samp{-fno-keep-static-consts} option. | |
1957 | ||
1958 | @item -fno-function-cse | |
1959 | Do not put function addresses in registers; make each instruction that | |
1960 | calls a constant function contain the function's address explicitly. | |
1961 | ||
1962 | This option results in less efficient code, but some strange hacks | |
1963 | that alter the assembler output may be confused by the optimizations | |
1964 | performed when this option is not used. | |
1965 | ||
1966 | @item -ffast-math | |
1967 | This option allows GCC to violate some ANSI or IEEE rules and/or | |
1968 | specifications in the interest of optimizing code for speed. For | |
1969 | example, it allows the compiler to assume arguments to the @code{sqrt} | |
1970 | function are non-negative numbers and that no floating-point values | |
1971 | are NaNs. | |
1972 | ||
1973 | This option should never be turned on by any @samp{-O} option since | |
1974 | it can result in incorrect output for programs which depend on | |
1975 | an exact implementation of IEEE or ANSI rules/specifications for | |
1976 | math functions. | |
1977 | @end table | |
1978 | ||
1979 | @c following causes underfulls.. they don't look great, but we deal. | |
1980 | @c --mew 26jan93 | |
1981 | The following options control specific optimizations. The @samp{-O2} | |
1982 | option turns on all of these optimizations except @samp{-funroll-loops} | |
1983 | and @samp{-funroll-all-loops}. On most machines, the @samp{-O} option | |
1984 | turns on the @samp{-fthread-jumps} and @samp{-fdelayed-branch} options, | |
1985 | but specific machines may handle it differently. | |
1986 | ||
1987 | You can use the following flags in the rare cases when ``fine-tuning'' | |
1988 | of optimizations to be performed is desired. | |
1989 | ||
1990 | @table @code | |
1991 | @item -fstrength-reduce | |
1992 | Perform the optimizations of loop strength reduction and | |
1993 | elimination of iteration variables. | |
1994 | ||
1995 | @item -fthread-jumps | |
1996 | Perform optimizations where we check to see if a jump branches to a | |
1997 | location where another comparison subsumed by the first is found. If | |
1998 | so, the first branch is redirected to either the destination of the | |
1999 | second branch or a point immediately following it, depending on whether | |
2000 | the condition is known to be true or false. | |
2001 | ||
2002 | @item -fcse-follow-jumps | |
2003 | In common subexpression elimination, scan through jump instructions | |
2004 | when the target of the jump is not reached by any other path. For | |
2005 | example, when CSE encounters an @code{if} statement with an | |
2006 | @code{else} clause, CSE will follow the jump when the condition | |
2007 | tested is false. | |
2008 | ||
2009 | @item -fcse-skip-blocks | |
2010 | This is similar to @samp{-fcse-follow-jumps}, but causes CSE to | |
2011 | follow jumps which conditionally skip over blocks. When CSE | |
2012 | encounters a simple @code{if} statement with no else clause, | |
2013 | @samp{-fcse-skip-blocks} causes CSE to follow the jump around the | |
2014 | body of the @code{if}. | |
2015 | ||
2016 | @item -frerun-cse-after-loop | |
2017 | Re-run common subexpression elimination after loop optimizations has been | |
2018 | performed. | |
2019 | ||
2020 | @item -fexpensive-optimizations | |
2021 | Perform a number of minor optimizations that are relatively expensive. | |
2022 | ||
2023 | @item -fdelayed-branch | |
2024 | If supported for the target machine, attempt to reorder instructions | |
2025 | to exploit instruction slots available after delayed branch | |
2026 | instructions. | |
2027 | ||
2028 | @item -fschedule-insns | |
2029 | If supported for the target machine, attempt to reorder instructions to | |
2030 | eliminate execution stalls due to required data being unavailable. This | |
2031 | helps machines that have slow floating point or memory load instructions | |
2032 | by allowing other instructions to be issued until the result of the load | |
2033 | or floating point instruction is required. | |
2034 | ||
2035 | @item -fschedule-insns2 | |
2036 | Similar to @samp{-fschedule-insns}, but requests an additional pass of | |
2037 | instruction scheduling after register allocation has been done. This is | |
2038 | especially useful on machines with a relatively small number of | |
2039 | registers and where memory load instructions take more than one cycle. | |
2040 | ||
2041 | @item -ffunction-sections | |
2042 | Place each function into its own section in the output file if the | |
2043 | target supports arbitrary sections. The function's name determines | |
2044 | the section's name in the output file. | |
2045 | ||
2046 | Use this option on systems where the linker can perform optimizations | |
2047 | to improve locality of reference in the instruction space. HPPA | |
2048 | processors running HP-UX and Sparc processors running Solaris 2 have | |
2049 | linkers with such optimizations. Other systems using the ELF object format | |
2050 | as well as AIX may have these optimizations in the future. | |
2051 | ||
2052 | Only use this option when there are significant benefits from doing | |
2053 | so. When you specify this option, the assembler and linker will | |
2054 | create larger object and executable files and will also be slower. | |
2055 | You will not be able to use @code{gprof} on all systems if you | |
2056 | specify this option and you may have problems with debugging if | |
2057 | you specify both this option and @samp{-g}. | |
2058 | ||
2059 | @item -fcaller-saves | |
2060 | Enable values to be allocated in registers that will be clobbered by | |
2061 | function calls, by emitting extra instructions to save and restore the | |
2062 | registers around such calls. Such allocation is done only when it | |
2063 | seems to result in better code than would otherwise be produced. | |
2064 | ||
2065 | This option is enabled by default on certain machines, usually those | |
2066 | which have no call-preserved registers to use instead. | |
2067 | ||
2068 | @item -funroll-loops | |
2069 | Perform the optimization of loop unrolling. This is only done for loops | |
2070 | whose number of iterations can be determined at compile time or run time. | |
2071 | @samp{-funroll-loop} implies both @samp{-fstrength-reduce} and | |
2072 | @samp{-frerun-cse-after-loop}. | |
2073 | ||
2074 | @item -funroll-all-loops | |
2075 | Perform the optimization of loop unrolling. This is done for all loops | |
2076 | and usually makes programs run more slowly. @samp{-funroll-all-loops} | |
2077 | implies @samp{-fstrength-reduce} as well as @samp{-frerun-cse-after-loop}. | |
2078 | ||
2079 | @item -fno-peephole | |
2080 | Disable any machine-specific peephole optimizations. | |
2081 | @end table | |
2082 | ||
2083 | @node Preprocessor Options | |
2084 | @section Options Controlling the Preprocessor | |
2085 | @cindex preprocessor options | |
2086 | @cindex options, preprocessor | |
2087 | ||
2088 | These options control the C preprocessor, which is run on each C source | |
2089 | file before actual compilation. | |
2090 | ||
2091 | If you use the @samp{-E} option, nothing is done except preprocessing. | |
2092 | Some of these options make sense only together with @samp{-E} because | |
2093 | they cause the preprocessor output to be unsuitable for actual | |
2094 | compilation. | |
2095 | ||
2096 | @table @code | |
2097 | @item -include @var{file} | |
2098 | Process @var{file} as input before processing the regular input file. | |
2099 | In effect, the contents of @var{file} are compiled first. Any @samp{-D} | |
2100 | and @samp{-U} options on the command line are always processed before | |
2101 | @samp{-include @var{file}}, regardless of the order in which they are | |
2102 | written. All the @samp{-include} and @samp{-imacros} options are | |
2103 | processed in the order in which they are written. | |
2104 | ||
2105 | @item -imacros @var{file} | |
2106 | Process @var{file} as input, discarding the resulting output, before | |
2107 | processing the regular input file. Because the output generated from | |
2108 | @var{file} is discarded, the only effect of @samp{-imacros @var{file}} | |
2109 | is to make the macros defined in @var{file} available for use in the | |
2110 | main input. | |
2111 | ||
2112 | Any @samp{-D} and @samp{-U} options on the command line are always | |
2113 | processed before @samp{-imacros @var{file}}, regardless of the order in | |
2114 | which they are written. All the @samp{-include} and @samp{-imacros} | |
2115 | options are processed in the order in which they are written. | |
2116 | ||
2117 | @item -idirafter @var{dir} | |
2118 | @cindex second include path | |
2119 | Add the directory @var{dir} to the second include path. The directories | |
2120 | on the second include path are searched when a header file is not found | |
2121 | in any of the directories in the main include path (the one that | |
2122 | @samp{-I} adds to). | |
2123 | ||
2124 | @item -iprefix @var{prefix} | |
2125 | Specify @var{prefix} as the prefix for subsequent @samp{-iwithprefix} | |
2126 | options. | |
2127 | ||
2128 | @item -iwithprefix @var{dir} | |
2129 | Add a directory to the second include path. The directory's name is | |
2130 | made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was | |
2131 | specified previously with @samp{-iprefix}. If you have not specified a | |
2132 | prefix yet, the directory containing the installed passes of the | |
2133 | compiler is used as the default. | |
2134 | ||
2135 | @item -iwithprefixbefore @var{dir} | |
2136 | Add a directory to the main include path. The directory's name is made | |
2137 | by concatenating @var{prefix} and @var{dir}, as in the case of | |
2138 | @samp{-iwithprefix}. | |
2139 | ||
2140 | @item -isystem @var{dir} | |
2141 | Add a directory to the beginning of the second include path, marking it | |
2142 | as a system directory, so that it gets the same special treatment as | |
2143 | is applied to the standard system directories. | |
2144 | ||
2145 | @item -nostdinc | |
2146 | Do not search the standard system directories for header files. Only | |
2147 | the directories you have specified with @samp{-I} options (and the | |
2148 | current directory, if appropriate) are searched. @xref{Directory | |
2149 | Options}, for information on @samp{-I}. | |
2150 | ||
2151 | By using both @samp{-nostdinc} and @samp{-I-}, you can limit the include-file | |
2152 | search path to only those directories you specify explicitly. | |
2153 | ||
2154 | @item -undef | |
2155 | Do not predefine any nonstandard macros. (Including architecture flags). | |
2156 | ||
2157 | @item -E | |
2158 | Run only the C preprocessor. Preprocess all the C source files | |
2159 | specified and output the results to standard output or to the | |
2160 | specified output file. | |
2161 | ||
2162 | @item -C | |
2163 | Tell the preprocessor not to discard comments. Used with the | |
2164 | @samp{-E} option. | |
2165 | ||
2166 | @item -P | |
2167 | Tell the preprocessor not to generate @samp{#line} directives. | |
2168 | Used with the @samp{-E} option. | |
2169 | ||
2170 | @cindex make | |
2171 | @cindex dependencies, make | |
2172 | @item -M | |
2173 | Tell the preprocessor to output a rule suitable for @code{make} | |
2174 | describing the dependencies of each object file. For each source file, | |
2175 | the preprocessor outputs one @code{make}-rule whose target is the object | |
2176 | file name for that source file and whose dependencies are all the | |
2177 | @code{#include} header files it uses. This rule may be a single line or | |
2178 | may be continued with @samp{\}-newline if it is long. The list of rules | |
2179 | is printed on standard output instead of the preprocessed C program. | |
2180 | ||
2181 | @samp{-M} implies @samp{-E}. | |
2182 | ||
2183 | Another way to specify output of a @code{make} rule is by setting | |
2184 | the environment variable @code{DEPENDENCIES_OUTPUT} (@pxref{Environment | |
2185 | Variables}). | |
2186 | ||
2187 | @item -MM | |
2188 | Like @samp{-M} but the output mentions only the user header files | |
2189 | included with @samp{#include "@var{file}"}. System header files | |
2190 | included with @samp{#include <@var{file}>} are omitted. | |
2191 | ||
2192 | @item -MD | |
2193 | Like @samp{-M} but the dependency information is written to a file made by | |
2194 | replacing ".c" with ".d" at the end of the input file names. | |
2195 | This is in addition to compiling the file as specified---@samp{-MD} does | |
2196 | not inhibit ordinary compilation the way @samp{-M} does. | |
2197 | ||
2198 | In Mach, you can use the utility @code{md} to merge multiple dependency | |
2199 | files into a single dependency file suitable for using with the @samp{make} | |
2200 | command. | |
2201 | ||
2202 | @item -MMD | |
2203 | Like @samp{-MD} except mention only user header files, not system | |
2204 | header files. | |
2205 | ||
2206 | @item -MG | |
2207 | Treat missing header files as generated files and assume they live in the | |
2208 | same directory as the source file. If you specify @samp{-MG}, you | |
2209 | must also specify either @samp{-M} or @samp{-MM}. @samp{-MG} is not | |
2210 | supported with @samp{-MD} or @samp{-MMD}. | |
2211 | ||
2212 | @item -H | |
2213 | Print the name of each header file used, in addition to other normal | |
2214 | activities. | |
2215 | ||
2216 | @item -A@var{question}(@var{answer}) | |
2217 | Assert the answer @var{answer} for @var{question}, in case it is tested | |
2218 | with a preprocessing conditional such as @samp{#if | |
2219 | #@var{question}(@var{answer})}. @samp{-A-} disables the standard | |
2220 | assertions that normally describe the target machine. | |
2221 | ||
2222 | @item -D@var{macro} | |
2223 | Define macro @var{macro} with the string @samp{1} as its definition. | |
2224 | ||
2225 | @item -D@var{macro}=@var{defn} | |
2226 | Define macro @var{macro} as @var{defn}. All instances of @samp{-D} on | |
2227 | the command line are processed before any @samp{-U} options. | |
2228 | ||
2229 | @item -U@var{macro} | |
2230 | Undefine macro @var{macro}. @samp{-U} options are evaluated after all | |
2231 | @samp{-D} options, but before any @samp{-include} and @samp{-imacros} | |
2232 | options. | |
2233 | ||
2234 | @item -dM | |
2235 | Tell the preprocessor to output only a list of the macro definitions | |
2236 | that are in effect at the end of preprocessing. Used with the @samp{-E} | |
2237 | option. | |
2238 | ||
2239 | @item -dD | |
2240 | Tell the preprocessing to pass all macro definitions into the output, in | |
2241 | their proper sequence in the rest of the output. | |
2242 | ||
2243 | @item -dN | |
2244 | Like @samp{-dD} except that the macro arguments and contents are omitted. | |
2245 | Only @samp{#define @var{name}} is included in the output. | |
2246 | ||
2247 | @item -trigraphs | |
2248 | Support ANSI C trigraphs. The @samp{-ansi} option also has this effect. | |
2249 | ||
2250 | @item -Wp,@var{option} | |
2251 | Pass @var{option} as an option to the preprocessor. If @var{option} | |
2252 | contains commas, it is split into multiple options at the commas. | |
2253 | @end table | |
2254 | ||
2255 | @node Assembler Options | |
2256 | @section Passing Options to the Assembler | |
2257 | ||
2258 | @c prevent bad page break with this line | |
2259 | You can pass options to the assembler. | |
2260 | ||
2261 | @table @code | |
2262 | @item -Wa,@var{option} | |
2263 | Pass @var{option} as an option to the assembler. If @var{option} | |
2264 | contains commas, it is split into multiple options at the commas. | |
2265 | @end table | |
2266 | ||
2267 | @node Link Options | |
2268 | @section Options for Linking | |
2269 | @cindex link options | |
2270 | @cindex options, linking | |
2271 | ||
2272 | These options come into play when the compiler links object files into | |
2273 | an executable output file. They are meaningless if the compiler is | |
2274 | not doing a link step. | |
2275 | ||
2276 | @table @code | |
2277 | @cindex file names | |
2278 | @item @var{object-file-name} | |
2279 | A file name that does not end in a special recognized suffix is | |
2280 | considered to name an object file or library. (Object files are | |
2281 | distinguished from libraries by the linker according to the file | |
2282 | contents.) If linking is done, these object files are used as input | |
2283 | to the linker. | |
2284 | ||
2285 | @item -c | |
2286 | @itemx -S | |
2287 | @itemx -E | |
2288 | If any of these options is used, then the linker is not run, and | |
2289 | object file names should not be used as arguments. @xref{Overall | |
2290 | Options}. | |
2291 | ||
2292 | @cindex Libraries | |
2293 | @item -l@var{library} | |
2294 | Search the library named @var{library} when linking. | |
2295 | ||
2296 | It makes a difference where in the command you write this option; the | |
2297 | linker searches processes libraries and object files in the order they | |
2298 | are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z} | |
2299 | after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers | |
2300 | to functions in @samp{z}, those functions may not be loaded. | |
2301 | ||
2302 | The linker searches a standard list of directories for the library, | |
2303 | which is actually a file named @file{lib@var{library}.a}. The linker | |
2304 | then uses this file as if it had been specified precisely by name. | |
2305 | ||
2306 | The directories searched include several standard system directories | |
2307 | plus any that you specify with @samp{-L}. | |
2308 | ||
2309 | Normally the files found this way are library files---archive files | |
2310 | whose members are object files. The linker handles an archive file by | |
2311 | scanning through it for members which define symbols that have so far | |
2312 | been referenced but not defined. But if the file that is found is an | |
2313 | ordinary object file, it is linked in the usual fashion. The only | |
2314 | difference between using an @samp{-l} option and specifying a file name | |
2315 | is that @samp{-l} surrounds @var{library} with @samp{lib} and @samp{.a} | |
2316 | and searches several directories. | |
2317 | ||
2318 | @item -lobjc | |
2319 | You need this special case of the @samp{-l} option in order to | |
2320 | link an Objective C program. | |
2321 | ||
2322 | @item -nostartfiles | |
2323 | Do not use the standard system startup files when linking. | |
2324 | The standard system libraries are used normally, unless @code{-nostdlib} | |
2325 | or @code{-nodefaultlibs} is used. | |
2326 | ||
2327 | @item -nodefaultlibs | |
2328 | Do not use the standard system libraries when linking. | |
2329 | Only the libraries you specify will be passed to the linker. | |
2330 | The standard startup files are used normally, unless @code{-nostartfiles} | |
2331 | is used. | |
2332 | ||
2333 | @item -nostdlib | |
2334 | Do not use the standard system startup files or libraries when linking. | |
2335 | No startup files and only the libraries you specify will be passed to | |
2336 | the linker. | |
2337 | ||
2338 | @cindex @code{-lgcc}, use with @code{-nostdlib} | |
2339 | @cindex @code{-nostdlib} and unresolved references | |
2340 | @cindex unresolved references and @code{-nostdlib} | |
2341 | @cindex @code{-lgcc}, use with @code{-nodefaultlibs} | |
2342 | @cindex @code{-nodefaultlibs} and unresolved references | |
2343 | @cindex unresolved references and @code{-nodefaultlibs} | |
2344 | One of the standard libraries bypassed by @samp{-nostdlib} and | |
2345 | @samp{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines | |
2346 | that GNU CC uses to overcome shortcomings of particular machines, or special | |
2347 | needs for some languages. | |
2348 | @ifset INTERNALS | |
2349 | (@xref{Interface,,Interfacing to GNU CC Output}, for more discussion of | |
2350 | @file{libgcc.a}.) | |
2351 | @end ifset | |
2352 | @ifclear INTERNALS | |
2353 | (@xref{Interface,,Interfacing to GNU CC Output,gcc.info,Porting GNU CC}, | |
2354 | for more discussion of @file{libgcc.a}.) | |
2355 | @end ifclear | |
2356 | In most cases, you need @file{libgcc.a} even when you want to avoid | |
2357 | other standard libraries. In other words, when you specify @samp{-nostdlib} | |
2358 | or @samp{-nodefaultlibs} you should usually specify @samp{-lgcc} as well. | |
2359 | This ensures that you have no unresolved references to internal GNU CC | |
2360 | library subroutines. (For example, @samp{__main}, used to ensure C++ | |
2361 | constructors will be called; @pxref{Collect2,,@code{collect2}}.) | |
2362 | ||
2363 | @item -s | |
2364 | Remove all symbol table and relocation information from the executable. | |
2365 | ||
2366 | @item -static | |
2367 | On systems that support dynamic linking, this prevents linking with the shared | |
2368 | libraries. On other systems, this option has no effect. | |
2369 | ||
2370 | @item -shared | |
2371 | Produce a shared object which can then be linked with other objects to | |
2372 | form an executable. Not all systems support this option. You must | |
2373 | also specify @samp{-fpic} or @samp{-fPIC} on some systems when | |
2374 | you specify this option. | |
2375 | ||
2376 | @item -symbolic | |
2377 | Bind references to global symbols when building a shared object. Warn | |
2378 | about any unresolved references (unless overridden by the link editor | |
2379 | option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support | |
2380 | this option. | |
2381 | ||
2382 | @item -Xlinker @var{option} | |
2383 | Pass @var{option} as an option to the linker. You can use this to | |
2384 | supply system-specific linker options which GNU CC does not know how to | |
2385 | recognize. | |
2386 | ||
2387 | If you want to pass an option that takes an argument, you must use | |
2388 | @samp{-Xlinker} twice, once for the option and once for the argument. | |
2389 | For example, to pass @samp{-assert definitions}, you must write | |
2390 | @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write | |
2391 | @samp{-Xlinker "-assert definitions"}, because this passes the entire | |
2392 | string as a single argument, which is not what the linker expects. | |
2393 | ||
2394 | @item -Wl,@var{option} | |
2395 | Pass @var{option} as an option to the linker. If @var{option} contains | |
2396 | commas, it is split into multiple options at the commas. | |
2397 | ||
2398 | @item -u @var{symbol} | |
2399 | Pretend the symbol @var{symbol} is undefined, to force linking of | |
2400 | library modules to define it. You can use @samp{-u} multiple times with | |
2401 | different symbols to force loading of additional library modules. | |
2402 | @end table | |
2403 | ||
2404 | @node Directory Options | |
2405 | @section Options for Directory Search | |
2406 | @cindex directory options | |
2407 | @cindex options, directory search | |
2408 | @cindex search path | |
2409 | ||
2410 | These options specify directories to search for header files, for | |
2411 | libraries and for parts of the compiler: | |
2412 | ||
2413 | @table @code | |
2414 | @item -I@var{dir} | |
2415 | Add the directory @var{directory} to the head of the list of directories | |
2416 | to be searched for header files. This can be used to override a system | |
2417 | header file, substituting your own version, since these directories are | |
2418 | searched before the system header file directories. If you use more | |
2419 | than one @samp{-I} option, the directories are scanned in left-to-right | |
2420 | order; the standard system directories come after. | |
2421 | ||
2422 | @item -I- | |
2423 | Any directories you specify with @samp{-I} options before the @samp{-I-} | |
2424 | option are searched only for the case of @samp{#include "@var{file}"}; | |
2425 | they are not searched for @samp{#include <@var{file}>}. | |
2426 | ||
2427 | If additional directories are specified with @samp{-I} options after | |
2428 | the @samp{-I-}, these directories are searched for all @samp{#include} | |
2429 | directives. (Ordinarily @emph{all} @samp{-I} directories are used | |
2430 | this way.) | |
2431 | ||
2432 | In addition, the @samp{-I-} option inhibits the use of the current | |
2433 | directory (where the current input file came from) as the first search | |
2434 | directory for @samp{#include "@var{file}"}. There is no way to | |
2435 | override this effect of @samp{-I-}. With @samp{-I.} you can specify | |
2436 | searching the directory which was current when the compiler was | |
2437 | invoked. That is not exactly the same as what the preprocessor does | |
2438 | by default, but it is often satisfactory. | |
2439 | ||
2440 | @samp{-I-} does not inhibit the use of the standard system directories | |
2441 | for header files. Thus, @samp{-I-} and @samp{-nostdinc} are | |
2442 | independent. | |
2443 | ||
2444 | @item -L@var{dir} | |
2445 | Add directory @var{dir} to the list of directories to be searched | |
2446 | for @samp{-l}. | |
2447 | ||
2448 | @item -B@var{prefix} | |
2449 | This option specifies where to find the executables, libraries, | |
2450 | include files, and data files of the compiler itself. | |
2451 | ||
2452 | The compiler driver program runs one or more of the subprograms | |
2453 | @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries | |
2454 | @var{prefix} as a prefix for each program it tries to run, both with and | |
2455 | without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}). | |
2456 | ||
2457 | For each subprogram to be run, the compiler driver first tries the | |
2458 | @samp{-B} prefix, if any. If that name is not found, or if @samp{-B} | |
2459 | was not specified, the driver tries two standard prefixes, which are | |
2460 | @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of | |
2461 | those results in a file name that is found, the unmodified program | |
2462 | name is searched for using the directories specified in your | |
2463 | @samp{PATH} environment variable. | |
2464 | ||
2465 | @samp{-B} prefixes that effectively specify directory names also apply | |
2466 | to libraries in the linker, because the compiler translates these | |
2467 | options into @samp{-L} options for the linker. They also apply to | |
2468 | includes files in the preprocessor, because the compiler translates these | |
2469 | options into @samp{-isystem} options for the preprocessor. In this case, | |
2470 | the compiler appends @samp{include} to the prefix. | |
2471 | ||
2472 | The run-time support file @file{libgcc.a} can also be searched for using | |
2473 | the @samp{-B} prefix, if needed. If it is not found there, the two | |
2474 | standard prefixes above are tried, and that is all. The file is left | |
2475 | out of the link if it is not found by those means. | |
2476 | ||
2477 | Another way to specify a prefix much like the @samp{-B} prefix is to use | |
2478 | the environment variable @code{GCC_EXEC_PREFIX}. @xref{Environment | |
2479 | Variables}. | |
2480 | @end table | |
2481 | ||
2482 | @node Target Options | |
2483 | @section Specifying Target Machine and Compiler Version | |
2484 | @cindex target options | |
2485 | @cindex cross compiling | |
2486 | @cindex specifying machine version | |
2487 | @cindex specifying compiler version and target machine | |
2488 | @cindex compiler version, specifying | |
2489 | @cindex target machine, specifying | |
2490 | ||
2491 | By default, GNU CC compiles code for the same type of machine that you | |
2492 | are using. However, it can also be installed as a cross-compiler, to | |
2493 | compile for some other type of machine. In fact, several different | |
2494 | configurations of GNU CC, for different target machines, can be | |
2495 | installed side by side. Then you specify which one to use with the | |
2496 | @samp{-b} option. | |
2497 | ||
2498 | In addition, older and newer versions of GNU CC can be installed side | |
2499 | by side. One of them (probably the newest) will be the default, but | |
2500 | you may sometimes wish to use another. | |
2501 | ||
2502 | @table @code | |
2503 | @item -b @var{machine} | |
2504 | The argument @var{machine} specifies the target machine for compilation. | |
2505 | This is useful when you have installed GNU CC as a cross-compiler. | |
2506 | ||
2507 | The value to use for @var{machine} is the same as was specified as the | |
2508 | machine type when configuring GNU CC as a cross-compiler. For | |
2509 | example, if a cross-compiler was configured with @samp{configure | |
2510 | i386v}, meaning to compile for an 80386 running System V, then you | |
2511 | would specify @samp{-b i386v} to run that cross compiler. | |
2512 | ||
2513 | When you do not specify @samp{-b}, it normally means to compile for | |
2514 | the same type of machine that you are using. | |
2515 | ||
2516 | @item -V @var{version} | |
2517 | The argument @var{version} specifies which version of GNU CC to run. | |
2518 | This is useful when multiple versions are installed. For example, | |
2519 | @var{version} might be @samp{2.0}, meaning to run GNU CC version 2.0. | |
2520 | ||
2521 | The default version, when you do not specify @samp{-V}, is the last | |
2522 | version of GNU CC that you installed. | |
2523 | @end table | |
2524 | ||
2525 | The @samp{-b} and @samp{-V} options actually work by controlling part of | |
2526 | the file name used for the executable files and libraries used for | |
2527 | compilation. A given version of GNU CC, for a given target machine, is | |
2528 | normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.@refill | |
2529 | ||
2530 | Thus, sites can customize the effect of @samp{-b} or @samp{-V} either by | |
2531 | changing the names of these directories or adding alternate names (or | |
2532 | symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the | |
2533 | file @file{80386} is a link to the file @file{i386v}, then @samp{-b | |
2534 | 80386} becomes an alias for @samp{-b i386v}. | |
2535 | ||
2536 | In one respect, the @samp{-b} or @samp{-V} do not completely change | |
2537 | to a different compiler: the top-level driver program @code{gcc} | |
2538 | that you originally invoked continues to run and invoke the other | |
2539 | executables (preprocessor, compiler per se, assembler and linker) | |
2540 | that do the real work. However, since no real work is done in the | |
2541 | driver program, it usually does not matter that the driver program | |
2542 | in use is not the one for the specified target and version. | |
2543 | ||
2544 | The only way that the driver program depends on the target machine is | |
2545 | in the parsing and handling of special machine-specific options. | |
2546 | However, this is controlled by a file which is found, along with the | |
2547 | other executables, in the directory for the specified version and | |
2548 | target machine. As a result, a single installed driver program adapts | |
2549 | to any specified target machine and compiler version. | |
2550 | ||
2551 | The driver program executable does control one significant thing, | |
2552 | however: the default version and target machine. Therefore, you can | |
2553 | install different instances of the driver program, compiled for | |
2554 | different targets or versions, under different names. | |
2555 | ||
2556 | For example, if the driver for version 2.0 is installed as @code{ogcc} | |
2557 | and that for version 2.1 is installed as @code{gcc}, then the command | |
2558 | @code{gcc} will use version 2.1 by default, while @code{ogcc} will use | |
2559 | 2.0 by default. However, you can choose either version with either | |
2560 | command with the @samp{-V} option. | |
2561 | ||
2562 | @node Submodel Options | |
2563 | @section Hardware Models and Configurations | |
2564 | @cindex submodel options | |
2565 | @cindex specifying hardware config | |
2566 | @cindex hardware models and configurations, specifying | |
2567 | @cindex machine dependent options | |
2568 | ||
2569 | Earlier we discussed the standard option @samp{-b} which chooses among | |
2570 | different installed compilers for completely different target | |
2571 | machines, such as Vax vs. 68000 vs. 80386. | |
2572 | ||
2573 | In addition, each of these target machine types can have its own | |
2574 | special options, starting with @samp{-m}, to choose among various | |
2575 | hardware models or configurations---for example, 68010 vs 68020, | |
2576 | floating coprocessor or none. A single installed version of the | |
2577 | compiler can compile for any model or configuration, according to the | |
2578 | options specified. | |
2579 | ||
2580 | Some configurations of the compiler also support additional special | |
2581 | options, usually for compatibility with other compilers on the same | |
2582 | platform. | |
2583 | ||
2584 | @ifset INTERNALS | |
2585 | These options are defined by the macro @code{TARGET_SWITCHES} in the | |
2586 | machine description. The default for the options is also defined by | |
2587 | that macro, which enables you to change the defaults. | |
2588 | @end ifset | |
2589 | ||
2590 | @menu | |
2591 | * M680x0 Options:: | |
2592 | * VAX Options:: | |
2593 | * SPARC Options:: | |
2594 | * Convex Options:: | |
2595 | * AMD29K Options:: | |
2596 | * ARM Options:: | |
2597 | * M88K Options:: | |
2598 | * RS/6000 and PowerPC Options:: | |
2599 | * RT Options:: | |
2600 | * MIPS Options:: | |
2601 | * i386 Options:: | |
2602 | * HPPA Options:: | |
2603 | * Intel 960 Options:: | |
2604 | * DEC Alpha Options:: | |
2605 | * Clipper Options:: | |
2606 | * H8/300 Options:: | |
2607 | * SH Options:: | |
2608 | * System V Options:: | |
2609 | @end menu | |
2610 | ||
2611 | @node M680x0 Options | |
2612 | @subsection M680x0 Options | |
2613 | @cindex M680x0 options | |
2614 | ||
2615 | These are the @samp{-m} options defined for the 68000 series. The default | |
2616 | values for these options depends on which style of 68000 was selected when | |
2617 | the compiler was configured; the defaults for the most common choices are | |
2618 | given below. | |
2619 | ||
2620 | @table @code | |
2621 | @item -m68000 | |
2622 | @itemx -mc68000 | |
2623 | Generate output for a 68000. This is the default | |
2624 | when the compiler is configured for 68000-based systems. | |
2625 | ||
2626 | @item -m68020 | |
2627 | @itemx -mc68020 | |
2628 | Generate output for a 68020. This is the default | |
2629 | when the compiler is configured for 68020-based systems. | |
2630 | ||
2631 | @item -m68881 | |
2632 | Generate output containing 68881 instructions for floating point. | |
2633 | This is the default for most 68020 systems unless @samp{-nfp} was | |
2634 | specified when the compiler was configured. | |
2635 | ||
2636 | @item -m68030 | |
2637 | Generate output for a 68030. This is the default when the compiler is | |
2638 | configured for 68030-based systems. | |
2639 | ||
2640 | @item -m68040 | |
2641 | Generate output for a 68040. This is the default when the compiler is | |
2642 | configured for 68040-based systems. | |
2643 | ||
2644 | This option inhibits the use of 68881/68882 instructions that have to be | |
2645 | emulated by software on the 68040. If your 68040 does not have code to | |
2646 | emulate those instructions, use @samp{-m68040}. | |
2647 | ||
2648 | @item -m68060 | |
2649 | Generate output for a 68060. This is the default when the compiler is | |
2650 | configured for 68060-based systems. | |
2651 | ||
2652 | This option inhibits the use of 68020 and 68881/68882 instructions that | |
2653 | have to be emulated by software on the 68060. If your 68060 does not | |
2654 | have code to emulate those instructions, use @samp{-m68060}. | |
2655 | ||
2656 | @item -m5200 | |
2657 | Generate output for a 520X "coldfire" family cpu. This is the default | |
2658 | when the compiler is configured for 520X-based systems. | |
2659 | ||
2660 | ||
2661 | @item -m68020-40 | |
2662 | Generate output for a 68040, without using any of the new instructions. | |
2663 | This results in code which can run relatively efficiently on either a | |
2664 | 68020/68881 or a 68030 or a 68040. The generated code does use the | |
2665 | 68881 instructions that are emulated on the 68040. | |
2666 | ||
2667 | @item -mfpa | |
2668 | Generate output containing Sun FPA instructions for floating point. | |
2669 | ||
2670 | @item -msoft-float | |
2671 | Generate output containing library calls for floating point. | |
2672 | @strong{Warning:} the requisite libraries are not available for all m68k | |
2673 | targets. Normally the facilities of the machine's usual C compiler are | |
2674 | used, but this can't be done directly in cross-compilation. You must | |
2675 | make your own arrangements to provide suitable library functions for | |
2676 | cross-compilation. The embedded targets @samp{m68k-*-aout} and | |
2677 | @samp{m68k-*-coff} do provide software floating point support. | |
2678 | ||
2679 | @item -mshort | |
2680 | Consider type @code{int} to be 16 bits wide, like @code{short int}. | |
2681 | ||
2682 | @item -mnobitfield | |
2683 | Do not use the bit-field instructions. The @samp{-m68000} option | |
2684 | implies @w{@samp{-mnobitfield}}. | |
2685 | ||
2686 | @item -mbitfield | |
2687 | Do use the bit-field instructions. The @samp{-m68020} option implies | |
2688 | @samp{-mbitfield}. This is the default if you use a configuration | |
2689 | designed for a 68020. | |
2690 | ||
2691 | @item -mrtd | |
2692 | Use a different function-calling convention, in which functions | |
2693 | that take a fixed number of arguments return with the @code{rtd} | |
2694 | instruction, which pops their arguments while returning. This | |
2695 | saves one instruction in the caller since there is no need to pop | |
2696 | the arguments there. | |
2697 | ||
2698 | This calling convention is incompatible with the one normally | |
2699 | used on Unix, so you cannot use it if you need to call libraries | |
2700 | compiled with the Unix compiler. | |
2701 | ||
2702 | Also, you must provide function prototypes for all functions that | |
2703 | take variable numbers of arguments (including @code{printf}); | |
2704 | otherwise incorrect code will be generated for calls to those | |
2705 | functions. | |
2706 | ||
2707 | In addition, seriously incorrect code will result if you call a | |
2708 | function with too many arguments. (Normally, extra arguments are | |
2709 | harmlessly ignored.) | |
2710 | ||
2711 | The @code{rtd} instruction is supported by the 68010, 68020, 68030, | |
2712 | 68040, and 68060 processors, but not by the 68000 or 5200. | |
2713 | @end table | |
2714 | ||
2715 | @node VAX Options | |
2716 | @subsection VAX Options | |
2717 | @cindex VAX options | |
2718 | ||
2719 | These @samp{-m} options are defined for the Vax: | |
2720 | ||
2721 | @table @code | |
2722 | @item -munix | |
2723 | Do not output certain jump instructions (@code{aobleq} and so on) | |
2724 | that the Unix assembler for the Vax cannot handle across long | |
2725 | ranges. | |
2726 | ||
2727 | @item -mgnu | |
2728 | Do output those jump instructions, on the assumption that you | |
2729 | will assemble with the GNU assembler. | |
2730 | ||
2731 | @item -mg | |
2732 | Output code for g-format floating point numbers instead of d-format. | |
2733 | @end table | |
2734 | ||
2735 | @node SPARC Options | |
2736 | @subsection SPARC Options | |
2737 | @cindex SPARC options | |
2738 | ||
2739 | These @samp{-m} switches are supported on the SPARC: | |
2740 | ||
2741 | @table @code | |
2742 | @item -mno-app-regs | |
2743 | @itemx -mapp-regs | |
2744 | Specify @samp{-mapp-regs} to generate output using the global registers | |
2745 | 2 through 4, which the SPARC SVR4 ABI reserves for applications. This | |
2746 | is the default. | |
2747 | ||
2748 | To be fully SVR4 ABI compliant at the cost of some performance loss, | |
2749 | specify @samp{-mno-app-regs}. You should compile libraries and system | |
2750 | software with this option. | |
2751 | ||
2752 | @item -mfpu | |
2753 | @itemx -mhard-float | |
2754 | Generate output containing floating point instructions. This is the | |
2755 | default. | |
2756 | ||
2757 | @item -mno-fpu | |
2758 | @itemx -msoft-float | |
2759 | Generate output containing library calls for floating point. | |
2760 | @strong{Warning:} the requisite libraries are not available for all SPARC | |
2761 | targets. Normally the facilities of the machine's usual C compiler are | |
2762 | used, but this cannot be done directly in cross-compilation. You must make | |
2763 | your own arrangements to provide suitable library functions for | |
2764 | cross-compilation. The embedded targets @samp{sparc-*-aout} and | |
2765 | @samp{sparclite-*-*} do provide software floating point support. | |
2766 | ||
2767 | @samp{-msoft-float} changes the calling convention in the output file; | |
2768 | therefore, it is only useful if you compile @emph{all} of a program with | |
2769 | this option. In particular, you need to compile @file{libgcc.a}, the | |
2770 | library that comes with GNU CC, with @samp{-msoft-float} in order for | |
2771 | this to work. | |
2772 | ||
2773 | @item -mhard-quad-float | |
2774 | Generate output containing quad-word (long double) floating point | |
2775 | instructions. | |
2776 | ||
2777 | @item -msoft-quad-float | |
2778 | Generate output containing library calls for quad-word (long double) | |
2779 | floating point instructions. The functions called are those specified | |
2780 | in the SPARC ABI. This is the default. | |
2781 | ||
2782 | As of this writing, there are no sparc implementations that have hardware | |
2783 | support for the quad-word floating point instructions. They all invoke | |
2784 | a trap handler for one of these instructions, and then the trap handler | |
2785 | emulates the effect of the instruction. Because of the trap handler overhead, | |
2786 | this is much slower than calling the ABI library routines. Thus the | |
2787 | @samp{-msoft-quad-float} option is the default. | |
2788 | ||
2789 | @item -mno-epilogue | |
2790 | @itemx -mepilogue | |
2791 | With @samp{-mepilogue} (the default), the compiler always emits code for | |
2792 | function exit at the end of each function. Any function exit in | |
2793 | the middle of the function (such as a return statement in C) will | |
2794 | generate a jump to the exit code at the end of the function. | |
2795 | ||
2796 | With @samp{-mno-epilogue}, the compiler tries to emit exit code inline | |
2797 | at every function exit. | |
2798 | ||
2799 | @item -mno-flat | |
2800 | @itemx -mflat | |
2801 | With @samp{-mflat}, the compiler does not generate save/restore instructions | |
2802 | and will use a "flat" or single register window calling convention. | |
2803 | This model uses %i7 as the frame pointer and is compatible with the normal | |
2804 | register window model. Code from either may be intermixed. | |
2805 | The local registers and the input registers (0-5) are still treated as | |
2806 | "call saved" registers and will be saved on the stack as necessary. | |
2807 | ||
2808 | With @samp{-mno-flat} (the default), the compiler emits save/restore | |
2809 | instructions (except for leaf functions) and is the normal mode of operation. | |
2810 | ||
2811 | @item -mno-unaligned-doubles | |
2812 | @itemx -munaligned-doubles | |
2813 | Assume that doubles have 8 byte alignment. This is the default. | |
2814 | ||
2815 | With @samp{-munaligned-doubles}, GNU CC assumes that doubles have 8 byte | |
2816 | alignment only if they are contained in another type, or if they have an | |
2817 | absolute address. Otherwise, it assumes they have 4 byte alignment. | |
2818 | Specifying this option avoids some rare compatibility problems with code | |
2819 | generated by other compilers. It is not the default because it results | |
2820 | in a performance loss, especially for floating point code. | |
2821 | ||
2822 | @item -mv8 | |
2823 | @itemx -msparclite | |
2824 | These two options select variations on the SPARC architecture. | |
2825 | ||
2826 | By default (unless specifically configured for the Fujitsu SPARClite), | |
2827 | GCC generates code for the v7 variant of the SPARC architecture. | |
2828 | ||
2829 | @samp{-mv8} will give you SPARC v8 code. The only difference from v7 | |
2830 | code is that the compiler emits the integer multiply and integer | |
2831 | divide instructions which exist in SPARC v8 but not in SPARC v7. | |
2832 | ||
2833 | @samp{-msparclite} will give you SPARClite code. This adds the integer | |
2834 | multiply, integer divide step and scan (@code{ffs}) instructions which | |
2835 | exist in SPARClite but not in SPARC v7. | |
2836 | ||
2837 | These options are deprecated and will be deleted in GNU CC 2.9. | |
2838 | They have been replaced with @samp{-mcpu=xxx}. | |
2839 | ||
2840 | @item -mcypress | |
2841 | @itemx -msupersparc | |
2842 | These two options select the processor for which the code is optimised. | |
2843 | ||
2844 | With @samp{-mcypress} (the default), the compiler optimizes code for the | |
2845 | Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series. | |
2846 | This is also appropriate for the older SparcStation 1, 2, IPX etc. | |
2847 | ||
2848 | With @samp{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as | |
2849 | used in the SparcStation 10, 1000 and 2000 series. This flag also enables use | |
2850 | of the full SPARC v8 instruction set. | |
2851 | ||
2852 | These options are deprecated and will be deleted in GNU CC 2.9. | |
2853 | They have been replaced with @samp{-mcpu=xxx}. | |
2854 | ||
2855 | @item -mcpu=@var{cpu_type} | |
2856 | Set architecture type and instruction scheduling parameters for machine | |
2857 | type @var{cpu_type}. Supported values for @var{cpu_type} are | |
2858 | @samp{common}, @samp{cypress}, @samp{v8}, @samp{supersparc}, | |
2859 | @samp{sparclite}, @samp{f930}, @samp{f934}, | |
2860 | @samp{sparclet}, @samp{90c701}, @samp{v8plus}, @samp{v9}, | |
2861 | and @samp{ultrasparc}. Specifying @samp{v9} is only supported on true | |
2862 | 64 bit targets. | |
2863 | ||
2864 | @item -mtune=@var{cpu_type} | |
2865 | Set the instruction scheduling parameters for machine type | |
2866 | @var{cpu_type}, but do not set the architecture type as the option | |
2867 | @samp{-mcpu=}@var{cpu_type} would. The same values for | |
2868 | @samp{-mcpu=}@var{cpu_type} are used for @samp{-tune=}@var{cpu_type}. | |
2869 | ||
2870 | @end table | |
2871 | ||
2872 | These @samp{-m} switches are supported in addition to the above | |
2873 | on the SPARCLET processor. | |
2874 | ||
2875 | @table @code | |
2876 | @item -mlittle-endian | |
2877 | Generate code for a processor running in little-endian mode. | |
2878 | ||
2879 | @item -mlive-g0 | |
2880 | Treat register @code{%g0} as a normal register. | |
2881 | GCC will continue to clobber it as necessary but will not assume | |
2882 | it always reads as 0. | |
2883 | ||
2884 | @item -mbroken-saverestore | |
2885 | Generate code that does not use non-trivial forms of the @code{save} and | |
2886 | @code{restore} instructions. Early versions of the SPARCLET processor do | |
2887 | not correctly handle @code{save} and @code{restore} instructions used with | |
2888 | arguments. They correctly handle them used without arguments. A @code{save} | |
2889 | instruction used without arguments increments the current window pointer | |
2890 | but does not allocate a new stack frame. It is assumed that the window | |
2891 | overflow trap handler will properly handle this case as will interrupt | |
2892 | handlers. | |
2893 | @end table | |
2894 | ||
2895 | These @samp{-m} switches are supported in addition to the above | |
2896 | on SPARC V9 processors in 64 bit environments. | |
2897 | ||
2898 | @table @code | |
2899 | @item -mlittle-endian | |
2900 | Generate code for a processor running in little-endian mode. | |
2901 | ||
2902 | @item -mmedlow | |
2903 | Generate code for the Medium/Low code model: assume a 32 bit address space. | |
2904 | Programs are statically linked, PIC is not supported. Pointers are still | |
2905 | 64 bits. | |
2906 | ||
2907 | It is very likely that a future version of GCC will rename this option. | |
2908 | ||
2909 | @item -mmedany | |
2910 | Generate code for the Medium/Anywhere code model: assume a 32 bit text | |
2911 | and a 32 bit data segment, both starting anywhere (determined at link time). | |
2912 | Programs are statically linked, PIC is not supported. Pointers are still | |
2913 | 64 bits. | |
2914 | ||
2915 | It is very likely that a future version of GCC will rename this option. | |
2916 | ||
2917 | @item -mfullany | |
2918 | Generate code for the Full/Anywhere code model: assume a full 64 bit | |
2919 | address space. PIC is not supported. | |
2920 | ||
2921 | It is very likely that a future version of GCC will rename this option. | |
2922 | ||
2923 | @item -mint64 | |
2924 | Types long and int are 64 bits. | |
2925 | ||
2926 | @item -mlong32 | |
2927 | Types long and int are 32 bits. | |
2928 | ||
2929 | @item -mlong64 | |
2930 | @itemx -mint32 | |
2931 | Type long is 64 bits, and type int is 32 bits. | |
2932 | ||
2933 | @item -mstack-bias | |
2934 | @itemx -mno-stack-bias | |
2935 | With @samp{-mstack-bias}, GNU CC assumes that the stack pointer, and | |
2936 | frame pointer if present, are offset by -2047 which must be added back | |
2937 | when making stack frame references. | |
2938 | Otherwise, assume no such offset is present. | |
2939 | @end table | |
2940 | ||
2941 | @node Convex Options | |
2942 | @subsection Convex Options | |
2943 | @cindex Convex options | |
2944 | ||
2945 | These @samp{-m} options are defined for Convex: | |
2946 | ||
2947 | @table @code | |
2948 | @item -mc1 | |
2949 | Generate output for C1. The code will run on any Convex machine. | |
2950 | The preprocessor symbol @code{__convex__c1__} is defined. | |
2951 | ||
2952 | @item -mc2 | |
2953 | Generate output for C2. Uses instructions not available on C1. | |
2954 | Scheduling and other optimizations are chosen for max performance on C2. | |
2955 | The preprocessor symbol @code{__convex_c2__} is defined. | |
2956 | ||
2957 | @item -mc32 | |
2958 | Generate output for C32xx. Uses instructions not available on C1. | |
2959 | Scheduling and other optimizations are chosen for max performance on C32. | |
2960 | The preprocessor symbol @code{__convex_c32__} is defined. | |
2961 | ||
2962 | @item -mc34 | |
2963 | Generate output for C34xx. Uses instructions not available on C1. | |
2964 | Scheduling and other optimizations are chosen for max performance on C34. | |
2965 | The preprocessor symbol @code{__convex_c34__} is defined. | |
2966 | ||
2967 | @item -mc38 | |
2968 | Generate output for C38xx. Uses instructions not available on C1. | |
2969 | Scheduling and other optimizations are chosen for max performance on C38. | |
2970 | The preprocessor symbol @code{__convex_c38__} is defined. | |
2971 | ||
2972 | @item -margcount | |
2973 | Generate code which puts an argument count in the word preceding each | |
2974 | argument list. This is compatible with regular CC, and a few programs | |
2975 | may need the argument count word. GDB and other source-level debuggers | |
2976 | do not need it; this info is in the symbol table. | |
2977 | ||
2978 | @item -mnoargcount | |
2979 | Omit the argument count word. This is the default. | |
2980 | ||
2981 | @item -mvolatile-cache | |
2982 | Allow volatile references to be cached. This is the default. | |
2983 | ||
2984 | @item -mvolatile-nocache | |
2985 | Volatile references bypass the data cache, going all the way to memory. | |
2986 | This is only needed for multi-processor code that does not use standard | |
2987 | synchronization instructions. Making non-volatile references to volatile | |
2988 | locations will not necessarily work. | |
2989 | ||
2990 | @item -mlong32 | |
2991 | Type long is 32 bits, the same as type int. This is the default. | |
2992 | ||
2993 | @item -mlong64 | |
2994 | Type long is 64 bits, the same as type long long. This option is useless, | |
2995 | because no library support exists for it. | |
2996 | @end table | |
2997 | ||
2998 | @node AMD29K Options | |
2999 | @subsection AMD29K Options | |
3000 | @cindex AMD29K options | |
3001 | ||
3002 | These @samp{-m} options are defined for the AMD Am29000: | |
3003 | ||
3004 | @table @code | |
3005 | @item -mdw | |
3006 | @kindex -mdw | |
3007 | @cindex DW bit (29k) | |
3008 | Generate code that assumes the @code{DW} bit is set, i.e., that byte and | |
3009 | halfword operations are directly supported by the hardware. This is the | |
3010 | default. | |
3011 | ||
3012 | @item -mndw | |
3013 | @kindex -mndw | |
3014 | Generate code that assumes the @code{DW} bit is not set. | |
3015 | ||
3016 | @item -mbw | |
3017 | @kindex -mbw | |
3018 | @cindex byte writes (29k) | |
3019 | Generate code that assumes the system supports byte and halfword write | |
3020 | operations. This is the default. | |
3021 | ||
3022 | @item -mnbw | |
3023 | @kindex -mnbw | |
3024 | Generate code that assumes the systems does not support byte and | |
3025 | halfword write operations. @samp{-mnbw} implies @samp{-mndw}. | |
3026 | ||
3027 | @item -msmall | |
3028 | @kindex -msmall | |
3029 | @cindex memory model (29k) | |
3030 | Use a small memory model that assumes that all function addresses are | |
3031 | either within a single 256 KB segment or at an absolute address of less | |
3032 | than 256k. This allows the @code{call} instruction to be used instead | |
3033 | of a @code{const}, @code{consth}, @code{calli} sequence. | |
3034 | ||
3035 | @item -mnormal | |
3036 | @kindex -mnormal | |
3037 | Use the normal memory model: Generate @code{call} instructions only when | |
3038 | calling functions in the same file and @code{calli} instructions | |
3039 | otherwise. This works if each file occupies less than 256 KB but allows | |
3040 | the entire executable to be larger than 256 KB. This is the default. | |
3041 | ||
3042 | @item -mlarge | |
3043 | Always use @code{calli} instructions. Specify this option if you expect | |
3044 | a single file to compile into more than 256 KB of code. | |
3045 | ||
3046 | @item -m29050 | |
3047 | @kindex -m29050 | |
3048 | @cindex processor selection (29k) | |
3049 | Generate code for the Am29050. | |
3050 | ||
3051 | @item -m29000 | |
3052 | @kindex -m29000 | |
3053 | Generate code for the Am29000. This is the default. | |
3054 | ||
3055 | @item -mkernel-registers | |
3056 | @kindex -mkernel-registers | |
3057 | @cindex kernel and user registers (29k) | |
3058 | Generate references to registers @code{gr64-gr95} instead of to | |
3059 | registers @code{gr96-gr127}. This option can be used when compiling | |
3060 | kernel code that wants a set of global registers disjoint from that used | |
3061 | by user-mode code. | |
3062 | ||
3063 | Note that when this option is used, register names in @samp{-f} flags | |
3064 | must use the normal, user-mode, names. | |
3065 | ||
3066 | @item -muser-registers | |
3067 | @kindex -muser-registers | |
3068 | Use the normal set of global registers, @code{gr96-gr127}. This is the | |
3069 | default. | |
3070 | ||
3071 | @item -mstack-check | |
3072 | @itemx -mno-stack-check | |
3073 | @kindex -mstack-check | |
3074 | @cindex stack checks (29k) | |
3075 | Insert (or do not insert) a call to @code{__msp_check} after each stack | |
3076 | adjustment. This is often used for kernel code. | |
3077 | ||
3078 | @item -mstorem-bug | |
3079 | @itemx -mno-storem-bug | |
3080 | @kindex -mstorem-bug | |
3081 | @cindex storem bug (29k) | |
3082 | @samp{-mstorem-bug} handles 29k processors which cannot handle the | |
3083 | separation of a mtsrim insn and a storem instruction (most 29000 chips | |
3084 | to date, but not the 29050). | |
3085 | ||
3086 | @item -mno-reuse-arg-regs | |
3087 | @itemx -mreuse-arg-regs | |
3088 | @kindex -mreuse-arg-regs | |
3089 | @samp{-mno-reuse-arg-regs} tells the compiler to only use incoming argument | |
3090 | registers for copying out arguments. This helps detect calling a function | |
3091 | with fewer arguments than it was declared with. | |
3092 | ||
3093 | @item -msoft-float | |
3094 | @kindex -msoft-float | |
3095 | Generate output containing library calls for floating point. | |
3096 | @strong{Warning:} the requisite libraries are not part of GNU CC. | |
3097 | Normally the facilities of the machine's usual C compiler are used, but | |
3098 | this can't be done directly in cross-compilation. You must make your | |
3099 | own arrangements to provide suitable library functions for | |
3100 | cross-compilation. | |
3101 | @end table | |
3102 | ||
3103 | @node ARM Options | |
3104 | @subsection ARM Options | |
3105 | @cindex ARM options | |
3106 | ||
3107 | These @samp{-m} options are defined for Advanced RISC Machines (ARM) | |
3108 | architectures: | |
3109 | ||
3110 | @table @code | |
3111 | @item -mapcs-frame | |
3112 | @kindex -mapcs-frame | |
3113 | Generate a stack frame that is compliant with the ARM Procedure Call | |
3114 | Standard for all functions, even if this is not strictly necessary for | |
3115 | correct execution of the code. | |
3116 | ||
3117 | @item -mapcs-26 | |
3118 | @kindex -mapcs-26 | |
3119 | Generate code for a processor running with a 26-bit program counter, | |
3120 | and conforming to the function calling standards for the APCS 26-bit | |
3121 | option. This option replaces the @samp{-m2} and @samp{-m3} options | |
3122 | of previous releases of the compiler. | |
3123 | ||
3124 | @item -mapcs-32 | |
3125 | @kindex -mapcs-32 | |
3126 | Generate code for a processor running with a 32-bit program counter, | |
3127 | and conforming to the function calling standards for the APCS 32-bit | |
3128 | option. This option replaces the @samp{-m6} option of previous releases | |
3129 | of the compiler. | |
3130 | ||
3131 | @item -mhard-float | |
3132 | Generate output containing floating point instructions. This is the | |
3133 | default. | |
3134 | ||
3135 | @item -msoft-float | |
3136 | Generate output containing library calls for floating point. | |
3137 | @strong{Warning:} the requisite libraries are not available for all ARM | |
3138 | targets. Normally the facilities of the machine's usual C compiler are | |
3139 | used, but this cannot be done directly in cross-compilation. You must make | |
3140 | your own arrangements to provide suitable library functions for | |
3141 | cross-compilation. | |
3142 | ||
3143 | @samp{-msoft-float} changes the calling convention in the output file; | |
3144 | therefore, it is only useful if you compile @emph{all} of a program with | |
3145 | this option. In particular, you need to compile @file{libgcc.a}, the | |
3146 | library that comes with GNU CC, with @samp{-msoft-float} in order for | |
3147 | this to work. | |
3148 | ||
3149 | @item -mlittle-endian | |
3150 | Generate code for a processor running in little-endian mode. This is | |
3151 | the default for all standard configurations. | |
3152 | ||
3153 | @item -mbig-endian | |
3154 | Generate code for a processor running in big-endian mode; the default is | |
3155 | to compile code for a little-endian processor. | |
3156 | ||
3157 | @item -mwords-little-endian | |
3158 | This option only applies when generating code for big-endian processors. | |
3159 | Generate code for a little-endian word order but a big-endian byte | |
3160 | order. That is, a byte order of the form @samp{32107654}. Note: this | |
3161 | option should only be used if you require compatibility with code for | |
3162 | big-endian ARM processors generated by versions of the compiler prior to | |
3163 | 2.8. | |
3164 | ||
3165 | @item -mshort-load-bytes | |
3166 | @kindex -mshort-load-bytes | |
3167 | Do not try to load half-words (eg @samp{short}s) by loading a word from | |
3168 | an unaligned address. For some targets the MMU is configured to trap | |
3169 | unaligned loads; use this option to generate code that is safe in these | |
3170 | environments. | |
3171 | ||
3172 | @item -mno-short-load-bytes | |
3173 | @kindex -mno-short-load-bytes | |
3174 | Use unaligned word loads to load half-words (eg @samp{short}s). This | |
3175 | option produces more efficient code, but the MMU is sometimes configured | |
3176 | to trap these instructions. | |
3177 | ||
3178 | @item -mbsd | |
3179 | @kindex -mbsd | |
3180 | This option only applies to RISC iX. Emulate the native BSD-mode | |
3181 | compiler. This is the default if @samp{-ansi} is not specified. | |
3182 | ||
3183 | @item -mxopen | |
3184 | @kindex -mxopen | |
3185 | This option only applies to RISC iX. Emulate the native X/Open-mode | |
3186 | compiler. | |
3187 | ||
3188 | @item -mno-symrename | |
3189 | @kindex -mno-symrename | |
3190 | This option only applies to RISC iX. Do not run the assembler | |
3191 | post-processor, @samp{symrename}, after code has been assembled. | |
3192 | Normally it is necessary to modify some of the standard symbols in | |
3193 | preparation for linking with the RISC iX C library; this option | |
3194 | suppresses this pass. The post-processor is never run when the | |
3195 | compiler is built for cross-compilation. | |
3196 | @end table | |
3197 | ||
3198 | @node M88K Options | |
3199 | @subsection M88K Options | |
3200 | @cindex M88k options | |
3201 | ||
3202 | These @samp{-m} options are defined for Motorola 88k architectures: | |
3203 | ||
3204 | @table @code | |
3205 | @item -m88000 | |
3206 | @kindex -m88000 | |
3207 | Generate code that works well on both the m88100 and the | |
3208 | m88110. | |
3209 | ||
3210 | @item -m88100 | |
3211 | @kindex -m88100 | |
3212 | Generate code that works best for the m88100, but that also | |
3213 | runs on the m88110. | |
3214 | ||
3215 | @item -m88110 | |
3216 | @kindex -m88110 | |
3217 | Generate code that works best for the m88110, and may not run | |
3218 | on the m88100. | |
3219 | ||
3220 | @item -mbig-pic | |
3221 | @kindex -mbig-pic | |
3222 | Obsolete option to be removed from the next revision. | |
3223 | Use @samp{-fPIC}. | |
3224 | ||
3225 | @item -midentify-revision | |
3226 | @kindex -midentify-revision | |
3227 | @kindex ident | |
3228 | @cindex identifying source, compiler (88k) | |
3229 | Include an @code{ident} directive in the assembler output recording the | |
3230 | source file name, compiler name and version, timestamp, and compilation | |
3231 | flags used. | |
3232 | ||
3233 | @item -mno-underscores | |
3234 | @kindex -mno-underscores | |
3235 | @cindex underscores, avoiding (88k) | |
3236 | In assembler output, emit symbol names without adding an underscore | |
3237 | character at the beginning of each name. The default is to use an | |
3238 | underscore as prefix on each name. | |
3239 | ||
3240 | @item -mocs-debug-info | |
3241 | @itemx -mno-ocs-debug-info | |
3242 | @kindex -mocs-debug-info | |
3243 | @kindex -mno-ocs-debug-info | |
3244 | @cindex OCS (88k) | |
3245 | @cindex debugging, 88k OCS | |
3246 | Include (or omit) additional debugging information (about registers used | |
3247 | in each stack frame) as specified in the 88open Object Compatibility | |
3248 | Standard, ``OCS''. This extra information allows debugging of code that | |
3249 | has had the frame pointer eliminated. The default for DG/UX, SVr4, and | |
3250 | Delta 88 SVr3.2 is to include this information; other 88k configurations | |
3251 | omit this information by default. | |
3252 | ||
3253 | @item -mocs-frame-position | |
3254 | @kindex -mocs-frame-position | |
3255 | @cindex register positions in frame (88k) | |
3256 | When emitting COFF debugging information for automatic variables and | |
3257 | parameters stored on the stack, use the offset from the canonical frame | |
3258 | address, which is the stack pointer (register 31) on entry to the | |
3259 | function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use | |
3260 | @samp{-mocs-frame-position}; other 88k configurations have the default | |
3261 | @samp{-mno-ocs-frame-position}. | |
3262 | ||
3263 | @item -mno-ocs-frame-position | |
3264 | @kindex -mno-ocs-frame-position | |
3265 | @cindex register positions in frame (88k) | |
3266 | When emitting COFF debugging information for automatic variables and | |
3267 | parameters stored on the stack, use the offset from the frame pointer | |
3268 | register (register 30). When this option is in effect, the frame | |
3269 | pointer is not eliminated when debugging information is selected by the | |
3270 | -g switch. | |
3271 | ||
3272 | @item -moptimize-arg-area | |
3273 | @itemx -mno-optimize-arg-area | |
3274 | @kindex -moptimize-arg-area | |
3275 | @kindex -mno-optimize-arg-area | |
3276 | @cindex arguments in frame (88k) | |
3277 | Control how function arguments are stored in stack frames. | |
3278 | @samp{-moptimize-arg-area} saves space by optimizing them, but this | |
3279 | conflicts with the 88open specifications. The opposite alternative, | |
3280 | @samp{-mno-optimize-arg-area}, agrees with 88open standards. By default | |
3281 | GNU CC does not optimize the argument area. | |
3282 | ||
3283 | @item -mshort-data-@var{num} | |
3284 | @kindex -mshort-data-@var{num} | |
3285 | @cindex smaller data references (88k) | |
3286 | @cindex r0-relative references (88k) | |
3287 | Generate smaller data references by making them relative to @code{r0}, | |
3288 | which allows loading a value using a single instruction (rather than the | |
3289 | usual two). You control which data references are affected by | |
3290 | specifying @var{num} with this option. For example, if you specify | |
3291 | @samp{-mshort-data-512}, then the data references affected are those | |
3292 | involving displacements of less than 512 bytes. | |
3293 | @samp{-mshort-data-@var{num}} is not effective for @var{num} greater | |
3294 | than 64k. | |
3295 | ||
3296 | @item -mserialize-volatile | |
3297 | @kindex -mserialize-volatile | |
3298 | @itemx -mno-serialize-volatile | |
3299 | @kindex -mno-serialize-volatile | |
3300 | @cindex sequential consistency on 88k | |
3301 | Do, or don't, generate code to guarantee sequential consistency | |
3302 | of volatile memory references. By default, consistency is | |
3303 | guaranteed. | |
3304 | ||
3305 | The order of memory references made by the MC88110 processor does | |
3306 | not always match the order of the instructions requesting those | |
3307 | references. In particular, a load instruction may execute before | |
3308 | a preceding store instruction. Such reordering violates | |
3309 | sequential consistency of volatile memory references, when there | |
3310 | are multiple processors. When consistency must be guaranteed, | |
3311 | GNU C generates special instructions, as needed, to force | |
3312 | execution in the proper order. | |
3313 | ||
3314 | The MC88100 processor does not reorder memory references and so | |
3315 | always provides sequential consistency. However, by default, GNU | |
3316 | C generates the special instructions to guarantee consistency | |
3317 | even when you use @samp{-m88100}, so that the code may be run on an | |
3318 | MC88110 processor. If you intend to run your code only on the | |
3319 | MC88100 processor, you may use @samp{-mno-serialize-volatile}. | |
3320 | ||
3321 | The extra code generated to guarantee consistency may affect the | |
3322 | performance of your application. If you know that you can safely | |
3323 | forgo this guarantee, you may use @samp{-mno-serialize-volatile}. | |
3324 | ||
3325 | @item -msvr4 | |
3326 | @itemx -msvr3 | |
3327 | @kindex -msvr4 | |
3328 | @kindex -msvr3 | |
3329 | @cindex assembler syntax, 88k | |
3330 | @cindex SVr4 | |
3331 | Turn on (@samp{-msvr4}) or off (@samp{-msvr3}) compiler extensions | |
3332 | related to System V release 4 (SVr4). This controls the following: | |
3333 | ||
3334 | @enumerate | |
3335 | @item | |
3336 | Which variant of the assembler syntax to emit. | |
3337 | @item | |
3338 | @samp{-msvr4} makes the C preprocessor recognize @samp{#pragma weak} | |
3339 | that is used on System V release 4. | |
3340 | @item | |
3341 | @samp{-msvr4} makes GNU CC issue additional declaration directives used in | |
3342 | SVr4. | |
3343 | @end enumerate | |
3344 | ||
3345 | @samp{-msvr4} is the default for the m88k-motorola-sysv4 and | |
3346 | m88k-dg-dgux m88k configurations. @samp{-msvr3} is the default for all | |
3347 | other m88k configurations. | |
3348 | ||
3349 | @item -mversion-03.00 | |
3350 | @kindex -mversion-03.00 | |
3351 | This option is obsolete, and is ignored. | |
3352 | @c ??? which asm syntax better for GAS? option there too? | |
3353 | ||
3354 | @item -mno-check-zero-division | |
3355 | @itemx -mcheck-zero-division | |
3356 | @kindex -mno-check-zero-division | |
3357 | @kindex -mcheck-zero-division | |
3358 | @cindex zero division on 88k | |
3359 | Do, or don't, generate code to guarantee that integer division by | |
3360 | zero will be detected. By default, detection is guaranteed. | |
3361 | ||
3362 | Some models of the MC88100 processor fail to trap upon integer | |
3363 | division by zero under certain conditions. By default, when | |
3364 | compiling code that might be run on such a processor, GNU C | |
3365 | generates code that explicitly checks for zero-valued divisors | |
3366 | and traps with exception number 503 when one is detected. Use of | |
3367 | mno-check-zero-division suppresses such checking for code | |
3368 | generated to run on an MC88100 processor. | |
3369 | ||
3370 | GNU C assumes that the MC88110 processor correctly detects all | |
3371 | instances of integer division by zero. When @samp{-m88110} is | |
3372 | specified, both @samp{-mcheck-zero-division} and | |
3373 | @samp{-mno-check-zero-division} are ignored, and no explicit checks for | |
3374 | zero-valued divisors are generated. | |
3375 | ||
3376 | @item -muse-div-instruction | |
3377 | @kindex -muse-div-instruction | |
3378 | @cindex divide instruction, 88k | |
3379 | Use the div instruction for signed integer division on the | |
3380 | MC88100 processor. By default, the div instruction is not used. | |
3381 | ||
3382 | On the MC88100 processor the signed integer division instruction | |
3383 | div) traps to the operating system on a negative operand. The | |
3384 | operating system transparently completes the operation, but at a | |
3385 | large cost in execution time. By default, when compiling code | |
3386 | that might be run on an MC88100 processor, GNU C emulates signed | |
3387 | integer division using the unsigned integer division instruction | |
3388 | divu), thereby avoiding the large penalty of a trap to the | |
3389 | operating system. Such emulation has its own, smaller, execution | |
3390 | cost in both time and space. To the extent that your code's | |
3391 | important signed integer division operations are performed on two | |
3392 | nonnegative operands, it may be desirable to use the div | |
3393 | instruction directly. | |
3394 | ||
3395 | On the MC88110 processor the div instruction (also known as the | |
3396 | divs instruction) processes negative operands without trapping to | |
3397 | the operating system. When @samp{-m88110} is specified, | |
3398 | @samp{-muse-div-instruction} is ignored, and the div instruction is used | |
3399 | for signed integer division. | |
3400 | ||
3401 | Note that the result of dividing INT_MIN by -1 is undefined. In | |
3402 | particular, the behavior of such a division with and without | |
3403 | @samp{-muse-div-instruction} may differ. | |
3404 | ||
3405 | @item -mtrap-large-shift | |
3406 | @itemx -mhandle-large-shift | |
3407 | @kindex -mtrap-large-shift | |
3408 | @kindex -mhandle-large-shift | |
3409 | @cindex bit shift overflow (88k) | |
3410 | @cindex large bit shifts (88k) | |
3411 | Include code to detect bit-shifts of more than 31 bits; respectively, | |
3412 | trap such shifts or emit code to handle them properly. By default GNU CC | |
3413 | makes no special provision for large bit shifts. | |
3414 | ||
3415 | @item -mwarn-passed-structs | |
3416 | @kindex -mwarn-passed-structs | |
3417 | @cindex structure passing (88k) | |
3418 | Warn when a function passes a struct as an argument or result. | |
3419 | Structure-passing conventions have changed during the evolution of the C | |
3420 | language, and are often the source of portability problems. By default, | |
3421 | GNU CC issues no such warning. | |
3422 | @end table | |
3423 | ||
3424 | @node RS/6000 and PowerPC Options | |
3425 | @subsection IBM RS/6000 and PowerPC Options | |
3426 | @cindex RS/6000 and PowerPC Options | |
3427 | @cindex IBM RS/6000 and PowerPC Options | |
3428 | ||
3429 | These @samp{-m} options are defined for the IBM RS/6000 and PowerPC: | |
3430 | @table @code | |
3431 | @item -mpower | |
3432 | @itemx -mno-power | |
3433 | @itemx -mpower2 | |
3434 | @itemx -mno-power2 | |
3435 | @itemx -mpowerpc | |
3436 | @itemx -mno-powerpc | |
3437 | @itemx -mpowerpc-gpopt | |
3438 | @itemx -mno-powerpc-gpopt | |
3439 | @itemx -mpowerpc-gfxopt | |
3440 | @itemx -mno-powerpc-gfxopt | |
3441 | @kindex -mpower | |
3442 | @kindex -mpower2 | |
3443 | @kindex -mpowerpc | |
3444 | @kindex -mpowerpc-gpopt | |
3445 | @kindex -mpowerpc-gfxopt | |
3446 | GNU CC supports two related instruction set architectures for the | |
3447 | RS/6000 and PowerPC. The @dfn{POWER} instruction set are those | |
3448 | instructions supported by the @samp{rios} chip set used in the original | |
3449 | RS/6000 systems and the @dfn{PowerPC} instruction set is the | |
3450 | architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and | |
3451 | the IBM 40x microprocessors. The PowerPC architecture defines 64-bit | |
3452 | instructions, but they are not supported by any current processors. | |
3453 | ||
3454 | Neither architecture is a subset of the other. However there is a | |
3455 | large common subset of instructions supported by both. An MQ | |
3456 | register is included in processors supporting the POWER architecture. | |
3457 | ||
3458 | You use these options to specify which instructions are available on the | |
3459 | processor you are using. The default value of these options is | |
3460 | determined when configuring GNU CC. Specifying the | |
3461 | @samp{-mcpu=@var{cpu_type}} overrides the specification of these | |
3462 | options. We recommend you use the @samp{-mcpu=@var{cpu_type}} option | |
3463 | rather than the options listed above. | |
3464 | ||
3465 | The @samp{-mpower} option allows GNU CC to generate instructions that | |
3466 | are found only in the POWER architecture and to use the MQ register. | |
3467 | Specifying @samp{-mpower2} implies @samp{-power} and also allows GNU CC | |
3468 | to generate instructions that are present in the POWER2 architecture but | |
3469 | not the original POWER architecture. | |
3470 | ||
3471 | The @samp{-mpowerpc} option allows GNU CC to generate instructions that | |
3472 | are found only in the 32-bit subset of the PowerPC architecture. | |
3473 | Specifying @samp{-mpowerpc-gpopt} implies @samp{-mpowerpc} and also allows | |
3474 | GNU CC to use the optional PowerPC architecture instructions in the | |
3475 | General Purpose group, including floating-point square root. Specifying | |
3476 | @samp{-mpowerpc-gfxopt} implies @samp{-mpowerpc} and also allows GNU CC to | |
3477 | use the optional PowerPC architecture instructions in the Graphics | |
3478 | group, including floating-point select. | |
3479 | ||
3480 | If you specify both @samp{-mno-power} and @samp{-mno-powerpc}, GNU CC | |
3481 | will use only the instructions in the common subset of both | |
3482 | architectures plus some special AIX common-mode calls, and will not use | |
3483 | the MQ register. Specifying both @samp{-mpower} and @samp{-mpowerpc} | |
3484 | permits GNU CC to use any instruction from either architecture and to | |
3485 | allow use of the MQ register; specify this for the Motorola MPC601. | |
3486 | ||
3487 | @item -mnew-mnemonics | |
3488 | @itemx -mold-mnemonics | |
3489 | @kindex -mnew-mnemonics | |
3490 | @kindex -mold-mnemonics | |
3491 | Select which mnemonics to use in the generated assembler code. | |
3492 | @samp{-mnew-mnemonics} requests output that uses the assembler mnemonics | |
3493 | defined for the PowerPC architecture, while @samp{-mold-mnemonics} | |
3494 | requests the assembler mnemonics defined for the POWER architecture. | |
3495 | Instructions defined in only one architecture have only one mnemonic; | |
3496 | GNU CC uses that mnemonic irrespective of which of these options is | |
3497 | specified. | |
3498 | ||
3499 | PowerPC assemblers support both the old and new mnemonics, as will later | |
3500 | POWER assemblers. Current POWER assemblers only support the old | |
3501 | mnemonics. Specify @samp{-mnew-mnemonics} if you have an assembler that | |
3502 | supports them, otherwise specify @samp{-mold-mnemonics}. | |
3503 | ||
3504 | The default value of these options depends on how GNU CC was configured. | |
3505 | Specifying @samp{-mcpu=@var{cpu_type}} sometimes overrides the value of | |
3506 | these option. Unless you are building a cross-compiler, you should | |
3507 | normally not specify either @samp{-mnew-mnemonics} or | |
3508 | @samp{-mold-mnemonics}, but should instead accept the default. | |
3509 | ||
3510 | @item -mcpu=@var{cpu_type} | |
3511 | Set architecture type, register usage, choice of mnemonics, and | |
3512 | instruction scheduling parameters for machine type @var{cpu_type}. | |
3513 | Supported values for @var{cpu_type} are @samp{rs6000}, @samp{rios1}, | |
3514 | @samp{rios2}, @samp{rsc}, @samp{601}, @samp{602}, @samp{603}, | |
3515 | @samp{603e}, @samp{604}, @samp{620}, @samp{power}, @samp{power2}, | |
3516 | @samp{powerpc}, @samp{403}, @samp{505}, @samp{821}, and @samp{860} and | |
3517 | @samp{common}. @samp{-mcpu=power}, @samp{-mcpu=power2}, and | |
3518 | @samp{-mcpu=powerpc} specify generic POWER, POWER2 and pure PowerPC | |
3519 | (i.e., not MPC601) architecture machine types, with an appropriate, | |
3520 | generic processor model assumed for scheduling purposes.@refill | |
3521 | ||
3522 | @c overfull hbox here --bob 22 jul96 | |
3523 | @c original text between ignore ... end ignore | |
3524 | @ignore | |
3525 | Specifying any of the @samp{-mcpu=rios1}, @samp{-mcpu=rios2}, @samp{-mcpu=rsc}, | |
3526 | @samp{-mcpu=power}, or @samp{-mcpu=power2} options enables the @samp{-mpower} | |
3527 | option and disables the @samp{-mpowerpc} option; @samp{-mcpu=601} | |
3528 | enables both the @samp{-mpower} and @samp{-mpowerpc} options; | |
3529 | all of @samp{-mcpu=602}, @samp{-mcpu=603}, @samp{-mcpu=603e}, @samp{-mcpu=604}, | |
3530 | @samp{-mcpu=620}, @samp{-mcpu=403}, @samp{-mcpu=505}, @samp{-mcpu=821}, | |
3531 | @samp{-mcpu=860} and @samp{-mcpu=powerpc} enable the @samp{-mpowerpc} | |
3532 | option and disable the @samp{-mpower} option; @samp{-mcpu=common} | |
3533 | disables both the @samp{-mpower} and @samp{-mpowerpc} options.@refill | |
3534 | @end ignore | |
3535 | @c changed paragraph | |
3536 | Specifying any of the following options: | |
3537 | @samp{-mcpu=rios1}, @samp{-mcpu=rios2}, @samp{-mcpu=rsc}, | |
3538 | @samp{-mcpu=power}, or @samp{-mcpu=power2} | |
3539 | enables the @samp{-mpower} option and disables the @samp{-mpowerpc} option; | |
3540 | @samp{-mcpu=601} enables both the @samp{-mpower} and @samp{-mpowerpc} options. | |
3541 | All of @samp{-mcpu=602}, @samp{-mcpu=603}, @samp{-mcpu=603e}, | |
3542 | @samp{-mcpu=604}, @samp{-mcpu=620}, | |
3543 | enable the @samp{-mpowerpc} option and disable the @samp{-mpower} option. | |
3544 | Exactly similarly, all of @samp{-mcpu=403}, | |
3545 | @samp{-mcpu=505}, @samp{-mcpu=821}, @samp{-mcpu=860} and @samp{-mcpu=powerpc} | |
3546 | enable the @samp{-mpowerpc} option and disable the @samp{-mpower} option. | |
3547 | @samp{-mcpu=common} disables both the | |
3548 | @samp{-mpower} and @samp{-mpowerpc} options.@refill | |
3549 | @c end changes to prevent overfull hboxes | |
3550 | ||
3551 | AIX versions 4 or greater selects @samp{-mcpu=common} by default, so | |
3552 | that code will operate on all members of the RS/6000 and PowerPC | |
3553 | families. In that case, GNU CC will use only the instructions in the | |
3554 | common subset of both architectures plus some special AIX common-mode | |
3555 | calls, and will not use the MQ register. GNU CC assumes a generic | |
3556 | processor model for scheduling purposes. | |
3557 | ||
3558 | Specifying any of the options @samp{-mcpu=rios1}, @samp{-mcpu=rios2}, | |
3559 | @samp{-mcpu=rsc}, @samp{-mcpu=power}, or @samp{-mcpu=power2} also | |
3560 | disables the @samp{new-mnemonics} option. Specifying @samp{-mcpu=601}, | |
3561 | @samp{-mcpu=602}, @samp{-mcpu=603}, @samp{-mcpu=603e}, @samp{-mcpu=604}, | |
3562 | @samp{620}, @samp{403}, or @samp{-mcpu=powerpc} also enables the | |
3563 | @samp{new-mnemonics} option.@refill | |
3564 | ||
3565 | Specifying @samp{-mcpu=403}, @samp{-mcpu=821}, or @samp{-mcpu=860} also | |
3566 | enables the @samp{-msoft-float} option. | |
3567 | ||
3568 | @item -mtune=@var{cpu_type} | |
3569 | Set the instruction scheduling parameters for machine type | |
3570 | @var{cpu_type}, but do not set the architecture type, register usage, | |
3571 | choice of mnemonics like @samp{-mcpu=}@var{cpu_type} would. The same | |
3572 | values for @var{cpu_type} are used for @samp{-mtune=}@var{cpu_type} as | |
3573 | for @samp{-mcpu=}@var{cpu_type}. The @samp{-mtune=}@var{cpu_type} | |
3574 | option overrides the @samp{-mcpu=}@var{cpu_type} option in terms of | |
3575 | instruction scheduling parameters. | |
3576 | ||
3577 | @item -mfull-toc | |
3578 | @itemx -mno-fp-in-toc | |
3579 | @itemx -mno-sum-in-toc | |
3580 | @itemx -mminimal-toc | |
3581 | Modify generation of the TOC (Table Of Contents), which is created for | |
3582 | every executable file. The @samp{-mfull-toc} option is selected by | |
3583 | default. In that case, GNU CC will allocate at least one TOC entry for | |
3584 | each unique non-automatic variable reference in your program. GNU CC | |
3585 | will also place floating-point constants in the TOC. However, only | |
3586 | 16,384 entries are available in the TOC. | |
3587 | ||
3588 | If you receive a linker error message that saying you have overflowed | |
3589 | the available TOC space, you can reduce the amount of TOC space used | |
3590 | with the @samp{-mno-fp-in-toc} and @samp{-mno-sum-in-toc} options. | |
3591 | @samp{-mno-fp-in-toc} prevents GNU CC from putting floating-point | |
3592 | constants in the TOC and @samp{-mno-sum-in-toc} forces GNU CC to | |
3593 | generate code to calculate the sum of an address and a constant at | |
3594 | run-time instead of putting that sum into the TOC. You may specify one | |
3595 | or both of these options. Each causes GNU CC to produce very slightly | |
3596 | slower and larger code at the expense of conserving TOC space. | |
3597 | ||
3598 | If you still run out of space in the TOC even when you specify both of | |
3599 | these options, specify @samp{-mminimal-toc} instead. This option causes | |
3600 | GNU CC to make only one TOC entry for every file. When you specify this | |
3601 | option, GNU CC will produce code that is slower and larger but which | |
3602 | uses extremely little TOC space. You may wish to use this option | |
3603 | only on files that contain less frequently executed code. @refill | |
3604 | ||
3605 | @item -mxl-call | |
3606 | @itemx -mno-xl-call | |
3607 | On AIX, pass floating-point arguments to prototyped functions beyond the | |
3608 | register save area (RSA) on the stack in addition to argument FPRs. The | |
3609 | AIX calling convention was extended but not initially documented to | |
3610 | handle an obscure K&R C case of calling a function that takes the | |
3611 | address of its arguments with fewer arguments than declared. AIX XL | |
3612 | compilers assume that floating point arguments which do not fit in the | |
3613 | RSA are on the stack when they compile a subroutine without | |
3614 | optimization. Because always storing floating-point arguments on the | |
3615 | stack is inefficient and rarely needed, this option is not enabled by | |
3616 | default and only is necessary when calling subroutines compiled by AIX | |
3617 | XL compilers without optimization. | |
3618 | ||
3619 | @item -msoft-float | |
3620 | @itemx -mhard-float | |
3621 | Generate code that does not use (uses) the floating-point register set. | |
3622 | Software floating point emulation is provided if you use the | |
3623 | @samp{-msoft-float} option, and pass the option to GNU CC when linking. | |
3624 | ||
3625 | @item -mmultiple | |
3626 | @itemx -mno-multiple | |
3627 | Generate code that uses (does not use) the load multiple word | |
3628 | instructions and the store multiple word instructions. These | |
3629 | instructions are generated by default on POWER systems, and not | |
3630 | generated on PowerPC systems. Do not use @samp{-mmultiple} on little | |
3631 | endian PowerPC systems, since those instructions do not work when the | |
3632 | processor is in little endian mode. | |
3633 | ||
3634 | @item -mstring | |
3635 | @itemx -mno-string | |
3636 | Generate code that uses (does not use) the load string instructions and the | |
3637 | store string word instructions to save multiple registers and do small block | |
3638 | moves. These instructions are generated by default on POWER systems, anod not | |
3639 | generated on PowerPC systems. Do not use @samp{-mstring} on little endian | |
3640 | PowerPC systems, since those instructions do not work when the processor is in | |
3641 | little endian mode. | |
3642 | ||
3643 | @item -mno-bit-align | |
3644 | @itemx -mbit-align | |
3645 | On System V.4 and embedded PowerPC systems do not (do) force structures | |
3646 | and unions that contain bit fields to be aligned to the base type of the | |
3647 | bit field. | |
3648 | ||
3649 | For example, by default a structure containing nothing but 8 | |
3650 | @code{unsigned} bitfields of length 1 would be aligned to a 4 byte | |
3651 | boundary and have a size of 4 bytes. By using @samp{-mno-bit-align}, | |
3652 | the structure would be aligned to a 1 byte boundary and be one byte in | |
3653 | size. | |
3654 | ||
3655 | @item -mno-strict-align | |
3656 | @itemx -mstrict-align | |
3657 | On System V.4 and embedded PowerPC systems do not (do) assume that | |
3658 | unaligned memory references will be handled by the system. | |
3659 | ||
3660 | @item -mrelocatable | |
3661 | @itemx -mno-relocatable | |
3662 | On embedded PowerPC systems generate code that allows (does not allow) | |
3663 | the program to be relocated to a different address at runtime. If you | |
3664 | use @samp{-mrelocatable} on any module, all objects linked together must | |
3665 | be compiled with @samp{-mrelocatable} or @samp{-mrelocatable-lib}. | |
3666 | ||
3667 | @item -mrelocatable-lib | |
3668 | @itemx -mno-relocatable-lib | |
3669 | On embedded PowerPC systems generate code that allows (does not allow) | |
3670 | the program to be relocated to a different address at runtime. Modules | |
3671 | compiled with @samp{-mreloctable-lib} can be linked with either modules | |
3672 | compiled without @samp{-mrelocatable} and @samp{-mrelocatable-lib} or | |
3673 | with modules compiled with the @samp{-mrelocatable} options. | |
3674 | ||
3675 | @item -mno-toc | |
3676 | @itemx -mtoc | |
3677 | On System V.4 and embedded PowerPC systems do not (do) assume that | |
3678 | register 2 contains a pointer to a global area pointing to the addresses | |
3679 | used in the program. | |
3680 | ||
3681 | @item -mno-traceback | |
3682 | @itemx -mtraceback | |
3683 | On embedded PowerPC systems do not (do) generate a traceback tag before | |
3684 | the start of the function. This tag can be used by the debugger to | |
3685 | identify where the start of a function is. | |
3686 | ||
3687 | @item -mlittle | |
3688 | @itemx -mlittle-endian | |
3689 | On System V.4 and embedded PowerPC systems compile code for the | |
3690 | processor in little endian mode. The @samp{-mlittle-endian} option is | |
3691 | the same as @samp{-mlittle}. | |
3692 | ||
3693 | @item -mbig | |
3694 | @itemx -mbig-endian | |
3695 | On System V.4 and embedded PowerPC systems compile code for the | |
3696 | processor in big endian mode. The @samp{-mbig-endian} option is | |
3697 | the same as @samp{-mbig}. | |
3698 | ||
3699 | @item -mcall-sysv | |
3700 | On System V.4 and embedded PowerPC systems compile code using calling | |
3701 | conventions that adheres to the March 1995 draft of the System V | |
3702 | Application Binary Interface, PowerPC processor supplement. This is the | |
3703 | default unless you configured GCC using @samp{powerpc-*-eabiaix}. | |
3704 | ||
3705 | @item -mcall-sysv-eabi | |
3706 | Specify both @samp{-mcall-sysv} and @samp{-meabi} options. | |
3707 | ||
3708 | @item -mcall-sysv-noeabi | |
3709 | Specify both @samp{-mcall-sysv} and @samp{-mno-eabi} options. | |
3710 | ||
3711 | @item -mcall-aix | |
3712 | On System V.4 and embedded PowerPC systems compile code using calling | |
3713 | conventions that are similar to those used on AIX. This is the | |
3714 | default if you configured GCC using @samp{powerpc-*-eabiaix}. | |
3715 | ||
3716 | @item -mcall-solaris | |
3717 | On System V.4 and embedded PowerPC systems compile code for the Solaris | |
3718 | operating system. | |
3719 | ||
3720 | @item -mcall-linux | |
3721 | On System V.4 and embedded PowerPC systems compile code for the Linux | |
3722 | operating system. | |
3723 | ||
3724 | @item -mprototype | |
3725 | @item -mno-prototype | |
3726 | On System V.4 and embedded PowerPC systems assume that all calls to | |
3727 | variable argument functions are properly prototyped. Otherwise, the | |
3728 | compiler must insert an instruction before every non prototyped call to | |
3729 | set or clear bit 6 of the condition code register (@var{CR}) to | |
3730 | indicate whether floating point values were passed in the floating point | |
3731 | registers in case the function takes a variable arguments. With | |
3732 | @samp{-mprototype}, only calls to prototyped variable argument functions | |
3733 | will set or clear the bit. | |
3734 | ||
3735 | @item -msim | |
3736 | On embedded PowerPC systems, assume that the startup module is called | |
3737 | @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and | |
3738 | @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}. | |
3739 | configurations. | |
3740 | ||
3741 | @item -mmvme | |
3742 | On embedded PowerPC systems, assume that the startup module is called | |
3743 | @file{mvme-crt0.o} and the standard C libraries are @file{libmvme.a} and | |
3744 | @file{libc.a}. | |
3745 | ||
3746 | @item -memb | |
3747 | On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags | |
3748 | header to indicate that @samp{eabi} extended relocations are used. | |
3749 | ||
3750 | @item -meabi | |
3751 | @itemx -mno-eabi | |
3752 | On System V.4 and embedded PowerPC systems do (do not) adhere to the | |
3753 | Embedded Applications Binary Interface (eabi) which is a set of | |
3754 | modifications to the System V.4 specifications. Selecting @code{-meabi} | |
3755 | means that the stack is aligned to an 8 byte boundary, a function | |
3756 | @code{__eabi} is called to from @code{main} to set up the eabi | |
3757 | environment, and the @samp{-msdata} option can use both @code{r2} and | |
3758 | @code{r13} to point to two separate small data areas. Selecting | |
3759 | @code{-mno-eabi} means that the stack is aligned to a 16 byte boundary, | |
3760 | do not call an initialization function from @code{main}, and the | |
3761 | @samp{-msdata} option will only use @code{r13} to point to a single | |
3762 | small data area. The @samp{-meabi} option is on by default if you | |
3763 | configured GCC using one of the @samp{powerpc*-*-eabi*} options. | |
3764 | ||
3765 | @item -msdata=eabi | |
3766 | On System V.4 and embedded PowerPC systems, put small initialized | |
3767 | @code{const} global and static data in the @samp{.sdata2} section, which | |
3768 | is pointed to by register @code{r2}. Put small initialized | |
3769 | non-@code{const} global and static data in the @samp{.sdata} section, | |
3770 | which is pointed to by register @code{r13}. Put small uninitialized | |
3771 | global and static data in the @samp{.sbss} section, which is adjacent to | |
3772 | the @samp{.sdata} section. The @samp{-msdata=eabi} option is | |
3773 | incompatible with the @samp{-mrelocatable} option. The | |
3774 | @samp{-msdata=eabi} option also sets the @samp{-memb} option. | |
3775 | ||
3776 | @item -msdata=sysv | |
3777 | On System V.4 and embedded PowerPC systems, put small global and static | |
3778 | data in the @samp{.sdata} section, which is pointed to by register | |
3779 | @code{r13}. Put small uninitialized global and static data in the | |
3780 | @samp{.sbss} section, which is adjacent to the @samp{.sdata} section. | |
3781 | The @samp{-msdata=sysv} option is incompatible with the | |
3782 | @samp{-mrelocatable} option. | |
3783 | ||
3784 | @item -msdata=default | |
3785 | @itemx -msdata | |
3786 | On System V.4 and embedded PowerPC systems, if @samp{-meabi} is used, | |
3787 | compile code the same as @samp{-msdata=eabi}, otherwise compile code the | |
3788 | same as @samp{-msdata=sysv}. | |
3789 | ||
3790 | @item -msdata-data | |
3791 | On System V.4 and embedded PowerPC systems, put small global and static | |
3792 | data in the @samp{.sdata} section. Put small uninitialized global and | |
3793 | static data in the @samp{.sbss} section. Do not use register @code{r13} | |
3794 | to address small data however. This is the default behavior unless | |
3795 | other @samp{-msdata} options are used. | |
3796 | ||
3797 | @item -msdata=none | |
3798 | @itemx -mno-sdata | |
3799 | On embedded PowerPC systems, put all initialized global and static data | |
3800 | in the @samp{.data} section, and all uninitialized data in the | |
3801 | @samp{.bss} section. | |
3802 | ||
3803 | @item -G @var{num} | |
3804 | @cindex smaller data references (PowerPC) | |
3805 | @cindex .sdata/.sdata2 references (PowerPC) | |
3806 | On embbeded PowerPC systems, put global and static items less than or | |
3807 | equal to @var{num} bytes into the small data or bss sections instead of | |
3808 | the normal data or bss section. By default, @var{num} is 8. The | |
3809 | @samp{-G @var{num}} switch is also passed to the linker. | |
3810 | All modules should be compiled with the same @samp{-G @var{num}} value. | |
3811 | ||
3812 | @item -mregnames | |
3813 | @itemx -mno-regnames | |
3814 | On System V.4 and embedded PowerPC systems do (do not) emit register | |
3815 | names in the assembly language output using symbolic forms. | |
3816 | @end table | |
3817 | @node RT Options | |
3818 | @subsection IBM RT Options | |
3819 | @cindex RT options | |
3820 | @cindex IBM RT options | |
3821 | ||
3822 | These @samp{-m} options are defined for the IBM RT PC: | |
3823 | ||
3824 | @table @code | |
3825 | @item -min-line-mul | |
3826 | Use an in-line code sequence for integer multiplies. This is the | |
3827 | default. | |
3828 | ||
3829 | @item -mcall-lib-mul | |
3830 | Call @code{lmul$$} for integer multiples. | |
3831 | ||
3832 | @item -mfull-fp-blocks | |
3833 | Generate full-size floating point data blocks, including the minimum | |
3834 | amount of scratch space recommended by IBM. This is the default. | |
3835 | ||
3836 | @item -mminimum-fp-blocks | |
3837 | Do not include extra scratch space in floating point data blocks. This | |
3838 | results in smaller code, but slower execution, since scratch space must | |
3839 | be allocated dynamically. | |
3840 | ||
3841 | @cindex @file{varargs.h} and RT PC | |
3842 | @cindex @file{stdarg.h} and RT PC | |
3843 | @item -mfp-arg-in-fpregs | |
3844 | Use a calling sequence incompatible with the IBM calling convention in | |
3845 | which floating point arguments are passed in floating point registers. | |
3846 | Note that @code{varargs.h} and @code{stdargs.h} will not work with | |
3847 | floating point operands if this option is specified. | |
3848 | ||
3849 | @item -mfp-arg-in-gregs | |
3850 | Use the normal calling convention for floating point arguments. This is | |
3851 | the default. | |
3852 | ||
3853 | @item -mhc-struct-return | |
3854 | Return structures of more than one word in memory, rather than in a | |
3855 | register. This provides compatibility with the MetaWare HighC (hc) | |
3856 | compiler. Use the option @samp{-fpcc-struct-return} for compatibility | |
3857 | with the Portable C Compiler (pcc). | |
3858 | ||
3859 | @item -mnohc-struct-return | |
3860 | Return some structures of more than one word in registers, when | |
3861 | convenient. This is the default. For compatibility with the | |
3862 | IBM-supplied compilers, use the option @samp{-fpcc-struct-return} or the | |
3863 | option @samp{-mhc-struct-return}. | |
3864 | @end table | |
3865 | ||
3866 | @node MIPS Options | |
3867 | @subsection MIPS Options | |
3868 | @cindex MIPS options | |
3869 | ||
3870 | These @samp{-m} options are defined for the MIPS family of computers: | |
3871 | ||
3872 | @table @code | |
3873 | @item -mcpu=@var{cpu type} | |
3874 | Assume the defaults for the machine type @var{cpu type} when scheduling | |
3875 | instructions. The choices for @var{cpu type} are @samp{r2000}, @samp{r3000}, | |
3876 | @samp{r4000}, @samp{r4400}, @samp{r4600}, and @samp{r6000}. While picking a | |
3877 | specific @var{cpu type} will schedule things appropriately for that | |
3878 | particular chip, the compiler will not generate any code that does not | |
3879 | meet level 1 of the MIPS ISA (instruction set architecture) without | |
3880 | the @samp{-mips2} or @samp{-mips3} switches being used. | |
3881 | ||
3882 | @item -mips1 | |
3883 | Issue instructions from level 1 of the MIPS ISA. This is the default. | |
3884 | @samp{r3000} is the default @var{cpu type} at this ISA level. | |
3885 | ||
3886 | @item -mips2 | |
3887 | Issue instructions from level 2 of the MIPS ISA (branch likely, square | |
3888 | root instructions). @samp{r6000} is the default @var{cpu type} at this | |
3889 | ISA level. | |
3890 | ||
3891 | @item -mips3 | |
3892 | Issue instructions from level 3 of the MIPS ISA (64 bit instructions). | |
3893 | @samp{r4000} is the default @var{cpu type} at this ISA level. | |
3894 | This option does not change the sizes of any of the C data types. | |
3895 | ||
3896 | @item -mfp32 | |
3897 | Assume that 32 32-bit floating point registers are available. This is | |
3898 | the default. | |
3899 | ||
3900 | @item -mfp64 | |
3901 | Assume that 32 64-bit floating point registers are available. This is | |
3902 | the default when the @samp{-mips3} option is used. | |
3903 | ||
3904 | @item -mgp32 | |
3905 | Assume that 32 32-bit general purpose registers are available. This is | |
3906 | the default. | |
3907 | ||
3908 | @item -mgp64 | |
3909 | Assume that 32 64-bit general purpose registers are available. This is | |
3910 | the default when the @samp{-mips3} option is used. | |
3911 | ||
3912 | @item -mint64 | |
3913 | Types long, int, and pointer are 64 bits. This works only if @samp{-mips3} | |
3914 | is also specified. | |
3915 | ||
3916 | @item -mlong64 | |
3917 | Types long and pointer are 64 bits, and type int is 32 bits. | |
3918 | This works only if @samp{-mips3} is also specified. | |
3919 | ||
3920 | @item -mmips-as | |
3921 | Generate code for the MIPS assembler, and invoke @file{mips-tfile} to | |
3922 | add normal debug information. This is the default for all | |
3923 | platforms except for the OSF/1 reference platform, using the OSF/rose | |
3924 | object format. If the either of the @samp{-gstabs} or @samp{-gstabs+} | |
3925 | switches are used, the @file{mips-tfile} program will encapsulate the | |
3926 | stabs within MIPS ECOFF. | |
3927 | ||
3928 | @item -mgas | |
3929 | Generate code for the GNU assembler. This is the default on the OSF/1 | |
3930 | reference platform, using the OSF/rose object format. | |
3931 | ||
3932 | @item -mrnames | |
3933 | @itemx -mno-rnames | |
3934 | The @samp{-mrnames} switch says to output code using the MIPS software | |
3935 | names for the registers, instead of the hardware names (ie, @var{a0} | |
3936 | instead of @var{$4}). The only known assembler that supports this option | |
3937 | is the Algorithmics assembler. | |
3938 | ||
3939 | @item -mgpopt | |
3940 | @itemx -mno-gpopt | |
3941 | The @samp{-mgpopt} switch says to write all of the data declarations | |
3942 | before the instructions in the text section, this allows the MIPS | |
3943 | assembler to generate one word memory references instead of using two | |
3944 | words for short global or static data items. This is on by default if | |
3945 | optimization is selected. | |
3946 | ||
3947 | @item -mstats | |
3948 | @itemx -mno-stats | |
3949 | For each non-inline function processed, the @samp{-mstats} switch | |
3950 | causes the compiler to emit one line to the standard error file to | |
3951 | print statistics about the program (number of registers saved, stack | |
3952 | size, etc.). | |
3953 | ||
3954 | @item -mmemcpy | |
3955 | @itemx -mno-memcpy | |
3956 | The @samp{-mmemcpy} switch makes all block moves call the appropriate | |
3957 | string function (@samp{memcpy} or @samp{bcopy}) instead of possibly | |
3958 | generating inline code. | |
3959 | ||
3960 | @item -mmips-tfile | |
3961 | @itemx -mno-mips-tfile | |
3962 | The @samp{-mno-mips-tfile} switch causes the compiler not | |
3963 | postprocess the object file with the @file{mips-tfile} program, | |
3964 | after the MIPS assembler has generated it to add debug support. If | |
3965 | @file{mips-tfile} is not run, then no local variables will be | |
3966 | available to the debugger. In addition, @file{stage2} and | |
3967 | @file{stage3} objects will have the temporary file names passed to the | |
3968 | assembler embedded in the object file, which means the objects will | |
3969 | not compare the same. The @samp{-mno-mips-tfile} switch should only | |
3970 | be used when there are bugs in the @file{mips-tfile} program that | |
3971 | prevents compilation. | |
3972 | ||
3973 | @item -msoft-float | |
3974 | Generate output containing library calls for floating point. | |
3975 | @strong{Warning:} the requisite libraries are not part of GNU CC. | |
3976 | Normally the facilities of the machine's usual C compiler are used, but | |
3977 | this can't be done directly in cross-compilation. You must make your | |
3978 | own arrangements to provide suitable library functions for | |
3979 | cross-compilation. | |
3980 | ||
3981 | @item -mhard-float | |
3982 | Generate output containing floating point instructions. This is the | |
3983 | default if you use the unmodified sources. | |
3984 | ||
3985 | @item -mabicalls | |
3986 | @itemx -mno-abicalls | |
3987 | Emit (or do not emit) the pseudo operations @samp{.abicalls}, | |
3988 | @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for | |
3989 | position independent code. | |
3990 | ||
3991 | @item -mlong-calls | |
3992 | @itemx -mno-long-calls | |
3993 | Do all calls with the @samp{JALR} instruction, which requires | |
3994 | loading up a function's address into a register before the call. | |
3995 | You need to use this switch, if you call outside of the current | |
3996 | 512 megabyte segment to functions that are not through pointers. | |
3997 | ||
3998 | @item -mhalf-pic | |
3999 | @itemx -mno-half-pic | |
4000 | Put pointers to extern references into the data section and load them | |
4001 | up, rather than put the references in the text section. | |
4002 | ||
4003 | @item -membedded-pic | |
4004 | @itemx -mno-embedded-pic | |
4005 | Generate PIC code suitable for some embedded systems. All calls are made | |
4006 | using PC relative address, and all data is addressed using the $gp register. | |
4007 | This requires GNU as and GNU ld which do most of the work. | |
4008 | ||
4009 | @item -membedded-data | |
4010 | @itemx -mno-embedded-data | |
4011 | Allocate variables to the read-only data section first if possible, then | |
4012 | next in the small data section if possible, otherwise in data. This gives | |
4013 | slightly slower code than the default, but reduces the amount of RAM required | |
4014 | when executing, and thus may be preferred for some embedded systems. | |
4015 | ||
4016 | @item -msingle-float | |
4017 | @itemx -mdouble-float | |
4018 | The @samp{-msingle-float} switch tells gcc to assume that the floating | |
4019 | point coprocessor only supports single precision operations, as on the | |
4020 | @samp{r4650} chip. The @samp{-mdouble-float} switch permits gcc to use | |
4021 | double precision operations. This is the default. | |
4022 | ||
4023 | @item -mmad | |
4024 | @itemx -mno-mad | |
4025 | Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions, | |
4026 | as on the @samp{r4650} chip. | |
4027 | ||
4028 | @item -m4650 | |
4029 | Turns on @samp{-msingle-float}, @samp{-mmad}, and, at least for now, | |
4030 | @samp{-mcpu=r4650}. | |
4031 | ||
4032 | @item -EL | |
4033 | Compile code for the processor in little endian mode. | |
4034 | The requisite libraries are assumed to exist. | |
4035 | ||
4036 | @item -EB | |
4037 | Compile code for the processor in big endian mode. | |
4038 | The requisite libraries are assumed to exist. | |
4039 | ||
4040 | @item -G @var{num} | |
4041 | @cindex smaller data references (MIPS) | |
4042 | @cindex gp-relative references (MIPS) | |
4043 | Put global and static items less than or equal to @var{num} bytes into | |
4044 | the small data or bss sections instead of the normal data or bss | |
4045 | section. This allows the assembler to emit one word memory reference | |
4046 | instructions based on the global pointer (@var{gp} or @var{$28}), | |
4047 | instead of the normal two words used. By default, @var{num} is 8 when | |
4048 | the MIPS assembler is used, and 0 when the GNU assembler is used. The | |
4049 | @samp{-G @var{num}} switch is also passed to the assembler and linker. | |
4050 | All modules should be compiled with the same @samp{-G @var{num}} | |
4051 | value. | |
4052 | ||
4053 | @item -nocpp | |
4054 | Tell the MIPS assembler to not run it's preprocessor over user | |
4055 | assembler files (with a @samp{.s} suffix) when assembling them. | |
4056 | @end table | |
4057 | ||
4058 | @ifset INTERNALS | |
4059 | These options are defined by the macro | |
4060 | @code{TARGET_SWITCHES} in the machine description. The default for the | |
4061 | options is also defined by that macro, which enables you to change the | |
4062 | defaults. | |
4063 | @end ifset | |
4064 | ||
4065 | @node i386 Options | |
4066 | @subsection Intel 386 Options | |
4067 | @cindex i386 Options | |
4068 | @cindex Intel 386 Options | |
4069 | ||
4070 | These @samp{-m} options are defined for the i386 family of computers: | |
4071 | ||
4072 | @table @code | |
4073 | @item -m486 | |
4074 | @itemx -m386 | |
4075 | Control whether or not code is optimized for a 486 instead of an | |
4076 | 386. Code generated for an 486 will run on a 386 and vice versa. | |
4077 | ||
4078 | @item -mieee-fp | |
4079 | @itemx -mno-ieee-fp | |
4080 | Control whether or not the compiler uses IEEE floating point | |
4081 | comparisons. These handle correctly the case where the result of a | |
4082 | comparison is unordered. | |
4083 | ||
4084 | @item -msoft-float | |
4085 | Generate output containing library calls for floating point. | |
4086 | @strong{Warning:} the requisite libraries are not part of GNU CC. | |
4087 | Normally the facilities of the machine's usual C compiler are used, but | |
4088 | this can't be done directly in cross-compilation. You must make your | |
4089 | own arrangements to provide suitable library functions for | |
4090 | cross-compilation. | |
4091 | ||
4092 | On machines where a function returns floating point results in the 80387 | |
4093 | register stack, some floating point opcodes may be emitted even if | |
4094 | @samp{-msoft-float} is used. | |
4095 | ||
4096 | @item -mno-fp-ret-in-387 | |
4097 | Do not use the FPU registers for return values of functions. | |
4098 | ||
4099 | The usual calling convention has functions return values of types | |
4100 | @code{float} and @code{double} in an FPU register, even if there | |
4101 | is no FPU. The idea is that the operating system should emulate | |
4102 | an FPU. | |
4103 | ||
4104 | The option @samp{-mno-fp-ret-in-387} causes such values to be returned | |
4105 | in ordinary CPU registers instead. | |
4106 | ||
4107 | @item -mno-fancy-math-387 | |
4108 | Some 387 emulators do not support the @code{sin}, @code{cos} and | |
4109 | @code{sqrt} instructions for the 387. Specify this option to avoid | |
4110 | generating those instructions. This option is the default on FreeBSD. | |
4111 | As of revision 2.6.1, these instructions are not generated unless you | |
4112 | also use the @samp{-ffast-math} switch. | |
4113 | ||
4114 | @item -malign-double | |
4115 | @itemx -mno-align-double | |
4116 | Control whether GNU CC aligns @code{double}, @code{long double}, and | |
4117 | @code{long long} variables on a two word boundary or a one word | |
4118 | boundary. Aligning @code{double} variables on a two word boundary will | |
4119 | produce code that runs somewhat faster on a @samp{Pentium} at the | |
4120 | expense of more memory. | |
4121 | ||
4122 | @strong{Warning:} if you use the @samp{-malign-double} switch, | |
4123 | structures containing the above types will be aligned differently than | |
4124 | the published application binary interface specifications for the 386. | |
4125 | ||
4126 | @item -msvr3-shlib | |
4127 | @itemx -mno-svr3-shlib | |
4128 | Control whether GNU CC places uninitialized locals into @code{bss} or | |
4129 | @code{data}. @samp{-msvr3-shlib} places these locals into @code{bss}. | |
4130 | These options are meaningful only on System V Release 3. | |
4131 | ||
4132 | @item -mno-wide-multiply | |
4133 | @itemx -mwide-multiply | |
4134 | Control whether GNU CC uses the @code{mul} and @code{imul} that produce | |
4135 | 64 bit results in @code{eax:edx} from 32 bit operands to do @code{long | |
4136 | long} multiplies and 32-bit division by constants. | |
4137 | ||
4138 | @item -mrtd | |
4139 | Use a different function-calling convention, in which functions that | |
4140 | take a fixed number of arguments return with the @code{ret} @var{num} | |
4141 | instruction, which pops their arguments while returning. This saves one | |
4142 | instruction in the caller since there is no need to pop the arguments | |
4143 | there. | |
4144 | ||
4145 | You can specify that an individual function is called with this calling | |
4146 | sequence with the function attribute @samp{stdcall}. You can also | |
4147 | override the @samp{-mrtd} option by using the function attribute | |
4148 | @samp{cdecl}. @xref{Function Attributes} | |
4149 | ||
4150 | @strong{Warning:} this calling convention is incompatible with the one | |
4151 | normally used on Unix, so you cannot use it if you need to call | |
4152 | libraries compiled with the Unix compiler. | |
4153 | ||
4154 | Also, you must provide function prototypes for all functions that | |
4155 | take variable numbers of arguments (including @code{printf}); | |
4156 | otherwise incorrect code will be generated for calls to those | |
4157 | functions. | |
4158 | ||
4159 | In addition, seriously incorrect code will result if you call a | |
4160 | function with too many arguments. (Normally, extra arguments are | |
4161 | harmlessly ignored.) | |
4162 | ||
4163 | @item -mreg-alloc=@var{regs} | |
4164 | Control the default allocation order of integer registers. The | |
4165 | string @var{regs} is a series of letters specifying a register. The | |
4166 | supported letters are: @code{a} allocate EAX; @code{b} allocate EBX; | |
4167 | @code{c} allocate ECX; @code{d} allocate EDX; @code{S} allocate ESI; | |
4168 | @code{D} allocate EDI; @code{B} allocate EBP. | |
4169 | ||
4170 | @item -mregparm=@var{num} | |
4171 | Control how many registers are used to pass integer arguments. By | |
4172 | default, no registers are used to pass arguments, and at most 3 | |
4173 | registers can be used. You can control this behavior for a specific | |
4174 | function by using the function attribute @samp{regparm}. @xref{Function Attributes} | |
4175 | ||
4176 | @strong{Warning:} if you use this switch, and | |
4177 | @var{num} is nonzero, then you must build all modules with the same | |
4178 | value, including any libraries. This includes the system libraries and | |
4179 | startup modules. | |
4180 | ||
4181 | @item -malign-loops=@var{num} | |
4182 | Align loops to a 2 raised to a @var{num} byte boundary. If | |
4183 | @samp{-malign-loops} is not specified, the default is 2. | |
4184 | ||
4185 | @item -malign-jumps=@var{num} | |
4186 | Align instructions that are only jumped to to a 2 raised to a @var{num} | |
4187 | byte boundary. If @samp{-malign-jumps} is not specified, the default is | |
4188 | 2 if optimizing for a 386, and 4 if optimizing for a 486. | |
4189 | ||
4190 | @item -malign-functions=@var{num} | |
4191 | Align the start of functions to a 2 raised to @var{num} byte boundary. | |
4192 | If @samp{-malign-jumps} is not specified, the default is 2 if optimizing | |
4193 | for a 386, and 4 if optimizing for a 486. | |
4194 | @end table | |
4195 | ||
4196 | @node HPPA Options | |
4197 | @subsection HPPA Options | |
4198 | @cindex HPPA Options | |
4199 | ||
4200 | These @samp{-m} options are defined for the HPPA family of computers: | |
4201 | ||
4202 | @table @code | |
4203 | @item -mpa-risc-1-0 | |
4204 | Generate code for a PA 1.0 processor. | |
4205 | ||
4206 | @item -mpa-risc-1-1 | |
4207 | Generate code for a PA 1.1 processor. | |
4208 | ||
4209 | @item -mjump-in-delay | |
4210 | Fill delay slots of function calls with unconditional jump instructions | |
4211 | by modifying the return pointer for the function call to be the target | |
4212 | of the conditional jump. | |
4213 | ||
4214 | @item -mdisable-fpregs | |
4215 | Prevent floating point registers from being used in any manner. This is | |
4216 | necessary for compiling kernels which perform lazy context switching of | |
4217 | floating point registers. If you use this option and attempt to perform | |
4218 | floating point operations, the compiler will abort. | |
4219 | ||
4220 | @item -mdisable-indexing | |
4221 | Prevent the compiler from using indexing address modes. This avoids some | |
4222 | rather obscure problems when compiling MIG generated code under MACH. | |
4223 | ||
4224 | @item -mno-space-regs | |
4225 | Generate code that assumes the target has no space registers. This allows | |
4226 | GCC to generate faster indirect calls and use unscaled index address modes. | |
4227 | ||
4228 | Such code is suitable for level 0 PA systems and kernels. | |
4229 | ||
4230 | @item -mfast-indirect-calls | |
4231 | Generate code that assumes calls never cross space boundaries. This | |
4232 | allows GCC to emit code which performs faster indirect calls. | |
4233 | ||
4234 | This option will not work in the presense of shared libraries or nested | |
4235 | functions. | |
4236 | ||
4237 | @item -mspace | |
4238 | Optimize for space rather than execution time. Currently this only | |
4239 | enables out of line function prologues and epilogues. This option is | |
4240 | incompatable with PIC code generation and profiling. | |
4241 | ||
4242 | @item -mlong-load-store | |
4243 | Generate 3-instruction load and store sequences as sometimes required by | |
4244 | the HP-UX 10 linker. This is equivalent to the @samp{+k} option to | |
4245 | the HP compilers. | |
4246 | ||
4247 | @item -mportable-runtime | |
4248 | Use the portable calling conventions proposed by HP for ELF systems. | |
4249 | ||
4250 | @item -mgas | |
4251 | Enable the use of assembler directives only GAS understands. | |
4252 | ||
4253 | @item -mschedule=@var{cpu type} | |
4254 | Schedule code according to the constraints for the machine type | |
4255 | @var{cpu type}. The choices for @var{cpu type} are @samp{700} for | |
4256 | 7@var{n}0 machines, @samp{7100} for 7@var{n}5 machines, and @samp{7100} | |
4257 | for 7@var{n}2 machines. @samp{7100} is the default for @var{cpu type}. | |
4258 | ||
4259 | Note the @samp{7100LC} scheduling information is incomplete and using | |
4260 | @samp{7100LC} often leads to bad schedules. For now it's probably best | |
4261 | to use @samp{7100} instead of @samp{7100LC} for the 7@var{n}2 machines. | |
4262 | ||
4263 | @item -mlinker-opt | |
4264 | Enable the optimization pass in the HPUX linker. Note this makes symbolic | |
4265 | debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers | |
4266 | in which they give bogus error messages when linking some programs. | |
4267 | ||
4268 | @item -msoft-float | |
4269 | Generate output containing library calls for floating point. | |
4270 | @strong{Warning:} the requisite libraries are not available for all HPPA | |
4271 | targets. Normally the facilities of the machine's usual C compiler are | |
4272 | used, but this cannot be done directly in cross-compilation. You must make | |
4273 | your own arrangements to provide suitable library functions for | |
4274 | cross-compilation. The embedded target @samp{hppa1.1-*-pro} | |
4275 | does provide software floating point support. | |
4276 | ||
4277 | @samp{-msoft-float} changes the calling convention in the output file; | |
4278 | therefore, it is only useful if you compile @emph{all} of a program with | |
4279 | this option. In particular, you need to compile @file{libgcc.a}, the | |
4280 | library that comes with GNU CC, with @samp{-msoft-float} in order for | |
4281 | this to work. | |
4282 | @end table | |
4283 | ||
4284 | @node Intel 960 Options | |
4285 | @subsection Intel 960 Options | |
4286 | ||
4287 | These @samp{-m} options are defined for the Intel 960 implementations: | |
4288 | ||
4289 | @table @code | |
4290 | @item -m@var{cpu type} | |
4291 | Assume the defaults for the machine type @var{cpu type} for some of | |
4292 | the other options, including instruction scheduling, floating point | |
4293 | support, and addressing modes. The choices for @var{cpu type} are | |
4294 | @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf}, | |
4295 | @samp{sa}, and @samp{sb}. | |
4296 | The default is | |
4297 | @samp{kb}. | |
4298 | ||
4299 | @item -mnumerics | |
4300 | @itemx -msoft-float | |
4301 | The @samp{-mnumerics} option indicates that the processor does support | |
4302 | floating-point instructions. The @samp{-msoft-float} option indicates | |
4303 | that floating-point support should not be assumed. | |
4304 | ||
4305 | @item -mleaf-procedures | |
4306 | @itemx -mno-leaf-procedures | |
4307 | Do (or do not) attempt to alter leaf procedures to be callable with the | |
4308 | @code{bal} instruction as well as @code{call}. This will result in more | |
4309 | efficient code for explicit calls when the @code{bal} instruction can be | |
4310 | substituted by the assembler or linker, but less efficient code in other | |
4311 | cases, such as calls via function pointers, or using a linker that doesn't | |
4312 | support this optimization. | |
4313 | ||
4314 | @item -mtail-call | |
4315 | @itemx -mno-tail-call | |
4316 | Do (or do not) make additional attempts (beyond those of the | |
4317 | machine-independent portions of the compiler) to optimize tail-recursive | |
4318 | calls into branches. You may not want to do this because the detection of | |
4319 | cases where this is not valid is not totally complete. The default is | |
4320 | @samp{-mno-tail-call}. | |
4321 | ||
4322 | @item -mcomplex-addr | |
4323 | @itemx -mno-complex-addr | |
4324 | Assume (or do not assume) that the use of a complex addressing mode is a | |
4325 | win on this implementation of the i960. Complex addressing modes may not | |
4326 | be worthwhile on the K-series, but they definitely are on the C-series. | |
4327 | The default is currently @samp{-mcomplex-addr} for all processors except | |
4328 | the CB and CC. | |
4329 | ||
4330 | @item -mcode-align | |
4331 | @itemx -mno-code-align | |
4332 | Align code to 8-byte boundaries for faster fetching (or don't bother). | |
4333 | Currently turned on by default for C-series implementations only. | |
4334 | ||
4335 | @ignore | |
4336 | @item -mclean-linkage | |
4337 | @itemx -mno-clean-linkage | |
4338 | These options are not fully implemented. | |
4339 | @end ignore | |
4340 | ||
4341 | @item -mic-compat | |
4342 | @itemx -mic2.0-compat | |
4343 | @itemx -mic3.0-compat | |
4344 | Enable compatibility with iC960 v2.0 or v3.0. | |
4345 | ||
4346 | @item -masm-compat | |
4347 | @itemx -mintel-asm | |
4348 | Enable compatibility with the iC960 assembler. | |
4349 | ||
4350 | @item -mstrict-align | |
4351 | @itemx -mno-strict-align | |
4352 | Do not permit (do permit) unaligned accesses. | |
4353 | ||
4354 | @item -mold-align | |
4355 | Enable structure-alignment compatibility with Intel's gcc release version | |
4356 | 1.3 (based on gcc 1.37). Currently this is buggy in that @samp{#pragma | |
4357 | align 1} is always assumed as well, and cannot be turned off. | |
4358 | @end table | |
4359 | ||
4360 | @node DEC Alpha Options | |
4361 | @subsection DEC Alpha Options | |
4362 | ||
4363 | These @samp{-m} options are defined for the DEC Alpha implementations: | |
4364 | ||
4365 | @table @code | |
4366 | @item -mno-soft-float | |
4367 | @itemx -msoft-float | |
4368 | Use (do not use) the hardware floating-point instructions for | |
4369 | floating-point operations. When @code{-msoft-float} is specified, | |
4370 | functions in @file{libgcc1.c} will be used to perform floating-point | |
4371 | operations. Unless they are replaced by routines that emulate the | |
4372 | floating-point operations, or compiled in such a way as to call such | |
4373 | emulations routines, these routines will issue floating-point | |
4374 | operations. If you are compiling for an Alpha without floating-point | |
4375 | operations, you must ensure that the library is built so as not to call | |
4376 | them. | |
4377 | ||
4378 | Note that Alpha implementations without floating-point operations are | |
4379 | required to have floating-point registers. | |
4380 | ||
4381 | @item -mfp-reg | |
4382 | @itemx -mno-fp-regs | |
4383 | Generate code that uses (does not use) the floating-point register set. | |
4384 | @code{-mno-fp-regs} implies @code{-msoft-float}. If the floating-point | |
4385 | register set is not used, floating point operands are passed in integer | |
4386 | registers as if they were integers and floating-point results are passed | |
4387 | in $0 instead of $f0. This is a non-standard calling sequence, so any | |
4388 | function with a floating-point argument or return value called by code | |
4389 | compiled with @code{-mno-fp-regs} must also be compiled with that | |
4390 | option. | |
4391 | ||
4392 | A typical use of this option is building a kernel that does not use, | |
4393 | and hence need not save and restore, any floating-point registers. | |
4394 | ||
4395 | @item -mieee | |
4396 | The Alpha architecture implements floating-point hardware optimized for | |
4397 | maximum performance. It is mostly compliant with the IEEE floating | |
4398 | point standard. However, for full compliance, software assistance is | |
4399 | required. This option generates code fully IEEE compliant code | |
4400 | @emph{except} that the @var{inexact flag} is not maintained (see below). | |
4401 | If this option is turned on, the CPP macro @code{_IEEE_FP} is defined | |
4402 | during compilation. The option is a shorthand for: @samp{-D_IEEE_FP | |
4403 | -mfp-trap-mode=su -mtrap-precision=i -mieee-conformant}. The resulting | |
4404 | code is less efficient but is able to correctly support denormalized | |
4405 | numbers and exceptional IEEE values such as not-a-number and plus/minus | |
4406 | infinity. Other Alpha compilers call this option | |
4407 | @code{-ieee_with_no_inexact}. | |
4408 | ||
4409 | @item -mieee-with-inexact | |
4410 | @c overfull hbox here --bob 22 jul96 | |
4411 | @c original text between ignore ... end ignore | |
4412 | @ignore | |
4413 | This is like @samp{-mieee} except the generated code also maintains the | |
4414 | IEEE @var{inexact flag}. Turning on this option causes the generated | |
4415 | code to implement fully-compliant IEEE math. The option is a shorthand | |
4416 | for @samp{-D_IEEE_FP -D_IEEE_FP_INEXACT} plus @samp{-mieee-conformant}, | |
4417 | @samp{-mfp-trap-mode=sui}, and @samp{-mtrap-precision=i}. On some Alpha | |
4418 | implementations the resulting code may execute significantly slower than | |
4419 | the code generated by default. Since there is very little code that | |
4420 | depends on the @var{inexact flag}, you should normally not specify this | |
4421 | option. Other Alpha compilers call this option | |
4422 | @samp{-ieee_with_inexact}. | |
4423 | @end ignore | |
4424 | @c changed paragraph | |
4425 | This is like @samp{-mieee} except the generated code also maintains the | |
4426 | IEEE @var{inexact flag}. Turning on this option causes the generated | |
4427 | code to implement fully-compliant IEEE math. The option is a shorthand | |
4428 | for @samp{-D_IEEE_FP -D_IEEE_FP_INEXACT} plus the three following: | |
4429 | @samp{-mieee-conformant}, | |
4430 | @samp{-mfp-trap-mode=sui}, | |
4431 | and @samp{-mtrap-precision=i}. | |
4432 | On some Alpha implementations the resulting code may execute | |
4433 | significantly slower than the code generated by default. Since there | |
4434 | is very little code that depends on the @var{inexact flag}, you should | |
4435 | normally not specify this option. Other Alpha compilers call this | |
4436 | option @samp{-ieee_with_inexact}. | |
4437 | @c end changes to prevent overfull hboxes | |
4438 | ||
4439 | @item -mfp-trap-mode=@var{trap mode} | |
4440 | This option controls what floating-point related traps are enabled. | |
4441 | Other Alpha compilers call this option @samp{-fptm }@var{trap mode}. | |
4442 | The trap mode can be set to one of four values: | |
4443 | ||
4444 | @table @samp | |
4445 | @item n | |
4446 | This is the default (normal) setting. The only traps that are enabled | |
4447 | are the ones that cannot be disabled in software (e.g., division by zero | |
4448 | trap). | |
4449 | ||
4450 | @item u | |
4451 | In addition to the traps enabled by @samp{n}, underflow traps are enabled | |
4452 | as well. | |
4453 | ||
4454 | @item su | |
4455 | Like @samp{su}, but the instructions are marked to be safe for software | |
4456 | completion (see Alpha architecture manual for details). | |
4457 | ||
4458 | @item sui | |
4459 | Like @samp{su}, but inexact traps are enabled as well. | |
4460 | @end table | |
4461 | ||
4462 | @item -mfp-rounding-mode=@var{rounding mode} | |
4463 | Selects the IEEE rounding mode. Other Alpha compilers call this option | |
4464 | @samp{-fprm }@var{rounding mode}. The @var{rounding mode} can be one | |
4465 | of: | |
4466 | ||
4467 | @table @samp | |
4468 | @item n | |
4469 | Normal IEEE rounding mode. Floating point numbers are rounded towards | |
4470 | the nearest machine number or towards the even machine number in case | |
4471 | of a tie. | |
4472 | ||
4473 | @item m | |
4474 | Round towards minus infinity. | |
4475 | ||
4476 | @item c | |
4477 | Chopped rounding mode. Floating point numbers are rounded towards zero. | |
4478 | ||
4479 | @item d | |
4480 | Dynamic rounding mode. A field in the floating point control register | |
4481 | (@var{fpcr}, see Alpha architecture reference manual) controls the | |
4482 | rounding mode in effect. The C library initializes this register for | |
4483 | rounding towards plus infinity. Thus, unless your program modifies the | |
4484 | @var{fpcr}, @samp{d} corresponds to round towards plus infinity.@end table | |
4485 | ||
4486 | @item -mtrap-precision=@var{trap precision} | |
4487 | In the Alpha architecture, floating point traps are imprecise. This | |
4488 | means without software assistance it is impossible to recover from a | |
4489 | floating trap and program execution normally needs to be terminated. | |
4490 | GNU CC can generate code that can assist operating system trap handlers | |
4491 | in determining the exact location that caused a floating point trap. | |
4492 | Depending on the requirements of an application, different levels of | |
4493 | precisions can be selected: | |
4494 | ||
4495 | @table @samp | |
4496 | @item p | |
4497 | Program precision. This option is the default and means a trap handler | |
4498 | can only identify which program caused a floating point exception. | |
4499 | ||
4500 | @item f | |
4501 | Function precision. The trap handler can determine the function that | |
4502 | caused a floating point exception. | |
4503 | ||
4504 | @item i | |
4505 | Instruction precision. The trap handler can determine the exact | |
4506 | instruction that caused a floating point exception. | |
4507 | @end table | |
4508 | ||
4509 | Other Alpha compilers provide the equivalent options called | |
4510 | @samp{-scope_safe} and @samp{-resumption_safe}. | |
4511 | ||
4512 | @item -mieee-conformant | |
4513 | This option marks the generated code as IEEE conformant. You must not | |
4514 | use this option unless you also specify @samp{-mtrap-precision=i} and either | |
4515 | @samp{-mfp-trap-mode=su} or @samp{-mfp-trap-mode=sui}. Its only effect | |
4516 | is to emit the line @samp{.eflag 48} in the function prologue of the | |
4517 | generated assembly file. Under DEC Unix, this has the effect that | |
4518 | IEEE-conformant math library routines will be linked in. | |
4519 | ||
4520 | @item -mbuild-constants | |
4521 | Normally GNU CC examines a 32- or 64-bit integer constant to | |
4522 | see if it can construct it from smaller constants in two or three | |
4523 | instructions. If it cannot, it will output the constant as a literal and | |
4524 | generate code to load it from the data segement at runtime. | |
4525 | ||
4526 | Use this option to require GNU CC to construct @emph{all} integer constants | |
4527 | using code, even if it takes more instructions (the maximum is six). | |
4528 | ||
4529 | You would typically use this option to build a shared library dynamic | |
4530 | loader. Itself a shared library, it must relocate itself in memory | |
4531 | before it can find the variables and constants in its own data segment. | |
4532 | @end table | |
4533 | ||
4534 | @node Clipper Options | |
4535 | @subsection Clipper Options | |
4536 | ||
4537 | These @samp{-m} options are defined for the Clipper implementations: | |
4538 | ||
4539 | @table @code | |
4540 | @item -mc300 | |
4541 | Produce code for a C300 Clipper processor. This is the default. | |
4542 | ||
4543 | @itemx -mc400 | |
4544 | Produce code for a C400 Clipper processor i.e. use floating point | |
4545 | registers f8..f15. | |
4546 | @end table | |
4547 | ||
4548 | @node H8/300 Options | |
4549 | @subsection H8/300 Options | |
4550 | ||
4551 | These @samp{-m} options are defined for the H8/300 implementations: | |
4552 | ||
4553 | @table @code | |
4554 | @item -mrelax | |
4555 | Shorten some address references at link time, when possible; uses the | |
4556 | linker option @samp{-relax}. @xref{H8/300,, @code{ld} and the H8/300, | |
4557 | ld.info, Using ld}, for a fuller description. | |
4558 | ||
4559 | @item -mh | |
4560 | Generate code for the H8/300H. | |
4561 | ||
4562 | @item -mint32 | |
4563 | Make @code{int} data 32 bits by default. | |
4564 | ||
4565 | @item -malign-300 | |
4566 | On the h8/300h, use the same alignment rules as for the h8/300. | |
4567 | The default for the h8/300h is to align longs and floats on 4 byte boundaries. | |
4568 | @samp{-malign-300} causes them to be aligned on 2 byte boundaries. | |
4569 | This option has no effect on the h8/300. | |
4570 | @end table | |
4571 | ||
4572 | @node SH Options | |
4573 | @subsection SH Options | |
4574 | ||
4575 | These @samp{-m} options are defined for the SH implementations: | |
4576 | ||
4577 | @table @code | |
4578 | @item -m1 | |
4579 | Generate code for the SH1. | |
4580 | ||
4581 | @item -m2 | |
4582 | Generate code for the SH2. | |
4583 | ||
4584 | @item -m3 | |
4585 | Generate code for the SH3. | |
4586 | ||
4587 | @item -m3e | |
4588 | Generate code for the SH3e. | |
4589 | ||
4590 | @item -mb | |
4591 | Compile code for the processor in big endian mode. | |
4592 | ||
4593 | @item -ml | |
4594 | Compile code for the processor in little endian mode. | |
4595 | ||
4596 | @item -mrelax | |
4597 | Shorten some address references at link time, when possible; uses the | |
4598 | linker option @samp{-relax}. | |
4599 | @end table | |
4600 | ||
4601 | @node System V Options | |
4602 | @subsection Options for System V | |
4603 | ||
4604 | These additional options are available on System V Release 4 for | |
4605 | compatibility with other compilers on those systems: | |
4606 | ||
4607 | @table @code | |
4608 | @item -G | |
4609 | Create a shared object. | |
4610 | It is recommended that @samp{-symbolic} or @samp{-shared} be used instead. | |
4611 | ||
4612 | @item -Qy | |
4613 | Identify the versions of each tool used by the compiler, in a | |
4614 | @code{.ident} assembler directive in the output. | |
4615 | ||
4616 | @item -Qn | |
4617 | Refrain from adding @code{.ident} directives to the output file (this is | |
4618 | the default). | |
4619 | ||
4620 | @item -YP,@var{dirs} | |
4621 | Search the directories @var{dirs}, and no others, for libraries | |
4622 | specified with @samp{-l}. | |
4623 | ||
4624 | @item -Ym,@var{dir} | |
4625 | Look in the directory @var{dir} to find the M4 preprocessor. | |
4626 | The assembler uses this option. | |
4627 | @c This is supposed to go with a -Yd for predefined M4 macro files, but | |
4628 | @c the generic assembler that comes with Solaris takes just -Ym. | |
4629 | @end table | |
4630 | ||
4631 | @node Code Gen Options | |
4632 | @section Options for Code Generation Conventions | |
4633 | @cindex code generation conventions | |
4634 | @cindex options, code generation | |
4635 | @cindex run-time options | |
4636 | ||
4637 | These machine-independent options control the interface conventions | |
4638 | used in code generation. | |
4639 | ||
4640 | Most of them have both positive and negative forms; the negative form | |
4641 | of @samp{-ffoo} would be @samp{-fno-foo}. In the table below, only | |
4642 | one of the forms is listed---the one which is not the default. You | |
4643 | can figure out the other form by either removing @samp{no-} or adding | |
4644 | it. | |
4645 | ||
4646 | @table @code | |
4647 | @item -fpcc-struct-return | |
4648 | Return ``short'' @code{struct} and @code{union} values in memory like | |
4649 | longer ones, rather than in registers. This convention is less | |
4650 | efficient, but it has the advantage of allowing intercallability between | |
4651 | GNU CC-compiled files and files compiled with other compilers. | |
4652 | ||
4653 | The precise convention for returning structures in memory depends | |
4654 | on the target configuration macros. | |
4655 | ||
4656 | Short structures and unions are those whose size and alignment match | |
4657 | that of some integer type. | |
4658 | ||
4659 | @item -freg-struct-return | |
4660 | Use the convention that @code{struct} and @code{union} values are | |
4661 | returned in registers when possible. This is more efficient for small | |
4662 | structures than @samp{-fpcc-struct-return}. | |
4663 | ||
4664 | If you specify neither @samp{-fpcc-struct-return} nor its contrary | |
4665 | @samp{-freg-struct-return}, GNU CC defaults to whichever convention is | |
4666 | standard for the target. If there is no standard convention, GNU CC | |
4667 | defaults to @samp{-fpcc-struct-return}, except on targets where GNU CC | |
4668 | is the principal compiler. In those cases, we can choose the standard, | |
4669 | and we chose the more efficient register return alternative. | |
4670 | ||
4671 | @item -fshort-enums | |
4672 | Allocate to an @code{enum} type only as many bytes as it needs for the | |
4673 | declared range of possible values. Specifically, the @code{enum} type | |
4674 | will be equivalent to the smallest integer type which has enough room. | |
4675 | ||
4676 | @item -fshort-double | |
4677 | Use the same size for @code{double} as for @code{float}. | |
4678 | ||
4679 | @item -fshared-data | |
4680 | Requests that the data and non-@code{const} variables of this | |
4681 | compilation be shared data rather than private data. The distinction | |
4682 | makes sense only on certain operating systems, where shared data is | |
4683 | shared between processes running the same program, while private data | |
4684 | exists in one copy per process. | |
4685 | ||
4686 | @item -fno-common | |
4687 | Allocate even uninitialized global variables in the bss section of the | |
4688 | object file, rather than generating them as common blocks. This has the | |
4689 | effect that if the same variable is declared (without @code{extern}) in | |
4690 | two different compilations, you will get an error when you link them. | |
4691 | The only reason this might be useful is if you wish to verify that the | |
4692 | program will work on other systems which always work this way. | |
4693 | ||
4694 | @item -fno-ident | |
4695 | Ignore the @samp{#ident} directive. | |
4696 | ||
4697 | @item -fno-gnu-linker | |
4698 | Do not output global initializations (such as C++ constructors and | |
4699 | destructors) in the form used by the GNU linker (on systems where the GNU | |
4700 | linker is the standard method of handling them). Use this option when | |
4701 | you want to use a non-GNU linker, which also requires using the | |
4702 | @code{collect2} program to make sure the system linker includes | |
4703 | constructors and destructors. (@code{collect2} is included in the GNU CC | |
4704 | distribution.) For systems which @emph{must} use @code{collect2}, the | |
4705 | compiler driver @code{gcc} is configured to do this automatically. | |
4706 | ||
4707 | @item -finhibit-size-directive | |
4708 | Don't output a @code{.size} assembler directive, or anything else that | |
4709 | would cause trouble if the function is split in the middle, and the | |
4710 | two halves are placed at locations far apart in memory. This option is | |
4711 | used when compiling @file{crtstuff.c}; you should not need to use it | |
4712 | for anything else. | |
4713 | ||
4714 | @item -fverbose-asm | |
4715 | Put extra commentary information in the generated assembly code to | |
4716 | make it more readable. This option is generally only of use to those | |
4717 | who actually need to read the generated assembly code (perhaps while | |
4718 | debugging the compiler itself). | |
4719 | ||
4720 | @samp{-fverbose-asm} is the default. @samp{-fno-verbose-asm} causes the | |
4721 | extra information to be omitted and is useful when comparing two assembler | |
4722 | files. | |
4723 | ||
4724 | @item -fvolatile | |
4725 | Consider all memory references through pointers to be volatile. | |
4726 | ||
4727 | @item -fvolatile-global | |
4728 | Consider all memory references to extern and global data items to | |
4729 | be volatile. | |
4730 | ||
4731 | @item -fpic | |
4732 | @cindex global offset table | |
4733 | @cindex PIC | |
4734 | Generate position-independent code (PIC) suitable for use in a shared | |
4735 | library, if supported for the target machine. Such code accesses all | |
4736 | constant addresses through a global offset table (GOT). If the GOT size | |
4737 | for the linked executable exceeds a machine-specific maximum size, you | |
4738 | get an error message from the linker indicating that @samp{-fpic} does | |
4739 | not work; in that case, recompile with @samp{-fPIC} instead. (These | |
4740 | maximums are 16k on the m88k, 8k on the Sparc, and 32k on the m68k and | |
4741 | RS/6000. The 386 has no such limit.) | |
4742 | ||
4743 | Position-independent code requires special support, and therefore works | |
4744 | only on certain machines. For the 386, GNU CC supports PIC for System V | |
4745 | but not for the Sun 386i. Code generated for the IBM RS/6000 is always | |
4746 | position-independent. | |
4747 | ||
4748 | The GNU assembler does not fully support PIC. Currently, you must use | |
4749 | some other assembler in order for PIC to work. We would welcome | |
4750 | volunteers to upgrade GAS to handle this; the first part of the job is | |
4751 | to figure out what the assembler must do differently. | |
4752 | ||
4753 | @item -fPIC | |
4754 | If supported for the target machine, emit position-independent code, | |
4755 | suitable for dynamic linking and avoiding any limit on the size of the | |
4756 | global offset table. This option makes a difference on the m68k, m88k, | |
4757 | and the Sparc. | |
4758 | ||
4759 | Position-independent code requires special support, and therefore works | |
4760 | only on certain machines. | |
4761 | ||
4762 | @item -ffixed-@var{reg} | |
4763 | Treat the register named @var{reg} as a fixed register; generated code | |
4764 | should never refer to it (except perhaps as a stack pointer, frame | |
4765 | pointer or in some other fixed role). | |
4766 | ||
4767 | @var{reg} must be the name of a register. The register names accepted | |
4768 | are machine-specific and are defined in the @code{REGISTER_NAMES} | |
4769 | macro in the machine description macro file. | |
4770 | ||
4771 | This flag does not have a negative form, because it specifies a | |
4772 | three-way choice. | |
4773 | ||
4774 | @item -fcall-used-@var{reg} | |
4775 | Treat the register named @var{reg} as an allocatable register that is | |
4776 | clobbered by function calls. It may be allocated for temporaries or | |
4777 | variables that do not live across a call. Functions compiled this way | |
4778 | will not save and restore the register @var{reg}. | |
4779 | ||
4780 | Use of this flag for a register that has a fixed pervasive role in the | |
4781 | machine's execution model, such as the stack pointer or frame pointer, | |
4782 | will produce disastrous results. | |
4783 | ||
4784 | This flag does not have a negative form, because it specifies a | |
4785 | three-way choice. | |
4786 | ||
4787 | @item -fcall-saved-@var{reg} | |
4788 | Treat the register named @var{reg} as an allocatable register saved by | |
4789 | functions. It may be allocated even for temporaries or variables that | |
4790 | live across a call. Functions compiled this way will save and restore | |
4791 | the register @var{reg} if they use it. | |
4792 | ||
4793 | Use of this flag for a register that has a fixed pervasive role in the | |
4794 | machine's execution model, such as the stack pointer or frame pointer, | |
4795 | will produce disastrous results. | |
4796 | ||
4797 | A different sort of disaster will result from the use of this flag for | |
4798 | a register in which function values may be returned. | |
4799 | ||
4800 | This flag does not have a negative form, because it specifies a | |
4801 | three-way choice. | |
4802 | ||
4803 | @item -fpack-struct | |
4804 | Pack all structure members together without holes. Usually you would | |
4805 | not want to use this option, since it makes the code suboptimal, and | |
4806 | the offsets of structure members won't agree with system libraries. | |
4807 | ||
4808 | @item +e0 | |
4809 | @itemx +e1 | |
4810 | Control whether virtual function definitions in classes are used to | |
4811 | generate code, or only to define interfaces for their callers. (C++ | |
4812 | only). | |
4813 | ||
4814 | These options are provided for compatibility with @code{cfront} 1.x | |
4815 | usage; the recommended alternative GNU C++ usage is in flux. @xref{C++ | |
4816 | Interface,,Declarations and Definitions in One Header}. | |
4817 | ||
4818 | With @samp{+e0}, virtual function definitions in classes are declared | |
4819 | @code{extern}; the declaration is used only as an interface | |
4820 | specification, not to generate code for the virtual functions (in this | |
4821 | compilation). | |
4822 | ||
4823 | With @samp{+e1}, G++ actually generates the code implementing virtual | |
4824 | functions defined in the code, and makes them publicly visible. | |
4825 | @end table | |
4826 | ||
4827 | @node Environment Variables | |
4828 | @section Environment Variables Affecting GNU CC | |
4829 | @cindex environment variables | |
4830 | ||
4831 | This section describes several environment variables that affect how GNU | |
4832 | CC operates. They work by specifying directories or prefixes to use | |
4833 | when searching for various kinds of files. | |
4834 | ||
4835 | @ifclear INTERNALS | |
4836 | Note that you can also specify places to search using options such as | |
4837 | @samp{-B}, @samp{-I} and @samp{-L} (@pxref{Directory Options}). These | |
4838 | take precedence over places specified using environment variables, which | |
4839 | in turn take precedence over those specified by the configuration of GNU | |
4840 | CC. | |
4841 | @end ifclear | |
4842 | @ifset INTERNALS | |
4843 | Note that you can also specify places to search using options such as | |
4844 | @samp{-B}, @samp{-I} and @samp{-L} (@pxref{Directory Options}). These | |
4845 | take precedence over places specified using environment variables, which | |
4846 | in turn take precedence over those specified by the configuration of GNU | |
4847 | CC. @xref{Driver}. | |
4848 | @end ifset | |
4849 | ||
4850 | @table @code | |
4851 | @item TMPDIR | |
4852 | @findex TMPDIR | |
4853 | If @code{TMPDIR} is set, it specifies the directory to use for temporary | |
4854 | files. GNU CC uses temporary files to hold the output of one stage of | |
4855 | compilation which is to be used as input to the next stage: for example, | |
4856 | the output of the preprocessor, which is the input to the compiler | |
4857 | proper. | |
4858 | ||
4859 | @item GCC_EXEC_PREFIX | |
4860 | @findex GCC_EXEC_PREFIX | |
4861 | If @code{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the | |
4862 | names of the subprograms executed by the compiler. No slash is added | |
4863 | when this prefix is combined with the name of a subprogram, but you can | |
4864 | specify a prefix that ends with a slash if you wish. | |
4865 | ||
4866 | If GNU CC cannot find the subprogram using the specified prefix, it | |
4867 | tries looking in the usual places for the subprogram. | |
4868 | ||
4869 | The default value of @code{GCC_EXEC_PREFIX} is | |
4870 | @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value | |
4871 | of @code{prefix} when you ran the @file{configure} script. | |
4872 | ||
4873 | Other prefixes specified with @samp{-B} take precedence over this prefix. | |
4874 | ||
4875 | This prefix is also used for finding files such as @file{crt0.o} that are | |
4876 | used for linking. | |
4877 | ||
4878 | In addition, the prefix is used in an unusual way in finding the | |
4879 | directories to search for header files. For each of the standard | |
4880 | directories whose name normally begins with @samp{/usr/local/lib/gcc-lib} | |
4881 | (more precisely, with the value of @code{GCC_INCLUDE_DIR}), GNU CC tries | |
4882 | replacing that beginning with the specified prefix to produce an | |
4883 | alternate directory name. Thus, with @samp{-Bfoo/}, GNU CC will search | |
4884 | @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}. | |
4885 | These alternate directories are searched first; the standard directories | |
4886 | come next. | |
4887 | ||
4888 | @item COMPILER_PATH | |
4889 | @findex COMPILER_PATH | |
4890 | The value of @code{COMPILER_PATH} is a colon-separated list of | |
4891 | directories, much like @code{PATH}. GNU CC tries the directories thus | |
4892 | specified when searching for subprograms, if it can't find the | |
4893 | subprograms using @code{GCC_EXEC_PREFIX}. | |
4894 | ||
4895 | @item LIBRARY_PATH | |
4896 | @findex LIBRARY_PATH | |
4897 | The value of @code{LIBRARY_PATH} is a colon-separated list of | |
4898 | directories, much like @code{PATH}. When configured as a native compiler, | |
4899 | GNU CC tries the directories thus specified when searching for special | |
4900 | linker files, if it can't find them using @code{GCC_EXEC_PREFIX}. Linking | |
4901 | using GNU CC also uses these directories when searching for ordinary | |
4902 | libraries for the @samp{-l} option (but directories specified with | |
4903 | @samp{-L} come first). | |
4904 | ||
4905 | @item C_INCLUDE_PATH | |
4906 | @itemx CPLUS_INCLUDE_PATH | |
4907 | @itemx OBJC_INCLUDE_PATH | |
4908 | @findex C_INCLUDE_PATH | |
4909 | @findex CPLUS_INCLUDE_PATH | |
4910 | @findex OBJC_INCLUDE_PATH | |
4911 | @c @itemx OBJCPLUS_INCLUDE_PATH | |
4912 | These environment variables pertain to particular languages. Each | |
4913 | variable's value is a colon-separated list of directories, much like | |
4914 | @code{PATH}. When GNU CC searches for header files, it tries the | |
4915 | directories listed in the variable for the language you are using, after | |
4916 | the directories specified with @samp{-I} but before the standard header | |
4917 | file directories. | |
4918 | ||
4919 | @item DEPENDENCIES_OUTPUT | |
4920 | @findex DEPENDENCIES_OUTPUT | |
4921 | @cindex dependencies for make as output | |
4922 | If this variable is set, its value specifies how to output dependencies | |
4923 | for Make based on the header files processed by the compiler. This | |
4924 | output looks much like the output from the @samp{-M} option | |
4925 | (@pxref{Preprocessor Options}), but it goes to a separate file, and is | |
4926 | in addition to the usual results of compilation. | |
4927 | ||
4928 | The value of @code{DEPENDENCIES_OUTPUT} can be just a file name, in | |
4929 | which case the Make rules are written to that file, guessing the target | |
4930 | name from the source file name. Or the value can have the form | |
4931 | @samp{@var{file} @var{target}}, in which case the rules are written to | |
4932 | file @var{file} using @var{target} as the target name. | |
4933 | @end table | |
4934 | ||
4935 | @node Running Protoize | |
4936 | @section Running Protoize | |
4937 | ||
4938 | The program @code{protoize} is an optional part of GNU C. You can use | |
4939 | it to add prototypes to a program, thus converting the program to ANSI | |
4940 | C in one respect. The companion program @code{unprotoize} does the | |
4941 | reverse: it removes argument types from any prototypes that are found. | |
4942 | ||
4943 | When you run these programs, you must specify a set of source files as | |
4944 | command line arguments. The conversion programs start out by compiling | |
4945 | these files to see what functions they define. The information gathered | |
4946 | about a file @var{foo} is saved in a file named @file{@var{foo}.X}. | |
4947 | ||
4948 | After scanning comes actual conversion. The specified files are all | |
4949 | eligible to be converted; any files they include (whether sources or | |
4950 | just headers) are eligible as well. | |
4951 | ||
4952 | But not all the eligible files are converted. By default, | |
4953 | @code{protoize} and @code{unprotoize} convert only source and header | |
4954 | files in the current directory. You can specify additional directories | |
4955 | whose files should be converted with the @samp{-d @var{directory}} | |
4956 | option. You can also specify particular files to exclude with the | |
4957 | @samp{-x @var{file}} option. A file is converted if it is eligible, its | |
4958 | directory name matches one of the specified directory names, and its | |
4959 | name within the directory has not been excluded. | |
4960 | ||
4961 | Basic conversion with @code{protoize} consists of rewriting most | |
4962 | function definitions and function declarations to specify the types of | |
4963 | the arguments. The only ones not rewritten are those for varargs | |
4964 | functions. | |
4965 | ||
4966 | @code{protoize} optionally inserts prototype declarations at the | |
4967 | beginning of the source file, to make them available for any calls that | |
4968 | precede the function's definition. Or it can insert prototype | |
4969 | declarations with block scope in the blocks where undeclared functions | |
4970 | are called. | |
4971 | ||
4972 | Basic conversion with @code{unprotoize} consists of rewriting most | |
4973 | function declarations to remove any argument types, and rewriting | |
4974 | function definitions to the old-style pre-ANSI form. | |
4975 | ||
4976 | Both conversion programs print a warning for any function declaration or | |
4977 | definition that they can't convert. You can suppress these warnings | |
4978 | with @samp{-q}. | |
4979 | ||
4980 | The output from @code{protoize} or @code{unprotoize} replaces the | |
4981 | original source file. The original file is renamed to a name ending | |
4982 | with @samp{.save}. If the @samp{.save} file already exists, then | |
4983 | the source file is simply discarded. | |
4984 | ||
4985 | @code{protoize} and @code{unprotoize} both depend on GNU CC itself to | |
4986 | scan the program and collect information about the functions it uses. | |
4987 | So neither of these programs will work until GNU CC is installed. | |
4988 | ||
4989 | Here is a table of the options you can use with @code{protoize} and | |
4990 | @code{unprotoize}. Each option works with both programs unless | |
4991 | otherwise stated. | |
4992 | ||
4993 | @table @code | |
4994 | @item -B @var{directory} | |
4995 | Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the | |
4996 | usual directory (normally @file{/usr/local/lib}). This file contains | |
4997 | prototype information about standard system functions. This option | |
4998 | applies only to @code{protoize}. | |
4999 | ||
5000 | @item -c @var{compilation-options} | |
5001 | Use @var{compilation-options} as the options when running @code{gcc} to | |
5002 | produce the @samp{.X} files. The special option @samp{-aux-info} is | |
5003 | always passed in addition, to tell @code{gcc} to write a @samp{.X} file. | |
5004 | ||
5005 | Note that the compilation options must be given as a single argument to | |
5006 | @code{protoize} or @code{unprotoize}. If you want to specify several | |
5007 | @code{gcc} options, you must quote the entire set of compilation options | |
5008 | to make them a single word in the shell. | |
5009 | ||
5010 | There are certain @code{gcc} arguments that you cannot use, because they | |
5011 | would produce the wrong kind of output. These include @samp{-g}, | |
5012 | @samp{-O}, @samp{-c}, @samp{-S}, and @samp{-o} If you include these in | |
5013 | the @var{compilation-options}, they are ignored. | |
5014 | ||
5015 | @item -C | |
5016 | Rename files to end in @samp{.C} instead of @samp{.c}. | |
5017 | This is convenient if you are converting a C program to C++. | |
5018 | This option applies only to @code{protoize}. | |
5019 | ||
5020 | @item -g | |
5021 | Add explicit global declarations. This means inserting explicit | |
5022 | declarations at the beginning of each source file for each function | |
5023 | that is called in the file and was not declared. These declarations | |
5024 | precede the first function definition that contains a call to an | |
5025 | undeclared function. This option applies only to @code{protoize}. | |
5026 | ||
5027 | @item -i @var{string} | |
5028 | Indent old-style parameter declarations with the string @var{string}. | |
5029 | This option applies only to @code{protoize}. | |
5030 | ||
5031 | @code{unprotoize} converts prototyped function definitions to old-style | |
5032 | function definitions, where the arguments are declared between the | |
5033 | argument list and the initial @samp{@{}. By default, @code{unprotoize} | |
5034 | uses five spaces as the indentation. If you want to indent with just | |
5035 | one space instead, use @samp{-i " "}. | |
5036 | ||
5037 | @item -k | |
5038 | Keep the @samp{.X} files. Normally, they are deleted after conversion | |
5039 | is finished. | |
5040 | ||
5041 | @item -l | |
5042 | Add explicit local declarations. @code{protoize} with @samp{-l} inserts | |
5043 | a prototype declaration for each function in each block which calls the | |
5044 | function without any declaration. This option applies only to | |
5045 | @code{protoize}. | |
5046 | ||
5047 | @item -n | |
5048 | Make no real changes. This mode just prints information about the conversions | |
5049 | that would have been done without @samp{-n}. | |
5050 | ||
5051 | @item -N | |
5052 | Make no @samp{.save} files. The original files are simply deleted. | |
5053 | Use this option with caution. | |
5054 | ||
5055 | @item -p @var{program} | |
5056 | Use the program @var{program} as the compiler. Normally, the name | |
5057 | @file{gcc} is used. | |
5058 | ||
5059 | @item -q | |
5060 | Work quietly. Most warnings are suppressed. | |
5061 | ||
5062 | @item -v | |
5063 | Print the version number, just like @samp{-v} for @code{gcc}. | |
5064 | @end table | |
5065 | ||
5066 | If you need special compiler options to compile one of your program's | |
5067 | source files, then you should generate that file's @samp{.X} file | |
5068 | specially, by running @code{gcc} on that source file with the | |
5069 | appropriate options and the option @samp{-aux-info}. Then run | |
5070 | @code{protoize} on the entire set of files. @code{protoize} will use | |
5071 | the existing @samp{.X} file because it is newer than the source file. | |
5072 | For example: | |
5073 | ||
5074 | @example | |
5075 | gcc -Dfoo=bar file1.c -aux-info | |
5076 | protoize *.c | |
5077 | @end example | |
5078 | ||
5079 | @noindent | |
5080 | You need to include the special files along with the rest in the | |
5081 | @code{protoize} command, even though their @samp{.X} files already | |
5082 | exist, because otherwise they won't get converted. | |
5083 | ||
5084 | @xref{Protoize Caveats}, for more information on how to use | |
5085 | @code{protoize} successfully. | |
5086 |