1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
8 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
9 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.1 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), g77(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{g77}, @file{as},
48 @file{ld}, @file{binutils} and @file{gdb}.
51 For instructions on reporting bugs, see
52 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
53 script to report bugs is recommended.
56 See the Info entry for @file{gcc}, or
57 @w{@uref{http://gcc.gnu.org/thanks.html}}, for contributors to GCC@.
62 @chapter GCC Command Options
63 @cindex GCC command options
64 @cindex command options
65 @cindex options, GCC command
67 @c man begin DESCRIPTION
69 When you invoke GCC, it normally does preprocessing, compilation,
70 assembly and linking. The ``overall options'' allow you to stop this
71 process at an intermediate stage. For example, the @option{-c} option
72 says not to run the linker. Then the output consists of object files
73 output by the assembler.
75 Other options are passed on to one stage of processing. Some options
76 control the preprocessor and others the compiler itself. Yet other
77 options control the assembler and linker; most of these are not
78 documented here, since you rarely need to use any of them.
80 @cindex C compilation options
81 Most of the command line options that you can use with GCC are useful
82 for C programs; when an option is only useful with another language
83 (usually C++), the explanation says so explicitly. If the description
84 for a particular option does not mention a source language, you can use
85 that option with all supported languages.
87 @cindex C++ compilation options
88 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
89 options for compiling C++ programs.
91 @cindex grouping options
92 @cindex options, grouping
93 The @command{gcc} program accepts options and file names as operands. Many
94 options have multi-letter names; therefore multiple single-letter options
95 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
98 @cindex order of options
99 @cindex options, order
100 You can mix options and other arguments. For the most part, the order
101 you use doesn't matter. Order does matter when you use several options
102 of the same kind; for example, if you specify @option{-L} more than once,
103 the directories are searched in the order specified.
105 Many options have long names starting with @samp{-f} or with
106 @samp{-W}---for example, @option{-fforce-mem},
107 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
108 these have both positive and negative forms; the negative form of
109 @option{-ffoo} would be @option{-fno-foo}. This manual documents
110 only one of these two forms, whichever one is not the default.
114 @xref{Option Index}, for an index to GCC's options.
117 * Option Summary:: Brief list of all options, without explanations.
118 * Overall Options:: Controlling the kind of output:
119 an executable, object files, assembler files,
120 or preprocessed source.
121 * Invoking G++:: Compiling C++ programs.
122 * C Dialect Options:: Controlling the variant of C language compiled.
123 * C++ Dialect Options:: Variations on C++.
124 * Objective-C Dialect Options:: Variations on Objective-C.
125 * Language Independent Options:: Controlling how diagnostics should be
127 * Warning Options:: How picky should the compiler be?
128 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
129 * Optimize Options:: How much optimization?
130 * Preprocessor Options:: Controlling header files and macro definitions.
131 Also, getting dependency information for Make.
132 * Assembler Options:: Passing options to the assembler.
133 * Link Options:: Specifying libraries and so on.
134 * Directory Options:: Where to find header files and libraries.
135 Where to find the compiler executable files.
136 * Spec Files:: How to pass switches to sub-processes.
137 * Target Options:: Running a cross-compiler, or an old version of GCC.
138 * Submodel Options:: Specifying minor hardware or convention variations,
139 such as 68010 vs 68020.
140 * Code Gen Options:: Specifying conventions for function calls, data layout
142 * Environment Variables:: Env vars that affect GCC.
143 * Running Protoize:: Automatically adding or removing function prototypes.
149 @section Option Summary
151 Here is a summary of all the options, grouped by type. Explanations are
152 in the following sections.
155 @item Overall Options
156 @xref{Overall Options,,Options Controlling the Kind of Output}.
158 -c -S -E -o @var{file} -pipe -pass-exit-codes -x @var{language} @gol
159 -v --target-help --help}
161 @item C Language Options
162 @xref{C Dialect Options,,Options Controlling C Dialect}.
164 -ansi -std=@var{standard} -aux-info @var{filename} @gol
165 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
166 -fhosted -ffreestanding @gol
167 -trigraphs -traditional -traditional-cpp @gol
168 -fallow-single-precision -fcond-mismatch @gol
169 -fsigned-bitfields -fsigned-char @gol
170 -funsigned-bitfields -funsigned-char @gol
171 -fwritable-strings -fshort-wchar}
173 @item C++ Language Options
174 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
176 -fno-access-control -fcheck-new -fconserve-space @gol
177 -fno-const-strings -fdollars-in-identifiers @gol
178 -fno-elide-constructors @gol
179 -fno-enforce-eh-specs -fexternal-templates @gol
180 -falt-external-templates @gol
181 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
182 -fno-implicit-templates @gol
183 -fno-implicit-inline-templates @gol
184 -fno-implement-inlines -fms-extensions @gol
185 -fno-nonansi-builtins -fno-operator-names @gol
186 -fno-optional-diags -fpermissive @gol
187 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
188 -fuse-cxa-atexit -fvtable-gc -fno-weak -nostdinc++ @gol
189 -fno-default-inline -Wctor-dtor-privacy @gol
190 -Wnon-virtual-dtor -Wreorder @gol
191 -Weffc++ -Wno-deprecated @gol
192 -Wno-non-template-friend -Wold-style-cast @gol
193 -Woverloaded-virtual -Wno-pmf-conversions @gol
194 -Wsign-promo -Wsynth}
196 @item Objective-C Language Options
197 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
199 -fconstant-string-class=@var{class-name} @gol
200 -fgnu-runtime -fnext-runtime -gen-decls @gol
201 -Wno-protocol -Wselector}
203 @item Language Independent Options
204 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
206 -fmessage-length=@var{n} @gol
207 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
209 @item Warning Options
210 @xref{Warning Options,,Options to Request or Suppress Warnings}.
212 -fsyntax-only -pedantic -pedantic-errors @gol
213 -w -W -Wall -Waggregate-return @gol
214 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
215 -Wconversion -Wdisabled-optimization -Wdiv-by-zero -Werror @gol
216 -Wfloat-equal -Wformat -Wformat=2 @gol
217 -Wformat-nonliteral -Wformat-security @gol
218 -Wimplicit -Wimplicit-int @gol
219 -Wimplicit-function-declaration @gol
220 -Werror-implicit-function-declaration @gol
221 -Wimport -Winline @gol
222 -Wlarger-than-@var{len} -Wlong-long @gol
223 -Wmain -Wmissing-braces -Wmissing-declarations @gol
224 -Wmissing-format-attribute -Wmissing-noreturn @gol
225 -Wmultichar -Wno-format-extra-args -Wno-format-y2k @gol
226 -Wno-import -Wpacked -Wpadded @gol
227 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
228 -Wreturn-type -Wsequence-point -Wshadow @gol
229 -Wsign-compare -Wswitch -Wsystem-headers @gol
230 -Wtrigraphs -Wundef -Wuninitialized @gol
231 -Wunknown-pragmas -Wunreachable-code @gol
232 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
233 -Wunused-value -Wunused-variable -Wwrite-strings}
235 @item C-only Warning Options
237 -Wbad-function-cast -Wmissing-prototypes -Wnested-externs @gol
238 -Wstrict-prototypes -Wtraditional}
240 @item Debugging Options
241 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
243 -d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
245 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
246 -fdump-tree-original@r{[}-@var{n}@r{]} -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
247 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
248 -fmem-report -fpretend-float @gol
249 -fprofile-arcs -ftest-coverage -ftime-report @gol
250 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
251 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
252 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
253 -print-multi-directory -print-multi-lib @gol
254 -print-prog-name=@var{program} -print-search-dirs -Q @gol
257 @item Optimization Options
258 @xref{Optimize Options,,Options that Control Optimization}.
260 -falign-functions=@var{n} -falign-jumps=@var{n} @gol
261 -falign-labels=@var{n} -falign-loops=@var{n} @gol
262 -fbranch-probabilities -fcaller-saves @gol
263 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
264 -fdelayed-branch -fdelete-null-pointer-checks @gol
265 -fexpensive-optimizations -ffast-math -ffloat-store @gol
266 -fforce-addr -fforce-mem -ffunction-sections @gol
267 -fgcse -fgcse-lm -fgcse-sm @gol
268 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
269 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
270 -fmove-all-movables -fno-default-inline -fno-defer-pop @gol
271 -fno-function-cse -fno-guess-branch-probability @gol
272 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
273 -funsafe-math-optimizations -fno-trapping-math @gol
274 -fomit-frame-pointer -foptimize-register-move @gol
275 -foptimize-sibling-calls -freduce-all-givs @gol
276 -fregmove -frename-registers @gol
277 -frerun-cse-after-loop -frerun-loop-opt @gol
278 -fschedule-insns -fschedule-insns2 @gol
279 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
280 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
281 -funroll-all-loops -funroll-loops @gol
282 --param @var{name}=@var{value}
283 -O -O0 -O1 -O2 -O3 -Os}
285 @item Preprocessor Options
286 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
288 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
289 -C -dD -dI -dM -dN @gol
290 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
291 -idirafter @var{dir} @gol
292 -include @var{file} -imacros @var{file} @gol
293 -iprefix @var{file} -iwithprefix @var{dir} @gol
294 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
295 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
296 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
298 @item Assembler Option
299 @xref{Assembler Options,,Passing Options to the Assembler}.
304 @xref{Link Options,,Options for Linking}.
306 @var{object-file-name} -l@var{library} @gol
307 -nostartfiles -nodefaultlibs -nostdlib @gol
308 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
309 -Wl,@var{option} -Xlinker @var{option} @gol
312 @item Directory Options
313 @xref{Directory Options,,Options for Directory Search}.
315 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
318 @c I wrote this xref this way to avoid overfull hbox. -- rms
319 @xref{Target Options}.
321 -b @var{machine} -V @var{version}}
323 @item Machine Dependent Options
324 @xref{Submodel Options,,Hardware Models and Configurations}.
325 @emph{M680x0 Options}
327 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
328 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
329 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
330 -malign-int -mstrict-align}
332 @emph{M68hc1x Options}
334 -m6811 -m6812 -m68hc11 -m68hc12 @gol
335 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
343 -mcpu=@var{cpu-type} @gol
344 -mtune=@var{cpu-type} @gol
345 -mcmodel=@var{code-model} @gol
347 -mapp-regs -mbroken-saverestore -mcypress @gol
348 -mepilogue -mfaster-structs -mflat @gol
349 -mfpu -mhard-float -mhard-quad-float @gol
350 -mimpure-text -mlive-g0 -mno-app-regs @gol
351 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu @gol
352 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
353 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
354 -msupersparc -munaligned-doubles -mv8}
356 @emph{Convex Options}
358 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
359 -margcount -mnoargcount @gol
360 -mlong32 -mlong64 @gol
361 -mvolatile-cache -mvolatile-nocache}
363 @emph{AMD29K Options}
365 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
366 -mlarge -mnormal -msmall @gol
367 -mkernel-registers -mno-reuse-arg-regs @gol
368 -mno-stack-check -mno-storem-bug @gol
369 -mreuse-arg-regs -msoft-float -mstack-check @gol
370 -mstorem-bug -muser-registers}
374 -mapcs-frame -mno-apcs-frame @gol
375 -mapcs-26 -mapcs-32 @gol
376 -mapcs-stack-check -mno-apcs-stack-check @gol
377 -mapcs-float -mno-apcs-float @gol
378 -mapcs-reentrant -mno-apcs-reentrant @gol
379 -msched-prolog -mno-sched-prolog @gol
380 -mlittle-endian -mbig-endian -mwords-little-endian @gol
381 -malignment-traps -mno-alignment-traps @gol
382 -msoft-float -mhard-float -mfpe @gol
383 -mthumb-interwork -mno-thumb-interwork @gol
384 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
385 -mstructure-size-boundary=@var{n} @gol
386 -mbsd -mxopen -mno-symrename @gol
387 -mabort-on-noreturn @gol
388 -mlong-calls -mno-long-calls @gol
389 -msingle-pic-base -mno-single-pic-base @gol
390 -mpic-register=@var{reg} @gol
391 -mnop-fun-dllimport @gol
392 -mpoke-function-name @gol
394 -mtpcs-frame -mtpcs-leaf-frame @gol
395 -mcaller-super-interworking -mcallee-super-interworking }
397 @emph{MN10200 Options}
401 @emph{MN10300 Options}
403 -mmult-bug -mno-mult-bug @gol
404 -mam33 -mno-am33 @gol
407 @emph{M32R/D Options}
409 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
414 -m88000 -m88100 -m88110 -mbig-pic @gol
415 -mcheck-zero-division -mhandle-large-shift @gol
416 -midentify-revision -mno-check-zero-division @gol
417 -mno-ocs-debug-info -mno-ocs-frame-position @gol
418 -mno-optimize-arg-area -mno-serialize-volatile @gol
419 -mno-underscores -mocs-debug-info @gol
420 -mocs-frame-position -moptimize-arg-area @gol
421 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
422 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
423 -mversion-03.00 -mwarn-passed-structs}
425 @emph{RS/6000 and PowerPC Options}
427 -mcpu=@var{cpu-type} @gol
428 -mtune=@var{cpu-type} @gol
429 -mpower -mno-power -mpower2 -mno-power2 @gol
430 -mpowerpc -mpowerpc64 -mno-powerpc @gol
431 -maltivec -mno-altivec @gol
432 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
433 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
434 -mnew-mnemonics -mold-mnemonics @gol
435 -mfull-toc -mminimal-toc -mno-fop-in-toc -mno-sum-in-toc @gol
436 -m64 -m32 -mxl-call -mno-xl-call -mthreads -mpe @gol
437 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
438 -mstring -mno-string -mupdate -mno-update @gol
439 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
440 -mstrict-align -mno-strict-align -mrelocatable @gol
441 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
442 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
443 -mcall-aix -mcall-sysv -mcall-netbsd @gol
444 -maix-struct-return -msvr4-struct-return
446 -mprototype -mno-prototype @gol
447 -msim -mmvme -mads -myellowknife -memb -msdata @gol
448 -msdata=@var{opt} -mvxworks -G @var{num}}
452 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
453 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
454 -mminimum-fp-blocks -mnohc-struct-return}
458 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
459 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
460 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
461 -mgas -mgp32 -mgp64 @gol
462 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
463 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
464 -mmips-as -mmips-tfile -mno-abicalls @gol
465 -mno-embedded-data -mno-uninit-const-in-rodata @gol
466 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
467 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
468 -mrnames -msoft-float @gol
469 -m4650 -msingle-float -mmad @gol
470 -mstats -EL -EB -G @var{num} -nocpp @gol
471 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
472 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
474 @emph{i386 and x86-64 Options}
476 -mcpu=@var{cpu-type} -march=@var{cpu-type} @gol
477 -mintel-syntax -mieee-fp -mno-fancy-math-387 @gol
478 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
479 -mno-wide-multiply -mrtd -malign-double @gol
480 -mpreferred-stack-boundary=@var{num} @gol
481 -mmmx -msse -m3dnow @gol
482 -mthreads -mno-align-stringops -minline-all-stringops @gol
483 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
484 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
490 -march=@var{architecture-type} @gol
491 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
492 -mfast-indirect-calls -mgas -mjump-in-delay @gol
493 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
494 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
495 -mno-jump-in-delay -mno-long-load-store @gol
496 -mno-portable-runtime -mno-soft-float @gol
497 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
498 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
499 -mschedule=@var{cpu-type} -mspace-regs}
501 @emph{Intel 960 Options}
503 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
504 -mcode-align -mcomplex-addr -mleaf-procedures @gol
505 -mic-compat -mic2.0-compat -mic3.0-compat @gol
506 -mintel-asm -mno-clean-linkage -mno-code-align @gol
507 -mno-complex-addr -mno-leaf-procedures @gol
508 -mno-old-align -mno-strict-align -mno-tail-call @gol
509 -mnumerics -mold-align -msoft-float -mstrict-align @gol
512 @emph{DEC Alpha Options}
514 -mfp-regs -mno-fp-regs -mno-soft-float -msoft-float @gol
515 -malpha-as -mgas @gol
516 -mieee -mieee-with-inexact -mieee-conformant @gol
517 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
518 -mtrap-precision=@var{mode} -mbuild-constants @gol
519 -mcpu=@var{cpu-type} @gol
520 -mbwx -mno-bwx -mcix -mno-cix -mmax -mno-max @gol
521 -mmemory-latency=@var{time}}
523 @emph{Clipper Options}
527 @emph{H8/300 Options}
529 -mrelax -mh -ms -mint32 -malign-300}
533 -m1 -m2 -m3 -m3e @gol
534 -m4-nofpu -m4-single-only -m4-single -m4 @gol
535 -mb -ml -mdalign -mrelax @gol
536 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
537 -mieee -misize -mpadstruct -mspace @gol
538 -mprefergot -musermode}
540 @emph{System V Options}
542 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
547 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
548 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
550 @emph{TMS320C3x/C4x Options}
552 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
553 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
554 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
555 -mparallel-insns -mparallel-mpy -mpreserve-float}
559 -mlong-calls -mno-long-calls -mep -mno-ep @gol
560 -mprolog-function -mno-prolog-function -mspace @gol
561 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
566 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
567 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
568 -mregparam -mnoregparam -msb -mnosb @gol
569 -mbitfield -mnobitfield -mhimem -mnohimem}
573 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
574 -mcall-prologues -mno-tablejump -mtiny-stack}
578 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
579 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
580 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
581 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
582 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
586 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
587 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
592 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
593 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
594 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
595 -minline-divide-max-throughput -mno-dwarf2-asm @gol
596 -mfixed-range=@var{register-range}}
598 @emph{S/390 and zSeries Options}
600 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
601 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
602 -m64 -m31 -mdebug -mno-debug}
606 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
607 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
608 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
609 -mstack-align -mdata-align -mconst-align @gol
610 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
611 -melf -maout -melinux -mlinux -sim -sim2}
613 @item Code Generation Options
614 @xref{Code Gen Options,,Options for Code Generation Conventions}.
616 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
617 -ffixed-@var{reg} -fexceptions @gol
618 -fnon-call-exceptions -funwind-tables -fasynchronous-unwind-tables @gol
619 -finhibit-size-directive -finstrument-functions @gol
620 -fno-common -fno-ident -fno-gnu-linker @gol
621 -fpcc-struct-return -fpic -fPIC @gol
622 -freg-struct-return -fshared-data -fshort-enums @gol
623 -fshort-double -fvolatile @gol
624 -fvolatile-global -fvolatile-static @gol
625 -fverbose-asm -fpack-struct -fstack-check @gol
626 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
627 -fargument-alias -fargument-noalias @gol
628 -fargument-noalias-global -fleading-underscore}
632 * Overall Options:: Controlling the kind of output:
633 an executable, object files, assembler files,
634 or preprocessed source.
635 * C Dialect Options:: Controlling the variant of C language compiled.
636 * C++ Dialect Options:: Variations on C++.
637 * Objective-C Dialect Options:: Variations on Objective-C.
638 * Language Independent Options:: Controlling how diagnostics should be
640 * Warning Options:: How picky should the compiler be?
641 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
642 * Optimize Options:: How much optimization?
643 * Preprocessor Options:: Controlling header files and macro definitions.
644 Also, getting dependency information for Make.
645 * Assembler Options:: Passing options to the assembler.
646 * Link Options:: Specifying libraries and so on.
647 * Directory Options:: Where to find header files and libraries.
648 Where to find the compiler executable files.
649 * Spec Files:: How to pass switches to sub-processes.
650 * Target Options:: Running a cross-compiler, or an old version of GCC.
653 @node Overall Options
654 @section Options Controlling the Kind of Output
656 Compilation can involve up to four stages: preprocessing, compilation
657 proper, assembly and linking, always in that order. The first three
658 stages apply to an individual source file, and end by producing an
659 object file; linking combines all the object files (those newly
660 compiled, and those specified as input) into an executable file.
662 @cindex file name suffix
663 For any given input file, the file name suffix determines what kind of
668 C source code which must be preprocessed.
671 C source code which should not be preprocessed.
674 C++ source code which should not be preprocessed.
677 Objective-C source code. Note that you must link with the library
678 @file{libobjc.a} to make an Objective-C program work.
681 Objective-C source code which should not be preprocessed.
684 C header file (not to be compiled or linked).
688 @itemx @var{file}.cxx
689 @itemx @var{file}.cpp
690 @itemx @var{file}.c++
692 C++ source code which must be preprocessed. Note that in @samp{.cxx},
693 the last two letters must both be literally @samp{x}. Likewise,
694 @samp{.C} refers to a literal capital C@.
697 @itemx @var{file}.for
698 @itemx @var{file}.FOR
699 Fortran source code which should not be preprocessed.
702 @itemx @var{file}.fpp
703 @itemx @var{file}.FPP
704 Fortran source code which must be preprocessed (with the traditional
708 Fortran source code which must be preprocessed with a RATFOR
709 preprocessor (not included with GCC)@.
711 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
712 Using and Porting GNU Fortran}, for more details of the handling of
715 @c FIXME: Descriptions of Java file types.
722 Ada source code file which contains a library unit declaration (a
723 declaration of a package, subprogram, or generic, or a generic
724 instantiation), or a library unit renaming declaration (a package,
725 generic, or subprogram renaming declaration). Such files are also
728 @itemx @var{file}.adb
729 Ada source code file containing a library unit body (a subprogram or
730 package body). Such files are also called @dfn{bodies}.
732 @c GCC also knows about some suffixes for languages not yet included:
738 @itemx @var{file}.chi
739 CHILL source code (preprocessed with the traditional preprocessor).
745 Assembler code which must be preprocessed.
748 An object file to be fed straight into linking.
749 Any file name with no recognized suffix is treated this way.
753 You can specify the input language explicitly with the @option{-x} option:
756 @item -x @var{language}
757 Specify explicitly the @var{language} for the following input files
758 (rather than letting the compiler choose a default based on the file
759 name suffix). This option applies to all following input files until
760 the next @option{-x} option. Possible values for @var{language} are:
762 c c-header cpp-output
764 objective-c objc-cpp-output
765 assembler assembler-with-cpp
768 f77 f77-cpp-input ratfor
773 Turn off any specification of a language, so that subsequent files are
774 handled according to their file name suffixes (as they are if @option{-x}
775 has not been used at all).
777 @item -pass-exit-codes
778 @opindex pass-exit-codes
779 Normally the @command{gcc} program will exit with the code of 1 if any
780 phase of the compiler returns a non-success return code. If you specify
781 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
782 numerically highest error produced by any phase that returned an error
786 If you only want some of the stages of compilation, you can use
787 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
788 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
789 @command{gcc} is to stop. Note that some combinations (for example,
790 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
795 Compile or assemble the source files, but do not link. The linking
796 stage simply is not done. The ultimate output is in the form of an
797 object file for each source file.
799 By default, the object file name for a source file is made by replacing
800 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
802 Unrecognized input files, not requiring compilation or assembly, are
807 Stop after the stage of compilation proper; do not assemble. The output
808 is in the form of an assembler code file for each non-assembler input
811 By default, the assembler file name for a source file is made by
812 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
814 Input files that don't require compilation are ignored.
818 Stop after the preprocessing stage; do not run the compiler proper. The
819 output is in the form of preprocessed source code, which is sent to the
822 Input files which don't require preprocessing are ignored.
824 @cindex output file option
827 Place output in file @var{file}. This applies regardless to whatever
828 sort of output is being produced, whether it be an executable file,
829 an object file, an assembler file or preprocessed C code.
831 Since only one output file can be specified, it does not make sense to
832 use @option{-o} when compiling more than one input file, unless you are
833 producing an executable file as output.
835 If @option{-o} is not specified, the default is to put an executable file
836 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
837 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
838 all preprocessed C source on standard output.
842 Print (on standard error output) the commands executed to run the stages
843 of compilation. Also print the version number of the compiler driver
844 program and of the preprocessor and the compiler proper.
848 Use pipes rather than temporary files for communication between the
849 various stages of compilation. This fails to work on some systems where
850 the assembler is unable to read from a pipe; but the GNU assembler has
855 Print (on the standard output) a description of the command line options
856 understood by @command{gcc}. If the @option{-v} option is also specified
857 then @option{--help} will also be passed on to the various processes
858 invoked by @command{gcc}, so that they can display the command line options
859 they accept. If the @option{-W} option is also specified then command
860 line options which have no documentation associated with them will also
865 Print (on the standard output) a description of target specific command
866 line options for each tool.
870 @section Compiling C++ Programs
872 @cindex suffixes for C++ source
873 @cindex C++ source file suffixes
874 C++ source files conventionally use one of the suffixes @samp{.C},
875 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
876 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
877 files with these names and compiles them as C++ programs even if you
878 call the compiler the same way as for compiling C programs (usually with
879 the name @command{gcc}).
883 However, C++ programs often require class libraries as well as a
884 compiler that understands the C++ language---and under some
885 circumstances, you might want to compile programs from standard input,
886 or otherwise without a suffix that flags them as C++ programs.
887 @command{g++} is a program that calls GCC with the default language
888 set to C++, and automatically specifies linking against the C++
889 library. On many systems, @command{g++} is also
890 installed with the name @command{c++}.
892 @cindex invoking @command{g++}
893 When you compile C++ programs, you may specify many of the same
894 command-line options that you use for compiling programs in any
895 language; or command-line options meaningful for C and related
896 languages; or options that are meaningful only for C++ programs.
897 @xref{C Dialect Options,,Options Controlling C Dialect}, for
898 explanations of options for languages related to C@.
899 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
900 explanations of options that are meaningful only for C++ programs.
902 @node C Dialect Options
903 @section Options Controlling C Dialect
904 @cindex dialect options
905 @cindex language dialect options
906 @cindex options, dialect
908 The following options control the dialect of C (or languages derived
909 from C, such as C++ and Objective-C) that the compiler accepts:
916 In C mode, support all ISO C89 programs. In C++ mode,
917 remove GNU extensions that conflict with ISO C++.
919 This turns off certain features of GCC that are incompatible with ISO
920 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
921 such as the @code{asm} and @code{typeof} keywords, and
922 predefined macros such as @code{unix} and @code{vax} that identify the
923 type of system you are using. It also enables the undesirable and
924 rarely used ISO trigraph feature. For the C compiler,
925 it disables recognition of C++ style @samp{//} comments as well as
926 the @code{inline} keyword.
928 The alternate keywords @code{__asm__}, @code{__extension__},
929 @code{__inline__} and @code{__typeof__} continue to work despite
930 @option{-ansi}. You would not want to use them in an ISO C program, of
931 course, but it is useful to put them in header files that might be included
932 in compilations done with @option{-ansi}. Alternate predefined macros
933 such as @code{__unix__} and @code{__vax__} are also available, with or
934 without @option{-ansi}.
936 The @option{-ansi} option does not cause non-ISO programs to be
937 rejected gratuitously. For that, @option{-pedantic} is required in
938 addition to @option{-ansi}. @xref{Warning Options}.
940 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
941 option is used. Some header files may notice this macro and refrain
942 from declaring certain functions or defining certain macros that the
943 ISO standard doesn't call for; this is to avoid interfering with any
944 programs that might use these names for other things.
946 Functions which would normally be built in but do not have semantics
947 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
948 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
949 built-in functions provided by GCC}, for details of the functions
954 Determine the language standard. This option is currently only
955 supported when compiling C@. A value for this option must be provided;
961 ISO C89 (same as @option{-ansi}).
964 ISO C89 as modified in amendment 1.
970 ISO C99. Note that this standard is not yet fully supported; see
971 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
972 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
975 Default, ISO C89 plus GNU extensions (including some C99 features).
979 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
980 this will become the default. The name @samp{gnu9x} is deprecated.
984 Even when this option is not specified, you can still use some of the
985 features of newer standards in so far as they do not conflict with
986 previous C standards. For example, you may use @code{__restrict__} even
987 when @option{-std=c99} is not specified.
989 The @option{-std} options specifying some version of ISO C have the same
990 effects as @option{-ansi}, except that features that were not in ISO C89
991 but are in the specified version (for example, @samp{//} comments and
992 the @code{inline} keyword in ISO C99) are not disabled.
994 @xref{Standards,,Language Standards Supported by GCC}, for details of
995 these standard versions.
997 @item -aux-info @var{filename}
999 Output to the given filename prototyped declarations for all functions
1000 declared and/or defined in a translation unit, including those in header
1001 files. This option is silently ignored in any language other than C@.
1003 Besides declarations, the file indicates, in comments, the origin of
1004 each declaration (source file and line), whether the declaration was
1005 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1006 @samp{O} for old, respectively, in the first character after the line
1007 number and the colon), and whether it came from a declaration or a
1008 definition (@samp{C} or @samp{F}, respectively, in the following
1009 character). In the case of function definitions, a K&R-style list of
1010 arguments followed by their declarations is also provided, inside
1011 comments, after the declaration.
1015 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1016 keyword, so that code can use these words as identifiers. You can use
1017 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1018 instead. @option{-ansi} implies @option{-fno-asm}.
1020 In C++, this switch only affects the @code{typeof} keyword, since
1021 @code{asm} and @code{inline} are standard keywords. You may want to
1022 use the @option{-fno-gnu-keywords} flag instead, which has the same
1023 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1024 switch only affects the @code{asm} and @code{typeof} keywords, since
1025 @code{inline} is a standard keyword in ISO C99.
1028 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1029 @opindex fno-builtin
1030 @cindex built-in functions
1031 Don't recognize built-in functions that do not begin with
1032 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1033 functions provided by GCC}, for details of the functions affected,
1034 including those which are not built-in functions when @option{-ansi} or
1035 @option{-std} options for strict ISO C conformance are used because they
1036 do not have an ISO standard meaning.
1038 GCC normally generates special code to handle certain built-in functions
1039 more efficiently; for instance, calls to @code{alloca} may become single
1040 instructions that adjust the stack directly, and calls to @code{memcpy}
1041 may become inline copy loops. The resulting code is often both smaller
1042 and faster, but since the function calls no longer appear as such, you
1043 cannot set a breakpoint on those calls, nor can you change the behavior
1044 of the functions by linking with a different library.
1046 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1047 option has no effect. Therefore, in C++, the only way to get the
1048 optimization benefits of built-in functions is to call the function
1049 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1050 built-in functions to implement many functions (like
1051 @code{std::strchr}), so that you automatically get efficient code.
1053 With the @option{-fno-builtin-@var{function}} option, not available
1054 when compiling C++, only the built-in function @var{function} is
1055 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1056 function is named this is not built-in in this version of GCC, this
1057 option is ignored. There is no corresponding
1058 @option{-fbuiltin-@var{function}} option; if you wish to enable
1059 built-in functions selectively when using @option{-fno-builtin} or
1060 @option{-ffreestanding}, you may define macros such as:
1063 #define abs(n) __builtin_abs ((n))
1064 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1069 @cindex hosted environment
1071 Assert that compilation takes place in a hosted environment. This implies
1072 @option{-fbuiltin}. A hosted environment is one in which the
1073 entire standard library is available, and in which @code{main} has a return
1074 type of @code{int}. Examples are nearly everything except a kernel.
1075 This is equivalent to @option{-fno-freestanding}.
1077 @item -ffreestanding
1078 @opindex ffreestanding
1079 @cindex hosted environment
1081 Assert that compilation takes place in a freestanding environment. This
1082 implies @option{-fno-builtin}. A freestanding environment
1083 is one in which the standard library may not exist, and program startup may
1084 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1085 This is equivalent to @option{-fno-hosted}.
1087 @xref{Standards,,Language Standards Supported by GCC}, for details of
1088 freestanding and hosted environments.
1092 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1093 options for strict ISO C conformance) implies @option{-trigraphs}.
1095 @cindex traditional C language
1096 @cindex C language, traditional
1098 @opindex traditional
1099 Attempt to support some aspects of traditional C compilers.
1104 All @code{extern} declarations take effect globally even if they
1105 are written inside of a function definition. This includes implicit
1106 declarations of functions.
1109 The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const}
1110 and @code{volatile} are not recognized. (You can still use the
1111 alternative keywords such as @code{__typeof__}, @code{__inline__}, and
1115 Comparisons between pointers and integers are always allowed.
1118 Integer types @code{unsigned short} and @code{unsigned char} promote
1119 to @code{unsigned int}.
1122 Out-of-range floating point literals are not an error.
1125 Certain constructs which ISO regards as a single invalid preprocessing
1126 number, such as @samp{0xe-0xd}, are treated as expressions instead.
1129 String ``constants'' are not necessarily constant; they are stored in
1130 writable space, and identical looking constants are allocated
1131 separately. (This is the same as the effect of
1132 @option{-fwritable-strings}.)
1134 @cindex @code{longjmp} and automatic variables
1136 All automatic variables not declared @code{register} are preserved by
1137 @code{longjmp}. Ordinarily, GNU C follows ISO C: automatic variables
1138 not declared @code{volatile} may be clobbered.
1143 @cindex escape sequences, traditional
1144 The character escape sequences @samp{\x} and @samp{\a} evaluate as the
1145 literal characters @samp{x} and @samp{a} respectively. Without
1146 @w{@option{-traditional}}, @samp{\x} is a prefix for the hexadecimal
1147 representation of a character, and @samp{\a} produces a bell.
1150 This option is deprecated and may be removed.
1152 You may wish to use @option{-fno-builtin} as well as @option{-traditional}
1153 if your program uses names that are normally GNU C built-in functions for
1154 other purposes of its own.
1156 You cannot use @option{-traditional} if you include any header files that
1157 rely on ISO C features. Some vendors are starting to ship systems with
1158 ISO C header files and you cannot use @option{-traditional} on such
1159 systems to compile files that include any system headers.
1161 The @option{-traditional} option also enables @option{-traditional-cpp}.
1163 @item -traditional-cpp
1164 @opindex traditional-cpp
1165 Attempt to support some aspects of traditional C preprocessors.
1166 See the GNU CPP manual for details.
1168 @item -fcond-mismatch
1169 @opindex fcond-mismatch
1170 Allow conditional expressions with mismatched types in the second and
1171 third arguments. The value of such an expression is void. This option
1172 is not supported for C++.
1174 @item -funsigned-char
1175 @opindex funsigned-char
1176 Let the type @code{char} be unsigned, like @code{unsigned char}.
1178 Each kind of machine has a default for what @code{char} should
1179 be. It is either like @code{unsigned char} by default or like
1180 @code{signed char} by default.
1182 Ideally, a portable program should always use @code{signed char} or
1183 @code{unsigned char} when it depends on the signedness of an object.
1184 But many programs have been written to use plain @code{char} and
1185 expect it to be signed, or expect it to be unsigned, depending on the
1186 machines they were written for. This option, and its inverse, let you
1187 make such a program work with the opposite default.
1189 The type @code{char} is always a distinct type from each of
1190 @code{signed char} or @code{unsigned char}, even though its behavior
1191 is always just like one of those two.
1194 @opindex fsigned-char
1195 Let the type @code{char} be signed, like @code{signed char}.
1197 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1198 the negative form of @option{-funsigned-char}. Likewise, the option
1199 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1201 @item -fsigned-bitfields
1202 @itemx -funsigned-bitfields
1203 @itemx -fno-signed-bitfields
1204 @itemx -fno-unsigned-bitfields
1205 @opindex fsigned-bitfields
1206 @opindex funsigned-bitfields
1207 @opindex fno-signed-bitfields
1208 @opindex fno-unsigned-bitfields
1209 These options control whether a bit-field is signed or unsigned, when the
1210 declaration does not use either @code{signed} or @code{unsigned}. By
1211 default, such a bit-field is signed, because this is consistent: the
1212 basic integer types such as @code{int} are signed types.
1214 However, when @option{-traditional} is used, bit-fields are all unsigned
1217 @item -fwritable-strings
1218 @opindex fwritable-strings
1219 Store string constants in the writable data segment and don't uniquize
1220 them. This is for compatibility with old programs which assume they can
1221 write into string constants. The option @option{-traditional} also has
1224 Writing into string constants is a very bad idea; ``constants'' should
1227 @item -fallow-single-precision
1228 @opindex fallow-single-precision
1229 Do not promote single precision math operations to double precision,
1230 even when compiling with @option{-traditional}.
1232 Traditional K&R C promotes all floating point operations to double
1233 precision, regardless of the sizes of the operands. On the
1234 architecture for which you are compiling, single precision may be faster
1235 than double precision. If you must use @option{-traditional}, but want
1236 to use single precision operations when the operands are single
1237 precision, use this option. This option has no effect when compiling
1238 with ISO or GNU C conventions (the default).
1241 @opindex fshort-wchar
1242 Override the underlying type for @samp{wchar_t} to be @samp{short
1243 unsigned int} instead of the default for the target. This option is
1244 useful for building programs to run under WINE@.
1247 @node C++ Dialect Options
1248 @section Options Controlling C++ Dialect
1250 @cindex compiler options, C++
1251 @cindex C++ options, command line
1252 @cindex options, C++
1253 This section describes the command-line options that are only meaningful
1254 for C++ programs; but you can also use most of the GNU compiler options
1255 regardless of what language your program is in. For example, you
1256 might compile a file @code{firstClass.C} like this:
1259 g++ -g -frepo -O -c firstClass.C
1263 In this example, only @option{-frepo} is an option meant
1264 only for C++ programs; you can use the other options with any
1265 language supported by GCC@.
1267 Here is a list of options that are @emph{only} for compiling C++ programs:
1270 @item -fno-access-control
1271 @opindex fno-access-control
1272 Turn off all access checking. This switch is mainly useful for working
1273 around bugs in the access control code.
1277 Check that the pointer returned by @code{operator new} is non-null
1278 before attempting to modify the storage allocated. The current Working
1279 Paper requires that @code{operator new} never return a null pointer, so
1280 this check is normally unnecessary.
1282 An alternative to using this option is to specify that your
1283 @code{operator new} does not throw any exceptions; if you declare it
1284 @samp{throw()}, G++ will check the return value. See also @samp{new
1287 @item -fconserve-space
1288 @opindex fconserve-space
1289 Put uninitialized or runtime-initialized global variables into the
1290 common segment, as C does. This saves space in the executable at the
1291 cost of not diagnosing duplicate definitions. If you compile with this
1292 flag and your program mysteriously crashes after @code{main()} has
1293 completed, you may have an object that is being destroyed twice because
1294 two definitions were merged.
1296 This option is no longer useful on most targets, now that support has
1297 been added for putting variables into BSS without making them common.
1299 @item -fno-const-strings
1300 @opindex fno-const-strings
1301 Give string constants type @code{char *} instead of type @code{const
1302 char *}. By default, G++ uses type @code{const char *} as required by
1303 the standard. Even if you use @option{-fno-const-strings}, you cannot
1304 actually modify the value of a string constant, unless you also use
1305 @option{-fwritable-strings}.
1307 This option might be removed in a future release of G++. For maximum
1308 portability, you should structure your code so that it works with
1309 string constants that have type @code{const char *}.
1311 @item -fdollars-in-identifiers
1312 @opindex fdollars-in-identifiers
1313 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1314 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1315 @samp{$} by default on most target systems, but there are a few exceptions.)
1316 Traditional C allowed the character @samp{$} to form part of
1317 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1319 @item -fno-elide-constructors
1320 @opindex fno-elide-constructors
1321 The C++ standard allows an implementation to omit creating a temporary
1322 which is only used to initialize another object of the same type.
1323 Specifying this option disables that optimization, and forces G++ to
1324 call the copy constructor in all cases.
1326 @item -fno-enforce-eh-specs
1327 @opindex fno-enforce-eh-specs
1328 Don't check for violation of exception specifications at runtime. This
1329 option violates the C++ standard, but may be useful for reducing code
1330 size in production builds, much like defining @samp{NDEBUG}. The compiler
1331 will still optimize based on the exception specifications.
1333 @item -fexternal-templates
1334 @opindex fexternal-templates
1336 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1337 template instantiation; template instances are emitted or not according
1338 to the location of the template definition. @xref{Template
1339 Instantiation}, for more information.
1341 This option is deprecated.
1343 @item -falt-external-templates
1344 @opindex falt-external-templates
1345 Similar to @option{-fexternal-templates}, but template instances are
1346 emitted or not according to the place where they are first instantiated.
1347 @xref{Template Instantiation}, for more information.
1349 This option is deprecated.
1352 @itemx -fno-for-scope
1354 @opindex fno-for-scope
1355 If @option{-ffor-scope} is specified, the scope of variables declared in
1356 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1357 as specified by the C++ standard.
1358 If @option{-fno-for-scope} is specified, the scope of variables declared in
1359 a @i{for-init-statement} extends to the end of the enclosing scope,
1360 as was the case in old versions of G++, and other (traditional)
1361 implementations of C++.
1363 The default if neither flag is given to follow the standard,
1364 but to allow and give a warning for old-style code that would
1365 otherwise be invalid, or have different behavior.
1367 @item -fno-gnu-keywords
1368 @opindex fno-gnu-keywords
1369 Do not recognize @code{typeof} as a keyword, so that code can use this
1370 word as an identifier. You can use the keyword @code{__typeof__} instead.
1371 @option{-ansi} implies @option{-fno-gnu-keywords}.
1373 @item -fno-implicit-templates
1374 @opindex fno-implicit-templates
1375 Never emit code for non-inline templates which are instantiated
1376 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1377 @xref{Template Instantiation}, for more information.
1379 @item -fno-implicit-inline-templates
1380 @opindex fno-implicit-inline-templates
1381 Don't emit code for implicit instantiations of inline templates, either.
1382 The default is to handle inlines differently so that compiles with and
1383 without optimization will need the same set of explicit instantiations.
1385 @item -fno-implement-inlines
1386 @opindex fno-implement-inlines
1387 To save space, do not emit out-of-line copies of inline functions
1388 controlled by @samp{#pragma implementation}. This will cause linker
1389 errors if these functions are not inlined everywhere they are called.
1391 @item -fms-extensions
1392 @opindex fms-extensions
1393 Disable pedantic warnings about constructs used in MFC, such as implicit
1394 int and getting a pointer to member function via non-standard syntax.
1396 @item -fno-nonansi-builtins
1397 @opindex fno-nonansi-builtins
1398 Disable built-in declarations of functions that are not mandated by
1399 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1400 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1402 @item -fno-operator-names
1403 @opindex fno-operator-names
1404 Do not treat the operator name keywords @code{and}, @code{bitand},
1405 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1406 synonyms as keywords.
1408 @item -fno-optional-diags
1409 @opindex fno-optional-diags
1410 Disable diagnostics that the standard says a compiler does not need to
1411 issue. Currently, the only such diagnostic issued by G++ is the one for
1412 a name having multiple meanings within a class.
1415 @opindex fpermissive
1416 Downgrade messages about nonconformant code from errors to warnings. By
1417 default, G++ effectively sets @option{-pedantic-errors} without
1418 @option{-pedantic}; this option reverses that. This behavior and this
1419 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1423 Enable automatic template instantiation at link time. This option also
1424 implies @option{-fno-implicit-templates}. @xref{Template
1425 Instantiation}, for more information.
1429 Disable generation of information about every class with virtual
1430 functions for use by the C++ runtime type identification features
1431 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1432 of the language, you can save some space by using this flag. Note that
1433 exception handling uses the same information, but it will generate it as
1438 Emit statistics about front-end processing at the end of the compilation.
1439 This information is generally only useful to the G++ development team.
1441 @item -ftemplate-depth-@var{n}
1442 @opindex ftemplate-depth
1443 Set the maximum instantiation depth for template classes to @var{n}.
1444 A limit on the template instantiation depth is needed to detect
1445 endless recursions during template class instantiation. ANSI/ISO C++
1446 conforming programs must not rely on a maximum depth greater than 17.
1448 @item -fuse-cxa-atexit
1449 @opindex fuse-cxa-atexit
1450 Register destructors for objects with static storage duration with the
1451 @code{__cxa_atexit} function rather than the @code{atexit} function.
1452 This option is required for fully standards-compliant handling of static
1453 destructors, but will only work if your C library supports
1454 @code{__cxa_atexit}.
1458 Emit special relocations for vtables and virtual function references
1459 so that the linker can identify unused virtual functions and zero out
1460 vtable slots that refer to them. This is most useful with
1461 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1462 also discard the functions themselves.
1464 This optimization requires GNU as and GNU ld. Not all systems support
1465 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1469 Do not use weak symbol support, even if it is provided by the linker.
1470 By default, G++ will use weak symbols if they are available. This
1471 option exists only for testing, and should not be used by end-users;
1472 it will result in inferior code and has no benefits. This option may
1473 be removed in a future release of G++.
1477 Do not search for header files in the standard directories specific to
1478 C++, but do still search the other standard directories. (This option
1479 is used when building the C++ library.)
1482 In addition, these optimization, warning, and code generation options
1483 have meanings only for C++ programs:
1486 @item -fno-default-inline
1487 @opindex fno-default-inline
1488 Do not assume @samp{inline} for functions defined inside a class scope.
1489 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1490 functions will have linkage like inline functions; they just won't be
1493 @item -Wctor-dtor-privacy @r{(C++ only)}
1494 @opindex Wctor-dtor-privacy
1495 Warn when a class seems unusable, because all the constructors or
1496 destructors in a class are private and the class has no friends or
1497 public static member functions.
1499 @item -Wnon-virtual-dtor @r{(C++ only)}
1500 @opindex Wnon-virtual-dtor
1501 Warn when a class declares a non-virtual destructor that should probably
1502 be virtual, because it looks like the class will be used polymorphically.
1504 @item -Wreorder @r{(C++ only)}
1506 @cindex reordering, warning
1507 @cindex warning for reordering of member initializers
1508 Warn when the order of member initializers given in the code does not
1509 match the order in which they must be executed. For instance:
1515 A(): j (0), i (1) @{ @}
1519 Here the compiler will warn that the member initializers for @samp{i}
1520 and @samp{j} will be rearranged to match the declaration order of the
1524 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1527 @item -Weffc++ @r{(C++ only)}
1529 Warn about violations of various style guidelines from Scott Meyers'
1530 @cite{Effective C++} books. If you use this option, you should be aware
1531 that the standard library headers do not obey all of these guidelines;
1532 you can use @samp{grep -v} to filter out those warnings.
1534 @item -Wno-deprecated @r{(C++ only)}
1535 @opindex Wno-deprecated
1536 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1538 @item -Wno-non-template-friend @r{(C++ only)}
1539 @opindex Wno-non-template-friend
1540 Disable warnings when non-templatized friend functions are declared
1541 within a template. With the advent of explicit template specification
1542 support in G++, if the name of the friend is an unqualified-id (i.e.,
1543 @samp{friend foo(int)}), the C++ language specification demands that the
1544 friend declare or define an ordinary, nontemplate function. (Section
1545 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1546 could be interpreted as a particular specialization of a templatized
1547 function. Because this non-conforming behavior is no longer the default
1548 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1549 check existing code for potential trouble spots, and is on by default.
1550 This new compiler behavior can be turned off with
1551 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1552 but disables the helpful warning.
1554 @item -Wold-style-cast @r{(C++ only)}
1555 @opindex Wold-style-cast
1556 Warn if an old-style (C-style) cast is used within a C++ program. The
1557 new-style casts (@samp{static_cast}, @samp{reinterpret_cast}, and
1558 @samp{const_cast}) are less vulnerable to unintended effects, and much
1561 @item -Woverloaded-virtual @r{(C++ only)}
1562 @opindex Woverloaded-virtual
1563 @cindex overloaded virtual fn, warning
1564 @cindex warning for overloaded virtual fn
1565 Warn when a function declaration hides virtual functions from a
1566 base class. For example, in:
1573 struct B: public A @{
1578 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1586 will fail to compile.
1588 @item -Wno-pmf-conversions @r{(C++ only)}
1589 @opindex Wno-pmf-conversions
1590 Disable the diagnostic for converting a bound pointer to member function
1593 @item -Wsign-promo @r{(C++ only)}
1594 @opindex Wsign-promo
1595 Warn when overload resolution chooses a promotion from unsigned or
1596 enumeral type to a signed type over a conversion to an unsigned type of
1597 the same size. Previous versions of G++ would try to preserve
1598 unsignedness, but the standard mandates the current behavior.
1600 @item -Wsynth @r{(C++ only)}
1602 @cindex warning for synthesized methods
1603 @cindex synthesized methods, warning
1604 Warn when G++'s synthesis behavior does not match that of cfront. For
1610 A& operator = (int);
1620 In this example, G++ will synthesize a default @samp{A& operator =
1621 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1624 @node Objective-C Dialect Options
1625 @section Options Controlling Objective-C Dialect
1627 @cindex compiler options, Objective-C
1628 @cindex Objective-C options, command line
1629 @cindex options, Objective-C
1630 This section describes the command-line options that are only meaningful
1631 for Objective-C programs; but you can also use most of the GNU compiler
1632 options regardless of what language your program is in. For example,
1633 you might compile a file @code{some_class.m} like this:
1636 gcc -g -fgnu-runtime -O -c some_class.m
1640 In this example, only @option{-fgnu-runtime} is an option meant only for
1641 Objective-C programs; you can use the other options with any language
1644 Here is a list of options that are @emph{only} for compiling Objective-C
1648 @item -fconstant-string-class=@var{class-name}
1649 @opindex fconstant-string-class
1650 Use @var{class-name} as the name of the class to instantiate for each
1651 literal string specified with the syntax @code{@@"@dots{}"}. The default
1652 class name is @code{NXConstantString}.
1655 @opindex fgnu-runtime
1656 Generate object code compatible with the standard GNU Objective-C
1657 runtime. This is the default for most types of systems.
1659 @item -fnext-runtime
1660 @opindex fnext-runtime
1661 Generate output compatible with the NeXT runtime. This is the default
1662 for NeXT-based systems, including Darwin and Mac OS X@.
1666 Dump interface declarations for all classes seen in the source file to a
1667 file named @file{@var{sourcename}.decl}.
1670 @opindex Wno-protocol
1671 Do not warn if methods required by a protocol are not implemented
1672 in the class adopting it.
1676 Warn if a selector has multiple methods of different types defined.
1678 @c not documented because only avail via -Wp
1679 @c @item -print-objc-runtime-info
1683 @node Language Independent Options
1684 @section Options to Control Diagnostic Messages Formatting
1685 @cindex options to control diagnostics formatting
1686 @cindex diagnostic messages
1687 @cindex message formatting
1689 Traditionally, diagnostic messages have been formatted irrespective of
1690 the output device's aspect (e.g.@: its width, @dots{}). The options described
1691 below can be used to control the diagnostic messages formatting
1692 algorithm, e.g.@: how many characters per line, how often source location
1693 information should be reported. Right now, only the C++ front end can
1694 honor these options. However it is expected, in the near future, that
1695 the remaining front ends would be able to digest them correctly.
1698 @item -fmessage-length=@var{n}
1699 @opindex fmessage-length
1700 Try to format error messages so that they fit on lines of about @var{n}
1701 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1702 the front ends supported by GCC@. If @var{n} is zero, then no
1703 line-wrapping will be done; each error message will appear on a single
1706 @opindex fdiagnostics-show-location
1707 @item -fdiagnostics-show-location=once
1708 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1709 reporter to emit @emph{once} source location information; that is, in
1710 case the message is too long to fit on a single physical line and has to
1711 be wrapped, the source location won't be emitted (as prefix) again,
1712 over and over, in subsequent continuation lines. This is the default
1715 @item -fdiagnostics-show-location=every-line
1716 Only meaningful in line-wrapping mode. Instructs the diagnostic
1717 messages reporter to emit the same source location information (as
1718 prefix) for physical lines that result from the process of breaking
1719 a message which is too long to fit on a single line.
1723 @node Warning Options
1724 @section Options to Request or Suppress Warnings
1725 @cindex options to control warnings
1726 @cindex warning messages
1727 @cindex messages, warning
1728 @cindex suppressing warnings
1730 Warnings are diagnostic messages that report constructions which
1731 are not inherently erroneous but which are risky or suggest there
1732 may have been an error.
1734 You can request many specific warnings with options beginning @samp{-W},
1735 for example @option{-Wimplicit} to request warnings on implicit
1736 declarations. Each of these specific warning options also has a
1737 negative form beginning @samp{-Wno-} to turn off warnings;
1738 for example, @option{-Wno-implicit}. This manual lists only one of the
1739 two forms, whichever is not the default.
1741 These options control the amount and kinds of warnings produced by GCC:
1744 @cindex syntax checking
1746 @opindex fsyntax-only
1747 Check the code for syntax errors, but don't do anything beyond that.
1751 Issue all the warnings demanded by strict ISO C and ISO C++;
1752 reject all programs that use forbidden extensions, and some other
1753 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1754 version of the ISO C standard specified by any @option{-std} option used.
1756 Valid ISO C and ISO C++ programs should compile properly with or without
1757 this option (though a rare few will require @option{-ansi} or a
1758 @option{-std} option specifying the required version of ISO C)@. However,
1759 without this option, certain GNU extensions and traditional C and C++
1760 features are supported as well. With this option, they are rejected.
1762 @option{-pedantic} does not cause warning messages for use of the
1763 alternate keywords whose names begin and end with @samp{__}. Pedantic
1764 warnings are also disabled in the expression that follows
1765 @code{__extension__}. However, only system header files should use
1766 these escape routes; application programs should avoid them.
1767 @xref{Alternate Keywords}.
1769 Some users try to use @option{-pedantic} to check programs for strict ISO
1770 C conformance. They soon find that it does not do quite what they want:
1771 it finds some non-ISO practices, but not all---only those for which
1772 ISO C @emph{requires} a diagnostic, and some others for which
1773 diagnostics have been added.
1775 A feature to report any failure to conform to ISO C might be useful in
1776 some instances, but would require considerable additional work and would
1777 be quite different from @option{-pedantic}. We don't have plans to
1778 support such a feature in the near future.
1780 Where the standard specified with @option{-std} represents a GNU
1781 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1782 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1783 extended dialect is based. Warnings from @option{-pedantic} are given
1784 where they are required by the base standard. (It would not make sense
1785 for such warnings to be given only for features not in the specified GNU
1786 C dialect, since by definition the GNU dialects of C include all
1787 features the compiler supports with the given option, and there would be
1788 nothing to warn about.)
1790 @item -pedantic-errors
1791 @opindex pedantic-errors
1792 Like @option{-pedantic}, except that errors are produced rather than
1797 Inhibit all warning messages.
1801 Inhibit warning messages about the use of @samp{#import}.
1803 @item -Wchar-subscripts
1804 @opindex Wchar-subscripts
1805 Warn if an array subscript has type @code{char}. This is a common cause
1806 of error, as programmers often forget that this type is signed on some
1811 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1812 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1816 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1817 the arguments supplied have types appropriate to the format string
1818 specified, and that the conversions specified in the format string make
1819 sense. This includes standard functions, and others specified by format
1820 attributes (@pxref{Function Attributes}), in the @code{printf},
1821 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1822 not in the C standard) families.
1824 The formats are checked against the format features supported by GNU
1825 libc version 2.2. These include all ISO C89 and C99 features, as well
1826 as features from the Single Unix Specification and some BSD and GNU
1827 extensions. Other library implementations may not support all these
1828 features; GCC does not support warning about features that go beyond a
1829 particular library's limitations. However, if @option{-pedantic} is used
1830 with @option{-Wformat}, warnings will be given about format features not
1831 in the selected standard version (but not for @code{strfmon} formats,
1832 since those are not in any version of the C standard). @xref{C Dialect
1833 Options,,Options Controlling C Dialect}.
1835 @option{-Wformat} is included in @option{-Wall}. For more control over some
1836 aspects of format checking, the options @option{-Wno-format-y2k},
1837 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1838 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1839 not included in @option{-Wall}.
1841 @item -Wno-format-y2k
1842 @opindex Wno-format-y2k
1843 If @option{-Wformat} is specified, do not warn about @code{strftime}
1844 formats which may yield only a two-digit year.
1846 @item -Wno-format-extra-args
1847 @opindex Wno-format-extra-args
1848 If @option{-Wformat} is specified, do not warn about excess arguments to a
1849 @code{printf} or @code{scanf} format function. The C standard specifies
1850 that such arguments are ignored.
1852 Where the unused arguments lie between used arguments that are
1853 specified with @samp{$} operand number specifications, normally
1854 warnings are still given, since the implementation could not know what
1855 type to pass to @code{va_arg} to skip the unused arguments. However,
1856 in the case of @code{scanf} formats, this option will suppress the
1857 warning if the unused arguments are all pointers, since the Single
1858 Unix Specification says that such unused arguments are allowed.
1860 @item -Wformat-nonliteral
1861 @opindex Wformat-nonliteral
1862 If @option{-Wformat} is specified, also warn if the format string is not a
1863 string literal and so cannot be checked, unless the format function
1864 takes its format arguments as a @code{va_list}.
1866 @item -Wformat-security
1867 @opindex Wformat-security
1868 If @option{-Wformat} is specified, also warn about uses of format
1869 functions that represent possible security problems. At present, this
1870 warns about calls to @code{printf} and @code{scanf} functions where the
1871 format string is not a string literal and there are no format arguments,
1872 as in @code{printf (foo);}. This may be a security hole if the format
1873 string came from untrusted input and contains @samp{%n}. (This is
1874 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1875 in future warnings may be added to @option{-Wformat-security} that are not
1876 included in @option{-Wformat-nonliteral}.)
1880 Enable @option{-Wformat} plus format checks not included in
1881 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1882 -Wformat-nonliteral -Wformat-security}.
1884 @item -Wimplicit-int
1885 @opindex Wimplicit-int
1886 Warn when a declaration does not specify a type.
1888 @item -Wimplicit-function-declaration
1889 @itemx -Werror-implicit-function-declaration
1890 @opindex Wimplicit-function-declaration
1891 @opindex Werror-implicit-function-declaration
1892 Give a warning (or error) whenever a function is used before being
1897 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1901 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1902 function with external linkage, returning int, taking either zero
1903 arguments, two, or three arguments of appropriate types.
1905 @item -Wmissing-braces
1906 @opindex Wmissing-braces
1907 Warn if an aggregate or union initializer is not fully bracketed. In
1908 the following example, the initializer for @samp{a} is not fully
1909 bracketed, but that for @samp{b} is fully bracketed.
1912 int a[2][2] = @{ 0, 1, 2, 3 @};
1913 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1917 @opindex Wparentheses
1918 Warn if parentheses are omitted in certain contexts, such
1919 as when there is an assignment in a context where a truth value
1920 is expected, or when operators are nested whose precedence people
1921 often get confused about.
1923 Also warn about constructions where there may be confusion to which
1924 @code{if} statement an @code{else} branch belongs. Here is an example of
1939 In C, every @code{else} branch belongs to the innermost possible @code{if}
1940 statement, which in this example is @code{if (b)}. This is often not
1941 what the programmer expected, as illustrated in the above example by
1942 indentation the programmer chose. When there is the potential for this
1943 confusion, GCC will issue a warning when this flag is specified.
1944 To eliminate the warning, add explicit braces around the innermost
1945 @code{if} statement so there is no way the @code{else} could belong to
1946 the enclosing @code{if}. The resulting code would look like this:
1962 @item -Wsequence-point
1963 @opindex Wsequence-point
1964 Warn about code that may have undefined semantics because of violations
1965 of sequence point rules in the C standard.
1967 The C standard defines the order in which expressions in a C program are
1968 evaluated in terms of @dfn{sequence points}, which represent a partial
1969 ordering between the execution of parts of the program: those executed
1970 before the sequence point, and those executed after it. These occur
1971 after the evaluation of a full expression (one which is not part of a
1972 larger expression), after the evaluation of the first operand of a
1973 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
1974 function is called (but after the evaluation of its arguments and the
1975 expression denoting the called function), and in certain other places.
1976 Other than as expressed by the sequence point rules, the order of
1977 evaluation of subexpressions of an expression is not specified. All
1978 these rules describe only a partial order rather than a total order,
1979 since, for example, if two functions are called within one expression
1980 with no sequence point between them, the order in which the functions
1981 are called is not specified. However, the standards committee have
1982 ruled that function calls do not overlap.
1984 It is not specified when between sequence points modifications to the
1985 values of objects take effect. Programs whose behavior depends on this
1986 have undefined behavior; the C standard specifies that ``Between the
1987 previous and next sequence point an object shall have its stored value
1988 modified at most once by the evaluation of an expression. Furthermore,
1989 the prior value shall be read only to determine the value to be
1990 stored.''. If a program breaks these rules, the results on any
1991 particular implementation are entirely unpredictable.
1993 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
1994 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
1995 diagnosed by this option, and it may give an occasional false positive
1996 result, but in general it has been found fairly effective at detecting
1997 this sort of problem in programs.
1999 The present implementation of this option only works for C programs. A
2000 future implementation may also work for C++ programs.
2002 The C standard is worded confusingly, therefore there is some debate
2003 over the precise meaning of the sequence point rules in subtle cases.
2004 Links to discussions of the problem, including proposed formal
2005 definitions, may be found on our readings page, at
2006 @w{@uref{http://gcc.gnu.org/readings.html}}.
2009 @opindex Wreturn-type
2010 Warn whenever a function is defined with a return-type that defaults to
2011 @code{int}. Also warn about any @code{return} statement with no
2012 return-value in a function whose return-type is not @code{void}.
2014 For C++, a function without return type always produces a diagnostic
2015 message, even when @option{-Wno-return-type} is specified. The only
2016 exceptions are @samp{main} and functions defined in system headers.
2020 Warn whenever a @code{switch} statement has an index of enumeral type
2021 and lacks a @code{case} for one or more of the named codes of that
2022 enumeration. (The presence of a @code{default} label prevents this
2023 warning.) @code{case} labels outside the enumeration range also
2024 provoke warnings when this option is used.
2028 Warn if any trigraphs are encountered that might change the meaning of
2029 the program (trigraphs within comments are not warned about).
2031 @item -Wunused-function
2032 @opindex Wunused-function
2033 Warn whenever a static function is declared but not defined or a
2034 non\-inline static function is unused.
2036 @item -Wunused-label
2037 @opindex Wunused-label
2038 Warn whenever a label is declared but not used.
2040 To suppress this warning use the @samp{unused} attribute
2041 (@pxref{Variable Attributes}).
2043 @item -Wunused-parameter
2044 @opindex Wunused-parameter
2045 Warn whenever a function parameter is unused aside from its declaration.
2047 To suppress this warning use the @samp{unused} attribute
2048 (@pxref{Variable Attributes}).
2050 @item -Wunused-variable
2051 @opindex Wunused-variable
2052 Warn whenever a local variable or non-constant static variable is unused
2053 aside from its declaration
2055 To suppress this warning use the @samp{unused} attribute
2056 (@pxref{Variable Attributes}).
2058 @item -Wunused-value
2059 @opindex Wunused-value
2060 Warn whenever a statement computes a result that is explicitly not used.
2062 To suppress this warning cast the expression to @samp{void}.
2066 All all the above @option{-Wunused} options combined.
2068 In order to get a warning about an unused function parameter, you must
2069 either specify @samp{-W -Wunused} or separately specify
2070 @option{-Wunused-parameter}.
2072 @item -Wuninitialized
2073 @opindex Wuninitialized
2074 Warn if an automatic variable is used without first being initialized or
2075 if a variable may be clobbered by a @code{setjmp} call.
2077 These warnings are possible only in optimizing compilation,
2078 because they require data flow information that is computed only
2079 when optimizing. If you don't specify @option{-O}, you simply won't
2082 These warnings occur only for variables that are candidates for
2083 register allocation. Therefore, they do not occur for a variable that
2084 is declared @code{volatile}, or whose address is taken, or whose size
2085 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2086 structures, unions or arrays, even when they are in registers.
2088 Note that there may be no warning about a variable that is used only
2089 to compute a value that itself is never used, because such
2090 computations may be deleted by data flow analysis before the warnings
2093 These warnings are made optional because GCC is not smart
2094 enough to see all the reasons why the code might be correct
2095 despite appearing to have an error. Here is one example of how
2116 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2117 always initialized, but GCC doesn't know this. Here is
2118 another common case:
2123 if (change_y) save_y = y, y = new_y;
2125 if (change_y) y = save_y;
2130 This has no bug because @code{save_y} is used only if it is set.
2132 @cindex @code{longjmp} warnings
2133 This option also warns when a non-volatile automatic variable might be
2134 changed by a call to @code{longjmp}. These warnings as well are possible
2135 only in optimizing compilation.
2137 The compiler sees only the calls to @code{setjmp}. It cannot know
2138 where @code{longjmp} will be called; in fact, a signal handler could
2139 call it at any point in the code. As a result, you may get a warning
2140 even when there is in fact no problem because @code{longjmp} cannot
2141 in fact be called at the place which would cause a problem.
2143 Some spurious warnings can be avoided if you declare all the functions
2144 you use that never return as @code{noreturn}. @xref{Function
2147 @item -Wreorder @r{(C++ only)}
2149 @cindex reordering, warning
2150 @cindex warning for reordering of member initializers
2151 Warn when the order of member initializers given in the code does not
2152 match the order in which they must be executed. For instance:
2154 @item -Wunknown-pragmas
2155 @opindex Wunknown-pragmas
2156 @cindex warning for unknown pragmas
2157 @cindex unknown pragmas, warning
2158 @cindex pragmas, warning of unknown
2159 Warn when a #pragma directive is encountered which is not understood by
2160 GCC@. If this command line option is used, warnings will even be issued
2161 for unknown pragmas in system header files. This is not the case if
2162 the warnings were only enabled by the @option{-Wall} command line option.
2166 All of the above @samp{-W} options combined. This enables all the
2167 warnings about constructions that some users consider questionable, and
2168 that are easy to avoid (or modify to prevent the warning), even in
2169 conjunction with macros.
2172 @opindex Wno-div-by-zero
2173 @opindex Wdiv-by-zero
2174 Warn about compile-time integer division by zero. This is default. To
2175 inhibit the warning messages, use @option{-Wno-div-by-zero}. Floating
2176 point division by zero is not warned about, as it can be a legitimate
2177 way of obtaining infinities and NaNs.
2180 @opindex Wno-multichar
2182 Warn if a multicharacter constant (@samp{'FOOF'}) is used. This is
2183 default. To inhibit the warning messages, use @option{-Wno-multichar}.
2184 Usually they indicate a typo in the user's code, as they have
2185 implementation-defined values, and should not be used in portable code.
2187 @item -Wsystem-headers
2188 @opindex Wsystem-headers
2189 @cindex warnings from system headers
2190 @cindex system headers, warnings from
2191 Print warning messages for constructs found in system header files.
2192 Warnings from system headers are normally suppressed, on the assumption
2193 that they usually do not indicate real problems and would only make the
2194 compiler output harder to read. Using this command line option tells
2195 GCC to emit warnings from system headers as if they occurred in user
2196 code. However, note that using @option{-Wall} in conjunction with this
2197 option will @emph{not} warn about unknown pragmas in system
2198 headers---for that, @option{-Wunknown-pragmas} must also be used.
2201 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2202 Some of them warn about constructions that users generally do not
2203 consider questionable, but which occasionally you might wish to check
2204 for; others warn about constructions that are necessary or hard to avoid
2205 in some cases, and there is no simple way to modify the code to suppress
2211 Print extra warning messages for these events:
2215 A function can return either with or without a value. (Falling
2216 off the end of the function body is considered returning without
2217 a value.) For example, this function would evoke such a
2231 An expression-statement or the left-hand side of a comma expression
2232 contains no side effects.
2233 To suppress the warning, cast the unused expression to void.
2234 For example, an expression such as @samp{x[i,j]} will cause a warning,
2235 but @samp{x[(void)i,j]} will not.
2238 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2241 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2242 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2243 that of ordinary mathematical notation.
2246 Storage-class specifiers like @code{static} are not the first things in
2247 a declaration. According to the C Standard, this usage is obsolescent.
2250 The return type of a function has a type qualifier such as @code{const}.
2251 Such a type qualifier has no effect, since the value returned by a
2252 function is not an lvalue. (But don't warn about the GNU extension of
2253 @code{volatile void} return types. That extension will be warned about
2254 if @option{-pedantic} is specified.)
2257 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2261 A comparison between signed and unsigned values could produce an
2262 incorrect result when the signed value is converted to unsigned.
2263 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2266 An aggregate has a partly bracketed initializer.
2267 For example, the following code would evoke such a warning,
2268 because braces are missing around the initializer for @code{x.h}:
2271 struct s @{ int f, g; @};
2272 struct t @{ struct s h; int i; @};
2273 struct t x = @{ 1, 2, 3 @};
2277 An aggregate has an initializer which does not initialize all members.
2278 For example, the following code would cause such a warning, because
2279 @code{x.h} would be implicitly initialized to zero:
2282 struct s @{ int f, g, h; @};
2283 struct s x = @{ 3, 4 @};
2288 @opindex Wfloat-equal
2289 Warn if floating point values are used in equality comparisons.
2291 The idea behind this is that sometimes it is convenient (for the
2292 programmer) to consider floating-point values as approximations to
2293 infinitely precise real numbers. If you are doing this, then you need
2294 to compute (by analysing the code, or in some other way) the maximum or
2295 likely maximum error that the computation introduces, and allow for it
2296 when performing comparisons (and when producing output, but that's a
2297 different problem). In particular, instead of testing for equality, you
2298 would check to see whether the two values have ranges that overlap; and
2299 this is done with the relational operators, so equality comparisons are
2302 @item -Wtraditional @r{(C only)}
2303 @opindex Wtraditional
2304 Warn about certain constructs that behave differently in traditional and
2305 ISO C@. Also warn about ISO C constructs that have no traditional C
2306 equivalent, and/or problematic constructs which should be avoided.
2310 Macro parameters that appear within string literals in the macro body.
2311 In traditional C macro replacement takes place within string literals,
2312 but does not in ISO C@.
2315 In traditional C, some preprocessor directives did not exist.
2316 Traditional preprocessors would only consider a line to be a directive
2317 if the @samp{#} appeared in column 1 on the line. Therefore
2318 @option{-Wtraditional} warns about directives that traditional C
2319 understands but would ignore because the @samp{#} does not appear as the
2320 first character on the line. It also suggests you hide directives like
2321 @samp{#pragma} not understood by traditional C by indenting them. Some
2322 traditional implementations would not recognize @samp{#elif}, so it
2323 suggests avoiding it altogether.
2326 A function-like macro that appears without arguments.
2329 The unary plus operator.
2332 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2333 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2334 constants.) Note, these suffixes appear in macros defined in the system
2335 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2336 Use of these macros in user code might normally lead to spurious
2337 warnings, however gcc's integrated preprocessor has enough context to
2338 avoid warning in these cases.
2341 A function declared external in one block and then used after the end of
2345 A @code{switch} statement has an operand of type @code{long}.
2348 A non-@code{static} function declaration follows a @code{static} one.
2349 This construct is not accepted by some traditional C compilers.
2352 The ISO type of an integer constant has a different width or
2353 signedness from its traditional type. This warning is only issued if
2354 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2355 typically represent bit patterns, are not warned about.
2358 Usage of ISO string concatenation is detected.
2361 Initialization of automatic aggregates.
2364 Identifier conflicts with labels. Traditional C lacks a separate
2365 namespace for labels.
2368 Initialization of unions. If the initializer is zero, the warning is
2369 omitted. This is done under the assumption that the zero initializer in
2370 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2371 initializer warnings and relies on default initialization to zero in the
2375 Conversions by prototypes between fixed/floating point values and vice
2376 versa. The absence of these prototypes when compiling with traditional
2377 C would cause serious problems. This is a subset of the possible
2378 conversion warnings, for the full set use @option{-Wconversion}.
2383 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2387 Warn whenever a local variable shadows another local variable, parameter or
2388 global variable or whenever a built-in function is shadowed.
2390 @item -Wlarger-than-@var{len}
2391 @opindex Wlarger-than
2392 Warn whenever an object of larger than @var{len} bytes is defined.
2394 @item -Wpointer-arith
2395 @opindex Wpointer-arith
2396 Warn about anything that depends on the ``size of'' a function type or
2397 of @code{void}. GNU C assigns these types a size of 1, for
2398 convenience in calculations with @code{void *} pointers and pointers
2401 @item -Wbad-function-cast @r{(C only)}
2402 @opindex Wbad-function-cast
2403 Warn whenever a function call is cast to a non-matching type.
2404 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2408 Warn whenever a pointer is cast so as to remove a type qualifier from
2409 the target type. For example, warn if a @code{const char *} is cast
2410 to an ordinary @code{char *}.
2413 @opindex Wcast-align
2414 Warn whenever a pointer is cast such that the required alignment of the
2415 target is increased. For example, warn if a @code{char *} is cast to
2416 an @code{int *} on machines where integers can only be accessed at
2417 two- or four-byte boundaries.
2419 @item -Wwrite-strings
2420 @opindex Wwrite-strings
2421 When compiling C, give string constants the type @code{const
2422 char[@var{length}]} so that
2423 copying the address of one into a non-@code{const} @code{char *}
2424 pointer will get a warning; when compiling C++, warn about the
2425 deprecated conversion from string constants to @code{char *}.
2426 These warnings will help you find at
2427 compile time code that can try to write into a string constant, but
2428 only if you have been very careful about using @code{const} in
2429 declarations and prototypes. Otherwise, it will just be a nuisance;
2430 this is why we did not make @option{-Wall} request these warnings.
2433 @opindex Wconversion
2434 Warn if a prototype causes a type conversion that is different from what
2435 would happen to the same argument in the absence of a prototype. This
2436 includes conversions of fixed point to floating and vice versa, and
2437 conversions changing the width or signedness of a fixed point argument
2438 except when the same as the default promotion.
2440 Also, warn if a negative integer constant expression is implicitly
2441 converted to an unsigned type. For example, warn about the assignment
2442 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2443 casts like @code{(unsigned) -1}.
2445 @item -Wsign-compare
2446 @opindex Wsign-compare
2447 @cindex warning for comparison of signed and unsigned values
2448 @cindex comparison of signed and unsigned values, warning
2449 @cindex signed and unsigned values, comparison warning
2450 Warn when a comparison between signed and unsigned values could produce
2451 an incorrect result when the signed value is converted to unsigned.
2452 This warning is also enabled by @option{-W}; to get the other warnings
2453 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2455 @item -Waggregate-return
2456 @opindex Waggregate-return
2457 Warn if any functions that return structures or unions are defined or
2458 called. (In languages where you can return an array, this also elicits
2461 @item -Wstrict-prototypes @r{(C only)}
2462 @opindex Wstrict-prototypes
2463 Warn if a function is declared or defined without specifying the
2464 argument types. (An old-style function definition is permitted without
2465 a warning if preceded by a declaration which specifies the argument
2468 @item -Wmissing-prototypes @r{(C only)}
2469 @opindex Wmissing-prototypes
2470 Warn if a global function is defined without a previous prototype
2471 declaration. This warning is issued even if the definition itself
2472 provides a prototype. The aim is to detect global functions that fail
2473 to be declared in header files.
2475 @item -Wmissing-declarations
2476 @opindex Wmissing-declarations
2477 Warn if a global function is defined without a previous declaration.
2478 Do so even if the definition itself provides a prototype.
2479 Use this option to detect global functions that are not declared in
2482 @item -Wmissing-noreturn
2483 @opindex Wmissing-noreturn
2484 Warn about functions which might be candidates for attribute @code{noreturn}.
2485 Note these are only possible candidates, not absolute ones. Care should
2486 be taken to manually verify functions actually do not ever return before
2487 adding the @code{noreturn} attribute, otherwise subtle code generation
2488 bugs could be introduced. You will not get a warning for @code{main} in
2489 hosted C environments.
2491 @item -Wmissing-format-attribute
2492 @opindex Wmissing-format-attribute
2494 If @option{-Wformat} is enabled, also warn about functions which might be
2495 candidates for @code{format} attributes. Note these are only possible
2496 candidates, not absolute ones. GCC will guess that @code{format}
2497 attributes might be appropriate for any function that calls a function
2498 like @code{vprintf} or @code{vscanf}, but this might not always be the
2499 case, and some functions for which @code{format} attributes are
2500 appropriate may not be detected. This option has no effect unless
2501 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2505 Warn if a structure is given the packed attribute, but the packed
2506 attribute has no effect on the layout or size of the structure.
2507 Such structures may be mis-aligned for little benefit. For
2508 instance, in this code, the variable @code{f.x} in @code{struct bar}
2509 will be misaligned even though @code{struct bar} does not itself
2510 have the packed attribute:
2517 @} __attribute__((packed));
2527 Warn if padding is included in a structure, either to align an element
2528 of the structure or to align the whole structure. Sometimes when this
2529 happens it is possible to rearrange the fields of the structure to
2530 reduce the padding and so make the structure smaller.
2532 @item -Wredundant-decls
2533 @opindex Wredundant-decls
2534 Warn if anything is declared more than once in the same scope, even in
2535 cases where multiple declaration is valid and changes nothing.
2537 @item -Wnested-externs @r{(C only)}
2538 @opindex Wnested-externs
2539 Warn if an @code{extern} declaration is encountered within a function.
2541 @item -Wunreachable-code
2542 @opindex Wunreachable-code
2543 Warn if the compiler detects that code will never be executed.
2545 This option is intended to warn when the compiler detects that at
2546 least a whole line of source code will never be executed, because
2547 some condition is never satisfied or because it is after a
2548 procedure that never returns.
2550 It is possible for this option to produce a warning even though there
2551 are circumstances under which part of the affected line can be executed,
2552 so care should be taken when removing apparently-unreachable code.
2554 For instance, when a function is inlined, a warning may mean that the
2555 line is unreachable in only one inlined copy of the function.
2557 This option is not made part of @option{-Wall} because in a debugging
2558 version of a program there is often substantial code which checks
2559 correct functioning of the program and is, hopefully, unreachable
2560 because the program does work. Another common use of unreachable
2561 code is to provide behavior which is selectable at compile-time.
2565 Warn if a function can not be inlined and it was declared as inline.
2569 @opindex Wno-long-long
2570 Warn if @samp{long long} type is used. This is default. To inhibit
2571 the warning messages, use @option{-Wno-long-long}. Flags
2572 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2573 only when @option{-pedantic} flag is used.
2575 @item -Wdisabled-optimization
2576 @opindex Wdisabled-optimization
2577 Warn if a requested optimization pass is disabled. This warning does
2578 not generally indicate that there is anything wrong with your code; it
2579 merely indicates that GCC's optimizers were unable to handle the code
2580 effectively. Often, the problem is that your code is too big or too
2581 complex; GCC will refuse to optimize programs when the optimization
2582 itself is likely to take inordinate amounts of time.
2586 Make all warnings into errors.
2589 @node Debugging Options
2590 @section Options for Debugging Your Program or GCC
2591 @cindex options, debugging
2592 @cindex debugging information options
2594 GCC has various special options that are used for debugging
2595 either your program or GCC:
2600 Produce debugging information in the operating system's native format
2601 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2604 On most systems that use stabs format, @option{-g} enables use of extra
2605 debugging information that only GDB can use; this extra information
2606 makes debugging work better in GDB but will probably make other debuggers
2608 refuse to read the program. If you want to control for certain whether
2609 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2610 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2611 or @option{-gvms} (see below).
2613 Unlike most other C compilers, GCC allows you to use @option{-g} with
2614 @option{-O}. The shortcuts taken by optimized code may occasionally
2615 produce surprising results: some variables you declared may not exist
2616 at all; flow of control may briefly move where you did not expect it;
2617 some statements may not be executed because they compute constant
2618 results or their values were already at hand; some statements may
2619 execute in different places because they were moved out of loops.
2621 Nevertheless it proves possible to debug optimized output. This makes
2622 it reasonable to use the optimizer for programs that might have bugs.
2624 The following options are useful when GCC is generated with the
2625 capability for more than one debugging format.
2629 Produce debugging information for use by GDB@. This means to use the
2630 most expressive format available (DWARF 2, stabs, or the native format
2631 if neither of those are supported), including GDB extensions if at all
2636 Produce debugging information in stabs format (if that is supported),
2637 without GDB extensions. This is the format used by DBX on most BSD
2638 systems. On MIPS, Alpha and System V Release 4 systems this option
2639 produces stabs debugging output which is not understood by DBX or SDB@.
2640 On System V Release 4 systems this option requires the GNU assembler.
2644 Produce debugging information in stabs format (if that is supported),
2645 using GNU extensions understood only by the GNU debugger (GDB)@. The
2646 use of these extensions is likely to make other debuggers crash or
2647 refuse to read the program.
2651 Produce debugging information in COFF format (if that is supported).
2652 This is the format used by SDB on most System V systems prior to
2657 Produce debugging information in XCOFF format (if that is supported).
2658 This is the format used by the DBX debugger on IBM RS/6000 systems.
2662 Produce debugging information in XCOFF format (if that is supported),
2663 using GNU extensions understood only by the GNU debugger (GDB)@. The
2664 use of these extensions is likely to make other debuggers crash or
2665 refuse to read the program, and may cause assemblers other than the GNU
2666 assembler (GAS) to fail with an error.
2670 Produce debugging information in DWARF version 1 format (if that is
2671 supported). This is the format used by SDB on most System V Release 4
2676 Produce debugging information in DWARF version 1 format (if that is
2677 supported), using GNU extensions understood only by the GNU debugger
2678 (GDB)@. The use of these extensions is likely to make other debuggers
2679 crash or refuse to read the program.
2683 Produce debugging information in DWARF version 2 format (if that is
2684 supported). This is the format used by DBX on IRIX 6.
2688 Produce debugging information in VMS debug format (if that is
2689 supported). This is the format used by DEBUG on VMS systems.
2692 @itemx -ggdb@var{level}
2693 @itemx -gstabs@var{level}
2694 @itemx -gcoff@var{level}
2695 @itemx -gxcoff@var{level}
2696 @itemx -gdwarf@var{level}
2697 @itemx -gdwarf-2@var{level}
2698 @itemx -gvms@var{level}
2699 Request debugging information and also use @var{level} to specify how
2700 much information. The default level is 2.
2702 Level 1 produces minimal information, enough for making backtraces in
2703 parts of the program that you don't plan to debug. This includes
2704 descriptions of functions and external variables, but no information
2705 about local variables and no line numbers.
2707 Level 3 includes extra information, such as all the macro definitions
2708 present in the program. Some debuggers support macro expansion when
2709 you use @option{-g3}.
2714 Generate extra code to write profile information suitable for the
2715 analysis program @code{prof}. You must use this option when compiling
2716 the source files you want data about, and you must also use it when
2719 @cindex @code{gprof}
2722 Generate extra code to write profile information suitable for the
2723 analysis program @code{gprof}. You must use this option when compiling
2724 the source files you want data about, and you must also use it when
2730 Generate extra code to write profile information for basic blocks, which will
2731 record the number of times each basic block is executed, the basic block start
2732 address, and the function name containing the basic block. If @option{-g} is
2733 used, the line number and filename of the start of the basic block will also be
2734 recorded. If not overridden by the machine description, the default action is
2735 to append to the text file @file{bb.out}.
2737 This data could be analyzed by a program like @code{tcov}. Note,
2738 however, that the format of the data is not what @code{tcov} expects.
2739 Eventually GNU @code{gprof} should be extended to process this data.
2743 Makes the compiler print out each function name as it is compiled, and
2744 print some statistics about each pass when it finishes.
2747 @opindex ftime-report
2748 Makes the compiler print some statistics about the time consumed by each
2749 pass when it finishes.
2752 @opindex fmem-report
2753 Makes the compiler print some statistics about permanent memory
2754 allocation when it finishes.
2756 @item -fprofile-arcs
2757 @opindex fprofile-arcs
2758 Instrument @dfn{arcs} during compilation to generate coverage data
2759 or for profile-directed block ordering. During execution the program
2760 records how many times each branch is executed and how many times it is
2761 taken. When the compiled program exits it saves this data to a file
2762 called @file{@var{sourcename}.da} for each source file.
2764 For profile-directed block ordering, compile the program with
2765 @option{-fprofile-arcs} plus optimization and code generation options,
2766 generate the arc profile information by running the program on a
2767 selected workload, and then compile the program again with the same
2768 optimization and code generation options plus
2769 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2770 Control Optimization}).
2772 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2773 when it is used with the @option{-ftest-coverage} option. GCC
2774 supports two methods of determining code coverage: the options that
2775 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2776 write information to text files. The options that support @code{gcov}
2777 do not need to instrument every arc in the program, so a program compiled
2778 with them runs faster than a program compiled with @option{-a}, which
2779 adds instrumentation code to every basic block in the program. The
2780 tradeoff: since @code{gcov} does not have execution counts for all
2781 branches, it must start with the execution counts for the instrumented
2782 branches, and then iterate over the program flow graph until the entire
2783 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2784 a program which uses information from @option{-a} and @option{-ax}.
2786 With @option{-fprofile-arcs}, for each function of your program GCC
2787 creates a program flow graph, then finds a spanning tree for the graph.
2788 Only arcs that are not on the spanning tree have to be instrumented: the
2789 compiler adds code to count the number of times that these arcs are
2790 executed. When an arc is the only exit or only entrance to a block, the
2791 instrumentation code can be added to the block; otherwise, a new basic
2792 block must be created to hold the instrumentation code.
2794 This option makes it possible to estimate branch probabilities and to
2795 calculate basic block execution counts. In general, basic block
2796 execution counts as provided by @option{-a} do not give enough
2797 information to estimate all branch probabilities.
2800 @item -ftest-coverage
2801 @opindex ftest-coverage
2802 Create data files for the @code{gcov} code-coverage utility
2803 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2804 The data file names begin with the name of your source file:
2807 @item @var{sourcename}.bb
2808 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2809 associate basic block execution counts with line numbers.
2811 @item @var{sourcename}.bbg
2812 A list of all arcs in the program flow graph. This allows @code{gcov}
2813 to reconstruct the program flow graph, so that it can compute all basic
2814 block and arc execution counts from the information in the
2815 @code{@var{sourcename}.da} file.
2818 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2819 option adds instrumentation to the program, which then writes
2820 execution counts to another data file:
2823 @item @var{sourcename}.da
2824 Runtime arc execution counts, used in conjunction with the arc
2825 information in the file @code{@var{sourcename}.bbg}.
2828 Coverage data will map better to the source files if
2829 @option{-ftest-coverage} is used without optimization.
2831 @item -d@var{letters}
2833 Says to make debugging dumps during compilation at times specified by
2834 @var{letters}. This is used for debugging the compiler. The file names
2835 for most of the dumps are made by appending a pass number and a word to
2836 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2837 Here are the possible letters for use in @var{letters}, and their meanings:
2842 Annotate the assembler output with miscellaneous debugging information.
2845 Dump after computing branch probabilities, to @file{@var{file}.14.bp}.
2848 Dump after block reordering, to @file{@var{file}.28.bbro}.
2851 Dump after instruction combination, to the file @file{@var{file}.16.combine}.
2854 Dump after the first if conversion, to the file @file{@var{file}.17.ce}.
2857 Dump after delayed branch scheduling, to @file{@var{file}.31.dbr}.
2860 Dump all macro definitions, at the end of preprocessing, in addition to
2864 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2865 @file{@var{file}.07.ussa}.
2868 Dump after the second if conversion, to @file{@var{file}.26.ce2}.
2871 Dump after life analysis, to @file{@var{file}.15.life}.
2874 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.09.addressof}.
2877 Dump after global register allocation, to @file{@var{file}.21.greg}.
2880 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2883 Dump after post-reload optimizations, to @file{@var{file}.22.postreload}.
2886 Dump after GCSE, to @file{@var{file}.10.gcse}.
2889 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2892 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2895 Dump after conversion from registers to stack, to @file{@var{file}.32.stack}.
2898 Dump after local register allocation, to @file{@var{file}.20.lreg}.
2901 Dump after loop optimization, to @file{@var{file}.11.loop}.
2904 Dump after performing the machine dependent reorganisation pass, to
2905 @file{@var{file}.30.mach}.
2908 Dump after register renumbering, to @file{@var{file}.25.rnreg}.
2911 Dump after the register move pass, to @file{@var{file}.18.regmove}.
2914 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2917 Dump after the second scheduling pass, to @file{@var{file}.27.sched2}.
2920 Dump after CSE (including the jump optimization that sometimes follows
2921 CSE), to @file{@var{file}.08.cse}.
2924 Dump after the first scheduling pass, to @file{@var{file}.19.sched}.
2927 Dump after the second CSE pass (including the jump optimization that
2928 sometimes follows CSE), to @file{@var{file}.12.cse2}.
2931 Dump after the second flow pass, to @file{@var{file}.23.flow2}.
2934 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2937 Dump after the peephole pass, to @file{@var{file}.24.peephole2}.
2940 Produce all the dumps listed above.
2943 Print statistics on memory usage, at the end of the run, to
2947 Annotate the assembler output with a comment indicating which
2948 pattern and alternative was used. The length of each instruction is
2952 Dump the RTL in the assembler output as a comment before each instruction.
2953 Also turns on @option{-dp} annotation.
2956 For each of the other indicated dump files (except for
2957 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
2958 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
2961 Just generate RTL for a function instead of compiling it. Usually used
2965 Dump debugging information during parsing, to standard error.
2968 @item -fdump-unnumbered
2969 @opindex fdump-unnumbered
2970 When doing debugging dumps (see @option{-d} option above), suppress instruction
2971 numbers and line number note output. This makes it more feasible to
2972 use diff on debugging dumps for compiler invocations with different
2973 options, in particular with and without @option{-g}.
2975 @item -fdump-class-hierarchy @r{(C++ only)}
2976 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
2977 @opindex fdump-class-hierarchy
2978 Dump a representation of each class's hierarchy and virtual function
2979 table layout to a file. The file name is made by appending @file{.class}
2980 to the source file name. If the @samp{-@var{options}} form is used,
2981 @var{options} controls the details of the dump as described for the
2982 @option{-fdump-tree} options.
2984 @item -fdump-tree-@var{switch} @r{(C++ only)}
2985 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
2987 Control the dumping at various stages of processing the intermediate
2988 language tree to a file. The file name is generated by appending a switch
2989 specific suffix to the source file name. If the @samp{-@var{options}}
2990 form is used, @var{options} is a list of @samp{-} separated options that
2991 control the details of the dump. Not all options are applicable to all
2992 dumps, those which are not meaningful will be ignored. The following
2993 options are available
2997 Print the address of each node. Usually this is not meaningful as it
2998 changes according to the environment and source file. Its primary use
2999 is for tying up a dump file with a debug environment.
3001 Inhibit dumping of members of a scope or body of a function merely
3002 because that scope has been reached. Only dump such items when they
3003 are directly reachable by some other path.
3005 Turn on all options.
3008 The following tree dumps are possible:
3011 Dump before any tree based optimization, to @file{@var{file}.original}.
3013 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3015 Dump after function inlining, to @file{@var{file}.inlined}.
3018 @item -fpretend-float
3019 @opindex fpretend-float
3020 When running a cross-compiler, pretend that the target machine uses the
3021 same floating point format as the host machine. This causes incorrect
3022 output of the actual floating constants, but the actual instruction
3023 sequence will probably be the same as GCC would make when running on
3028 Store the usual ``temporary'' intermediate files permanently; place them
3029 in the current directory and name them based on the source file. Thus,
3030 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3031 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3032 preprocessed @file{foo.i} output file even though the compiler now
3033 normally uses an integrated preprocessor.
3037 Report the CPU time taken by each subprocess in the compilation
3038 sequence. For C source files, this is the compiler proper and assembler
3039 (plus the linker if linking is done). The output looks like this:
3046 The first number on each line is the ``user time,'' that is time spent
3047 executing the program itself. The second number is ``system time,''
3048 time spent executing operating system routines on behalf of the program.
3049 Both numbers are in seconds.
3051 @item -print-file-name=@var{library}
3052 @opindex print-file-name
3053 Print the full absolute name of the library file @var{library} that
3054 would be used when linking---and don't do anything else. With this
3055 option, GCC does not compile or link anything; it just prints the
3058 @item -print-multi-directory
3059 @opindex print-multi-directory
3060 Print the directory name corresponding to the multilib selected by any
3061 other switches present in the command line. This directory is supposed
3062 to exist in @env{GCC_EXEC_PREFIX}.
3064 @item -print-multi-lib
3065 @opindex print-multi-lib
3066 Print the mapping from multilib directory names to compiler switches
3067 that enable them. The directory name is separated from the switches by
3068 @samp{;}, and each switch starts with an @samp{@@} instead of the
3069 @samp{-}, without spaces between multiple switches. This is supposed to
3070 ease shell-processing.
3072 @item -print-prog-name=@var{program}
3073 @opindex print-prog-name
3074 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3076 @item -print-libgcc-file-name
3077 @opindex print-libgcc-file-name
3078 Same as @option{-print-file-name=libgcc.a}.
3080 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3081 but you do want to link with @file{libgcc.a}. You can do
3084 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3087 @item -print-search-dirs
3088 @opindex print-search-dirs
3089 Print the name of the configured installation directory and a list of
3090 program and library directories gcc will search---and don't do anything else.
3092 This is useful when gcc prints the error message
3093 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3094 To resolve this you either need to put @file{cpp0} and the other compiler
3095 components where gcc expects to find them, or you can set the environment
3096 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3097 Don't forget the trailing '/'.
3098 @xref{Environment Variables}.
3101 @opindex dumpmachine
3102 Print the compiler's target machine (for example,
3103 @samp{i686-pc-linux-gnu})---and don't do anything else.
3106 @opindex dumpversion
3107 Print the compiler version (for example, @samp{3.0})---and don't do
3112 Print the compiler's built-in specs---and don't do anything else. (This
3113 is used when GCC itself is being built.) @xref{Spec Files}.
3116 @node Optimize Options
3117 @section Options That Control Optimization
3118 @cindex optimize options
3119 @cindex options, optimization
3121 These options control various sorts of optimizations:
3128 Optimize. Optimizing compilation takes somewhat more time, and a lot
3129 more memory for a large function.
3131 Without @option{-O}, the compiler's goal is to reduce the cost of
3132 compilation and to make debugging produce the expected results.
3133 Statements are independent: if you stop the program with a breakpoint
3134 between statements, you can then assign a new value to any variable or
3135 change the program counter to any other statement in the function and
3136 get exactly the results you would expect from the source code.
3138 With @option{-O}, the compiler tries to reduce code size and execution
3139 time, without performing any optimizations that take a great deal of
3144 Optimize even more. GCC performs nearly all supported optimizations
3145 that do not involve a space-speed tradeoff. The compiler does not
3146 perform loop unrolling or function inlining when you specify @option{-O2}.
3147 As compared to @option{-O}, this option increases both compilation time
3148 and the performance of the generated code.
3150 @option{-O2} turns on all optional optimizations except for loop unrolling,
3151 function inlining, and register renaming. It also turns on the
3152 @option{-fforce-mem} option on all machines and frame pointer elimination
3153 on machines where doing so does not interfere with debugging.
3155 Please note the warning under @option{-fgcse} about
3156 invoking @option{-O2} on programs that use computed gotos.
3160 Optimize yet more. @option{-O3} turns on all optimizations specified by
3161 @option{-O2} and also turns on the @option{-finline-functions} and
3162 @option{-frename-registers} options.
3170 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3171 do not typically increase code size. It also performs further
3172 optimizations designed to reduce code size.
3174 If you use multiple @option{-O} options, with or without level numbers,
3175 the last such option is the one that is effective.
3178 Options of the form @option{-f@var{flag}} specify machine-independent
3179 flags. Most flags have both positive and negative forms; the negative
3180 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3181 only one of the forms is listed---the one which is not the default.
3182 You can figure out the other form by either removing @samp{no-} or
3187 @opindex ffloat-store
3188 Do not store floating point variables in registers, and inhibit other
3189 options that might change whether a floating point value is taken from a
3192 @cindex floating point precision
3193 This option prevents undesirable excess precision on machines such as
3194 the 68000 where the floating registers (of the 68881) keep more
3195 precision than a @code{double} is supposed to have. Similarly for the
3196 x86 architecture. For most programs, the excess precision does only
3197 good, but a few programs rely on the precise definition of IEEE floating
3198 point. Use @option{-ffloat-store} for such programs, after modifying
3199 them to store all pertinent intermediate computations into variables.
3201 @item -fno-default-inline
3202 @opindex fno-default-inline
3203 Do not make member functions inline by default merely because they are
3204 defined inside the class scope (C++ only). Otherwise, when you specify
3205 @w{@option{-O}}, member functions defined inside class scope are compiled
3206 inline by default; i.e., you don't need to add @samp{inline} in front of
3207 the member function name.
3209 @item -fno-defer-pop
3210 @opindex fno-defer-pop
3211 Always pop the arguments to each function call as soon as that function
3212 returns. For machines which must pop arguments after a function call,
3213 the compiler normally lets arguments accumulate on the stack for several
3214 function calls and pops them all at once.
3218 Force memory operands to be copied into registers before doing
3219 arithmetic on them. This produces better code by making all memory
3220 references potential common subexpressions. When they are not common
3221 subexpressions, instruction combination should eliminate the separate
3222 register-load. The @option{-O2} option turns on this option.
3225 @opindex fforce-addr
3226 Force memory address constants to be copied into registers before
3227 doing arithmetic on them. This may produce better code just as
3228 @option{-fforce-mem} may.
3230 @item -fomit-frame-pointer
3231 @opindex fomit-frame-pointer
3232 Don't keep the frame pointer in a register for functions that
3233 don't need one. This avoids the instructions to save, set up and
3234 restore frame pointers; it also makes an extra register available
3235 in many functions. @strong{It also makes debugging impossible on
3239 On some machines, such as the VAX, this flag has no effect, because
3240 the standard calling sequence automatically handles the frame pointer
3241 and nothing is saved by pretending it doesn't exist. The
3242 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3243 whether a target machine supports this flag. @xref{Registers}.
3246 On some machines, such as the VAX, this flag has no effect, because
3247 the standard calling sequence automatically handles the frame pointer
3248 and nothing is saved by pretending it doesn't exist. The
3249 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3250 whether a target machine supports this flag. @xref{Registers,,Register
3251 Usage, gcc.info, Using and Porting GCC}.
3254 @item -foptimize-sibling-calls
3255 @opindex foptimize-sibling-calls
3256 Optimize sibling and tail recursive calls.
3260 This option generates traps for signed overflow on addition, subtraction,
3261 multiplication operations.
3265 Don't pay attention to the @code{inline} keyword. Normally this option
3266 is used to keep the compiler from expanding any functions inline.
3267 Note that if you are not optimizing, no functions can be expanded inline.
3269 @item -finline-functions
3270 @opindex finline-functions
3271 Integrate all simple functions into their callers. The compiler
3272 heuristically decides which functions are simple enough to be worth
3273 integrating in this way.
3275 If all calls to a given function are integrated, and the function is
3276 declared @code{static}, then the function is normally not output as
3277 assembler code in its own right.
3279 @item -finline-limit=@var{n}
3280 @opindex finline-limit
3281 By default, gcc limits the size of functions that can be inlined. This flag
3282 allows the control of this limit for functions that are explicitly marked as
3283 inline (ie marked with the inline keyword or defined within the class
3284 definition in c++). @var{n} is the size of functions that can be inlined in
3285 number of pseudo instructions (not counting parameter handling). The default
3286 value of @var{n} is 600.
3287 Increasing this value can result in more inlined code at
3288 the cost of compilation time and memory consumption. Decreasing usually makes
3289 the compilation faster and less code will be inlined (which presumably
3290 means slower programs). This option is particularly useful for programs that
3291 use inlining heavily such as those based on recursive templates with C++.
3293 @emph{Note:} pseudo instruction represents, in this particular context, an
3294 abstract measurement of function's size. In no way, it represents a count
3295 of assembly instructions and as such its exact meaning might change from one
3296 release to an another.
3298 @item -fkeep-inline-functions
3299 @opindex fkeep-inline-functions
3300 Even if all calls to a given function are integrated, and the function
3301 is declared @code{static}, nevertheless output a separate run-time
3302 callable version of the function. This switch does not affect
3303 @code{extern inline} functions.
3305 @item -fkeep-static-consts
3306 @opindex fkeep-static-consts
3307 Emit variables declared @code{static const} when optimization isn't turned
3308 on, even if the variables aren't referenced.
3310 GCC enables this option by default. If you want to force the compiler to
3311 check if the variable was referenced, regardless of whether or not
3312 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3314 @item -fmerge-constants
3315 Attempt to merge identical constants (string constants and floating point
3316 constants) accross compilation units.
3318 This option is default for optimized compilation if assembler and linker
3319 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3321 @item -fmerge-all-constants
3322 Attempt to merge identical constants and identical variables.
3324 This option implies @option{-fmerge-constants}. In addition to
3325 @option{-fmerge-constants} this considers e.g. even constant initialized
3326 arrays or initialized constant variables with integral or floating point
3327 types. Languages like C or C++ require each non-automatic variable to
3328 have distinct location, so using this option will result in non-conforming
3331 @item -fno-function-cse
3332 @opindex fno-function-cse
3333 Do not put function addresses in registers; make each instruction that
3334 calls a constant function contain the function's address explicitly.
3336 This option results in less efficient code, but some strange hacks
3337 that alter the assembler output may be confused by the optimizations
3338 performed when this option is not used.
3342 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3343 @option{-fno-trapping-math}.
3345 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3347 This option should never be turned on by any @option{-O} option since
3348 it can result in incorrect output for programs which depend on
3349 an exact implementation of IEEE or ISO rules/specifications for
3352 @item -fno-math-errno
3353 @opindex fno-math-errno
3354 Do not set ERRNO after calling math functions that are executed
3355 with a single instruction, e.g., sqrt. A program that relies on
3356 IEEE exceptions for math error handling may want to use this flag
3357 for speed while maintaining IEEE arithmetic compatibility.
3359 This option should never be turned on by any @option{-O} option since
3360 it can result in incorrect output for programs which depend on
3361 an exact implementation of IEEE or ISO rules/specifications for
3364 The default is @option{-fmath-errno}.
3366 @item -funsafe-math-optimizations
3367 @opindex funsafe-math-optimizations
3368 Allow optimizations for floating-point arithmetic that (a) assume
3369 that arguments and results are valid and (b) may violate IEEE or
3370 ANSI standards. When used at link-time, it may include libraries
3371 or startup files that change the default FPU control word or other
3372 similar optimizations.
3374 This option should never be turned on by any @option{-O} option since
3375 it can result in incorrect output for programs which depend on
3376 an exact implementation of IEEE or ISO rules/specifications for
3379 The default is @option{-fno-unsafe-math-optimizations}.
3381 @item -fno-trapping-math
3382 @opindex fno-trapping-math
3383 Compile code assuming that floating-point operations cannot generate
3384 user-visible traps. Setting this option may allow faster code
3385 if one relies on ``non-stop'' IEEE arithmetic, for example.
3387 This option should never be turned on by any @option{-O} option since
3388 it can result in incorrect output for programs which depend on
3389 an exact implementation of IEEE or ISO rules/specifications for
3392 The default is @option{-ftrapping-math}.
3395 The following options control specific optimizations. The @option{-O2}
3396 option turns on all of these optimizations except @option{-funroll-loops}
3397 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3398 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3399 but specific machines may handle it differently.
3401 You can use the following flags in the rare cases when ``fine-tuning''
3402 of optimizations to be performed is desired.
3404 Not all of the optimizations performed by GCC have @option{-f} options
3408 @item -fstrength-reduce
3409 @opindex fstrength-reduce
3410 Perform the optimizations of loop strength reduction and
3411 elimination of iteration variables.
3413 @item -fthread-jumps
3414 @opindex fthread-jumps
3415 Perform optimizations where we check to see if a jump branches to a
3416 location where another comparison subsumed by the first is found. If
3417 so, the first branch is redirected to either the destination of the
3418 second branch or a point immediately following it, depending on whether
3419 the condition is known to be true or false.
3421 @item -fcse-follow-jumps
3422 @opindex fcse-follow-jumps
3423 In common subexpression elimination, scan through jump instructions
3424 when the target of the jump is not reached by any other path. For
3425 example, when CSE encounters an @code{if} statement with an
3426 @code{else} clause, CSE will follow the jump when the condition
3429 @item -fcse-skip-blocks
3430 @opindex fcse-skip-blocks
3431 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3432 follow jumps which conditionally skip over blocks. When CSE
3433 encounters a simple @code{if} statement with no else clause,
3434 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3435 body of the @code{if}.
3437 @item -frerun-cse-after-loop
3438 @opindex frerun-cse-after-loop
3439 Re-run common subexpression elimination after loop optimizations has been
3442 @item -frerun-loop-opt
3443 @opindex frerun-loop-opt
3444 Run the loop optimizer twice.
3448 Perform a global common subexpression elimination pass.
3449 This pass also performs global constant and copy propagation.
3451 @emph{Note:} When compiling a program using computed gotos, a GCC
3452 extension, you may get better runtime performance if you disable
3453 the global common subexpression elmination pass by adding
3454 @option{-fno-gcse} to the command line.
3458 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3459 attempt to move loads which are only killed by stores into themselves. This
3460 allows a loop containing a load/store sequence to be changed to a load outside
3461 the loop, and a copy/store within the loop.
3465 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3466 subexpression elimination. This pass will attempt to move stores out of loops.
3467 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3468 can be changed to a load before the loop and a store after the loop.
3470 @item -fdelete-null-pointer-checks
3471 @opindex fdelete-null-pointer-checks
3472 Use global dataflow analysis to identify and eliminate useless checks
3473 for null pointers. The compiler assumes that dereferencing a null
3474 pointer would have halted the program. If a pointer is checked after
3475 it has already been dereferenced, it cannot be null.
3477 In some environments, this assumption is not true, and programs can
3478 safely dereference null pointers. Use
3479 @option{-fno-delete-null-pointer-checks} to disable this optimization
3480 for programs which depend on that behavior.
3482 @item -fexpensive-optimizations
3483 @opindex fexpensive-optimizations
3484 Perform a number of minor optimizations that are relatively expensive.
3486 @item -foptimize-register-move
3488 @opindex foptimize-register-move
3490 Attempt to reassign register numbers in move instructions and as
3491 operands of other simple instructions in order to maximize the amount of
3492 register tying. This is especially helpful on machines with two-operand
3493 instructions. GCC enables this optimization by default with @option{-O2}
3496 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3499 @item -fdelayed-branch
3500 @opindex fdelayed-branch
3501 If supported for the target machine, attempt to reorder instructions
3502 to exploit instruction slots available after delayed branch
3505 @item -fschedule-insns
3506 @opindex fschedule-insns
3507 If supported for the target machine, attempt to reorder instructions to
3508 eliminate execution stalls due to required data being unavailable. This
3509 helps machines that have slow floating point or memory load instructions
3510 by allowing other instructions to be issued until the result of the load
3511 or floating point instruction is required.
3513 @item -fschedule-insns2
3514 @opindex fschedule-insns2
3515 Similar to @option{-fschedule-insns}, but requests an additional pass of
3516 instruction scheduling after register allocation has been done. This is
3517 especially useful on machines with a relatively small number of
3518 registers and where memory load instructions take more than one cycle.
3520 @item -ffunction-sections
3521 @itemx -fdata-sections
3522 @opindex ffunction-sections
3523 @opindex fdata-sections
3524 Place each function or data item into its own section in the output
3525 file if the target supports arbitrary sections. The name of the
3526 function or the name of the data item determines the section's name
3529 Use these options on systems where the linker can perform optimizations
3530 to improve locality of reference in the instruction space. HPPA
3531 processors running HP-UX and Sparc processors running Solaris 2 have
3532 linkers with such optimizations. Other systems using the ELF object format
3533 as well as AIX may have these optimizations in the future.
3535 Only use these options when there are significant benefits from doing
3536 so. When you specify these options, the assembler and linker will
3537 create larger object and executable files and will also be slower.
3538 You will not be able to use @code{gprof} on all systems if you
3539 specify this option and you may have problems with debugging if
3540 you specify both this option and @option{-g}.
3542 @item -fcaller-saves
3543 @opindex fcaller-saves
3544 Enable values to be allocated in registers that will be clobbered by
3545 function calls, by emitting extra instructions to save and restore the
3546 registers around such calls. Such allocation is done only when it
3547 seems to result in better code than would otherwise be produced.
3549 This option is always enabled by default on certain machines, usually
3550 those which have no call-preserved registers to use instead.
3552 For all machines, optimization level 2 and higher enables this flag by
3555 @item -funroll-loops
3556 @opindex funroll-loops
3557 Unroll loops whose number of iterations can be determined at compile
3558 time or upon entry to the loop. @option{-funroll-loops} implies both
3559 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3560 option makes code larger, and may or may not make it run faster.
3562 @item -funroll-all-loops
3563 @opindex funroll-all-loops
3564 Unroll all loops, even if their number of iterations is uncertain when
3565 the loop is entered. This usually makes programs run more slowly.
3566 @option{-funroll-all-loops} implies the same options as
3567 @option{-funroll-loops},
3570 @item -fmove-all-movables
3571 @opindex fmove-all-movables
3572 Forces all invariant computations in loops to be moved
3575 @item -freduce-all-givs
3576 @opindex freduce-all-givs
3577 Forces all general-induction variables in loops to be
3580 @emph{Note:} When compiling programs written in Fortran,
3581 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3582 by default when you use the optimizer.
3584 These options may generate better or worse code; results are highly
3585 dependent on the structure of loops within the source code.
3587 These two options are intended to be removed someday, once
3588 they have helped determine the efficacy of various
3589 approaches to improving loop optimizations.
3591 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3592 know how use of these options affects
3593 the performance of your production code.
3594 We're very interested in code that runs @emph{slower}
3595 when these options are @emph{enabled}.
3598 @itemx -fno-peephole2
3599 @opindex fno-peephole
3600 @opindex fno-peephole2
3601 Disable any machine-specific peephole optimizations. The difference
3602 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3603 are implemented in the compiler; some targets use one, some use the
3604 other, a few use both.
3606 @item -fbranch-probabilities
3607 @opindex fbranch-probabilities
3608 After running a program compiled with @option{-fprofile-arcs}
3609 (@pxref{Debugging Options,, Options for Debugging Your Program or
3610 @command{gcc}}), you can compile it a second time using
3611 @option{-fbranch-probabilities}, to improve optimizations based on
3612 the number of times each branch was taken. When the program
3613 compiled with @option{-fprofile-arcs} exits it saves arc execution
3614 counts to a file called @file{@var{sourcename}.da} for each source
3615 file The information in this data file is very dependent on the
3616 structure of the generated code, so you must use the same source code
3617 and the same optimization options for both compilations.
3620 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3621 note on the first instruction of each basic block, and a
3622 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3623 These can be used to improve optimization. Currently, they are only
3624 used in one place: in @file{reorg.c}, instead of guessing which path a
3625 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3626 exactly determine which path is taken more often.
3629 @item -fno-guess-branch-probability
3630 @opindex fno-guess-branch-probability
3631 Do not guess branch probabilities using a randomized model.
3633 Sometimes gcc will opt to use a randomized model to guess branch
3634 probabilities, when none are available from either profiling feedback
3635 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3636 different runs of the compiler on the same program may produce different
3639 In a hard real-time system, people don't want different runs of the
3640 compiler to produce code that has different behavior; minimizing
3641 non-determinism is of paramount import. This switch allows users to
3642 reduce non-determinism, possibly at the expense of inferior
3645 @item -fstrict-aliasing
3646 @opindex fstrict-aliasing
3647 Allows the compiler to assume the strictest aliasing rules applicable to
3648 the language being compiled. For C (and C++), this activates
3649 optimizations based on the type of expressions. In particular, an
3650 object of one type is assumed never to reside at the same address as an
3651 object of a different type, unless the types are almost the same. For
3652 example, an @code{unsigned int} can alias an @code{int}, but not a
3653 @code{void*} or a @code{double}. A character type may alias any other
3656 Pay special attention to code like this:
3669 The practice of reading from a different union member than the one most
3670 recently written to (called ``type-punning'') is common. Even with
3671 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3672 is accessed through the union type. So, the code above will work as
3673 expected. However, this code might not:
3685 Every language that wishes to perform language-specific alias analysis
3686 should define a function that computes, given an @code{tree}
3687 node, an alias set for the node. Nodes in different alias sets are not
3688 allowed to alias. For an example, see the C front-end function
3689 @code{c_get_alias_set}.
3692 @item -falign-functions
3693 @itemx -falign-functions=@var{n}
3694 @opindex falign-functions
3695 Align the start of functions to the next power-of-two greater than
3696 @var{n}, skipping up to @var{n} bytes. For instance,
3697 @option{-falign-functions=32} aligns functions to the next 32-byte
3698 boundary, but @option{-falign-functions=24} would align to the next
3699 32-byte boundary only if this can be done by skipping 23 bytes or less.
3701 @option{-fno-align-functions} and @option{-falign-functions=1} are
3702 equivalent and mean that functions will not be aligned.
3704 Some assemblers only support this flag when @var{n} is a power of two;
3705 in that case, it is rounded up.
3707 If @var{n} is not specified, use a machine-dependent default.
3709 @item -falign-labels
3710 @itemx -falign-labels=@var{n}
3711 @opindex falign-labels
3712 Align all branch targets to a power-of-two boundary, skipping up to
3713 @var{n} bytes like @option{-falign-functions}. This option can easily
3714 make code slower, because it must insert dummy operations for when the
3715 branch target is reached in the usual flow of the code.
3717 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3718 are greater than this value, then their values are used instead.
3720 If @var{n} is not specified, use a machine-dependent default which is
3721 very likely to be @samp{1}, meaning no alignment.
3724 @itemx -falign-loops=@var{n}
3725 @opindex falign-loops
3726 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3727 like @option{-falign-functions}. The hope is that the loop will be
3728 executed many times, which will make up for any execution of the dummy
3731 If @var{n} is not specified, use a machine-dependent default.
3734 @itemx -falign-jumps=@var{n}
3735 @opindex falign-jumps
3736 Align branch targets to a power-of-two boundary, for branch targets
3737 where the targets can only be reached by jumping, skipping up to @var{n}
3738 bytes like @option{-falign-functions}. In this case, no dummy operations
3741 If @var{n} is not specified, use a machine-dependent default.
3745 Perform optimizations in static single assignment form. Each function's
3746 flow graph is translated into SSA form, optimizations are performed, and
3747 the flow graph is translated back from SSA form. Users should not
3748 specify this option, since it is not yet ready for production use.
3752 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3753 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3757 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3758 Like @option{-fssa}, this is an experimental feature.
3760 @item -fsingle-precision-constant
3761 @opindex fsingle-precision-constant
3762 Treat floating point constant as single precision constant instead of
3763 implicitly converting it to double precision constant.
3765 @item -frename-registers
3766 @opindex frename-registers
3767 Attempt to avoid false dependencies in scheduled code by making use
3768 of registers left over after register allocation. This optimization
3769 will most benefit processors with lots of registers. It can, however,
3770 make debugging impossible, since variables will no longer stay in
3771 a ``home register''.
3773 @item --param @var{name}=@var{value}
3775 In some places, GCC uses various constants to control the amount of
3776 optimization that is done. For example, GCC will not inline functions
3777 that contain more that a certain number of instructions. You can
3778 control some of these constants on the command-line using the
3779 @option{--param} option.
3781 In each case, the @var{value} is an integer. The allowable choices for
3782 @var{name} are given in the following table:
3785 @item max-delay-slot-insn-search
3786 The maximum number of instructions to consider when looking for an
3787 instruction to fill a delay slot. If more than this arbitrary number of
3788 instructions is searched, the time savings from filling the delay slot
3789 will be minimal so stop searching. Increasing values mean more
3790 aggressive optimization, making the compile time increase with probably
3791 small improvement in executable run time.
3793 @item max-delay-slot-live-search
3794 When trying to fill delay slots, the maximum number of instructions to
3795 consider when searching for a block with valid live register
3796 information. Increasing this arbitrarily chosen value means more
3797 aggressive optimization, increasing the compile time. This parameter
3798 should be removed when the delay slot code is rewritten to maintain the
3801 @item max-gcse-memory
3802 The approximate maximum amount of memory that will be allocated in
3803 order to perform the global common subexpression elimination
3804 optimization. If more memory than specified is required, the
3805 optimization will not be done.
3807 @item max-gcse-passes
3808 The maximum number of passes of GCSE to run.
3810 @item max-pending-list-length
3811 The maximum number of pending dependencies scheduling will allow
3812 before flushing the current state and starting over. Large functions
3813 with few branches or calls can create excessively large lists which
3814 needlessly consume memory and resources.
3816 @item max-inline-insns
3817 If an function contains more than this many instructions, it
3818 will not be inlined. This option is precisely equivalent to
3819 @option{-finline-limit}.
3824 @node Preprocessor Options
3825 @section Options Controlling the Preprocessor
3826 @cindex preprocessor options
3827 @cindex options, preprocessor
3829 These options control the C preprocessor, which is run on each C source
3830 file before actual compilation.
3832 If you use the @option{-E} option, nothing is done except preprocessing.
3833 Some of these options make sense only together with @option{-E} because
3834 they cause the preprocessor output to be unsuitable for actual
3838 @item -include @var{file}
3840 Process @var{file} as input before processing the regular input file.
3841 In effect, the contents of @var{file} are compiled first. Any @option{-D}
3842 and @option{-U} options on the command line are always processed before
3843 @option{-include @var{file}}, regardless of the order in which they are
3844 written. All the @option{-include} and @option{-imacros} options are
3845 processed in the order in which they are written.
3847 @item -imacros @var{file}
3849 Process @var{file} as input, discarding the resulting output, before
3850 processing the regular input file. Because the output generated from
3851 @var{file} is discarded, the only effect of @option{-imacros @var{file}}
3852 is to make the macros defined in @var{file} available for use in the
3853 main input. All the @option{-include} and @option{-imacros} options are
3854 processed in the order in which they are written.
3856 @item -idirafter @var{dir}
3858 @cindex second include path
3859 Add the directory @var{dir} to the second include path. The directories
3860 on the second include path are searched when a header file is not found
3861 in any of the directories in the main include path (the one that
3862 @option{-I} adds to).
3864 @item -iprefix @var{prefix}
3866 Specify @var{prefix} as the prefix for subsequent @option{-iwithprefix}
3869 @item -iwithprefix @var{dir}
3870 @opindex iwithprefix
3871 Add a directory to the second include path. The directory's name is
3872 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
3873 specified previously with @option{-iprefix}. If you have not specified a
3874 prefix yet, the directory containing the installed passes of the
3875 compiler is used as the default.
3877 @item -iwithprefixbefore @var{dir}
3878 @opindex iwithprefixbefore
3879 Add a directory to the main include path. The directory's name is made
3880 by concatenating @var{prefix} and @var{dir}, as in the case of
3881 @option{-iwithprefix}.
3883 @item -isystem @var{dir}
3885 Add a directory to the beginning of the second include path, marking it
3886 as a system directory, so that it gets the same special treatment as
3887 is applied to the standard system directories.
3891 Do not search the standard system directories for header files. Only
3892 the directories you have specified with @option{-I} options (and the
3893 current directory, if appropriate) are searched. @xref{Directory
3894 Options}, for information on @option{-I}.
3896 By using both @option{-nostdinc} and @option{-I-}, you can limit the include-file
3897 search path to only those directories you specify explicitly.
3901 When searching for a header file in a directory, remap file names if a
3902 file named @file{header.gcc} exists in that directory. This can be used
3903 to work around limitations of file systems with file name restrictions.
3904 The @file{header.gcc} file should contain a series of lines with two
3905 tokens on each line: the first token is the name to map, and the second
3906 token is the actual name to use.
3910 Do not predefine any nonstandard macros. (Including architecture flags).
3914 Run only the C preprocessor. Preprocess all the C source files
3915 specified and output the results to standard output or to the
3916 specified output file.
3920 Tell the preprocessor not to discard comments. Used with the
3925 Tell the preprocessor not to generate @samp{#line} directives.
3926 Used with the @option{-E} option.
3929 @cindex dependencies, make
3932 Instead of outputting the result of preprocessing, output a rule
3933 suitable for @code{make} describing the dependencies of the main source
3934 file. The preprocessor outputs one @code{make} rule containing the
3935 object file name for that source file, a colon, and the names of all the
3936 included files. Unless overridden explicitly, the object file name
3937 consists of the basename of the source file with any suffix replaced with
3938 object file suffix. If there are many included files then the
3939 rule is split into several lines using @samp{\}-newline.
3941 @option{-M} implies @option{-E}.
3945 Like @option{-M}, but mention only the files included with @samp{#include
3946 "@var{file}"}. System header files included with @samp{#include
3947 <@var{file}>} are omitted.
3951 Like @option{-M} but the dependency information is written to a file
3952 rather than stdout. @code{gcc} will use the same file name and
3953 directory as the object file, but with the suffix @file{.d} instead.
3955 This is in addition to compiling the main file as specified---@option{-MD}
3956 does not inhibit ordinary compilation the way @option{-M} does,
3957 unless you also specify @option{-MG}.
3959 With Mach, you can use the utility @code{md} to merge multiple
3960 dependency files into a single dependency file suitable for using with
3961 the @samp{make} command.
3965 Like @option{-MD} except mention only user header files, not system
3968 @item -MF @var{file}
3970 When used with @option{-M} or @option{-MM}, specifies a file to write the
3971 dependencies to. This allows the preprocessor to write the preprocessed
3972 file to stdout normally. If no @option{-MF} switch is given, CPP sends
3973 the rules to stdout and suppresses normal preprocessed output.
3975 Another way to specify output of a @code{make} rule is by setting
3976 the environment variable @env{DEPENDENCIES_OUTPUT} (@pxref{Environment
3981 When used with @option{-M} or @option{-MM}, @option{-MG} says to treat missing
3982 header files as generated files and assume they live in the same
3983 directory as the source file. It suppresses preprocessed output, as a
3984 missing header file is ordinarily an error.
3986 This feature is used in automatic updating of makefiles.
3990 This option instructs CPP to add a phony target for each dependency
3991 other than the main file, causing each to depend on nothing. These
3992 dummy rules work around errors @code{make} gives if you remove header
3993 files without updating the @code{Makefile} to match.
3995 This is typical output:-
3998 /tmp/test.o: /tmp/test.c /tmp/test.h
4003 @item -MQ @var{target}
4004 @item -MT @var{target}
4007 By default CPP uses the main file name, including any path, and appends
4008 the object suffix, normally ``.o'', to it to obtain the name of the
4009 target for dependency generation. With @option{-MT} you can specify a
4010 target yourself, overriding the default one.
4012 If you want multiple targets, you can specify them as a single argument
4013 to @option{-MT}, or use multiple @option{-MT} options.
4015 The targets you specify are output in the order they appear on the
4016 command line. @option{-MQ} is identical to @option{-MT}, except that the
4017 target name is quoted for Make, but with @option{-MT} it isn't. For
4018 example, @option{-MT '$(objpfx)foo.o'} gives
4021 $(objpfx)foo.o: /tmp/foo.c
4024 but @option{-MQ '$(objpfx)foo.o'} gives
4027 $$(objpfx)foo.o: /tmp/foo.c
4030 The default target is automatically quoted, as if it were given with
4035 Print the name of each header file used, in addition to other normal
4038 @item -A@var{question}(@var{answer})
4040 Assert the answer @var{answer} for @var{question}, in case it is tested
4041 with a preprocessing conditional such as @samp{#if
4042 #@var{question}(@var{answer})}. @option{-A-} disables the standard
4043 assertions that normally describe the target machine.
4047 Define macro @var{macro} with the string @samp{1} as its definition.
4049 @item -D@var{macro}=@var{defn}
4050 Define macro @var{macro} as @var{defn}. All instances of @option{-D} on
4051 the command line are processed before any @option{-U} options.
4053 Any @option{-D} and @option{-U} options on the command line are processed in
4054 order, and always before @option{-imacros @var{file}}, regardless of the
4055 order in which they are written.
4059 Undefine macro @var{macro}. @option{-U} options are evaluated after all
4060 @option{-D} options, but before any @option{-include} and @option{-imacros}
4063 Any @option{-D} and @option{-U} options on the command line are processed in
4064 order, and always before @option{-imacros @var{file}}, regardless of the
4065 order in which they are written.
4069 Tell the preprocessor to output only a list of the macro definitions
4070 that are in effect at the end of preprocessing. Used with the @option{-E}
4075 Tell the preprocessing to pass all macro definitions into the output, in
4076 their proper sequence in the rest of the output.
4080 Like @option{-dD} except that the macro arguments and contents are omitted.
4081 Only @samp{#define @var{name}} is included in the output.
4085 Output @samp{#include} directives in addition to the result of
4088 @item -fpreprocessed
4089 @opindex fpreprocessed
4090 Indicate to the preprocessor that the input file has already been
4091 preprocessed. This suppresses things like macro expansion, trigraph
4092 conversion, escaped newline splicing, and processing of most directives.
4093 The preprocessor still recognizes and removes comments, so that you can
4094 pass a file preprocessed with @option{-C} to the compiler without
4095 problems. In this mode the integrated preprocessor is little more than
4096 a tokenizer for the front ends.
4098 @option{-fpreprocessed} is implicit if the input file has one of the
4099 extensions @samp{i}, @samp{ii} or @samp{mi}. These are the extensions
4100 that GCC uses for preprocessed files created by @option{-save-temps}.
4104 Process ISO standard trigraph sequences. These are three-character
4105 sequences, all starting with @samp{??}, that are defined by ISO C to
4106 stand for single characters. For example, @samp{??/} stands for
4107 @samp{\}, so @samp{'??/n'} is a character constant for a newline. By
4108 default, GCC ignores trigraphs, but in standard-conforming modes it
4109 converts them. See the @option{-std} and @option{-ansi} options.
4111 The nine trigraph sequences are
4114 @expansion{} @samp{[}
4117 @expansion{} @samp{]}
4120 @expansion{} @samp{@{}
4123 @expansion{} @samp{@}}
4126 @expansion{} @samp{#}
4129 @expansion{} @samp{\}
4132 @expansion{} @samp{^}
4135 @expansion{} @samp{|}
4138 @expansion{} @samp{~}
4142 Trigraph support is not popular, so many compilers do not implement it
4143 properly. Portable code should not rely on trigraphs being either
4144 converted or ignored.
4146 @item -Wp,@var{option}
4148 Pass @var{option} as an option to the preprocessor. If @var{option}
4149 contains commas, it is split into multiple options at the commas.
4152 @node Assembler Options
4153 @section Passing Options to the Assembler
4155 @c prevent bad page break with this line
4156 You can pass options to the assembler.
4159 @item -Wa,@var{option}
4161 Pass @var{option} as an option to the assembler. If @var{option}
4162 contains commas, it is split into multiple options at the commas.
4166 @section Options for Linking
4167 @cindex link options
4168 @cindex options, linking
4170 These options come into play when the compiler links object files into
4171 an executable output file. They are meaningless if the compiler is
4172 not doing a link step.
4176 @item @var{object-file-name}
4177 A file name that does not end in a special recognized suffix is
4178 considered to name an object file or library. (Object files are
4179 distinguished from libraries by the linker according to the file
4180 contents.) If linking is done, these object files are used as input
4189 If any of these options is used, then the linker is not run, and
4190 object file names should not be used as arguments. @xref{Overall
4194 @item -l@var{library}
4195 @itemx -l @var{library}
4197 Search the library named @var{library} when linking. (The second
4198 alternative with the library as a separate argument is only for
4199 POSIX compliance and is not recommended.)
4201 It makes a difference where in the command you write this option; the
4202 linker searches and processes libraries and object files in the order they
4203 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4204 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4205 to functions in @samp{z}, those functions may not be loaded.
4207 The linker searches a standard list of directories for the library,
4208 which is actually a file named @file{lib@var{library}.a}. The linker
4209 then uses this file as if it had been specified precisely by name.
4211 The directories searched include several standard system directories
4212 plus any that you specify with @option{-L}.
4214 Normally the files found this way are library files---archive files
4215 whose members are object files. The linker handles an archive file by
4216 scanning through it for members which define symbols that have so far
4217 been referenced but not defined. But if the file that is found is an
4218 ordinary object file, it is linked in the usual fashion. The only
4219 difference between using an @option{-l} option and specifying a file name
4220 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4221 and searches several directories.
4225 You need this special case of the @option{-l} option in order to
4226 link an Objective-C program.
4229 @opindex nostartfiles
4230 Do not use the standard system startup files when linking.
4231 The standard system libraries are used normally, unless @option{-nostdlib}
4232 or @option{-nodefaultlibs} is used.
4234 @item -nodefaultlibs
4235 @opindex nodefaultlibs
4236 Do not use the standard system libraries when linking.
4237 Only the libraries you specify will be passed to the linker.
4238 The standard startup files are used normally, unless @option{-nostartfiles}
4239 is used. The compiler may generate calls to memcmp, memset, and memcpy
4240 for System V (and ISO C) environments or to bcopy and bzero for
4241 BSD environments. These entries are usually resolved by entries in
4242 libc. These entry points should be supplied through some other
4243 mechanism when this option is specified.
4247 Do not use the standard system startup files or libraries when linking.
4248 No startup files and only the libraries you specify will be passed to
4249 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4250 for System V (and ISO C) environments or to bcopy and bzero for
4251 BSD environments. These entries are usually resolved by entries in
4252 libc. These entry points should be supplied through some other
4253 mechanism when this option is specified.
4255 @cindex @option{-lgcc}, use with @option{-nostdlib}
4256 @cindex @option{-nostdlib} and unresolved references
4257 @cindex unresolved references and @option{-nostdlib}
4258 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4259 @cindex @option{-nodefaultlibs} and unresolved references
4260 @cindex unresolved references and @option{-nodefaultlibs}
4261 One of the standard libraries bypassed by @option{-nostdlib} and
4262 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4263 that GCC uses to overcome shortcomings of particular machines, or special
4264 needs for some languages.
4266 (@xref{Interface,,Interfacing to GCC Output}, for more discussion of
4270 (@xref{Interface,,Interfacing to GCC Output,gcc.info,Porting GCC},
4271 for more discussion of @file{libgcc.a}.)
4273 In most cases, you need @file{libgcc.a} even when you want to avoid
4274 other standard libraries. In other words, when you specify @option{-nostdlib}
4275 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4276 This ensures that you have no unresolved references to internal GCC
4277 library subroutines. (For example, @samp{__main}, used to ensure C++
4278 constructors will be called; @pxref{Collect2,,@command{collect2}}.)
4282 Remove all symbol table and relocation information from the executable.
4286 On systems that support dynamic linking, this prevents linking with the shared
4287 libraries. On other systems, this option has no effect.
4291 Produce a shared object which can then be linked with other objects to
4292 form an executable. Not all systems support this option. For predictable
4293 results, you must also specify the same set of options that were used to
4294 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4295 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4296 needs to build supplementary stub code for constructors to work. On
4297 multi-libbed systems, @samp{gcc -shared} must select the correct support
4298 libraries to link against. Failing to supply the correct flags may lead
4299 to subtle defects. Supplying them in cases where they are not necessary
4302 @item -shared-libgcc
4303 @itemx -static-libgcc
4304 @opindex shared-libgcc
4305 @opindex static-libgcc
4306 On systems that provide @file{libgcc} as a shared library, these options
4307 force the use of either the shared or static version respectively.
4308 If no shared version of @file{libgcc} was built when the compiler was
4309 configured, these options have no effect.
4311 There are several situations in which an application should use the
4312 shared @file{libgcc} instead of the static version. The most common
4313 of these is when the application wishes to throw and catch exceptions
4314 across different shared libraries. In that case, each of the libraries
4315 as well as the application itself should use the shared @file{libgcc}.
4317 Therefore, whenever you specify the @option{-shared} option, the GCC
4318 driver automatically adds @option{-shared-libgcc}, unless you explicitly
4319 specify @option{-static-libgcc}. The G++ driver automatically adds
4320 @option{-shared-libgcc} when you build a main executable as well because
4321 for C++ programs that is typically the right thing to do.
4322 (Exception-handling will not work reliably otherwise.)
4324 However, when linking a main executable written in C, you must
4325 explicitly say @option{-shared-libgcc} if you want to use the shared
4330 Bind references to global symbols when building a shared object. Warn
4331 about any unresolved references (unless overridden by the link editor
4332 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4335 @item -Xlinker @var{option}
4337 Pass @var{option} as an option to the linker. You can use this to
4338 supply system-specific linker options which GCC does not know how to
4341 If you want to pass an option that takes an argument, you must use
4342 @option{-Xlinker} twice, once for the option and once for the argument.
4343 For example, to pass @option{-assert definitions}, you must write
4344 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4345 @option{-Xlinker "-assert definitions"}, because this passes the entire
4346 string as a single argument, which is not what the linker expects.
4348 @item -Wl,@var{option}
4350 Pass @var{option} as an option to the linker. If @var{option} contains
4351 commas, it is split into multiple options at the commas.
4353 @item -u @var{symbol}
4355 Pretend the symbol @var{symbol} is undefined, to force linking of
4356 library modules to define it. You can use @option{-u} multiple times with
4357 different symbols to force loading of additional library modules.
4360 @node Directory Options
4361 @section Options for Directory Search
4362 @cindex directory options
4363 @cindex options, directory search
4366 These options specify directories to search for header files, for
4367 libraries and for parts of the compiler:
4372 Add the directory @var{dir} to the head of the list of directories to be
4373 searched for header files. This can be used to override a system header
4374 file, substituting your own version, since these directories are
4375 searched before the system header file directories. However, you should
4376 not use this option to add directories that contain vendor-supplied
4377 system header files (use @option{-isystem} for that). If you use more than
4378 one @option{-I} option, the directories are scanned in left-to-right
4379 order; the standard system directories come after.
4381 If a standard system include directory, or a directory specified with
4382 @option{-isystem}, is also specified with @option{-I}, it will be
4383 searched only in the position requested by @option{-I}. Also, it will
4384 not be considered a system include directory. If that directory really
4385 does contain system headers, there is a good chance that they will
4386 break. For instance, if GCC's installation procedure edited the headers
4387 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4388 original, buggy headers to be found instead of the corrected ones. GCC
4389 will issue a warning when a system include directory is hidden in this
4394 Any directories you specify with @option{-I} options before the @option{-I-}
4395 option are searched only for the case of @samp{#include "@var{file}"};
4396 they are not searched for @samp{#include <@var{file}>}.
4398 If additional directories are specified with @option{-I} options after
4399 the @option{-I-}, these directories are searched for all @samp{#include}
4400 directives. (Ordinarily @emph{all} @option{-I} directories are used
4403 In addition, the @option{-I-} option inhibits the use of the current
4404 directory (where the current input file came from) as the first search
4405 directory for @samp{#include "@var{file}"}. There is no way to
4406 override this effect of @option{-I-}. With @option{-I.} you can specify
4407 searching the directory which was current when the compiler was
4408 invoked. That is not exactly the same as what the preprocessor does
4409 by default, but it is often satisfactory.
4411 @option{-I-} does not inhibit the use of the standard system directories
4412 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4417 Add directory @var{dir} to the list of directories to be searched
4420 @item -B@var{prefix}
4422 This option specifies where to find the executables, libraries,
4423 include files, and data files of the compiler itself.
4425 The compiler driver program runs one or more of the subprograms
4426 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4427 @var{prefix} as a prefix for each program it tries to run, both with and
4428 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4430 For each subprogram to be run, the compiler driver first tries the
4431 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4432 was not specified, the driver tries two standard prefixes, which are
4433 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4434 those results in a file name that is found, the unmodified program
4435 name is searched for using the directories specified in your
4436 @env{PATH} environment variable.
4438 The compiler will check to see if the path provided by the @option{-B}
4439 refers to a directory, and if necessary it will add a directory
4440 separator character at the end of the path.
4442 @option{-B} prefixes that effectively specify directory names also apply
4443 to libraries in the linker, because the compiler translates these
4444 options into @option{-L} options for the linker. They also apply to
4445 includes files in the preprocessor, because the compiler translates these
4446 options into @option{-isystem} options for the preprocessor. In this case,
4447 the compiler appends @samp{include} to the prefix.
4449 The run-time support file @file{libgcc.a} can also be searched for using
4450 the @option{-B} prefix, if needed. If it is not found there, the two
4451 standard prefixes above are tried, and that is all. The file is left
4452 out of the link if it is not found by those means.
4454 Another way to specify a prefix much like the @option{-B} prefix is to use
4455 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4458 As a special kludge, if the path provided by @option{-B} is
4459 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4460 9, then it will be replaced by @file{[dir/]include}. This is to help
4461 with boot-strapping the compiler.
4463 @item -specs=@var{file}
4465 Process @var{file} after the compiler reads in the standard @file{specs}
4466 file, in order to override the defaults that the @file{gcc} driver
4467 program uses when determining what switches to pass to @file{cc1},
4468 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4469 @option{-specs=@var{file}} can be specified on the command line, and they
4470 are processed in order, from left to right.
4476 @section Specifying subprocesses and the switches to pass to them
4478 @command{gcc} is a driver program. It performs its job by invoking a
4479 sequence of other programs to do the work of compiling, assembling and
4480 linking. GCC interprets its command-line parameters and uses these to
4481 deduce which programs it should invoke, and which command-line options
4482 it ought to place on their command lines. This behavior is controlled
4483 by @dfn{spec strings}. In most cases there is one spec string for each
4484 program that GCC can invoke, but a few programs have multiple spec
4485 strings to control their behavior. The spec strings built into GCC can
4486 be overridden by using the @option{-specs=} command-line switch to specify
4489 @dfn{Spec files} are plaintext files that are used to construct spec
4490 strings. They consist of a sequence of directives separated by blank
4491 lines. The type of directive is determined by the first non-whitespace
4492 character on the line and it can be one of the following:
4495 @item %@var{command}
4496 Issues a @var{command} to the spec file processor. The commands that can
4500 @item %include <@var{file}>
4502 Search for @var{file} and insert its text at the current point in the
4505 @item %include_noerr <@var{file}>
4506 @cindex %include_noerr
4507 Just like @samp{%include}, but do not generate an error message if the include
4508 file cannot be found.
4510 @item %rename @var{old_name} @var{new_name}
4512 Rename the spec string @var{old_name} to @var{new_name}.
4516 @item *[@var{spec_name}]:
4517 This tells the compiler to create, override or delete the named spec
4518 string. All lines after this directive up to the next directive or
4519 blank line are considered to be the text for the spec string. If this
4520 results in an empty string then the spec will be deleted. (Or, if the
4521 spec did not exist, then nothing will happened.) Otherwise, if the spec
4522 does not currently exist a new spec will be created. If the spec does
4523 exist then its contents will be overridden by the text of this
4524 directive, unless the first character of that text is the @samp{+}
4525 character, in which case the text will be appended to the spec.
4527 @item [@var{suffix}]:
4528 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4529 and up to the next directive or blank line are considered to make up the
4530 spec string for the indicated suffix. When the compiler encounters an
4531 input file with the named suffix, it will processes the spec string in
4532 order to work out how to compile that file. For example:
4539 This says that any input file whose name ends in @samp{.ZZ} should be
4540 passed to the program @samp{z-compile}, which should be invoked with the
4541 command-line switch @option{-input} and with the result of performing the
4542 @samp{%i} substitution. (See below.)
4544 As an alternative to providing a spec string, the text that follows a
4545 suffix directive can be one of the following:
4548 @item @@@var{language}
4549 This says that the suffix is an alias for a known @var{language}. This is
4550 similar to using the @option{-x} command-line switch to GCC to specify a
4551 language explicitly. For example:
4558 Says that .ZZ files are, in fact, C++ source files.
4561 This causes an error messages saying:
4564 @var{name} compiler not installed on this system.
4568 GCC already has an extensive list of suffixes built into it.
4569 This directive will add an entry to the end of the list of suffixes, but
4570 since the list is searched from the end backwards, it is effectively
4571 possible to override earlier entries using this technique.
4575 GCC has the following spec strings built into it. Spec files can
4576 override these strings or create their own. Note that individual
4577 targets can also add their own spec strings to this list.
4580 asm Options to pass to the assembler
4581 asm_final Options to pass to the assembler post-processor
4582 cpp Options to pass to the C preprocessor
4583 cc1 Options to pass to the C compiler
4584 cc1plus Options to pass to the C++ compiler
4585 endfile Object files to include at the end of the link
4586 link Options to pass to the linker
4587 lib Libraries to include on the command line to the linker
4588 libgcc Decides which GCC support library to pass to the linker
4589 linker Sets the name of the linker
4590 predefines Defines to be passed to the C preprocessor
4591 signed_char Defines to pass to CPP to say whether @code{char} is signed
4593 startfile Object files to include at the start of the link
4596 Here is a small example of a spec file:
4602 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4605 This example renames the spec called @samp{lib} to @samp{old_lib} and
4606 then overrides the previous definition of @samp{lib} with a new one.
4607 The new definition adds in some extra command-line options before
4608 including the text of the old definition.
4610 @dfn{Spec strings} are a list of command-line options to be passed to their
4611 corresponding program. In addition, the spec strings can contain
4612 @samp{%}-prefixed sequences to substitute variable text or to
4613 conditionally insert text into the command line. Using these constructs
4614 it is possible to generate quite complex command lines.
4616 Here is a table of all defined @samp{%}-sequences for spec
4617 strings. Note that spaces are not generated automatically around the
4618 results of expanding these sequences. Therefore you can concatenate them
4619 together or combine them with constant text in a single argument.
4623 Substitute one @samp{%} into the program name or argument.
4626 Substitute the name of the input file being processed.
4629 Substitute the basename of the input file being processed.
4630 This is the substring up to (and not including) the last period
4631 and not including the directory.
4634 This is the same as @samp{%b}, but include the file suffix (text after
4638 Marks the argument containing or following the @samp{%d} as a
4639 temporary file name, so that that file will be deleted if GCC exits
4640 successfully. Unlike @samp{%g}, this contributes no text to the
4643 @item %g@var{suffix}
4644 Substitute a file name that has suffix @var{suffix} and is chosen
4645 once per compilation, and mark the argument in the same way as
4646 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4647 name is now chosen in a way that is hard to predict even when previously
4648 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4649 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4650 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4651 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4652 was simply substituted with a file name chosen once per compilation,
4653 without regard to any appended suffix (which was therefore treated
4654 just like ordinary text), making such attacks more likely to succeed.
4656 @item %u@var{suffix}
4657 Like @samp{%g}, but generates a new temporary file name even if
4658 @samp{%u@var{suffix}} was already seen.
4660 @item %U@var{suffix}
4661 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4662 new one if there is no such last file name. In the absence of any
4663 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4664 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4665 would involve the generation of two distinct file names, one
4666 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4667 simply substituted with a file name chosen for the previous @samp{%u},
4668 without regard to any appended suffix.
4670 @item %j@var{SUFFIX}
4671 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4672 writable, and if save-temps is off; otherwise, substitute the name
4673 of a temporary file, just like @samp{%u}. This temporary file is not
4674 meant for communication between processes, but rather as a junk
4677 @item %.@var{SUFFIX}
4678 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4679 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4680 terminated by the next space or %.
4683 Marks the argument containing or following the @samp{%w} as the
4684 designated output file of this compilation. This puts the argument
4685 into the sequence of arguments that @samp{%o} will substitute later.
4688 Substitutes the names of all the output files, with spaces
4689 automatically placed around them. You should write spaces
4690 around the @samp{%o} as well or the results are undefined.
4691 @samp{%o} is for use in the specs for running the linker.
4692 Input files whose names have no recognized suffix are not compiled
4693 at all, but they are included among the output files, so they will
4697 Substitutes the suffix for object files. Note that this is
4698 handled specially when it immediately follows @samp{%g, %u, or %U},
4699 because of the need for those to form complete file names. The
4700 handling is such that @samp{%O} is treated exactly as if it had already
4701 been substituted, except that @samp{%g, %u, and %U} do not currently
4702 support additional @var{suffix} characters following @samp{%O} as they would
4703 following, for example, @samp{.o}.
4706 Substitutes the standard macro predefinitions for the
4707 current target machine. Use this when running @code{cpp}.
4710 Like @samp{%p}, but puts @samp{__} before and after the name of each
4711 predefined macro, except for macros that start with @samp{__} or with
4712 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4716 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4719 Current argument is the name of a library or startup file of some sort.
4720 Search for that file in a standard list of directories and substitute
4721 the full name found.
4724 Print @var{str} as an error message. @var{str} is terminated by a newline.
4725 Use this when inconsistent options are detected.
4728 Output @samp{-} if the input for the current command is coming from a pipe.
4731 Substitute the contents of spec string @var{name} at this point.
4734 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4736 @item %x@{@var{option}@}
4737 Accumulate an option for @samp{%X}.
4740 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4744 Output the accumulated assembler options specified by @option{-Wa}.
4747 Output the accumulated preprocessor options specified by @option{-Wp}.
4750 Substitute the major version number of GCC@.
4751 (For version 2.9.5, this is 2.)
4754 Substitute the minor version number of GCC@.
4755 (For version 2.9.5, this is 9.)
4758 Substitute the patch level number of GCC@.
4759 (For version 2.9.5, this is 5.)
4762 Process the @code{asm} spec. This is used to compute the
4763 switches to be passed to the assembler.
4766 Process the @code{asm_final} spec. This is a spec string for
4767 passing switches to an assembler post-processor, if such a program is
4771 Process the @code{link} spec. This is the spec for computing the
4772 command line passed to the linker. Typically it will make use of the
4773 @samp{%L %G %S %D and %E} sequences.
4776 Dump out a @option{-L} option for each directory that GCC believes might
4777 contain startup files. If the target supports multilibs then the
4778 current multilib directory will be prepended to each of these paths.
4781 Output the multilib directory with directory separators replaced with
4782 @samp{_}. If multilib directories are not set, or the multilib directory is
4783 @file{.} then this option emits nothing.
4786 Process the @code{lib} spec. This is a spec string for deciding which
4787 libraries should be included on the command line to the linker.
4790 Process the @code{libgcc} spec. This is a spec string for deciding
4791 which GCC support library should be included on the command line to the linker.
4794 Process the @code{startfile} spec. This is a spec for deciding which
4795 object files should be the first ones passed to the linker. Typically
4796 this might be a file named @file{crt0.o}.
4799 Process the @code{endfile} spec. This is a spec string that specifies
4800 the last object files that will be passed to the linker.
4803 Process the @code{cpp} spec. This is used to construct the arguments
4804 to be passed to the C preprocessor.
4807 Process the @code{signed_char} spec. This is intended to be used
4808 to tell cpp whether a char is signed. It typically has the definition:
4810 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4814 Process the @code{cc1} spec. This is used to construct the options to be
4815 passed to the actual C compiler (@samp{cc1}).
4818 Process the @code{cc1plus} spec. This is used to construct the options to be
4819 passed to the actual C++ compiler (@samp{cc1plus}).
4822 Substitute the variable part of a matched option. See below.
4823 Note that each comma in the substituted string is replaced by
4827 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4828 If that switch was not specified, this substitutes nothing. Note that
4829 the leading dash is omitted when specifying this option, and it is
4830 automatically inserted if the substitution is performed. Thus the spec
4831 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4832 and would output the command line option @option{-foo}.
4834 @item %W@{@code{S}@}
4835 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4838 @item %@{@code{S}*@}
4839 Substitutes all the switches specified to GCC whose names start
4840 with @code{-S}, but which also take an argument. This is used for
4841 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4842 GCC considers @option{-o foo} as being
4843 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4844 text, including the space. Thus two arguments would be generated.
4846 @item %@{^@code{S}*@}
4847 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4848 argument. Thus %@{^o*@} would only generate one argument, not two.
4850 @item %@{@code{S}*&@code{T}*@}
4851 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4852 (the order of @code{S} and @code{T} in the spec is not significant).
4853 There can be any number of ampersand-separated variables; for each the
4854 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4856 @item %@{<@code{S}@}
4857 Remove all occurrences of @code{-S} from the command line. Note---this
4858 command is position dependent. @samp{%} commands in the spec string
4859 before this option will see @code{-S}, @samp{%} commands in the spec
4860 string after this option will not.
4862 @item %@{@code{S}*:@code{X}@}
4863 Substitutes @code{X} if one or more switches whose names start with
4864 @code{-S} are specified to GCC@. Note that the tail part of the
4865 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4866 for each occurrence of @samp{%*} within @code{X}.
4868 @item %@{@code{S}:@code{X}@}
4869 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4871 @item %@{!@code{S}:@code{X}@}
4872 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4874 @item %@{|@code{S}:@code{X}@}
4875 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4877 @item %@{|!@code{S}:@code{X}@}
4878 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4880 @item %@{.@code{S}:@code{X}@}
4881 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4883 @item %@{!.@code{S}:@code{X}@}
4884 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4886 @item %@{@code{S}|@code{P}:@code{X}@}
4887 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4888 combined with @samp{!} and @samp{.} sequences as well, although they
4889 have a stronger binding than the @samp{|}. For example a spec string
4893 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4896 will output the following command-line options from the following input
4897 command-line options:
4902 -d fred.c -foo -baz -boggle
4903 -d jim.d -bar -baz -boggle
4908 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4909 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4910 or spaces, or even newlines. They are processed as usual, as described
4913 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4914 switches are handled specifically in these
4915 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4916 @option{-W} switch is found later in the command line, the earlier switch
4917 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4918 letter, which passes all matching options.
4920 The character @samp{|} at the beginning of the predicate text is used to indicate
4921 that a command should be piped to the following command, but only if @option{-pipe}
4924 It is built into GCC which switches take arguments and which do not.
4925 (You might think it would be useful to generalize this to allow each
4926 compiler's spec to say which switches take arguments. But this cannot
4927 be done in a consistent fashion. GCC cannot even decide which input
4928 files have been specified without knowing which switches take arguments,
4929 and it must know which input files to compile in order to tell which
4932 GCC also knows implicitly that arguments starting in @option{-l} are to be
4933 treated as compiler output files, and passed to the linker in their
4934 proper position among the other output files.
4936 @c man begin OPTIONS
4938 @node Target Options
4939 @section Specifying Target Machine and Compiler Version
4940 @cindex target options
4941 @cindex cross compiling
4942 @cindex specifying machine version
4943 @cindex specifying compiler version and target machine
4944 @cindex compiler version, specifying
4945 @cindex target machine, specifying
4947 By default, GCC compiles code for the same type of machine that you
4948 are using. However, it can also be installed as a cross-compiler, to
4949 compile for some other type of machine. In fact, several different
4950 configurations of GCC, for different target machines, can be
4951 installed side by side. Then you specify which one to use with the
4954 In addition, older and newer versions of GCC can be installed side
4955 by side. One of them (probably the newest) will be the default, but
4956 you may sometimes wish to use another.
4959 @item -b @var{machine}
4961 The argument @var{machine} specifies the target machine for compilation.
4962 This is useful when you have installed GCC as a cross-compiler.
4964 The value to use for @var{machine} is the same as was specified as the
4965 machine type when configuring GCC as a cross-compiler. For
4966 example, if a cross-compiler was configured with @samp{configure
4967 i386v}, meaning to compile for an 80386 running System V, then you
4968 would specify @option{-b i386v} to run that cross compiler.
4970 When you do not specify @option{-b}, it normally means to compile for
4971 the same type of machine that you are using.
4973 @item -V @var{version}
4975 The argument @var{version} specifies which version of GCC to run.
4976 This is useful when multiple versions are installed. For example,
4977 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
4979 The default version, when you do not specify @option{-V}, is the last
4980 version of GCC that you installed.
4983 The @option{-b} and @option{-V} options actually work by controlling part of
4984 the file name used for the executable files and libraries used for
4985 compilation. A given version of GCC, for a given target machine, is
4986 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
4988 Thus, sites can customize the effect of @option{-b} or @option{-V} either by
4989 changing the names of these directories or adding alternate names (or
4990 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
4991 file @file{80386} is a link to the file @file{i386v}, then @option{-b
4992 80386} becomes an alias for @option{-b i386v}.
4994 In one respect, the @option{-b} or @option{-V} do not completely change
4995 to a different compiler: the top-level driver program @command{gcc}
4996 that you originally invoked continues to run and invoke the other
4997 executables (preprocessor, compiler per se, assembler and linker)
4998 that do the real work. However, since no real work is done in the
4999 driver program, it usually does not matter that the driver program
5000 in use is not the one for the specified target. It is common for the
5001 interface to the other executables to change incompatibly between
5002 compiler versions, so unless the version specified is very close to that
5003 of the driver (for example, @option{-V 3.0} with a driver program from GCC
5004 version 3.0.1), use of @option{-V} may not work; for example, using
5005 @option{-V 2.95.2} will not work with a driver program from GCC 3.0.
5007 The only way that the driver program depends on the target machine is
5008 in the parsing and handling of special machine-specific options.
5009 However, this is controlled by a file which is found, along with the
5010 other executables, in the directory for the specified version and
5011 target machine. As a result, a single installed driver program adapts
5012 to any specified target machine, and sufficiently similar compiler
5015 The driver program executable does control one significant thing,
5016 however: the default version and target machine. Therefore, you can
5017 install different instances of the driver program, compiled for
5018 different targets or versions, under different names.
5020 For example, if the driver for version 2.0 is installed as @command{ogcc}
5021 and that for version 2.1 is installed as @command{gcc}, then the command
5022 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
5023 2.0 by default. However, you can choose either version with either
5024 command with the @option{-V} option.
5026 @node Submodel Options
5027 @section Hardware Models and Configurations
5028 @cindex submodel options
5029 @cindex specifying hardware config
5030 @cindex hardware models and configurations, specifying
5031 @cindex machine dependent options
5033 Earlier we discussed the standard option @option{-b} which chooses among
5034 different installed compilers for completely different target
5035 machines, such as VAX vs.@: 68000 vs.@: 80386.
5037 In addition, each of these target machine types can have its own
5038 special options, starting with @samp{-m}, to choose among various
5039 hardware models or configurations---for example, 68010 vs 68020,
5040 floating coprocessor or none. A single installed version of the
5041 compiler can compile for any model or configuration, according to the
5044 Some configurations of the compiler also support additional special
5045 options, usually for compatibility with other compilers on the same
5049 These options are defined by the macro @code{TARGET_SWITCHES} in the
5050 machine description. The default for the options is also defined by
5051 that macro, which enables you to change the defaults.
5066 * RS/6000 and PowerPC Options::
5069 * i386 and x86-64 Options::
5071 * Intel 960 Options::
5072 * DEC Alpha Options::
5076 * System V Options::
5077 * TMS320C3x/C4x Options::
5085 * S/390 and zSeries Options::
5090 @node M680x0 Options
5091 @subsection M680x0 Options
5092 @cindex M680x0 options
5094 These are the @samp{-m} options defined for the 68000 series. The default
5095 values for these options depends on which style of 68000 was selected when
5096 the compiler was configured; the defaults for the most common choices are
5104 Generate output for a 68000. This is the default
5105 when the compiler is configured for 68000-based systems.
5107 Use this option for microcontrollers with a 68000 or EC000 core,
5108 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5114 Generate output for a 68020. This is the default
5115 when the compiler is configured for 68020-based systems.
5119 Generate output containing 68881 instructions for floating point.
5120 This is the default for most 68020 systems unless @option{--nfp} was
5121 specified when the compiler was configured.
5125 Generate output for a 68030. This is the default when the compiler is
5126 configured for 68030-based systems.
5130 Generate output for a 68040. This is the default when the compiler is
5131 configured for 68040-based systems.
5133 This option inhibits the use of 68881/68882 instructions that have to be
5134 emulated by software on the 68040. Use this option if your 68040 does not
5135 have code to emulate those instructions.
5139 Generate output for a 68060. This is the default when the compiler is
5140 configured for 68060-based systems.
5142 This option inhibits the use of 68020 and 68881/68882 instructions that
5143 have to be emulated by software on the 68060. Use this option if your 68060
5144 does not have code to emulate those instructions.
5148 Generate output for a CPU32. This is the default
5149 when the compiler is configured for CPU32-based systems.
5151 Use this option for microcontrollers with a
5152 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5153 68336, 68340, 68341, 68349 and 68360.
5157 Generate output for a 520X ``coldfire'' family cpu. This is the default
5158 when the compiler is configured for 520X-based systems.
5160 Use this option for microcontroller with a 5200 core, including
5161 the MCF5202, MCF5203, MCF5204 and MCF5202.
5166 Generate output for a 68040, without using any of the new instructions.
5167 This results in code which can run relatively efficiently on either a
5168 68020/68881 or a 68030 or a 68040. The generated code does use the
5169 68881 instructions that are emulated on the 68040.
5173 Generate output for a 68060, without using any of the new instructions.
5174 This results in code which can run relatively efficiently on either a
5175 68020/68881 or a 68030 or a 68040. The generated code does use the
5176 68881 instructions that are emulated on the 68060.
5180 Generate output containing Sun FPA instructions for floating point.
5183 @opindex msoft-float
5184 Generate output containing library calls for floating point.
5185 @strong{Warning:} the requisite libraries are not available for all m68k
5186 targets. Normally the facilities of the machine's usual C compiler are
5187 used, but this can't be done directly in cross-compilation. You must
5188 make your own arrangements to provide suitable library functions for
5189 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5190 @samp{m68k-*-coff} do provide software floating point support.
5194 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5197 @opindex mnobitfield
5198 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5199 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5203 Do use the bit-field instructions. The @option{-m68020} option implies
5204 @option{-mbitfield}. This is the default if you use a configuration
5205 designed for a 68020.
5209 Use a different function-calling convention, in which functions
5210 that take a fixed number of arguments return with the @code{rtd}
5211 instruction, which pops their arguments while returning. This
5212 saves one instruction in the caller since there is no need to pop
5213 the arguments there.
5215 This calling convention is incompatible with the one normally
5216 used on Unix, so you cannot use it if you need to call libraries
5217 compiled with the Unix compiler.
5219 Also, you must provide function prototypes for all functions that
5220 take variable numbers of arguments (including @code{printf});
5221 otherwise incorrect code will be generated for calls to those
5224 In addition, seriously incorrect code will result if you call a
5225 function with too many arguments. (Normally, extra arguments are
5226 harmlessly ignored.)
5228 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5229 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5232 @itemx -mno-align-int
5234 @opindex mno-align-int
5235 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5236 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5237 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5238 Aligning variables on 32-bit boundaries produces code that runs somewhat
5239 faster on processors with 32-bit busses at the expense of more memory.
5241 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5242 align structures containing the above types differently than
5243 most published application binary interface specifications for the m68k.
5247 Use the pc-relative addressing mode of the 68000 directly, instead of
5248 using a global offset table. At present, this option implies @option{-fpic},
5249 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5250 not presently supported with @option{-mpcrel}, though this could be supported for
5251 68020 and higher processors.
5253 @item -mno-strict-align
5254 @itemx -mstrict-align
5255 @opindex mno-strict-align
5256 @opindex mstrict-align
5257 Do not (do) assume that unaligned memory references will be handled by
5262 @node M68hc1x Options
5263 @subsection M68hc1x Options
5264 @cindex M68hc1x options
5266 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5267 microcontrollers. The default values for these options depends on
5268 which style of microcontroller was selected when the compiler was configured;
5269 the defaults for the most common choices are given below.
5276 Generate output for a 68HC11. This is the default
5277 when the compiler is configured for 68HC11-based systems.
5283 Generate output for a 68HC12. This is the default
5284 when the compiler is configured for 68HC12-based systems.
5287 @opindex mauto-incdec
5288 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5293 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5295 @item -msoft-reg-count=@var{count}
5296 @opindex msoft-reg-count
5297 Specify the number of pseudo-soft registers which are used for the
5298 code generation. The maximum number is 32. Using more pseudo-soft
5299 register may or may not result in better code depending on the program.
5300 The default is 4 for 68HC11 and 2 for 68HC12.
5305 @subsection VAX Options
5308 These @samp{-m} options are defined for the VAX:
5313 Do not output certain jump instructions (@code{aobleq} and so on)
5314 that the Unix assembler for the VAX cannot handle across long
5319 Do output those jump instructions, on the assumption that you
5320 will assemble with the GNU assembler.
5324 Output code for g-format floating point numbers instead of d-format.
5328 @subsection SPARC Options
5329 @cindex SPARC options
5331 These @samp{-m} switches are supported on the SPARC:
5336 @opindex mno-app-regs
5338 Specify @option{-mapp-regs} to generate output using the global registers
5339 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5342 To be fully SVR4 ABI compliant at the cost of some performance loss,
5343 specify @option{-mno-app-regs}. You should compile libraries and system
5344 software with this option.
5349 @opindex mhard-float
5350 Generate output containing floating point instructions. This is the
5356 @opindex msoft-float
5357 Generate output containing library calls for floating point.
5358 @strong{Warning:} the requisite libraries are not available for all SPARC
5359 targets. Normally the facilities of the machine's usual C compiler are
5360 used, but this cannot be done directly in cross-compilation. You must make
5361 your own arrangements to provide suitable library functions for
5362 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5363 @samp{sparclite-*-*} do provide software floating point support.
5365 @option{-msoft-float} changes the calling convention in the output file;
5366 therefore, it is only useful if you compile @emph{all} of a program with
5367 this option. In particular, you need to compile @file{libgcc.a}, the
5368 library that comes with GCC, with @option{-msoft-float} in order for
5371 @item -mhard-quad-float
5372 @opindex mhard-quad-float
5373 Generate output containing quad-word (long double) floating point
5376 @item -msoft-quad-float
5377 @opindex msoft-quad-float
5378 Generate output containing library calls for quad-word (long double)
5379 floating point instructions. The functions called are those specified
5380 in the SPARC ABI@. This is the default.
5382 As of this writing, there are no sparc implementations that have hardware
5383 support for the quad-word floating point instructions. They all invoke
5384 a trap handler for one of these instructions, and then the trap handler
5385 emulates the effect of the instruction. Because of the trap handler overhead,
5386 this is much slower than calling the ABI library routines. Thus the
5387 @option{-msoft-quad-float} option is the default.
5391 @opindex mno-epilogue
5393 With @option{-mepilogue} (the default), the compiler always emits code for
5394 function exit at the end of each function. Any function exit in
5395 the middle of the function (such as a return statement in C) will
5396 generate a jump to the exit code at the end of the function.
5398 With @option{-mno-epilogue}, the compiler tries to emit exit code inline
5399 at every function exit.
5405 With @option{-mflat}, the compiler does not generate save/restore instructions
5406 and will use a ``flat'' or single register window calling convention.
5407 This model uses %i7 as the frame pointer and is compatible with the normal
5408 register window model. Code from either may be intermixed.
5409 The local registers and the input registers (0--5) are still treated as
5410 ``call saved'' registers and will be saved on the stack as necessary.
5412 With @option{-mno-flat} (the default), the compiler emits save/restore
5413 instructions (except for leaf functions) and is the normal mode of operation.
5415 @item -mno-unaligned-doubles
5416 @itemx -munaligned-doubles
5417 @opindex mno-unaligned-doubles
5418 @opindex munaligned-doubles
5419 Assume that doubles have 8 byte alignment. This is the default.
5421 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5422 alignment only if they are contained in another type, or if they have an
5423 absolute address. Otherwise, it assumes they have 4 byte alignment.
5424 Specifying this option avoids some rare compatibility problems with code
5425 generated by other compilers. It is not the default because it results
5426 in a performance loss, especially for floating point code.
5428 @item -mno-faster-structs
5429 @itemx -mfaster-structs
5430 @opindex mno-faster-structs
5431 @opindex mfaster-structs
5432 With @option{-mfaster-structs}, the compiler assumes that structures
5433 should have 8 byte alignment. This enables the use of pairs of
5434 @code{ldd} and @code{std} instructions for copies in structure
5435 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5436 However, the use of this changed alignment directly violates the Sparc
5437 ABI@. Thus, it's intended only for use on targets where the developer
5438 acknowledges that their resulting code will not be directly in line with
5439 the rules of the ABI@.
5445 These two options select variations on the SPARC architecture.
5447 By default (unless specifically configured for the Fujitsu SPARClite),
5448 GCC generates code for the v7 variant of the SPARC architecture.
5450 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5451 code is that the compiler emits the integer multiply and integer
5452 divide instructions which exist in SPARC v8 but not in SPARC v7.
5454 @option{-msparclite} will give you SPARClite code. This adds the integer
5455 multiply, integer divide step and scan (@code{ffs}) instructions which
5456 exist in SPARClite but not in SPARC v7.
5458 These options are deprecated and will be deleted in a future GCC release.
5459 They have been replaced with @option{-mcpu=xxx}.
5464 @opindex msupersparc
5465 These two options select the processor for which the code is optimized.
5467 With @option{-mcypress} (the default), the compiler optimizes code for the
5468 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5469 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5471 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5472 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5473 of the full SPARC v8 instruction set.
5475 These options are deprecated and will be deleted in a future GCC release.
5476 They have been replaced with @option{-mcpu=xxx}.
5478 @item -mcpu=@var{cpu_type}
5480 Set the instruction set, register set, and instruction scheduling parameters
5481 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5482 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5483 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5484 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5486 Default instruction scheduling parameters are used for values that select
5487 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5488 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5490 Here is a list of each supported architecture and their supported
5495 v8: supersparc, hypersparc
5496 sparclite: f930, f934, sparclite86x
5501 @item -mtune=@var{cpu_type}
5503 Set the instruction scheduling parameters for machine type
5504 @var{cpu_type}, but do not set the instruction set or register set that the
5505 option @option{-mcpu=@var{cpu_type}} would.
5507 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5508 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5509 that select a particular cpu implementation. Those are @samp{cypress},
5510 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5511 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5515 These @samp{-m} switches are supported in addition to the above
5516 on the SPARCLET processor.
5519 @item -mlittle-endian
5520 @opindex mlittle-endian
5521 Generate code for a processor running in little-endian mode.
5525 Treat register @code{%g0} as a normal register.
5526 GCC will continue to clobber it as necessary but will not assume
5527 it always reads as 0.
5529 @item -mbroken-saverestore
5530 @opindex mbroken-saverestore
5531 Generate code that does not use non-trivial forms of the @code{save} and
5532 @code{restore} instructions. Early versions of the SPARCLET processor do
5533 not correctly handle @code{save} and @code{restore} instructions used with
5534 arguments. They correctly handle them used without arguments. A @code{save}
5535 instruction used without arguments increments the current window pointer
5536 but does not allocate a new stack frame. It is assumed that the window
5537 overflow trap handler will properly handle this case as will interrupt
5541 These @samp{-m} switches are supported in addition to the above
5542 on SPARC V9 processors in 64-bit environments.
5545 @item -mlittle-endian
5546 @opindex mlittle-endian
5547 Generate code for a processor running in little-endian mode.
5553 Generate code for a 32-bit or 64-bit environment.
5554 The 32-bit environment sets int, long and pointer to 32 bits.
5555 The 64-bit environment sets int to 32 bits and long and pointer
5558 @item -mcmodel=medlow
5559 @opindex mcmodel=medlow
5560 Generate code for the Medium/Low code model: the program must be linked
5561 in the low 32 bits of the address space. Pointers are 64 bits.
5562 Programs can be statically or dynamically linked.
5564 @item -mcmodel=medmid
5565 @opindex mcmodel=medmid
5566 Generate code for the Medium/Middle code model: the program must be linked
5567 in the low 44 bits of the address space, the text segment must be less than
5568 2G bytes, and data segment must be within 2G of the text segment.
5569 Pointers are 64 bits.
5571 @item -mcmodel=medany
5572 @opindex mcmodel=medany
5573 Generate code for the Medium/Anywhere code model: the program may be linked
5574 anywhere in the address space, the text segment must be less than
5575 2G bytes, and data segment must be within 2G of the text segment.
5576 Pointers are 64 bits.
5578 @item -mcmodel=embmedany
5579 @opindex mcmodel=embmedany
5580 Generate code for the Medium/Anywhere code model for embedded systems:
5581 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5582 (determined at link time). Register %g4 points to the base of the
5583 data segment. Pointers are still 64 bits.
5584 Programs are statically linked, PIC is not supported.
5587 @itemx -mno-stack-bias
5588 @opindex mstack-bias
5589 @opindex mno-stack-bias
5590 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5591 frame pointer if present, are offset by @minus{}2047 which must be added back
5592 when making stack frame references.
5593 Otherwise, assume no such offset is present.
5596 @node Convex Options
5597 @subsection Convex Options
5598 @cindex Convex options
5600 These @samp{-m} options are defined for Convex:
5605 Generate output for C1. The code will run on any Convex machine.
5606 The preprocessor symbol @code{__convex__c1__} is defined.
5610 Generate output for C2. Uses instructions not available on C1.
5611 Scheduling and other optimizations are chosen for max performance on C2.
5612 The preprocessor symbol @code{__convex_c2__} is defined.
5616 Generate output for C32xx. Uses instructions not available on C1.
5617 Scheduling and other optimizations are chosen for max performance on C32.
5618 The preprocessor symbol @code{__convex_c32__} is defined.
5622 Generate output for C34xx. Uses instructions not available on C1.
5623 Scheduling and other optimizations are chosen for max performance on C34.
5624 The preprocessor symbol @code{__convex_c34__} is defined.
5628 Generate output for C38xx. Uses instructions not available on C1.
5629 Scheduling and other optimizations are chosen for max performance on C38.
5630 The preprocessor symbol @code{__convex_c38__} is defined.
5634 Generate code which puts an argument count in the word preceding each
5635 argument list. This is compatible with regular CC, and a few programs
5636 may need the argument count word. GDB and other source-level debuggers
5637 do not need it; this info is in the symbol table.
5640 @opindex mnoargcount
5641 Omit the argument count word. This is the default.
5643 @item -mvolatile-cache
5644 @opindex mvolatile-cache
5645 Allow volatile references to be cached. This is the default.
5647 @item -mvolatile-nocache
5648 @opindex mvolatile-nocache
5649 Volatile references bypass the data cache, going all the way to memory.
5650 This is only needed for multi-processor code that does not use standard
5651 synchronization instructions. Making non-volatile references to volatile
5652 locations will not necessarily work.
5656 Type long is 32 bits, the same as type int. This is the default.
5660 Type long is 64 bits, the same as type long long. This option is useless,
5661 because no library support exists for it.
5664 @node AMD29K Options
5665 @subsection AMD29K Options
5666 @cindex AMD29K options
5668 These @samp{-m} options are defined for the AMD Am29000:
5673 @cindex DW bit (29k)
5674 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5675 halfword operations are directly supported by the hardware. This is the
5680 Generate code that assumes the @code{DW} bit is not set.
5684 @cindex byte writes (29k)
5685 Generate code that assumes the system supports byte and halfword write
5686 operations. This is the default.
5690 Generate code that assumes the systems does not support byte and
5691 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5695 @cindex memory model (29k)
5696 Use a small memory model that assumes that all function addresses are
5697 either within a single 256 KB segment or at an absolute address of less
5698 than 256k. This allows the @code{call} instruction to be used instead
5699 of a @code{const}, @code{consth}, @code{calli} sequence.
5703 Use the normal memory model: Generate @code{call} instructions only when
5704 calling functions in the same file and @code{calli} instructions
5705 otherwise. This works if each file occupies less than 256 KB but allows
5706 the entire executable to be larger than 256 KB@. This is the default.
5710 Always use @code{calli} instructions. Specify this option if you expect
5711 a single file to compile into more than 256 KB of code.
5715 @cindex processor selection (29k)
5716 Generate code for the Am29050.
5720 Generate code for the Am29000. This is the default.
5722 @item -mkernel-registers
5723 @opindex mkernel-registers
5724 @cindex kernel and user registers (29k)
5725 Generate references to registers @code{gr64-gr95} instead of to
5726 registers @code{gr96-gr127}. This option can be used when compiling
5727 kernel code that wants a set of global registers disjoint from that used
5730 Note that when this option is used, register names in @samp{-f} flags
5731 must use the normal, user-mode, names.
5733 @item -muser-registers
5734 @opindex muser-registers
5735 Use the normal set of global registers, @code{gr96-gr127}. This is the
5739 @itemx -mno-stack-check
5740 @opindex mstack-check
5741 @opindex mno-stack-check
5742 @cindex stack checks (29k)
5743 Insert (or do not insert) a call to @code{__msp_check} after each stack
5744 adjustment. This is often used for kernel code.
5747 @itemx -mno-storem-bug
5748 @opindex mstorem-bug
5749 @opindex mno-storem-bug
5750 @cindex storem bug (29k)
5751 @option{-mstorem-bug} handles 29k processors which cannot handle the
5752 separation of a mtsrim insn and a storem instruction (most 29000 chips
5753 to date, but not the 29050).
5755 @item -mno-reuse-arg-regs
5756 @itemx -mreuse-arg-regs
5757 @opindex mno-reuse-arg-regs
5758 @opindex mreuse-arg-regs
5759 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5760 registers for copying out arguments. This helps detect calling a function
5761 with fewer arguments than it was declared with.
5763 @item -mno-impure-text
5764 @itemx -mimpure-text
5765 @opindex mno-impure-text
5766 @opindex mimpure-text
5767 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5768 not pass @option{-assert pure-text} to the linker when linking a shared object.
5771 @opindex msoft-float
5772 Generate output containing library calls for floating point.
5773 @strong{Warning:} the requisite libraries are not part of GCC@.
5774 Normally the facilities of the machine's usual C compiler are used, but
5775 this can't be done directly in cross-compilation. You must make your
5776 own arrangements to provide suitable library functions for
5781 Do not generate multm or multmu instructions. This is useful for some embedded
5782 systems which do not have trap handlers for these instructions.
5786 @subsection ARM Options
5789 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5794 @opindex mapcs-frame
5795 Generate a stack frame that is compliant with the ARM Procedure Call
5796 Standard for all functions, even if this is not strictly necessary for
5797 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5798 with this option will cause the stack frames not to be generated for
5799 leaf functions. The default is @option{-mno-apcs-frame}.
5803 This is a synonym for @option{-mapcs-frame}.
5807 Generate code for a processor running with a 26-bit program counter,
5808 and conforming to the function calling standards for the APCS 26-bit
5809 option. This option replaces the @option{-m2} and @option{-m3} options
5810 of previous releases of the compiler.
5814 Generate code for a processor running with a 32-bit program counter,
5815 and conforming to the function calling standards for the APCS 32-bit
5816 option. This option replaces the @option{-m6} option of previous releases
5820 @c not currently implemented
5821 @item -mapcs-stack-check
5822 @opindex mapcs-stack-check
5823 Generate code to check the amount of stack space available upon entry to
5824 every function (that actually uses some stack space). If there is
5825 insufficient space available then either the function
5826 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5827 called, depending upon the amount of stack space required. The run time
5828 system is required to provide these functions. The default is
5829 @option{-mno-apcs-stack-check}, since this produces smaller code.
5831 @c not currently implemented
5833 @opindex mapcs-float
5834 Pass floating point arguments using the float point registers. This is
5835 one of the variants of the APCS@. This option is recommended if the
5836 target hardware has a floating point unit or if a lot of floating point
5837 arithmetic is going to be performed by the code. The default is
5838 @option{-mno-apcs-float}, since integer only code is slightly increased in
5839 size if @option{-mapcs-float} is used.
5841 @c not currently implemented
5842 @item -mapcs-reentrant
5843 @opindex mapcs-reentrant
5844 Generate reentrant, position independent code. The default is
5845 @option{-mno-apcs-reentrant}.
5848 @item -mthumb-interwork
5849 @opindex mthumb-interwork
5850 Generate code which supports calling between the ARM and Thumb
5851 instruction sets. Without this option the two instruction sets cannot
5852 be reliably used inside one program. The default is
5853 @option{-mno-thumb-interwork}, since slightly larger code is generated
5854 when @option{-mthumb-interwork} is specified.
5856 @item -mno-sched-prolog
5857 @opindex mno-sched-prolog
5858 Prevent the reordering of instructions in the function prolog, or the
5859 merging of those instruction with the instructions in the function's
5860 body. This means that all functions will start with a recognizable set
5861 of instructions (or in fact one of a choice from a small set of
5862 different function prologues), and this information can be used to
5863 locate the start if functions inside an executable piece of code. The
5864 default is @option{-msched-prolog}.
5867 @opindex mhard-float
5868 Generate output containing floating point instructions. This is the
5872 @opindex msoft-float
5873 Generate output containing library calls for floating point.
5874 @strong{Warning:} the requisite libraries are not available for all ARM
5875 targets. Normally the facilities of the machine's usual C compiler are
5876 used, but this cannot be done directly in cross-compilation. You must make
5877 your own arrangements to provide suitable library functions for
5880 @option{-msoft-float} changes the calling convention in the output file;
5881 therefore, it is only useful if you compile @emph{all} of a program with
5882 this option. In particular, you need to compile @file{libgcc.a}, the
5883 library that comes with GCC, with @option{-msoft-float} in order for
5886 @item -mlittle-endian
5887 @opindex mlittle-endian
5888 Generate code for a processor running in little-endian mode. This is
5889 the default for all standard configurations.
5892 @opindex mbig-endian
5893 Generate code for a processor running in big-endian mode; the default is
5894 to compile code for a little-endian processor.
5896 @item -mwords-little-endian
5897 @opindex mwords-little-endian
5898 This option only applies when generating code for big-endian processors.
5899 Generate code for a little-endian word order but a big-endian byte
5900 order. That is, a byte order of the form @samp{32107654}. Note: this
5901 option should only be used if you require compatibility with code for
5902 big-endian ARM processors generated by versions of the compiler prior to
5905 @item -malignment-traps
5906 @opindex malignment-traps
5907 Generate code that will not trap if the MMU has alignment traps enabled.
5908 On ARM architectures prior to ARMv4, there were no instructions to
5909 access half-word objects stored in memory. However, when reading from
5910 memory a feature of the ARM architecture allows a word load to be used,
5911 even if the address is unaligned, and the processor core will rotate the
5912 data as it is being loaded. This option tells the compiler that such
5913 misaligned accesses will cause a MMU trap and that it should instead
5914 synthesise the access as a series of byte accesses. The compiler can
5915 still use word accesses to load half-word data if it knows that the
5916 address is aligned to a word boundary.
5918 This option is ignored when compiling for ARM architecture 4 or later,
5919 since these processors have instructions to directly access half-word
5922 @item -mno-alignment-traps
5923 @opindex mno-alignment-traps
5924 Generate code that assumes that the MMU will not trap unaligned
5925 accesses. This produces better code when the target instruction set
5926 does not have half-word memory operations (i.e.@: implementations prior to
5929 Note that you cannot use this option to access unaligned word objects,
5930 since the processor will only fetch one 32-bit aligned object from
5933 The default setting for most targets is @option{-mno-alignment-traps}, since
5934 this produces better code when there are no half-word memory
5935 instructions available.
5937 @item -mshort-load-bytes
5938 @itemx -mno-short-load-words
5939 @opindex mshort-load-bytes
5940 @opindex mno-short-load-words
5941 These are deprecated aliases for @option{-malignment-traps}.
5943 @item -mno-short-load-bytes
5944 @itemx -mshort-load-words
5945 @opindex mno-short-load-bytes
5946 @opindex mshort-load-words
5947 This are deprecated aliases for @option{-mno-alignment-traps}.
5951 This option only applies to RISC iX@. Emulate the native BSD-mode
5952 compiler. This is the default if @option{-ansi} is not specified.
5956 This option only applies to RISC iX@. Emulate the native X/Open-mode
5959 @item -mno-symrename
5960 @opindex mno-symrename
5961 This option only applies to RISC iX@. Do not run the assembler
5962 post-processor, @samp{symrename}, after code has been assembled.
5963 Normally it is necessary to modify some of the standard symbols in
5964 preparation for linking with the RISC iX C library; this option
5965 suppresses this pass. The post-processor is never run when the
5966 compiler is built for cross-compilation.
5968 @item -mcpu=@var{name}
5970 This specifies the name of the target ARM processor. GCC uses this name
5971 to determine what kind of instructions it can emit when generating
5972 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
5973 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
5974 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
5975 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
5976 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
5977 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
5978 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
5979 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
5980 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
5981 @samp{arm1020t}, @samp{xscale}.
5983 @itemx -mtune=@var{name}
5985 This option is very similar to the @option{-mcpu=} option, except that
5986 instead of specifying the actual target processor type, and hence
5987 restricting which instructions can be used, it specifies that GCC should
5988 tune the performance of the code as if the target were of the type
5989 specified in this option, but still choosing the instructions that it
5990 will generate based on the cpu specified by a @option{-mcpu=} option.
5991 For some ARM implementations better performance can be obtained by using
5994 @item -march=@var{name}
5996 This specifies the name of the target ARM architecture. GCC uses this
5997 name to determine what kind of instructions it can emit when generating
5998 assembly code. This option can be used in conjunction with or instead
5999 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6000 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6001 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6003 @item -mfpe=@var{number}
6004 @itemx -mfp=@var{number}
6007 This specifies the version of the floating point emulation available on
6008 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6009 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6011 @item -mstructure-size-boundary=@var{n}
6012 @opindex mstructure-size-boundary
6013 The size of all structures and unions will be rounded up to a multiple
6014 of the number of bits set by this option. Permissible values are 8 and
6015 32. The default value varies for different toolchains. For the COFF
6016 targeted toolchain the default value is 8. Specifying the larger number
6017 can produce faster, more efficient code, but can also increase the size
6018 of the program. The two values are potentially incompatible. Code
6019 compiled with one value cannot necessarily expect to work with code or
6020 libraries compiled with the other value, if they exchange information
6021 using structures or unions.
6023 @item -mabort-on-noreturn
6024 @opindex mabort-on-noreturn
6025 Generate a call to the function @code{abort} at the end of a
6026 @code{noreturn} function. It will be executed if the function tries to
6030 @itemx -mno-long-calls
6031 @opindex mlong-calls
6032 @opindex mno-long-calls
6033 Tells the compiler to perform function calls by first loading the
6034 address of the function into a register and then performing a subroutine
6035 call on this register. This switch is needed if the target function
6036 will lie outside of the 64 megabyte addressing range of the offset based
6037 version of subroutine call instruction.
6039 Even if this switch is enabled, not all function calls will be turned
6040 into long calls. The heuristic is that static functions, functions
6041 which have the @samp{short-call} attribute, functions that are inside
6042 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6043 definitions have already been compiled within the current compilation
6044 unit, will not be turned into long calls. The exception to this rule is
6045 that weak function definitions, functions with the @samp{long-call}
6046 attribute or the @samp{section} attribute, and functions that are within
6047 the scope of a @samp{#pragma long_calls} directive, will always be
6048 turned into long calls.
6050 This feature is not enabled by default. Specifying
6051 @option{-mno-long-calls} will restore the default behavior, as will
6052 placing the function calls within the scope of a @samp{#pragma
6053 long_calls_off} directive. Note these switches have no effect on how
6054 the compiler generates code to handle function calls via function
6057 @item -mnop-fun-dllimport
6058 @opindex mnop-fun-dllimport
6059 Disable support for the @code{dllimport} attribute.
6061 @item -msingle-pic-base
6062 @opindex msingle-pic-base
6063 Treat the register used for PIC addressing as read-only, rather than
6064 loading it in the prologue for each function. The run-time system is
6065 responsible for initializing this register with an appropriate value
6066 before execution begins.
6068 @item -mpic-register=@var{reg}
6069 @opindex mpic-register
6070 Specify the register to be used for PIC addressing. The default is R10
6071 unless stack-checking is enabled, when R9 is used.
6073 @item -mpoke-function-name
6074 @opindex mpoke-function-name
6075 Write the name of each function into the text section, directly
6076 preceding the function prologue. The generated code is similar to this:
6080 .ascii "arm_poke_function_name", 0
6083 .word 0xff000000 + (t1 - t0)
6084 arm_poke_function_name
6086 stmfd sp!, @{fp, ip, lr, pc@}
6090 When performing a stack backtrace, code can inspect the value of
6091 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6092 location @code{pc - 12} and the top 8 bits are set, then we know that
6093 there is a function name embedded immediately preceding this location
6094 and has length @code{((pc[-3]) & 0xff000000)}.
6098 Generate code for the 16-bit Thumb instruction set. The default is to
6099 use the 32-bit ARM instruction set.
6102 @opindex mtpcs-frame
6103 Generate a stack frame that is compliant with the Thumb Procedure Call
6104 Standard for all non-leaf functions. (A leaf function is one that does
6105 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6107 @item -mtpcs-leaf-frame
6108 @opindex mtpcs-leaf-frame
6109 Generate a stack frame that is compliant with the Thumb Procedure Call
6110 Standard for all leaf functions. (A leaf function is one that does
6111 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6113 @item -mcallee-super-interworking
6114 @opindex mcallee-super-interworking
6115 Gives all externally visible functions in the file being compiled an ARM
6116 instruction set header which switches to Thumb mode before executing the
6117 rest of the function. This allows these functions to be called from
6118 non-interworking code.
6120 @item -mcaller-super-interworking
6121 @opindex mcaller-super-interworking
6122 Allows calls via function pointers (including virtual functions) to
6123 execute correctly regardless of whether the target code has been
6124 compiled for interworking or not. There is a small overhead in the cost
6125 of executing a function pointer if this option is enabled.
6129 @node MN10200 Options
6130 @subsection MN10200 Options
6131 @cindex MN10200 options
6132 These @option{-m} options are defined for Matsushita MN10200 architectures:
6137 Indicate to the linker that it should perform a relaxation optimization pass
6138 to shorten branches, calls and absolute memory addresses. This option only
6139 has an effect when used on the command line for the final link step.
6141 This option makes symbolic debugging impossible.
6144 @node MN10300 Options
6145 @subsection MN10300 Options
6146 @cindex MN10300 options
6147 These @option{-m} options are defined for Matsushita MN10300 architectures:
6152 Generate code to avoid bugs in the multiply instructions for the MN10300
6153 processors. This is the default.
6156 @opindex mno-mult-bug
6157 Do not generate code to avoid bugs in the multiply instructions for the
6162 Generate code which uses features specific to the AM33 processor.
6166 Do not generate code which uses features specific to the AM33 processor. This
6171 Do not link in the C run-time initialization object file.
6175 Indicate to the linker that it should perform a relaxation optimization pass
6176 to shorten branches, calls and absolute memory addresses. This option only
6177 has an effect when used on the command line for the final link step.
6179 This option makes symbolic debugging impossible.
6183 @node M32R/D Options
6184 @subsection M32R/D Options
6185 @cindex M32R/D options
6187 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6192 Generate code for the M32R/X@.
6196 Generate code for the M32R@. This is the default.
6198 @item -mcode-model=small
6199 @opindex mcode-model=small
6200 Assume all objects live in the lower 16MB of memory (so that their addresses
6201 can be loaded with the @code{ld24} instruction), and assume all subroutines
6202 are reachable with the @code{bl} instruction.
6203 This is the default.
6205 The addressability of a particular object can be set with the
6206 @code{model} attribute.
6208 @item -mcode-model=medium
6209 @opindex mcode-model=medium
6210 Assume objects may be anywhere in the 32-bit address space (the compiler
6211 will generate @code{seth/add3} instructions to load their addresses), and
6212 assume all subroutines are reachable with the @code{bl} instruction.
6214 @item -mcode-model=large
6215 @opindex mcode-model=large
6216 Assume objects may be anywhere in the 32-bit address space (the compiler
6217 will generate @code{seth/add3} instructions to load their addresses), and
6218 assume subroutines may not be reachable with the @code{bl} instruction
6219 (the compiler will generate the much slower @code{seth/add3/jl}
6220 instruction sequence).
6223 @opindex msdata=none
6224 Disable use of the small data area. Variables will be put into
6225 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6226 @code{section} attribute has been specified).
6227 This is the default.
6229 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6230 Objects may be explicitly put in the small data area with the
6231 @code{section} attribute using one of these sections.
6234 @opindex msdata=sdata
6235 Put small global and static data in the small data area, but do not
6236 generate special code to reference them.
6240 Put small global and static data in the small data area, and generate
6241 special instructions to reference them.
6245 @cindex smaller data references
6246 Put global and static objects less than or equal to @var{num} bytes
6247 into the small data or bss sections instead of the normal data or bss
6248 sections. The default value of @var{num} is 8.
6249 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6250 for this option to have any effect.
6252 All modules should be compiled with the same @option{-G @var{num}} value.
6253 Compiling with different values of @var{num} may or may not work; if it
6254 doesn't the linker will give an error message---incorrect code will not be
6260 @subsection M88K Options
6261 @cindex M88k options
6263 These @samp{-m} options are defined for Motorola 88k architectures:
6268 Generate code that works well on both the m88100 and the
6273 Generate code that works best for the m88100, but that also
6278 Generate code that works best for the m88110, and may not run
6283 Obsolete option to be removed from the next revision.
6286 @item -midentify-revision
6287 @opindex midentify-revision
6288 @cindex identifying source, compiler (88k)
6289 Include an @code{ident} directive in the assembler output recording the
6290 source file name, compiler name and version, timestamp, and compilation
6293 @item -mno-underscores
6294 @opindex mno-underscores
6295 @cindex underscores, avoiding (88k)
6296 In assembler output, emit symbol names without adding an underscore
6297 character at the beginning of each name. The default is to use an
6298 underscore as prefix on each name.
6300 @item -mocs-debug-info
6301 @itemx -mno-ocs-debug-info
6302 @opindex mocs-debug-info
6303 @opindex mno-ocs-debug-info
6305 @cindex debugging, 88k OCS
6306 Include (or omit) additional debugging information (about registers used
6307 in each stack frame) as specified in the 88open Object Compatibility
6308 Standard, ``OCS''@. This extra information allows debugging of code that
6309 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6310 Delta 88 SVr3.2 is to include this information; other 88k configurations
6311 omit this information by default.
6313 @item -mocs-frame-position
6314 @opindex mocs-frame-position
6315 @cindex register positions in frame (88k)
6316 When emitting COFF debugging information for automatic variables and
6317 parameters stored on the stack, use the offset from the canonical frame
6318 address, which is the stack pointer (register 31) on entry to the
6319 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6320 @option{-mocs-frame-position}; other 88k configurations have the default
6321 @option{-mno-ocs-frame-position}.
6323 @item -mno-ocs-frame-position
6324 @opindex mno-ocs-frame-position
6325 @cindex register positions in frame (88k)
6326 When emitting COFF debugging information for automatic variables and
6327 parameters stored on the stack, use the offset from the frame pointer
6328 register (register 30). When this option is in effect, the frame
6329 pointer is not eliminated when debugging information is selected by the
6332 @item -moptimize-arg-area
6333 @opindex moptimize-arg-area
6334 @cindex arguments in frame (88k)
6335 Save space by reorganizing the stack frame. This option generates code
6336 that does not agree with the 88open specifications, but uses less
6339 @itemx -mno-optimize-arg-area
6340 @opindex mno-optimize-arg-area
6341 Do not reorganize the stack frame to save space. This is the default.
6342 The generated conforms to the specification, but uses more memory.
6344 @item -mshort-data-@var{num}
6345 @opindex mshort-data
6346 @cindex smaller data references (88k)
6347 @cindex r0-relative references (88k)
6348 Generate smaller data references by making them relative to @code{r0},
6349 which allows loading a value using a single instruction (rather than the
6350 usual two). You control which data references are affected by
6351 specifying @var{num} with this option. For example, if you specify
6352 @option{-mshort-data-512}, then the data references affected are those
6353 involving displacements of less than 512 bytes.
6354 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6357 @item -mserialize-volatile
6358 @opindex mserialize-volatile
6359 @itemx -mno-serialize-volatile
6360 @opindex mno-serialize-volatile
6361 @cindex sequential consistency on 88k
6362 Do, or don't, generate code to guarantee sequential consistency
6363 of volatile memory references. By default, consistency is
6366 The order of memory references made by the MC88110 processor does
6367 not always match the order of the instructions requesting those
6368 references. In particular, a load instruction may execute before
6369 a preceding store instruction. Such reordering violates
6370 sequential consistency of volatile memory references, when there
6371 are multiple processors. When consistency must be guaranteed,
6372 GCC generates special instructions, as needed, to force
6373 execution in the proper order.
6375 The MC88100 processor does not reorder memory references and so
6376 always provides sequential consistency. However, by default, GCC
6377 generates the special instructions to guarantee consistency
6378 even when you use @option{-m88100}, so that the code may be run on an
6379 MC88110 processor. If you intend to run your code only on the
6380 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6382 The extra code generated to guarantee consistency may affect the
6383 performance of your application. If you know that you can safely
6384 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6390 @cindex assembler syntax, 88k
6392 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6393 related to System V release 4 (SVr4). This controls the following:
6397 Which variant of the assembler syntax to emit.
6399 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6400 that is used on System V release 4.
6402 @option{-msvr4} makes GCC issue additional declaration directives used in
6406 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6407 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6408 other m88k configurations.
6410 @item -mversion-03.00
6411 @opindex mversion-03.00
6412 This option is obsolete, and is ignored.
6413 @c ??? which asm syntax better for GAS? option there too?
6415 @item -mno-check-zero-division
6416 @itemx -mcheck-zero-division
6417 @opindex mno-check-zero-division
6418 @opindex mcheck-zero-division
6419 @cindex zero division on 88k
6420 Do, or don't, generate code to guarantee that integer division by
6421 zero will be detected. By default, detection is guaranteed.
6423 Some models of the MC88100 processor fail to trap upon integer
6424 division by zero under certain conditions. By default, when
6425 compiling code that might be run on such a processor, GCC
6426 generates code that explicitly checks for zero-valued divisors
6427 and traps with exception number 503 when one is detected. Use of
6428 @option{-mno-check-zero-division} suppresses such checking for code
6429 generated to run on an MC88100 processor.
6431 GCC assumes that the MC88110 processor correctly detects all instances
6432 of integer division by zero. When @option{-m88110} is specified, no
6433 explicit checks for zero-valued divisors are generated, and both
6434 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6437 @item -muse-div-instruction
6438 @opindex muse-div-instruction
6439 @cindex divide instruction, 88k
6440 Use the div instruction for signed integer division on the
6441 MC88100 processor. By default, the div instruction is not used.
6443 On the MC88100 processor the signed integer division instruction
6444 div) traps to the operating system on a negative operand. The
6445 operating system transparently completes the operation, but at a
6446 large cost in execution time. By default, when compiling code
6447 that might be run on an MC88100 processor, GCC emulates signed
6448 integer division using the unsigned integer division instruction
6449 divu), thereby avoiding the large penalty of a trap to the
6450 operating system. Such emulation has its own, smaller, execution
6451 cost in both time and space. To the extent that your code's
6452 important signed integer division operations are performed on two
6453 nonnegative operands, it may be desirable to use the div
6454 instruction directly.
6456 On the MC88110 processor the div instruction (also known as the
6457 divs instruction) processes negative operands without trapping to
6458 the operating system. When @option{-m88110} is specified,
6459 @option{-muse-div-instruction} is ignored, and the div instruction is used
6460 for signed integer division.
6462 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6463 particular, the behavior of such a division with and without
6464 @option{-muse-div-instruction} may differ.
6466 @item -mtrap-large-shift
6467 @itemx -mhandle-large-shift
6468 @opindex mtrap-large-shift
6469 @opindex mhandle-large-shift
6470 @cindex bit shift overflow (88k)
6471 @cindex large bit shifts (88k)
6472 Include code to detect bit-shifts of more than 31 bits; respectively,
6473 trap such shifts or emit code to handle them properly. By default GCC
6474 makes no special provision for large bit shifts.
6476 @item -mwarn-passed-structs
6477 @opindex mwarn-passed-structs
6478 @cindex structure passing (88k)
6479 Warn when a function passes a struct as an argument or result.
6480 Structure-passing conventions have changed during the evolution of the C
6481 language, and are often the source of portability problems. By default,
6482 GCC issues no such warning.
6485 @c break page here to avoid unsightly interparagraph stretch.
6489 @node RS/6000 and PowerPC Options
6490 @subsection IBM RS/6000 and PowerPC Options
6491 @cindex RS/6000 and PowerPC Options
6492 @cindex IBM RS/6000 and PowerPC Options
6494 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6502 @itemx -mpowerpc-gpopt
6503 @itemx -mno-powerpc-gpopt
6504 @itemx -mpowerpc-gfxopt
6505 @itemx -mno-powerpc-gfxopt
6507 @itemx -mno-powerpc64
6513 @opindex mno-powerpc
6514 @opindex mpowerpc-gpopt
6515 @opindex mno-powerpc-gpopt
6516 @opindex mpowerpc-gfxopt
6517 @opindex mno-powerpc-gfxopt
6519 @opindex mno-powerpc64
6520 GCC supports two related instruction set architectures for the
6521 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6522 instructions supported by the @samp{rios} chip set used in the original
6523 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6524 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6525 the IBM 4xx microprocessors.
6527 Neither architecture is a subset of the other. However there is a
6528 large common subset of instructions supported by both. An MQ
6529 register is included in processors supporting the POWER architecture.
6531 You use these options to specify which instructions are available on the
6532 processor you are using. The default value of these options is
6533 determined when configuring GCC@. Specifying the
6534 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6535 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6536 rather than the options listed above.
6538 The @option{-mpower} option allows GCC to generate instructions that
6539 are found only in the POWER architecture and to use the MQ register.
6540 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6541 to generate instructions that are present in the POWER2 architecture but
6542 not the original POWER architecture.
6544 The @option{-mpowerpc} option allows GCC to generate instructions that
6545 are found only in the 32-bit subset of the PowerPC architecture.
6546 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6547 GCC to use the optional PowerPC architecture instructions in the
6548 General Purpose group, including floating-point square root. Specifying
6549 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6550 use the optional PowerPC architecture instructions in the Graphics
6551 group, including floating-point select.
6553 The @option{-mpowerpc64} option allows GCC to generate the additional
6554 64-bit instructions that are found in the full PowerPC64 architecture
6555 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6556 @option{-mno-powerpc64}.
6558 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6559 will use only the instructions in the common subset of both
6560 architectures plus some special AIX common-mode calls, and will not use
6561 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6562 permits GCC to use any instruction from either architecture and to
6563 allow use of the MQ register; specify this for the Motorola MPC601.
6565 @item -mnew-mnemonics
6566 @itemx -mold-mnemonics
6567 @opindex mnew-mnemonics
6568 @opindex mold-mnemonics
6569 Select which mnemonics to use in the generated assembler code. With
6570 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6571 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6572 assembler mnemonics defined for the POWER architecture. Instructions
6573 defined in only one architecture have only one mnemonic; GCC uses that
6574 mnemonic irrespective of which of these options is specified.
6576 GCC defaults to the mnemonics appropriate for the architecture in
6577 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6578 value of these option. Unless you are building a cross-compiler, you
6579 should normally not specify either @option{-mnew-mnemonics} or
6580 @option{-mold-mnemonics}, but should instead accept the default.
6582 @item -mcpu=@var{cpu_type}
6584 Set architecture type, register usage, choice of mnemonics, and
6585 instruction scheduling parameters for machine type @var{cpu_type}.
6586 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6587 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6588 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6589 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6590 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6591 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6593 @option{-mcpu=common} selects a completely generic processor. Code
6594 generated under this option will run on any POWER or PowerPC processor.
6595 GCC will use only the instructions in the common subset of both
6596 architectures, and will not use the MQ register. GCC assumes a generic
6597 processor model for scheduling purposes.
6599 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6600 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6601 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6602 types, with an appropriate, generic processor model assumed for
6603 scheduling purposes.
6605 The other options specify a specific processor. Code generated under
6606 those options will run best on that processor, and may not run at all on
6609 The @option{-mcpu} options automatically enable or disable other
6610 @option{-m} options as follows:
6614 @option{-mno-power}, @option{-mno-powerc}
6621 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6636 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6639 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6644 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6647 @item -mtune=@var{cpu_type}
6649 Set the instruction scheduling parameters for machine type
6650 @var{cpu_type}, but do not set the architecture type, register usage, or
6651 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6652 values for @var{cpu_type} are used for @option{-mtune} as for
6653 @option{-mcpu}. If both are specified, the code generated will use the
6654 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6655 scheduling parameters set by @option{-mtune}.
6660 @opindex mno-altivec
6661 These switches enable or disable the use of built-in functions that
6662 allow access to the AltiVec instruction set. You may also need to set
6663 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6667 @itemx -mno-fp-in-toc
6668 @itemx -mno-sum-in-toc
6669 @itemx -mminimal-toc
6671 @opindex mno-fp-in-toc
6672 @opindex mno-sum-in-toc
6673 @opindex mminimal-toc
6674 Modify generation of the TOC (Table Of Contents), which is created for
6675 every executable file. The @option{-mfull-toc} option is selected by
6676 default. In that case, GCC will allocate at least one TOC entry for
6677 each unique non-automatic variable reference in your program. GCC
6678 will also place floating-point constants in the TOC@. However, only
6679 16,384 entries are available in the TOC@.
6681 If you receive a linker error message that saying you have overflowed
6682 the available TOC space, you can reduce the amount of TOC space used
6683 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6684 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6685 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6686 generate code to calculate the sum of an address and a constant at
6687 run-time instead of putting that sum into the TOC@. You may specify one
6688 or both of these options. Each causes GCC to produce very slightly
6689 slower and larger code at the expense of conserving TOC space.
6691 If you still run out of space in the TOC even when you specify both of
6692 these options, specify @option{-mminimal-toc} instead. This option causes
6693 GCC to make only one TOC entry for every file. When you specify this
6694 option, GCC will produce code that is slower and larger but which
6695 uses extremely little TOC space. You may wish to use this option
6696 only on files that contain less frequently executed code.
6702 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6703 @code{long} type, and the infrastructure needed to support them.
6704 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6705 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6706 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6711 @opindex mno-xl-call
6712 On AIX, pass floating-point arguments to prototyped functions beyond the
6713 register save area (RSA) on the stack in addition to argument FPRs. The
6714 AIX calling convention was extended but not initially documented to
6715 handle an obscure K&R C case of calling a function that takes the
6716 address of its arguments with fewer arguments than declared. AIX XL
6717 compilers access floating point arguments which do not fit in the
6718 RSA from the stack when a subroutine is compiled without
6719 optimization. Because always storing floating-point arguments on the
6720 stack is inefficient and rarely needed, this option is not enabled by
6721 default and only is necessary when calling subroutines compiled by AIX
6722 XL compilers without optimization.
6726 Support @dfn{AIX Threads}. Link an application written to use
6727 @dfn{pthreads} with special libraries and startup code to enable the
6732 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6733 application written to use message passing with special startup code to
6734 enable the application to run. The system must have PE installed in the
6735 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6736 must be overridden with the @option{-specs=} option to specify the
6737 appropriate directory location. The Parallel Environment does not
6738 support threads, so the @option{-mpe} option and the @option{-mthreads}
6739 option are incompatible.
6743 @opindex msoft-float
6744 @opindex mhard-float
6745 Generate code that does not use (uses) the floating-point register set.
6746 Software floating point emulation is provided if you use the
6747 @option{-msoft-float} option, and pass the option to GCC when linking.
6750 @itemx -mno-multiple
6752 @opindex mno-multiple
6753 Generate code that uses (does not use) the load multiple word
6754 instructions and the store multiple word instructions. These
6755 instructions are generated by default on POWER systems, and not
6756 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6757 endian PowerPC systems, since those instructions do not work when the
6758 processor is in little endian mode. The exceptions are PPC740 and
6759 PPC750 which permit the instructions usage in little endian mode.
6765 Generate code that uses (does not use) the load string instructions
6766 and the store string word instructions to save multiple registers and
6767 do small block moves. These instructions are generated by default on
6768 POWER systems, and not generated on PowerPC systems. Do not use
6769 @option{-mstring} on little endian PowerPC systems, since those
6770 instructions do not work when the processor is in little endian mode.
6771 The exceptions are PPC740 and PPC750 which permit the instructions
6772 usage in little endian mode.
6778 Generate code that uses (does not use) the load or store instructions
6779 that update the base register to the address of the calculated memory
6780 location. These instructions are generated by default. If you use
6781 @option{-mno-update}, there is a small window between the time that the
6782 stack pointer is updated and the address of the previous frame is
6783 stored, which means code that walks the stack frame across interrupts or
6784 signals may get corrupted data.
6787 @itemx -mno-fused-madd
6788 @opindex mfused-madd
6789 @opindex mno-fused-madd
6790 Generate code that uses (does not use) the floating point multiply and
6791 accumulate instructions. These instructions are generated by default if
6792 hardware floating is used.
6794 @item -mno-bit-align
6796 @opindex mno-bit-align
6798 On System V.4 and embedded PowerPC systems do not (do) force structures
6799 and unions that contain bit-fields to be aligned to the base type of the
6802 For example, by default a structure containing nothing but 8
6803 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6804 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6805 the structure would be aligned to a 1 byte boundary and be one byte in
6808 @item -mno-strict-align
6809 @itemx -mstrict-align
6810 @opindex mno-strict-align
6811 @opindex mstrict-align
6812 On System V.4 and embedded PowerPC systems do not (do) assume that
6813 unaligned memory references will be handled by the system.
6816 @itemx -mno-relocatable
6817 @opindex mrelocatable
6818 @opindex mno-relocatable
6819 On embedded PowerPC systems generate code that allows (does not allow)
6820 the program to be relocated to a different address at runtime. If you
6821 use @option{-mrelocatable} on any module, all objects linked together must
6822 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6824 @item -mrelocatable-lib
6825 @itemx -mno-relocatable-lib
6826 @opindex mrelocatable-lib
6827 @opindex mno-relocatable-lib
6828 On embedded PowerPC systems generate code that allows (does not allow)
6829 the program to be relocated to a different address at runtime. Modules
6830 compiled with @option{-mrelocatable-lib} can be linked with either modules
6831 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6832 with modules compiled with the @option{-mrelocatable} options.
6838 On System V.4 and embedded PowerPC systems do not (do) assume that
6839 register 2 contains a pointer to a global area pointing to the addresses
6840 used in the program.
6843 @itemx -mlittle-endian
6845 @opindex mlittle-endian
6846 On System V.4 and embedded PowerPC systems compile code for the
6847 processor in little endian mode. The @option{-mlittle-endian} option is
6848 the same as @option{-mlittle}.
6853 @opindex mbig-endian
6854 On System V.4 and embedded PowerPC systems compile code for the
6855 processor in big endian mode. The @option{-mbig-endian} option is
6856 the same as @option{-mbig}.
6860 On System V.4 and embedded PowerPC systems compile code using calling
6861 conventions that adheres to the March 1995 draft of the System V
6862 Application Binary Interface, PowerPC processor supplement. This is the
6863 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6865 @item -mcall-sysv-eabi
6866 @opindex mcall-sysv-eabi
6867 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6869 @item -mcall-sysv-noeabi
6870 @opindex mcall-sysv-noeabi
6871 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6875 On System V.4 and embedded PowerPC systems compile code using calling
6876 conventions that are similar to those used on AIX@. This is the
6877 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6879 @item -mcall-solaris
6880 @opindex mcall-solaris
6881 On System V.4 and embedded PowerPC systems compile code for the Solaris
6885 @opindex mcall-linux
6886 On System V.4 and embedded PowerPC systems compile code for the
6887 Linux-based GNU system.
6890 @opindex mcall-netbsd
6891 On System V.4 and embedded PowerPC systems compile code for the
6892 NetBSD operating system.
6894 @item -maix-struct-return
6895 @opindex maix-struct-return
6896 Return all structures in memory (as specified by the AIX ABI)@.
6898 @item -msvr4-struct-return
6899 @opindex msvr4-struct-return
6900 Return structures smaller than 8 bytes in registers (as specified by the
6904 @opindex mabi=altivec
6905 Extend the current ABI with AltiVec ABI extensions. This does not
6906 change the default ABI, instead it adds the AltiVec ABI extensions to
6910 @itemx -mno-prototype
6912 @opindex mno-prototype
6913 On System V.4 and embedded PowerPC systems assume that all calls to
6914 variable argument functions are properly prototyped. Otherwise, the
6915 compiler must insert an instruction before every non prototyped call to
6916 set or clear bit 6 of the condition code register (@var{CR}) to
6917 indicate whether floating point values were passed in the floating point
6918 registers in case the function takes a variable arguments. With
6919 @option{-mprototype}, only calls to prototyped variable argument functions
6920 will set or clear the bit.
6924 On embedded PowerPC systems, assume that the startup module is called
6925 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6926 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6931 On embedded PowerPC systems, assume that the startup module is called
6932 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6937 On embedded PowerPC systems, assume that the startup module is called
6938 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6942 @opindex myellowknife
6943 On embedded PowerPC systems, assume that the startup module is called
6944 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6949 On System V.4 and embedded PowerPC systems, specify that you are
6950 compiling for a VxWorks system.
6954 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6955 header to indicate that @samp{eabi} extended relocations are used.
6961 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6962 Embedded Applications Binary Interface (eabi) which is a set of
6963 modifications to the System V.4 specifications. Selecting @option{-meabi}
6964 means that the stack is aligned to an 8 byte boundary, a function
6965 @code{__eabi} is called to from @code{main} to set up the eabi
6966 environment, and the @option{-msdata} option can use both @code{r2} and
6967 @code{r13} to point to two separate small data areas. Selecting
6968 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6969 do not call an initialization function from @code{main}, and the
6970 @option{-msdata} option will only use @code{r13} to point to a single
6971 small data area. The @option{-meabi} option is on by default if you
6972 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
6975 @opindex msdata=eabi
6976 On System V.4 and embedded PowerPC systems, put small initialized
6977 @code{const} global and static data in the @samp{.sdata2} section, which
6978 is pointed to by register @code{r2}. Put small initialized
6979 non-@code{const} global and static data in the @samp{.sdata} section,
6980 which is pointed to by register @code{r13}. Put small uninitialized
6981 global and static data in the @samp{.sbss} section, which is adjacent to
6982 the @samp{.sdata} section. The @option{-msdata=eabi} option is
6983 incompatible with the @option{-mrelocatable} option. The
6984 @option{-msdata=eabi} option also sets the @option{-memb} option.
6987 @opindex msdata=sysv
6988 On System V.4 and embedded PowerPC systems, put small global and static
6989 data in the @samp{.sdata} section, which is pointed to by register
6990 @code{r13}. Put small uninitialized global and static data in the
6991 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
6992 The @option{-msdata=sysv} option is incompatible with the
6993 @option{-mrelocatable} option.
6995 @item -msdata=default
6997 @opindex msdata=default
6999 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7000 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7001 same as @option{-msdata=sysv}.
7004 @opindex msdata-data
7005 On System V.4 and embedded PowerPC systems, put small global and static
7006 data in the @samp{.sdata} section. Put small uninitialized global and
7007 static data in the @samp{.sbss} section. Do not use register @code{r13}
7008 to address small data however. This is the default behavior unless
7009 other @option{-msdata} options are used.
7013 @opindex msdata=none
7015 On embedded PowerPC systems, put all initialized global and static data
7016 in the @samp{.data} section, and all uninitialized data in the
7017 @samp{.bss} section.
7021 @cindex smaller data references (PowerPC)
7022 @cindex .sdata/.sdata2 references (PowerPC)
7023 On embedded PowerPC systems, put global and static items less than or
7024 equal to @var{num} bytes into the small data or bss sections instead of
7025 the normal data or bss section. By default, @var{num} is 8. The
7026 @option{-G @var{num}} switch is also passed to the linker.
7027 All modules should be compiled with the same @option{-G @var{num}} value.
7030 @itemx -mno-regnames
7032 @opindex mno-regnames
7033 On System V.4 and embedded PowerPC systems do (do not) emit register
7034 names in the assembly language output using symbolic forms.
7039 @subsection IBM RT Options
7041 @cindex IBM RT options
7043 These @samp{-m} options are defined for the IBM RT PC:
7047 @opindex min-line-mul
7048 Use an in-line code sequence for integer multiplies. This is the
7051 @item -mcall-lib-mul
7052 @opindex mcall-lib-mul
7053 Call @code{lmul$$} for integer multiples.
7055 @item -mfull-fp-blocks
7056 @opindex mfull-fp-blocks
7057 Generate full-size floating point data blocks, including the minimum
7058 amount of scratch space recommended by IBM@. This is the default.
7060 @item -mminimum-fp-blocks
7061 @opindex mminimum-fp-blocks
7062 Do not include extra scratch space in floating point data blocks. This
7063 results in smaller code, but slower execution, since scratch space must
7064 be allocated dynamically.
7066 @cindex @file{varargs.h} and RT PC
7067 @cindex @file{stdarg.h} and RT PC
7068 @item -mfp-arg-in-fpregs
7069 @opindex mfp-arg-in-fpregs
7070 Use a calling sequence incompatible with the IBM calling convention in
7071 which floating point arguments are passed in floating point registers.
7072 Note that @code{varargs.h} and @code{stdarg.h} will not work with
7073 floating point operands if this option is specified.
7075 @item -mfp-arg-in-gregs
7076 @opindex mfp-arg-in-gregs
7077 Use the normal calling convention for floating point arguments. This is
7080 @item -mhc-struct-return
7081 @opindex mhc-struct-return
7082 Return structures of more than one word in memory, rather than in a
7083 register. This provides compatibility with the MetaWare HighC (hc)
7084 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7085 with the Portable C Compiler (pcc).
7087 @item -mnohc-struct-return
7088 @opindex mnohc-struct-return
7089 Return some structures of more than one word in registers, when
7090 convenient. This is the default. For compatibility with the
7091 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7092 option @option{-mhc-struct-return}.
7096 @subsection MIPS Options
7097 @cindex MIPS options
7099 These @samp{-m} options are defined for the MIPS family of computers:
7103 @item -march=@var{cpu-type}
7105 Assume the defaults for the machine type @var{cpu-type} when generating
7106 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7107 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7108 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7109 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7110 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7111 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
7113 @item -mtune=@var{cpu-type}
7115 Assume the defaults for the machine type @var{cpu-type} when scheduling
7116 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7117 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7118 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7119 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7120 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7121 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
7122 @var{cpu-type} will schedule things appropriately for that particular
7123 chip, the compiler will not generate any code that does not meet level 1
7124 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
7125 or @option{-mabi} switch being used.
7127 @item -mcpu=@var{cpu-type}
7129 This is identical to specifying both @option{-march} and @option{-mtune}.
7133 Issue instructions from level 1 of the MIPS ISA@. This is the default.
7134 @samp{r3000} is the default @var{cpu-type} at this ISA level.
7138 Issue instructions from level 2 of the MIPS ISA (branch likely, square
7139 root instructions). @samp{r6000} is the default @var{cpu-type} at this
7144 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
7145 @samp{r4000} is the default @var{cpu-type} at this ISA level.
7149 Issue instructions from level 4 of the MIPS ISA (conditional move,
7150 prefetch, enhanced FPU instructions). @samp{r8000} is the default
7151 @var{cpu-type} at this ISA level.
7155 Assume that 32 32-bit floating point registers are available. This is
7160 Assume that 32 64-bit floating point registers are available. This is
7161 the default when the @option{-mips3} option is used.
7164 @itemx -mno-fused-madd
7165 @opindex mfused-madd
7166 @opindex mno-fused-madd
7167 Generate code that uses (does not use) the floating point multiply and
7168 accumulate instructions, when they are available. These instructions
7169 are generated by default if they are available, but this may be
7170 undesirable if the extra precision causes problems or on certain chips
7171 in the mode where denormals are rounded to zero where denormals
7172 generated by multiply and accumulate instructions cause exceptions
7177 Assume that 32 32-bit general purpose registers are available. This is
7182 Assume that 32 64-bit general purpose registers are available. This is
7183 the default when the @option{-mips3} option is used.
7187 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7188 explanation of the default, and the width of pointers.
7192 Force long types to be 64 bits wide. See @option{-mlong32} for an
7193 explanation of the default, and the width of pointers.
7197 Force long, int, and pointer types to be 32 bits wide.
7199 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
7200 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
7201 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
7202 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
7203 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
7204 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
7205 are 32 bits, and longs are 64 bits wide. The width of pointer types is
7206 the smaller of the width of longs or the width of general purpose
7207 registers (which in turn depends on the ISA)@.
7219 Generate code for the indicated ABI@. The default instruction level is
7220 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7221 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7222 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7227 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7228 add normal debug information. This is the default for all
7229 platforms except for the OSF/1 reference platform, using the OSF/rose
7230 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7231 switches are used, the @file{mips-tfile} program will encapsulate the
7232 stabs within MIPS ECOFF@.
7236 Generate code for the GNU assembler. This is the default on the OSF/1
7237 reference platform, using the OSF/rose object format. Also, this is
7238 the default if the configure option @option{--with-gnu-as} is used.
7240 @item -msplit-addresses
7241 @itemx -mno-split-addresses
7242 @opindex msplit-addresses
7243 @opindex mno-split-addresses
7244 Generate code to load the high and low parts of address constants separately.
7245 This allows GCC to optimize away redundant loads of the high order
7246 bits of addresses. This optimization requires GNU as and GNU ld.
7247 This optimization is enabled by default for some embedded targets where
7248 GNU as and GNU ld are standard.
7254 The @option{-mrnames} switch says to output code using the MIPS software
7255 names for the registers, instead of the hardware names (ie, @var{a0}
7256 instead of @var{$4}). The only known assembler that supports this option
7257 is the Algorithmics assembler.
7263 The @option{-mgpopt} switch says to write all of the data declarations
7264 before the instructions in the text section, this allows the MIPS
7265 assembler to generate one word memory references instead of using two
7266 words for short global or static data items. This is on by default if
7267 optimization is selected.
7273 For each non-inline function processed, the @option{-mstats} switch
7274 causes the compiler to emit one line to the standard error file to
7275 print statistics about the program (number of registers saved, stack
7282 The @option{-mmemcpy} switch makes all block moves call the appropriate
7283 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7284 generating inline code.
7287 @itemx -mno-mips-tfile
7288 @opindex mmips-tfile
7289 @opindex mno-mips-tfile
7290 The @option{-mno-mips-tfile} switch causes the compiler not
7291 postprocess the object file with the @file{mips-tfile} program,
7292 after the MIPS assembler has generated it to add debug support. If
7293 @file{mips-tfile} is not run, then no local variables will be
7294 available to the debugger. In addition, @file{stage2} and
7295 @file{stage3} objects will have the temporary file names passed to the
7296 assembler embedded in the object file, which means the objects will
7297 not compare the same. The @option{-mno-mips-tfile} switch should only
7298 be used when there are bugs in the @file{mips-tfile} program that
7299 prevents compilation.
7302 @opindex msoft-float
7303 Generate output containing library calls for floating point.
7304 @strong{Warning:} the requisite libraries are not part of GCC@.
7305 Normally the facilities of the machine's usual C compiler are used, but
7306 this can't be done directly in cross-compilation. You must make your
7307 own arrangements to provide suitable library functions for
7311 @opindex mhard-float
7312 Generate output containing floating point instructions. This is the
7313 default if you use the unmodified sources.
7316 @itemx -mno-abicalls
7318 @opindex mno-abicalls
7319 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7320 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7321 position independent code.
7324 @itemx -mno-long-calls
7325 @opindex mlong-calls
7326 @opindex mno-long-calls
7327 Do all calls with the @samp{JALR} instruction, which requires
7328 loading up a function's address into a register before the call.
7329 You need to use this switch, if you call outside of the current
7330 512 megabyte segment to functions that are not through pointers.
7333 @itemx -mno-half-pic
7335 @opindex mno-half-pic
7336 Put pointers to extern references into the data section and load them
7337 up, rather than put the references in the text section.
7339 @item -membedded-pic
7340 @itemx -mno-embedded-pic
7341 @opindex membedded-pic
7342 @opindex mno-embedded-pic
7343 Generate PIC code suitable for some embedded systems. All calls are
7344 made using PC relative address, and all data is addressed using the $gp
7345 register. No more than 65536 bytes of global data may be used. This
7346 requires GNU as and GNU ld which do most of the work. This currently
7347 only works on targets which use ECOFF; it does not work with ELF@.
7349 @item -membedded-data
7350 @itemx -mno-embedded-data
7351 @opindex membedded-data
7352 @opindex mno-embedded-data
7353 Allocate variables to the read-only data section first if possible, then
7354 next in the small data section if possible, otherwise in data. This gives
7355 slightly slower code than the default, but reduces the amount of RAM required
7356 when executing, and thus may be preferred for some embedded systems.
7358 @item -muninit-const-in-rodata
7359 @itemx -mno-uninit-const-in-rodata
7360 @opindex muninit-const-in-rodata
7361 @opindex mno-uninit-const-in-rodata
7362 When used together with @option{-membedded-data}, it will always store uninitialized
7363 const variables in the read-only data section.
7365 @item -msingle-float
7366 @itemx -mdouble-float
7367 @opindex msingle-float
7368 @opindex mdouble-float
7369 The @option{-msingle-float} switch tells gcc to assume that the floating
7370 point coprocessor only supports single precision operations, as on the
7371 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7372 double precision operations. This is the default.
7378 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7379 as on the @samp{r4650} chip.
7383 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7384 @option{-mcpu=r4650}.
7390 Enable 16-bit instructions.
7394 Use the entry and exit pseudo ops. This option can only be used with
7399 Compile code for the processor in little endian mode.
7400 The requisite libraries are assumed to exist.
7404 Compile code for the processor in big endian mode.
7405 The requisite libraries are assumed to exist.
7409 @cindex smaller data references (MIPS)
7410 @cindex gp-relative references (MIPS)
7411 Put global and static items less than or equal to @var{num} bytes into
7412 the small data or bss sections instead of the normal data or bss
7413 section. This allows the assembler to emit one word memory reference
7414 instructions based on the global pointer (@var{gp} or @var{$28}),
7415 instead of the normal two words used. By default, @var{num} is 8 when
7416 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7417 @option{-G @var{num}} switch is also passed to the assembler and linker.
7418 All modules should be compiled with the same @option{-G @var{num}}
7423 Tell the MIPS assembler to not run its preprocessor over user
7424 assembler files (with a @samp{.s} suffix) when assembling them.
7428 Pass an option to gas which will cause nops to be inserted if
7429 the read of the destination register of an mfhi or mflo instruction
7430 occurs in the following two instructions.
7434 Do not include the default crt0.
7436 @item -mflush-func=@var{func}
7437 @itemx -mno-flush-func
7438 @opindex mflush-func
7439 Specifies the function to call to flush the I and D caches, or to not
7440 call any such function. If called, the function must take the same
7441 arguments as the common @code{_flush_func()}, that is, the address of the
7442 memory range for which the cache is being flushed, the size of the
7443 memory range, and the number 3 (to flush both caches). The default
7444 depends on the target gcc was configured for, but commonly is either
7445 @samp{_flush_func} or @samp{__cpu_flush}.
7449 These options are defined by the macro
7450 @code{TARGET_SWITCHES} in the machine description. The default for the
7451 options is also defined by that macro, which enables you to change the
7455 @node i386 and x86-64 Options
7456 @subsection Intel 386 and AMD x86-64 Options
7457 @cindex i386 Options
7458 @cindex x86-64 Options
7459 @cindex Intel 386 Options
7460 @cindex AMD x86-64 Options
7462 These @samp{-m} options are defined for the i386 and x86-64 family of
7466 @item -mcpu=@var{cpu-type}
7468 Assume the defaults for the machine type @var{cpu-type} when scheduling
7469 instructions. The choices for @var{cpu-type} are @samp{i386},
7470 @samp{i486}, @samp{i586}, @samp{i686}, @samp{pentium},
7471 @samp{pentiumpro}, @samp{pentium4}, @samp{k6}, and @samp{athlon}
7473 While picking a specific @var{cpu-type} will schedule things appropriately
7474 for that particular chip, the compiler will not generate any code that
7475 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7476 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7477 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7478 AMD chips as opposed to the Intel ones.
7480 @item -march=@var{cpu-type}
7482 Generate instructions for the machine type @var{cpu-type}. The choices
7483 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7484 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7493 @opindex mpentiumpro
7494 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7495 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7496 These synonyms are deprecated.
7498 @item -mintel-syntax
7499 @opindex mintel-syntax
7500 Emit assembly using Intel syntax opcodes instead of AT&T syntax.
7505 @opindex mno-ieee-fp
7506 Control whether or not the compiler uses IEEE floating point
7507 comparisons. These handle correctly the case where the result of a
7508 comparison is unordered.
7511 @opindex msoft-float
7512 Generate output containing library calls for floating point.
7513 @strong{Warning:} the requisite libraries are not part of GCC@.
7514 Normally the facilities of the machine's usual C compiler are used, but
7515 this can't be done directly in cross-compilation. You must make your
7516 own arrangements to provide suitable library functions for
7519 On machines where a function returns floating point results in the 80387
7520 register stack, some floating point opcodes may be emitted even if
7521 @option{-msoft-float} is used.
7523 @item -mno-fp-ret-in-387
7524 @opindex mno-fp-ret-in-387
7525 Do not use the FPU registers for return values of functions.
7527 The usual calling convention has functions return values of types
7528 @code{float} and @code{double} in an FPU register, even if there
7529 is no FPU@. The idea is that the operating system should emulate
7532 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7533 in ordinary CPU registers instead.
7535 @item -mno-fancy-math-387
7536 @opindex mno-fancy-math-387
7537 Some 387 emulators do not support the @code{sin}, @code{cos} and
7538 @code{sqrt} instructions for the 387. Specify this option to avoid
7539 generating those instructions. This option is the default on FreeBSD@.
7540 As of revision 2.6.1, these instructions are not generated unless you
7541 also use the @option{-funsafe-math-optimizations} switch.
7543 @item -malign-double
7544 @itemx -mno-align-double
7545 @opindex malign-double
7546 @opindex mno-align-double
7547 Control whether GCC aligns @code{double}, @code{long double}, and
7548 @code{long long} variables on a two word boundary or a one word
7549 boundary. Aligning @code{double} variables on a two word boundary will
7550 produce code that runs somewhat faster on a @samp{Pentium} at the
7551 expense of more memory.
7553 @item -m128bit-long-double
7554 @opindex m128bit-long-double
7555 Control the size of @code{long double} type. i386 application binary interface
7556 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7557 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7558 impossible to reach with 12 byte long doubles in the array accesses.
7560 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7561 structures and arrays containing @code{long double} will change their size as
7562 well as function calling convention for function taking @code{long double}
7565 @item -m96bit-long-double
7566 @opindex m96bit-long-double
7567 Set the size of @code{long double} to 96 bits as required by the i386
7568 application binary interface. This is the default.
7571 @itemx -mno-svr3-shlib
7572 @opindex msvr3-shlib
7573 @opindex mno-svr3-shlib
7574 Control whether GCC places uninitialized local variables into the
7575 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7576 into @code{bss}. These options are meaningful only on System V Release 3.
7580 Use a different function-calling convention, in which functions that
7581 take a fixed number of arguments return with the @code{ret} @var{num}
7582 instruction, which pops their arguments while returning. This saves one
7583 instruction in the caller since there is no need to pop the arguments
7586 You can specify that an individual function is called with this calling
7587 sequence with the function attribute @samp{stdcall}. You can also
7588 override the @option{-mrtd} option by using the function attribute
7589 @samp{cdecl}. @xref{Function Attributes}.
7591 @strong{Warning:} this calling convention is incompatible with the one
7592 normally used on Unix, so you cannot use it if you need to call
7593 libraries compiled with the Unix compiler.
7595 Also, you must provide function prototypes for all functions that
7596 take variable numbers of arguments (including @code{printf});
7597 otherwise incorrect code will be generated for calls to those
7600 In addition, seriously incorrect code will result if you call a
7601 function with too many arguments. (Normally, extra arguments are
7602 harmlessly ignored.)
7604 @item -mregparm=@var{num}
7606 Control how many registers are used to pass integer arguments. By
7607 default, no registers are used to pass arguments, and at most 3
7608 registers can be used. You can control this behavior for a specific
7609 function by using the function attribute @samp{regparm}.
7610 @xref{Function Attributes}.
7612 @strong{Warning:} if you use this switch, and
7613 @var{num} is nonzero, then you must build all modules with the same
7614 value, including any libraries. This includes the system libraries and
7617 @item -mpreferred-stack-boundary=@var{num}
7618 @opindex mpreferred-stack-boundary
7619 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7620 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7621 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7622 size (@option{-Os}), in which case the default is the minimum correct
7623 alignment (4 bytes for x86, and 8 bytes for x86-64).
7625 On Pentium and PentiumPro, @code{double} and @code{long double} values
7626 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7627 suffer significant run time performance penalties. On Pentium III, the
7628 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7629 penalties if it is not 16 byte aligned.
7631 To ensure proper alignment of this values on the stack, the stack boundary
7632 must be as aligned as that required by any value stored on the stack.
7633 Further, every function must be generated such that it keeps the stack
7634 aligned. Thus calling a function compiled with a higher preferred
7635 stack boundary from a function compiled with a lower preferred stack
7636 boundary will most likely misalign the stack. It is recommended that
7637 libraries that use callbacks always use the default setting.
7639 This extra alignment does consume extra stack space, and generally
7640 increases code size. Code that is sensitive to stack space usage, such
7641 as embedded systems and operating system kernels, may want to reduce the
7642 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7656 These switches enable or disable the use of built-in functions that allow
7657 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7659 The following machine modes are available for use with MMX built-in functions
7660 (@pxref{Vector Extensions}): @code{V2SI} for a vector of two 32 bit integers,
7661 @code{V4HI} for a vector of four 16 bit integers, and @code{V8QI} for a
7662 vector of eight 8 bit integers. Some of the built-in functions operate on
7663 MMX registers as a whole 64 bit entity, these use @code{DI} as their mode.
7665 If 3Dnow extensions are enabled, @code{V2SF} is used as a mode for a vector
7666 of two 32 bit floating point values.
7668 If SSE extensions are enabled, @code{V4SF} is used for a vector of four 32 bit
7669 floating point values. Some instructions use a vector of four 32 bit
7670 integers, these use @code{V4SI}. Finally, some instructions operate on an
7671 entire vector register, interpreting it as a 128 bit integer, these use mode
7674 The following built-in functions are made available by @option{-mmmx}:
7676 @item v8qi __builtin_ia32_paddb (v8qi, v8qi)
7677 Generates the @code{paddb} machine instruction.
7678 @item v4hi __builtin_ia32_paddw (v4hi, v4hi)
7679 Generates the @code{paddw} machine instruction.
7680 @item v2si __builtin_ia32_paddd (v2si, v2si)
7681 Generates the @code{paddd} machine instruction.
7682 @item v8qi __builtin_ia32_psubb (v8qi, v8qi)
7683 Generates the @code{psubb} machine instruction.
7684 @item v4hi __builtin_ia32_psubw (v4hi, v4hi)
7685 Generates the @code{psubw} machine instruction.
7686 @item v2si __builtin_ia32_psubd (v2si, v2si)
7687 Generates the @code{psubd} machine instruction.
7689 @item v8qi __builtin_ia32_paddsb (v8qi, v8qi)
7690 Generates the @code{paddsb} machine instruction.
7691 @item v4hi __builtin_ia32_paddsw (v4hi, v4hi)
7692 Generates the @code{paddsw} machine instruction.
7693 @item v8qi __builtin_ia32_psubsb (v8qi, v8qi)
7694 Generates the @code{psubsb} machine instruction.
7695 @item v4hi __builtin_ia32_psubsw (v4hi, v4hi)
7696 Generates the @code{psubsw} machine instruction.
7698 @item v8qi __builtin_ia32_paddusb (v8qi, v8qi)
7699 Generates the @code{paddusb} machine instruction.
7700 @item v4hi __builtin_ia32_paddusw (v4hi, v4hi)
7701 Generates the @code{paddusw} machine instruction.
7702 @item v8qi __builtin_ia32_psubusb (v8qi, v8qi)
7703 Generates the @code{psubusb} machine instruction.
7704 @item v4hi __builtin_ia32_psubusw (v4hi, v4hi)
7705 Generates the @code{psubusw} machine instruction.
7707 @item v4hi __builtin_ia32_pmullw (v4hi, v4hi)
7708 Generates the @code{pmullw} machine instruction.
7709 @item v4hi __builtin_ia32_pmulhw (v4hi, v4hi)
7710 Generates the @code{pmulhw} machine instruction.
7712 @item di __builtin_ia32_pand (di, di)
7713 Generates the @code{pand} machine instruction.
7714 @item di __builtin_ia32_pandn (di,di)
7715 Generates the @code{pandn} machine instruction.
7716 @item di __builtin_ia32_por (di, di)
7717 Generates the @code{por} machine instruction.
7718 @item di __builtin_ia32_pxor (di, di)
7719 Generates the @code{pxor} machine instruction.
7721 @item v8qi __builtin_ia32_pcmpeqb (v8qi, v8qi)
7722 Generates the @code{pcmpeqb} machine instruction.
7723 @item v4hi __builtin_ia32_pcmpeqw (v4hi, v4hi)
7724 Generates the @code{pcmpeqw} machine instruction.
7725 @item v2si __builtin_ia32_pcmpeqd (v2si, v2si)
7726 Generates the @code{pcmpeqd} machine instruction.
7727 @item v8qi __builtin_ia32_pcmpgtb (v8qi, v8qi)
7728 Generates the @code{pcmpgtb} machine instruction.
7729 @item v4hi __builtin_ia32_pcmpgtw (v4hi, v4hi)
7730 Generates the @code{pcmpgtw} machine instruction.
7731 @item v2si __builtin_ia32_pcmpgtd (v2si, v2si)
7732 Generates the @code{pcmpgtd} machine instruction.
7734 @item v8qi __builtin_ia32_punpckhbw (v8qi, v8qi)
7735 Generates the @code{punpckhbw} machine instruction.
7736 @item v4hi __builtin_ia32_punpckhwd (v4hi, v4hi)
7737 Generates the @code{punpckhwd} machine instruction.
7738 @item v2si __builtin_ia32_punpckhdq (v2si, v2si)
7739 Generates the @code{punpckhdq} machine instruction.
7740 @item v8qi __builtin_ia32_punpcklbw (v8qi, v8qi)
7741 Generates the @code{punpcklbw} machine instruction.
7742 @item v4hi __builtin_ia32_punpcklwd (v4hi, v4hi)
7743 Generates the @code{punpcklwd} machine instruction.
7744 @item v2si __builtin_ia32_punpckldq (v2si, v2si)
7745 Generates the @code{punpckldq} machine instruction.
7747 @item v8qi __builtin_ia32_packsswb (v4hi, v4hi)
7748 Generates the @code{packsswb} machine instruction.
7749 @item v4hi __builtin_ia32_packssdw (v2si, v2si)
7750 Generates the @code{packssdw} machine instruction.
7751 @item v8qi __builtin_ia32_packuswb (v4hi, v4hi)
7752 Generates the @code{packuswb} machine instruction.
7756 The following built-in functions are made available either with @option{-msse}, or
7757 with a combination of @option{-m3dnow} and @option{-march=athlon}.
7760 @item v4hi __builtin_ia32_pmulhuw (v4hi, v4hi)
7761 Generates the @code{pmulhuw} machine instruction.
7763 @item v8qi __builtin_ia32_pavgb (v8qi, v8qi)
7764 Generates the @code{pavgb} machine instruction.
7765 @item v4hi __builtin_ia32_pavgw (v4hi, v4hi)
7766 Generates the @code{pavgw} machine instruction.
7767 @item v4hi __builtin_ia32_psadbw (v8qi, v8qi)
7768 Generates the @code{psadbw} machine instruction.
7770 @item v8qi __builtin_ia32_pmaxub (v8qi, v8qi)
7771 Generates the @code{pmaxub} machine instruction.
7772 @item v4hi __builtin_ia32_pmaxsw (v4hi, v4hi)
7773 Generates the @code{pmaxsw} machine instruction.
7774 @item v8qi __builtin_ia32_pminub (v8qi, v8qi)
7775 Generates the @code{pminub} machine instruction.
7776 @item v4hi __builtin_ia32_pminsw (v4hi, v4hi)
7777 Generates the @code{pminsw} machine instruction.
7779 @item int __builtin_ia32_pextrw (v4hi, int)
7780 Generates the @code{pextrw} machine instruction.
7781 @item v4hi __builtin_ia32_pinsrw (v4hi, int, int)
7782 Generates the @code{pinsrw} machine instruction.
7784 @item int __builtin_ia32_pmovmskb (v8qi)
7785 Generates the @code{pmovmskb} machine instruction.
7786 @item void __builtin_ia32_maskmovq (v8qi, v8qi, char *)
7787 Generates the @code{maskmovq} machine instruction.
7788 @item void __builtin_ia32_movntq (di *, di)
7789 Generates the @code{movntq} machine instruction.
7790 @item void __builtin_ia32_sfence (void)
7791 Generates the @code{sfence} machine instruction.
7792 @item void __builtin_ia32_prefetch (char *, int selector)
7793 Generates a prefetch machine instruction, depending on the value of
7794 selector. If @code{selector} is 0, it generates @code{prefetchnta}; for
7795 a value of 1, it generates @code{prefetcht0}; for a value of 2, it generates
7796 @code{prefetcht1}; and for a value of 3 it generates @code{prefetcht2}.
7800 The following built-in functions are available when @option{-msse} is used.
7803 @item int __builtin_ia32_comieq (v4sf, v4sf)
7804 Generates the @code{comiss} machine instruction and performs an equality
7805 comparison. The return value is the truth value of that comparison.
7806 @item int __builtin_ia32_comineq (v4sf, v4sf)
7807 Generates the @code{comiss} machine instruction and performs an inequality
7808 comparison. The return value is the truth value of that comparison.
7809 @item int __builtin_ia32_comilt (v4sf, v4sf)
7810 Generates the @code{comiss} machine instruction and performs a ``less than''
7811 comparison. The return value is the truth value of that comparison.
7812 @item int __builtin_ia32_comile (v4sf, v4sf)
7813 Generates the @code{comiss} machine instruction and performs a ``less or
7814 equal'' comparison. The return value is the truth value of that comparison.
7815 @item int __builtin_ia32_comigt (v4sf, v4sf)
7816 Generates the @code{comiss} machine instruction and performs a ``greater than''
7817 comparison. The return value is the truth value of that comparison.
7818 @item int __builtin_ia32_comige (v4sf, v4sf)
7819 Generates the @code{comiss} machine instruction and performs a ``greater or
7820 equal'' comparison. The return value is the truth value of that comparison.
7822 @item int __builtin_ia32_ucomieq (v4sf, v4sf)
7823 Generates the @code{ucomiss} machine instruction and performs an equality
7824 comparison. The return value is the truth value of that comparison.
7825 @item int __builtin_ia32_ucomineq (v4sf, v4sf)
7826 Generates the @code{ucomiss} machine instruction and performs an inequality
7827 comparison. The return value is the truth value of that comparison.
7828 @item int __builtin_ia32_ucomilt (v4sf, v4sf)
7829 Generates the @code{ucomiss} machine instruction and performs a ``less than''
7830 comparison. The return value is the truth value of that comparison.
7831 @item int __builtin_ia32_ucomile (v4sf, v4sf)
7832 Generates the @code{ucomiss} machine instruction and performs a ``less or
7833 equal'' comparison. The return value is the truth value of that comparison.
7834 @item int __builtin_ia32_ucomigt (v4sf, v4sf)
7835 Generates the @code{ucomiss} machine instruction and performs a ``greater than''
7836 comparison. The return value is the truth value of that comparison.
7837 @item int __builtin_ia32_ucomige (v4sf, v4sf)
7838 Generates the @code{ucomiss} machine instruction and performs a ``greater or
7839 equal'' comparison. The return value is the truth value of that comparison.
7841 @item v4sf __builtin_ia32_addps (v4sf, v4sf)
7842 Generates the @code{addps} machine instruction.
7843 @item v4sf __builtin_ia32_addss (v4sf, v4sf)
7844 Generates the @code{addss} machine instruction.
7845 @item v4sf __builtin_ia32_subps (v4sf, v4sf)
7846 Generates the @code{subps} machine instruction.
7847 @item v4sf __builtin_ia32_subss (v4sf, v4sf)
7848 Generates the @code{subss} machine instruction.
7849 @item v4sf __builtin_ia32_mulps (v4sf, v4sf)
7850 Generates the @code{mulps} machine instruction.
7851 @item v4sf __builtin_ia32_mulss (v4sf, v4sf)
7852 Generates the @code{mulss} machine instruction.
7853 @item v4sf __builtin_ia32_divps (v4sf, v4sf)
7854 Generates the @code{divps} machine instruction.
7855 @item v4sf __builtin_ia32_divss (v4sf, v4sf)
7856 Generates the @code{divss} machine instruction.
7858 @item v4si __builtin_ia32_cmpeqps (v4sf, v4sf)
7859 Generates the @code{cmpeqps} machine instruction.
7860 @item v4si __builtin_ia32_cmplts (v4sf, v4sf)
7861 Generates the @code{cmpltps} machine instruction.
7862 @item v4si __builtin_ia32_cmpleps (v4sf, v4sf)
7863 Generates the @code{cmpleps} machine instruction.
7864 @item v4si __builtin_ia32_cmpgtps (v4sf, v4sf)
7865 Generates the @code{cmpgtps} machine instruction.
7866 @item v4si __builtin_ia32_cmpgeps (v4sf, v4sf)
7867 Generates the @code{cmpgeps} machine instruction.
7868 @item v4si __builtin_ia32_cmpunordps (v4sf, v4sf)
7869 Generates the @code{cmpunodps} machine instruction.
7870 @item v4si __builtin_ia32_cmpneqps (v4sf, v4sf)
7871 Generates the @code{cmpeqps} machine instruction.
7872 @item v4si __builtin_ia32_cmpnltps (v4sf, v4sf)
7873 Generates the @code{cmpltps} machine instruction.
7874 @item v4si __builtin_ia32_cmpnleps (v4sf, v4sf)
7875 Generates the @code{cmpleps} machine instruction.
7876 @item v4si __builtin_ia32_cmpngtps (v4sf, v4sf)
7877 Generates the @code{cmpgtps} machine instruction.
7878 @item v4si __builtin_ia32_cmpngeps (v4sf, v4sf)
7879 Generates the @code{cmpgeps} machine instruction.
7880 @item v4si __builtin_ia32_cmpordps (v4sf, v4sf)
7881 Generates the @code{cmpunodps} machine instruction.
7883 @item v4si __builtin_ia32_cmpeqss (v4sf, v4sf)
7884 Generates the @code{cmpeqss} machine instruction.
7885 @item v4si __builtin_ia32_cmpltss (v4sf, v4sf)
7886 Generates the @code{cmpltss} machine instruction.
7887 @item v4si __builtin_ia32_cmpless (v4sf, v4sf)
7888 Generates the @code{cmpless} machine instruction.
7889 @item v4si __builtin_ia32_cmpgtss (v4sf, v4sf)
7890 Generates the @code{cmpgtss} machine instruction.
7891 @item v4si __builtin_ia32_cmpgess (v4sf, v4sf)
7892 Generates the @code{cmpgess} machine instruction.
7893 @item v4si __builtin_ia32_cmpunordss (v4sf, v4sf)
7894 Generates the @code{cmpunodss} machine instruction.
7895 @item v4si __builtin_ia32_cmpneqss (v4sf, v4sf)
7896 Generates the @code{cmpeqss} machine instruction.
7897 @item v4si __builtin_ia32_cmpnlts (v4sf, v4sf)
7898 Generates the @code{cmpltss} machine instruction.
7899 @item v4si __builtin_ia32_cmpnless (v4sf, v4sf)
7900 Generates the @code{cmpless} machine instruction.
7901 @item v4si __builtin_ia32_cmpngtss (v4sf, v4sf)
7902 Generates the @code{cmpgtss} machine instruction.
7903 @item v4si __builtin_ia32_cmpngess (v4sf, v4sf)
7904 Generates the @code{cmpgess} machine instruction.
7905 @item v4si __builtin_ia32_cmpordss (v4sf, v4sf)
7906 Generates the @code{cmpunodss} machine instruction.
7908 @item v4sf __builtin_ia32_maxps (v4sf, v4sf)
7909 Generates the @code{maxps} machine instruction.
7910 @item v4sf __builtin_ia32_maxsss (v4sf, v4sf)
7911 Generates the @code{maxss} machine instruction.
7912 @item v4sf __builtin_ia32_minps (v4sf, v4sf)
7913 Generates the @code{minps} machine instruction.
7914 @item v4sf __builtin_ia32_minsss (v4sf, v4sf)
7915 Generates the @code{minss} machine instruction.
7917 @item ti __builtin_ia32_andps (ti, ti)
7918 Generates the @code{andps} machine instruction.
7919 @item ti __builtin_ia32_andnps (ti, ti)
7920 Generates the @code{andnps} machine instruction.
7921 @item ti __builtin_ia32_orps (ti, ti)
7922 Generates the @code{orps} machine instruction.
7923 @item ti __builtin_ia32_xorps (ti, ti)
7924 Generates the @code{xorps} machine instruction.
7926 @item v4sf __builtin_ia32_movps (v4sf, v4sf)
7927 Generates the @code{movps} machine instruction.
7928 @item v4sf __builtin_ia32_movhlps (v4sf, v4sf)
7929 Generates the @code{movhlps} machine instruction.
7930 @item v4sf __builtin_ia32_movlhps (v4sf, v4sf)
7931 Generates the @code{movlhps} machine instruction.
7932 @item v4sf __builtin_ia32_unpckhps (v4sf, v4sf)
7933 Generates the @code{unpckhps} machine instruction.
7934 @item v4sf __builtin_ia32_unpcklps (v4sf, v4sf)
7935 Generates the @code{unpcklps} machine instruction.
7937 @item v4sf __builtin_ia32_cvtpi2ps (v4sf, v2si)
7938 Generates the @code{cvtpi2ps} machine instruction.
7939 @item v2si __builtin_ia32_cvtps2pi (v4sf)
7940 Generates the @code{cvtps2pi} machine instruction.
7941 @item v4sf __builtin_ia32_cvtsi2ss (v4sf, int)
7942 Generates the @code{cvtsi2ss} machine instruction.
7943 @item int __builtin_ia32_cvtss2si (v4sf)
7944 Generates the @code{cvtsi2ss} machine instruction.
7945 @item v2si __builtin_ia32_cvttps2pi (v4sf)
7946 Generates the @code{cvttps2pi} machine instruction.
7947 @item int __builtin_ia32_cvttss2si (v4sf)
7948 Generates the @code{cvttsi2ss} machine instruction.
7950 @item v4sf __builtin_ia32_rcpps (v4sf)
7951 Generates the @code{rcpps} machine instruction.
7952 @item v4sf __builtin_ia32_rsqrtps (v4sf)
7953 Generates the @code{rsqrtps} machine instruction.
7954 @item v4sf __builtin_ia32_sqrtps (v4sf)
7955 Generates the @code{sqrtps} machine instruction.
7956 @item v4sf __builtin_ia32_rcpss (v4sf)
7957 Generates the @code{rcpss} machine instruction.
7958 @item v4sf __builtin_ia32_rsqrtss (v4sf)
7959 Generates the @code{rsqrtss} machine instruction.
7960 @item v4sf __builtin_ia32_sqrtss (v4sf)
7961 Generates the @code{sqrtss} machine instruction.
7963 @item v4sf __builtin_ia32_shufps (v4sf, v4sf, int)
7964 Generates the @code{shufps} machine instruction.
7966 @item v4sf __builtin_ia32_loadaps (float *)
7967 Generates the @code{movaps} machine instruction as a load from memory.
7968 @item void __builtin_ia32_storeaps (float *, v4sf)
7969 Generates the @code{movaps} machine instruction as a store to memory.
7970 @item v4sf __builtin_ia32_loadups (float *)
7971 Generates the @code{movups} machine instruction as a load from memory.
7972 @item void __builtin_ia32_storeups (float *, v4sf)
7973 Generates the @code{movups} machine instruction as a store to memory.
7974 @item v4sf __builtin_ia32_loadsss (float *)
7975 Generates the @code{movss} machine instruction as a load from memory.
7976 @item void __builtin_ia32_storess (float *, v4sf)
7977 Generates the @code{movss} machine instruction as a store to memory.
7979 @item v4sf __builtin_ia32_loadhps (v4sf, v2si *)
7980 Generates the @code{movhps} machine instruction as a load from memory.
7981 @item v4sf __builtin_ia32_loadlps (v4sf, v2si *)
7982 Generates the @code{movlps} machine instruction as a load from memory
7983 @item void __builtin_ia32_storehps (v4sf, v2si *)
7984 Generates the @code{movhps} machine instruction as a store to memory.
7985 @item void __builtin_ia32_storelps (v4sf, v2si *)
7986 Generates the @code{movlps} machine instruction as a store to memory.
7988 @item void __builtin_ia32_movntps (float *, v4sf)
7989 Generates the @code{movntps} machine instruction.
7990 @item int __builtin_ia32_movmskps (v4sf)
7991 Generates the @code{movntps} machine instruction.
7993 @item void __builtin_ia32_storeps1 (float *, v4sf)
7994 Generates the @code{movaps} machine instruction as a store to memory.
7995 Before storing, the value is modified with a @code{shufps} instruction
7996 so that the lowest of the four floating point elements is replicated
7997 across the entire vector that is stored.
7998 @item void __builtin_ia32_storerps (float *, v4sf)
7999 Generates the @code{movaps} machine instruction as a store to memory.
8000 Before storing, the value is modified with a @code{shufps} instruction
8001 so that the order of the four floating point elements in the vector is
8003 @item v4sf __builtin_ia32_loadps1 (float *)
8004 Generates a @code{movss} machine instruction to load a floating point
8005 value from memory, and a @code{shufps} instruction to replicate the
8006 loaded value across all four elements of the result vector.
8007 @item v4sf __builtin_ia32_loadrps (float *)
8008 Generates a @code{movaps} machine instruction to load a vector from
8009 memory, and a @code{shufps} instruction to reverse the order of the
8010 four floating point elements in the result vector.
8011 @item v4sf __builtin_ia32_setps (float, float, float, float)
8012 Constructs a vector from four single floating point values. The return
8013 value is equal to the value that would result from storing the four
8014 arguments into consecutive memory locations and then executing a
8015 @code{movaps} to load the vector from memory.
8016 @item v4sf __builtin_ia32_setps1 (float)
8017 Constructs a vector from a single floating point value by replicating
8018 it across all four elements of the result vector.
8022 @itemx -mno-push-args
8024 @opindex mno-push-args
8025 Use PUSH operations to store outgoing parameters. This method is shorter
8026 and usually equally fast as method using SUB/MOV operations and is enabled
8027 by default. In some cases disabling it may improve performance because of
8028 improved scheduling and reduced dependencies.
8030 @item -maccumulate-outgoing-args
8031 @opindex maccumulate-outgoing-args
8032 If enabled, the maximum amount of space required for outgoing arguments will be
8033 computed in the function prologue. This is faster on most modern CPUs
8034 because of reduced dependencies, improved scheduling and reduced stack usage
8035 when preferred stack boundary is not equal to 2. The drawback is a notable
8036 increase in code size. This switch implies @option{-mno-push-args}.
8040 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8041 on thread-safe exception handling must compile and link all code with the
8042 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8043 @option{-D_MT}; when linking, it links in a special thread helper library
8044 @option{-lmingwthrd} which cleans up per thread exception handling data.
8046 @item -mno-align-stringops
8047 @opindex mno-align-stringops
8048 Do not align destination of inlined string operations. This switch reduces
8049 code size and improves performance in case the destination is already aligned,
8050 but gcc don't know about it.
8052 @item -minline-all-stringops
8053 @opindex minline-all-stringops
8054 By default GCC inlines string operations only when destination is known to be
8055 aligned at least to 4 byte boundary. This enables more inlining, increase code
8056 size, but may improve performance of code that depends on fast memcpy, strlen
8057 and memset for short lengths.
8059 @item -momit-leaf-frame-pointer
8060 @opindex momit-leaf-frame-pointer
8061 Don't keep the frame pointer in a register for leaf functions. This
8062 avoids the instructions to save, set up and restore frame pointers and
8063 makes an extra register available in leaf functions. The option
8064 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8065 which might make debugging harder.
8068 These @samp{-m} switches are supported in addition to the above
8069 on AMD x86-64 processors in 64-bit environments.
8076 Generate code for a 32-bit or 64-bit environment.
8077 The 32-bit environment sets int, long and pointer to 32 bits and
8078 generates code that runs on any i386 system.
8079 The 64-bit environment sets int to 32 bits and long and pointer
8080 to 64 bits and generates code for AMD's x86-64 architecture.
8083 @opindex no-red-zone
8084 Do not use a so called red zone for x86-64 code. The red zone is mandated
8085 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8086 stack pointer that will not be modified by signal or interrupt handlers
8087 and therefore can be used for temporary data without adjusting the stack
8088 pointer. The flag @option{-mno-red-zone} disables this red zone.
8092 @subsection HPPA Options
8093 @cindex HPPA Options
8095 These @samp{-m} options are defined for the HPPA family of computers:
8098 @item -march=@var{architecture-type}
8100 Generate code for the specified architecture. The choices for
8101 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8102 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8103 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8104 architecture option for your machine. Code compiled for lower numbered
8105 architectures will run on higher numbered architectures, but not the
8108 PA 2.0 support currently requires gas snapshot 19990413 or later. The
8109 next release of binutils (current is 2.9.1) will probably contain PA 2.0
8113 @itemx -mpa-risc-1-1
8114 @itemx -mpa-risc-2-0
8115 @opindex mpa-risc-1-0
8116 @opindex mpa-risc-1-1
8117 @opindex mpa-risc-2-0
8118 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8121 @opindex mbig-switch
8122 Generate code suitable for big switch tables. Use this option only if
8123 the assembler/linker complain about out of range branches within a switch
8126 @item -mjump-in-delay
8127 @opindex mjump-in-delay
8128 Fill delay slots of function calls with unconditional jump instructions
8129 by modifying the return pointer for the function call to be the target
8130 of the conditional jump.
8132 @item -mdisable-fpregs
8133 @opindex mdisable-fpregs
8134 Prevent floating point registers from being used in any manner. This is
8135 necessary for compiling kernels which perform lazy context switching of
8136 floating point registers. If you use this option and attempt to perform
8137 floating point operations, the compiler will abort.
8139 @item -mdisable-indexing
8140 @opindex mdisable-indexing
8141 Prevent the compiler from using indexing address modes. This avoids some
8142 rather obscure problems when compiling MIG generated code under MACH@.
8144 @item -mno-space-regs
8145 @opindex mno-space-regs
8146 Generate code that assumes the target has no space registers. This allows
8147 GCC to generate faster indirect calls and use unscaled index address modes.
8149 Such code is suitable for level 0 PA systems and kernels.
8151 @item -mfast-indirect-calls
8152 @opindex mfast-indirect-calls
8153 Generate code that assumes calls never cross space boundaries. This
8154 allows GCC to emit code which performs faster indirect calls.
8156 This option will not work in the presence of shared libraries or nested
8159 @item -mlong-load-store
8160 @opindex mlong-load-store
8161 Generate 3-instruction load and store sequences as sometimes required by
8162 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8165 @item -mportable-runtime
8166 @opindex mportable-runtime
8167 Use the portable calling conventions proposed by HP for ELF systems.
8171 Enable the use of assembler directives only GAS understands.
8173 @item -mschedule=@var{cpu-type}
8175 Schedule code according to the constraints for the machine type
8176 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8177 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
8178 @file{/usr/lib/sched.models} on an HP-UX system to determine the
8179 proper scheduling option for your machine.
8182 @opindex mlinker-opt
8183 Enable the optimization pass in the HPUX linker. Note this makes symbolic
8184 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
8185 in which they give bogus error messages when linking some programs.
8188 @opindex msoft-float
8189 Generate output containing library calls for floating point.
8190 @strong{Warning:} the requisite libraries are not available for all HPPA
8191 targets. Normally the facilities of the machine's usual C compiler are
8192 used, but this cannot be done directly in cross-compilation. You must make
8193 your own arrangements to provide suitable library functions for
8194 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8195 does provide software floating point support.
8197 @option{-msoft-float} changes the calling convention in the output file;
8198 therefore, it is only useful if you compile @emph{all} of a program with
8199 this option. In particular, you need to compile @file{libgcc.a}, the
8200 library that comes with GCC, with @option{-msoft-float} in order for
8204 @node Intel 960 Options
8205 @subsection Intel 960 Options
8207 These @samp{-m} options are defined for the Intel 960 implementations:
8210 @item -m@var{cpu-type}
8218 Assume the defaults for the machine type @var{cpu-type} for some of
8219 the other options, including instruction scheduling, floating point
8220 support, and addressing modes. The choices for @var{cpu-type} are
8221 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
8222 @samp{sa}, and @samp{sb}.
8229 @opindex msoft-float
8230 The @option{-mnumerics} option indicates that the processor does support
8231 floating-point instructions. The @option{-msoft-float} option indicates
8232 that floating-point support should not be assumed.
8234 @item -mleaf-procedures
8235 @itemx -mno-leaf-procedures
8236 @opindex mleaf-procedures
8237 @opindex mno-leaf-procedures
8238 Do (or do not) attempt to alter leaf procedures to be callable with the
8239 @code{bal} instruction as well as @code{call}. This will result in more
8240 efficient code for explicit calls when the @code{bal} instruction can be
8241 substituted by the assembler or linker, but less efficient code in other
8242 cases, such as calls via function pointers, or using a linker that doesn't
8243 support this optimization.
8246 @itemx -mno-tail-call
8248 @opindex mno-tail-call
8249 Do (or do not) make additional attempts (beyond those of the
8250 machine-independent portions of the compiler) to optimize tail-recursive
8251 calls into branches. You may not want to do this because the detection of
8252 cases where this is not valid is not totally complete. The default is
8253 @option{-mno-tail-call}.
8255 @item -mcomplex-addr
8256 @itemx -mno-complex-addr
8257 @opindex mcomplex-addr
8258 @opindex mno-complex-addr
8259 Assume (or do not assume) that the use of a complex addressing mode is a
8260 win on this implementation of the i960. Complex addressing modes may not
8261 be worthwhile on the K-series, but they definitely are on the C-series.
8262 The default is currently @option{-mcomplex-addr} for all processors except
8266 @itemx -mno-code-align
8267 @opindex mcode-align
8268 @opindex mno-code-align
8269 Align code to 8-byte boundaries for faster fetching (or don't bother).
8270 Currently turned on by default for C-series implementations only.
8273 @item -mclean-linkage
8274 @itemx -mno-clean-linkage
8275 @opindex mclean-linkage
8276 @opindex mno-clean-linkage
8277 These options are not fully implemented.
8281 @itemx -mic2.0-compat
8282 @itemx -mic3.0-compat
8284 @opindex mic2.0-compat
8285 @opindex mic3.0-compat
8286 Enable compatibility with iC960 v2.0 or v3.0.
8290 @opindex masm-compat
8292 Enable compatibility with the iC960 assembler.
8294 @item -mstrict-align
8295 @itemx -mno-strict-align
8296 @opindex mstrict-align
8297 @opindex mno-strict-align
8298 Do not permit (do permit) unaligned accesses.
8302 Enable structure-alignment compatibility with Intel's gcc release version
8303 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8305 @item -mlong-double-64
8306 @opindex mlong-double-64
8307 Implement type @samp{long double} as 64-bit floating point numbers.
8308 Without the option @samp{long double} is implemented by 80-bit
8309 floating point numbers. The only reason we have it because there is
8310 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8311 is only useful for people using soft-float targets. Otherwise, we
8312 should recommend against use of it.
8316 @node DEC Alpha Options
8317 @subsection DEC Alpha Options
8319 These @samp{-m} options are defined for the DEC Alpha implementations:
8322 @item -mno-soft-float
8324 @opindex mno-soft-float
8325 @opindex msoft-float
8326 Use (do not use) the hardware floating-point instructions for
8327 floating-point operations. When @option{-msoft-float} is specified,
8328 functions in @file{libgcc.a} will be used to perform floating-point
8329 operations. Unless they are replaced by routines that emulate the
8330 floating-point operations, or compiled in such a way as to call such
8331 emulations routines, these routines will issue floating-point
8332 operations. If you are compiling for an Alpha without floating-point
8333 operations, you must ensure that the library is built so as not to call
8336 Note that Alpha implementations without floating-point operations are
8337 required to have floating-point registers.
8342 @opindex mno-fp-regs
8343 Generate code that uses (does not use) the floating-point register set.
8344 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8345 register set is not used, floating point operands are passed in integer
8346 registers as if they were integers and floating-point results are passed
8347 in $0 instead of $f0. This is a non-standard calling sequence, so any
8348 function with a floating-point argument or return value called by code
8349 compiled with @option{-mno-fp-regs} must also be compiled with that
8352 A typical use of this option is building a kernel that does not use,
8353 and hence need not save and restore, any floating-point registers.
8357 The Alpha architecture implements floating-point hardware optimized for
8358 maximum performance. It is mostly compliant with the IEEE floating
8359 point standard. However, for full compliance, software assistance is
8360 required. This option generates code fully IEEE compliant code
8361 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8362 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8363 defined during compilation. The resulting code is less efficient but is
8364 able to correctly support denormalized numbers and exceptional IEEE
8365 values such as not-a-number and plus/minus infinity. Other Alpha
8366 compilers call this option @option{-ieee_with_no_inexact}.
8368 @item -mieee-with-inexact
8369 @opindex mieee-with-inexact
8370 This is like @option{-mieee} except the generated code also maintains
8371 the IEEE @var{inexact-flag}. Turning on this option causes the
8372 generated code to implement fully-compliant IEEE math. In addition to
8373 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8374 macro. On some Alpha implementations the resulting code may execute
8375 significantly slower than the code generated by default. Since there is
8376 very little code that depends on the @var{inexact-flag}, you should
8377 normally not specify this option. Other Alpha compilers call this
8378 option @option{-ieee_with_inexact}.
8380 @item -mfp-trap-mode=@var{trap-mode}
8381 @opindex mfp-trap-mode
8382 This option controls what floating-point related traps are enabled.
8383 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8384 The trap mode can be set to one of four values:
8388 This is the default (normal) setting. The only traps that are enabled
8389 are the ones that cannot be disabled in software (e.g., division by zero
8393 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8397 Like @samp{su}, but the instructions are marked to be safe for software
8398 completion (see Alpha architecture manual for details).
8401 Like @samp{su}, but inexact traps are enabled as well.
8404 @item -mfp-rounding-mode=@var{rounding-mode}
8405 @opindex mfp-rounding-mode
8406 Selects the IEEE rounding mode. Other Alpha compilers call this option
8407 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8412 Normal IEEE rounding mode. Floating point numbers are rounded towards
8413 the nearest machine number or towards the even machine number in case
8417 Round towards minus infinity.
8420 Chopped rounding mode. Floating point numbers are rounded towards zero.
8423 Dynamic rounding mode. A field in the floating point control register
8424 (@var{fpcr}, see Alpha architecture reference manual) controls the
8425 rounding mode in effect. The C library initializes this register for
8426 rounding towards plus infinity. Thus, unless your program modifies the
8427 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8430 @item -mtrap-precision=@var{trap-precision}
8431 @opindex mtrap-precision
8432 In the Alpha architecture, floating point traps are imprecise. This
8433 means without software assistance it is impossible to recover from a
8434 floating trap and program execution normally needs to be terminated.
8435 GCC can generate code that can assist operating system trap handlers
8436 in determining the exact location that caused a floating point trap.
8437 Depending on the requirements of an application, different levels of
8438 precisions can be selected:
8442 Program precision. This option is the default and means a trap handler
8443 can only identify which program caused a floating point exception.
8446 Function precision. The trap handler can determine the function that
8447 caused a floating point exception.
8450 Instruction precision. The trap handler can determine the exact
8451 instruction that caused a floating point exception.
8454 Other Alpha compilers provide the equivalent options called
8455 @option{-scope_safe} and @option{-resumption_safe}.
8457 @item -mieee-conformant
8458 @opindex mieee-conformant
8459 This option marks the generated code as IEEE conformant. You must not
8460 use this option unless you also specify @option{-mtrap-precision=i} and either
8461 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8462 is to emit the line @samp{.eflag 48} in the function prologue of the
8463 generated assembly file. Under DEC Unix, this has the effect that
8464 IEEE-conformant math library routines will be linked in.
8466 @item -mbuild-constants
8467 @opindex mbuild-constants
8468 Normally GCC examines a 32- or 64-bit integer constant to
8469 see if it can construct it from smaller constants in two or three
8470 instructions. If it cannot, it will output the constant as a literal and
8471 generate code to load it from the data segment at runtime.
8473 Use this option to require GCC to construct @emph{all} integer constants
8474 using code, even if it takes more instructions (the maximum is six).
8476 You would typically use this option to build a shared library dynamic
8477 loader. Itself a shared library, it must relocate itself in memory
8478 before it can find the variables and constants in its own data segment.
8484 Select whether to generate code to be assembled by the vendor-supplied
8485 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8499 Indicate whether GCC should generate code to use the optional BWX,
8500 CIX, and MAX instruction sets. The default is to use the instruction sets
8501 supported by the CPU type specified via @option{-mcpu=} option or that
8502 of the CPU on which GCC was built if none was specified.
8504 @item -mcpu=@var{cpu_type}
8506 Set the instruction set, register set, and instruction scheduling
8507 parameters for machine type @var{cpu_type}. You can specify either the
8508 @samp{EV} style name or the corresponding chip number. GCC
8509 supports scheduling parameters for the EV4 and EV5 family of processors
8510 and will choose the default values for the instruction set from
8511 the processor you specify. If you do not specify a processor type,
8512 GCC will default to the processor on which the compiler was built.
8514 Supported values for @var{cpu_type} are
8519 Schedules as an EV4 and has no instruction set extensions.
8523 Schedules as an EV5 and has no instruction set extensions.
8527 Schedules as an EV5 and supports the BWX extension.
8532 Schedules as an EV5 and supports the BWX and MAX extensions.
8536 Schedules as an EV5 (until Digital releases the scheduling parameters
8537 for the EV6) and supports the BWX, CIX, and MAX extensions.
8540 @item -mmemory-latency=@var{time}
8541 @opindex mmemory-latency
8542 Sets the latency the scheduler should assume for typical memory
8543 references as seen by the application. This number is highly
8544 dependent on the memory access patterns used by the application
8545 and the size of the external cache on the machine.
8547 Valid options for @var{time} are
8551 A decimal number representing clock cycles.
8557 The compiler contains estimates of the number of clock cycles for
8558 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8559 (also called Dcache, Scache, and Bcache), as well as to main memory.
8560 Note that L3 is only valid for EV5.
8565 @node Clipper Options
8566 @subsection Clipper Options
8568 These @samp{-m} options are defined for the Clipper implementations:
8573 Produce code for a C300 Clipper processor. This is the default.
8577 Produce code for a C400 Clipper processor, i.e.@: use floating point
8581 @node H8/300 Options
8582 @subsection H8/300 Options
8584 These @samp{-m} options are defined for the H8/300 implementations:
8589 Shorten some address references at link time, when possible; uses the
8590 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8591 ld.info, Using ld}, for a fuller description.
8595 Generate code for the H8/300H@.
8599 Generate code for the H8/S@.
8603 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8607 Make @code{int} data 32 bits by default.
8611 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8612 The default for the H8/300H and H8/S is to align longs and floats on 4
8614 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8615 This option has no effect on the H8/300.
8619 @subsection SH Options
8621 These @samp{-m} options are defined for the SH implementations:
8626 Generate code for the SH1.
8630 Generate code for the SH2.
8634 Generate code for the SH3.
8638 Generate code for the SH3e.
8642 Generate code for the SH4 without a floating-point unit.
8644 @item -m4-single-only
8645 @opindex m4-single-only
8646 Generate code for the SH4 with a floating-point unit that only
8647 supports single-precision arithmetic.
8651 Generate code for the SH4 assuming the floating-point unit is in
8652 single-precision mode by default.
8656 Generate code for the SH4.
8660 Compile code for the processor in big endian mode.
8664 Compile code for the processor in little endian mode.
8668 Align doubles at 64-bit boundaries. Note that this changes the calling
8669 conventions, and thus some functions from the standard C library will
8670 not work unless you recompile it first with @option{-mdalign}.
8674 Shorten some address references at link time, when possible; uses the
8675 linker option @option{-relax}.
8679 Use 32-bit offsets in @code{switch} tables. The default is to use
8684 Enable the use of the instruction @code{fmovd}.
8688 Comply with the calling conventions defined by Hitachi.
8692 Mark the @code{MAC} register as call-clobbered, even if
8693 @option{-mhitachi} is given.
8697 Increase IEEE-compliance of floating-point code.
8701 Dump instruction size and location in the assembly code.
8705 This option is deprecated. It pads structures to multiple of 4 bytes,
8706 which is incompatible with the SH ABI@.
8710 Optimize for space instead of speed. Implied by @option{-Os}.
8714 When generating position-independent code, emit function calls using
8715 the Global Offset Table instead of the Procedure Linkage Table.
8719 Generate a library function call to invalidate instruction cache
8720 entries, after fixing up a trampoline. This library function call
8721 doesn't assume it can write to the whole memory address space. This
8722 is the default when the target is @code{sh-*-linux*}.
8725 @node System V Options
8726 @subsection Options for System V
8728 These additional options are available on System V Release 4 for
8729 compatibility with other compilers on those systems:
8734 Create a shared object.
8735 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8739 Identify the versions of each tool used by the compiler, in a
8740 @code{.ident} assembler directive in the output.
8744 Refrain from adding @code{.ident} directives to the output file (this is
8747 @item -YP,@var{dirs}
8749 Search the directories @var{dirs}, and no others, for libraries
8750 specified with @option{-l}.
8754 Look in the directory @var{dir} to find the M4 preprocessor.
8755 The assembler uses this option.
8756 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8757 @c the generic assembler that comes with Solaris takes just -Ym.
8760 @node TMS320C3x/C4x Options
8761 @subsection TMS320C3x/C4x Options
8762 @cindex TMS320C3x/C4x Options
8764 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8768 @item -mcpu=@var{cpu_type}
8770 Set the instruction set, register set, and instruction scheduling
8771 parameters for machine type @var{cpu_type}. Supported values for
8772 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8773 @samp{c44}. The default is @samp{c40} to generate code for the
8778 @itemx -msmall-memory
8780 @opindex mbig-memory
8782 @opindex msmall-memory
8784 Generates code for the big or small memory model. The small memory
8785 model assumed that all data fits into one 64K word page. At run-time
8786 the data page (DP) register must be set to point to the 64K page
8787 containing the .bss and .data program sections. The big memory model is
8788 the default and requires reloading of the DP register for every direct
8795 Allow (disallow) allocation of general integer operands into the block
8802 Enable (disable) generation of code using decrement and branch,
8803 DBcond(D), instructions. This is enabled by default for the C4x. To be
8804 on the safe side, this is disabled for the C3x, since the maximum
8805 iteration count on the C3x is @math{2^23 + 1} (but who iterates loops more than
8806 @math{2^23} times on the C3x?). Note that GCC will try to reverse a loop so
8807 that it can utilise the decrement and branch instruction, but will give
8808 up if there is more than one memory reference in the loop. Thus a loop
8809 where the loop counter is decremented can generate slightly more
8810 efficient code, in cases where the RPTB instruction cannot be utilised.
8812 @item -mdp-isr-reload
8814 @opindex mdp-isr-reload
8816 Force the DP register to be saved on entry to an interrupt service
8817 routine (ISR), reloaded to point to the data section, and restored on
8818 exit from the ISR@. This should not be required unless someone has
8819 violated the small memory model by modifying the DP register, say within
8826 For the C3x use the 24-bit MPYI instruction for integer multiplies
8827 instead of a library call to guarantee 32-bit results. Note that if one
8828 of the operands is a constant, then the multiplication will be performed
8829 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8830 then squaring operations are performed inline instead of a library call.
8833 @itemx -mno-fast-fix
8835 @opindex mno-fast-fix
8836 The C3x/C4x FIX instruction to convert a floating point value to an
8837 integer value chooses the nearest integer less than or equal to the
8838 floating point value rather than to the nearest integer. Thus if the
8839 floating point number is negative, the result will be incorrectly
8840 truncated an additional code is necessary to detect and correct this
8841 case. This option can be used to disable generation of the additional
8842 code required to correct the result.
8848 Enable (disable) generation of repeat block sequences using the RPTB
8849 instruction for zero overhead looping. The RPTB construct is only used
8850 for innermost loops that do not call functions or jump across the loop
8851 boundaries. There is no advantage having nested RPTB loops due to the
8852 overhead required to save and restore the RC, RS, and RE registers.
8853 This is enabled by default with @option{-O2}.
8855 @item -mrpts=@var{count}
8859 Enable (disable) the use of the single instruction repeat instruction
8860 RPTS@. If a repeat block contains a single instruction, and the loop
8861 count can be guaranteed to be less than the value @var{count}, GCC will
8862 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8863 then a RPTS will be emitted even if the loop count cannot be determined
8864 at compile time. Note that the repeated instruction following RPTS does
8865 not have to be reloaded from memory each iteration, thus freeing up the
8866 CPU buses for operands. However, since interrupts are blocked by this
8867 instruction, it is disabled by default.
8869 @item -mloop-unsigned
8870 @itemx -mno-loop-unsigned
8871 @opindex mloop-unsigned
8872 @opindex mno-loop-unsigned
8873 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8874 is @math{2^31 + 1} since these instructions test if the iteration count is
8875 negative to terminate the loop. If the iteration count is unsigned
8876 there is a possibility than the @math{2^31 + 1} maximum iteration count may be
8877 exceeded. This switch allows an unsigned iteration count.
8881 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8882 with. This also enforces compatibility with the API employed by the TI
8883 C3x C compiler. For example, long doubles are passed as structures
8884 rather than in floating point registers.
8890 Generate code that uses registers (stack) for passing arguments to functions.
8891 By default, arguments are passed in registers where possible rather
8892 than by pushing arguments on to the stack.
8894 @item -mparallel-insns
8895 @itemx -mno-parallel-insns
8896 @opindex mparallel-insns
8897 @opindex mno-parallel-insns
8898 Allow the generation of parallel instructions. This is enabled by
8899 default with @option{-O2}.
8901 @item -mparallel-mpy
8902 @itemx -mno-parallel-mpy
8903 @opindex mparallel-mpy
8904 @opindex mno-parallel-mpy
8905 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8906 provided @option{-mparallel-insns} is also specified. These instructions have
8907 tight register constraints which can pessimize the code generation
8913 @subsection V850 Options
8914 @cindex V850 Options
8916 These @samp{-m} options are defined for V850 implementations:
8920 @itemx -mno-long-calls
8921 @opindex mlong-calls
8922 @opindex mno-long-calls
8923 Treat all calls as being far away (near). If calls are assumed to be
8924 far away, the compiler will always load the functions address up into a
8925 register, and call indirect through the pointer.
8931 Do not optimize (do optimize) basic blocks that use the same index
8932 pointer 4 or more times to copy pointer into the @code{ep} register, and
8933 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8934 option is on by default if you optimize.
8936 @item -mno-prolog-function
8937 @itemx -mprolog-function
8938 @opindex mno-prolog-function
8939 @opindex mprolog-function
8940 Do not use (do use) external functions to save and restore registers at
8941 the prolog and epilog of a function. The external functions are slower,
8942 but use less code space if more than one function saves the same number
8943 of registers. The @option{-mprolog-function} option is on by default if
8948 Try to make the code as small as possible. At present, this just turns
8949 on the @option{-mep} and @option{-mprolog-function} options.
8953 Put static or global variables whose size is @var{n} bytes or less into
8954 the tiny data area that register @code{ep} points to. The tiny data
8955 area can hold up to 256 bytes in total (128 bytes for byte references).
8959 Put static or global variables whose size is @var{n} bytes or less into
8960 the small data area that register @code{gp} points to. The small data
8961 area can hold up to 64 kilobytes.
8965 Put static or global variables whose size is @var{n} bytes or less into
8966 the first 32 kilobytes of memory.
8970 Specify that the target processor is the V850.
8973 @opindex mbig-switch
8974 Generate code suitable for big switch tables. Use this option only if
8975 the assembler/linker complain about out of range branches within a switch
8980 @subsection ARC Options
8983 These options are defined for ARC implementations:
8988 Compile code for little endian mode. This is the default.
8992 Compile code for big endian mode.
8995 @opindex mmangle-cpu
8996 Prepend the name of the cpu to all public symbol names.
8997 In multiple-processor systems, there are many ARC variants with different
8998 instruction and register set characteristics. This flag prevents code
8999 compiled for one cpu to be linked with code compiled for another.
9000 No facility exists for handling variants that are ``almost identical''.
9001 This is an all or nothing option.
9003 @item -mcpu=@var{cpu}
9005 Compile code for ARC variant @var{cpu}.
9006 Which variants are supported depend on the configuration.
9007 All variants support @option{-mcpu=base}, this is the default.
9009 @item -mtext=@var{text-section}
9010 @itemx -mdata=@var{data-section}
9011 @itemx -mrodata=@var{readonly-data-section}
9015 Put functions, data, and readonly data in @var{text-section},
9016 @var{data-section}, and @var{readonly-data-section} respectively
9017 by default. This can be overridden with the @code{section} attribute.
9018 @xref{Variable Attributes}.
9023 @subsection NS32K Options
9024 @cindex NS32K options
9026 These are the @samp{-m} options defined for the 32000 series. The default
9027 values for these options depends on which style of 32000 was selected when
9028 the compiler was configured; the defaults for the most common choices are
9036 Generate output for a 32032. This is the default
9037 when the compiler is configured for 32032 and 32016 based systems.
9043 Generate output for a 32332. This is the default
9044 when the compiler is configured for 32332-based systems.
9050 Generate output for a 32532. This is the default
9051 when the compiler is configured for 32532-based systems.
9055 Generate output containing 32081 instructions for floating point.
9056 This is the default for all systems.
9060 Generate output containing 32381 instructions for floating point. This
9061 also implies @option{-m32081}. The 32381 is only compatible with the 32332
9062 and 32532 cpus. This is the default for the pc532-netbsd configuration.
9066 Try and generate multiply-add floating point instructions @code{polyF}
9067 and @code{dotF}. This option is only available if the @option{-m32381}
9068 option is in effect. Using these instructions requires changes to
9069 register allocation which generally has a negative impact on
9070 performance. This option should only be enabled when compiling code
9071 particularly likely to make heavy use of multiply-add instructions.
9074 @opindex mnomulti-add
9075 Do not try and generate multiply-add floating point instructions
9076 @code{polyF} and @code{dotF}. This is the default on all platforms.
9079 @opindex msoft-float
9080 Generate output containing library calls for floating point.
9081 @strong{Warning:} the requisite libraries may not be available.
9084 @opindex mnobitfield
9085 Do not use the bit-field instructions. On some machines it is faster to
9086 use shifting and masking operations. This is the default for the pc532.
9090 Do use the bit-field instructions. This is the default for all platforms
9095 Use a different function-calling convention, in which functions
9096 that take a fixed number of arguments return pop their
9097 arguments on return with the @code{ret} instruction.
9099 This calling convention is incompatible with the one normally
9100 used on Unix, so you cannot use it if you need to call libraries
9101 compiled with the Unix compiler.
9103 Also, you must provide function prototypes for all functions that
9104 take variable numbers of arguments (including @code{printf});
9105 otherwise incorrect code will be generated for calls to those
9108 In addition, seriously incorrect code will result if you call a
9109 function with too many arguments. (Normally, extra arguments are
9110 harmlessly ignored.)
9112 This option takes its name from the 680x0 @code{rtd} instruction.
9117 Use a different function-calling convention where the first two arguments
9118 are passed in registers.
9120 This calling convention is incompatible with the one normally
9121 used on Unix, so you cannot use it if you need to call libraries
9122 compiled with the Unix compiler.
9125 @opindex mnoregparam
9126 Do not pass any arguments in registers. This is the default for all
9131 It is OK to use the sb as an index register which is always loaded with
9132 zero. This is the default for the pc532-netbsd target.
9136 The sb register is not available for use or has not been initialized to
9137 zero by the run time system. This is the default for all targets except
9138 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
9139 @option{-fpic} is set.
9143 Many ns32000 series addressing modes use displacements of up to 512MB@.
9144 If an address is above 512MB then displacements from zero can not be used.
9145 This option causes code to be generated which can be loaded above 512MB@.
9146 This may be useful for operating systems or ROM code.
9150 Assume code will be loaded in the first 512MB of virtual address space.
9151 This is the default for all platforms.
9157 @subsection AVR Options
9160 These options are defined for AVR implementations:
9163 @item -mmcu=@var{mcu}
9165 Specify ATMEL AVR instruction set or MCU type.
9167 Instruction set avr1 is for the minimal AVR core, not supported by the C
9168 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9169 attiny11, attiny12, attiny15, attiny28).
9171 Instruction set avr2 (default) is for the classic AVR core with up to
9172 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9173 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9174 at90c8534, at90s8535).
9176 Instruction set avr3 is for the classic AVR core with up to 128K program
9177 memory space (MCU types: atmega103, atmega603).
9179 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9180 memory space (MCU types: atmega83, atmega85).
9182 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9183 memory space (MCU types: atmega161, atmega163, atmega32, at94k).
9187 Output instruction sizes to the asm file.
9189 @item -minit-stack=@var{N}
9190 @opindex minit-stack
9191 Specify the initial stack address, which may be a symbol or numeric value,
9192 @samp{__stack} is the default.
9194 @item -mno-interrupts
9195 @opindex mno-interrupts
9196 Generated code is not compatible with hardware interrupts.
9197 Code size will be smaller.
9199 @item -mcall-prologues
9200 @opindex mcall-prologues
9201 Functions prologues/epilogues expanded as call to appropriate
9202 subroutines. Code size will be smaller.
9204 @item -mno-tablejump
9205 @opindex mno-tablejump
9206 Do not generate tablejump insns which sometimes increase code size.
9209 @opindex mtiny-stack
9210 Change only the low 8 bits of the stack pointer.
9214 @subsection MCore Options
9215 @cindex MCore options
9217 These are the @samp{-m} options defined for the Motorola M*Core
9227 @opindex mno-hardlit
9228 Inline constants into the code stream if it can be done in two
9229 instructions or less.
9237 Use the divide instruction. (Enabled by default).
9239 @item -mrelax-immediate
9240 @itemx -mrelax-immediate
9241 @itemx -mno-relax-immediate
9242 @opindex mrelax-immediate
9243 @opindex mrelax-immediate
9244 @opindex mno-relax-immediate
9245 Allow arbitrary sized immediates in bit operations.
9247 @item -mwide-bitfields
9248 @itemx -mwide-bitfields
9249 @itemx -mno-wide-bitfields
9250 @opindex mwide-bitfields
9251 @opindex mwide-bitfields
9252 @opindex mno-wide-bitfields
9253 Always treat bit-fields as int-sized.
9255 @item -m4byte-functions
9256 @itemx -m4byte-functions
9257 @itemx -mno-4byte-functions
9258 @opindex m4byte-functions
9259 @opindex m4byte-functions
9260 @opindex mno-4byte-functions
9261 Force all functions to be aligned to a four byte boundary.
9263 @item -mcallgraph-data
9264 @itemx -mcallgraph-data
9265 @itemx -mno-callgraph-data
9266 @opindex mcallgraph-data
9267 @opindex mcallgraph-data
9268 @opindex mno-callgraph-data
9269 Emit callgraph information.
9273 @itemx -mno-slow-bytes
9274 @opindex mslow-bytes
9275 @opindex mslow-bytes
9276 @opindex mno-slow-bytes
9277 Prefer word access when reading byte quantities.
9279 @item -mlittle-endian
9280 @itemx -mlittle-endian
9282 @opindex mlittle-endian
9283 @opindex mlittle-endian
9284 @opindex mbig-endian
9285 Generate code for a little endian target.
9293 Generate code for the 210 processor.
9297 @subsection IA-64 Options
9298 @cindex IA-64 Options
9300 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9304 @opindex mbig-endian
9305 Generate code for a big endian target. This is the default for HPUX@.
9307 @item -mlittle-endian
9308 @opindex mlittle-endian
9309 Generate code for a little endian target. This is the default for AIX5
9316 Generate (or don't) code for the GNU assembler. This is the default.
9317 @c Also, this is the default if the configure option @option{--with-gnu-as}
9324 Generate (or don't) code for the GNU linker. This is the default.
9325 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9330 Generate code that does not use a global pointer register. The result
9331 is not position independent code, and violates the IA-64 ABI@.
9333 @item -mvolatile-asm-stop
9334 @itemx -mno-volatile-asm-stop
9335 @opindex mvolatile-asm-stop
9336 @opindex mno-volatile-asm-stop
9337 Generate (or don't) a stop bit immediately before and after volatile asm
9342 Generate code that works around Itanium B step errata.
9344 @item -mregister-names
9345 @itemx -mno-register-names
9346 @opindex mregister-names
9347 @opindex mno-register-names
9348 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9349 the stacked registers. This may make assembler output more readable.
9355 Disable (or enable) optimizations that use the small data section. This may
9356 be useful for working around optimizer bugs.
9359 @opindex mconstant-gp
9360 Generate code that uses a single constant global pointer value. This is
9361 useful when compiling kernel code.
9365 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9366 This is useful when compiling firmware code.
9368 @item -minline-divide-min-latency
9369 @opindex minline-divide-min-latency
9370 Generate code for inline divides using the minimum latency algorithm.
9372 @item -minline-divide-max-throughput
9373 @opindex minline-divide-max-throughput
9374 Generate code for inline divides using the maximum throughput algorithm.
9376 @item -mno-dwarf2-asm
9378 @opindex mno-dwarf2-asm
9379 @opindex mdwarf2-asm
9380 Don't (or do) generate assembler code for the DWARF2 line number debugging
9381 info. This may be useful when not using the GNU assembler.
9383 @item -mfixed-range=@var{register-range}
9384 @opindex mfixed-range
9385 Generate code treating the given register range as fixed registers.
9386 A fixed register is one that the register allocator can not use. This is
9387 useful when compiling kernel code. A register range is specified as
9388 two registers separated by a dash. Multiple register ranges can be
9389 specified separated by a comma.
9393 @subsection D30V Options
9394 @cindex D30V Options
9396 These @samp{-m} options are defined for D30V implementations:
9401 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9402 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9403 memory, which starts at location @code{0x80000000}.
9407 Same as the @option{-mextmem} switch.
9411 Link the @samp{.text} section into onchip text memory, which starts at
9412 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9413 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9414 into onchip data memory, which starts at location @code{0x20000000}.
9416 @item -mno-asm-optimize
9417 @itemx -masm-optimize
9418 @opindex mno-asm-optimize
9419 @opindex masm-optimize
9420 Disable (enable) passing @option{-O} to the assembler when optimizing.
9421 The assembler uses the @option{-O} option to automatically parallelize
9422 adjacent short instructions where possible.
9424 @item -mbranch-cost=@var{n}
9425 @opindex mbranch-cost
9426 Increase the internal costs of branches to @var{n}. Higher costs means
9427 that the compiler will issue more instructions to avoid doing a branch.
9430 @item -mcond-exec=@var{n}
9432 Specify the maximum number of conditionally executed instructions that
9433 replace a branch. The default is 4.
9436 @node S/390 and zSeries Options
9437 @subsection S/390 and zSeries Options
9438 @cindex S/390 and zSeries Options
9440 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9445 @opindex mhard-float
9446 @opindex msoft-float
9447 Use (do not use) the hardware floating-point instructions and registers
9448 for floating-point operations. When @option{-msoft-float} is specified,
9449 functions in @file{libgcc.a} will be used to perform floating-point
9450 operations. When @option{-mhard-float} is specified, the compiler
9451 generates IEEE floating-point instructions. This is the default.
9454 @itemx -mno-backchain
9456 @opindex mno-backchain
9457 Generate (or do not generate) code which maintains an explicit
9458 backchain within the stack frame that points to the caller's frame.
9459 This is currently needed to allow debugging. The default is to
9460 generate the backchain.
9463 @itemx -mno-small-exec
9464 @opindex msmall-exec
9465 @opindex mno-small-exec
9466 Generate (or do not generate) code using the @code{bras} instruction
9467 to do subroutine calls.
9468 This only works reliably if the total executable size does not
9469 exceed 64k. The default is to use the @code{basr} instruction instead,
9470 which does not have this limitation.
9476 When @option{-m31} is specified, generate code compliant to the
9477 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9478 code compliant to the Linux for zSeries ABI@. This allows GCC in
9479 particular to generate 64-bit instructions. For the @samp{s390}
9480 targets, the default is @option{-m31}, while the @samp{s390x}
9481 targets default to @option{-m64}.
9487 Generate (or do not generate) code using the @code{mvcle} instruction
9488 to perform block moves. When @option{-mno-mvcle} is specifed,
9489 use a @code{mvc} loop instead. This is the default.
9495 Print (or do not print) additional debug information when compiling.
9496 The default is to not print debug information.
9501 @subsection CRIS Options
9502 @cindex CRIS Options
9504 These options are defined specifically for the CRIS ports.
9507 @item -march=@var{architecture-type}
9508 @itemx -mcpu=@var{architecture-type}
9511 Generate code for the specified architecture. The choices for
9512 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9513 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9514 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9517 @item -mtune=@var{architecture-type}
9519 Tune to @var{architecture-type} everything applicable about the generated
9520 code, except for the ABI and the set of available instructions. The
9521 choices for @var{architecture-type} are the same as for
9522 @option{-march=@var{architecture-type}}.
9524 @item -mmax-stack-frame=@var{n}
9525 @opindex mmax-stack-frame
9526 Warn when the stack frame of a function exceeds @var{n} bytes.
9528 @item -melinux-stacksize=@var{n}
9529 @opindex melinux-stacksize
9530 Only available with the @samp{cris-axis-aout} target. Arranges for
9531 indications in the program to the kernel loader that the stack of the
9532 program should be set to @var{n} bytes.
9538 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9539 @option{-march=v3} and @option{-march=v8} respectively.
9543 Enable CRIS-specific verbose debug-related information in the assembly
9544 code. This option also has the effect to turn off the @samp{#NO_APP}
9545 formatted-code indicator to the assembler at the beginning of the
9550 Do not use condition-code results from previous instruction; always emit
9551 compare and test instructions before use of condition codes.
9553 @item -mno-side-effects
9554 @opindex mno-side-effects
9555 Do not emit instructions with side-effects in addressing modes other than
9559 @itemx -mno-stack-align
9561 @itemx -mno-data-align
9562 @itemx -mconst-align
9563 @itemx -mno-const-align
9564 @opindex mstack-align
9565 @opindex mno-stack-align
9566 @opindex mdata-align
9567 @opindex mno-data-align
9568 @opindex mconst-align
9569 @opindex mno-const-align
9570 These options (no-options) arranges (eliminate arrangements) for the
9571 stack-frame, individual data and constants to be aligned for the maximum
9572 single data access size for the chosen CPU model. The default is to
9573 arrange for 32-bit alignment. ABI details such as structure layout are
9574 not affected by these options.
9582 Similar to the stack- data- and const-align options above, these options
9583 arrange for stack-frame, writable data and constants to all be 32-bit,
9584 16-bit or 8-bit aligned. The default is 32-bit alignment.
9586 @item -mno-prologue-epilogue
9587 @itemx -mprologue-epilogue
9588 @opindex mno-prologue-epilogue
9589 @opindex mprologue-epilogue
9590 With @option{-mno-prologue-epilogue}, the normal function prologue and
9591 epilogue that sets up the stack-frame are omitted and no return
9592 instructions or return sequences are generated in the code. Use this
9593 option only together with visual inspection of the compiled code: no
9594 warnings or errors are generated when call-saved registers must be saved,
9595 or storage for local variable needs to be allocated.
9601 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9602 instruction sequences that load addresses for functions from the PLT part
9603 of the GOT rather than (traditional on other architectures) calls to the
9604 PLT. The default is @option{-mgotplt}.
9608 Legacy no-op option only recognized with the cris-axis-aout target.
9612 Legacy no-op option only recognized with the cris-axis-elf and
9613 cris-axis-linux-gnu targets.
9617 Only recognized with the cris-axis-aout target, where it selects a
9618 GNU/linux-like multilib, include files and instruction set for
9623 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9627 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9628 to link with input-output functions from a simulator library. Code,
9629 initialized data and zero-initialized data are allocated consecutively.
9633 Like @option{-sim}, but pass linker options to locate initialized data at
9634 0x40000000 and zero-initialized data at 0x80000000.
9638 @subsection MMIX Options
9639 @cindex MMIX Options
9641 These options are defined for the MMIX:
9645 @itemx -mno-libfuncs
9646 Specify that intrinsic library functions are being compiled, passing all
9647 values in registers, no matter the size.
9651 Generate floating-point comparison instructions that compare with respect
9652 to the @code{rE} epsilon register.
9654 @item -mabi=mmixware
9656 Generate code that passes function parameters and return values that (in
9657 the called function) are seen as registers @code{$0} and up, as opposed to
9658 the GNU ABI which uses global registers @code{$231} and up.
9661 @item -mno-zero-extend
9662 When reading data from memory in sizes shorter than 64 bits, use (do not
9663 use) zero-extending load instructions by default, rather than
9664 sign-extending ones.
9667 @itemx -mno-knuthdiv
9668 Make the result of a division yielding a remainder have the same sign as
9669 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9670 remainder follows the sign of the dividend. Both methods are
9671 arithmetically valid, the latter being almost exclusively used.
9673 @item -mtoplevel-symbols
9674 @itemx -mno-toplevel-symbols
9675 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9676 code can be used with the @code{PREFIX} assembly directive.
9679 Generate an executable in the ELF format, rather than the default
9680 @samp{mmo} format used by the @command{mmix} simulator.
9683 @node Code Gen Options
9684 @section Options for Code Generation Conventions
9685 @cindex code generation conventions
9686 @cindex options, code generation
9687 @cindex run-time options
9689 These machine-independent options control the interface conventions
9690 used in code generation.
9692 Most of them have both positive and negative forms; the negative form
9693 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9694 one of the forms is listed---the one which is not the default. You
9695 can figure out the other form by either removing @samp{no-} or adding
9700 @opindex fexceptions
9701 Enable exception handling. Generates extra code needed to propagate
9702 exceptions. For some targets, this implies GCC will generate frame
9703 unwind information for all functions, which can produce significant data
9704 size overhead, although it does not affect execution. If you do not
9705 specify this option, GCC will enable it by default for languages like
9706 C++ which normally require exception handling, and disable it for
9707 languages like C that do not normally require it. However, you may need
9708 to enable this option when compiling C code that needs to interoperate
9709 properly with exception handlers written in C++. You may also wish to
9710 disable this option if you are compiling older C++ programs that don't
9711 use exception handling.
9713 @item -fnon-call-exceptions
9714 @opindex fnon-call-exceptions
9715 Generate code that allows trapping instructions to throw exceptions.
9716 Note that this requires platform-specific runtime support that does
9717 not exist everywhere. Moreover, it only allows @emph{trapping}
9718 instructions to throw exceptions, i.e.@: memory references or floating
9719 point instructions. It does not allow exceptions to be thrown from
9720 arbitrary signal handlers such as @code{SIGALRM}.
9722 @item -funwind-tables
9723 @opindex funwind-tables
9724 Similar to @option{-fexceptions}, except that it will just generate any needed
9725 static data, but will not affect the generated code in any other way.
9726 You will normally not enable this option; instead, a language processor
9727 that needs this handling would enable it on your behalf.
9729 @item -fasynchronous-unwind-tables
9730 @opindex funwind-tables
9731 Generate unwind table in dwarf2 format, if supported by target machine. The
9732 table is exact at each instruction boundary, so it can be used for stack
9733 unwinding from asynchronous events (such as debugger or garbage collector).
9735 @item -fpcc-struct-return
9736 @opindex fpcc-struct-return
9737 Return ``short'' @code{struct} and @code{union} values in memory like
9738 longer ones, rather than in registers. This convention is less
9739 efficient, but it has the advantage of allowing intercallability between
9740 GCC-compiled files and files compiled with other compilers.
9742 The precise convention for returning structures in memory depends
9743 on the target configuration macros.
9745 Short structures and unions are those whose size and alignment match
9746 that of some integer type.
9748 @item -freg-struct-return
9749 @opindex freg-struct-return
9750 Return @code{struct} and @code{union} values in registers when possible.
9751 This is more efficient for small structures than
9752 @option{-fpcc-struct-return}.
9754 If you specify neither @option{-fpcc-struct-return} nor
9755 @option{-freg-struct-return}, GCC defaults to whichever convention is
9756 standard for the target. If there is no standard convention, GCC
9757 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9758 the principal compiler. In those cases, we can choose the standard, and
9759 we chose the more efficient register return alternative.
9762 @opindex fshort-enums
9763 Allocate to an @code{enum} type only as many bytes as it needs for the
9764 declared range of possible values. Specifically, the @code{enum} type
9765 will be equivalent to the smallest integer type which has enough room.
9767 @item -fshort-double
9768 @opindex fshort-double
9769 Use the same size for @code{double} as for @code{float}.
9772 @opindex fshared-data
9773 Requests that the data and non-@code{const} variables of this
9774 compilation be shared data rather than private data. The distinction
9775 makes sense only on certain operating systems, where shared data is
9776 shared between processes running the same program, while private data
9777 exists in one copy per process.
9781 In C, allocate even uninitialized global variables in the data section of the
9782 object file, rather than generating them as common blocks. This has the
9783 effect that if the same variable is declared (without @code{extern}) in
9784 two different compilations, you will get an error when you link them.
9785 The only reason this might be useful is if you wish to verify that the
9786 program will work on other systems which always work this way.
9790 Ignore the @samp{#ident} directive.
9792 @item -fno-gnu-linker
9793 @opindex fno-gnu-linker
9794 Do not output global initializations (such as C++ constructors and
9795 destructors) in the form used by the GNU linker (on systems where the GNU
9796 linker is the standard method of handling them). Use this option when
9797 you want to use a non-GNU linker, which also requires using the
9798 @command{collect2} program to make sure the system linker includes
9799 constructors and destructors. (@command{collect2} is included in the GCC
9800 distribution.) For systems which @emph{must} use @command{collect2}, the
9801 compiler driver @command{gcc} is configured to do this automatically.
9803 @item -finhibit-size-directive
9804 @opindex finhibit-size-directive
9805 Don't output a @code{.size} assembler directive, or anything else that
9806 would cause trouble if the function is split in the middle, and the
9807 two halves are placed at locations far apart in memory. This option is
9808 used when compiling @file{crtstuff.c}; you should not need to use it
9812 @opindex fverbose-asm
9813 Put extra commentary information in the generated assembly code to
9814 make it more readable. This option is generally only of use to those
9815 who actually need to read the generated assembly code (perhaps while
9816 debugging the compiler itself).
9818 @option{-fno-verbose-asm}, the default, causes the
9819 extra information to be omitted and is useful when comparing two assembler
9824 Consider all memory references through pointers to be volatile.
9826 @item -fvolatile-global
9827 @opindex fvolatile-global
9828 Consider all memory references to extern and global data items to
9829 be volatile. GCC does not consider static data items to be volatile
9830 because of this switch.
9832 @item -fvolatile-static
9833 @opindex fvolatile-static
9834 Consider all memory references to static data to be volatile.
9838 @cindex global offset table
9840 Generate position-independent code (PIC) suitable for use in a shared
9841 library, if supported for the target machine. Such code accesses all
9842 constant addresses through a global offset table (GOT)@. The dynamic
9843 loader resolves the GOT entries when the program starts (the dynamic
9844 loader is not part of GCC; it is part of the operating system). If
9845 the GOT size for the linked executable exceeds a machine-specific
9846 maximum size, you get an error message from the linker indicating that
9847 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9848 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9849 on the m68k and RS/6000. The 386 has no such limit.)
9851 Position-independent code requires special support, and therefore works
9852 only on certain machines. For the 386, GCC supports PIC for System V
9853 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9854 position-independent.
9858 If supported for the target machine, emit position-independent code,
9859 suitable for dynamic linking and avoiding any limit on the size of the
9860 global offset table. This option makes a difference on the m68k, m88k,
9863 Position-independent code requires special support, and therefore works
9864 only on certain machines.
9866 @item -ffixed-@var{reg}
9868 Treat the register named @var{reg} as a fixed register; generated code
9869 should never refer to it (except perhaps as a stack pointer, frame
9870 pointer or in some other fixed role).
9872 @var{reg} must be the name of a register. The register names accepted
9873 are machine-specific and are defined in the @code{REGISTER_NAMES}
9874 macro in the machine description macro file.
9876 This flag does not have a negative form, because it specifies a
9879 @item -fcall-used-@var{reg}
9881 Treat the register named @var{reg} as an allocable register that is
9882 clobbered by function calls. It may be allocated for temporaries or
9883 variables that do not live across a call. Functions compiled this way
9884 will not save and restore the register @var{reg}.
9886 It is an error to used this flag with the frame pointer or stack pointer.
9887 Use of this flag for other registers that have fixed pervasive roles in
9888 the machine's execution model will produce disastrous results.
9890 This flag does not have a negative form, because it specifies a
9893 @item -fcall-saved-@var{reg}
9894 @opindex fcall-saved
9895 Treat the register named @var{reg} as an allocable register saved by
9896 functions. It may be allocated even for temporaries or variables that
9897 live across a call. Functions compiled this way will save and restore
9898 the register @var{reg} if they use it.
9900 It is an error to used this flag with the frame pointer or stack pointer.
9901 Use of this flag for other registers that have fixed pervasive roles in
9902 the machine's execution model will produce disastrous results.
9904 A different sort of disaster will result from the use of this flag for
9905 a register in which function values may be returned.
9907 This flag does not have a negative form, because it specifies a
9911 @opindex fpack-struct
9912 Pack all structure members together without holes. Usually you would
9913 not want to use this option, since it makes the code suboptimal, and
9914 the offsets of structure members won't agree with system libraries.
9916 @item -finstrument-functions
9917 @opindex finstrument-functions
9918 Generate instrumentation calls for entry and exit to functions. Just
9919 after function entry and just before function exit, the following
9920 profiling functions will be called with the address of the current
9921 function and its call site. (On some platforms,
9922 @code{__builtin_return_address} does not work beyond the current
9923 function, so the call site information may not be available to the
9924 profiling functions otherwise.)
9927 void __cyg_profile_func_enter (void *this_fn,
9929 void __cyg_profile_func_exit (void *this_fn,
9933 The first argument is the address of the start of the current function,
9934 which may be looked up exactly in the symbol table.
9936 This instrumentation is also done for functions expanded inline in other
9937 functions. The profiling calls will indicate where, conceptually, the
9938 inline function is entered and exited. This means that addressable
9939 versions of such functions must be available. If all your uses of a
9940 function are expanded inline, this may mean an additional expansion of
9941 code size. If you use @samp{extern inline} in your C code, an
9942 addressable version of such functions must be provided. (This is
9943 normally the case anyways, but if you get lucky and the optimizer always
9944 expands the functions inline, you might have gotten away without
9945 providing static copies.)
9947 A function may be given the attribute @code{no_instrument_function}, in
9948 which case this instrumentation will not be done. This can be used, for
9949 example, for the profiling functions listed above, high-priority
9950 interrupt routines, and any functions from which the profiling functions
9951 cannot safely be called (perhaps signal handlers, if the profiling
9952 routines generate output or allocate memory).
9955 @opindex fstack-check
9956 Generate code to verify that you do not go beyond the boundary of the
9957 stack. You should specify this flag if you are running in an
9958 environment with multiple threads, but only rarely need to specify it in
9959 a single-threaded environment since stack overflow is automatically
9960 detected on nearly all systems if there is only one stack.
9962 Note that this switch does not actually cause checking to be done; the
9963 operating system must do that. The switch causes generation of code
9964 to ensure that the operating system sees the stack being extended.
9966 @item -fstack-limit-register=@var{reg}
9967 @itemx -fstack-limit-symbol=@var{sym}
9968 @itemx -fno-stack-limit
9969 @opindex fstack-limit-register
9970 @opindex fstack-limit-symbol
9971 @opindex fno-stack-limit
9972 Generate code to ensure that the stack does not grow beyond a certain value,
9973 either the value of a register or the address of a symbol. If the stack
9974 would grow beyond the value, a signal is raised. For most targets,
9975 the signal is raised before the stack overruns the boundary, so
9976 it is possible to catch the signal without taking special precautions.
9978 For instance, if the stack starts at absolute address @samp{0x80000000}
9979 and grows downwards, you can use the flags
9980 @option{-fstack-limit-symbol=__stack_limit} and
9981 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
9982 of 128KB@. Note that this may only work with the GNU linker.
9984 @cindex aliasing of parameters
9985 @cindex parameters, aliased
9986 @item -fargument-alias
9987 @itemx -fargument-noalias
9988 @itemx -fargument-noalias-global
9989 @opindex fargument-alias
9990 @opindex fargument-noalias
9991 @opindex fargument-noalias-global
9992 Specify the possible relationships among parameters and between
9993 parameters and global data.
9995 @option{-fargument-alias} specifies that arguments (parameters) may
9996 alias each other and may alias global storage.@*
9997 @option{-fargument-noalias} specifies that arguments do not alias
9998 each other, but may alias global storage.@*
9999 @option{-fargument-noalias-global} specifies that arguments do not
10000 alias each other and do not alias global storage.
10002 Each language will automatically use whatever option is required by
10003 the language standard. You should not need to use these options yourself.
10005 @item -fleading-underscore
10006 @opindex fleading-underscore
10007 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
10008 change the way C symbols are represented in the object file. One use
10009 is to help link with legacy assembly code.
10011 Be warned that you should know what you are doing when invoking this
10012 option, and that not all targets provide complete support for it.
10017 @node Environment Variables
10018 @section Environment Variables Affecting GCC
10019 @cindex environment variables
10021 @c man begin ENVIRONMENT
10023 This section describes several environment variables that affect how GCC
10024 operates. Some of them work by specifying directories or prefixes to use
10025 when searching for various kinds of files. Some are used to specify other
10026 aspects of the compilation environment.
10029 Note that you can also specify places to search using options such as
10030 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10031 take precedence over places specified using environment variables, which
10032 in turn take precedence over those specified by the configuration of GCC@.
10036 Note that you can also specify places to search using options such as
10037 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10038 take precedence over places specified using environment variables, which
10039 in turn take precedence over those specified by the configuration of GCC@.
10046 @c @itemx LC_COLLATE
10048 @c @itemx LC_MONETARY
10049 @c @itemx LC_NUMERIC
10054 @c @findex LC_COLLATE
10055 @findex LC_MESSAGES
10056 @c @findex LC_MONETARY
10057 @c @findex LC_NUMERIC
10061 These environment variables control the way that GCC uses
10062 localization information that allow GCC to work with different
10063 national conventions. GCC inspects the locale categories
10064 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
10065 so. These locale categories can be set to any value supported by your
10066 installation. A typical value is @samp{en_UK} for English in the United
10069 The @env{LC_CTYPE} environment variable specifies character
10070 classification. GCC uses it to determine the character boundaries in
10071 a string; this is needed for some multibyte encodings that contain quote
10072 and escape characters that would otherwise be interpreted as a string
10075 The @env{LC_MESSAGES} environment variable specifies the language to
10076 use in diagnostic messages.
10078 If the @env{LC_ALL} environment variable is set, it overrides the value
10079 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10080 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10081 environment variable. If none of these variables are set, GCC
10082 defaults to traditional C English behavior.
10086 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10087 files. GCC uses temporary files to hold the output of one stage of
10088 compilation which is to be used as input to the next stage: for example,
10089 the output of the preprocessor, which is the input to the compiler
10092 @item GCC_EXEC_PREFIX
10093 @findex GCC_EXEC_PREFIX
10094 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10095 names of the subprograms executed by the compiler. No slash is added
10096 when this prefix is combined with the name of a subprogram, but you can
10097 specify a prefix that ends with a slash if you wish.
10099 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10100 an appropriate prefix to use based on the pathname it was invoked with.
10102 If GCC cannot find the subprogram using the specified prefix, it
10103 tries looking in the usual places for the subprogram.
10105 The default value of @env{GCC_EXEC_PREFIX} is
10106 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10107 of @code{prefix} when you ran the @file{configure} script.
10109 Other prefixes specified with @option{-B} take precedence over this prefix.
10111 This prefix is also used for finding files such as @file{crt0.o} that are
10114 In addition, the prefix is used in an unusual way in finding the
10115 directories to search for header files. For each of the standard
10116 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10117 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10118 replacing that beginning with the specified prefix to produce an
10119 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10120 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10121 These alternate directories are searched first; the standard directories
10124 @item COMPILER_PATH
10125 @findex COMPILER_PATH
10126 The value of @env{COMPILER_PATH} is a colon-separated list of
10127 directories, much like @env{PATH}. GCC tries the directories thus
10128 specified when searching for subprograms, if it can't find the
10129 subprograms using @env{GCC_EXEC_PREFIX}.
10132 @findex LIBRARY_PATH
10133 The value of @env{LIBRARY_PATH} is a colon-separated list of
10134 directories, much like @env{PATH}. When configured as a native compiler,
10135 GCC tries the directories thus specified when searching for special
10136 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10137 using GCC also uses these directories when searching for ordinary
10138 libraries for the @option{-l} option (but directories specified with
10139 @option{-L} come first).
10141 @item C_INCLUDE_PATH
10142 @itemx CPLUS_INCLUDE_PATH
10143 @itemx OBJC_INCLUDE_PATH
10144 @findex C_INCLUDE_PATH
10145 @findex CPLUS_INCLUDE_PATH
10146 @findex OBJC_INCLUDE_PATH
10147 @c @itemx OBJCPLUS_INCLUDE_PATH
10148 These environment variables pertain to particular languages. Each
10149 variable's value is a colon-separated list of directories, much like
10150 @env{PATH}. When GCC searches for header files, it tries the
10151 directories listed in the variable for the language you are using, after
10152 the directories specified with @option{-I} but before the standard header
10155 @item DEPENDENCIES_OUTPUT
10156 @findex DEPENDENCIES_OUTPUT
10157 @cindex dependencies for make as output
10158 If this variable is set, its value specifies how to output dependencies
10159 for Make based on the header files processed by the compiler. This
10160 output looks much like the output from the @option{-M} option
10161 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
10162 in addition to the usual results of compilation.
10164 The value of @env{DEPENDENCIES_OUTPUT} can be just a file name, in
10165 which case the Make rules are written to that file, guessing the target
10166 name from the source file name. Or the value can have the form
10167 @samp{@var{file} @var{target}}, in which case the rules are written to
10168 file @var{file} using @var{target} as the target name.
10172 @cindex locale definition
10173 This variable is used to pass locale information to the compiler. One way in
10174 which this information is used is to determine the character set to be used
10175 when character literals, string literals and comments are parsed in C and C++.
10176 When the compiler is configured to allow multibyte characters,
10177 the following values for @env{LANG} are recognized:
10181 Recognize JIS characters.
10183 Recognize SJIS characters.
10185 Recognize EUCJP characters.
10188 If @env{LANG} is not defined, or if it has some other value, then the
10189 compiler will use mblen and mbtowc as defined by the default locale to
10190 recognize and translate multibyte characters.
10195 @node Running Protoize
10196 @section Running Protoize
10198 The program @code{protoize} is an optional part of GCC@. You can use
10199 it to add prototypes to a program, thus converting the program to ISO
10200 C in one respect. The companion program @code{unprotoize} does the
10201 reverse: it removes argument types from any prototypes that are found.
10203 When you run these programs, you must specify a set of source files as
10204 command line arguments. The conversion programs start out by compiling
10205 these files to see what functions they define. The information gathered
10206 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10208 After scanning comes actual conversion. The specified files are all
10209 eligible to be converted; any files they include (whether sources or
10210 just headers) are eligible as well.
10212 But not all the eligible files are converted. By default,
10213 @code{protoize} and @code{unprotoize} convert only source and header
10214 files in the current directory. You can specify additional directories
10215 whose files should be converted with the @option{-d @var{directory}}
10216 option. You can also specify particular files to exclude with the
10217 @option{-x @var{file}} option. A file is converted if it is eligible, its
10218 directory name matches one of the specified directory names, and its
10219 name within the directory has not been excluded.
10221 Basic conversion with @code{protoize} consists of rewriting most
10222 function definitions and function declarations to specify the types of
10223 the arguments. The only ones not rewritten are those for varargs
10226 @code{protoize} optionally inserts prototype declarations at the
10227 beginning of the source file, to make them available for any calls that
10228 precede the function's definition. Or it can insert prototype
10229 declarations with block scope in the blocks where undeclared functions
10232 Basic conversion with @code{unprotoize} consists of rewriting most
10233 function declarations to remove any argument types, and rewriting
10234 function definitions to the old-style pre-ISO form.
10236 Both conversion programs print a warning for any function declaration or
10237 definition that they can't convert. You can suppress these warnings
10240 The output from @code{protoize} or @code{unprotoize} replaces the
10241 original source file. The original file is renamed to a name ending
10242 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10243 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10244 for DOS) file already exists, then the source file is simply discarded.
10246 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10247 scan the program and collect information about the functions it uses.
10248 So neither of these programs will work until GCC is installed.
10250 Here is a table of the options you can use with @code{protoize} and
10251 @code{unprotoize}. Each option works with both programs unless
10255 @item -B @var{directory}
10256 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10257 usual directory (normally @file{/usr/local/lib}). This file contains
10258 prototype information about standard system functions. This option
10259 applies only to @code{protoize}.
10261 @item -c @var{compilation-options}
10262 Use @var{compilation-options} as the options when running @code{gcc} to
10263 produce the @samp{.X} files. The special option @option{-aux-info} is
10264 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10266 Note that the compilation options must be given as a single argument to
10267 @code{protoize} or @code{unprotoize}. If you want to specify several
10268 @code{gcc} options, you must quote the entire set of compilation options
10269 to make them a single word in the shell.
10271 There are certain @code{gcc} arguments that you cannot use, because they
10272 would produce the wrong kind of output. These include @option{-g},
10273 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10274 the @var{compilation-options}, they are ignored.
10277 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10278 systems) instead of @samp{.c}. This is convenient if you are converting
10279 a C program to C++. This option applies only to @code{protoize}.
10282 Add explicit global declarations. This means inserting explicit
10283 declarations at the beginning of each source file for each function
10284 that is called in the file and was not declared. These declarations
10285 precede the first function definition that contains a call to an
10286 undeclared function. This option applies only to @code{protoize}.
10288 @item -i @var{string}
10289 Indent old-style parameter declarations with the string @var{string}.
10290 This option applies only to @code{protoize}.
10292 @code{unprotoize} converts prototyped function definitions to old-style
10293 function definitions, where the arguments are declared between the
10294 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10295 uses five spaces as the indentation. If you want to indent with just
10296 one space instead, use @option{-i " "}.
10299 Keep the @samp{.X} files. Normally, they are deleted after conversion
10303 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10304 a prototype declaration for each function in each block which calls the
10305 function without any declaration. This option applies only to
10309 Make no real changes. This mode just prints information about the conversions
10310 that would have been done without @option{-n}.
10313 Make no @samp{.save} files. The original files are simply deleted.
10314 Use this option with caution.
10316 @item -p @var{program}
10317 Use the program @var{program} as the compiler. Normally, the name
10318 @file{gcc} is used.
10321 Work quietly. Most warnings are suppressed.
10324 Print the version number, just like @option{-v} for @code{gcc}.
10327 If you need special compiler options to compile one of your program's
10328 source files, then you should generate that file's @samp{.X} file
10329 specially, by running @code{gcc} on that source file with the
10330 appropriate options and the option @option{-aux-info}. Then run
10331 @code{protoize} on the entire set of files. @code{protoize} will use
10332 the existing @samp{.X} file because it is newer than the source file.
10336 gcc -Dfoo=bar file1.c -aux-info file1.X
10341 You need to include the special files along with the rest in the
10342 @code{protoize} command, even though their @samp{.X} files already
10343 exist, because otherwise they won't get converted.
10345 @xref{Protoize Caveats}, for more information on how to use
10346 @code{protoize} successfully.