1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002 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, 2002 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 @command{gcc}, or
57 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
58 for contributors to GCC@.
63 @chapter GCC Command Options
64 @cindex GCC command options
65 @cindex command options
66 @cindex options, GCC command
68 @c man begin DESCRIPTION
70 When you invoke GCC, it normally does preprocessing, compilation,
71 assembly and linking. The ``overall options'' allow you to stop this
72 process at an intermediate stage. For example, the @option{-c} option
73 says not to run the linker. Then the output consists of object files
74 output by the assembler.
76 Other options are passed on to one stage of processing. Some options
77 control the preprocessor and others the compiler itself. Yet other
78 options control the assembler and linker; most of these are not
79 documented here, since you rarely need to use any of them.
81 @cindex C compilation options
82 Most of the command line options that you can use with GCC are useful
83 for C programs; when an option is only useful with another language
84 (usually C++), the explanation says so explicitly. If the description
85 for a particular option does not mention a source language, you can use
86 that option with all supported languages.
88 @cindex C++ compilation options
89 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
90 options for compiling C++ programs.
92 @cindex grouping options
93 @cindex options, grouping
94 The @command{gcc} program accepts options and file names as operands. Many
95 options have multi-letter names; therefore multiple single-letter options
96 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
99 @cindex order of options
100 @cindex options, order
101 You can mix options and other arguments. For the most part, the order
102 you use doesn't matter. Order does matter when you use several options
103 of the same kind; for example, if you specify @option{-L} more than once,
104 the directories are searched in the order specified.
106 Many options have long names starting with @samp{-f} or with
107 @samp{-W}---for example, @option{-fforce-mem},
108 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
109 these have both positive and negative forms; the negative form of
110 @option{-ffoo} would be @option{-fno-foo}. This manual documents
111 only one of these two forms, whichever one is not the default.
115 @xref{Option Index}, for an index to GCC's options.
118 * Option Summary:: Brief list of all options, without explanations.
119 * Overall Options:: Controlling the kind of output:
120 an executable, object files, assembler files,
121 or preprocessed source.
122 * Invoking G++:: Compiling C++ programs.
123 * C Dialect Options:: Controlling the variant of C language compiled.
124 * C++ Dialect Options:: Variations on C++.
125 * Objective-C Dialect Options:: Variations on Objective-C.
126 * Language Independent Options:: Controlling how diagnostics should be
128 * Warning Options:: How picky should the compiler be?
129 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
130 * Optimize Options:: How much optimization?
131 * Preprocessor Options:: Controlling header files and macro definitions.
132 Also, getting dependency information for Make.
133 * Assembler Options:: Passing options to the assembler.
134 * Link Options:: Specifying libraries and so on.
135 * Directory Options:: Where to find header files and libraries.
136 Where to find the compiler executable files.
137 * Spec Files:: How to pass switches to sub-processes.
138 * Target Options:: Running a cross-compiler, or an old version of GCC.
139 * Submodel Options:: Specifying minor hardware or convention variations,
140 such as 68010 vs 68020.
141 * Code Gen Options:: Specifying conventions for function calls, data layout
143 * Environment Variables:: Env vars that affect GCC.
144 * Running Protoize:: Automatically adding or removing function prototypes.
150 @section Option Summary
152 Here is a summary of all the options, grouped by type. Explanations are
153 in the following sections.
156 @item Overall Options
157 @xref{Overall Options,,Options Controlling the Kind of Output}.
159 -c -S -E -o @var{file} -pipe -pass-exit-codes -x @var{language} @gol
160 -v --target-help --help}
162 @item C Language Options
163 @xref{C Dialect Options,,Options Controlling C Dialect}.
165 -ansi -std=@var{standard} -aux-info @var{filename} @gol
166 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
167 -fhosted -ffreestanding @gol
168 -trigraphs -traditional -traditional-cpp @gol
169 -fallow-single-precision -fcond-mismatch @gol
170 -fsigned-bitfields -fsigned-char @gol
171 -funsigned-bitfields -funsigned-char @gol
172 -fwritable-strings -fshort-wchar}
174 @item C++ Language Options
175 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
177 -fno-access-control -fcheck-new -fconserve-space @gol
178 -fno-const-strings -fdollars-in-identifiers @gol
179 -fno-elide-constructors @gol
180 -fno-enforce-eh-specs -fexternal-templates @gol
181 -falt-external-templates @gol
182 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
183 -fno-implicit-templates @gol
184 -fno-implicit-inline-templates @gol
185 -fno-implement-inlines -fms-extensions @gol
186 -fno-nonansi-builtins -fno-operator-names @gol
187 -fno-optional-diags -fpermissive @gol
188 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
189 -fuse-cxa-atexit -fvtable-gc -fno-weak -nostdinc++ @gol
190 -fno-default-inline -Wctor-dtor-privacy @gol
191 -Wnon-virtual-dtor -Wreorder @gol
192 -Weffc++ -Wno-deprecated @gol
193 -Wno-non-template-friend -Wold-style-cast @gol
194 -Woverloaded-virtual -Wno-pmf-conversions @gol
195 -Wsign-promo -Wsynth}
197 @item Objective-C Language Options
198 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
200 -fconstant-string-class=@var{class-name} @gol
201 -fgnu-runtime -fnext-runtime -gen-decls @gol
202 -Wno-protocol -Wselector}
204 @item Language Independent Options
205 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
207 -fmessage-length=@var{n} @gol
208 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
210 @item Warning Options
211 @xref{Warning Options,,Options to Request or Suppress Warnings}.
213 -fsyntax-only -pedantic -pedantic-errors @gol
214 -w -W -Wall -Waggregate-return @gol
215 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
216 -Wconversion -Wno-deprecated-declarations @gol
217 -Wdisabled-optimization -Wdiv-by-zero -Werror @gol
218 -Wfloat-equal -Wformat -Wformat=2 @gol
219 -Wformat-nonliteral -Wformat-security @gol
220 -Wimplicit -Wimplicit-int @gol
221 -Wimplicit-function-declaration @gol
222 -Werror-implicit-function-declaration @gol
223 -Wimport -Winline @gol
224 -Wlarger-than-@var{len} -Wlong-long @gol
225 -Wmain -Wmissing-braces -Wmissing-declarations @gol
226 -Wmissing-format-attribute -Wmissing-noreturn @gol
227 -Wmultichar -Wno-format-extra-args -Wno-format-y2k @gol
228 -Wno-import -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wswitch -Wsystem-headers @gol
232 -Wtrigraphs -Wundef -Wuninitialized @gol
233 -Wunknown-pragmas -Wunreachable-code @gol
234 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
235 -Wunused-value -Wunused-variable -Wwrite-strings}
237 @item C-only Warning Options
239 -Wbad-function-cast -Wmissing-prototypes -Wnested-externs @gol
240 -Wstrict-prototypes -Wtraditional}
242 @item Debugging Options
243 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
245 -d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
247 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
248 -fdump-tree-original@r{[}-@var{n}@r{]} -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
250 -fmem-report -fpretend-float @gol
251 -fprofile-arcs -ftest-coverage -ftime-report @gol
252 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
253 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
254 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
255 -print-multi-directory -print-multi-lib @gol
256 -print-prog-name=@var{program} -print-search-dirs -Q @gol
259 @item Optimization Options
260 @xref{Optimize Options,,Options that Control Optimization}.
262 -falign-functions=@var{n} -falign-jumps=@var{n} @gol
263 -falign-labels=@var{n} -falign-loops=@var{n} @gol
264 -fbranch-probabilities -fcaller-saves -fcprop-registers @gol
265 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
266 -fdelayed-branch -fdelete-null-pointer-checks @gol
267 -fexpensive-optimizations -ffast-math -ffloat-store @gol
268 -fforce-addr -fforce-mem -ffunction-sections @gol
269 -fgcse -fgcse-lm -fgcse-sm @gol
270 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
271 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
272 -fmove-all-movables -fno-default-inline -fno-defer-pop @gol
273 -fno-function-cse -fno-guess-branch-probability @gol
274 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
275 -funsafe-math-optimizations -fno-trapping-math @gol
276 -fomit-frame-pointer -foptimize-register-move @gol
277 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
278 -freduce-all-givs -fregmove -frename-registers @gol
279 -frerun-cse-after-loop -frerun-loop-opt @gol
280 -fschedule-insns -fschedule-insns2 @gol
281 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
282 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
283 -funroll-all-loops -funroll-loops @gol
284 --param @var{name}=@var{value}
285 -O -O0 -O1 -O2 -O3 -Os}
287 @item Preprocessor Options
288 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
290 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
291 -C -dD -dI -dM -dN @gol
292 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
293 -idirafter @var{dir} @gol
294 -include @var{file} -imacros @var{file} @gol
295 -iprefix @var{file} -iwithprefix @var{dir} @gol
296 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
297 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
298 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
300 @item Assembler Option
301 @xref{Assembler Options,,Passing Options to the Assembler}.
306 @xref{Link Options,,Options for Linking}.
308 @var{object-file-name} -l@var{library} @gol
309 -nostartfiles -nodefaultlibs -nostdlib @gol
310 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
311 -Wl,@var{option} -Xlinker @var{option} @gol
314 @item Directory Options
315 @xref{Directory Options,,Options for Directory Search}.
317 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
320 @c I wrote this xref this way to avoid overfull hbox. -- rms
321 @xref{Target Options}.
323 -b @var{machine} -V @var{version}}
325 @item Machine Dependent Options
326 @xref{Submodel Options,,Hardware Models and Configurations}.
328 @emph{M680x0 Options}
330 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
331 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
332 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
333 -malign-int -mstrict-align}
335 @emph{M68hc1x Options}
337 -m6811 -m6812 -m68hc11 -m68hc12 @gol
338 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
346 -mcpu=@var{cpu-type} @gol
347 -mtune=@var{cpu-type} @gol
348 -mcmodel=@var{code-model} @gol
350 -mapp-regs -mbroken-saverestore -mcypress @gol
351 -mepilogue -mfaster-structs -mflat @gol
352 -mfpu -mhard-float -mhard-quad-float @gol
353 -mimpure-text -mlive-g0 -mno-app-regs @gol
354 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu @gol
355 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
356 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
357 -msupersparc -munaligned-doubles -mv8}
359 @emph{Convex Options}
361 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
362 -margcount -mnoargcount @gol
363 -mlong32 -mlong64 @gol
364 -mvolatile-cache -mvolatile-nocache}
366 @emph{AMD29K Options}
368 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
369 -mlarge -mnormal -msmall @gol
370 -mkernel-registers -mno-reuse-arg-regs @gol
371 -mno-stack-check -mno-storem-bug @gol
372 -mreuse-arg-regs -msoft-float -mstack-check @gol
373 -mstorem-bug -muser-registers}
377 -mapcs-frame -mno-apcs-frame @gol
378 -mapcs-26 -mapcs-32 @gol
379 -mapcs-stack-check -mno-apcs-stack-check @gol
380 -mapcs-float -mno-apcs-float @gol
381 -mapcs-reentrant -mno-apcs-reentrant @gol
382 -msched-prolog -mno-sched-prolog @gol
383 -mlittle-endian -mbig-endian -mwords-little-endian @gol
384 -malignment-traps -mno-alignment-traps @gol
385 -msoft-float -mhard-float -mfpe @gol
386 -mthumb-interwork -mno-thumb-interwork @gol
387 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
388 -mstructure-size-boundary=@var{n} @gol
389 -mbsd -mxopen -mno-symrename @gol
390 -mabort-on-noreturn @gol
391 -mlong-calls -mno-long-calls @gol
392 -msingle-pic-base -mno-single-pic-base @gol
393 -mpic-register=@var{reg} @gol
394 -mnop-fun-dllimport @gol
395 -mpoke-function-name @gol
397 -mtpcs-frame -mtpcs-leaf-frame @gol
398 -mcaller-super-interworking -mcallee-super-interworking }
400 @emph{MN10200 Options}
404 @emph{MN10300 Options}
406 -mmult-bug -mno-mult-bug @gol
407 -mam33 -mno-am33 @gol
410 @emph{M32R/D Options}
412 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
417 -m88000 -m88100 -m88110 -mbig-pic @gol
418 -mcheck-zero-division -mhandle-large-shift @gol
419 -midentify-revision -mno-check-zero-division @gol
420 -mno-ocs-debug-info -mno-ocs-frame-position @gol
421 -mno-optimize-arg-area -mno-serialize-volatile @gol
422 -mno-underscores -mocs-debug-info @gol
423 -mocs-frame-position -moptimize-arg-area @gol
424 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
425 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
426 -mversion-03.00 -mwarn-passed-structs}
428 @emph{RS/6000 and PowerPC Options}
430 -mcpu=@var{cpu-type} @gol
431 -mtune=@var{cpu-type} @gol
432 -mpower -mno-power -mpower2 -mno-power2 @gol
433 -mpowerpc -mpowerpc64 -mno-powerpc @gol
434 -maltivec -mno-altivec @gol
435 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
436 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
437 -mnew-mnemonics -mold-mnemonics @gol
438 -mfull-toc -mminimal-toc -mno-fop-in-toc -mno-sum-in-toc @gol
439 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
440 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
441 -mstring -mno-string -mupdate -mno-update @gol
442 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
443 -mstrict-align -mno-strict-align -mrelocatable @gol
444 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
445 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
446 -mcall-aix -mcall-sysv -mcall-netbsd @gol
447 -maix-struct-return -msvr4-struct-return
449 -mprototype -mno-prototype @gol
450 -msim -mmvme -mads -myellowknife -memb -msdata @gol
451 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
455 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
456 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
457 -mminimum-fp-blocks -mnohc-struct-return}
461 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
462 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
463 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
464 -mgas -mgp32 -mgp64 @gol
465 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
466 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
467 -mmips-as -mmips-tfile -mno-abicalls @gol
468 -mno-embedded-data -mno-uninit-const-in-rodata @gol
469 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
470 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
471 -mrnames -msoft-float @gol
472 -m4650 -msingle-float -mmad @gol
473 -mstats -EL -EB -G @var{num} -nocpp @gol
474 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
475 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
477 @emph{i386 and x86-64 Options}
479 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
480 -masm=@var{dialect} -mno-fancy-math-387 @gol
481 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
482 -mno-wide-multiply -mrtd -malign-double @gol
483 -mpreferred-stack-boundary=@var{num} @gol
484 -mmmx -msse -msse2 -msse-math -m3dnow @gol
485 -mthreads -mno-align-stringops -minline-all-stringops @gol
486 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
487 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
493 -march=@var{architecture-type} @gol
494 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
495 -mfast-indirect-calls -mgas -mjump-in-delay @gol
496 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
497 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
498 -mno-jump-in-delay -mno-long-load-store @gol
499 -mno-portable-runtime -mno-soft-float @gol
500 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
501 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
502 -mschedule=@var{cpu-type} -mspace-regs}
504 @emph{Intel 960 Options}
506 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
507 -mcode-align -mcomplex-addr -mleaf-procedures @gol
508 -mic-compat -mic2.0-compat -mic3.0-compat @gol
509 -mintel-asm -mno-clean-linkage -mno-code-align @gol
510 -mno-complex-addr -mno-leaf-procedures @gol
511 -mno-old-align -mno-strict-align -mno-tail-call @gol
512 -mnumerics -mold-align -msoft-float -mstrict-align @gol
515 @emph{DEC Alpha Options}
517 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
518 -mieee -mieee-with-inexact -mieee-conformant @gol
519 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
520 -mtrap-precision=@var{mode} -mbuild-constants @gol
521 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
522 -mbwx -mmax -mfix -mcix @gol
523 -mfloat-vax -mfloat-ieee @gol
524 -mexplicit-relocs -msmall-data -mlarge-data @gol
525 -mmemory-latency=@var{time}}
527 @emph{DEC Alpha/VMS Options}
531 @emph{Clipper Options}
535 @emph{H8/300 Options}
537 -mrelax -mh -ms -mint32 -malign-300}
541 -m1 -m2 -m3 -m3e @gol
542 -m4-nofpu -m4-single-only -m4-single -m4 @gol
543 -mb -ml -mdalign -mrelax @gol
544 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
545 -mieee -misize -mpadstruct -mspace @gol
546 -mprefergot -musermode}
548 @emph{System V Options}
550 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
555 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
556 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
558 @emph{TMS320C3x/C4x Options}
560 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
561 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
562 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
563 -mparallel-insns -mparallel-mpy -mpreserve-float}
567 -mlong-calls -mno-long-calls -mep -mno-ep @gol
568 -mprolog-function -mno-prolog-function -mspace @gol
569 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
574 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
575 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
576 -mregparam -mnoregparam -msb -mnosb @gol
577 -mbitfield -mnobitfield -mhimem -mnohimem}
581 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
582 -mcall-prologues -mno-tablejump -mtiny-stack}
586 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
587 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
588 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
589 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
590 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
594 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
595 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
596 -melf -mbranch-predict -mreg-stack-fill-bug-workaround @gol
597 -mno-branch-predict -mno-reg-stack-fill-bug-workaround}
601 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
602 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
603 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
604 -minline-divide-max-throughput -mno-dwarf2-asm @gol
605 -mfixed-range=@var{register-range}}
609 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
610 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
612 @emph{S/390 and zSeries Options}
614 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
615 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
616 -m64 -m31 -mdebug -mno-debug}
620 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
621 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
622 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
623 -mstack-align -mdata-align -mconst-align @gol
624 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
625 -melf -maout -melinux -mlinux -sim -sim2}
627 @emph{PDP-11 Options}
629 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
630 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
631 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
632 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
633 -mbranch-expensive -mbranch-cheap @gol
634 -msplit -mno-split -munix-asm -mdec-asm}
636 @emph{Xstormy16 Options}
640 @item Code Generation Options
641 @xref{Code Gen Options,,Options for Code Generation Conventions}.
643 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
644 -ffixed-@var{reg} -fexceptions @gol
645 -fnon-call-exceptions -funwind-tables @gol
646 -fasynchronous-unwind-tables @gol
647 -finhibit-size-directive -finstrument-functions @gol
648 -fno-common -fno-ident -fno-gnu-linker @gol
649 -fpcc-struct-return -fpic -fPIC @gol
650 -freg-struct-return -fshared-data -fshort-enums @gol
651 -fshort-double -fvolatile @gol
652 -fvolatile-global -fvolatile-static @gol
653 -fverbose-asm -fpack-struct -fstack-check @gol
654 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
655 -fargument-alias -fargument-noalias @gol
656 -fargument-noalias-global -fleading-underscore}
660 * Overall Options:: Controlling the kind of output:
661 an executable, object files, assembler files,
662 or preprocessed source.
663 * C Dialect Options:: Controlling the variant of C language compiled.
664 * C++ Dialect Options:: Variations on C++.
665 * Objective-C Dialect Options:: Variations on Objective-C.
666 * Language Independent Options:: Controlling how diagnostics should be
668 * Warning Options:: How picky should the compiler be?
669 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
670 * Optimize Options:: How much optimization?
671 * Preprocessor Options:: Controlling header files and macro definitions.
672 Also, getting dependency information for Make.
673 * Assembler Options:: Passing options to the assembler.
674 * Link Options:: Specifying libraries and so on.
675 * Directory Options:: Where to find header files and libraries.
676 Where to find the compiler executable files.
677 * Spec Files:: How to pass switches to sub-processes.
678 * Target Options:: Running a cross-compiler, or an old version of GCC.
681 @node Overall Options
682 @section Options Controlling the Kind of Output
684 Compilation can involve up to four stages: preprocessing, compilation
685 proper, assembly and linking, always in that order. The first three
686 stages apply to an individual source file, and end by producing an
687 object file; linking combines all the object files (those newly
688 compiled, and those specified as input) into an executable file.
690 @cindex file name suffix
691 For any given input file, the file name suffix determines what kind of
696 C source code which must be preprocessed.
699 C source code which should not be preprocessed.
702 C++ source code which should not be preprocessed.
705 Objective-C source code. Note that you must link with the library
706 @file{libobjc.a} to make an Objective-C program work.
709 Objective-C source code which should not be preprocessed.
712 C header file (not to be compiled or linked).
716 @itemx @var{file}.cxx
717 @itemx @var{file}.cpp
718 @itemx @var{file}.c++
720 C++ source code which must be preprocessed. Note that in @samp{.cxx},
721 the last two letters must both be literally @samp{x}. Likewise,
722 @samp{.C} refers to a literal capital C@.
725 @itemx @var{file}.for
726 @itemx @var{file}.FOR
727 Fortran source code which should not be preprocessed.
730 @itemx @var{file}.fpp
731 @itemx @var{file}.FPP
732 Fortran source code which must be preprocessed (with the traditional
736 Fortran source code which must be preprocessed with a RATFOR
737 preprocessor (not included with GCC)@.
739 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
740 Using and Porting GNU Fortran}, for more details of the handling of
743 @c FIXME: Descriptions of Java file types.
750 Ada source code file which contains a library unit declaration (a
751 declaration of a package, subprogram, or generic, or a generic
752 instantiation), or a library unit renaming declaration (a package,
753 generic, or subprogram renaming declaration). Such files are also
756 @itemx @var{file}.adb
757 Ada source code file containing a library unit body (a subprogram or
758 package body). Such files are also called @dfn{bodies}.
760 @c GCC also knows about some suffixes for languages not yet included:
766 @itemx @var{file}.chi
767 CHILL source code (preprocessed with the traditional preprocessor).
773 Assembler code which must be preprocessed.
776 An object file to be fed straight into linking.
777 Any file name with no recognized suffix is treated this way.
781 You can specify the input language explicitly with the @option{-x} option:
784 @item -x @var{language}
785 Specify explicitly the @var{language} for the following input files
786 (rather than letting the compiler choose a default based on the file
787 name suffix). This option applies to all following input files until
788 the next @option{-x} option. Possible values for @var{language} are:
790 c c-header cpp-output
792 objective-c objc-cpp-output
793 assembler assembler-with-cpp
796 f77 f77-cpp-input ratfor
801 Turn off any specification of a language, so that subsequent files are
802 handled according to their file name suffixes (as they are if @option{-x}
803 has not been used at all).
805 @item -pass-exit-codes
806 @opindex pass-exit-codes
807 Normally the @command{gcc} program will exit with the code of 1 if any
808 phase of the compiler returns a non-success return code. If you specify
809 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
810 numerically highest error produced by any phase that returned an error
814 If you only want some of the stages of compilation, you can use
815 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
816 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
817 @command{gcc} is to stop. Note that some combinations (for example,
818 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
823 Compile or assemble the source files, but do not link. The linking
824 stage simply is not done. The ultimate output is in the form of an
825 object file for each source file.
827 By default, the object file name for a source file is made by replacing
828 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
830 Unrecognized input files, not requiring compilation or assembly, are
835 Stop after the stage of compilation proper; do not assemble. The output
836 is in the form of an assembler code file for each non-assembler input
839 By default, the assembler file name for a source file is made by
840 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
842 Input files that don't require compilation are ignored.
846 Stop after the preprocessing stage; do not run the compiler proper. The
847 output is in the form of preprocessed source code, which is sent to the
850 Input files which don't require preprocessing are ignored.
852 @cindex output file option
855 Place output in file @var{file}. This applies regardless to whatever
856 sort of output is being produced, whether it be an executable file,
857 an object file, an assembler file or preprocessed C code.
859 Since only one output file can be specified, it does not make sense to
860 use @option{-o} when compiling more than one input file, unless you are
861 producing an executable file as output.
863 If @option{-o} is not specified, the default is to put an executable file
864 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
865 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
866 all preprocessed C source on standard output.
870 Print (on standard error output) the commands executed to run the stages
871 of compilation. Also print the version number of the compiler driver
872 program and of the preprocessor and the compiler proper.
876 Use pipes rather than temporary files for communication between the
877 various stages of compilation. This fails to work on some systems where
878 the assembler is unable to read from a pipe; but the GNU assembler has
883 Print (on the standard output) a description of the command line options
884 understood by @command{gcc}. If the @option{-v} option is also specified
885 then @option{--help} will also be passed on to the various processes
886 invoked by @command{gcc}, so that they can display the command line options
887 they accept. If the @option{-W} option is also specified then command
888 line options which have no documentation associated with them will also
893 Print (on the standard output) a description of target specific command
894 line options for each tool.
898 @section Compiling C++ Programs
900 @cindex suffixes for C++ source
901 @cindex C++ source file suffixes
902 C++ source files conventionally use one of the suffixes @samp{.C},
903 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
904 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
905 files with these names and compiles them as C++ programs even if you
906 call the compiler the same way as for compiling C programs (usually with
907 the name @command{gcc}).
911 However, C++ programs often require class libraries as well as a
912 compiler that understands the C++ language---and under some
913 circumstances, you might want to compile programs from standard input,
914 or otherwise without a suffix that flags them as C++ programs.
915 @command{g++} is a program that calls GCC with the default language
916 set to C++, and automatically specifies linking against the C++
917 library. On many systems, @command{g++} is also
918 installed with the name @command{c++}.
920 @cindex invoking @command{g++}
921 When you compile C++ programs, you may specify many of the same
922 command-line options that you use for compiling programs in any
923 language; or command-line options meaningful for C and related
924 languages; or options that are meaningful only for C++ programs.
925 @xref{C Dialect Options,,Options Controlling C Dialect}, for
926 explanations of options for languages related to C@.
927 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
928 explanations of options that are meaningful only for C++ programs.
930 @node C Dialect Options
931 @section Options Controlling C Dialect
932 @cindex dialect options
933 @cindex language dialect options
934 @cindex options, dialect
936 The following options control the dialect of C (or languages derived
937 from C, such as C++ and Objective-C) that the compiler accepts:
944 In C mode, support all ISO C89 programs. In C++ mode,
945 remove GNU extensions that conflict with ISO C++.
947 This turns off certain features of GCC that are incompatible with ISO
948 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
949 such as the @code{asm} and @code{typeof} keywords, and
950 predefined macros such as @code{unix} and @code{vax} that identify the
951 type of system you are using. It also enables the undesirable and
952 rarely used ISO trigraph feature. For the C compiler,
953 it disables recognition of C++ style @samp{//} comments as well as
954 the @code{inline} keyword.
956 The alternate keywords @code{__asm__}, @code{__extension__},
957 @code{__inline__} and @code{__typeof__} continue to work despite
958 @option{-ansi}. You would not want to use them in an ISO C program, of
959 course, but it is useful to put them in header files that might be included
960 in compilations done with @option{-ansi}. Alternate predefined macros
961 such as @code{__unix__} and @code{__vax__} are also available, with or
962 without @option{-ansi}.
964 The @option{-ansi} option does not cause non-ISO programs to be
965 rejected gratuitously. For that, @option{-pedantic} is required in
966 addition to @option{-ansi}. @xref{Warning Options}.
968 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
969 option is used. Some header files may notice this macro and refrain
970 from declaring certain functions or defining certain macros that the
971 ISO standard doesn't call for; this is to avoid interfering with any
972 programs that might use these names for other things.
974 Functions which would normally be built in but do not have semantics
975 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
976 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
977 built-in functions provided by GCC}, for details of the functions
982 Determine the language standard. This option is currently only
983 supported when compiling C@. A value for this option must be provided;
989 ISO C89 (same as @option{-ansi}).
992 ISO C89 as modified in amendment 1.
998 ISO C99. Note that this standard is not yet fully supported; see
999 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1000 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1003 Default, ISO C89 plus GNU extensions (including some C99 features).
1007 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1008 this will become the default. The name @samp{gnu9x} is deprecated.
1012 Even when this option is not specified, you can still use some of the
1013 features of newer standards in so far as they do not conflict with
1014 previous C standards. For example, you may use @code{__restrict__} even
1015 when @option{-std=c99} is not specified.
1017 The @option{-std} options specifying some version of ISO C have the same
1018 effects as @option{-ansi}, except that features that were not in ISO C89
1019 but are in the specified version (for example, @samp{//} comments and
1020 the @code{inline} keyword in ISO C99) are not disabled.
1022 @xref{Standards,,Language Standards Supported by GCC}, for details of
1023 these standard versions.
1025 @item -aux-info @var{filename}
1027 Output to the given filename prototyped declarations for all functions
1028 declared and/or defined in a translation unit, including those in header
1029 files. This option is silently ignored in any language other than C@.
1031 Besides declarations, the file indicates, in comments, the origin of
1032 each declaration (source file and line), whether the declaration was
1033 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1034 @samp{O} for old, respectively, in the first character after the line
1035 number and the colon), and whether it came from a declaration or a
1036 definition (@samp{C} or @samp{F}, respectively, in the following
1037 character). In the case of function definitions, a K&R-style list of
1038 arguments followed by their declarations is also provided, inside
1039 comments, after the declaration.
1043 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1044 keyword, so that code can use these words as identifiers. You can use
1045 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1046 instead. @option{-ansi} implies @option{-fno-asm}.
1048 In C++, this switch only affects the @code{typeof} keyword, since
1049 @code{asm} and @code{inline} are standard keywords. You may want to
1050 use the @option{-fno-gnu-keywords} flag instead, which has the same
1051 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1052 switch only affects the @code{asm} and @code{typeof} keywords, since
1053 @code{inline} is a standard keyword in ISO C99.
1056 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1057 @opindex fno-builtin
1058 @cindex built-in functions
1059 Don't recognize built-in functions that do not begin with
1060 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1061 functions provided by GCC}, for details of the functions affected,
1062 including those which are not built-in functions when @option{-ansi} or
1063 @option{-std} options for strict ISO C conformance are used because they
1064 do not have an ISO standard meaning.
1066 GCC normally generates special code to handle certain built-in functions
1067 more efficiently; for instance, calls to @code{alloca} may become single
1068 instructions that adjust the stack directly, and calls to @code{memcpy}
1069 may become inline copy loops. The resulting code is often both smaller
1070 and faster, but since the function calls no longer appear as such, you
1071 cannot set a breakpoint on those calls, nor can you change the behavior
1072 of the functions by linking with a different library.
1074 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1075 option has no effect. Therefore, in C++, the only way to get the
1076 optimization benefits of built-in functions is to call the function
1077 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1078 built-in functions to implement many functions (like
1079 @code{std::strchr}), so that you automatically get efficient code.
1081 With the @option{-fno-builtin-@var{function}} option, not available
1082 when compiling C++, only the built-in function @var{function} is
1083 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1084 function is named this is not built-in in this version of GCC, this
1085 option is ignored. There is no corresponding
1086 @option{-fbuiltin-@var{function}} option; if you wish to enable
1087 built-in functions selectively when using @option{-fno-builtin} or
1088 @option{-ffreestanding}, you may define macros such as:
1091 #define abs(n) __builtin_abs ((n))
1092 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1097 @cindex hosted environment
1099 Assert that compilation takes place in a hosted environment. This implies
1100 @option{-fbuiltin}. A hosted environment is one in which the
1101 entire standard library is available, and in which @code{main} has a return
1102 type of @code{int}. Examples are nearly everything except a kernel.
1103 This is equivalent to @option{-fno-freestanding}.
1105 @item -ffreestanding
1106 @opindex ffreestanding
1107 @cindex hosted environment
1109 Assert that compilation takes place in a freestanding environment. This
1110 implies @option{-fno-builtin}. A freestanding environment
1111 is one in which the standard library may not exist, and program startup may
1112 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1113 This is equivalent to @option{-fno-hosted}.
1115 @xref{Standards,,Language Standards Supported by GCC}, for details of
1116 freestanding and hosted environments.
1120 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1121 options for strict ISO C conformance) implies @option{-trigraphs}.
1123 @cindex traditional C language
1124 @cindex C language, traditional
1126 @opindex traditional
1127 Attempt to support some aspects of traditional C compilers.
1132 All @code{extern} declarations take effect globally even if they
1133 are written inside of a function definition. This includes implicit
1134 declarations of functions.
1137 The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const}
1138 and @code{volatile} are not recognized. (You can still use the
1139 alternative keywords such as @code{__typeof__}, @code{__inline__}, and
1143 Comparisons between pointers and integers are always allowed.
1146 Integer types @code{unsigned short} and @code{unsigned char} promote
1147 to @code{unsigned int}.
1150 Out-of-range floating point literals are not an error.
1153 Certain constructs which ISO regards as a single invalid preprocessing
1154 number, such as @samp{0xe-0xd}, are treated as expressions instead.
1157 String ``constants'' are not necessarily constant; they are stored in
1158 writable space, and identical looking constants are allocated
1159 separately. (This is the same as the effect of
1160 @option{-fwritable-strings}.)
1162 @cindex @code{longjmp} and automatic variables
1164 All automatic variables not declared @code{register} are preserved by
1165 @code{longjmp}. Ordinarily, GNU C follows ISO C: automatic variables
1166 not declared @code{volatile} may be clobbered.
1171 @cindex escape sequences, traditional
1172 The character escape sequences @samp{\x} and @samp{\a} evaluate as the
1173 literal characters @samp{x} and @samp{a} respectively. Without
1174 @w{@option{-traditional}}, @samp{\x} is a prefix for the hexadecimal
1175 representation of a character, and @samp{\a} produces a bell.
1178 This option is deprecated and may be removed.
1180 You may wish to use @option{-fno-builtin} as well as @option{-traditional}
1181 if your program uses names that are normally GNU C built-in functions for
1182 other purposes of its own.
1184 You cannot use @option{-traditional} if you include any header files that
1185 rely on ISO C features. Some vendors are starting to ship systems with
1186 ISO C header files and you cannot use @option{-traditional} on such
1187 systems to compile files that include any system headers.
1189 The @option{-traditional} option also enables @option{-traditional-cpp}.
1191 @item -traditional-cpp
1192 @opindex traditional-cpp
1193 Attempt to support some aspects of traditional C preprocessors.
1194 See the GNU CPP manual for details.
1196 @item -fcond-mismatch
1197 @opindex fcond-mismatch
1198 Allow conditional expressions with mismatched types in the second and
1199 third arguments. The value of such an expression is void. This option
1200 is not supported for C++.
1202 @item -funsigned-char
1203 @opindex funsigned-char
1204 Let the type @code{char} be unsigned, like @code{unsigned char}.
1206 Each kind of machine has a default for what @code{char} should
1207 be. It is either like @code{unsigned char} by default or like
1208 @code{signed char} by default.
1210 Ideally, a portable program should always use @code{signed char} or
1211 @code{unsigned char} when it depends on the signedness of an object.
1212 But many programs have been written to use plain @code{char} and
1213 expect it to be signed, or expect it to be unsigned, depending on the
1214 machines they were written for. This option, and its inverse, let you
1215 make such a program work with the opposite default.
1217 The type @code{char} is always a distinct type from each of
1218 @code{signed char} or @code{unsigned char}, even though its behavior
1219 is always just like one of those two.
1222 @opindex fsigned-char
1223 Let the type @code{char} be signed, like @code{signed char}.
1225 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1226 the negative form of @option{-funsigned-char}. Likewise, the option
1227 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1229 @item -fsigned-bitfields
1230 @itemx -funsigned-bitfields
1231 @itemx -fno-signed-bitfields
1232 @itemx -fno-unsigned-bitfields
1233 @opindex fsigned-bitfields
1234 @opindex funsigned-bitfields
1235 @opindex fno-signed-bitfields
1236 @opindex fno-unsigned-bitfields
1237 These options control whether a bit-field is signed or unsigned, when the
1238 declaration does not use either @code{signed} or @code{unsigned}. By
1239 default, such a bit-field is signed, because this is consistent: the
1240 basic integer types such as @code{int} are signed types.
1242 However, when @option{-traditional} is used, bit-fields are all unsigned
1245 @item -fwritable-strings
1246 @opindex fwritable-strings
1247 Store string constants in the writable data segment and don't uniquize
1248 them. This is for compatibility with old programs which assume they can
1249 write into string constants. The option @option{-traditional} also has
1252 Writing into string constants is a very bad idea; ``constants'' should
1255 @item -fallow-single-precision
1256 @opindex fallow-single-precision
1257 Do not promote single precision math operations to double precision,
1258 even when compiling with @option{-traditional}.
1260 Traditional K&R C promotes all floating point operations to double
1261 precision, regardless of the sizes of the operands. On the
1262 architecture for which you are compiling, single precision may be faster
1263 than double precision. If you must use @option{-traditional}, but want
1264 to use single precision operations when the operands are single
1265 precision, use this option. This option has no effect when compiling
1266 with ISO or GNU C conventions (the default).
1269 @opindex fshort-wchar
1270 Override the underlying type for @samp{wchar_t} to be @samp{short
1271 unsigned int} instead of the default for the target. This option is
1272 useful for building programs to run under WINE@.
1275 @node C++ Dialect Options
1276 @section Options Controlling C++ Dialect
1278 @cindex compiler options, C++
1279 @cindex C++ options, command line
1280 @cindex options, C++
1281 This section describes the command-line options that are only meaningful
1282 for C++ programs; but you can also use most of the GNU compiler options
1283 regardless of what language your program is in. For example, you
1284 might compile a file @code{firstClass.C} like this:
1287 g++ -g -frepo -O -c firstClass.C
1291 In this example, only @option{-frepo} is an option meant
1292 only for C++ programs; you can use the other options with any
1293 language supported by GCC@.
1295 Here is a list of options that are @emph{only} for compiling C++ programs:
1298 @item -fno-access-control
1299 @opindex fno-access-control
1300 Turn off all access checking. This switch is mainly useful for working
1301 around bugs in the access control code.
1305 Check that the pointer returned by @code{operator new} is non-null
1306 before attempting to modify the storage allocated. The current Working
1307 Paper requires that @code{operator new} never return a null pointer, so
1308 this check is normally unnecessary.
1310 An alternative to using this option is to specify that your
1311 @code{operator new} does not throw any exceptions; if you declare it
1312 @samp{throw()}, G++ will check the return value. See also @samp{new
1315 @item -fconserve-space
1316 @opindex fconserve-space
1317 Put uninitialized or runtime-initialized global variables into the
1318 common segment, as C does. This saves space in the executable at the
1319 cost of not diagnosing duplicate definitions. If you compile with this
1320 flag and your program mysteriously crashes after @code{main()} has
1321 completed, you may have an object that is being destroyed twice because
1322 two definitions were merged.
1324 This option is no longer useful on most targets, now that support has
1325 been added for putting variables into BSS without making them common.
1327 @item -fno-const-strings
1328 @opindex fno-const-strings
1329 Give string constants type @code{char *} instead of type @code{const
1330 char *}. By default, G++ uses type @code{const char *} as required by
1331 the standard. Even if you use @option{-fno-const-strings}, you cannot
1332 actually modify the value of a string constant, unless you also use
1333 @option{-fwritable-strings}.
1335 This option might be removed in a future release of G++. For maximum
1336 portability, you should structure your code so that it works with
1337 string constants that have type @code{const char *}.
1339 @item -fdollars-in-identifiers
1340 @opindex fdollars-in-identifiers
1341 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1342 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1343 @samp{$} by default on most target systems, but there are a few exceptions.)
1344 Traditional C allowed the character @samp{$} to form part of
1345 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1347 @item -fno-elide-constructors
1348 @opindex fno-elide-constructors
1349 The C++ standard allows an implementation to omit creating a temporary
1350 which is only used to initialize another object of the same type.
1351 Specifying this option disables that optimization, and forces G++ to
1352 call the copy constructor in all cases.
1354 @item -fno-enforce-eh-specs
1355 @opindex fno-enforce-eh-specs
1356 Don't check for violation of exception specifications at runtime. This
1357 option violates the C++ standard, but may be useful for reducing code
1358 size in production builds, much like defining @samp{NDEBUG}. The compiler
1359 will still optimize based on the exception specifications.
1361 @item -fexternal-templates
1362 @opindex fexternal-templates
1364 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1365 template instantiation; template instances are emitted or not according
1366 to the location of the template definition. @xref{Template
1367 Instantiation}, for more information.
1369 This option is deprecated.
1371 @item -falt-external-templates
1372 @opindex falt-external-templates
1373 Similar to @option{-fexternal-templates}, but template instances are
1374 emitted or not according to the place where they are first instantiated.
1375 @xref{Template Instantiation}, for more information.
1377 This option is deprecated.
1380 @itemx -fno-for-scope
1382 @opindex fno-for-scope
1383 If @option{-ffor-scope} is specified, the scope of variables declared in
1384 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1385 as specified by the C++ standard.
1386 If @option{-fno-for-scope} is specified, the scope of variables declared in
1387 a @i{for-init-statement} extends to the end of the enclosing scope,
1388 as was the case in old versions of G++, and other (traditional)
1389 implementations of C++.
1391 The default if neither flag is given to follow the standard,
1392 but to allow and give a warning for old-style code that would
1393 otherwise be invalid, or have different behavior.
1395 @item -fno-gnu-keywords
1396 @opindex fno-gnu-keywords
1397 Do not recognize @code{typeof} as a keyword, so that code can use this
1398 word as an identifier. You can use the keyword @code{__typeof__} instead.
1399 @option{-ansi} implies @option{-fno-gnu-keywords}.
1401 @item -fno-implicit-templates
1402 @opindex fno-implicit-templates
1403 Never emit code for non-inline templates which are instantiated
1404 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1405 @xref{Template Instantiation}, for more information.
1407 @item -fno-implicit-inline-templates
1408 @opindex fno-implicit-inline-templates
1409 Don't emit code for implicit instantiations of inline templates, either.
1410 The default is to handle inlines differently so that compiles with and
1411 without optimization will need the same set of explicit instantiations.
1413 @item -fno-implement-inlines
1414 @opindex fno-implement-inlines
1415 To save space, do not emit out-of-line copies of inline functions
1416 controlled by @samp{#pragma implementation}. This will cause linker
1417 errors if these functions are not inlined everywhere they are called.
1419 @item -fms-extensions
1420 @opindex fms-extensions
1421 Disable pedantic warnings about constructs used in MFC, such as implicit
1422 int and getting a pointer to member function via non-standard syntax.
1424 @item -fno-nonansi-builtins
1425 @opindex fno-nonansi-builtins
1426 Disable built-in declarations of functions that are not mandated by
1427 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1428 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1430 @item -fno-operator-names
1431 @opindex fno-operator-names
1432 Do not treat the operator name keywords @code{and}, @code{bitand},
1433 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1434 synonyms as keywords.
1436 @item -fno-optional-diags
1437 @opindex fno-optional-diags
1438 Disable diagnostics that the standard says a compiler does not need to
1439 issue. Currently, the only such diagnostic issued by G++ is the one for
1440 a name having multiple meanings within a class.
1443 @opindex fpermissive
1444 Downgrade messages about nonconformant code from errors to warnings. By
1445 default, G++ effectively sets @option{-pedantic-errors} without
1446 @option{-pedantic}; this option reverses that. This behavior and this
1447 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1451 Enable automatic template instantiation at link time. This option also
1452 implies @option{-fno-implicit-templates}. @xref{Template
1453 Instantiation}, for more information.
1457 Disable generation of information about every class with virtual
1458 functions for use by the C++ runtime type identification features
1459 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1460 of the language, you can save some space by using this flag. Note that
1461 exception handling uses the same information, but it will generate it as
1466 Emit statistics about front-end processing at the end of the compilation.
1467 This information is generally only useful to the G++ development team.
1469 @item -ftemplate-depth-@var{n}
1470 @opindex ftemplate-depth
1471 Set the maximum instantiation depth for template classes to @var{n}.
1472 A limit on the template instantiation depth is needed to detect
1473 endless recursions during template class instantiation. ANSI/ISO C++
1474 conforming programs must not rely on a maximum depth greater than 17.
1476 @item -fuse-cxa-atexit
1477 @opindex fuse-cxa-atexit
1478 Register destructors for objects with static storage duration with the
1479 @code{__cxa_atexit} function rather than the @code{atexit} function.
1480 This option is required for fully standards-compliant handling of static
1481 destructors, but will only work if your C library supports
1482 @code{__cxa_atexit}.
1486 Emit special relocations for vtables and virtual function references
1487 so that the linker can identify unused virtual functions and zero out
1488 vtable slots that refer to them. This is most useful with
1489 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1490 also discard the functions themselves.
1492 This optimization requires GNU as and GNU ld. Not all systems support
1493 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1497 Do not use weak symbol support, even if it is provided by the linker.
1498 By default, G++ will use weak symbols if they are available. This
1499 option exists only for testing, and should not be used by end-users;
1500 it will result in inferior code and has no benefits. This option may
1501 be removed in a future release of G++.
1505 Do not search for header files in the standard directories specific to
1506 C++, but do still search the other standard directories. (This option
1507 is used when building the C++ library.)
1510 In addition, these optimization, warning, and code generation options
1511 have meanings only for C++ programs:
1514 @item -fno-default-inline
1515 @opindex fno-default-inline
1516 Do not assume @samp{inline} for functions defined inside a class scope.
1517 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1518 functions will have linkage like inline functions; they just won't be
1521 @item -Wctor-dtor-privacy @r{(C++ only)}
1522 @opindex Wctor-dtor-privacy
1523 Warn when a class seems unusable, because all the constructors or
1524 destructors in a class are private and the class has no friends or
1525 public static member functions.
1527 @item -Wnon-virtual-dtor @r{(C++ only)}
1528 @opindex Wnon-virtual-dtor
1529 Warn when a class declares a non-virtual destructor that should probably
1530 be virtual, because it looks like the class will be used polymorphically.
1532 @item -Wreorder @r{(C++ only)}
1534 @cindex reordering, warning
1535 @cindex warning for reordering of member initializers
1536 Warn when the order of member initializers given in the code does not
1537 match the order in which they must be executed. For instance:
1543 A(): j (0), i (1) @{ @}
1547 Here the compiler will warn that the member initializers for @samp{i}
1548 and @samp{j} will be rearranged to match the declaration order of the
1552 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1555 @item -Weffc++ @r{(C++ only)}
1557 Warn about violations of various style guidelines from Scott Meyers'
1558 @cite{Effective C++} books. If you use this option, you should be aware
1559 that the standard library headers do not obey all of these guidelines;
1560 you can use @samp{grep -v} to filter out those warnings.
1562 @item -Wno-deprecated @r{(C++ only)}
1563 @opindex Wno-deprecated
1564 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1566 @item -Wno-non-template-friend @r{(C++ only)}
1567 @opindex Wno-non-template-friend
1568 Disable warnings when non-templatized friend functions are declared
1569 within a template. With the advent of explicit template specification
1570 support in G++, if the name of the friend is an unqualified-id (i.e.,
1571 @samp{friend foo(int)}), the C++ language specification demands that the
1572 friend declare or define an ordinary, nontemplate function. (Section
1573 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1574 could be interpreted as a particular specialization of a templatized
1575 function. Because this non-conforming behavior is no longer the default
1576 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1577 check existing code for potential trouble spots, and is on by default.
1578 This new compiler behavior can be turned off with
1579 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1580 but disables the helpful warning.
1582 @item -Wold-style-cast @r{(C++ only)}
1583 @opindex Wold-style-cast
1584 Warn if an old-style (C-style) cast to a non-void type is used within
1585 a C++ program. The new-style casts (@samp{static_cast},
1586 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1587 unintended effects, and much easier to grep for.
1589 @item -Woverloaded-virtual @r{(C++ only)}
1590 @opindex Woverloaded-virtual
1591 @cindex overloaded virtual fn, warning
1592 @cindex warning for overloaded virtual fn
1593 Warn when a function declaration hides virtual functions from a
1594 base class. For example, in:
1601 struct B: public A @{
1606 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1614 will fail to compile.
1616 @item -Wno-pmf-conversions @r{(C++ only)}
1617 @opindex Wno-pmf-conversions
1618 Disable the diagnostic for converting a bound pointer to member function
1621 @item -Wsign-promo @r{(C++ only)}
1622 @opindex Wsign-promo
1623 Warn when overload resolution chooses a promotion from unsigned or
1624 enumeral type to a signed type over a conversion to an unsigned type of
1625 the same size. Previous versions of G++ would try to preserve
1626 unsignedness, but the standard mandates the current behavior.
1628 @item -Wsynth @r{(C++ only)}
1630 @cindex warning for synthesized methods
1631 @cindex synthesized methods, warning
1632 Warn when G++'s synthesis behavior does not match that of cfront. For
1638 A& operator = (int);
1648 In this example, G++ will synthesize a default @samp{A& operator =
1649 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1652 @node Objective-C Dialect Options
1653 @section Options Controlling Objective-C Dialect
1655 @cindex compiler options, Objective-C
1656 @cindex Objective-C options, command line
1657 @cindex options, Objective-C
1658 This section describes the command-line options that are only meaningful
1659 for Objective-C programs; but you can also use most of the GNU compiler
1660 options regardless of what language your program is in. For example,
1661 you might compile a file @code{some_class.m} like this:
1664 gcc -g -fgnu-runtime -O -c some_class.m
1668 In this example, only @option{-fgnu-runtime} is an option meant only for
1669 Objective-C programs; you can use the other options with any language
1672 Here is a list of options that are @emph{only} for compiling Objective-C
1676 @item -fconstant-string-class=@var{class-name}
1677 @opindex fconstant-string-class
1678 Use @var{class-name} as the name of the class to instantiate for each
1679 literal string specified with the syntax @code{@@"@dots{}"}. The default
1680 class name is @code{NXConstantString}.
1683 @opindex fgnu-runtime
1684 Generate object code compatible with the standard GNU Objective-C
1685 runtime. This is the default for most types of systems.
1687 @item -fnext-runtime
1688 @opindex fnext-runtime
1689 Generate output compatible with the NeXT runtime. This is the default
1690 for NeXT-based systems, including Darwin and Mac OS X@.
1694 Dump interface declarations for all classes seen in the source file to a
1695 file named @file{@var{sourcename}.decl}.
1698 @opindex Wno-protocol
1699 Do not warn if methods required by a protocol are not implemented
1700 in the class adopting it.
1704 Warn if a selector has multiple methods of different types defined.
1706 @c not documented because only avail via -Wp
1707 @c @item -print-objc-runtime-info
1711 @node Language Independent Options
1712 @section Options to Control Diagnostic Messages Formatting
1713 @cindex options to control diagnostics formatting
1714 @cindex diagnostic messages
1715 @cindex message formatting
1717 Traditionally, diagnostic messages have been formatted irrespective of
1718 the output device's aspect (e.g.@: its width, @dots{}). The options described
1719 below can be used to control the diagnostic messages formatting
1720 algorithm, e.g.@: how many characters per line, how often source location
1721 information should be reported. Right now, only the C++ front end can
1722 honor these options. However it is expected, in the near future, that
1723 the remaining front ends would be able to digest them correctly.
1726 @item -fmessage-length=@var{n}
1727 @opindex fmessage-length
1728 Try to format error messages so that they fit on lines of about @var{n}
1729 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1730 the front ends supported by GCC@. If @var{n} is zero, then no
1731 line-wrapping will be done; each error message will appear on a single
1734 @opindex fdiagnostics-show-location
1735 @item -fdiagnostics-show-location=once
1736 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1737 reporter to emit @emph{once} source location information; that is, in
1738 case the message is too long to fit on a single physical line and has to
1739 be wrapped, the source location won't be emitted (as prefix) again,
1740 over and over, in subsequent continuation lines. This is the default
1743 @item -fdiagnostics-show-location=every-line
1744 Only meaningful in line-wrapping mode. Instructs the diagnostic
1745 messages reporter to emit the same source location information (as
1746 prefix) for physical lines that result from the process of breaking
1747 a message which is too long to fit on a single line.
1751 @node Warning Options
1752 @section Options to Request or Suppress Warnings
1753 @cindex options to control warnings
1754 @cindex warning messages
1755 @cindex messages, warning
1756 @cindex suppressing warnings
1758 Warnings are diagnostic messages that report constructions which
1759 are not inherently erroneous but which are risky or suggest there
1760 may have been an error.
1762 You can request many specific warnings with options beginning @samp{-W},
1763 for example @option{-Wimplicit} to request warnings on implicit
1764 declarations. Each of these specific warning options also has a
1765 negative form beginning @samp{-Wno-} to turn off warnings;
1766 for example, @option{-Wno-implicit}. This manual lists only one of the
1767 two forms, whichever is not the default.
1769 These options control the amount and kinds of warnings produced by GCC:
1772 @cindex syntax checking
1774 @opindex fsyntax-only
1775 Check the code for syntax errors, but don't do anything beyond that.
1779 Issue all the warnings demanded by strict ISO C and ISO C++;
1780 reject all programs that use forbidden extensions, and some other
1781 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1782 version of the ISO C standard specified by any @option{-std} option used.
1784 Valid ISO C and ISO C++ programs should compile properly with or without
1785 this option (though a rare few will require @option{-ansi} or a
1786 @option{-std} option specifying the required version of ISO C)@. However,
1787 without this option, certain GNU extensions and traditional C and C++
1788 features are supported as well. With this option, they are rejected.
1790 @option{-pedantic} does not cause warning messages for use of the
1791 alternate keywords whose names begin and end with @samp{__}. Pedantic
1792 warnings are also disabled in the expression that follows
1793 @code{__extension__}. However, only system header files should use
1794 these escape routes; application programs should avoid them.
1795 @xref{Alternate Keywords}.
1797 Some users try to use @option{-pedantic} to check programs for strict ISO
1798 C conformance. They soon find that it does not do quite what they want:
1799 it finds some non-ISO practices, but not all---only those for which
1800 ISO C @emph{requires} a diagnostic, and some others for which
1801 diagnostics have been added.
1803 A feature to report any failure to conform to ISO C might be useful in
1804 some instances, but would require considerable additional work and would
1805 be quite different from @option{-pedantic}. We don't have plans to
1806 support such a feature in the near future.
1808 Where the standard specified with @option{-std} represents a GNU
1809 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1810 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1811 extended dialect is based. Warnings from @option{-pedantic} are given
1812 where they are required by the base standard. (It would not make sense
1813 for such warnings to be given only for features not in the specified GNU
1814 C dialect, since by definition the GNU dialects of C include all
1815 features the compiler supports with the given option, and there would be
1816 nothing to warn about.)
1818 @item -pedantic-errors
1819 @opindex pedantic-errors
1820 Like @option{-pedantic}, except that errors are produced rather than
1825 Inhibit all warning messages.
1829 Inhibit warning messages about the use of @samp{#import}.
1831 @item -Wchar-subscripts
1832 @opindex Wchar-subscripts
1833 Warn if an array subscript has type @code{char}. This is a common cause
1834 of error, as programmers often forget that this type is signed on some
1839 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1840 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1844 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1845 the arguments supplied have types appropriate to the format string
1846 specified, and that the conversions specified in the format string make
1847 sense. This includes standard functions, and others specified by format
1848 attributes (@pxref{Function Attributes}), in the @code{printf},
1849 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1850 not in the C standard) families.
1852 The formats are checked against the format features supported by GNU
1853 libc version 2.2. These include all ISO C89 and C99 features, as well
1854 as features from the Single Unix Specification and some BSD and GNU
1855 extensions. Other library implementations may not support all these
1856 features; GCC does not support warning about features that go beyond a
1857 particular library's limitations. However, if @option{-pedantic} is used
1858 with @option{-Wformat}, warnings will be given about format features not
1859 in the selected standard version (but not for @code{strfmon} formats,
1860 since those are not in any version of the C standard). @xref{C Dialect
1861 Options,,Options Controlling C Dialect}.
1863 @option{-Wformat} is included in @option{-Wall}. For more control over some
1864 aspects of format checking, the options @option{-Wno-format-y2k},
1865 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1866 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1867 not included in @option{-Wall}.
1869 @item -Wno-format-y2k
1870 @opindex Wno-format-y2k
1871 If @option{-Wformat} is specified, do not warn about @code{strftime}
1872 formats which may yield only a two-digit year.
1874 @item -Wno-format-extra-args
1875 @opindex Wno-format-extra-args
1876 If @option{-Wformat} is specified, do not warn about excess arguments to a
1877 @code{printf} or @code{scanf} format function. The C standard specifies
1878 that such arguments are ignored.
1880 Where the unused arguments lie between used arguments that are
1881 specified with @samp{$} operand number specifications, normally
1882 warnings are still given, since the implementation could not know what
1883 type to pass to @code{va_arg} to skip the unused arguments. However,
1884 in the case of @code{scanf} formats, this option will suppress the
1885 warning if the unused arguments are all pointers, since the Single
1886 Unix Specification says that such unused arguments are allowed.
1888 @item -Wformat-nonliteral
1889 @opindex Wformat-nonliteral
1890 If @option{-Wformat} is specified, also warn if the format string is not a
1891 string literal and so cannot be checked, unless the format function
1892 takes its format arguments as a @code{va_list}.
1894 @item -Wformat-security
1895 @opindex Wformat-security
1896 If @option{-Wformat} is specified, also warn about uses of format
1897 functions that represent possible security problems. At present, this
1898 warns about calls to @code{printf} and @code{scanf} functions where the
1899 format string is not a string literal and there are no format arguments,
1900 as in @code{printf (foo);}. This may be a security hole if the format
1901 string came from untrusted input and contains @samp{%n}. (This is
1902 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1903 in future warnings may be added to @option{-Wformat-security} that are not
1904 included in @option{-Wformat-nonliteral}.)
1908 Enable @option{-Wformat} plus format checks not included in
1909 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1910 -Wformat-nonliteral -Wformat-security}.
1912 @item -Wimplicit-int
1913 @opindex Wimplicit-int
1914 Warn when a declaration does not specify a type.
1916 @item -Wimplicit-function-declaration
1917 @itemx -Werror-implicit-function-declaration
1918 @opindex Wimplicit-function-declaration
1919 @opindex Werror-implicit-function-declaration
1920 Give a warning (or error) whenever a function is used before being
1925 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1929 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1930 function with external linkage, returning int, taking either zero
1931 arguments, two, or three arguments of appropriate types.
1933 @item -Wmissing-braces
1934 @opindex Wmissing-braces
1935 Warn if an aggregate or union initializer is not fully bracketed. In
1936 the following example, the initializer for @samp{a} is not fully
1937 bracketed, but that for @samp{b} is fully bracketed.
1940 int a[2][2] = @{ 0, 1, 2, 3 @};
1941 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1945 @opindex Wparentheses
1946 Warn if parentheses are omitted in certain contexts, such
1947 as when there is an assignment in a context where a truth value
1948 is expected, or when operators are nested whose precedence people
1949 often get confused about.
1951 Also warn about constructions where there may be confusion to which
1952 @code{if} statement an @code{else} branch belongs. Here is an example of
1967 In C, every @code{else} branch belongs to the innermost possible @code{if}
1968 statement, which in this example is @code{if (b)}. This is often not
1969 what the programmer expected, as illustrated in the above example by
1970 indentation the programmer chose. When there is the potential for this
1971 confusion, GCC will issue a warning when this flag is specified.
1972 To eliminate the warning, add explicit braces around the innermost
1973 @code{if} statement so there is no way the @code{else} could belong to
1974 the enclosing @code{if}. The resulting code would look like this:
1990 @item -Wsequence-point
1991 @opindex Wsequence-point
1992 Warn about code that may have undefined semantics because of violations
1993 of sequence point rules in the C standard.
1995 The C standard defines the order in which expressions in a C program are
1996 evaluated in terms of @dfn{sequence points}, which represent a partial
1997 ordering between the execution of parts of the program: those executed
1998 before the sequence point, and those executed after it. These occur
1999 after the evaluation of a full expression (one which is not part of a
2000 larger expression), after the evaluation of the first operand of a
2001 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2002 function is called (but after the evaluation of its arguments and the
2003 expression denoting the called function), and in certain other places.
2004 Other than as expressed by the sequence point rules, the order of
2005 evaluation of subexpressions of an expression is not specified. All
2006 these rules describe only a partial order rather than a total order,
2007 since, for example, if two functions are called within one expression
2008 with no sequence point between them, the order in which the functions
2009 are called is not specified. However, the standards committee have
2010 ruled that function calls do not overlap.
2012 It is not specified when between sequence points modifications to the
2013 values of objects take effect. Programs whose behavior depends on this
2014 have undefined behavior; the C standard specifies that ``Between the
2015 previous and next sequence point an object shall have its stored value
2016 modified at most once by the evaluation of an expression. Furthermore,
2017 the prior value shall be read only to determine the value to be
2018 stored.''. If a program breaks these rules, the results on any
2019 particular implementation are entirely unpredictable.
2021 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2022 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2023 diagnosed by this option, and it may give an occasional false positive
2024 result, but in general it has been found fairly effective at detecting
2025 this sort of problem in programs.
2027 The present implementation of this option only works for C programs. A
2028 future implementation may also work for C++ programs.
2030 The C standard is worded confusingly, therefore there is some debate
2031 over the precise meaning of the sequence point rules in subtle cases.
2032 Links to discussions of the problem, including proposed formal
2033 definitions, may be found on our readings page, at
2034 @w{@uref{http://gcc.gnu.org/readings.html}}.
2037 @opindex Wreturn-type
2038 Warn whenever a function is defined with a return-type that defaults to
2039 @code{int}. Also warn about any @code{return} statement with no
2040 return-value in a function whose return-type is not @code{void}.
2042 For C++, a function without return type always produces a diagnostic
2043 message, even when @option{-Wno-return-type} is specified. The only
2044 exceptions are @samp{main} and functions defined in system headers.
2048 Warn whenever a @code{switch} statement has an index of enumeral type
2049 and lacks a @code{case} for one or more of the named codes of that
2050 enumeration. (The presence of a @code{default} label prevents this
2051 warning.) @code{case} labels outside the enumeration range also
2052 provoke warnings when this option is used.
2056 Warn if any trigraphs are encountered that might change the meaning of
2057 the program (trigraphs within comments are not warned about).
2059 @item -Wunused-function
2060 @opindex Wunused-function
2061 Warn whenever a static function is declared but not defined or a
2062 non\-inline static function is unused.
2064 @item -Wunused-label
2065 @opindex Wunused-label
2066 Warn whenever a label is declared but not used.
2068 To suppress this warning use the @samp{unused} attribute
2069 (@pxref{Variable Attributes}).
2071 @item -Wunused-parameter
2072 @opindex Wunused-parameter
2073 Warn whenever a function parameter is unused aside from its declaration.
2075 To suppress this warning use the @samp{unused} attribute
2076 (@pxref{Variable Attributes}).
2078 @item -Wunused-variable
2079 @opindex Wunused-variable
2080 Warn whenever a local variable or non-constant static variable is unused
2081 aside from its declaration
2083 To suppress this warning use the @samp{unused} attribute
2084 (@pxref{Variable Attributes}).
2086 @item -Wunused-value
2087 @opindex Wunused-value
2088 Warn whenever a statement computes a result that is explicitly not used.
2090 To suppress this warning cast the expression to @samp{void}.
2094 All all the above @option{-Wunused} options combined.
2096 In order to get a warning about an unused function parameter, you must
2097 either specify @samp{-W -Wunused} or separately specify
2098 @option{-Wunused-parameter}.
2100 @item -Wuninitialized
2101 @opindex Wuninitialized
2102 Warn if an automatic variable is used without first being initialized or
2103 if a variable may be clobbered by a @code{setjmp} call.
2105 These warnings are possible only in optimizing compilation,
2106 because they require data flow information that is computed only
2107 when optimizing. If you don't specify @option{-O}, you simply won't
2110 These warnings occur only for variables that are candidates for
2111 register allocation. Therefore, they do not occur for a variable that
2112 is declared @code{volatile}, or whose address is taken, or whose size
2113 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2114 structures, unions or arrays, even when they are in registers.
2116 Note that there may be no warning about a variable that is used only
2117 to compute a value that itself is never used, because such
2118 computations may be deleted by data flow analysis before the warnings
2121 These warnings are made optional because GCC is not smart
2122 enough to see all the reasons why the code might be correct
2123 despite appearing to have an error. Here is one example of how
2144 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2145 always initialized, but GCC doesn't know this. Here is
2146 another common case:
2151 if (change_y) save_y = y, y = new_y;
2153 if (change_y) y = save_y;
2158 This has no bug because @code{save_y} is used only if it is set.
2160 @cindex @code{longjmp} warnings
2161 This option also warns when a non-volatile automatic variable might be
2162 changed by a call to @code{longjmp}. These warnings as well are possible
2163 only in optimizing compilation.
2165 The compiler sees only the calls to @code{setjmp}. It cannot know
2166 where @code{longjmp} will be called; in fact, a signal handler could
2167 call it at any point in the code. As a result, you may get a warning
2168 even when there is in fact no problem because @code{longjmp} cannot
2169 in fact be called at the place which would cause a problem.
2171 Some spurious warnings can be avoided if you declare all the functions
2172 you use that never return as @code{noreturn}. @xref{Function
2175 @item -Wreorder @r{(C++ only)}
2177 @cindex reordering, warning
2178 @cindex warning for reordering of member initializers
2179 Warn when the order of member initializers given in the code does not
2180 match the order in which they must be executed. For instance:
2182 @item -Wunknown-pragmas
2183 @opindex Wunknown-pragmas
2184 @cindex warning for unknown pragmas
2185 @cindex unknown pragmas, warning
2186 @cindex pragmas, warning of unknown
2187 Warn when a #pragma directive is encountered which is not understood by
2188 GCC@. If this command line option is used, warnings will even be issued
2189 for unknown pragmas in system header files. This is not the case if
2190 the warnings were only enabled by the @option{-Wall} command line option.
2194 All of the above @samp{-W} options combined. This enables all the
2195 warnings about constructions that some users consider questionable, and
2196 that are easy to avoid (or modify to prevent the warning), even in
2197 conjunction with macros.
2200 @opindex Wno-div-by-zero
2201 @opindex Wdiv-by-zero
2202 Warn about compile-time integer division by zero. This is default. To
2203 inhibit the warning messages, use @option{-Wno-div-by-zero}. Floating
2204 point division by zero is not warned about, as it can be a legitimate
2205 way of obtaining infinities and NaNs.
2208 @opindex Wno-multichar
2210 Warn if a multicharacter constant (@samp{'FOOF'}) is used. This is
2211 default. To inhibit the warning messages, use @option{-Wno-multichar}.
2212 Usually they indicate a typo in the user's code, as they have
2213 implementation-defined values, and should not be used in portable code.
2215 @item -Wsystem-headers
2216 @opindex Wsystem-headers
2217 @cindex warnings from system headers
2218 @cindex system headers, warnings from
2219 Print warning messages for constructs found in system header files.
2220 Warnings from system headers are normally suppressed, on the assumption
2221 that they usually do not indicate real problems and would only make the
2222 compiler output harder to read. Using this command line option tells
2223 GCC to emit warnings from system headers as if they occurred in user
2224 code. However, note that using @option{-Wall} in conjunction with this
2225 option will @emph{not} warn about unknown pragmas in system
2226 headers---for that, @option{-Wunknown-pragmas} must also be used.
2229 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2230 Some of them warn about constructions that users generally do not
2231 consider questionable, but which occasionally you might wish to check
2232 for; others warn about constructions that are necessary or hard to avoid
2233 in some cases, and there is no simple way to modify the code to suppress
2239 Print extra warning messages for these events:
2243 A function can return either with or without a value. (Falling
2244 off the end of the function body is considered returning without
2245 a value.) For example, this function would evoke such a
2259 An expression-statement or the left-hand side of a comma expression
2260 contains no side effects.
2261 To suppress the warning, cast the unused expression to void.
2262 For example, an expression such as @samp{x[i,j]} will cause a warning,
2263 but @samp{x[(void)i,j]} will not.
2266 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2269 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2270 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2271 that of ordinary mathematical notation.
2274 Storage-class specifiers like @code{static} are not the first things in
2275 a declaration. According to the C Standard, this usage is obsolescent.
2278 The return type of a function has a type qualifier such as @code{const}.
2279 Such a type qualifier has no effect, since the value returned by a
2280 function is not an lvalue. (But don't warn about the GNU extension of
2281 @code{volatile void} return types. That extension will be warned about
2282 if @option{-pedantic} is specified.)
2285 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2289 A comparison between signed and unsigned values could produce an
2290 incorrect result when the signed value is converted to unsigned.
2291 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2294 An aggregate has a partly bracketed initializer.
2295 For example, the following code would evoke such a warning,
2296 because braces are missing around the initializer for @code{x.h}:
2299 struct s @{ int f, g; @};
2300 struct t @{ struct s h; int i; @};
2301 struct t x = @{ 1, 2, 3 @};
2305 An aggregate has an initializer which does not initialize all members.
2306 For example, the following code would cause such a warning, because
2307 @code{x.h} would be implicitly initialized to zero:
2310 struct s @{ int f, g, h; @};
2311 struct s x = @{ 3, 4 @};
2316 @opindex Wfloat-equal
2317 Warn if floating point values are used in equality comparisons.
2319 The idea behind this is that sometimes it is convenient (for the
2320 programmer) to consider floating-point values as approximations to
2321 infinitely precise real numbers. If you are doing this, then you need
2322 to compute (by analysing the code, or in some other way) the maximum or
2323 likely maximum error that the computation introduces, and allow for it
2324 when performing comparisons (and when producing output, but that's a
2325 different problem). In particular, instead of testing for equality, you
2326 would check to see whether the two values have ranges that overlap; and
2327 this is done with the relational operators, so equality comparisons are
2330 @item -Wtraditional @r{(C only)}
2331 @opindex Wtraditional
2332 Warn about certain constructs that behave differently in traditional and
2333 ISO C@. Also warn about ISO C constructs that have no traditional C
2334 equivalent, and/or problematic constructs which should be avoided.
2338 Macro parameters that appear within string literals in the macro body.
2339 In traditional C macro replacement takes place within string literals,
2340 but does not in ISO C@.
2343 In traditional C, some preprocessor directives did not exist.
2344 Traditional preprocessors would only consider a line to be a directive
2345 if the @samp{#} appeared in column 1 on the line. Therefore
2346 @option{-Wtraditional} warns about directives that traditional C
2347 understands but would ignore because the @samp{#} does not appear as the
2348 first character on the line. It also suggests you hide directives like
2349 @samp{#pragma} not understood by traditional C by indenting them. Some
2350 traditional implementations would not recognize @samp{#elif}, so it
2351 suggests avoiding it altogether.
2354 A function-like macro that appears without arguments.
2357 The unary plus operator.
2360 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2361 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2362 constants.) Note, these suffixes appear in macros defined in the system
2363 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2364 Use of these macros in user code might normally lead to spurious
2365 warnings, however gcc's integrated preprocessor has enough context to
2366 avoid warning in these cases.
2369 A function declared external in one block and then used after the end of
2373 A @code{switch} statement has an operand of type @code{long}.
2376 A non-@code{static} function declaration follows a @code{static} one.
2377 This construct is not accepted by some traditional C compilers.
2380 The ISO type of an integer constant has a different width or
2381 signedness from its traditional type. This warning is only issued if
2382 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2383 typically represent bit patterns, are not warned about.
2386 Usage of ISO string concatenation is detected.
2389 Initialization of automatic aggregates.
2392 Identifier conflicts with labels. Traditional C lacks a separate
2393 namespace for labels.
2396 Initialization of unions. If the initializer is zero, the warning is
2397 omitted. This is done under the assumption that the zero initializer in
2398 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2399 initializer warnings and relies on default initialization to zero in the
2403 Conversions by prototypes between fixed/floating point values and vice
2404 versa. The absence of these prototypes when compiling with traditional
2405 C would cause serious problems. This is a subset of the possible
2406 conversion warnings, for the full set use @option{-Wconversion}.
2411 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2415 Warn whenever a local variable shadows another local variable, parameter or
2416 global variable or whenever a built-in function is shadowed.
2418 @item -Wlarger-than-@var{len}
2419 @opindex Wlarger-than
2420 Warn whenever an object of larger than @var{len} bytes is defined.
2422 @item -Wpointer-arith
2423 @opindex Wpointer-arith
2424 Warn about anything that depends on the ``size of'' a function type or
2425 of @code{void}. GNU C assigns these types a size of 1, for
2426 convenience in calculations with @code{void *} pointers and pointers
2429 @item -Wbad-function-cast @r{(C only)}
2430 @opindex Wbad-function-cast
2431 Warn whenever a function call is cast to a non-matching type.
2432 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2436 Warn whenever a pointer is cast so as to remove a type qualifier from
2437 the target type. For example, warn if a @code{const char *} is cast
2438 to an ordinary @code{char *}.
2441 @opindex Wcast-align
2442 Warn whenever a pointer is cast such that the required alignment of the
2443 target is increased. For example, warn if a @code{char *} is cast to
2444 an @code{int *} on machines where integers can only be accessed at
2445 two- or four-byte boundaries.
2447 @item -Wwrite-strings
2448 @opindex Wwrite-strings
2449 When compiling C, give string constants the type @code{const
2450 char[@var{length}]} so that
2451 copying the address of one into a non-@code{const} @code{char *}
2452 pointer will get a warning; when compiling C++, warn about the
2453 deprecated conversion from string constants to @code{char *}.
2454 These warnings will help you find at
2455 compile time code that can try to write into a string constant, but
2456 only if you have been very careful about using @code{const} in
2457 declarations and prototypes. Otherwise, it will just be a nuisance;
2458 this is why we did not make @option{-Wall} request these warnings.
2461 @opindex Wconversion
2462 Warn if a prototype causes a type conversion that is different from what
2463 would happen to the same argument in the absence of a prototype. This
2464 includes conversions of fixed point to floating and vice versa, and
2465 conversions changing the width or signedness of a fixed point argument
2466 except when the same as the default promotion.
2468 Also, warn if a negative integer constant expression is implicitly
2469 converted to an unsigned type. For example, warn about the assignment
2470 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2471 casts like @code{(unsigned) -1}.
2473 @item -Wsign-compare
2474 @opindex Wsign-compare
2475 @cindex warning for comparison of signed and unsigned values
2476 @cindex comparison of signed and unsigned values, warning
2477 @cindex signed and unsigned values, comparison warning
2478 Warn when a comparison between signed and unsigned values could produce
2479 an incorrect result when the signed value is converted to unsigned.
2480 This warning is also enabled by @option{-W}; to get the other warnings
2481 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2483 @item -Waggregate-return
2484 @opindex Waggregate-return
2485 Warn if any functions that return structures or unions are defined or
2486 called. (In languages where you can return an array, this also elicits
2489 @item -Wstrict-prototypes @r{(C only)}
2490 @opindex Wstrict-prototypes
2491 Warn if a function is declared or defined without specifying the
2492 argument types. (An old-style function definition is permitted without
2493 a warning if preceded by a declaration which specifies the argument
2496 @item -Wmissing-prototypes @r{(C only)}
2497 @opindex Wmissing-prototypes
2498 Warn if a global function is defined without a previous prototype
2499 declaration. This warning is issued even if the definition itself
2500 provides a prototype. The aim is to detect global functions that fail
2501 to be declared in header files.
2503 @item -Wmissing-declarations
2504 @opindex Wmissing-declarations
2505 Warn if a global function is defined without a previous declaration.
2506 Do so even if the definition itself provides a prototype.
2507 Use this option to detect global functions that are not declared in
2510 @item -Wmissing-noreturn
2511 @opindex Wmissing-noreturn
2512 Warn about functions which might be candidates for attribute @code{noreturn}.
2513 Note these are only possible candidates, not absolute ones. Care should
2514 be taken to manually verify functions actually do not ever return before
2515 adding the @code{noreturn} attribute, otherwise subtle code generation
2516 bugs could be introduced. You will not get a warning for @code{main} in
2517 hosted C environments.
2519 @item -Wmissing-format-attribute
2520 @opindex Wmissing-format-attribute
2522 If @option{-Wformat} is enabled, also warn about functions which might be
2523 candidates for @code{format} attributes. Note these are only possible
2524 candidates, not absolute ones. GCC will guess that @code{format}
2525 attributes might be appropriate for any function that calls a function
2526 like @code{vprintf} or @code{vscanf}, but this might not always be the
2527 case, and some functions for which @code{format} attributes are
2528 appropriate may not be detected. This option has no effect unless
2529 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2531 @item -Wno-deprecated-declarations
2532 @opindex Wno-deprecated-declarations
2533 Do not warn about uses of functions, variables, and types marked as
2534 deprecated by using the @code{deprecated} attribute.
2535 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2536 @pxref{Type Attributes}.)
2540 Warn if a structure is given the packed attribute, but the packed
2541 attribute has no effect on the layout or size of the structure.
2542 Such structures may be mis-aligned for little benefit. For
2543 instance, in this code, the variable @code{f.x} in @code{struct bar}
2544 will be misaligned even though @code{struct bar} does not itself
2545 have the packed attribute:
2552 @} __attribute__((packed));
2562 Warn if padding is included in a structure, either to align an element
2563 of the structure or to align the whole structure. Sometimes when this
2564 happens it is possible to rearrange the fields of the structure to
2565 reduce the padding and so make the structure smaller.
2567 @item -Wredundant-decls
2568 @opindex Wredundant-decls
2569 Warn if anything is declared more than once in the same scope, even in
2570 cases where multiple declaration is valid and changes nothing.
2572 @item -Wnested-externs @r{(C only)}
2573 @opindex Wnested-externs
2574 Warn if an @code{extern} declaration is encountered within a function.
2576 @item -Wunreachable-code
2577 @opindex Wunreachable-code
2578 Warn if the compiler detects that code will never be executed.
2580 This option is intended to warn when the compiler detects that at
2581 least a whole line of source code will never be executed, because
2582 some condition is never satisfied or because it is after a
2583 procedure that never returns.
2585 It is possible for this option to produce a warning even though there
2586 are circumstances under which part of the affected line can be executed,
2587 so care should be taken when removing apparently-unreachable code.
2589 For instance, when a function is inlined, a warning may mean that the
2590 line is unreachable in only one inlined copy of the function.
2592 This option is not made part of @option{-Wall} because in a debugging
2593 version of a program there is often substantial code which checks
2594 correct functioning of the program and is, hopefully, unreachable
2595 because the program does work. Another common use of unreachable
2596 code is to provide behavior which is selectable at compile-time.
2600 Warn if a function can not be inlined and it was declared as inline.
2604 @opindex Wno-long-long
2605 Warn if @samp{long long} type is used. This is default. To inhibit
2606 the warning messages, use @option{-Wno-long-long}. Flags
2607 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2608 only when @option{-pedantic} flag is used.
2610 @item -Wdisabled-optimization
2611 @opindex Wdisabled-optimization
2612 Warn if a requested optimization pass is disabled. This warning does
2613 not generally indicate that there is anything wrong with your code; it
2614 merely indicates that GCC's optimizers were unable to handle the code
2615 effectively. Often, the problem is that your code is too big or too
2616 complex; GCC will refuse to optimize programs when the optimization
2617 itself is likely to take inordinate amounts of time.
2621 Make all warnings into errors.
2624 @node Debugging Options
2625 @section Options for Debugging Your Program or GCC
2626 @cindex options, debugging
2627 @cindex debugging information options
2629 GCC has various special options that are used for debugging
2630 either your program or GCC:
2635 Produce debugging information in the operating system's native format
2636 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2639 On most systems that use stabs format, @option{-g} enables use of extra
2640 debugging information that only GDB can use; this extra information
2641 makes debugging work better in GDB but will probably make other debuggers
2643 refuse to read the program. If you want to control for certain whether
2644 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2645 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2646 or @option{-gvms} (see below).
2648 Unlike most other C compilers, GCC allows you to use @option{-g} with
2649 @option{-O}. The shortcuts taken by optimized code may occasionally
2650 produce surprising results: some variables you declared may not exist
2651 at all; flow of control may briefly move where you did not expect it;
2652 some statements may not be executed because they compute constant
2653 results or their values were already at hand; some statements may
2654 execute in different places because they were moved out of loops.
2656 Nevertheless it proves possible to debug optimized output. This makes
2657 it reasonable to use the optimizer for programs that might have bugs.
2659 The following options are useful when GCC is generated with the
2660 capability for more than one debugging format.
2664 Produce debugging information for use by GDB@. This means to use the
2665 most expressive format available (DWARF 2, stabs, or the native format
2666 if neither of those are supported), including GDB extensions if at all
2671 Produce debugging information in stabs format (if that is supported),
2672 without GDB extensions. This is the format used by DBX on most BSD
2673 systems. On MIPS, Alpha and System V Release 4 systems this option
2674 produces stabs debugging output which is not understood by DBX or SDB@.
2675 On System V Release 4 systems this option requires the GNU assembler.
2679 Produce debugging information in stabs format (if that is supported),
2680 using GNU extensions understood only by the GNU debugger (GDB)@. The
2681 use of these extensions is likely to make other debuggers crash or
2682 refuse to read the program.
2686 Produce debugging information in COFF format (if that is supported).
2687 This is the format used by SDB on most System V systems prior to
2692 Produce debugging information in XCOFF format (if that is supported).
2693 This is the format used by the DBX debugger on IBM RS/6000 systems.
2697 Produce debugging information in XCOFF format (if that is supported),
2698 using GNU extensions understood only by the GNU debugger (GDB)@. The
2699 use of these extensions is likely to make other debuggers crash or
2700 refuse to read the program, and may cause assemblers other than the GNU
2701 assembler (GAS) to fail with an error.
2705 Produce debugging information in DWARF version 1 format (if that is
2706 supported). This is the format used by SDB on most System V Release 4
2711 Produce debugging information in DWARF version 1 format (if that is
2712 supported), using GNU extensions understood only by the GNU debugger
2713 (GDB)@. The use of these extensions is likely to make other debuggers
2714 crash or refuse to read the program.
2718 Produce debugging information in DWARF version 2 format (if that is
2719 supported). This is the format used by DBX on IRIX 6.
2723 Produce debugging information in VMS debug format (if that is
2724 supported). This is the format used by DEBUG on VMS systems.
2727 @itemx -ggdb@var{level}
2728 @itemx -gstabs@var{level}
2729 @itemx -gcoff@var{level}
2730 @itemx -gxcoff@var{level}
2731 @itemx -gdwarf@var{level}
2732 @itemx -gdwarf-2@var{level}
2733 @itemx -gvms@var{level}
2734 Request debugging information and also use @var{level} to specify how
2735 much information. The default level is 2.
2737 Level 1 produces minimal information, enough for making backtraces in
2738 parts of the program that you don't plan to debug. This includes
2739 descriptions of functions and external variables, but no information
2740 about local variables and no line numbers.
2742 Level 3 includes extra information, such as all the macro definitions
2743 present in the program. Some debuggers support macro expansion when
2744 you use @option{-g3}.
2749 Generate extra code to write profile information suitable for the
2750 analysis program @code{prof}. You must use this option when compiling
2751 the source files you want data about, and you must also use it when
2754 @cindex @code{gprof}
2757 Generate extra code to write profile information suitable for the
2758 analysis program @code{gprof}. You must use this option when compiling
2759 the source files you want data about, and you must also use it when
2765 Generate extra code to write profile information for basic blocks, which will
2766 record the number of times each basic block is executed, the basic block start
2767 address, and the function name containing the basic block. If @option{-g} is
2768 used, the line number and filename of the start of the basic block will also be
2769 recorded. If not overridden by the machine description, the default action is
2770 to append to the text file @file{bb.out}.
2772 This data could be analyzed by a program like @code{tcov}. Note,
2773 however, that the format of the data is not what @code{tcov} expects.
2774 Eventually GNU @code{gprof} should be extended to process this data.
2778 Makes the compiler print out each function name as it is compiled, and
2779 print some statistics about each pass when it finishes.
2782 @opindex ftime-report
2783 Makes the compiler print some statistics about the time consumed by each
2784 pass when it finishes.
2787 @opindex fmem-report
2788 Makes the compiler print some statistics about permanent memory
2789 allocation when it finishes.
2791 @item -fprofile-arcs
2792 @opindex fprofile-arcs
2793 Instrument @dfn{arcs} during compilation to generate coverage data
2794 or for profile-directed block ordering. During execution the program
2795 records how many times each branch is executed and how many times it is
2796 taken. When the compiled program exits it saves this data to a file
2797 called @file{@var{sourcename}.da} for each source file.
2799 For profile-directed block ordering, compile the program with
2800 @option{-fprofile-arcs} plus optimization and code generation options,
2801 generate the arc profile information by running the program on a
2802 selected workload, and then compile the program again with the same
2803 optimization and code generation options plus
2804 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2805 Control Optimization}).
2807 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2808 when it is used with the @option{-ftest-coverage} option. GCC
2809 supports two methods of determining code coverage: the options that
2810 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2811 write information to text files. The options that support @code{gcov}
2812 do not need to instrument every arc in the program, so a program compiled
2813 with them runs faster than a program compiled with @option{-a}, which
2814 adds instrumentation code to every basic block in the program. The
2815 tradeoff: since @code{gcov} does not have execution counts for all
2816 branches, it must start with the execution counts for the instrumented
2817 branches, and then iterate over the program flow graph until the entire
2818 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2819 a program which uses information from @option{-a} and @option{-ax}.
2821 With @option{-fprofile-arcs}, for each function of your program GCC
2822 creates a program flow graph, then finds a spanning tree for the graph.
2823 Only arcs that are not on the spanning tree have to be instrumented: the
2824 compiler adds code to count the number of times that these arcs are
2825 executed. When an arc is the only exit or only entrance to a block, the
2826 instrumentation code can be added to the block; otherwise, a new basic
2827 block must be created to hold the instrumentation code.
2829 This option makes it possible to estimate branch probabilities and to
2830 calculate basic block execution counts. In general, basic block
2831 execution counts as provided by @option{-a} do not give enough
2832 information to estimate all branch probabilities.
2835 @item -ftest-coverage
2836 @opindex ftest-coverage
2837 Create data files for the @code{gcov} code-coverage utility
2838 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2839 The data file names begin with the name of your source file:
2842 @item @var{sourcename}.bb
2843 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2844 associate basic block execution counts with line numbers.
2846 @item @var{sourcename}.bbg
2847 A list of all arcs in the program flow graph. This allows @code{gcov}
2848 to reconstruct the program flow graph, so that it can compute all basic
2849 block and arc execution counts from the information in the
2850 @code{@var{sourcename}.da} file.
2853 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2854 option adds instrumentation to the program, which then writes
2855 execution counts to another data file:
2858 @item @var{sourcename}.da
2859 Runtime arc execution counts, used in conjunction with the arc
2860 information in the file @code{@var{sourcename}.bbg}.
2863 Coverage data will map better to the source files if
2864 @option{-ftest-coverage} is used without optimization.
2866 @item -d@var{letters}
2868 Says to make debugging dumps during compilation at times specified by
2869 @var{letters}. This is used for debugging the compiler. The file names
2870 for most of the dumps are made by appending a pass number and a word to
2871 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2872 Here are the possible letters for use in @var{letters}, and their meanings:
2877 Annotate the assembler output with miscellaneous debugging information.
2880 Dump after computing branch probabilities, to @file{@var{file}.14.bp}.
2883 Dump after block reordering, to @file{@var{file}.28.bbro}.
2886 Dump after instruction combination, to the file @file{@var{file}.16.combine}.
2889 Dump after the first if conversion, to the file @file{@var{file}.17.ce}.
2892 Dump after delayed branch scheduling, to @file{@var{file}.31.dbr}.
2895 Dump all macro definitions, at the end of preprocessing, in addition to
2899 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2900 @file{@var{file}.07.ussa}.
2903 Dump after the second if conversion, to @file{@var{file}.26.ce2}.
2906 Dump after life analysis, to @file{@var{file}.15.life}.
2909 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.09.addressof}.
2912 Dump after global register allocation, to @file{@var{file}.21.greg}.
2915 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2918 Dump after post-reload optimizations, to @file{@var{file}.22.postreload}.
2921 Dump after GCSE, to @file{@var{file}.10.gcse}.
2924 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2927 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2930 Dump after conversion from registers to stack, to @file{@var{file}.32.stack}.
2933 Dump after local register allocation, to @file{@var{file}.20.lreg}.
2936 Dump after loop optimization, to @file{@var{file}.11.loop}.
2939 Dump after performing the machine dependent reorganisation pass, to
2940 @file{@var{file}.30.mach}.
2943 Dump after register renumbering, to @file{@var{file}.25.rnreg}.
2946 Dump after the register move pass, to @file{@var{file}.18.regmove}.
2949 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2952 Dump after the second scheduling pass, to @file{@var{file}.27.sched2}.
2955 Dump after CSE (including the jump optimization that sometimes follows
2956 CSE), to @file{@var{file}.08.cse}.
2959 Dump after the first scheduling pass, to @file{@var{file}.19.sched}.
2962 Dump after the second CSE pass (including the jump optimization that
2963 sometimes follows CSE), to @file{@var{file}.12.cse2}.
2966 Dump after the second flow pass, to @file{@var{file}.23.flow2}.
2969 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2972 Dump after the peephole pass, to @file{@var{file}.24.peephole2}.
2975 Produce all the dumps listed above.
2978 Print statistics on memory usage, at the end of the run, to
2982 Annotate the assembler output with a comment indicating which
2983 pattern and alternative was used. The length of each instruction is
2987 Dump the RTL in the assembler output as a comment before each instruction.
2988 Also turns on @option{-dp} annotation.
2991 For each of the other indicated dump files (except for
2992 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
2993 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
2996 Just generate RTL for a function instead of compiling it. Usually used
3000 Dump debugging information during parsing, to standard error.
3003 @item -fdump-unnumbered
3004 @opindex fdump-unnumbered
3005 When doing debugging dumps (see @option{-d} option above), suppress instruction
3006 numbers and line number note output. This makes it more feasible to
3007 use diff on debugging dumps for compiler invocations with different
3008 options, in particular with and without @option{-g}.
3010 @item -fdump-class-hierarchy @r{(C++ only)}
3011 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3012 @opindex fdump-class-hierarchy
3013 Dump a representation of each class's hierarchy and virtual function
3014 table layout to a file. The file name is made by appending @file{.class}
3015 to the source file name. If the @samp{-@var{options}} form is used,
3016 @var{options} controls the details of the dump as described for the
3017 @option{-fdump-tree} options.
3019 @item -fdump-tree-@var{switch} @r{(C++ only)}
3020 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3022 Control the dumping at various stages of processing the intermediate
3023 language tree to a file. The file name is generated by appending a switch
3024 specific suffix to the source file name. If the @samp{-@var{options}}
3025 form is used, @var{options} is a list of @samp{-} separated options that
3026 control the details of the dump. Not all options are applicable to all
3027 dumps, those which are not meaningful will be ignored. The following
3028 options are available
3032 Print the address of each node. Usually this is not meaningful as it
3033 changes according to the environment and source file. Its primary use
3034 is for tying up a dump file with a debug environment.
3036 Inhibit dumping of members of a scope or body of a function merely
3037 because that scope has been reached. Only dump such items when they
3038 are directly reachable by some other path.
3040 Turn on all options.
3043 The following tree dumps are possible:
3046 Dump before any tree based optimization, to @file{@var{file}.original}.
3048 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3050 Dump after function inlining, to @file{@var{file}.inlined}.
3053 @item -fpretend-float
3054 @opindex fpretend-float
3055 When running a cross-compiler, pretend that the target machine uses the
3056 same floating point format as the host machine. This causes incorrect
3057 output of the actual floating constants, but the actual instruction
3058 sequence will probably be the same as GCC would make when running on
3063 Store the usual ``temporary'' intermediate files permanently; place them
3064 in the current directory and name them based on the source file. Thus,
3065 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3066 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3067 preprocessed @file{foo.i} output file even though the compiler now
3068 normally uses an integrated preprocessor.
3072 Report the CPU time taken by each subprocess in the compilation
3073 sequence. For C source files, this is the compiler proper and assembler
3074 (plus the linker if linking is done). The output looks like this:
3081 The first number on each line is the ``user time,'' that is time spent
3082 executing the program itself. The second number is ``system time,''
3083 time spent executing operating system routines on behalf of the program.
3084 Both numbers are in seconds.
3086 @item -print-file-name=@var{library}
3087 @opindex print-file-name
3088 Print the full absolute name of the library file @var{library} that
3089 would be used when linking---and don't do anything else. With this
3090 option, GCC does not compile or link anything; it just prints the
3093 @item -print-multi-directory
3094 @opindex print-multi-directory
3095 Print the directory name corresponding to the multilib selected by any
3096 other switches present in the command line. This directory is supposed
3097 to exist in @env{GCC_EXEC_PREFIX}.
3099 @item -print-multi-lib
3100 @opindex print-multi-lib
3101 Print the mapping from multilib directory names to compiler switches
3102 that enable them. The directory name is separated from the switches by
3103 @samp{;}, and each switch starts with an @samp{@@} instead of the
3104 @samp{-}, without spaces between multiple switches. This is supposed to
3105 ease shell-processing.
3107 @item -print-prog-name=@var{program}
3108 @opindex print-prog-name
3109 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3111 @item -print-libgcc-file-name
3112 @opindex print-libgcc-file-name
3113 Same as @option{-print-file-name=libgcc.a}.
3115 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3116 but you do want to link with @file{libgcc.a}. You can do
3119 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3122 @item -print-search-dirs
3123 @opindex print-search-dirs
3124 Print the name of the configured installation directory and a list of
3125 program and library directories gcc will search---and don't do anything else.
3127 This is useful when gcc prints the error message
3128 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3129 To resolve this you either need to put @file{cpp0} and the other compiler
3130 components where gcc expects to find them, or you can set the environment
3131 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3132 Don't forget the trailing '/'.
3133 @xref{Environment Variables}.
3136 @opindex dumpmachine
3137 Print the compiler's target machine (for example,
3138 @samp{i686-pc-linux-gnu})---and don't do anything else.
3141 @opindex dumpversion
3142 Print the compiler version (for example, @samp{3.0})---and don't do
3147 Print the compiler's built-in specs---and don't do anything else. (This
3148 is used when GCC itself is being built.) @xref{Spec Files}.
3151 @node Optimize Options
3152 @section Options That Control Optimization
3153 @cindex optimize options
3154 @cindex options, optimization
3156 These options control various sorts of optimizations:
3163 Optimize. Optimizing compilation takes somewhat more time, and a lot
3164 more memory for a large function.
3166 Without @option{-O}, the compiler's goal is to reduce the cost of
3167 compilation and to make debugging produce the expected results.
3168 Statements are independent: if you stop the program with a breakpoint
3169 between statements, you can then assign a new value to any variable or
3170 change the program counter to any other statement in the function and
3171 get exactly the results you would expect from the source code.
3173 With @option{-O}, the compiler tries to reduce code size and execution
3174 time, without performing any optimizations that take a great deal of
3179 Optimize even more. GCC performs nearly all supported optimizations
3180 that do not involve a space-speed tradeoff. The compiler does not
3181 perform loop unrolling or function inlining when you specify @option{-O2}.
3182 As compared to @option{-O}, this option increases both compilation time
3183 and the performance of the generated code.
3185 @option{-O2} turns on all optional optimizations except for loop unrolling,
3186 function inlining, and register renaming. It also turns on the
3187 @option{-fforce-mem} option on all machines and frame pointer elimination
3188 on machines where doing so does not interfere with debugging.
3190 Please note the warning under @option{-fgcse} about
3191 invoking @option{-O2} on programs that use computed gotos.
3195 Optimize yet more. @option{-O3} turns on all optimizations specified by
3196 @option{-O2} and also turns on the @option{-finline-functions} and
3197 @option{-frename-registers} options.
3205 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3206 do not typically increase code size. It also performs further
3207 optimizations designed to reduce code size.
3209 If you use multiple @option{-O} options, with or without level numbers,
3210 the last such option is the one that is effective.
3213 Options of the form @option{-f@var{flag}} specify machine-independent
3214 flags. Most flags have both positive and negative forms; the negative
3215 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3216 only one of the forms is listed---the one which is not the default.
3217 You can figure out the other form by either removing @samp{no-} or
3222 @opindex ffloat-store
3223 Do not store floating point variables in registers, and inhibit other
3224 options that might change whether a floating point value is taken from a
3227 @cindex floating point precision
3228 This option prevents undesirable excess precision on machines such as
3229 the 68000 where the floating registers (of the 68881) keep more
3230 precision than a @code{double} is supposed to have. Similarly for the
3231 x86 architecture. For most programs, the excess precision does only
3232 good, but a few programs rely on the precise definition of IEEE floating
3233 point. Use @option{-ffloat-store} for such programs, after modifying
3234 them to store all pertinent intermediate computations into variables.
3236 @item -fno-default-inline
3237 @opindex fno-default-inline
3238 Do not make member functions inline by default merely because they are
3239 defined inside the class scope (C++ only). Otherwise, when you specify
3240 @w{@option{-O}}, member functions defined inside class scope are compiled
3241 inline by default; i.e., you don't need to add @samp{inline} in front of
3242 the member function name.
3244 @item -fno-defer-pop
3245 @opindex fno-defer-pop
3246 Always pop the arguments to each function call as soon as that function
3247 returns. For machines which must pop arguments after a function call,
3248 the compiler normally lets arguments accumulate on the stack for several
3249 function calls and pops them all at once.
3253 Force memory operands to be copied into registers before doing
3254 arithmetic on them. This produces better code by making all memory
3255 references potential common subexpressions. When they are not common
3256 subexpressions, instruction combination should eliminate the separate
3257 register-load. The @option{-O2} option turns on this option.
3260 @opindex fforce-addr
3261 Force memory address constants to be copied into registers before
3262 doing arithmetic on them. This may produce better code just as
3263 @option{-fforce-mem} may.
3265 @item -fomit-frame-pointer
3266 @opindex fomit-frame-pointer
3267 Don't keep the frame pointer in a register for functions that
3268 don't need one. This avoids the instructions to save, set up and
3269 restore frame pointers; it also makes an extra register available
3270 in many functions. @strong{It also makes debugging impossible on
3273 On some machines, such as the VAX, this flag has no effect, because
3274 the standard calling sequence automatically handles the frame pointer
3275 and nothing is saved by pretending it doesn't exist. The
3276 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3277 whether a target machine supports this flag. @xref{Registers,,Register
3278 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3280 @item -foptimize-sibling-calls
3281 @opindex foptimize-sibling-calls
3282 Optimize sibling and tail recursive calls.
3286 This option generates traps for signed overflow on addition, subtraction,
3287 multiplication operations.
3291 Don't pay attention to the @code{inline} keyword. Normally this option
3292 is used to keep the compiler from expanding any functions inline.
3293 Note that if you are not optimizing, no functions can be expanded inline.
3295 @item -finline-functions
3296 @opindex finline-functions
3297 Integrate all simple functions into their callers. The compiler
3298 heuristically decides which functions are simple enough to be worth
3299 integrating in this way.
3301 If all calls to a given function are integrated, and the function is
3302 declared @code{static}, then the function is normally not output as
3303 assembler code in its own right.
3305 @item -finline-limit=@var{n}
3306 @opindex finline-limit
3307 By default, gcc limits the size of functions that can be inlined. This flag
3308 allows the control of this limit for functions that are explicitly marked as
3309 inline (ie marked with the inline keyword or defined within the class
3310 definition in c++). @var{n} is the size of functions that can be inlined in
3311 number of pseudo instructions (not counting parameter handling). The default
3312 value of @var{n} is 600.
3313 Increasing this value can result in more inlined code at
3314 the cost of compilation time and memory consumption. Decreasing usually makes
3315 the compilation faster and less code will be inlined (which presumably
3316 means slower programs). This option is particularly useful for programs that
3317 use inlining heavily such as those based on recursive templates with C++.
3319 @emph{Note:} pseudo instruction represents, in this particular context, an
3320 abstract measurement of function's size. In no way, it represents a count
3321 of assembly instructions and as such its exact meaning might change from one
3322 release to an another.
3324 @item -fkeep-inline-functions
3325 @opindex fkeep-inline-functions
3326 Even if all calls to a given function are integrated, and the function
3327 is declared @code{static}, nevertheless output a separate run-time
3328 callable version of the function. This switch does not affect
3329 @code{extern inline} functions.
3331 @item -fkeep-static-consts
3332 @opindex fkeep-static-consts
3333 Emit variables declared @code{static const} when optimization isn't turned
3334 on, even if the variables aren't referenced.
3336 GCC enables this option by default. If you want to force the compiler to
3337 check if the variable was referenced, regardless of whether or not
3338 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3340 @item -fmerge-constants
3341 Attempt to merge identical constants (string constants and floating point
3342 constants) accross compilation units.
3344 This option is default for optimized compilation if assembler and linker
3345 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3347 @item -fmerge-all-constants
3348 Attempt to merge identical constants and identical variables.
3350 This option implies @option{-fmerge-constants}. In addition to
3351 @option{-fmerge-constants} this considers e.g. even constant initialized
3352 arrays or initialized constant variables with integral or floating point
3353 types. Languages like C or C++ require each non-automatic variable to
3354 have distinct location, so using this option will result in non-conforming
3357 @item -fno-function-cse
3358 @opindex fno-function-cse
3359 Do not put function addresses in registers; make each instruction that
3360 calls a constant function contain the function's address explicitly.
3362 This option results in less efficient code, but some strange hacks
3363 that alter the assembler output may be confused by the optimizations
3364 performed when this option is not used.
3368 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3369 @option{-fno-trapping-math}.
3371 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3373 This option should never be turned on by any @option{-O} option since
3374 it can result in incorrect output for programs which depend on
3375 an exact implementation of IEEE or ISO rules/specifications for
3378 @item -fno-math-errno
3379 @opindex fno-math-errno
3380 Do not set ERRNO after calling math functions that are executed
3381 with a single instruction, e.g., sqrt. A program that relies on
3382 IEEE exceptions for math error handling may want to use this flag
3383 for speed while maintaining IEEE arithmetic compatibility.
3385 This option should never be turned on by any @option{-O} option since
3386 it can result in incorrect output for programs which depend on
3387 an exact implementation of IEEE or ISO rules/specifications for
3390 The default is @option{-fmath-errno}.
3392 @item -funsafe-math-optimizations
3393 @opindex funsafe-math-optimizations
3394 Allow optimizations for floating-point arithmetic that (a) assume
3395 that arguments and results are valid and (b) may violate IEEE or
3396 ANSI standards. When used at link-time, it may include libraries
3397 or startup files that change the default FPU control word or other
3398 similar optimizations.
3400 This option should never be turned on by any @option{-O} option since
3401 it can result in incorrect output for programs which depend on
3402 an exact implementation of IEEE or ISO rules/specifications for
3405 The default is @option{-fno-unsafe-math-optimizations}.
3407 @item -fno-trapping-math
3408 @opindex fno-trapping-math
3409 Compile code assuming that floating-point operations cannot generate
3410 user-visible traps. Setting this option may allow faster code
3411 if one relies on ``non-stop'' IEEE arithmetic, for example.
3413 This option should never be turned on by any @option{-O} option since
3414 it can result in incorrect output for programs which depend on
3415 an exact implementation of IEEE or ISO rules/specifications for
3418 The default is @option{-ftrapping-math}.
3421 The following options control specific optimizations. The @option{-O2}
3422 option turns on all of these optimizations except @option{-funroll-loops}
3423 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3424 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3425 but specific machines may handle it differently.
3427 You can use the following flags in the rare cases when ``fine-tuning''
3428 of optimizations to be performed is desired.
3430 Not all of the optimizations performed by GCC have @option{-f} options
3434 @item -fstrength-reduce
3435 @opindex fstrength-reduce
3436 Perform the optimizations of loop strength reduction and
3437 elimination of iteration variables.
3439 @item -fthread-jumps
3440 @opindex fthread-jumps
3441 Perform optimizations where we check to see if a jump branches to a
3442 location where another comparison subsumed by the first is found. If
3443 so, the first branch is redirected to either the destination of the
3444 second branch or a point immediately following it, depending on whether
3445 the condition is known to be true or false.
3447 @item -fcse-follow-jumps
3448 @opindex fcse-follow-jumps
3449 In common subexpression elimination, scan through jump instructions
3450 when the target of the jump is not reached by any other path. For
3451 example, when CSE encounters an @code{if} statement with an
3452 @code{else} clause, CSE will follow the jump when the condition
3455 @item -fcse-skip-blocks
3456 @opindex fcse-skip-blocks
3457 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3458 follow jumps which conditionally skip over blocks. When CSE
3459 encounters a simple @code{if} statement with no else clause,
3460 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3461 body of the @code{if}.
3463 @item -frerun-cse-after-loop
3464 @opindex frerun-cse-after-loop
3465 Re-run common subexpression elimination after loop optimizations has been
3468 @item -frerun-loop-opt
3469 @opindex frerun-loop-opt
3470 Run the loop optimizer twice.
3474 Perform a global common subexpression elimination pass.
3475 This pass also performs global constant and copy propagation.
3477 @emph{Note:} When compiling a program using computed gotos, a GCC
3478 extension, you may get better runtime performance if you disable
3479 the global common subexpression elmination pass by adding
3480 @option{-fno-gcse} to the command line.
3484 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3485 attempt to move loads which are only killed by stores into themselves. This
3486 allows a loop containing a load/store sequence to be changed to a load outside
3487 the loop, and a copy/store within the loop.
3491 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3492 subexpression elimination. This pass will attempt to move stores out of loops.
3493 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3494 can be changed to a load before the loop and a store after the loop.
3496 @item -fdelete-null-pointer-checks
3497 @opindex fdelete-null-pointer-checks
3498 Use global dataflow analysis to identify and eliminate useless checks
3499 for null pointers. The compiler assumes that dereferencing a null
3500 pointer would have halted the program. If a pointer is checked after
3501 it has already been dereferenced, it cannot be null.
3503 In some environments, this assumption is not true, and programs can
3504 safely dereference null pointers. Use
3505 @option{-fno-delete-null-pointer-checks} to disable this optimization
3506 for programs which depend on that behavior.
3508 @item -fexpensive-optimizations
3509 @opindex fexpensive-optimizations
3510 Perform a number of minor optimizations that are relatively expensive.
3512 @item -foptimize-register-move
3514 @opindex foptimize-register-move
3516 Attempt to reassign register numbers in move instructions and as
3517 operands of other simple instructions in order to maximize the amount of
3518 register tying. This is especially helpful on machines with two-operand
3519 instructions. GCC enables this optimization by default with @option{-O2}
3522 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3525 @item -fdelayed-branch
3526 @opindex fdelayed-branch
3527 If supported for the target machine, attempt to reorder instructions
3528 to exploit instruction slots available after delayed branch
3531 @item -fschedule-insns
3532 @opindex fschedule-insns
3533 If supported for the target machine, attempt to reorder instructions to
3534 eliminate execution stalls due to required data being unavailable. This
3535 helps machines that have slow floating point or memory load instructions
3536 by allowing other instructions to be issued until the result of the load
3537 or floating point instruction is required.
3539 @item -fschedule-insns2
3540 @opindex fschedule-insns2
3541 Similar to @option{-fschedule-insns}, but requests an additional pass of
3542 instruction scheduling after register allocation has been done. This is
3543 especially useful on machines with a relatively small number of
3544 registers and where memory load instructions take more than one cycle.
3546 @item -ffunction-sections
3547 @itemx -fdata-sections
3548 @opindex ffunction-sections
3549 @opindex fdata-sections
3550 Place each function or data item into its own section in the output
3551 file if the target supports arbitrary sections. The name of the
3552 function or the name of the data item determines the section's name
3555 Use these options on systems where the linker can perform optimizations
3556 to improve locality of reference in the instruction space. HPPA
3557 processors running HP-UX and Sparc processors running Solaris 2 have
3558 linkers with such optimizations. Other systems using the ELF object format
3559 as well as AIX may have these optimizations in the future.
3561 Only use these options when there are significant benefits from doing
3562 so. When you specify these options, the assembler and linker will
3563 create larger object and executable files and will also be slower.
3564 You will not be able to use @code{gprof} on all systems if you
3565 specify this option and you may have problems with debugging if
3566 you specify both this option and @option{-g}.
3568 @item -fcaller-saves
3569 @opindex fcaller-saves
3570 Enable values to be allocated in registers that will be clobbered by
3571 function calls, by emitting extra instructions to save and restore the
3572 registers around such calls. Such allocation is done only when it
3573 seems to result in better code than would otherwise be produced.
3575 This option is always enabled by default on certain machines, usually
3576 those which have no call-preserved registers to use instead.
3578 For all machines, optimization level 2 and higher enables this flag by
3581 @item -funroll-loops
3582 @opindex funroll-loops
3583 Unroll loops whose number of iterations can be determined at compile
3584 time or upon entry to the loop. @option{-funroll-loops} implies both
3585 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3586 option makes code larger, and may or may not make it run faster.
3588 @item -funroll-all-loops
3589 @opindex funroll-all-loops
3590 Unroll all loops, even if their number of iterations is uncertain when
3591 the loop is entered. This usually makes programs run more slowly.
3592 @option{-funroll-all-loops} implies the same options as
3593 @option{-funroll-loops},
3595 @item -fprefetch-loop-arrays
3596 @opindex fprefetch-loop-arrays
3597 If supported by the target machine, generate instructions to prefetch
3598 memory to improve the performance of loops that access large arrays.
3600 @item -fmove-all-movables
3601 @opindex fmove-all-movables
3602 Forces all invariant computations in loops to be moved
3605 @item -freduce-all-givs
3606 @opindex freduce-all-givs
3607 Forces all general-induction variables in loops to be
3610 @emph{Note:} When compiling programs written in Fortran,
3611 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3612 by default when you use the optimizer.
3614 These options may generate better or worse code; results are highly
3615 dependent on the structure of loops within the source code.
3617 These two options are intended to be removed someday, once
3618 they have helped determine the efficacy of various
3619 approaches to improving loop optimizations.
3621 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3622 know how use of these options affects
3623 the performance of your production code.
3624 We're very interested in code that runs @emph{slower}
3625 when these options are @emph{enabled}.
3628 @itemx -fno-peephole2
3629 @opindex fno-peephole
3630 @opindex fno-peephole2
3631 Disable any machine-specific peephole optimizations. The difference
3632 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3633 are implemented in the compiler; some targets use one, some use the
3634 other, a few use both.
3636 @item -fbranch-probabilities
3637 @opindex fbranch-probabilities
3638 After running a program compiled with @option{-fprofile-arcs}
3639 (@pxref{Debugging Options,, Options for Debugging Your Program or
3640 @command{gcc}}), you can compile it a second time using
3641 @option{-fbranch-probabilities}, to improve optimizations based on
3642 the number of times each branch was taken. When the program
3643 compiled with @option{-fprofile-arcs} exits it saves arc execution
3644 counts to a file called @file{@var{sourcename}.da} for each source
3645 file The information in this data file is very dependent on the
3646 structure of the generated code, so you must use the same source code
3647 and the same optimization options for both compilations.
3649 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3650 note on the first instruction of each basic block, and a
3651 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3652 These can be used to improve optimization. Currently, they are only
3653 used in one place: in @file{reorg.c}, instead of guessing which path a
3654 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3655 exactly determine which path is taken more often.
3657 @item -fno-guess-branch-probability
3658 @opindex fno-guess-branch-probability
3659 Do not guess branch probabilities using a randomized model.
3661 Sometimes gcc will opt to use a randomized model to guess branch
3662 probabilities, when none are available from either profiling feedback
3663 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3664 different runs of the compiler on the same program may produce different
3667 In a hard real-time system, people don't want different runs of the
3668 compiler to produce code that has different behavior; minimizing
3669 non-determinism is of paramount import. This switch allows users to
3670 reduce non-determinism, possibly at the expense of inferior
3673 @item -fstrict-aliasing
3674 @opindex fstrict-aliasing
3675 Allows the compiler to assume the strictest aliasing rules applicable to
3676 the language being compiled. For C (and C++), this activates
3677 optimizations based on the type of expressions. In particular, an
3678 object of one type is assumed never to reside at the same address as an
3679 object of a different type, unless the types are almost the same. For
3680 example, an @code{unsigned int} can alias an @code{int}, but not a
3681 @code{void*} or a @code{double}. A character type may alias any other
3684 Pay special attention to code like this:
3697 The practice of reading from a different union member than the one most
3698 recently written to (called ``type-punning'') is common. Even with
3699 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3700 is accessed through the union type. So, the code above will work as
3701 expected. However, this code might not:
3712 Every language that wishes to perform language-specific alias analysis
3713 should define a function that computes, given an @code{tree}
3714 node, an alias set for the node. Nodes in different alias sets are not
3715 allowed to alias. For an example, see the C front-end function
3716 @code{c_get_alias_set}.
3718 @item -falign-functions
3719 @itemx -falign-functions=@var{n}
3720 @opindex falign-functions
3721 Align the start of functions to the next power-of-two greater than
3722 @var{n}, skipping up to @var{n} bytes. For instance,
3723 @option{-falign-functions=32} aligns functions to the next 32-byte
3724 boundary, but @option{-falign-functions=24} would align to the next
3725 32-byte boundary only if this can be done by skipping 23 bytes or less.
3727 @option{-fno-align-functions} and @option{-falign-functions=1} are
3728 equivalent and mean that functions will not be aligned.
3730 Some assemblers only support this flag when @var{n} is a power of two;
3731 in that case, it is rounded up.
3733 If @var{n} is not specified, use a machine-dependent default.
3735 @item -falign-labels
3736 @itemx -falign-labels=@var{n}
3737 @opindex falign-labels
3738 Align all branch targets to a power-of-two boundary, skipping up to
3739 @var{n} bytes like @option{-falign-functions}. This option can easily
3740 make code slower, because it must insert dummy operations for when the
3741 branch target is reached in the usual flow of the code.
3743 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3744 are greater than this value, then their values are used instead.
3746 If @var{n} is not specified, use a machine-dependent default which is
3747 very likely to be @samp{1}, meaning no alignment.
3750 @itemx -falign-loops=@var{n}
3751 @opindex falign-loops
3752 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3753 like @option{-falign-functions}. The hope is that the loop will be
3754 executed many times, which will make up for any execution of the dummy
3757 If @var{n} is not specified, use a machine-dependent default.
3760 @itemx -falign-jumps=@var{n}
3761 @opindex falign-jumps
3762 Align branch targets to a power-of-two boundary, for branch targets
3763 where the targets can only be reached by jumping, skipping up to @var{n}
3764 bytes like @option{-falign-functions}. In this case, no dummy operations
3767 If @var{n} is not specified, use a machine-dependent default.
3771 Perform optimizations in static single assignment form. Each function's
3772 flow graph is translated into SSA form, optimizations are performed, and
3773 the flow graph is translated back from SSA form. Users should not
3774 specify this option, since it is not yet ready for production use.
3778 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3779 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3783 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3784 Like @option{-fssa}, this is an experimental feature.
3786 @item -fsingle-precision-constant
3787 @opindex fsingle-precision-constant
3788 Treat floating point constant as single precision constant instead of
3789 implicitly converting it to double precision constant.
3791 @item -frename-registers
3792 @opindex frename-registers
3793 Attempt to avoid false dependencies in scheduled code by making use
3794 of registers left over after register allocation. This optimization
3795 will most benefit processors with lots of registers. It can, however,
3796 make debugging impossible, since variables will no longer stay in
3797 a ``home register''.
3799 @item -fno-cprop-registers
3800 @opindex fno-cprop-registers
3801 After register allocation and post-register allocation instruction splitting,
3802 we perform a copy-propagation pass to try to reduce scheduling dependencies
3803 and occasionally eliminate the copy.
3805 @item --param @var{name}=@var{value}
3807 In some places, GCC uses various constants to control the amount of
3808 optimization that is done. For example, GCC will not inline functions
3809 that contain more that a certain number of instructions. You can
3810 control some of these constants on the command-line using the
3811 @option{--param} option.
3813 In each case, the @var{value} is an integer. The allowable choices for
3814 @var{name} are given in the following table:
3817 @item max-delay-slot-insn-search
3818 The maximum number of instructions to consider when looking for an
3819 instruction to fill a delay slot. If more than this arbitrary number of
3820 instructions is searched, the time savings from filling the delay slot
3821 will be minimal so stop searching. Increasing values mean more
3822 aggressive optimization, making the compile time increase with probably
3823 small improvement in executable run time.
3825 @item max-delay-slot-live-search
3826 When trying to fill delay slots, the maximum number of instructions to
3827 consider when searching for a block with valid live register
3828 information. Increasing this arbitrarily chosen value means more
3829 aggressive optimization, increasing the compile time. This parameter
3830 should be removed when the delay slot code is rewritten to maintain the
3833 @item max-gcse-memory
3834 The approximate maximum amount of memory that will be allocated in
3835 order to perform the global common subexpression elimination
3836 optimization. If more memory than specified is required, the
3837 optimization will not be done.
3839 @item max-gcse-passes
3840 The maximum number of passes of GCSE to run.
3842 @item max-pending-list-length
3843 The maximum number of pending dependencies scheduling will allow
3844 before flushing the current state and starting over. Large functions
3845 with few branches or calls can create excessively large lists which
3846 needlessly consume memory and resources.
3848 @item max-inline-insns
3849 If an function contains more than this many instructions, it
3850 will not be inlined. This option is precisely equivalent to
3851 @option{-finline-limit}.
3856 @node Preprocessor Options
3857 @section Options Controlling the Preprocessor
3858 @cindex preprocessor options
3859 @cindex options, preprocessor
3861 These options control the C preprocessor, which is run on each C source
3862 file before actual compilation.
3864 If you use the @option{-E} option, nothing is done except preprocessing.
3865 Some of these options make sense only together with @option{-E} because
3866 they cause the preprocessor output to be unsuitable for actual
3870 @item -include @var{file}
3872 Process @var{file} as input before processing the regular input file.
3873 In effect, the contents of @var{file} are compiled first. Any @option{-D}
3874 and @option{-U} options on the command line are always processed before
3875 @option{-include @var{file}}, regardless of the order in which they are
3876 written. All the @option{-include} and @option{-imacros} options are
3877 processed in the order in which they are written.
3879 @item -imacros @var{file}
3881 Process @var{file} as input, discarding the resulting output, before
3882 processing the regular input file. Because the output generated from
3883 @var{file} is discarded, the only effect of @option{-imacros @var{file}}
3884 is to make the macros defined in @var{file} available for use in the
3885 main input. All the @option{-include} and @option{-imacros} options are
3886 processed in the order in which they are written.
3888 @item -idirafter @var{dir}
3890 @cindex second include path
3891 Add the directory @var{dir} to the second include path. The directories
3892 on the second include path are searched when a header file is not found
3893 in any of the directories in the main include path (the one that
3894 @option{-I} adds to).
3896 @item -iprefix @var{prefix}
3898 Specify @var{prefix} as the prefix for subsequent @option{-iwithprefix}
3901 @item -iwithprefix @var{dir}
3902 @opindex iwithprefix
3903 Add a directory to the second include path. The directory's name is
3904 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
3905 specified previously with @option{-iprefix}. If you have not specified a
3906 prefix yet, the directory containing the installed passes of the
3907 compiler is used as the default.
3909 @item -iwithprefixbefore @var{dir}
3910 @opindex iwithprefixbefore
3911 Add a directory to the main include path. The directory's name is made
3912 by concatenating @var{prefix} and @var{dir}, as in the case of
3913 @option{-iwithprefix}.
3915 @item -isystem @var{dir}
3917 Add a directory to the beginning of the second include path, marking it
3918 as a system directory, so that it gets the same special treatment as
3919 is applied to the standard system directories.
3923 Do not search the standard system directories for header files. Only
3924 the directories you have specified with @option{-I} options (and the
3925 current directory, if appropriate) are searched. @xref{Directory
3926 Options}, for information on @option{-I}.
3928 By using both @option{-nostdinc} and @option{-I-}, you can limit the include-file
3929 search path to only those directories you specify explicitly.
3933 When searching for a header file in a directory, remap file names if a
3934 file named @file{header.gcc} exists in that directory. This can be used
3935 to work around limitations of file systems with file name restrictions.
3936 The @file{header.gcc} file should contain a series of lines with two
3937 tokens on each line: the first token is the name to map, and the second
3938 token is the actual name to use.
3942 Do not predefine any nonstandard macros. (Including architecture flags).
3946 Run only the C preprocessor. Preprocess all the C source files
3947 specified and output the results to standard output or to the
3948 specified output file.
3952 Tell the preprocessor not to discard comments. Used with the
3957 Tell the preprocessor not to generate @samp{#line} directives.
3958 Used with the @option{-E} option.
3961 @cindex dependencies, make
3964 Instead of outputting the result of preprocessing, output a rule
3965 suitable for @code{make} describing the dependencies of the main source
3966 file. The preprocessor outputs one @code{make} rule containing the
3967 object file name for that source file, a colon, and the names of all the
3968 included files. Unless overridden explicitly, the object file name
3969 consists of the basename of the source file with any suffix replaced with
3970 object file suffix. If there are many included files then the
3971 rule is split into several lines using @samp{\}-newline.
3973 @option{-M} implies @option{-E}.
3977 Like @option{-M}, but mention only the files included with @samp{#include
3978 "@var{file}"}. System header files included with @samp{#include
3979 <@var{file}>} are omitted.
3983 Like @option{-M} but the dependency information is written to a file
3984 rather than stdout. @code{gcc} will use the same file name and
3985 directory as the object file, but with the suffix @file{.d} instead.
3987 This is in addition to compiling the main file as specified---@option{-MD}
3988 does not inhibit ordinary compilation the way @option{-M} does,
3989 unless you also specify @option{-MG}.
3991 With Mach, you can use the utility @code{md} to merge multiple
3992 dependency files into a single dependency file suitable for using with
3993 the @samp{make} command.
3997 Like @option{-MD} except mention only user header files, not system
4000 @item -MF @var{file}
4002 When used with @option{-M} or @option{-MM}, specifies a file to write the
4003 dependencies to. This allows the preprocessor to write the preprocessed
4004 file to stdout normally. If no @option{-MF} switch is given, CPP sends
4005 the rules to stdout and suppresses normal preprocessed output.
4007 Another way to specify output of a @code{make} rule is by setting
4008 the environment variable @env{DEPENDENCIES_OUTPUT} (@pxref{Environment
4013 When used with @option{-M} or @option{-MM}, @option{-MG} says to treat missing
4014 header files as generated files and assume they live in the same
4015 directory as the source file. It suppresses preprocessed output, as a
4016 missing header file is ordinarily an error.
4018 This feature is used in automatic updating of makefiles.
4022 This option instructs CPP to add a phony target for each dependency
4023 other than the main file, causing each to depend on nothing. These
4024 dummy rules work around errors @code{make} gives if you remove header
4025 files without updating the @code{Makefile} to match.
4027 This is typical output:-
4030 /tmp/test.o: /tmp/test.c /tmp/test.h
4035 @item -MQ @var{target}
4036 @item -MT @var{target}
4039 By default CPP uses the main file name, including any path, and appends
4040 the object suffix, normally ``.o'', to it to obtain the name of the
4041 target for dependency generation. With @option{-MT} you can specify a
4042 target yourself, overriding the default one.
4044 If you want multiple targets, you can specify them as a single argument
4045 to @option{-MT}, or use multiple @option{-MT} options.
4047 The targets you specify are output in the order they appear on the
4048 command line. @option{-MQ} is identical to @option{-MT}, except that the
4049 target name is quoted for Make, but with @option{-MT} it isn't. For
4050 example, @option{-MT '$(objpfx)foo.o'} gives
4053 $(objpfx)foo.o: /tmp/foo.c
4056 but @option{-MQ '$(objpfx)foo.o'} gives
4059 $$(objpfx)foo.o: /tmp/foo.c
4062 The default target is automatically quoted, as if it were given with
4067 Print the name of each header file used, in addition to other normal
4070 @item -A@var{question}(@var{answer})
4072 Assert the answer @var{answer} for @var{question}, in case it is tested
4073 with a preprocessing conditional such as @samp{#if
4074 #@var{question}(@var{answer})}. @option{-A-} disables the standard
4075 assertions that normally describe the target machine.
4079 Define macro @var{macro} with the string @samp{1} as its definition.
4081 @item -D@var{macro}=@var{defn}
4082 Define macro @var{macro} as @var{defn}. All instances of @option{-D} on
4083 the command line are processed before any @option{-U} options.
4085 Any @option{-D} and @option{-U} options on the command line are processed in
4086 order, and always before @option{-imacros @var{file}}, regardless of the
4087 order in which they are written.
4091 Undefine macro @var{macro}. @option{-U} options are evaluated after all
4092 @option{-D} options, but before any @option{-include} and @option{-imacros}
4095 Any @option{-D} and @option{-U} options on the command line are processed in
4096 order, and always before @option{-imacros @var{file}}, regardless of the
4097 order in which they are written.
4101 Tell the preprocessor to output only a list of the macro definitions
4102 that are in effect at the end of preprocessing. Used with the @option{-E}
4107 Tell the preprocessing to pass all macro definitions into the output, in
4108 their proper sequence in the rest of the output.
4112 Like @option{-dD} except that the macro arguments and contents are omitted.
4113 Only @samp{#define @var{name}} is included in the output.
4117 Output @samp{#include} directives in addition to the result of
4120 @item -fpreprocessed
4121 @opindex fpreprocessed
4122 Indicate to the preprocessor that the input file has already been
4123 preprocessed. This suppresses things like macro expansion, trigraph
4124 conversion, escaped newline splicing, and processing of most directives.
4125 The preprocessor still recognizes and removes comments, so that you can
4126 pass a file preprocessed with @option{-C} to the compiler without
4127 problems. In this mode the integrated preprocessor is little more than
4128 a tokenizer for the front ends.
4130 @option{-fpreprocessed} is implicit if the input file has one of the
4131 extensions @samp{i}, @samp{ii} or @samp{mi}. These are the extensions
4132 that GCC uses for preprocessed files created by @option{-save-temps}.
4136 Process ISO standard trigraph sequences. These are three-character
4137 sequences, all starting with @samp{??}, that are defined by ISO C to
4138 stand for single characters. For example, @samp{??/} stands for
4139 @samp{\}, so @samp{'??/n'} is a character constant for a newline. By
4140 default, GCC ignores trigraphs, but in standard-conforming modes it
4141 converts them. See the @option{-std} and @option{-ansi} options.
4143 The nine trigraph sequences are
4146 @expansion{} @samp{[}
4149 @expansion{} @samp{]}
4152 @expansion{} @samp{@{}
4155 @expansion{} @samp{@}}
4158 @expansion{} @samp{#}
4161 @expansion{} @samp{\}
4164 @expansion{} @samp{^}
4167 @expansion{} @samp{|}
4170 @expansion{} @samp{~}
4174 Trigraph support is not popular, so many compilers do not implement it
4175 properly. Portable code should not rely on trigraphs being either
4176 converted or ignored.
4178 @item -Wp,@var{option}
4180 Pass @var{option} as an option to the preprocessor. If @var{option}
4181 contains commas, it is split into multiple options at the commas.
4184 @node Assembler Options
4185 @section Passing Options to the Assembler
4187 @c prevent bad page break with this line
4188 You can pass options to the assembler.
4191 @item -Wa,@var{option}
4193 Pass @var{option} as an option to the assembler. If @var{option}
4194 contains commas, it is split into multiple options at the commas.
4198 @section Options for Linking
4199 @cindex link options
4200 @cindex options, linking
4202 These options come into play when the compiler links object files into
4203 an executable output file. They are meaningless if the compiler is
4204 not doing a link step.
4208 @item @var{object-file-name}
4209 A file name that does not end in a special recognized suffix is
4210 considered to name an object file or library. (Object files are
4211 distinguished from libraries by the linker according to the file
4212 contents.) If linking is done, these object files are used as input
4221 If any of these options is used, then the linker is not run, and
4222 object file names should not be used as arguments. @xref{Overall
4226 @item -l@var{library}
4227 @itemx -l @var{library}
4229 Search the library named @var{library} when linking. (The second
4230 alternative with the library as a separate argument is only for
4231 POSIX compliance and is not recommended.)
4233 It makes a difference where in the command you write this option; the
4234 linker searches and processes libraries and object files in the order they
4235 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4236 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4237 to functions in @samp{z}, those functions may not be loaded.
4239 The linker searches a standard list of directories for the library,
4240 which is actually a file named @file{lib@var{library}.a}. The linker
4241 then uses this file as if it had been specified precisely by name.
4243 The directories searched include several standard system directories
4244 plus any that you specify with @option{-L}.
4246 Normally the files found this way are library files---archive files
4247 whose members are object files. The linker handles an archive file by
4248 scanning through it for members which define symbols that have so far
4249 been referenced but not defined. But if the file that is found is an
4250 ordinary object file, it is linked in the usual fashion. The only
4251 difference between using an @option{-l} option and specifying a file name
4252 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4253 and searches several directories.
4257 You need this special case of the @option{-l} option in order to
4258 link an Objective-C program.
4261 @opindex nostartfiles
4262 Do not use the standard system startup files when linking.
4263 The standard system libraries are used normally, unless @option{-nostdlib}
4264 or @option{-nodefaultlibs} is used.
4266 @item -nodefaultlibs
4267 @opindex nodefaultlibs
4268 Do not use the standard system libraries when linking.
4269 Only the libraries you specify will be passed to the linker.
4270 The standard startup files are used normally, unless @option{-nostartfiles}
4271 is used. The compiler may generate calls to memcmp, memset, and memcpy
4272 for System V (and ISO C) environments or to bcopy and bzero for
4273 BSD environments. These entries are usually resolved by entries in
4274 libc. These entry points should be supplied through some other
4275 mechanism when this option is specified.
4279 Do not use the standard system startup files or libraries when linking.
4280 No startup files and only the libraries you specify will be passed to
4281 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4282 for System V (and ISO C) environments or to bcopy and bzero for
4283 BSD environments. These entries are usually resolved by entries in
4284 libc. These entry points should be supplied through some other
4285 mechanism when this option is specified.
4287 @cindex @option{-lgcc}, use with @option{-nostdlib}
4288 @cindex @option{-nostdlib} and unresolved references
4289 @cindex unresolved references and @option{-nostdlib}
4290 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4291 @cindex @option{-nodefaultlibs} and unresolved references
4292 @cindex unresolved references and @option{-nodefaultlibs}
4293 One of the standard libraries bypassed by @option{-nostdlib} and
4294 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4295 that GCC uses to overcome shortcomings of particular machines, or special
4296 needs for some languages.
4297 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4298 Collection (GCC) Internals},
4299 for more discussion of @file{libgcc.a}.)
4300 In most cases, you need @file{libgcc.a} even when you want to avoid
4301 other standard libraries. In other words, when you specify @option{-nostdlib}
4302 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4303 This ensures that you have no unresolved references to internal GCC
4304 library subroutines. (For example, @samp{__main}, used to ensure C++
4305 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4306 GNU Compiler Collection (GCC) Internals}.)
4310 Remove all symbol table and relocation information from the executable.
4314 On systems that support dynamic linking, this prevents linking with the shared
4315 libraries. On other systems, this option has no effect.
4319 Produce a shared object which can then be linked with other objects to
4320 form an executable. Not all systems support this option. For predictable
4321 results, you must also specify the same set of options that were used to
4322 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4323 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4324 needs to build supplementary stub code for constructors to work. On
4325 multi-libbed systems, @samp{gcc -shared} must select the correct support
4326 libraries to link against. Failing to supply the correct flags may lead
4327 to subtle defects. Supplying them in cases where they are not necessary
4330 @item -shared-libgcc
4331 @itemx -static-libgcc
4332 @opindex shared-libgcc
4333 @opindex static-libgcc
4334 On systems that provide @file{libgcc} as a shared library, these options
4335 force the use of either the shared or static version respectively.
4336 If no shared version of @file{libgcc} was built when the compiler was
4337 configured, these options have no effect.
4339 There are several situations in which an application should use the
4340 shared @file{libgcc} instead of the static version. The most common
4341 of these is when the application wishes to throw and catch exceptions
4342 across different shared libraries. In that case, each of the libraries
4343 as well as the application itself should use the shared @file{libgcc}.
4345 Therefore, whenever you specify the @option{-shared} option, the GCC
4346 driver automatically adds @option{-shared-libgcc}, unless you explicitly
4347 specify @option{-static-libgcc}. The G++ driver automatically adds
4348 @option{-shared-libgcc} when you build a main executable as well because
4349 for C++ programs that is typically the right thing to do.
4350 (Exception-handling will not work reliably otherwise.)
4352 However, when linking a main executable written in C, you must
4353 explicitly say @option{-shared-libgcc} if you want to use the shared
4358 Bind references to global symbols when building a shared object. Warn
4359 about any unresolved references (unless overridden by the link editor
4360 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4363 @item -Xlinker @var{option}
4365 Pass @var{option} as an option to the linker. You can use this to
4366 supply system-specific linker options which GCC does not know how to
4369 If you want to pass an option that takes an argument, you must use
4370 @option{-Xlinker} twice, once for the option and once for the argument.
4371 For example, to pass @option{-assert definitions}, you must write
4372 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4373 @option{-Xlinker "-assert definitions"}, because this passes the entire
4374 string as a single argument, which is not what the linker expects.
4376 @item -Wl,@var{option}
4378 Pass @var{option} as an option to the linker. If @var{option} contains
4379 commas, it is split into multiple options at the commas.
4381 @item -u @var{symbol}
4383 Pretend the symbol @var{symbol} is undefined, to force linking of
4384 library modules to define it. You can use @option{-u} multiple times with
4385 different symbols to force loading of additional library modules.
4388 @node Directory Options
4389 @section Options for Directory Search
4390 @cindex directory options
4391 @cindex options, directory search
4394 These options specify directories to search for header files, for
4395 libraries and for parts of the compiler:
4400 Add the directory @var{dir} to the head of the list of directories to be
4401 searched for header files. This can be used to override a system header
4402 file, substituting your own version, since these directories are
4403 searched before the system header file directories. However, you should
4404 not use this option to add directories that contain vendor-supplied
4405 system header files (use @option{-isystem} for that). If you use more than
4406 one @option{-I} option, the directories are scanned in left-to-right
4407 order; the standard system directories come after.
4409 If a standard system include directory, or a directory specified with
4410 @option{-isystem}, is also specified with @option{-I}, it will be
4411 searched only in the position requested by @option{-I}. Also, it will
4412 not be considered a system include directory. If that directory really
4413 does contain system headers, there is a good chance that they will
4414 break. For instance, if GCC's installation procedure edited the headers
4415 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4416 original, buggy headers to be found instead of the corrected ones. GCC
4417 will issue a warning when a system include directory is hidden in this
4422 Any directories you specify with @option{-I} options before the @option{-I-}
4423 option are searched only for the case of @samp{#include "@var{file}"};
4424 they are not searched for @samp{#include <@var{file}>}.
4426 If additional directories are specified with @option{-I} options after
4427 the @option{-I-}, these directories are searched for all @samp{#include}
4428 directives. (Ordinarily @emph{all} @option{-I} directories are used
4431 In addition, the @option{-I-} option inhibits the use of the current
4432 directory (where the current input file came from) as the first search
4433 directory for @samp{#include "@var{file}"}. There is no way to
4434 override this effect of @option{-I-}. With @option{-I.} you can specify
4435 searching the directory which was current when the compiler was
4436 invoked. That is not exactly the same as what the preprocessor does
4437 by default, but it is often satisfactory.
4439 @option{-I-} does not inhibit the use of the standard system directories
4440 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4445 Add directory @var{dir} to the list of directories to be searched
4448 @item -B@var{prefix}
4450 This option specifies where to find the executables, libraries,
4451 include files, and data files of the compiler itself.
4453 The compiler driver program runs one or more of the subprograms
4454 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4455 @var{prefix} as a prefix for each program it tries to run, both with and
4456 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4458 For each subprogram to be run, the compiler driver first tries the
4459 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4460 was not specified, the driver tries two standard prefixes, which are
4461 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4462 those results in a file name that is found, the unmodified program
4463 name is searched for using the directories specified in your
4464 @env{PATH} environment variable.
4466 The compiler will check to see if the path provided by the @option{-B}
4467 refers to a directory, and if necessary it will add a directory
4468 separator character at the end of the path.
4470 @option{-B} prefixes that effectively specify directory names also apply
4471 to libraries in the linker, because the compiler translates these
4472 options into @option{-L} options for the linker. They also apply to
4473 includes files in the preprocessor, because the compiler translates these
4474 options into @option{-isystem} options for the preprocessor. In this case,
4475 the compiler appends @samp{include} to the prefix.
4477 The run-time support file @file{libgcc.a} can also be searched for using
4478 the @option{-B} prefix, if needed. If it is not found there, the two
4479 standard prefixes above are tried, and that is all. The file is left
4480 out of the link if it is not found by those means.
4482 Another way to specify a prefix much like the @option{-B} prefix is to use
4483 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4486 As a special kludge, if the path provided by @option{-B} is
4487 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4488 9, then it will be replaced by @file{[dir/]include}. This is to help
4489 with boot-strapping the compiler.
4491 @item -specs=@var{file}
4493 Process @var{file} after the compiler reads in the standard @file{specs}
4494 file, in order to override the defaults that the @file{gcc} driver
4495 program uses when determining what switches to pass to @file{cc1},
4496 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4497 @option{-specs=@var{file}} can be specified on the command line, and they
4498 are processed in order, from left to right.
4504 @section Specifying subprocesses and the switches to pass to them
4506 @command{gcc} is a driver program. It performs its job by invoking a
4507 sequence of other programs to do the work of compiling, assembling and
4508 linking. GCC interprets its command-line parameters and uses these to
4509 deduce which programs it should invoke, and which command-line options
4510 it ought to place on their command lines. This behavior is controlled
4511 by @dfn{spec strings}. In most cases there is one spec string for each
4512 program that GCC can invoke, but a few programs have multiple spec
4513 strings to control their behavior. The spec strings built into GCC can
4514 be overridden by using the @option{-specs=} command-line switch to specify
4517 @dfn{Spec files} are plaintext files that are used to construct spec
4518 strings. They consist of a sequence of directives separated by blank
4519 lines. The type of directive is determined by the first non-whitespace
4520 character on the line and it can be one of the following:
4523 @item %@var{command}
4524 Issues a @var{command} to the spec file processor. The commands that can
4528 @item %include <@var{file}>
4530 Search for @var{file} and insert its text at the current point in the
4533 @item %include_noerr <@var{file}>
4534 @cindex %include_noerr
4535 Just like @samp{%include}, but do not generate an error message if the include
4536 file cannot be found.
4538 @item %rename @var{old_name} @var{new_name}
4540 Rename the spec string @var{old_name} to @var{new_name}.
4544 @item *[@var{spec_name}]:
4545 This tells the compiler to create, override or delete the named spec
4546 string. All lines after this directive up to the next directive or
4547 blank line are considered to be the text for the spec string. If this
4548 results in an empty string then the spec will be deleted. (Or, if the
4549 spec did not exist, then nothing will happened.) Otherwise, if the spec
4550 does not currently exist a new spec will be created. If the spec does
4551 exist then its contents will be overridden by the text of this
4552 directive, unless the first character of that text is the @samp{+}
4553 character, in which case the text will be appended to the spec.
4555 @item [@var{suffix}]:
4556 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4557 and up to the next directive or blank line are considered to make up the
4558 spec string for the indicated suffix. When the compiler encounters an
4559 input file with the named suffix, it will processes the spec string in
4560 order to work out how to compile that file. For example:
4567 This says that any input file whose name ends in @samp{.ZZ} should be
4568 passed to the program @samp{z-compile}, which should be invoked with the
4569 command-line switch @option{-input} and with the result of performing the
4570 @samp{%i} substitution. (See below.)
4572 As an alternative to providing a spec string, the text that follows a
4573 suffix directive can be one of the following:
4576 @item @@@var{language}
4577 This says that the suffix is an alias for a known @var{language}. This is
4578 similar to using the @option{-x} command-line switch to GCC to specify a
4579 language explicitly. For example:
4586 Says that .ZZ files are, in fact, C++ source files.
4589 This causes an error messages saying:
4592 @var{name} compiler not installed on this system.
4596 GCC already has an extensive list of suffixes built into it.
4597 This directive will add an entry to the end of the list of suffixes, but
4598 since the list is searched from the end backwards, it is effectively
4599 possible to override earlier entries using this technique.
4603 GCC has the following spec strings built into it. Spec files can
4604 override these strings or create their own. Note that individual
4605 targets can also add their own spec strings to this list.
4608 asm Options to pass to the assembler
4609 asm_final Options to pass to the assembler post-processor
4610 cpp Options to pass to the C preprocessor
4611 cc1 Options to pass to the C compiler
4612 cc1plus Options to pass to the C++ compiler
4613 endfile Object files to include at the end of the link
4614 link Options to pass to the linker
4615 lib Libraries to include on the command line to the linker
4616 libgcc Decides which GCC support library to pass to the linker
4617 linker Sets the name of the linker
4618 predefines Defines to be passed to the C preprocessor
4619 signed_char Defines to pass to CPP to say whether @code{char} is signed
4621 startfile Object files to include at the start of the link
4624 Here is a small example of a spec file:
4630 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4633 This example renames the spec called @samp{lib} to @samp{old_lib} and
4634 then overrides the previous definition of @samp{lib} with a new one.
4635 The new definition adds in some extra command-line options before
4636 including the text of the old definition.
4638 @dfn{Spec strings} are a list of command-line options to be passed to their
4639 corresponding program. In addition, the spec strings can contain
4640 @samp{%}-prefixed sequences to substitute variable text or to
4641 conditionally insert text into the command line. Using these constructs
4642 it is possible to generate quite complex command lines.
4644 Here is a table of all defined @samp{%}-sequences for spec
4645 strings. Note that spaces are not generated automatically around the
4646 results of expanding these sequences. Therefore you can concatenate them
4647 together or combine them with constant text in a single argument.
4651 Substitute one @samp{%} into the program name or argument.
4654 Substitute the name of the input file being processed.
4657 Substitute the basename of the input file being processed.
4658 This is the substring up to (and not including) the last period
4659 and not including the directory.
4662 This is the same as @samp{%b}, but include the file suffix (text after
4666 Marks the argument containing or following the @samp{%d} as a
4667 temporary file name, so that that file will be deleted if GCC exits
4668 successfully. Unlike @samp{%g}, this contributes no text to the
4671 @item %g@var{suffix}
4672 Substitute a file name that has suffix @var{suffix} and is chosen
4673 once per compilation, and mark the argument in the same way as
4674 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4675 name is now chosen in a way that is hard to predict even when previously
4676 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4677 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4678 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4679 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4680 was simply substituted with a file name chosen once per compilation,
4681 without regard to any appended suffix (which was therefore treated
4682 just like ordinary text), making such attacks more likely to succeed.
4684 @item %u@var{suffix}
4685 Like @samp{%g}, but generates a new temporary file name even if
4686 @samp{%u@var{suffix}} was already seen.
4688 @item %U@var{suffix}
4689 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4690 new one if there is no such last file name. In the absence of any
4691 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4692 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4693 would involve the generation of two distinct file names, one
4694 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4695 simply substituted with a file name chosen for the previous @samp{%u},
4696 without regard to any appended suffix.
4698 @item %j@var{SUFFIX}
4699 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4700 writable, and if save-temps is off; otherwise, substitute the name
4701 of a temporary file, just like @samp{%u}. This temporary file is not
4702 meant for communication between processes, but rather as a junk
4705 @item %.@var{SUFFIX}
4706 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4707 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4708 terminated by the next space or %.
4711 Marks the argument containing or following the @samp{%w} as the
4712 designated output file of this compilation. This puts the argument
4713 into the sequence of arguments that @samp{%o} will substitute later.
4716 Substitutes the names of all the output files, with spaces
4717 automatically placed around them. You should write spaces
4718 around the @samp{%o} as well or the results are undefined.
4719 @samp{%o} is for use in the specs for running the linker.
4720 Input files whose names have no recognized suffix are not compiled
4721 at all, but they are included among the output files, so they will
4725 Substitutes the suffix for object files. Note that this is
4726 handled specially when it immediately follows @samp{%g, %u, or %U},
4727 because of the need for those to form complete file names. The
4728 handling is such that @samp{%O} is treated exactly as if it had already
4729 been substituted, except that @samp{%g, %u, and %U} do not currently
4730 support additional @var{suffix} characters following @samp{%O} as they would
4731 following, for example, @samp{.o}.
4734 Substitutes the standard macro predefinitions for the
4735 current target machine. Use this when running @code{cpp}.
4738 Like @samp{%p}, but puts @samp{__} before and after the name of each
4739 predefined macro, except for macros that start with @samp{__} or with
4740 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4744 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4747 Current argument is the name of a library or startup file of some sort.
4748 Search for that file in a standard list of directories and substitute
4749 the full name found.
4752 Print @var{str} as an error message. @var{str} is terminated by a newline.
4753 Use this when inconsistent options are detected.
4756 Output @samp{-} if the input for the current command is coming from a pipe.
4759 Substitute the contents of spec string @var{name} at this point.
4762 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4764 @item %x@{@var{option}@}
4765 Accumulate an option for @samp{%X}.
4768 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4772 Output the accumulated assembler options specified by @option{-Wa}.
4775 Output the accumulated preprocessor options specified by @option{-Wp}.
4778 Substitute the major version number of GCC@.
4779 (For version 2.9.5, this is 2.)
4782 Substitute the minor version number of GCC@.
4783 (For version 2.9.5, this is 9.)
4786 Substitute the patch level number of GCC@.
4787 (For version 2.9.5, this is 5.)
4790 Process the @code{asm} spec. This is used to compute the
4791 switches to be passed to the assembler.
4794 Process the @code{asm_final} spec. This is a spec string for
4795 passing switches to an assembler post-processor, if such a program is
4799 Process the @code{link} spec. This is the spec for computing the
4800 command line passed to the linker. Typically it will make use of the
4801 @samp{%L %G %S %D and %E} sequences.
4804 Dump out a @option{-L} option for each directory that GCC believes might
4805 contain startup files. If the target supports multilibs then the
4806 current multilib directory will be prepended to each of these paths.
4809 Output the multilib directory with directory separators replaced with
4810 @samp{_}. If multilib directories are not set, or the multilib directory is
4811 @file{.} then this option emits nothing.
4814 Process the @code{lib} spec. This is a spec string for deciding which
4815 libraries should be included on the command line to the linker.
4818 Process the @code{libgcc} spec. This is a spec string for deciding
4819 which GCC support library should be included on the command line to the linker.
4822 Process the @code{startfile} spec. This is a spec for deciding which
4823 object files should be the first ones passed to the linker. Typically
4824 this might be a file named @file{crt0.o}.
4827 Process the @code{endfile} spec. This is a spec string that specifies
4828 the last object files that will be passed to the linker.
4831 Process the @code{cpp} spec. This is used to construct the arguments
4832 to be passed to the C preprocessor.
4835 Process the @code{signed_char} spec. This is intended to be used
4836 to tell cpp whether a char is signed. It typically has the definition:
4838 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4842 Process the @code{cc1} spec. This is used to construct the options to be
4843 passed to the actual C compiler (@samp{cc1}).
4846 Process the @code{cc1plus} spec. This is used to construct the options to be
4847 passed to the actual C++ compiler (@samp{cc1plus}).
4850 Substitute the variable part of a matched option. See below.
4851 Note that each comma in the substituted string is replaced by
4855 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4856 If that switch was not specified, this substitutes nothing. Note that
4857 the leading dash is omitted when specifying this option, and it is
4858 automatically inserted if the substitution is performed. Thus the spec
4859 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4860 and would output the command line option @option{-foo}.
4862 @item %W@{@code{S}@}
4863 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4866 @item %@{@code{S}*@}
4867 Substitutes all the switches specified to GCC whose names start
4868 with @code{-S}, but which also take an argument. This is used for
4869 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4870 GCC considers @option{-o foo} as being
4871 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4872 text, including the space. Thus two arguments would be generated.
4874 @item %@{^@code{S}*@}
4875 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4876 argument. Thus %@{^o*@} would only generate one argument, not two.
4878 @item %@{@code{S}*&@code{T}*@}
4879 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4880 (the order of @code{S} and @code{T} in the spec is not significant).
4881 There can be any number of ampersand-separated variables; for each the
4882 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4884 @item %@{<@code{S}@}
4885 Remove all occurrences of @code{-S} from the command line. Note---this
4886 command is position dependent. @samp{%} commands in the spec string
4887 before this option will see @code{-S}, @samp{%} commands in the spec
4888 string after this option will not.
4890 @item %@{@code{S}*:@code{X}@}
4891 Substitutes @code{X} if one or more switches whose names start with
4892 @code{-S} are specified to GCC@. Note that the tail part of the
4893 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4894 for each occurrence of @samp{%*} within @code{X}.
4896 @item %@{@code{S}:@code{X}@}
4897 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4899 @item %@{!@code{S}:@code{X}@}
4900 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4902 @item %@{|@code{S}:@code{X}@}
4903 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4905 @item %@{|!@code{S}:@code{X}@}
4906 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4908 @item %@{.@code{S}:@code{X}@}
4909 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4911 @item %@{!.@code{S}:@code{X}@}
4912 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4914 @item %@{@code{S}|@code{P}:@code{X}@}
4915 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4916 combined with @samp{!} and @samp{.} sequences as well, although they
4917 have a stronger binding than the @samp{|}. For example a spec string
4921 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4924 will output the following command-line options from the following input
4925 command-line options:
4930 -d fred.c -foo -baz -boggle
4931 -d jim.d -bar -baz -boggle
4936 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4937 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4938 or spaces, or even newlines. They are processed as usual, as described
4941 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4942 switches are handled specifically in these
4943 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4944 @option{-W} switch is found later in the command line, the earlier switch
4945 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4946 letter, which passes all matching options.
4948 The character @samp{|} at the beginning of the predicate text is used to indicate
4949 that a command should be piped to the following command, but only if @option{-pipe}
4952 It is built into GCC which switches take arguments and which do not.
4953 (You might think it would be useful to generalize this to allow each
4954 compiler's spec to say which switches take arguments. But this cannot
4955 be done in a consistent fashion. GCC cannot even decide which input
4956 files have been specified without knowing which switches take arguments,
4957 and it must know which input files to compile in order to tell which
4960 GCC also knows implicitly that arguments starting in @option{-l} are to be
4961 treated as compiler output files, and passed to the linker in their
4962 proper position among the other output files.
4964 @c man begin OPTIONS
4966 @node Target Options
4967 @section Specifying Target Machine and Compiler Version
4968 @cindex target options
4969 @cindex cross compiling
4970 @cindex specifying machine version
4971 @cindex specifying compiler version and target machine
4972 @cindex compiler version, specifying
4973 @cindex target machine, specifying
4975 By default, GCC compiles code for the same type of machine that you
4976 are using. However, it can also be installed as a cross-compiler, to
4977 compile for some other type of machine. In fact, several different
4978 configurations of GCC, for different target machines, can be
4979 installed side by side. Then you specify which one to use with the
4982 In addition, older and newer versions of GCC can be installed side
4983 by side. One of them (probably the newest) will be the default, but
4984 you may sometimes wish to use another.
4987 @item -b @var{machine}
4989 The argument @var{machine} specifies the target machine for compilation.
4990 This is useful when you have installed GCC as a cross-compiler.
4992 The value to use for @var{machine} is the same as was specified as the
4993 machine type when configuring GCC as a cross-compiler. For
4994 example, if a cross-compiler was configured with @samp{configure
4995 i386v}, meaning to compile for an 80386 running System V, then you
4996 would specify @option{-b i386v} to run that cross compiler.
4998 When you do not specify @option{-b}, it normally means to compile for
4999 the same type of machine that you are using.
5001 @item -V @var{version}
5003 The argument @var{version} specifies which version of GCC to run.
5004 This is useful when multiple versions are installed. For example,
5005 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5007 The default version, when you do not specify @option{-V}, is the last
5008 version of GCC that you installed.
5011 The @option{-b} and @option{-V} options actually work by controlling part of
5012 the file name used for the executable files and libraries used for
5013 compilation. A given version of GCC, for a given target machine, is
5014 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
5016 Thus, sites can customize the effect of @option{-b} or @option{-V} either by
5017 changing the names of these directories or adding alternate names (or
5018 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
5019 file @file{80386} is a link to the file @file{i386v}, then @option{-b
5020 80386} becomes an alias for @option{-b i386v}.
5022 In one respect, the @option{-b} or @option{-V} do not completely change
5023 to a different compiler: the top-level driver program @command{gcc}
5024 that you originally invoked continues to run and invoke the other
5025 executables (preprocessor, compiler per se, assembler and linker)
5026 that do the real work. However, since no real work is done in the
5027 driver program, it usually does not matter that the driver program
5028 in use is not the one for the specified target. It is common for the
5029 interface to the other executables to change incompatibly between
5030 compiler versions, so unless the version specified is very close to that
5031 of the driver (for example, @option{-V 3.0} with a driver program from GCC
5032 version 3.0.1), use of @option{-V} may not work; for example, using
5033 @option{-V 2.95.2} will not work with a driver program from GCC 3.0.
5035 The only way that the driver program depends on the target machine is
5036 in the parsing and handling of special machine-specific options.
5037 However, this is controlled by a file which is found, along with the
5038 other executables, in the directory for the specified version and
5039 target machine. As a result, a single installed driver program adapts
5040 to any specified target machine, and sufficiently similar compiler
5043 The driver program executable does control one significant thing,
5044 however: the default version and target machine. Therefore, you can
5045 install different instances of the driver program, compiled for
5046 different targets or versions, under different names.
5048 For example, if the driver for version 2.0 is installed as @command{ogcc}
5049 and that for version 2.1 is installed as @command{gcc}, then the command
5050 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
5051 2.0 by default. However, you can choose either version with either
5052 command with the @option{-V} option.
5054 @node Submodel Options
5055 @section Hardware Models and Configurations
5056 @cindex submodel options
5057 @cindex specifying hardware config
5058 @cindex hardware models and configurations, specifying
5059 @cindex machine dependent options
5061 Earlier we discussed the standard option @option{-b} which chooses among
5062 different installed compilers for completely different target
5063 machines, such as VAX vs.@: 68000 vs.@: 80386.
5065 In addition, each of these target machine types can have its own
5066 special options, starting with @samp{-m}, to choose among various
5067 hardware models or configurations---for example, 68010 vs 68020,
5068 floating coprocessor or none. A single installed version of the
5069 compiler can compile for any model or configuration, according to the
5072 Some configurations of the compiler also support additional special
5073 options, usually for compatibility with other compilers on the same
5076 These options are defined by the macro @code{TARGET_SWITCHES} in the
5077 machine description. The default for the options is also defined by
5078 that macro, which enables you to change the defaults.
5092 * RS/6000 and PowerPC Options::
5095 * i386 and x86-64 Options::
5097 * Intel 960 Options::
5098 * DEC Alpha Options::
5099 * DEC Alpha/VMS Options::
5103 * System V Options::
5104 * TMS320C3x/C4x Options::
5112 * S/390 and zSeries Options::
5116 * Xstormy16 Options::
5119 @node M680x0 Options
5120 @subsection M680x0 Options
5121 @cindex M680x0 options
5123 These are the @samp{-m} options defined for the 68000 series. The default
5124 values for these options depends on which style of 68000 was selected when
5125 the compiler was configured; the defaults for the most common choices are
5133 Generate output for a 68000. This is the default
5134 when the compiler is configured for 68000-based systems.
5136 Use this option for microcontrollers with a 68000 or EC000 core,
5137 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5143 Generate output for a 68020. This is the default
5144 when the compiler is configured for 68020-based systems.
5148 Generate output containing 68881 instructions for floating point.
5149 This is the default for most 68020 systems unless @option{--nfp} was
5150 specified when the compiler was configured.
5154 Generate output for a 68030. This is the default when the compiler is
5155 configured for 68030-based systems.
5159 Generate output for a 68040. This is the default when the compiler is
5160 configured for 68040-based systems.
5162 This option inhibits the use of 68881/68882 instructions that have to be
5163 emulated by software on the 68040. Use this option if your 68040 does not
5164 have code to emulate those instructions.
5168 Generate output for a 68060. This is the default when the compiler is
5169 configured for 68060-based systems.
5171 This option inhibits the use of 68020 and 68881/68882 instructions that
5172 have to be emulated by software on the 68060. Use this option if your 68060
5173 does not have code to emulate those instructions.
5177 Generate output for a CPU32. This is the default
5178 when the compiler is configured for CPU32-based systems.
5180 Use this option for microcontrollers with a
5181 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5182 68336, 68340, 68341, 68349 and 68360.
5186 Generate output for a 520X ``coldfire'' family cpu. This is the default
5187 when the compiler is configured for 520X-based systems.
5189 Use this option for microcontroller with a 5200 core, including
5190 the MCF5202, MCF5203, MCF5204 and MCF5202.
5195 Generate output for a 68040, without using any of the new instructions.
5196 This results in code which can run relatively efficiently on either a
5197 68020/68881 or a 68030 or a 68040. The generated code does use the
5198 68881 instructions that are emulated on the 68040.
5202 Generate output for a 68060, without using any of the new instructions.
5203 This results in code which can run relatively efficiently on either a
5204 68020/68881 or a 68030 or a 68040. The generated code does use the
5205 68881 instructions that are emulated on the 68060.
5209 Generate output containing Sun FPA instructions for floating point.
5212 @opindex msoft-float
5213 Generate output containing library calls for floating point.
5214 @strong{Warning:} the requisite libraries are not available for all m68k
5215 targets. Normally the facilities of the machine's usual C compiler are
5216 used, but this can't be done directly in cross-compilation. You must
5217 make your own arrangements to provide suitable library functions for
5218 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5219 @samp{m68k-*-coff} do provide software floating point support.
5223 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5226 @opindex mnobitfield
5227 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5228 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5232 Do use the bit-field instructions. The @option{-m68020} option implies
5233 @option{-mbitfield}. This is the default if you use a configuration
5234 designed for a 68020.
5238 Use a different function-calling convention, in which functions
5239 that take a fixed number of arguments return with the @code{rtd}
5240 instruction, which pops their arguments while returning. This
5241 saves one instruction in the caller since there is no need to pop
5242 the arguments there.
5244 This calling convention is incompatible with the one normally
5245 used on Unix, so you cannot use it if you need to call libraries
5246 compiled with the Unix compiler.
5248 Also, you must provide function prototypes for all functions that
5249 take variable numbers of arguments (including @code{printf});
5250 otherwise incorrect code will be generated for calls to those
5253 In addition, seriously incorrect code will result if you call a
5254 function with too many arguments. (Normally, extra arguments are
5255 harmlessly ignored.)
5257 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5258 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5261 @itemx -mno-align-int
5263 @opindex mno-align-int
5264 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5265 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5266 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5267 Aligning variables on 32-bit boundaries produces code that runs somewhat
5268 faster on processors with 32-bit busses at the expense of more memory.
5270 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5271 align structures containing the above types differently than
5272 most published application binary interface specifications for the m68k.
5276 Use the pc-relative addressing mode of the 68000 directly, instead of
5277 using a global offset table. At present, this option implies @option{-fpic},
5278 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5279 not presently supported with @option{-mpcrel}, though this could be supported for
5280 68020 and higher processors.
5282 @item -mno-strict-align
5283 @itemx -mstrict-align
5284 @opindex mno-strict-align
5285 @opindex mstrict-align
5286 Do not (do) assume that unaligned memory references will be handled by
5291 @node M68hc1x Options
5292 @subsection M68hc1x Options
5293 @cindex M68hc1x options
5295 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5296 microcontrollers. The default values for these options depends on
5297 which style of microcontroller was selected when the compiler was configured;
5298 the defaults for the most common choices are given below.
5305 Generate output for a 68HC11. This is the default
5306 when the compiler is configured for 68HC11-based systems.
5312 Generate output for a 68HC12. This is the default
5313 when the compiler is configured for 68HC12-based systems.
5316 @opindex mauto-incdec
5317 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5322 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5324 @item -msoft-reg-count=@var{count}
5325 @opindex msoft-reg-count
5326 Specify the number of pseudo-soft registers which are used for the
5327 code generation. The maximum number is 32. Using more pseudo-soft
5328 register may or may not result in better code depending on the program.
5329 The default is 4 for 68HC11 and 2 for 68HC12.
5334 @subsection VAX Options
5337 These @samp{-m} options are defined for the VAX:
5342 Do not output certain jump instructions (@code{aobleq} and so on)
5343 that the Unix assembler for the VAX cannot handle across long
5348 Do output those jump instructions, on the assumption that you
5349 will assemble with the GNU assembler.
5353 Output code for g-format floating point numbers instead of d-format.
5357 @subsection SPARC Options
5358 @cindex SPARC options
5360 These @samp{-m} switches are supported on the SPARC:
5365 @opindex mno-app-regs
5367 Specify @option{-mapp-regs} to generate output using the global registers
5368 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5371 To be fully SVR4 ABI compliant at the cost of some performance loss,
5372 specify @option{-mno-app-regs}. You should compile libraries and system
5373 software with this option.
5378 @opindex mhard-float
5379 Generate output containing floating point instructions. This is the
5385 @opindex msoft-float
5386 Generate output containing library calls for floating point.
5387 @strong{Warning:} the requisite libraries are not available for all SPARC
5388 targets. Normally the facilities of the machine's usual C compiler are
5389 used, but this cannot be done directly in cross-compilation. You must make
5390 your own arrangements to provide suitable library functions for
5391 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5392 @samp{sparclite-*-*} do provide software floating point support.
5394 @option{-msoft-float} changes the calling convention in the output file;
5395 therefore, it is only useful if you compile @emph{all} of a program with
5396 this option. In particular, you need to compile @file{libgcc.a}, the
5397 library that comes with GCC, with @option{-msoft-float} in order for
5400 @item -mhard-quad-float
5401 @opindex mhard-quad-float
5402 Generate output containing quad-word (long double) floating point
5405 @item -msoft-quad-float
5406 @opindex msoft-quad-float
5407 Generate output containing library calls for quad-word (long double)
5408 floating point instructions. The functions called are those specified
5409 in the SPARC ABI@. This is the default.
5411 As of this writing, there are no sparc implementations that have hardware
5412 support for the quad-word floating point instructions. They all invoke
5413 a trap handler for one of these instructions, and then the trap handler
5414 emulates the effect of the instruction. Because of the trap handler overhead,
5415 this is much slower than calling the ABI library routines. Thus the
5416 @option{-msoft-quad-float} option is the default.
5420 @opindex mno-epilogue
5422 With @option{-mepilogue} (the default), the compiler always emits code for
5423 function exit at the end of each function. Any function exit in
5424 the middle of the function (such as a return statement in C) will
5425 generate a jump to the exit code at the end of the function.
5427 With @option{-mno-epilogue}, the compiler tries to emit exit code inline
5428 at every function exit.
5434 With @option{-mflat}, the compiler does not generate save/restore instructions
5435 and will use a ``flat'' or single register window calling convention.
5436 This model uses %i7 as the frame pointer and is compatible with the normal
5437 register window model. Code from either may be intermixed.
5438 The local registers and the input registers (0--5) are still treated as
5439 ``call saved'' registers and will be saved on the stack as necessary.
5441 With @option{-mno-flat} (the default), the compiler emits save/restore
5442 instructions (except for leaf functions) and is the normal mode of operation.
5444 @item -mno-unaligned-doubles
5445 @itemx -munaligned-doubles
5446 @opindex mno-unaligned-doubles
5447 @opindex munaligned-doubles
5448 Assume that doubles have 8 byte alignment. This is the default.
5450 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5451 alignment only if they are contained in another type, or if they have an
5452 absolute address. Otherwise, it assumes they have 4 byte alignment.
5453 Specifying this option avoids some rare compatibility problems with code
5454 generated by other compilers. It is not the default because it results
5455 in a performance loss, especially for floating point code.
5457 @item -mno-faster-structs
5458 @itemx -mfaster-structs
5459 @opindex mno-faster-structs
5460 @opindex mfaster-structs
5461 With @option{-mfaster-structs}, the compiler assumes that structures
5462 should have 8 byte alignment. This enables the use of pairs of
5463 @code{ldd} and @code{std} instructions for copies in structure
5464 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5465 However, the use of this changed alignment directly violates the Sparc
5466 ABI@. Thus, it's intended only for use on targets where the developer
5467 acknowledges that their resulting code will not be directly in line with
5468 the rules of the ABI@.
5474 These two options select variations on the SPARC architecture.
5476 By default (unless specifically configured for the Fujitsu SPARClite),
5477 GCC generates code for the v7 variant of the SPARC architecture.
5479 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5480 code is that the compiler emits the integer multiply and integer
5481 divide instructions which exist in SPARC v8 but not in SPARC v7.
5483 @option{-msparclite} will give you SPARClite code. This adds the integer
5484 multiply, integer divide step and scan (@code{ffs}) instructions which
5485 exist in SPARClite but not in SPARC v7.
5487 These options are deprecated and will be deleted in a future GCC release.
5488 They have been replaced with @option{-mcpu=xxx}.
5493 @opindex msupersparc
5494 These two options select the processor for which the code is optimized.
5496 With @option{-mcypress} (the default), the compiler optimizes code for the
5497 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5498 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5500 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5501 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5502 of the full SPARC v8 instruction set.
5504 These options are deprecated and will be deleted in a future GCC release.
5505 They have been replaced with @option{-mcpu=xxx}.
5507 @item -mcpu=@var{cpu_type}
5509 Set the instruction set, register set, and instruction scheduling parameters
5510 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5511 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5512 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5513 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5515 Default instruction scheduling parameters are used for values that select
5516 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5517 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5519 Here is a list of each supported architecture and their supported
5524 v8: supersparc, hypersparc
5525 sparclite: f930, f934, sparclite86x
5530 @item -mtune=@var{cpu_type}
5532 Set the instruction scheduling parameters for machine type
5533 @var{cpu_type}, but do not set the instruction set or register set that the
5534 option @option{-mcpu=@var{cpu_type}} would.
5536 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5537 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5538 that select a particular cpu implementation. Those are @samp{cypress},
5539 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5540 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5544 These @samp{-m} switches are supported in addition to the above
5545 on the SPARCLET processor.
5548 @item -mlittle-endian
5549 @opindex mlittle-endian
5550 Generate code for a processor running in little-endian mode.
5554 Treat register @code{%g0} as a normal register.
5555 GCC will continue to clobber it as necessary but will not assume
5556 it always reads as 0.
5558 @item -mbroken-saverestore
5559 @opindex mbroken-saverestore
5560 Generate code that does not use non-trivial forms of the @code{save} and
5561 @code{restore} instructions. Early versions of the SPARCLET processor do
5562 not correctly handle @code{save} and @code{restore} instructions used with
5563 arguments. They correctly handle them used without arguments. A @code{save}
5564 instruction used without arguments increments the current window pointer
5565 but does not allocate a new stack frame. It is assumed that the window
5566 overflow trap handler will properly handle this case as will interrupt
5570 These @samp{-m} switches are supported in addition to the above
5571 on SPARC V9 processors in 64-bit environments.
5574 @item -mlittle-endian
5575 @opindex mlittle-endian
5576 Generate code for a processor running in little-endian mode.
5582 Generate code for a 32-bit or 64-bit environment.
5583 The 32-bit environment sets int, long and pointer to 32 bits.
5584 The 64-bit environment sets int to 32 bits and long and pointer
5587 @item -mcmodel=medlow
5588 @opindex mcmodel=medlow
5589 Generate code for the Medium/Low code model: the program must be linked
5590 in the low 32 bits of the address space. Pointers are 64 bits.
5591 Programs can be statically or dynamically linked.
5593 @item -mcmodel=medmid
5594 @opindex mcmodel=medmid
5595 Generate code for the Medium/Middle code model: the program must be linked
5596 in the low 44 bits of the address space, the text segment must be less than
5597 2G bytes, and data segment must be within 2G of the text segment.
5598 Pointers are 64 bits.
5600 @item -mcmodel=medany
5601 @opindex mcmodel=medany
5602 Generate code for the Medium/Anywhere code model: the program may be linked
5603 anywhere in the address space, the text segment must be less than
5604 2G bytes, and data segment must be within 2G of the text segment.
5605 Pointers are 64 bits.
5607 @item -mcmodel=embmedany
5608 @opindex mcmodel=embmedany
5609 Generate code for the Medium/Anywhere code model for embedded systems:
5610 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5611 (determined at link time). Register %g4 points to the base of the
5612 data segment. Pointers are still 64 bits.
5613 Programs are statically linked, PIC is not supported.
5616 @itemx -mno-stack-bias
5617 @opindex mstack-bias
5618 @opindex mno-stack-bias
5619 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5620 frame pointer if present, are offset by @minus{}2047 which must be added back
5621 when making stack frame references.
5622 Otherwise, assume no such offset is present.
5625 @node Convex Options
5626 @subsection Convex Options
5627 @cindex Convex options
5629 These @samp{-m} options are defined for Convex:
5634 Generate output for C1. The code will run on any Convex machine.
5635 The preprocessor symbol @code{__convex__c1__} is defined.
5639 Generate output for C2. Uses instructions not available on C1.
5640 Scheduling and other optimizations are chosen for max performance on C2.
5641 The preprocessor symbol @code{__convex_c2__} is defined.
5645 Generate output for C32xx. Uses instructions not available on C1.
5646 Scheduling and other optimizations are chosen for max performance on C32.
5647 The preprocessor symbol @code{__convex_c32__} is defined.
5651 Generate output for C34xx. Uses instructions not available on C1.
5652 Scheduling and other optimizations are chosen for max performance on C34.
5653 The preprocessor symbol @code{__convex_c34__} is defined.
5657 Generate output for C38xx. Uses instructions not available on C1.
5658 Scheduling and other optimizations are chosen for max performance on C38.
5659 The preprocessor symbol @code{__convex_c38__} is defined.
5663 Generate code which puts an argument count in the word preceding each
5664 argument list. This is compatible with regular CC, and a few programs
5665 may need the argument count word. GDB and other source-level debuggers
5666 do not need it; this info is in the symbol table.
5669 @opindex mnoargcount
5670 Omit the argument count word. This is the default.
5672 @item -mvolatile-cache
5673 @opindex mvolatile-cache
5674 Allow volatile references to be cached. This is the default.
5676 @item -mvolatile-nocache
5677 @opindex mvolatile-nocache
5678 Volatile references bypass the data cache, going all the way to memory.
5679 This is only needed for multi-processor code that does not use standard
5680 synchronization instructions. Making non-volatile references to volatile
5681 locations will not necessarily work.
5685 Type long is 32 bits, the same as type int. This is the default.
5689 Type long is 64 bits, the same as type long long. This option is useless,
5690 because no library support exists for it.
5693 @node AMD29K Options
5694 @subsection AMD29K Options
5695 @cindex AMD29K options
5697 These @samp{-m} options are defined for the AMD Am29000:
5702 @cindex DW bit (29k)
5703 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5704 halfword operations are directly supported by the hardware. This is the
5709 Generate code that assumes the @code{DW} bit is not set.
5713 @cindex byte writes (29k)
5714 Generate code that assumes the system supports byte and halfword write
5715 operations. This is the default.
5719 Generate code that assumes the systems does not support byte and
5720 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5724 @cindex memory model (29k)
5725 Use a small memory model that assumes that all function addresses are
5726 either within a single 256 KB segment or at an absolute address of less
5727 than 256k. This allows the @code{call} instruction to be used instead
5728 of a @code{const}, @code{consth}, @code{calli} sequence.
5732 Use the normal memory model: Generate @code{call} instructions only when
5733 calling functions in the same file and @code{calli} instructions
5734 otherwise. This works if each file occupies less than 256 KB but allows
5735 the entire executable to be larger than 256 KB@. This is the default.
5739 Always use @code{calli} instructions. Specify this option if you expect
5740 a single file to compile into more than 256 KB of code.
5744 @cindex processor selection (29k)
5745 Generate code for the Am29050.
5749 Generate code for the Am29000. This is the default.
5751 @item -mkernel-registers
5752 @opindex mkernel-registers
5753 @cindex kernel and user registers (29k)
5754 Generate references to registers @code{gr64-gr95} instead of to
5755 registers @code{gr96-gr127}. This option can be used when compiling
5756 kernel code that wants a set of global registers disjoint from that used
5759 Note that when this option is used, register names in @samp{-f} flags
5760 must use the normal, user-mode, names.
5762 @item -muser-registers
5763 @opindex muser-registers
5764 Use the normal set of global registers, @code{gr96-gr127}. This is the
5768 @itemx -mno-stack-check
5769 @opindex mstack-check
5770 @opindex mno-stack-check
5771 @cindex stack checks (29k)
5772 Insert (or do not insert) a call to @code{__msp_check} after each stack
5773 adjustment. This is often used for kernel code.
5776 @itemx -mno-storem-bug
5777 @opindex mstorem-bug
5778 @opindex mno-storem-bug
5779 @cindex storem bug (29k)
5780 @option{-mstorem-bug} handles 29k processors which cannot handle the
5781 separation of a mtsrim insn and a storem instruction (most 29000 chips
5782 to date, but not the 29050).
5784 @item -mno-reuse-arg-regs
5785 @itemx -mreuse-arg-regs
5786 @opindex mno-reuse-arg-regs
5787 @opindex mreuse-arg-regs
5788 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5789 registers for copying out arguments. This helps detect calling a function
5790 with fewer arguments than it was declared with.
5792 @item -mno-impure-text
5793 @itemx -mimpure-text
5794 @opindex mno-impure-text
5795 @opindex mimpure-text
5796 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5797 not pass @option{-assert pure-text} to the linker when linking a shared object.
5800 @opindex msoft-float
5801 Generate output containing library calls for floating point.
5802 @strong{Warning:} the requisite libraries are not part of GCC@.
5803 Normally the facilities of the machine's usual C compiler are used, but
5804 this can't be done directly in cross-compilation. You must make your
5805 own arrangements to provide suitable library functions for
5810 Do not generate multm or multmu instructions. This is useful for some embedded
5811 systems which do not have trap handlers for these instructions.
5815 @subsection ARM Options
5818 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5823 @opindex mapcs-frame
5824 Generate a stack frame that is compliant with the ARM Procedure Call
5825 Standard for all functions, even if this is not strictly necessary for
5826 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5827 with this option will cause the stack frames not to be generated for
5828 leaf functions. The default is @option{-mno-apcs-frame}.
5832 This is a synonym for @option{-mapcs-frame}.
5836 Generate code for a processor running with a 26-bit program counter,
5837 and conforming to the function calling standards for the APCS 26-bit
5838 option. This option replaces the @option{-m2} and @option{-m3} options
5839 of previous releases of the compiler.
5843 Generate code for a processor running with a 32-bit program counter,
5844 and conforming to the function calling standards for the APCS 32-bit
5845 option. This option replaces the @option{-m6} option of previous releases
5849 @c not currently implemented
5850 @item -mapcs-stack-check
5851 @opindex mapcs-stack-check
5852 Generate code to check the amount of stack space available upon entry to
5853 every function (that actually uses some stack space). If there is
5854 insufficient space available then either the function
5855 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5856 called, depending upon the amount of stack space required. The run time
5857 system is required to provide these functions. The default is
5858 @option{-mno-apcs-stack-check}, since this produces smaller code.
5860 @c not currently implemented
5862 @opindex mapcs-float
5863 Pass floating point arguments using the float point registers. This is
5864 one of the variants of the APCS@. This option is recommended if the
5865 target hardware has a floating point unit or if a lot of floating point
5866 arithmetic is going to be performed by the code. The default is
5867 @option{-mno-apcs-float}, since integer only code is slightly increased in
5868 size if @option{-mapcs-float} is used.
5870 @c not currently implemented
5871 @item -mapcs-reentrant
5872 @opindex mapcs-reentrant
5873 Generate reentrant, position independent code. The default is
5874 @option{-mno-apcs-reentrant}.
5877 @item -mthumb-interwork
5878 @opindex mthumb-interwork
5879 Generate code which supports calling between the ARM and Thumb
5880 instruction sets. Without this option the two instruction sets cannot
5881 be reliably used inside one program. The default is
5882 @option{-mno-thumb-interwork}, since slightly larger code is generated
5883 when @option{-mthumb-interwork} is specified.
5885 @item -mno-sched-prolog
5886 @opindex mno-sched-prolog
5887 Prevent the reordering of instructions in the function prolog, or the
5888 merging of those instruction with the instructions in the function's
5889 body. This means that all functions will start with a recognizable set
5890 of instructions (or in fact one of a choice from a small set of
5891 different function prologues), and this information can be used to
5892 locate the start if functions inside an executable piece of code. The
5893 default is @option{-msched-prolog}.
5896 @opindex mhard-float
5897 Generate output containing floating point instructions. This is the
5901 @opindex msoft-float
5902 Generate output containing library calls for floating point.
5903 @strong{Warning:} the requisite libraries are not available for all ARM
5904 targets. Normally the facilities of the machine's usual C compiler are
5905 used, but this cannot be done directly in cross-compilation. You must make
5906 your own arrangements to provide suitable library functions for
5909 @option{-msoft-float} changes the calling convention in the output file;
5910 therefore, it is only useful if you compile @emph{all} of a program with
5911 this option. In particular, you need to compile @file{libgcc.a}, the
5912 library that comes with GCC, with @option{-msoft-float} in order for
5915 @item -mlittle-endian
5916 @opindex mlittle-endian
5917 Generate code for a processor running in little-endian mode. This is
5918 the default for all standard configurations.
5921 @opindex mbig-endian
5922 Generate code for a processor running in big-endian mode; the default is
5923 to compile code for a little-endian processor.
5925 @item -mwords-little-endian
5926 @opindex mwords-little-endian
5927 This option only applies when generating code for big-endian processors.
5928 Generate code for a little-endian word order but a big-endian byte
5929 order. That is, a byte order of the form @samp{32107654}. Note: this
5930 option should only be used if you require compatibility with code for
5931 big-endian ARM processors generated by versions of the compiler prior to
5934 @item -malignment-traps
5935 @opindex malignment-traps
5936 Generate code that will not trap if the MMU has alignment traps enabled.
5937 On ARM architectures prior to ARMv4, there were no instructions to
5938 access half-word objects stored in memory. However, when reading from
5939 memory a feature of the ARM architecture allows a word load to be used,
5940 even if the address is unaligned, and the processor core will rotate the
5941 data as it is being loaded. This option tells the compiler that such
5942 misaligned accesses will cause a MMU trap and that it should instead
5943 synthesise the access as a series of byte accesses. The compiler can
5944 still use word accesses to load half-word data if it knows that the
5945 address is aligned to a word boundary.
5947 This option is ignored when compiling for ARM architecture 4 or later,
5948 since these processors have instructions to directly access half-word
5951 @item -mno-alignment-traps
5952 @opindex mno-alignment-traps
5953 Generate code that assumes that the MMU will not trap unaligned
5954 accesses. This produces better code when the target instruction set
5955 does not have half-word memory operations (i.e.@: implementations prior to
5958 Note that you cannot use this option to access unaligned word objects,
5959 since the processor will only fetch one 32-bit aligned object from
5962 The default setting for most targets is @option{-mno-alignment-traps}, since
5963 this produces better code when there are no half-word memory
5964 instructions available.
5966 @item -mshort-load-bytes
5967 @itemx -mno-short-load-words
5968 @opindex mshort-load-bytes
5969 @opindex mno-short-load-words
5970 These are deprecated aliases for @option{-malignment-traps}.
5972 @item -mno-short-load-bytes
5973 @itemx -mshort-load-words
5974 @opindex mno-short-load-bytes
5975 @opindex mshort-load-words
5976 This are deprecated aliases for @option{-mno-alignment-traps}.
5980 This option only applies to RISC iX@. Emulate the native BSD-mode
5981 compiler. This is the default if @option{-ansi} is not specified.
5985 This option only applies to RISC iX@. Emulate the native X/Open-mode
5988 @item -mno-symrename
5989 @opindex mno-symrename
5990 This option only applies to RISC iX@. Do not run the assembler
5991 post-processor, @samp{symrename}, after code has been assembled.
5992 Normally it is necessary to modify some of the standard symbols in
5993 preparation for linking with the RISC iX C library; this option
5994 suppresses this pass. The post-processor is never run when the
5995 compiler is built for cross-compilation.
5997 @item -mcpu=@var{name}
5999 This specifies the name of the target ARM processor. GCC uses this name
6000 to determine what kind of instructions it can emit when generating
6001 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6002 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6003 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6004 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6005 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6006 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6007 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6008 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6009 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6010 @samp{arm1020t}, @samp{xscale}.
6012 @itemx -mtune=@var{name}
6014 This option is very similar to the @option{-mcpu=} option, except that
6015 instead of specifying the actual target processor type, and hence
6016 restricting which instructions can be used, it specifies that GCC should
6017 tune the performance of the code as if the target were of the type
6018 specified in this option, but still choosing the instructions that it
6019 will generate based on the cpu specified by a @option{-mcpu=} option.
6020 For some ARM implementations better performance can be obtained by using
6023 @item -march=@var{name}
6025 This specifies the name of the target ARM architecture. GCC uses this
6026 name to determine what kind of instructions it can emit when generating
6027 assembly code. This option can be used in conjunction with or instead
6028 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6029 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6030 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6032 @item -mfpe=@var{number}
6033 @itemx -mfp=@var{number}
6036 This specifies the version of the floating point emulation available on
6037 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6038 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6040 @item -mstructure-size-boundary=@var{n}
6041 @opindex mstructure-size-boundary
6042 The size of all structures and unions will be rounded up to a multiple
6043 of the number of bits set by this option. Permissible values are 8 and
6044 32. The default value varies for different toolchains. For the COFF
6045 targeted toolchain the default value is 8. Specifying the larger number
6046 can produce faster, more efficient code, but can also increase the size
6047 of the program. The two values are potentially incompatible. Code
6048 compiled with one value cannot necessarily expect to work with code or
6049 libraries compiled with the other value, if they exchange information
6050 using structures or unions.
6052 @item -mabort-on-noreturn
6053 @opindex mabort-on-noreturn
6054 Generate a call to the function @code{abort} at the end of a
6055 @code{noreturn} function. It will be executed if the function tries to
6059 @itemx -mno-long-calls
6060 @opindex mlong-calls
6061 @opindex mno-long-calls
6062 Tells the compiler to perform function calls by first loading the
6063 address of the function into a register and then performing a subroutine
6064 call on this register. This switch is needed if the target function
6065 will lie outside of the 64 megabyte addressing range of the offset based
6066 version of subroutine call instruction.
6068 Even if this switch is enabled, not all function calls will be turned
6069 into long calls. The heuristic is that static functions, functions
6070 which have the @samp{short-call} attribute, functions that are inside
6071 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6072 definitions have already been compiled within the current compilation
6073 unit, will not be turned into long calls. The exception to this rule is
6074 that weak function definitions, functions with the @samp{long-call}
6075 attribute or the @samp{section} attribute, and functions that are within
6076 the scope of a @samp{#pragma long_calls} directive, will always be
6077 turned into long calls.
6079 This feature is not enabled by default. Specifying
6080 @option{-mno-long-calls} will restore the default behavior, as will
6081 placing the function calls within the scope of a @samp{#pragma
6082 long_calls_off} directive. Note these switches have no effect on how
6083 the compiler generates code to handle function calls via function
6086 @item -mnop-fun-dllimport
6087 @opindex mnop-fun-dllimport
6088 Disable support for the @code{dllimport} attribute.
6090 @item -msingle-pic-base
6091 @opindex msingle-pic-base
6092 Treat the register used for PIC addressing as read-only, rather than
6093 loading it in the prologue for each function. The run-time system is
6094 responsible for initializing this register with an appropriate value
6095 before execution begins.
6097 @item -mpic-register=@var{reg}
6098 @opindex mpic-register
6099 Specify the register to be used for PIC addressing. The default is R10
6100 unless stack-checking is enabled, when R9 is used.
6102 @item -mpoke-function-name
6103 @opindex mpoke-function-name
6104 Write the name of each function into the text section, directly
6105 preceding the function prologue. The generated code is similar to this:
6109 .ascii "arm_poke_function_name", 0
6112 .word 0xff000000 + (t1 - t0)
6113 arm_poke_function_name
6115 stmfd sp!, @{fp, ip, lr, pc@}
6119 When performing a stack backtrace, code can inspect the value of
6120 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6121 location @code{pc - 12} and the top 8 bits are set, then we know that
6122 there is a function name embedded immediately preceding this location
6123 and has length @code{((pc[-3]) & 0xff000000)}.
6127 Generate code for the 16-bit Thumb instruction set. The default is to
6128 use the 32-bit ARM instruction set.
6131 @opindex mtpcs-frame
6132 Generate a stack frame that is compliant with the Thumb Procedure Call
6133 Standard for all non-leaf functions. (A leaf function is one that does
6134 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6136 @item -mtpcs-leaf-frame
6137 @opindex mtpcs-leaf-frame
6138 Generate a stack frame that is compliant with the Thumb Procedure Call
6139 Standard for all leaf functions. (A leaf function is one that does
6140 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6142 @item -mcallee-super-interworking
6143 @opindex mcallee-super-interworking
6144 Gives all externally visible functions in the file being compiled an ARM
6145 instruction set header which switches to Thumb mode before executing the
6146 rest of the function. This allows these functions to be called from
6147 non-interworking code.
6149 @item -mcaller-super-interworking
6150 @opindex mcaller-super-interworking
6151 Allows calls via function pointers (including virtual functions) to
6152 execute correctly regardless of whether the target code has been
6153 compiled for interworking or not. There is a small overhead in the cost
6154 of executing a function pointer if this option is enabled.
6158 @node MN10200 Options
6159 @subsection MN10200 Options
6160 @cindex MN10200 options
6161 These @option{-m} options are defined for Matsushita MN10200 architectures:
6166 Indicate to the linker that it should perform a relaxation optimization pass
6167 to shorten branches, calls and absolute memory addresses. This option only
6168 has an effect when used on the command line for the final link step.
6170 This option makes symbolic debugging impossible.
6173 @node MN10300 Options
6174 @subsection MN10300 Options
6175 @cindex MN10300 options
6176 These @option{-m} options are defined for Matsushita MN10300 architectures:
6181 Generate code to avoid bugs in the multiply instructions for the MN10300
6182 processors. This is the default.
6185 @opindex mno-mult-bug
6186 Do not generate code to avoid bugs in the multiply instructions for the
6191 Generate code which uses features specific to the AM33 processor.
6195 Do not generate code which uses features specific to the AM33 processor. This
6200 Do not link in the C run-time initialization object file.
6204 Indicate to the linker that it should perform a relaxation optimization pass
6205 to shorten branches, calls and absolute memory addresses. This option only
6206 has an effect when used on the command line for the final link step.
6208 This option makes symbolic debugging impossible.
6212 @node M32R/D Options
6213 @subsection M32R/D Options
6214 @cindex M32R/D options
6216 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6221 Generate code for the M32R/X@.
6225 Generate code for the M32R@. This is the default.
6227 @item -mcode-model=small
6228 @opindex mcode-model=small
6229 Assume all objects live in the lower 16MB of memory (so that their addresses
6230 can be loaded with the @code{ld24} instruction), and assume all subroutines
6231 are reachable with the @code{bl} instruction.
6232 This is the default.
6234 The addressability of a particular object can be set with the
6235 @code{model} attribute.
6237 @item -mcode-model=medium
6238 @opindex mcode-model=medium
6239 Assume objects may be anywhere in the 32-bit address space (the compiler
6240 will generate @code{seth/add3} instructions to load their addresses), and
6241 assume all subroutines are reachable with the @code{bl} instruction.
6243 @item -mcode-model=large
6244 @opindex mcode-model=large
6245 Assume objects may be anywhere in the 32-bit address space (the compiler
6246 will generate @code{seth/add3} instructions to load their addresses), and
6247 assume subroutines may not be reachable with the @code{bl} instruction
6248 (the compiler will generate the much slower @code{seth/add3/jl}
6249 instruction sequence).
6252 @opindex msdata=none
6253 Disable use of the small data area. Variables will be put into
6254 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6255 @code{section} attribute has been specified).
6256 This is the default.
6258 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6259 Objects may be explicitly put in the small data area with the
6260 @code{section} attribute using one of these sections.
6263 @opindex msdata=sdata
6264 Put small global and static data in the small data area, but do not
6265 generate special code to reference them.
6269 Put small global and static data in the small data area, and generate
6270 special instructions to reference them.
6274 @cindex smaller data references
6275 Put global and static objects less than or equal to @var{num} bytes
6276 into the small data or bss sections instead of the normal data or bss
6277 sections. The default value of @var{num} is 8.
6278 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6279 for this option to have any effect.
6281 All modules should be compiled with the same @option{-G @var{num}} value.
6282 Compiling with different values of @var{num} may or may not work; if it
6283 doesn't the linker will give an error message---incorrect code will not be
6289 @subsection M88K Options
6290 @cindex M88k options
6292 These @samp{-m} options are defined for Motorola 88k architectures:
6297 Generate code that works well on both the m88100 and the
6302 Generate code that works best for the m88100, but that also
6307 Generate code that works best for the m88110, and may not run
6312 Obsolete option to be removed from the next revision.
6315 @item -midentify-revision
6316 @opindex midentify-revision
6317 @cindex identifying source, compiler (88k)
6318 Include an @code{ident} directive in the assembler output recording the
6319 source file name, compiler name and version, timestamp, and compilation
6322 @item -mno-underscores
6323 @opindex mno-underscores
6324 @cindex underscores, avoiding (88k)
6325 In assembler output, emit symbol names without adding an underscore
6326 character at the beginning of each name. The default is to use an
6327 underscore as prefix on each name.
6329 @item -mocs-debug-info
6330 @itemx -mno-ocs-debug-info
6331 @opindex mocs-debug-info
6332 @opindex mno-ocs-debug-info
6334 @cindex debugging, 88k OCS
6335 Include (or omit) additional debugging information (about registers used
6336 in each stack frame) as specified in the 88open Object Compatibility
6337 Standard, ``OCS''@. This extra information allows debugging of code that
6338 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6339 Delta 88 SVr3.2 is to include this information; other 88k configurations
6340 omit this information by default.
6342 @item -mocs-frame-position
6343 @opindex mocs-frame-position
6344 @cindex register positions in frame (88k)
6345 When emitting COFF debugging information for automatic variables and
6346 parameters stored on the stack, use the offset from the canonical frame
6347 address, which is the stack pointer (register 31) on entry to the
6348 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6349 @option{-mocs-frame-position}; other 88k configurations have the default
6350 @option{-mno-ocs-frame-position}.
6352 @item -mno-ocs-frame-position
6353 @opindex mno-ocs-frame-position
6354 @cindex register positions in frame (88k)
6355 When emitting COFF debugging information for automatic variables and
6356 parameters stored on the stack, use the offset from the frame pointer
6357 register (register 30). When this option is in effect, the frame
6358 pointer is not eliminated when debugging information is selected by the
6361 @item -moptimize-arg-area
6362 @opindex moptimize-arg-area
6363 @cindex arguments in frame (88k)
6364 Save space by reorganizing the stack frame. This option generates code
6365 that does not agree with the 88open specifications, but uses less
6368 @itemx -mno-optimize-arg-area
6369 @opindex mno-optimize-arg-area
6370 Do not reorganize the stack frame to save space. This is the default.
6371 The generated conforms to the specification, but uses more memory.
6373 @item -mshort-data-@var{num}
6374 @opindex mshort-data
6375 @cindex smaller data references (88k)
6376 @cindex r0-relative references (88k)
6377 Generate smaller data references by making them relative to @code{r0},
6378 which allows loading a value using a single instruction (rather than the
6379 usual two). You control which data references are affected by
6380 specifying @var{num} with this option. For example, if you specify
6381 @option{-mshort-data-512}, then the data references affected are those
6382 involving displacements of less than 512 bytes.
6383 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6386 @item -mserialize-volatile
6387 @opindex mserialize-volatile
6388 @itemx -mno-serialize-volatile
6389 @opindex mno-serialize-volatile
6390 @cindex sequential consistency on 88k
6391 Do, or don't, generate code to guarantee sequential consistency
6392 of volatile memory references. By default, consistency is
6395 The order of memory references made by the MC88110 processor does
6396 not always match the order of the instructions requesting those
6397 references. In particular, a load instruction may execute before
6398 a preceding store instruction. Such reordering violates
6399 sequential consistency of volatile memory references, when there
6400 are multiple processors. When consistency must be guaranteed,
6401 GCC generates special instructions, as needed, to force
6402 execution in the proper order.
6404 The MC88100 processor does not reorder memory references and so
6405 always provides sequential consistency. However, by default, GCC
6406 generates the special instructions to guarantee consistency
6407 even when you use @option{-m88100}, so that the code may be run on an
6408 MC88110 processor. If you intend to run your code only on the
6409 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6411 The extra code generated to guarantee consistency may affect the
6412 performance of your application. If you know that you can safely
6413 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6419 @cindex assembler syntax, 88k
6421 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6422 related to System V release 4 (SVr4). This controls the following:
6426 Which variant of the assembler syntax to emit.
6428 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6429 that is used on System V release 4.
6431 @option{-msvr4} makes GCC issue additional declaration directives used in
6435 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6436 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6437 other m88k configurations.
6439 @item -mversion-03.00
6440 @opindex mversion-03.00
6441 This option is obsolete, and is ignored.
6442 @c ??? which asm syntax better for GAS? option there too?
6444 @item -mno-check-zero-division
6445 @itemx -mcheck-zero-division
6446 @opindex mno-check-zero-division
6447 @opindex mcheck-zero-division
6448 @cindex zero division on 88k
6449 Do, or don't, generate code to guarantee that integer division by
6450 zero will be detected. By default, detection is guaranteed.
6452 Some models of the MC88100 processor fail to trap upon integer
6453 division by zero under certain conditions. By default, when
6454 compiling code that might be run on such a processor, GCC
6455 generates code that explicitly checks for zero-valued divisors
6456 and traps with exception number 503 when one is detected. Use of
6457 @option{-mno-check-zero-division} suppresses such checking for code
6458 generated to run on an MC88100 processor.
6460 GCC assumes that the MC88110 processor correctly detects all instances
6461 of integer division by zero. When @option{-m88110} is specified, no
6462 explicit checks for zero-valued divisors are generated, and both
6463 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6466 @item -muse-div-instruction
6467 @opindex muse-div-instruction
6468 @cindex divide instruction, 88k
6469 Use the div instruction for signed integer division on the
6470 MC88100 processor. By default, the div instruction is not used.
6472 On the MC88100 processor the signed integer division instruction
6473 div) traps to the operating system on a negative operand. The
6474 operating system transparently completes the operation, but at a
6475 large cost in execution time. By default, when compiling code
6476 that might be run on an MC88100 processor, GCC emulates signed
6477 integer division using the unsigned integer division instruction
6478 divu), thereby avoiding the large penalty of a trap to the
6479 operating system. Such emulation has its own, smaller, execution
6480 cost in both time and space. To the extent that your code's
6481 important signed integer division operations are performed on two
6482 nonnegative operands, it may be desirable to use the div
6483 instruction directly.
6485 On the MC88110 processor the div instruction (also known as the
6486 divs instruction) processes negative operands without trapping to
6487 the operating system. When @option{-m88110} is specified,
6488 @option{-muse-div-instruction} is ignored, and the div instruction is used
6489 for signed integer division.
6491 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6492 particular, the behavior of such a division with and without
6493 @option{-muse-div-instruction} may differ.
6495 @item -mtrap-large-shift
6496 @itemx -mhandle-large-shift
6497 @opindex mtrap-large-shift
6498 @opindex mhandle-large-shift
6499 @cindex bit shift overflow (88k)
6500 @cindex large bit shifts (88k)
6501 Include code to detect bit-shifts of more than 31 bits; respectively,
6502 trap such shifts or emit code to handle them properly. By default GCC
6503 makes no special provision for large bit shifts.
6505 @item -mwarn-passed-structs
6506 @opindex mwarn-passed-structs
6507 @cindex structure passing (88k)
6508 Warn when a function passes a struct as an argument or result.
6509 Structure-passing conventions have changed during the evolution of the C
6510 language, and are often the source of portability problems. By default,
6511 GCC issues no such warning.
6514 @c break page here to avoid unsightly interparagraph stretch.
6518 @node RS/6000 and PowerPC Options
6519 @subsection IBM RS/6000 and PowerPC Options
6520 @cindex RS/6000 and PowerPC Options
6521 @cindex IBM RS/6000 and PowerPC Options
6523 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6531 @itemx -mpowerpc-gpopt
6532 @itemx -mno-powerpc-gpopt
6533 @itemx -mpowerpc-gfxopt
6534 @itemx -mno-powerpc-gfxopt
6536 @itemx -mno-powerpc64
6542 @opindex mno-powerpc
6543 @opindex mpowerpc-gpopt
6544 @opindex mno-powerpc-gpopt
6545 @opindex mpowerpc-gfxopt
6546 @opindex mno-powerpc-gfxopt
6548 @opindex mno-powerpc64
6549 GCC supports two related instruction set architectures for the
6550 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6551 instructions supported by the @samp{rios} chip set used in the original
6552 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6553 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6554 the IBM 4xx microprocessors.
6556 Neither architecture is a subset of the other. However there is a
6557 large common subset of instructions supported by both. An MQ
6558 register is included in processors supporting the POWER architecture.
6560 You use these options to specify which instructions are available on the
6561 processor you are using. The default value of these options is
6562 determined when configuring GCC@. Specifying the
6563 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6564 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6565 rather than the options listed above.
6567 The @option{-mpower} option allows GCC to generate instructions that
6568 are found only in the POWER architecture and to use the MQ register.
6569 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6570 to generate instructions that are present in the POWER2 architecture but
6571 not the original POWER architecture.
6573 The @option{-mpowerpc} option allows GCC to generate instructions that
6574 are found only in the 32-bit subset of the PowerPC architecture.
6575 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6576 GCC to use the optional PowerPC architecture instructions in the
6577 General Purpose group, including floating-point square root. Specifying
6578 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6579 use the optional PowerPC architecture instructions in the Graphics
6580 group, including floating-point select.
6582 The @option{-mpowerpc64} option allows GCC to generate the additional
6583 64-bit instructions that are found in the full PowerPC64 architecture
6584 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6585 @option{-mno-powerpc64}.
6587 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6588 will use only the instructions in the common subset of both
6589 architectures plus some special AIX common-mode calls, and will not use
6590 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6591 permits GCC to use any instruction from either architecture and to
6592 allow use of the MQ register; specify this for the Motorola MPC601.
6594 @item -mnew-mnemonics
6595 @itemx -mold-mnemonics
6596 @opindex mnew-mnemonics
6597 @opindex mold-mnemonics
6598 Select which mnemonics to use in the generated assembler code. With
6599 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6600 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6601 assembler mnemonics defined for the POWER architecture. Instructions
6602 defined in only one architecture have only one mnemonic; GCC uses that
6603 mnemonic irrespective of which of these options is specified.
6605 GCC defaults to the mnemonics appropriate for the architecture in
6606 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6607 value of these option. Unless you are building a cross-compiler, you
6608 should normally not specify either @option{-mnew-mnemonics} or
6609 @option{-mold-mnemonics}, but should instead accept the default.
6611 @item -mcpu=@var{cpu_type}
6613 Set architecture type, register usage, choice of mnemonics, and
6614 instruction scheduling parameters for machine type @var{cpu_type}.
6615 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6616 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6617 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6618 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6619 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6620 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6622 @option{-mcpu=common} selects a completely generic processor. Code
6623 generated under this option will run on any POWER or PowerPC processor.
6624 GCC will use only the instructions in the common subset of both
6625 architectures, and will not use the MQ register. GCC assumes a generic
6626 processor model for scheduling purposes.
6628 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6629 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6630 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6631 types, with an appropriate, generic processor model assumed for
6632 scheduling purposes.
6634 The other options specify a specific processor. Code generated under
6635 those options will run best on that processor, and may not run at all on
6638 The @option{-mcpu} options automatically enable or disable other
6639 @option{-m} options as follows:
6643 @option{-mno-power}, @option{-mno-powerc}
6650 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6665 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6668 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6673 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6676 @item -mtune=@var{cpu_type}
6678 Set the instruction scheduling parameters for machine type
6679 @var{cpu_type}, but do not set the architecture type, register usage, or
6680 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6681 values for @var{cpu_type} are used for @option{-mtune} as for
6682 @option{-mcpu}. If both are specified, the code generated will use the
6683 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6684 scheduling parameters set by @option{-mtune}.
6689 @opindex mno-altivec
6690 These switches enable or disable the use of built-in functions that
6691 allow access to the AltiVec instruction set. You may also need to set
6692 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6696 @itemx -mno-fp-in-toc
6697 @itemx -mno-sum-in-toc
6698 @itemx -mminimal-toc
6700 @opindex mno-fp-in-toc
6701 @opindex mno-sum-in-toc
6702 @opindex mminimal-toc
6703 Modify generation of the TOC (Table Of Contents), which is created for
6704 every executable file. The @option{-mfull-toc} option is selected by
6705 default. In that case, GCC will allocate at least one TOC entry for
6706 each unique non-automatic variable reference in your program. GCC
6707 will also place floating-point constants in the TOC@. However, only
6708 16,384 entries are available in the TOC@.
6710 If you receive a linker error message that saying you have overflowed
6711 the available TOC space, you can reduce the amount of TOC space used
6712 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6713 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6714 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6715 generate code to calculate the sum of an address and a constant at
6716 run-time instead of putting that sum into the TOC@. You may specify one
6717 or both of these options. Each causes GCC to produce very slightly
6718 slower and larger code at the expense of conserving TOC space.
6720 If you still run out of space in the TOC even when you specify both of
6721 these options, specify @option{-mminimal-toc} instead. This option causes
6722 GCC to make only one TOC entry for every file. When you specify this
6723 option, GCC will produce code that is slower and larger but which
6724 uses extremely little TOC space. You may wish to use this option
6725 only on files that contain less frequently executed code.
6731 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6732 @code{long} type, and the infrastructure needed to support them.
6733 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6734 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6735 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6740 @opindex mno-xl-call
6741 On AIX, pass floating-point arguments to prototyped functions beyond the
6742 register save area (RSA) on the stack in addition to argument FPRs. The
6743 AIX calling convention was extended but not initially documented to
6744 handle an obscure K&R C case of calling a function that takes the
6745 address of its arguments with fewer arguments than declared. AIX XL
6746 compilers access floating point arguments which do not fit in the
6747 RSA from the stack when a subroutine is compiled without
6748 optimization. Because always storing floating-point arguments on the
6749 stack is inefficient and rarely needed, this option is not enabled by
6750 default and only is necessary when calling subroutines compiled by AIX
6751 XL compilers without optimization.
6755 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6756 application written to use message passing with special startup code to
6757 enable the application to run. The system must have PE installed in the
6758 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6759 must be overridden with the @option{-specs=} option to specify the
6760 appropriate directory location. The Parallel Environment does not
6761 support threads, so the @option{-mpe} option and the @option{-pthread}
6762 option are incompatible.
6766 @opindex msoft-float
6767 @opindex mhard-float
6768 Generate code that does not use (uses) the floating-point register set.
6769 Software floating point emulation is provided if you use the
6770 @option{-msoft-float} option, and pass the option to GCC when linking.
6773 @itemx -mno-multiple
6775 @opindex mno-multiple
6776 Generate code that uses (does not use) the load multiple word
6777 instructions and the store multiple word instructions. These
6778 instructions are generated by default on POWER systems, and not
6779 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6780 endian PowerPC systems, since those instructions do not work when the
6781 processor is in little endian mode. The exceptions are PPC740 and
6782 PPC750 which permit the instructions usage in little endian mode.
6788 Generate code that uses (does not use) the load string instructions
6789 and the store string word instructions to save multiple registers and
6790 do small block moves. These instructions are generated by default on
6791 POWER systems, and not generated on PowerPC systems. Do not use
6792 @option{-mstring} on little endian PowerPC systems, since those
6793 instructions do not work when the processor is in little endian mode.
6794 The exceptions are PPC740 and PPC750 which permit the instructions
6795 usage in little endian mode.
6801 Generate code that uses (does not use) the load or store instructions
6802 that update the base register to the address of the calculated memory
6803 location. These instructions are generated by default. If you use
6804 @option{-mno-update}, there is a small window between the time that the
6805 stack pointer is updated and the address of the previous frame is
6806 stored, which means code that walks the stack frame across interrupts or
6807 signals may get corrupted data.
6810 @itemx -mno-fused-madd
6811 @opindex mfused-madd
6812 @opindex mno-fused-madd
6813 Generate code that uses (does not use) the floating point multiply and
6814 accumulate instructions. These instructions are generated by default if
6815 hardware floating is used.
6817 @item -mno-bit-align
6819 @opindex mno-bit-align
6821 On System V.4 and embedded PowerPC systems do not (do) force structures
6822 and unions that contain bit-fields to be aligned to the base type of the
6825 For example, by default a structure containing nothing but 8
6826 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6827 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6828 the structure would be aligned to a 1 byte boundary and be one byte in
6831 @item -mno-strict-align
6832 @itemx -mstrict-align
6833 @opindex mno-strict-align
6834 @opindex mstrict-align
6835 On System V.4 and embedded PowerPC systems do not (do) assume that
6836 unaligned memory references will be handled by the system.
6839 @itemx -mno-relocatable
6840 @opindex mrelocatable
6841 @opindex mno-relocatable
6842 On embedded PowerPC systems generate code that allows (does not allow)
6843 the program to be relocated to a different address at runtime. If you
6844 use @option{-mrelocatable} on any module, all objects linked together must
6845 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6847 @item -mrelocatable-lib
6848 @itemx -mno-relocatable-lib
6849 @opindex mrelocatable-lib
6850 @opindex mno-relocatable-lib
6851 On embedded PowerPC systems generate code that allows (does not allow)
6852 the program to be relocated to a different address at runtime. Modules
6853 compiled with @option{-mrelocatable-lib} can be linked with either modules
6854 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6855 with modules compiled with the @option{-mrelocatable} options.
6861 On System V.4 and embedded PowerPC systems do not (do) assume that
6862 register 2 contains a pointer to a global area pointing to the addresses
6863 used in the program.
6866 @itemx -mlittle-endian
6868 @opindex mlittle-endian
6869 On System V.4 and embedded PowerPC systems compile code for the
6870 processor in little endian mode. The @option{-mlittle-endian} option is
6871 the same as @option{-mlittle}.
6876 @opindex mbig-endian
6877 On System V.4 and embedded PowerPC systems compile code for the
6878 processor in big endian mode. The @option{-mbig-endian} option is
6879 the same as @option{-mbig}.
6883 On System V.4 and embedded PowerPC systems compile code using calling
6884 conventions that adheres to the March 1995 draft of the System V
6885 Application Binary Interface, PowerPC processor supplement. This is the
6886 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6888 @item -mcall-sysv-eabi
6889 @opindex mcall-sysv-eabi
6890 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6892 @item -mcall-sysv-noeabi
6893 @opindex mcall-sysv-noeabi
6894 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6898 On System V.4 and embedded PowerPC systems compile code using calling
6899 conventions that are similar to those used on AIX@. This is the
6900 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6902 @item -mcall-solaris
6903 @opindex mcall-solaris
6904 On System V.4 and embedded PowerPC systems compile code for the Solaris
6908 @opindex mcall-linux
6909 On System V.4 and embedded PowerPC systems compile code for the
6910 Linux-based GNU system.
6914 On System V.4 and embedded PowerPC systems compile code for the
6915 Hurd-based GNU system.
6918 @opindex mcall-netbsd
6919 On System V.4 and embedded PowerPC systems compile code for the
6920 NetBSD operating system.
6922 @item -maix-struct-return
6923 @opindex maix-struct-return
6924 Return all structures in memory (as specified by the AIX ABI)@.
6926 @item -msvr4-struct-return
6927 @opindex msvr4-struct-return
6928 Return structures smaller than 8 bytes in registers (as specified by the
6932 @opindex mabi=altivec
6933 Extend the current ABI with AltiVec ABI extensions. This does not
6934 change the default ABI, instead it adds the AltiVec ABI extensions to
6938 @itemx -mno-prototype
6940 @opindex mno-prototype
6941 On System V.4 and embedded PowerPC systems assume that all calls to
6942 variable argument functions are properly prototyped. Otherwise, the
6943 compiler must insert an instruction before every non prototyped call to
6944 set or clear bit 6 of the condition code register (@var{CR}) to
6945 indicate whether floating point values were passed in the floating point
6946 registers in case the function takes a variable arguments. With
6947 @option{-mprototype}, only calls to prototyped variable argument functions
6948 will set or clear the bit.
6952 On embedded PowerPC systems, assume that the startup module is called
6953 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6954 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6959 On embedded PowerPC systems, assume that the startup module is called
6960 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6965 On embedded PowerPC systems, assume that the startup module is called
6966 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6970 @opindex myellowknife
6971 On embedded PowerPC systems, assume that the startup module is called
6972 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6977 On System V.4 and embedded PowerPC systems, specify that you are
6978 compiling for a VxWorks system.
6982 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6983 header to indicate that @samp{eabi} extended relocations are used.
6989 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6990 Embedded Applications Binary Interface (eabi) which is a set of
6991 modifications to the System V.4 specifications. Selecting @option{-meabi}
6992 means that the stack is aligned to an 8 byte boundary, a function
6993 @code{__eabi} is called to from @code{main} to set up the eabi
6994 environment, and the @option{-msdata} option can use both @code{r2} and
6995 @code{r13} to point to two separate small data areas. Selecting
6996 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6997 do not call an initialization function from @code{main}, and the
6998 @option{-msdata} option will only use @code{r13} to point to a single
6999 small data area. The @option{-meabi} option is on by default if you
7000 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7003 @opindex msdata=eabi
7004 On System V.4 and embedded PowerPC systems, put small initialized
7005 @code{const} global and static data in the @samp{.sdata2} section, which
7006 is pointed to by register @code{r2}. Put small initialized
7007 non-@code{const} global and static data in the @samp{.sdata} section,
7008 which is pointed to by register @code{r13}. Put small uninitialized
7009 global and static data in the @samp{.sbss} section, which is adjacent to
7010 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7011 incompatible with the @option{-mrelocatable} option. The
7012 @option{-msdata=eabi} option also sets the @option{-memb} option.
7015 @opindex msdata=sysv
7016 On System V.4 and embedded PowerPC systems, put small global and static
7017 data in the @samp{.sdata} section, which is pointed to by register
7018 @code{r13}. Put small uninitialized global and static data in the
7019 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7020 The @option{-msdata=sysv} option is incompatible with the
7021 @option{-mrelocatable} option.
7023 @item -msdata=default
7025 @opindex msdata=default
7027 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7028 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7029 same as @option{-msdata=sysv}.
7032 @opindex msdata-data
7033 On System V.4 and embedded PowerPC systems, put small global and static
7034 data in the @samp{.sdata} section. Put small uninitialized global and
7035 static data in the @samp{.sbss} section. Do not use register @code{r13}
7036 to address small data however. This is the default behavior unless
7037 other @option{-msdata} options are used.
7041 @opindex msdata=none
7043 On embedded PowerPC systems, put all initialized global and static data
7044 in the @samp{.data} section, and all uninitialized data in the
7045 @samp{.bss} section.
7049 @cindex smaller data references (PowerPC)
7050 @cindex .sdata/.sdata2 references (PowerPC)
7051 On embedded PowerPC systems, put global and static items less than or
7052 equal to @var{num} bytes into the small data or bss sections instead of
7053 the normal data or bss section. By default, @var{num} is 8. The
7054 @option{-G @var{num}} switch is also passed to the linker.
7055 All modules should be compiled with the same @option{-G @var{num}} value.
7058 @itemx -mno-regnames
7060 @opindex mno-regnames
7061 On System V.4 and embedded PowerPC systems do (do not) emit register
7062 names in the assembly language output using symbolic forms.
7066 Adds support for multithreading with the @dfn{pthreads} library.
7067 This option sets flags for both the preprocessor and linker.
7072 @subsection IBM RT Options
7074 @cindex IBM RT options
7076 These @samp{-m} options are defined for the IBM RT PC:
7080 @opindex min-line-mul
7081 Use an in-line code sequence for integer multiplies. This is the
7084 @item -mcall-lib-mul
7085 @opindex mcall-lib-mul
7086 Call @code{lmul$$} for integer multiples.
7088 @item -mfull-fp-blocks
7089 @opindex mfull-fp-blocks
7090 Generate full-size floating point data blocks, including the minimum
7091 amount of scratch space recommended by IBM@. This is the default.
7093 @item -mminimum-fp-blocks
7094 @opindex mminimum-fp-blocks
7095 Do not include extra scratch space in floating point data blocks. This
7096 results in smaller code, but slower execution, since scratch space must
7097 be allocated dynamically.
7099 @cindex @file{varargs.h} and RT PC
7100 @cindex @file{stdarg.h} and RT PC
7101 @item -mfp-arg-in-fpregs
7102 @opindex mfp-arg-in-fpregs
7103 Use a calling sequence incompatible with the IBM calling convention in
7104 which floating point arguments are passed in floating point registers.
7105 Note that @code{varargs.h} and @code{stdarg.h} will not work with
7106 floating point operands if this option is specified.
7108 @item -mfp-arg-in-gregs
7109 @opindex mfp-arg-in-gregs
7110 Use the normal calling convention for floating point arguments. This is
7113 @item -mhc-struct-return
7114 @opindex mhc-struct-return
7115 Return structures of more than one word in memory, rather than in a
7116 register. This provides compatibility with the MetaWare HighC (hc)
7117 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7118 with the Portable C Compiler (pcc).
7120 @item -mnohc-struct-return
7121 @opindex mnohc-struct-return
7122 Return some structures of more than one word in registers, when
7123 convenient. This is the default. For compatibility with the
7124 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7125 option @option{-mhc-struct-return}.
7129 @subsection MIPS Options
7130 @cindex MIPS options
7132 These @samp{-m} options are defined for the MIPS family of computers:
7136 @item -march=@var{cpu-type}
7138 Assume the defaults for the machine type @var{cpu-type} when generating
7139 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7140 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7141 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7142 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7143 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7144 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
7146 @item -mtune=@var{cpu-type}
7148 Assume the defaults for the machine type @var{cpu-type} when scheduling
7149 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7150 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7151 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7152 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7153 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7154 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
7155 @var{cpu-type} will schedule things appropriately for that particular
7156 chip, the compiler will not generate any code that does not meet level 1
7157 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
7158 or @option{-mabi} switch being used.
7160 @item -mcpu=@var{cpu-type}
7162 This is identical to specifying both @option{-march} and @option{-mtune}.
7166 Issue instructions from level 1 of the MIPS ISA@. This is the default.
7167 @samp{r3000} is the default @var{cpu-type} at this ISA level.
7171 Issue instructions from level 2 of the MIPS ISA (branch likely, square
7172 root instructions). @samp{r6000} is the default @var{cpu-type} at this
7177 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
7178 @samp{r4000} is the default @var{cpu-type} at this ISA level.
7182 Issue instructions from level 4 of the MIPS ISA (conditional move,
7183 prefetch, enhanced FPU instructions). @samp{r8000} is the default
7184 @var{cpu-type} at this ISA level.
7188 Assume that 32 32-bit floating point registers are available. This is
7193 Assume that 32 64-bit floating point registers are available. This is
7194 the default when the @option{-mips3} option is used.
7197 @itemx -mno-fused-madd
7198 @opindex mfused-madd
7199 @opindex mno-fused-madd
7200 Generate code that uses (does not use) the floating point multiply and
7201 accumulate instructions, when they are available. These instructions
7202 are generated by default if they are available, but this may be
7203 undesirable if the extra precision causes problems or on certain chips
7204 in the mode where denormals are rounded to zero where denormals
7205 generated by multiply and accumulate instructions cause exceptions
7210 Assume that 32 32-bit general purpose registers are available. This is
7215 Assume that 32 64-bit general purpose registers are available. This is
7216 the default when the @option{-mips3} option is used.
7220 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7221 explanation of the default, and the width of pointers.
7225 Force long types to be 64 bits wide. See @option{-mlong32} for an
7226 explanation of the default, and the width of pointers.
7230 Force long, int, and pointer types to be 32 bits wide.
7232 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
7233 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
7234 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
7235 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
7236 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
7237 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
7238 are 32 bits, and longs are 64 bits wide. The width of pointer types is
7239 the smaller of the width of longs or the width of general purpose
7240 registers (which in turn depends on the ISA)@.
7252 Generate code for the indicated ABI@. The default instruction level is
7253 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7254 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7255 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7260 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7261 add normal debug information. This is the default for all
7262 platforms except for the OSF/1 reference platform, using the OSF/rose
7263 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7264 switches are used, the @file{mips-tfile} program will encapsulate the
7265 stabs within MIPS ECOFF@.
7269 Generate code for the GNU assembler. This is the default on the OSF/1
7270 reference platform, using the OSF/rose object format. Also, this is
7271 the default if the configure option @option{--with-gnu-as} is used.
7273 @item -msplit-addresses
7274 @itemx -mno-split-addresses
7275 @opindex msplit-addresses
7276 @opindex mno-split-addresses
7277 Generate code to load the high and low parts of address constants separately.
7278 This allows GCC to optimize away redundant loads of the high order
7279 bits of addresses. This optimization requires GNU as and GNU ld.
7280 This optimization is enabled by default for some embedded targets where
7281 GNU as and GNU ld are standard.
7287 The @option{-mrnames} switch says to output code using the MIPS software
7288 names for the registers, instead of the hardware names (ie, @var{a0}
7289 instead of @var{$4}). The only known assembler that supports this option
7290 is the Algorithmics assembler.
7296 The @option{-mgpopt} switch says to write all of the data declarations
7297 before the instructions in the text section, this allows the MIPS
7298 assembler to generate one word memory references instead of using two
7299 words for short global or static data items. This is on by default if
7300 optimization is selected.
7306 For each non-inline function processed, the @option{-mstats} switch
7307 causes the compiler to emit one line to the standard error file to
7308 print statistics about the program (number of registers saved, stack
7315 The @option{-mmemcpy} switch makes all block moves call the appropriate
7316 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7317 generating inline code.
7320 @itemx -mno-mips-tfile
7321 @opindex mmips-tfile
7322 @opindex mno-mips-tfile
7323 The @option{-mno-mips-tfile} switch causes the compiler not
7324 postprocess the object file with the @file{mips-tfile} program,
7325 after the MIPS assembler has generated it to add debug support. If
7326 @file{mips-tfile} is not run, then no local variables will be
7327 available to the debugger. In addition, @file{stage2} and
7328 @file{stage3} objects will have the temporary file names passed to the
7329 assembler embedded in the object file, which means the objects will
7330 not compare the same. The @option{-mno-mips-tfile} switch should only
7331 be used when there are bugs in the @file{mips-tfile} program that
7332 prevents compilation.
7335 @opindex msoft-float
7336 Generate output containing library calls for floating point.
7337 @strong{Warning:} the requisite libraries are not part of GCC@.
7338 Normally the facilities of the machine's usual C compiler are used, but
7339 this can't be done directly in cross-compilation. You must make your
7340 own arrangements to provide suitable library functions for
7344 @opindex mhard-float
7345 Generate output containing floating point instructions. This is the
7346 default if you use the unmodified sources.
7349 @itemx -mno-abicalls
7351 @opindex mno-abicalls
7352 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7353 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7354 position independent code.
7357 @itemx -mno-long-calls
7358 @opindex mlong-calls
7359 @opindex mno-long-calls
7360 Do all calls with the @samp{JALR} instruction, which requires
7361 loading up a function's address into a register before the call.
7362 You need to use this switch, if you call outside of the current
7363 512 megabyte segment to functions that are not through pointers.
7366 @itemx -mno-half-pic
7368 @opindex mno-half-pic
7369 Put pointers to extern references into the data section and load them
7370 up, rather than put the references in the text section.
7372 @item -membedded-pic
7373 @itemx -mno-embedded-pic
7374 @opindex membedded-pic
7375 @opindex mno-embedded-pic
7376 Generate PIC code suitable for some embedded systems. All calls are
7377 made using PC relative address, and all data is addressed using the $gp
7378 register. No more than 65536 bytes of global data may be used. This
7379 requires GNU as and GNU ld which do most of the work. This currently
7380 only works on targets which use ECOFF; it does not work with ELF@.
7382 @item -membedded-data
7383 @itemx -mno-embedded-data
7384 @opindex membedded-data
7385 @opindex mno-embedded-data
7386 Allocate variables to the read-only data section first if possible, then
7387 next in the small data section if possible, otherwise in data. This gives
7388 slightly slower code than the default, but reduces the amount of RAM required
7389 when executing, and thus may be preferred for some embedded systems.
7391 @item -muninit-const-in-rodata
7392 @itemx -mno-uninit-const-in-rodata
7393 @opindex muninit-const-in-rodata
7394 @opindex mno-uninit-const-in-rodata
7395 When used together with @option{-membedded-data}, it will always store uninitialized
7396 const variables in the read-only data section.
7398 @item -msingle-float
7399 @itemx -mdouble-float
7400 @opindex msingle-float
7401 @opindex mdouble-float
7402 The @option{-msingle-float} switch tells gcc to assume that the floating
7403 point coprocessor only supports single precision operations, as on the
7404 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7405 double precision operations. This is the default.
7411 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7412 as on the @samp{r4650} chip.
7416 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7417 @option{-mcpu=r4650}.
7423 Enable 16-bit instructions.
7427 Use the entry and exit pseudo ops. This option can only be used with
7432 Compile code for the processor in little endian mode.
7433 The requisite libraries are assumed to exist.
7437 Compile code for the processor in big endian mode.
7438 The requisite libraries are assumed to exist.
7442 @cindex smaller data references (MIPS)
7443 @cindex gp-relative references (MIPS)
7444 Put global and static items less than or equal to @var{num} bytes into
7445 the small data or bss sections instead of the normal data or bss
7446 section. This allows the assembler to emit one word memory reference
7447 instructions based on the global pointer (@var{gp} or @var{$28}),
7448 instead of the normal two words used. By default, @var{num} is 8 when
7449 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7450 @option{-G @var{num}} switch is also passed to the assembler and linker.
7451 All modules should be compiled with the same @option{-G @var{num}}
7456 Tell the MIPS assembler to not run its preprocessor over user
7457 assembler files (with a @samp{.s} suffix) when assembling them.
7461 Pass an option to gas which will cause nops to be inserted if
7462 the read of the destination register of an mfhi or mflo instruction
7463 occurs in the following two instructions.
7467 Do not include the default crt0.
7469 @item -mflush-func=@var{func}
7470 @itemx -mno-flush-func
7471 @opindex mflush-func
7472 Specifies the function to call to flush the I and D caches, or to not
7473 call any such function. If called, the function must take the same
7474 arguments as the common @code{_flush_func()}, that is, the address of the
7475 memory range for which the cache is being flushed, the size of the
7476 memory range, and the number 3 (to flush both caches). The default
7477 depends on the target gcc was configured for, but commonly is either
7478 @samp{_flush_func} or @samp{__cpu_flush}.
7481 These options are defined by the macro
7482 @code{TARGET_SWITCHES} in the machine description. The default for the
7483 options is also defined by that macro, which enables you to change the
7486 @node i386 and x86-64 Options
7487 @subsection Intel 386 and AMD x86-64 Options
7488 @cindex i386 Options
7489 @cindex x86-64 Options
7490 @cindex Intel 386 Options
7491 @cindex AMD x86-64 Options
7493 These @samp{-m} options are defined for the i386 and x86-64 family of
7497 @item -mcpu=@var{cpu-type}
7499 Tune to @var{cpu-type} everything applicable about the generated code, except
7500 for the ABI and the set of available instructions. The choices for
7501 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7502 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7503 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7504 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7505 and @samp{athlon-mp}.
7507 While picking a specific @var{cpu-type} will schedule things appropriately
7508 for that particular chip, the compiler will not generate any code that
7509 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7510 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7511 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7512 AMD chips as opposed to the Intel ones.
7514 @item -march=@var{cpu-type}
7516 Generate instructions for the machine type @var{cpu-type}. The choices
7517 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7518 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7527 @opindex mpentiumpro
7528 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7529 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7530 These synonyms are deprecated.
7532 @item -mfpmath=@var{unit}
7534 generate floating point arithmetics for selected unit @var{unit}. the choices
7539 Use the standard 387 floating point coprocessor present majority of chips and
7540 emulated otherwise. Code compiled with this option will run almost everywhere.
7541 The temporary results are computed in 80bit precesion instead of precision
7542 specified by the type resulting in slightly different results compared to most
7543 of other chips. See @option{-ffloat-store} for more detailed description.
7545 This is the default choice for i386 compiler.
7548 Use scalar floating point instructions present in the SSE instruction set.
7549 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7550 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7551 instruction set supports only single precision arithmetics, thus the double and
7552 extended precision arithmetics is still done using 387. Later version, present
7553 only in Pentium4 and the future AMD x86-64 chips supports double precision
7556 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7557 @option{-msse2} switches to enable SSE extensions and make this option
7558 effective. For x86-64 compiler, these extensions are enabled by default.
7560 The resulting code should be considerably faster in majority of cases and avoid
7561 the numerical instability problems of 387 code, but may break some existing
7562 code that expects temporaries to be 80bit.
7564 This is the default choice for x86-64 compiler.
7567 Attempt to utilize both instruction sets at once. This effectivly double the
7568 amount of available registers and on chips with separate execution units for
7569 387 and SSE the execution resources too. Use this option with care, as it is
7570 still experimental, because gcc register allocator does not model separate
7571 functional units well resulting in instable performance.
7574 @item -masm=@var{dialect}
7575 @opindex masm=@var{dialect}
7576 Output asm instructions using selected @var{dialect}. Supported choices are
7577 @samp{intel} or @samp{att} (the default one).
7582 @opindex mno-ieee-fp
7583 Control whether or not the compiler uses IEEE floating point
7584 comparisons. These handle correctly the case where the result of a
7585 comparison is unordered.
7588 @opindex msoft-float
7589 Generate output containing library calls for floating point.
7590 @strong{Warning:} the requisite libraries are not part of GCC@.
7591 Normally the facilities of the machine's usual C compiler are used, but
7592 this can't be done directly in cross-compilation. You must make your
7593 own arrangements to provide suitable library functions for
7596 On machines where a function returns floating point results in the 80387
7597 register stack, some floating point opcodes may be emitted even if
7598 @option{-msoft-float} is used.
7600 @item -mno-fp-ret-in-387
7601 @opindex mno-fp-ret-in-387
7602 Do not use the FPU registers for return values of functions.
7604 The usual calling convention has functions return values of types
7605 @code{float} and @code{double} in an FPU register, even if there
7606 is no FPU@. The idea is that the operating system should emulate
7609 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7610 in ordinary CPU registers instead.
7612 @item -mno-fancy-math-387
7613 @opindex mno-fancy-math-387
7614 Some 387 emulators do not support the @code{sin}, @code{cos} and
7615 @code{sqrt} instructions for the 387. Specify this option to avoid
7616 generating those instructions. This option is the default on FreeBSD@.
7617 As of revision 2.6.1, these instructions are not generated unless you
7618 also use the @option{-funsafe-math-optimizations} switch.
7620 @item -malign-double
7621 @itemx -mno-align-double
7622 @opindex malign-double
7623 @opindex mno-align-double
7624 Control whether GCC aligns @code{double}, @code{long double}, and
7625 @code{long long} variables on a two word boundary or a one word
7626 boundary. Aligning @code{double} variables on a two word boundary will
7627 produce code that runs somewhat faster on a @samp{Pentium} at the
7628 expense of more memory.
7630 @item -m128bit-long-double
7631 @opindex m128bit-long-double
7632 Control the size of @code{long double} type. i386 application binary interface
7633 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7634 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7635 impossible to reach with 12 byte long doubles in the array accesses.
7637 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7638 structures and arrays containing @code{long double} will change their size as
7639 well as function calling convention for function taking @code{long double}
7642 @item -m96bit-long-double
7643 @opindex m96bit-long-double
7644 Set the size of @code{long double} to 96 bits as required by the i386
7645 application binary interface. This is the default.
7648 @itemx -mno-svr3-shlib
7649 @opindex msvr3-shlib
7650 @opindex mno-svr3-shlib
7651 Control whether GCC places uninitialized local variables into the
7652 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7653 into @code{bss}. These options are meaningful only on System V Release 3.
7657 Use a different function-calling convention, in which functions that
7658 take a fixed number of arguments return with the @code{ret} @var{num}
7659 instruction, which pops their arguments while returning. This saves one
7660 instruction in the caller since there is no need to pop the arguments
7663 You can specify that an individual function is called with this calling
7664 sequence with the function attribute @samp{stdcall}. You can also
7665 override the @option{-mrtd} option by using the function attribute
7666 @samp{cdecl}. @xref{Function Attributes}.
7668 @strong{Warning:} this calling convention is incompatible with the one
7669 normally used on Unix, so you cannot use it if you need to call
7670 libraries compiled with the Unix compiler.
7672 Also, you must provide function prototypes for all functions that
7673 take variable numbers of arguments (including @code{printf});
7674 otherwise incorrect code will be generated for calls to those
7677 In addition, seriously incorrect code will result if you call a
7678 function with too many arguments. (Normally, extra arguments are
7679 harmlessly ignored.)
7681 @item -mregparm=@var{num}
7683 Control how many registers are used to pass integer arguments. By
7684 default, no registers are used to pass arguments, and at most 3
7685 registers can be used. You can control this behavior for a specific
7686 function by using the function attribute @samp{regparm}.
7687 @xref{Function Attributes}.
7689 @strong{Warning:} if you use this switch, and
7690 @var{num} is nonzero, then you must build all modules with the same
7691 value, including any libraries. This includes the system libraries and
7694 @item -mpreferred-stack-boundary=@var{num}
7695 @opindex mpreferred-stack-boundary
7696 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7697 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7698 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7699 size (@option{-Os}), in which case the default is the minimum correct
7700 alignment (4 bytes for x86, and 8 bytes for x86-64).
7702 On Pentium and PentiumPro, @code{double} and @code{long double} values
7703 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7704 suffer significant run time performance penalties. On Pentium III, the
7705 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7706 penalties if it is not 16 byte aligned.
7708 To ensure proper alignment of this values on the stack, the stack boundary
7709 must be as aligned as that required by any value stored on the stack.
7710 Further, every function must be generated such that it keeps the stack
7711 aligned. Thus calling a function compiled with a higher preferred
7712 stack boundary from a function compiled with a lower preferred stack
7713 boundary will most likely misalign the stack. It is recommended that
7714 libraries that use callbacks always use the default setting.
7716 This extra alignment does consume extra stack space, and generally
7717 increases code size. Code that is sensitive to stack space usage, such
7718 as embedded systems and operating system kernels, may want to reduce the
7719 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7735 These switches enable or disable the use of built-in functions that allow
7736 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7738 @xref{X86 Built-in Functions}, for details of the functions enabled
7739 and disabled by these switches.
7742 @itemx -mno-push-args
7744 @opindex mno-push-args
7745 Use PUSH operations to store outgoing parameters. This method is shorter
7746 and usually equally fast as method using SUB/MOV operations and is enabled
7747 by default. In some cases disabling it may improve performance because of
7748 improved scheduling and reduced dependencies.
7750 @item -maccumulate-outgoing-args
7751 @opindex maccumulate-outgoing-args
7752 If enabled, the maximum amount of space required for outgoing arguments will be
7753 computed in the function prologue. This is faster on most modern CPUs
7754 because of reduced dependencies, improved scheduling and reduced stack usage
7755 when preferred stack boundary is not equal to 2. The drawback is a notable
7756 increase in code size. This switch implies @option{-mno-push-args}.
7760 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7761 on thread-safe exception handling must compile and link all code with the
7762 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7763 @option{-D_MT}; when linking, it links in a special thread helper library
7764 @option{-lmingwthrd} which cleans up per thread exception handling data.
7766 @item -mno-align-stringops
7767 @opindex mno-align-stringops
7768 Do not align destination of inlined string operations. This switch reduces
7769 code size and improves performance in case the destination is already aligned,
7770 but gcc don't know about it.
7772 @item -minline-all-stringops
7773 @opindex minline-all-stringops
7774 By default GCC inlines string operations only when destination is known to be
7775 aligned at least to 4 byte boundary. This enables more inlining, increase code
7776 size, but may improve performance of code that depends on fast memcpy, strlen
7777 and memset for short lengths.
7779 @item -momit-leaf-frame-pointer
7780 @opindex momit-leaf-frame-pointer
7781 Don't keep the frame pointer in a register for leaf functions. This
7782 avoids the instructions to save, set up and restore frame pointers and
7783 makes an extra register available in leaf functions. The option
7784 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7785 which might make debugging harder.
7788 These @samp{-m} switches are supported in addition to the above
7789 on AMD x86-64 processors in 64-bit environments.
7796 Generate code for a 32-bit or 64-bit environment.
7797 The 32-bit environment sets int, long and pointer to 32 bits and
7798 generates code that runs on any i386 system.
7799 The 64-bit environment sets int to 32 bits and long and pointer
7800 to 64 bits and generates code for AMD's x86-64 architecture.
7803 @opindex no-red-zone
7804 Do not use a so called red zone for x86-64 code. The red zone is mandated
7805 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7806 stack pointer that will not be modified by signal or interrupt handlers
7807 and therefore can be used for temporary data without adjusting the stack
7808 pointer. The flag @option{-mno-red-zone} disables this red zone.
7812 @subsection HPPA Options
7813 @cindex HPPA Options
7815 These @samp{-m} options are defined for the HPPA family of computers:
7818 @item -march=@var{architecture-type}
7820 Generate code for the specified architecture. The choices for
7821 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7822 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7823 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7824 architecture option for your machine. Code compiled for lower numbered
7825 architectures will run on higher numbered architectures, but not the
7828 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7829 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7833 @itemx -mpa-risc-1-1
7834 @itemx -mpa-risc-2-0
7835 @opindex mpa-risc-1-0
7836 @opindex mpa-risc-1-1
7837 @opindex mpa-risc-2-0
7838 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7841 @opindex mbig-switch
7842 Generate code suitable for big switch tables. Use this option only if
7843 the assembler/linker complain about out of range branches within a switch
7846 @item -mjump-in-delay
7847 @opindex mjump-in-delay
7848 Fill delay slots of function calls with unconditional jump instructions
7849 by modifying the return pointer for the function call to be the target
7850 of the conditional jump.
7852 @item -mdisable-fpregs
7853 @opindex mdisable-fpregs
7854 Prevent floating point registers from being used in any manner. This is
7855 necessary for compiling kernels which perform lazy context switching of
7856 floating point registers. If you use this option and attempt to perform
7857 floating point operations, the compiler will abort.
7859 @item -mdisable-indexing
7860 @opindex mdisable-indexing
7861 Prevent the compiler from using indexing address modes. This avoids some
7862 rather obscure problems when compiling MIG generated code under MACH@.
7864 @item -mno-space-regs
7865 @opindex mno-space-regs
7866 Generate code that assumes the target has no space registers. This allows
7867 GCC to generate faster indirect calls and use unscaled index address modes.
7869 Such code is suitable for level 0 PA systems and kernels.
7871 @item -mfast-indirect-calls
7872 @opindex mfast-indirect-calls
7873 Generate code that assumes calls never cross space boundaries. This
7874 allows GCC to emit code which performs faster indirect calls.
7876 This option will not work in the presence of shared libraries or nested
7879 @item -mlong-load-store
7880 @opindex mlong-load-store
7881 Generate 3-instruction load and store sequences as sometimes required by
7882 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7885 @item -mportable-runtime
7886 @opindex mportable-runtime
7887 Use the portable calling conventions proposed by HP for ELF systems.
7891 Enable the use of assembler directives only GAS understands.
7893 @item -mschedule=@var{cpu-type}
7895 Schedule code according to the constraints for the machine type
7896 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7897 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
7898 @file{/usr/lib/sched.models} on an HP-UX system to determine the
7899 proper scheduling option for your machine.
7902 @opindex mlinker-opt
7903 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7904 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7905 in which they give bogus error messages when linking some programs.
7908 @opindex msoft-float
7909 Generate output containing library calls for floating point.
7910 @strong{Warning:} the requisite libraries are not available for all HPPA
7911 targets. Normally the facilities of the machine's usual C compiler are
7912 used, but this cannot be done directly in cross-compilation. You must make
7913 your own arrangements to provide suitable library functions for
7914 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7915 does provide software floating point support.
7917 @option{-msoft-float} changes the calling convention in the output file;
7918 therefore, it is only useful if you compile @emph{all} of a program with
7919 this option. In particular, you need to compile @file{libgcc.a}, the
7920 library that comes with GCC, with @option{-msoft-float} in order for
7924 @node Intel 960 Options
7925 @subsection Intel 960 Options
7927 These @samp{-m} options are defined for the Intel 960 implementations:
7930 @item -m@var{cpu-type}
7938 Assume the defaults for the machine type @var{cpu-type} for some of
7939 the other options, including instruction scheduling, floating point
7940 support, and addressing modes. The choices for @var{cpu-type} are
7941 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7942 @samp{sa}, and @samp{sb}.
7949 @opindex msoft-float
7950 The @option{-mnumerics} option indicates that the processor does support
7951 floating-point instructions. The @option{-msoft-float} option indicates
7952 that floating-point support should not be assumed.
7954 @item -mleaf-procedures
7955 @itemx -mno-leaf-procedures
7956 @opindex mleaf-procedures
7957 @opindex mno-leaf-procedures
7958 Do (or do not) attempt to alter leaf procedures to be callable with the
7959 @code{bal} instruction as well as @code{call}. This will result in more
7960 efficient code for explicit calls when the @code{bal} instruction can be
7961 substituted by the assembler or linker, but less efficient code in other
7962 cases, such as calls via function pointers, or using a linker that doesn't
7963 support this optimization.
7966 @itemx -mno-tail-call
7968 @opindex mno-tail-call
7969 Do (or do not) make additional attempts (beyond those of the
7970 machine-independent portions of the compiler) to optimize tail-recursive
7971 calls into branches. You may not want to do this because the detection of
7972 cases where this is not valid is not totally complete. The default is
7973 @option{-mno-tail-call}.
7975 @item -mcomplex-addr
7976 @itemx -mno-complex-addr
7977 @opindex mcomplex-addr
7978 @opindex mno-complex-addr
7979 Assume (or do not assume) that the use of a complex addressing mode is a
7980 win on this implementation of the i960. Complex addressing modes may not
7981 be worthwhile on the K-series, but they definitely are on the C-series.
7982 The default is currently @option{-mcomplex-addr} for all processors except
7986 @itemx -mno-code-align
7987 @opindex mcode-align
7988 @opindex mno-code-align
7989 Align code to 8-byte boundaries for faster fetching (or don't bother).
7990 Currently turned on by default for C-series implementations only.
7993 @item -mclean-linkage
7994 @itemx -mno-clean-linkage
7995 @opindex mclean-linkage
7996 @opindex mno-clean-linkage
7997 These options are not fully implemented.
8001 @itemx -mic2.0-compat
8002 @itemx -mic3.0-compat
8004 @opindex mic2.0-compat
8005 @opindex mic3.0-compat
8006 Enable compatibility with iC960 v2.0 or v3.0.
8010 @opindex masm-compat
8012 Enable compatibility with the iC960 assembler.
8014 @item -mstrict-align
8015 @itemx -mno-strict-align
8016 @opindex mstrict-align
8017 @opindex mno-strict-align
8018 Do not permit (do permit) unaligned accesses.
8022 Enable structure-alignment compatibility with Intel's gcc release version
8023 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8025 @item -mlong-double-64
8026 @opindex mlong-double-64
8027 Implement type @samp{long double} as 64-bit floating point numbers.
8028 Without the option @samp{long double} is implemented by 80-bit
8029 floating point numbers. The only reason we have it because there is
8030 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8031 is only useful for people using soft-float targets. Otherwise, we
8032 should recommend against use of it.
8036 @node DEC Alpha Options
8037 @subsection DEC Alpha Options
8039 These @samp{-m} options are defined for the DEC Alpha implementations:
8042 @item -mno-soft-float
8044 @opindex mno-soft-float
8045 @opindex msoft-float
8046 Use (do not use) the hardware floating-point instructions for
8047 floating-point operations. When @option{-msoft-float} is specified,
8048 functions in @file{libgcc.a} will be used to perform floating-point
8049 operations. Unless they are replaced by routines that emulate the
8050 floating-point operations, or compiled in such a way as to call such
8051 emulations routines, these routines will issue floating-point
8052 operations. If you are compiling for an Alpha without floating-point
8053 operations, you must ensure that the library is built so as not to call
8056 Note that Alpha implementations without floating-point operations are
8057 required to have floating-point registers.
8062 @opindex mno-fp-regs
8063 Generate code that uses (does not use) the floating-point register set.
8064 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8065 register set is not used, floating point operands are passed in integer
8066 registers as if they were integers and floating-point results are passed
8067 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8068 so any function with a floating-point argument or return value called by code
8069 compiled with @option{-mno-fp-regs} must also be compiled with that
8072 A typical use of this option is building a kernel that does not use,
8073 and hence need not save and restore, any floating-point registers.
8077 The Alpha architecture implements floating-point hardware optimized for
8078 maximum performance. It is mostly compliant with the IEEE floating
8079 point standard. However, for full compliance, software assistance is
8080 required. This option generates code fully IEEE compliant code
8081 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8082 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8083 defined during compilation. The resulting code is less efficient but is
8084 able to correctly support denormalized numbers and exceptional IEEE
8085 values such as not-a-number and plus/minus infinity. Other Alpha
8086 compilers call this option @option{-ieee_with_no_inexact}.
8088 @item -mieee-with-inexact
8089 @opindex mieee-with-inexact
8090 This is like @option{-mieee} except the generated code also maintains
8091 the IEEE @var{inexact-flag}. Turning on this option causes the
8092 generated code to implement fully-compliant IEEE math. In addition to
8093 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8094 macro. On some Alpha implementations the resulting code may execute
8095 significantly slower than the code generated by default. Since there is
8096 very little code that depends on the @var{inexact-flag}, you should
8097 normally not specify this option. Other Alpha compilers call this
8098 option @option{-ieee_with_inexact}.
8100 @item -mfp-trap-mode=@var{trap-mode}
8101 @opindex mfp-trap-mode
8102 This option controls what floating-point related traps are enabled.
8103 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8104 The trap mode can be set to one of four values:
8108 This is the default (normal) setting. The only traps that are enabled
8109 are the ones that cannot be disabled in software (e.g., division by zero
8113 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8117 Like @samp{su}, but the instructions are marked to be safe for software
8118 completion (see Alpha architecture manual for details).
8121 Like @samp{su}, but inexact traps are enabled as well.
8124 @item -mfp-rounding-mode=@var{rounding-mode}
8125 @opindex mfp-rounding-mode
8126 Selects the IEEE rounding mode. Other Alpha compilers call this option
8127 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8132 Normal IEEE rounding mode. Floating point numbers are rounded towards
8133 the nearest machine number or towards the even machine number in case
8137 Round towards minus infinity.
8140 Chopped rounding mode. Floating point numbers are rounded towards zero.
8143 Dynamic rounding mode. A field in the floating point control register
8144 (@var{fpcr}, see Alpha architecture reference manual) controls the
8145 rounding mode in effect. The C library initializes this register for
8146 rounding towards plus infinity. Thus, unless your program modifies the
8147 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8150 @item -mtrap-precision=@var{trap-precision}
8151 @opindex mtrap-precision
8152 In the Alpha architecture, floating point traps are imprecise. This
8153 means without software assistance it is impossible to recover from a
8154 floating trap and program execution normally needs to be terminated.
8155 GCC can generate code that can assist operating system trap handlers
8156 in determining the exact location that caused a floating point trap.
8157 Depending on the requirements of an application, different levels of
8158 precisions can be selected:
8162 Program precision. This option is the default and means a trap handler
8163 can only identify which program caused a floating point exception.
8166 Function precision. The trap handler can determine the function that
8167 caused a floating point exception.
8170 Instruction precision. The trap handler can determine the exact
8171 instruction that caused a floating point exception.
8174 Other Alpha compilers provide the equivalent options called
8175 @option{-scope_safe} and @option{-resumption_safe}.
8177 @item -mieee-conformant
8178 @opindex mieee-conformant
8179 This option marks the generated code as IEEE conformant. You must not
8180 use this option unless you also specify @option{-mtrap-precision=i} and either
8181 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8182 is to emit the line @samp{.eflag 48} in the function prologue of the
8183 generated assembly file. Under DEC Unix, this has the effect that
8184 IEEE-conformant math library routines will be linked in.
8186 @item -mbuild-constants
8187 @opindex mbuild-constants
8188 Normally GCC examines a 32- or 64-bit integer constant to
8189 see if it can construct it from smaller constants in two or three
8190 instructions. If it cannot, it will output the constant as a literal and
8191 generate code to load it from the data segment at runtime.
8193 Use this option to require GCC to construct @emph{all} integer constants
8194 using code, even if it takes more instructions (the maximum is six).
8196 You would typically use this option to build a shared library dynamic
8197 loader. Itself a shared library, it must relocate itself in memory
8198 before it can find the variables and constants in its own data segment.
8204 Select whether to generate code to be assembled by the vendor-supplied
8205 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8223 Indicate whether GCC should generate code to use the optional BWX,
8224 CIX, FIX and MAX instruction sets. The default is to use the instruction
8225 sets supported by the CPU type specified via @option{-mcpu=} option or that
8226 of the CPU on which GCC was built if none was specified.
8231 @opindex mfloat-ieee
8232 Generate code that uses (does not use) VAX F and G floating point
8233 arithmetic instead of IEEE single and double precision.
8235 @item -mexplicit-relocs
8236 @itemx -mno-explicit-relocs
8237 @opindex mexplicit-relocs
8238 @opindex mno-explicit-relocs
8239 Older Alpha assemblers provided no way to generate symbol relocations
8240 except via assembler macros. Use of these macros does not allow
8241 optimial instruction scheduling. GNU binutils as of version 2.12
8242 supports a new syntax that allows the compiler to explicitly mark
8243 which relocations should apply to which instructions. This option
8244 is mostly useful for debugging, as GCC detects the capabilities of
8245 the assembler when it is built and sets the default accordingly.
8249 @opindex msmall-data
8250 @opindex mlarge-data
8251 When @option{-mexplicit-relocs} is in effect, static data is
8252 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8253 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8254 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8255 16-bit relocations off of the @code{$gp} register. This limits the
8256 size of the small data area to 64KB, but allows the variables to be
8257 directly accessed via a single instruction.
8259 The default is @option{-mlarge-data}. With this option the data area
8260 is limited to just below 2GB. Programs that require more than 2GB of
8261 data must use @code{malloc} or @code{mmap} to allocate the data in the
8262 heap instead of in the program's data segment.
8264 When generating code for shared libraries, @option{-fpic} implies
8265 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8267 @item -mcpu=@var{cpu_type}
8269 Set the instruction set and instruction scheduling parameters for
8270 machine type @var{cpu_type}. You can specify either the @samp{EV}
8271 style name or the corresponding chip number. GCC supports scheduling
8272 parameters for the EV4, EV5 and EV6 family of processors and will
8273 choose the default values for the instruction set from the processor
8274 you specify. If you do not specify a processor type, GCC will default
8275 to the processor on which the compiler was built.
8277 Supported values for @var{cpu_type} are
8283 Schedules as an EV4 and has no instruction set extensions.
8287 Schedules as an EV5 and has no instruction set extensions.
8291 Schedules as an EV5 and supports the BWX extension.
8296 Schedules as an EV5 and supports the BWX and MAX extensions.
8300 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8304 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8307 @item -mtune=@var{cpu_type}
8309 Set only the instruction scheduling parameters for machine type
8310 @var{cpu_type}. The instruction set is not changed.
8312 @item -mmemory-latency=@var{time}
8313 @opindex mmemory-latency
8314 Sets the latency the scheduler should assume for typical memory
8315 references as seen by the application. This number is highly
8316 dependent on the memory access patterns used by the application
8317 and the size of the external cache on the machine.
8319 Valid options for @var{time} are
8323 A decimal number representing clock cycles.
8329 The compiler contains estimates of the number of clock cycles for
8330 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8331 (also called Dcache, Scache, and Bcache), as well as to main memory.
8332 Note that L3 is only valid for EV5.
8337 @node DEC Alpha/VMS Options
8338 @subsection DEC Alpha/VMS Options
8340 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8343 @item -mvms-return-codes
8344 @opindex mvms-return-codes
8345 Return VMS condition codes from main. The default is to return POSIX
8346 style condition (e.g.@ error) codes.
8349 @node Clipper Options
8350 @subsection Clipper Options
8352 These @samp{-m} options are defined for the Clipper implementations:
8357 Produce code for a C300 Clipper processor. This is the default.
8361 Produce code for a C400 Clipper processor, i.e.@: use floating point
8365 @node H8/300 Options
8366 @subsection H8/300 Options
8368 These @samp{-m} options are defined for the H8/300 implementations:
8373 Shorten some address references at link time, when possible; uses the
8374 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8375 ld.info, Using ld}, for a fuller description.
8379 Generate code for the H8/300H@.
8383 Generate code for the H8/S@.
8387 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8391 Make @code{int} data 32 bits by default.
8395 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8396 The default for the H8/300H and H8/S is to align longs and floats on 4
8398 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8399 This option has no effect on the H8/300.
8403 @subsection SH Options
8405 These @samp{-m} options are defined for the SH implementations:
8410 Generate code for the SH1.
8414 Generate code for the SH2.
8418 Generate code for the SH3.
8422 Generate code for the SH3e.
8426 Generate code for the SH4 without a floating-point unit.
8428 @item -m4-single-only
8429 @opindex m4-single-only
8430 Generate code for the SH4 with a floating-point unit that only
8431 supports single-precision arithmetic.
8435 Generate code for the SH4 assuming the floating-point unit is in
8436 single-precision mode by default.
8440 Generate code for the SH4.
8444 Compile code for the processor in big endian mode.
8448 Compile code for the processor in little endian mode.
8452 Align doubles at 64-bit boundaries. Note that this changes the calling
8453 conventions, and thus some functions from the standard C library will
8454 not work unless you recompile it first with @option{-mdalign}.
8458 Shorten some address references at link time, when possible; uses the
8459 linker option @option{-relax}.
8463 Use 32-bit offsets in @code{switch} tables. The default is to use
8468 Enable the use of the instruction @code{fmovd}.
8472 Comply with the calling conventions defined by Hitachi.
8476 Mark the @code{MAC} register as call-clobbered, even if
8477 @option{-mhitachi} is given.
8481 Increase IEEE-compliance of floating-point code.
8485 Dump instruction size and location in the assembly code.
8489 This option is deprecated. It pads structures to multiple of 4 bytes,
8490 which is incompatible with the SH ABI@.
8494 Optimize for space instead of speed. Implied by @option{-Os}.
8498 When generating position-independent code, emit function calls using
8499 the Global Offset Table instead of the Procedure Linkage Table.
8503 Generate a library function call to invalidate instruction cache
8504 entries, after fixing up a trampoline. This library function call
8505 doesn't assume it can write to the whole memory address space. This
8506 is the default when the target is @code{sh-*-linux*}.
8509 @node System V Options
8510 @subsection Options for System V
8512 These additional options are available on System V Release 4 for
8513 compatibility with other compilers on those systems:
8518 Create a shared object.
8519 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8523 Identify the versions of each tool used by the compiler, in a
8524 @code{.ident} assembler directive in the output.
8528 Refrain from adding @code{.ident} directives to the output file (this is
8531 @item -YP,@var{dirs}
8533 Search the directories @var{dirs}, and no others, for libraries
8534 specified with @option{-l}.
8538 Look in the directory @var{dir} to find the M4 preprocessor.
8539 The assembler uses this option.
8540 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8541 @c the generic assembler that comes with Solaris takes just -Ym.
8544 @node TMS320C3x/C4x Options
8545 @subsection TMS320C3x/C4x Options
8546 @cindex TMS320C3x/C4x Options
8548 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8552 @item -mcpu=@var{cpu_type}
8554 Set the instruction set, register set, and instruction scheduling
8555 parameters for machine type @var{cpu_type}. Supported values for
8556 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8557 @samp{c44}. The default is @samp{c40} to generate code for the
8562 @itemx -msmall-memory
8564 @opindex mbig-memory
8566 @opindex msmall-memory
8568 Generates code for the big or small memory model. The small memory
8569 model assumed that all data fits into one 64K word page. At run-time
8570 the data page (DP) register must be set to point to the 64K page
8571 containing the .bss and .data program sections. The big memory model is
8572 the default and requires reloading of the DP register for every direct
8579 Allow (disallow) allocation of general integer operands into the block
8586 Enable (disable) generation of code using decrement and branch,
8587 DBcond(D), instructions. This is enabled by default for the C4x. To be
8588 on the safe side, this is disabled for the C3x, since the maximum
8589 iteration count on the C3x is @math{2^23 + 1} (but who iterates loops more than
8590 @math{2^23} times on the C3x?). Note that GCC will try to reverse a loop so
8591 that it can utilise the decrement and branch instruction, but will give
8592 up if there is more than one memory reference in the loop. Thus a loop
8593 where the loop counter is decremented can generate slightly more
8594 efficient code, in cases where the RPTB instruction cannot be utilised.
8596 @item -mdp-isr-reload
8598 @opindex mdp-isr-reload
8600 Force the DP register to be saved on entry to an interrupt service
8601 routine (ISR), reloaded to point to the data section, and restored on
8602 exit from the ISR@. This should not be required unless someone has
8603 violated the small memory model by modifying the DP register, say within
8610 For the C3x use the 24-bit MPYI instruction for integer multiplies
8611 instead of a library call to guarantee 32-bit results. Note that if one
8612 of the operands is a constant, then the multiplication will be performed
8613 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8614 then squaring operations are performed inline instead of a library call.
8617 @itemx -mno-fast-fix
8619 @opindex mno-fast-fix
8620 The C3x/C4x FIX instruction to convert a floating point value to an
8621 integer value chooses the nearest integer less than or equal to the
8622 floating point value rather than to the nearest integer. Thus if the
8623 floating point number is negative, the result will be incorrectly
8624 truncated an additional code is necessary to detect and correct this
8625 case. This option can be used to disable generation of the additional
8626 code required to correct the result.
8632 Enable (disable) generation of repeat block sequences using the RPTB
8633 instruction for zero overhead looping. The RPTB construct is only used
8634 for innermost loops that do not call functions or jump across the loop
8635 boundaries. There is no advantage having nested RPTB loops due to the
8636 overhead required to save and restore the RC, RS, and RE registers.
8637 This is enabled by default with @option{-O2}.
8639 @item -mrpts=@var{count}
8643 Enable (disable) the use of the single instruction repeat instruction
8644 RPTS@. If a repeat block contains a single instruction, and the loop
8645 count can be guaranteed to be less than the value @var{count}, GCC will
8646 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8647 then a RPTS will be emitted even if the loop count cannot be determined
8648 at compile time. Note that the repeated instruction following RPTS does
8649 not have to be reloaded from memory each iteration, thus freeing up the
8650 CPU buses for operands. However, since interrupts are blocked by this
8651 instruction, it is disabled by default.
8653 @item -mloop-unsigned
8654 @itemx -mno-loop-unsigned
8655 @opindex mloop-unsigned
8656 @opindex mno-loop-unsigned
8657 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8658 is @math{2^31 + 1} since these instructions test if the iteration count is
8659 negative to terminate the loop. If the iteration count is unsigned
8660 there is a possibility than the @math{2^31 + 1} maximum iteration count may be
8661 exceeded. This switch allows an unsigned iteration count.
8665 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8666 with. This also enforces compatibility with the API employed by the TI
8667 C3x C compiler. For example, long doubles are passed as structures
8668 rather than in floating point registers.
8674 Generate code that uses registers (stack) for passing arguments to functions.
8675 By default, arguments are passed in registers where possible rather
8676 than by pushing arguments on to the stack.
8678 @item -mparallel-insns
8679 @itemx -mno-parallel-insns
8680 @opindex mparallel-insns
8681 @opindex mno-parallel-insns
8682 Allow the generation of parallel instructions. This is enabled by
8683 default with @option{-O2}.
8685 @item -mparallel-mpy
8686 @itemx -mno-parallel-mpy
8687 @opindex mparallel-mpy
8688 @opindex mno-parallel-mpy
8689 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8690 provided @option{-mparallel-insns} is also specified. These instructions have
8691 tight register constraints which can pessimize the code generation
8697 @subsection V850 Options
8698 @cindex V850 Options
8700 These @samp{-m} options are defined for V850 implementations:
8704 @itemx -mno-long-calls
8705 @opindex mlong-calls
8706 @opindex mno-long-calls
8707 Treat all calls as being far away (near). If calls are assumed to be
8708 far away, the compiler will always load the functions address up into a
8709 register, and call indirect through the pointer.
8715 Do not optimize (do optimize) basic blocks that use the same index
8716 pointer 4 or more times to copy pointer into the @code{ep} register, and
8717 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8718 option is on by default if you optimize.
8720 @item -mno-prolog-function
8721 @itemx -mprolog-function
8722 @opindex mno-prolog-function
8723 @opindex mprolog-function
8724 Do not use (do use) external functions to save and restore registers at
8725 the prolog and epilog of a function. The external functions are slower,
8726 but use less code space if more than one function saves the same number
8727 of registers. The @option{-mprolog-function} option is on by default if
8732 Try to make the code as small as possible. At present, this just turns
8733 on the @option{-mep} and @option{-mprolog-function} options.
8737 Put static or global variables whose size is @var{n} bytes or less into
8738 the tiny data area that register @code{ep} points to. The tiny data
8739 area can hold up to 256 bytes in total (128 bytes for byte references).
8743 Put static or global variables whose size is @var{n} bytes or less into
8744 the small data area that register @code{gp} points to. The small data
8745 area can hold up to 64 kilobytes.
8749 Put static or global variables whose size is @var{n} bytes or less into
8750 the first 32 kilobytes of memory.
8754 Specify that the target processor is the V850.
8757 @opindex mbig-switch
8758 Generate code suitable for big switch tables. Use this option only if
8759 the assembler/linker complain about out of range branches within a switch
8764 @subsection ARC Options
8767 These options are defined for ARC implementations:
8772 Compile code for little endian mode. This is the default.
8776 Compile code for big endian mode.
8779 @opindex mmangle-cpu
8780 Prepend the name of the cpu to all public symbol names.
8781 In multiple-processor systems, there are many ARC variants with different
8782 instruction and register set characteristics. This flag prevents code
8783 compiled for one cpu to be linked with code compiled for another.
8784 No facility exists for handling variants that are ``almost identical''.
8785 This is an all or nothing option.
8787 @item -mcpu=@var{cpu}
8789 Compile code for ARC variant @var{cpu}.
8790 Which variants are supported depend on the configuration.
8791 All variants support @option{-mcpu=base}, this is the default.
8793 @item -mtext=@var{text-section}
8794 @itemx -mdata=@var{data-section}
8795 @itemx -mrodata=@var{readonly-data-section}
8799 Put functions, data, and readonly data in @var{text-section},
8800 @var{data-section}, and @var{readonly-data-section} respectively
8801 by default. This can be overridden with the @code{section} attribute.
8802 @xref{Variable Attributes}.
8807 @subsection NS32K Options
8808 @cindex NS32K options
8810 These are the @samp{-m} options defined for the 32000 series. The default
8811 values for these options depends on which style of 32000 was selected when
8812 the compiler was configured; the defaults for the most common choices are
8820 Generate output for a 32032. This is the default
8821 when the compiler is configured for 32032 and 32016 based systems.
8827 Generate output for a 32332. This is the default
8828 when the compiler is configured for 32332-based systems.
8834 Generate output for a 32532. This is the default
8835 when the compiler is configured for 32532-based systems.
8839 Generate output containing 32081 instructions for floating point.
8840 This is the default for all systems.
8844 Generate output containing 32381 instructions for floating point. This
8845 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8846 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8850 Try and generate multiply-add floating point instructions @code{polyF}
8851 and @code{dotF}. This option is only available if the @option{-m32381}
8852 option is in effect. Using these instructions requires changes to
8853 register allocation which generally has a negative impact on
8854 performance. This option should only be enabled when compiling code
8855 particularly likely to make heavy use of multiply-add instructions.
8858 @opindex mnomulti-add
8859 Do not try and generate multiply-add floating point instructions
8860 @code{polyF} and @code{dotF}. This is the default on all platforms.
8863 @opindex msoft-float
8864 Generate output containing library calls for floating point.
8865 @strong{Warning:} the requisite libraries may not be available.
8868 @opindex mnobitfield
8869 Do not use the bit-field instructions. On some machines it is faster to
8870 use shifting and masking operations. This is the default for the pc532.
8874 Do use the bit-field instructions. This is the default for all platforms
8879 Use a different function-calling convention, in which functions
8880 that take a fixed number of arguments return pop their
8881 arguments on return with the @code{ret} instruction.
8883 This calling convention is incompatible with the one normally
8884 used on Unix, so you cannot use it if you need to call libraries
8885 compiled with the Unix compiler.
8887 Also, you must provide function prototypes for all functions that
8888 take variable numbers of arguments (including @code{printf});
8889 otherwise incorrect code will be generated for calls to those
8892 In addition, seriously incorrect code will result if you call a
8893 function with too many arguments. (Normally, extra arguments are
8894 harmlessly ignored.)
8896 This option takes its name from the 680x0 @code{rtd} instruction.
8901 Use a different function-calling convention where the first two arguments
8902 are passed in registers.
8904 This calling convention is incompatible with the one normally
8905 used on Unix, so you cannot use it if you need to call libraries
8906 compiled with the Unix compiler.
8909 @opindex mnoregparam
8910 Do not pass any arguments in registers. This is the default for all
8915 It is OK to use the sb as an index register which is always loaded with
8916 zero. This is the default for the pc532-netbsd target.
8920 The sb register is not available for use or has not been initialized to
8921 zero by the run time system. This is the default for all targets except
8922 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8923 @option{-fpic} is set.
8927 Many ns32000 series addressing modes use displacements of up to 512MB@.
8928 If an address is above 512MB then displacements from zero can not be used.
8929 This option causes code to be generated which can be loaded above 512MB@.
8930 This may be useful for operating systems or ROM code.
8934 Assume code will be loaded in the first 512MB of virtual address space.
8935 This is the default for all platforms.
8941 @subsection AVR Options
8944 These options are defined for AVR implementations:
8947 @item -mmcu=@var{mcu}
8949 Specify ATMEL AVR instruction set or MCU type.
8951 Instruction set avr1 is for the minimal AVR core, not supported by the C
8952 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8953 attiny11, attiny12, attiny15, attiny28).
8955 Instruction set avr2 (default) is for the classic AVR core with up to
8956 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8957 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8958 at90c8534, at90s8535).
8960 Instruction set avr3 is for the classic AVR core with up to 128K program
8961 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8963 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8964 memory space (MCU types: atmega8, atmega83, atmega85).
8966 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8967 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8968 atmega64, atmega128, at43usb355, at94k).
8972 Output instruction sizes to the asm file.
8974 @item -minit-stack=@var{N}
8975 @opindex minit-stack
8976 Specify the initial stack address, which may be a symbol or numeric value,
8977 @samp{__stack} is the default.
8979 @item -mno-interrupts
8980 @opindex mno-interrupts
8981 Generated code is not compatible with hardware interrupts.
8982 Code size will be smaller.
8984 @item -mcall-prologues
8985 @opindex mcall-prologues
8986 Functions prologues/epilogues expanded as call to appropriate
8987 subroutines. Code size will be smaller.
8989 @item -mno-tablejump
8990 @opindex mno-tablejump
8991 Do not generate tablejump insns which sometimes increase code size.
8994 @opindex mtiny-stack
8995 Change only the low 8 bits of the stack pointer.
8999 @subsection MCore Options
9000 @cindex MCore options
9002 These are the @samp{-m} options defined for the Motorola M*Core
9012 @opindex mno-hardlit
9013 Inline constants into the code stream if it can be done in two
9014 instructions or less.
9022 Use the divide instruction. (Enabled by default).
9024 @item -mrelax-immediate
9025 @itemx -mrelax-immediate
9026 @itemx -mno-relax-immediate
9027 @opindex mrelax-immediate
9028 @opindex mrelax-immediate
9029 @opindex mno-relax-immediate
9030 Allow arbitrary sized immediates in bit operations.
9032 @item -mwide-bitfields
9033 @itemx -mwide-bitfields
9034 @itemx -mno-wide-bitfields
9035 @opindex mwide-bitfields
9036 @opindex mwide-bitfields
9037 @opindex mno-wide-bitfields
9038 Always treat bit-fields as int-sized.
9040 @item -m4byte-functions
9041 @itemx -m4byte-functions
9042 @itemx -mno-4byte-functions
9043 @opindex m4byte-functions
9044 @opindex m4byte-functions
9045 @opindex mno-4byte-functions
9046 Force all functions to be aligned to a four byte boundary.
9048 @item -mcallgraph-data
9049 @itemx -mcallgraph-data
9050 @itemx -mno-callgraph-data
9051 @opindex mcallgraph-data
9052 @opindex mcallgraph-data
9053 @opindex mno-callgraph-data
9054 Emit callgraph information.
9058 @itemx -mno-slow-bytes
9059 @opindex mslow-bytes
9060 @opindex mslow-bytes
9061 @opindex mno-slow-bytes
9062 Prefer word access when reading byte quantities.
9064 @item -mlittle-endian
9065 @itemx -mlittle-endian
9067 @opindex mlittle-endian
9068 @opindex mlittle-endian
9069 @opindex mbig-endian
9070 Generate code for a little endian target.
9078 Generate code for the 210 processor.
9082 @subsection IA-64 Options
9083 @cindex IA-64 Options
9085 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9089 @opindex mbig-endian
9090 Generate code for a big endian target. This is the default for HPUX@.
9092 @item -mlittle-endian
9093 @opindex mlittle-endian
9094 Generate code for a little endian target. This is the default for AIX5
9101 Generate (or don't) code for the GNU assembler. This is the default.
9102 @c Also, this is the default if the configure option @option{--with-gnu-as}
9109 Generate (or don't) code for the GNU linker. This is the default.
9110 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9115 Generate code that does not use a global pointer register. The result
9116 is not position independent code, and violates the IA-64 ABI@.
9118 @item -mvolatile-asm-stop
9119 @itemx -mno-volatile-asm-stop
9120 @opindex mvolatile-asm-stop
9121 @opindex mno-volatile-asm-stop
9122 Generate (or don't) a stop bit immediately before and after volatile asm
9127 Generate code that works around Itanium B step errata.
9129 @item -mregister-names
9130 @itemx -mno-register-names
9131 @opindex mregister-names
9132 @opindex mno-register-names
9133 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9134 the stacked registers. This may make assembler output more readable.
9140 Disable (or enable) optimizations that use the small data section. This may
9141 be useful for working around optimizer bugs.
9144 @opindex mconstant-gp
9145 Generate code that uses a single constant global pointer value. This is
9146 useful when compiling kernel code.
9150 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9151 This is useful when compiling firmware code.
9153 @item -minline-divide-min-latency
9154 @opindex minline-divide-min-latency
9155 Generate code for inline divides using the minimum latency algorithm.
9157 @item -minline-divide-max-throughput
9158 @opindex minline-divide-max-throughput
9159 Generate code for inline divides using the maximum throughput algorithm.
9161 @item -mno-dwarf2-asm
9163 @opindex mno-dwarf2-asm
9164 @opindex mdwarf2-asm
9165 Don't (or do) generate assembler code for the DWARF2 line number debugging
9166 info. This may be useful when not using the GNU assembler.
9168 @item -mfixed-range=@var{register-range}
9169 @opindex mfixed-range
9170 Generate code treating the given register range as fixed registers.
9171 A fixed register is one that the register allocator can not use. This is
9172 useful when compiling kernel code. A register range is specified as
9173 two registers separated by a dash. Multiple register ranges can be
9174 specified separated by a comma.
9178 @subsection D30V Options
9179 @cindex D30V Options
9181 These @samp{-m} options are defined for D30V implementations:
9186 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9187 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9188 memory, which starts at location @code{0x80000000}.
9192 Same as the @option{-mextmem} switch.
9196 Link the @samp{.text} section into onchip text memory, which starts at
9197 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9198 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9199 into onchip data memory, which starts at location @code{0x20000000}.
9201 @item -mno-asm-optimize
9202 @itemx -masm-optimize
9203 @opindex mno-asm-optimize
9204 @opindex masm-optimize
9205 Disable (enable) passing @option{-O} to the assembler when optimizing.
9206 The assembler uses the @option{-O} option to automatically parallelize
9207 adjacent short instructions where possible.
9209 @item -mbranch-cost=@var{n}
9210 @opindex mbranch-cost
9211 Increase the internal costs of branches to @var{n}. Higher costs means
9212 that the compiler will issue more instructions to avoid doing a branch.
9215 @item -mcond-exec=@var{n}
9217 Specify the maximum number of conditionally executed instructions that
9218 replace a branch. The default is 4.
9221 @node S/390 and zSeries Options
9222 @subsection S/390 and zSeries Options
9223 @cindex S/390 and zSeries Options
9225 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9230 @opindex mhard-float
9231 @opindex msoft-float
9232 Use (do not use) the hardware floating-point instructions and registers
9233 for floating-point operations. When @option{-msoft-float} is specified,
9234 functions in @file{libgcc.a} will be used to perform floating-point
9235 operations. When @option{-mhard-float} is specified, the compiler
9236 generates IEEE floating-point instructions. This is the default.
9239 @itemx -mno-backchain
9241 @opindex mno-backchain
9242 Generate (or do not generate) code which maintains an explicit
9243 backchain within the stack frame that points to the caller's frame.
9244 This is currently needed to allow debugging. The default is to
9245 generate the backchain.
9248 @itemx -mno-small-exec
9249 @opindex msmall-exec
9250 @opindex mno-small-exec
9251 Generate (or do not generate) code using the @code{bras} instruction
9252 to do subroutine calls.
9253 This only works reliably if the total executable size does not
9254 exceed 64k. The default is to use the @code{basr} instruction instead,
9255 which does not have this limitation.
9261 When @option{-m31} is specified, generate code compliant to the
9262 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9263 code compliant to the Linux for zSeries ABI@. This allows GCC in
9264 particular to generate 64-bit instructions. For the @samp{s390}
9265 targets, the default is @option{-m31}, while the @samp{s390x}
9266 targets default to @option{-m64}.
9272 Generate (or do not generate) code using the @code{mvcle} instruction
9273 to perform block moves. When @option{-mno-mvcle} is specifed,
9274 use a @code{mvc} loop instead. This is the default.
9280 Print (or do not print) additional debug information when compiling.
9281 The default is to not print debug information.
9286 @subsection CRIS Options
9287 @cindex CRIS Options
9289 These options are defined specifically for the CRIS ports.
9292 @item -march=@var{architecture-type}
9293 @itemx -mcpu=@var{architecture-type}
9296 Generate code for the specified architecture. The choices for
9297 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9298 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9299 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9302 @item -mtune=@var{architecture-type}
9304 Tune to @var{architecture-type} everything applicable about the generated
9305 code, except for the ABI and the set of available instructions. The
9306 choices for @var{architecture-type} are the same as for
9307 @option{-march=@var{architecture-type}}.
9309 @item -mmax-stack-frame=@var{n}
9310 @opindex mmax-stack-frame
9311 Warn when the stack frame of a function exceeds @var{n} bytes.
9313 @item -melinux-stacksize=@var{n}
9314 @opindex melinux-stacksize
9315 Only available with the @samp{cris-axis-aout} target. Arranges for
9316 indications in the program to the kernel loader that the stack of the
9317 program should be set to @var{n} bytes.
9323 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9324 @option{-march=v3} and @option{-march=v8} respectively.
9328 Enable CRIS-specific verbose debug-related information in the assembly
9329 code. This option also has the effect to turn off the @samp{#NO_APP}
9330 formatted-code indicator to the assembler at the beginning of the
9335 Do not use condition-code results from previous instruction; always emit
9336 compare and test instructions before use of condition codes.
9338 @item -mno-side-effects
9339 @opindex mno-side-effects
9340 Do not emit instructions with side-effects in addressing modes other than
9344 @itemx -mno-stack-align
9346 @itemx -mno-data-align
9347 @itemx -mconst-align
9348 @itemx -mno-const-align
9349 @opindex mstack-align
9350 @opindex mno-stack-align
9351 @opindex mdata-align
9352 @opindex mno-data-align
9353 @opindex mconst-align
9354 @opindex mno-const-align
9355 These options (no-options) arranges (eliminate arrangements) for the
9356 stack-frame, individual data and constants to be aligned for the maximum
9357 single data access size for the chosen CPU model. The default is to
9358 arrange for 32-bit alignment. ABI details such as structure layout are
9359 not affected by these options.
9367 Similar to the stack- data- and const-align options above, these options
9368 arrange for stack-frame, writable data and constants to all be 32-bit,
9369 16-bit or 8-bit aligned. The default is 32-bit alignment.
9371 @item -mno-prologue-epilogue
9372 @itemx -mprologue-epilogue
9373 @opindex mno-prologue-epilogue
9374 @opindex mprologue-epilogue
9375 With @option{-mno-prologue-epilogue}, the normal function prologue and
9376 epilogue that sets up the stack-frame are omitted and no return
9377 instructions or return sequences are generated in the code. Use this
9378 option only together with visual inspection of the compiled code: no
9379 warnings or errors are generated when call-saved registers must be saved,
9380 or storage for local variable needs to be allocated.
9386 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9387 instruction sequences that load addresses for functions from the PLT part
9388 of the GOT rather than (traditional on other architectures) calls to the
9389 PLT. The default is @option{-mgotplt}.
9393 Legacy no-op option only recognized with the cris-axis-aout target.
9397 Legacy no-op option only recognized with the cris-axis-elf and
9398 cris-axis-linux-gnu targets.
9402 Only recognized with the cris-axis-aout target, where it selects a
9403 GNU/linux-like multilib, include files and instruction set for
9408 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9412 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9413 to link with input-output functions from a simulator library. Code,
9414 initialized data and zero-initialized data are allocated consecutively.
9418 Like @option{-sim}, but pass linker options to locate initialized data at
9419 0x40000000 and zero-initialized data at 0x80000000.
9423 @subsection MMIX Options
9424 @cindex MMIX Options
9426 These options are defined for the MMIX:
9430 @itemx -mno-libfuncs
9432 @opindex mno-libfuncs
9433 Specify that intrinsic library functions are being compiled, passing all
9434 values in registers, no matter the size.
9439 @opindex mno-epsilon
9440 Generate floating-point comparison instructions that compare with respect
9441 to the @code{rE} epsilon register.
9443 @item -mabi=mmixware
9445 @opindex mabi-mmixware
9447 Generate code that passes function parameters and return values that (in
9448 the called function) are seen as registers @code{$0} and up, as opposed to
9449 the GNU ABI which uses global registers @code{$231} and up.
9452 @itemx -mno-zero-extend
9453 @opindex mzero-extend
9454 @opindex mno-zero-extend
9455 When reading data from memory in sizes shorter than 64 bits, use (do not
9456 use) zero-extending load instructions by default, rather than
9457 sign-extending ones.
9460 @itemx -mno-knuthdiv
9462 @opindex mno-knuthdiv
9463 Make the result of a division yielding a remainder have the same sign as
9464 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9465 remainder follows the sign of the dividend. Both methods are
9466 arithmetically valid, the latter being almost exclusively used.
9468 @item -mtoplevel-symbols
9469 @itemx -mno-toplevel-symbols
9470 @opindex mtoplevel-symbols
9471 @opindex mno-toplevel-symbols
9472 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9473 code can be used with the @code{PREFIX} assembly directive.
9477 Generate an executable in the ELF format, rather than the default
9478 @samp{mmo} format used by the @command{mmix} simulator.
9480 @item -mbranch-predict
9481 @itemx -mno-branch-predict
9482 @opindex mbranch-predict
9483 @opindex mno-branch-predict
9484 Use (do not use) the probable-branch instructions, when static branch
9485 prediction indicates a probable branch.
9487 @item -mreg-stack-fill-bug-workaround
9488 @itemx -mno-reg-stack-fill-bug-workaround
9489 @opindex mreg-stack-fill-bug-workaround
9490 @opindex mno-reg-stack-fill-bug-workaround
9491 Work around (do not work around) an inconsistency in the circular
9492 register stack mechanism in the @command{mmix} simulator, which
9493 causes entries in the register stack to not be flushed to memory if
9494 the instruction causing the fill-up is @code{PUSHJ} or @code{PUSHGO}.
9497 @node PDP-11 Options
9498 @subsection PDP-11 Options
9499 @cindex PDP-11 Options
9501 These options are defined for the PDP-11:
9506 Use hardware FPP floating point. This is the default. (FIS floating
9507 point on the PDP-11/40 is not supported.)
9510 @opindex msoft-float
9511 Do not use hardware floating point.
9515 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9519 Return floating-point results in memory. This is the default.
9523 Generate code for a PDP-11/40.
9527 Generate code for a PDP-11/45. This is the default.
9531 Generate code for a PDP-11/10.
9533 @item -mbcopy-builtin
9534 @opindex bcopy-builtin
9535 Use inline @code{movstrhi} patterns for copying memory. This is the
9540 Do not use inline @code{movstrhi} patterns for copying memory.
9546 Use 16-bit @code{int}. This is the default.
9552 Use 32-bit @code{int}.
9557 @opindex mno-float32
9558 Use 64-bit @code{float}. This is the default.
9563 @opindex mno-float64
9564 Use 32-bit @code{float}.
9568 Use @code{abshi2} pattern. This is the default.
9572 Do not use @code{abshi2} pattern.
9574 @item -mbranch-expensive
9575 @opindex mbranch-expensive
9576 Pretend that branches are expensive. This is for experimenting with
9577 code generation only.
9579 @item -mbranch-cheap
9580 @opindex mbranch-cheap
9581 Do not pretend that branches are expensive. This is the default.
9585 Generate code for a system with split I&D.
9589 Generate code for a system without split I&D. This is the default.
9593 Use Unix assembler syntax. This is the default when configured for
9598 Use DEC assembler syntax. This is the default when configured for any
9599 PDP-11 target other than @samp{pdp11-*-bsd}.
9602 @node Xstormy16 Options
9603 @subsection Xstormy16 Options
9604 @cindex Xstormy16 Options
9606 These options are defined for Xstormy16:
9611 Choose startup files and linker script suitable for the simulator.
9614 @node Code Gen Options
9615 @section Options for Code Generation Conventions
9616 @cindex code generation conventions
9617 @cindex options, code generation
9618 @cindex run-time options
9620 These machine-independent options control the interface conventions
9621 used in code generation.
9623 Most of them have both positive and negative forms; the negative form
9624 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9625 one of the forms is listed---the one which is not the default. You
9626 can figure out the other form by either removing @samp{no-} or adding
9631 @opindex fexceptions
9632 Enable exception handling. Generates extra code needed to propagate
9633 exceptions. For some targets, this implies GCC will generate frame
9634 unwind information for all functions, which can produce significant data
9635 size overhead, although it does not affect execution. If you do not
9636 specify this option, GCC will enable it by default for languages like
9637 C++ which normally require exception handling, and disable it for
9638 languages like C that do not normally require it. However, you may need
9639 to enable this option when compiling C code that needs to interoperate
9640 properly with exception handlers written in C++. You may also wish to
9641 disable this option if you are compiling older C++ programs that don't
9642 use exception handling.
9644 @item -fnon-call-exceptions
9645 @opindex fnon-call-exceptions
9646 Generate code that allows trapping instructions to throw exceptions.
9647 Note that this requires platform-specific runtime support that does
9648 not exist everywhere. Moreover, it only allows @emph{trapping}
9649 instructions to throw exceptions, i.e.@: memory references or floating
9650 point instructions. It does not allow exceptions to be thrown from
9651 arbitrary signal handlers such as @code{SIGALRM}.
9653 @item -funwind-tables
9654 @opindex funwind-tables
9655 Similar to @option{-fexceptions}, except that it will just generate any needed
9656 static data, but will not affect the generated code in any other way.
9657 You will normally not enable this option; instead, a language processor
9658 that needs this handling would enable it on your behalf.
9660 @item -fasynchronous-unwind-tables
9661 @opindex funwind-tables
9662 Generate unwind table in dwarf2 format, if supported by target machine. The
9663 table is exact at each instruction boundary, so it can be used for stack
9664 unwinding from asynchronous events (such as debugger or garbage collector).
9666 @item -fpcc-struct-return
9667 @opindex fpcc-struct-return
9668 Return ``short'' @code{struct} and @code{union} values in memory like
9669 longer ones, rather than in registers. This convention is less
9670 efficient, but it has the advantage of allowing intercallability between
9671 GCC-compiled files and files compiled with other compilers.
9673 The precise convention for returning structures in memory depends
9674 on the target configuration macros.
9676 Short structures and unions are those whose size and alignment match
9677 that of some integer type.
9679 @item -freg-struct-return
9680 @opindex freg-struct-return
9681 Return @code{struct} and @code{union} values in registers when possible.
9682 This is more efficient for small structures than
9683 @option{-fpcc-struct-return}.
9685 If you specify neither @option{-fpcc-struct-return} nor
9686 @option{-freg-struct-return}, GCC defaults to whichever convention is
9687 standard for the target. If there is no standard convention, GCC
9688 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9689 the principal compiler. In those cases, we can choose the standard, and
9690 we chose the more efficient register return alternative.
9693 @opindex fshort-enums
9694 Allocate to an @code{enum} type only as many bytes as it needs for the
9695 declared range of possible values. Specifically, the @code{enum} type
9696 will be equivalent to the smallest integer type which has enough room.
9698 @item -fshort-double
9699 @opindex fshort-double
9700 Use the same size for @code{double} as for @code{float}.
9703 @opindex fshared-data
9704 Requests that the data and non-@code{const} variables of this
9705 compilation be shared data rather than private data. The distinction
9706 makes sense only on certain operating systems, where shared data is
9707 shared between processes running the same program, while private data
9708 exists in one copy per process.
9712 In C, allocate even uninitialized global variables in the data section of the
9713 object file, rather than generating them as common blocks. This has the
9714 effect that if the same variable is declared (without @code{extern}) in
9715 two different compilations, you will get an error when you link them.
9716 The only reason this might be useful is if you wish to verify that the
9717 program will work on other systems which always work this way.
9721 Ignore the @samp{#ident} directive.
9723 @item -fno-gnu-linker
9724 @opindex fno-gnu-linker
9725 Do not output global initializations (such as C++ constructors and
9726 destructors) in the form used by the GNU linker (on systems where the GNU
9727 linker is the standard method of handling them). Use this option when
9728 you want to use a non-GNU linker, which also requires using the
9729 @command{collect2} program to make sure the system linker includes
9730 constructors and destructors. (@command{collect2} is included in the GCC
9731 distribution.) For systems which @emph{must} use @command{collect2}, the
9732 compiler driver @command{gcc} is configured to do this automatically.
9734 @item -finhibit-size-directive
9735 @opindex finhibit-size-directive
9736 Don't output a @code{.size} assembler directive, or anything else that
9737 would cause trouble if the function is split in the middle, and the
9738 two halves are placed at locations far apart in memory. This option is
9739 used when compiling @file{crtstuff.c}; you should not need to use it
9743 @opindex fverbose-asm
9744 Put extra commentary information in the generated assembly code to
9745 make it more readable. This option is generally only of use to those
9746 who actually need to read the generated assembly code (perhaps while
9747 debugging the compiler itself).
9749 @option{-fno-verbose-asm}, the default, causes the
9750 extra information to be omitted and is useful when comparing two assembler
9755 Consider all memory references through pointers to be volatile.
9757 @item -fvolatile-global
9758 @opindex fvolatile-global
9759 Consider all memory references to extern and global data items to
9760 be volatile. GCC does not consider static data items to be volatile
9761 because of this switch.
9763 @item -fvolatile-static
9764 @opindex fvolatile-static
9765 Consider all memory references to static data to be volatile.
9769 @cindex global offset table
9771 Generate position-independent code (PIC) suitable for use in a shared
9772 library, if supported for the target machine. Such code accesses all
9773 constant addresses through a global offset table (GOT)@. The dynamic
9774 loader resolves the GOT entries when the program starts (the dynamic
9775 loader is not part of GCC; it is part of the operating system). If
9776 the GOT size for the linked executable exceeds a machine-specific
9777 maximum size, you get an error message from the linker indicating that
9778 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9779 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9780 on the m68k and RS/6000. The 386 has no such limit.)
9782 Position-independent code requires special support, and therefore works
9783 only on certain machines. For the 386, GCC supports PIC for System V
9784 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9785 position-independent.
9789 If supported for the target machine, emit position-independent code,
9790 suitable for dynamic linking and avoiding any limit on the size of the
9791 global offset table. This option makes a difference on the m68k, m88k,
9794 Position-independent code requires special support, and therefore works
9795 only on certain machines.
9797 @item -ffixed-@var{reg}
9799 Treat the register named @var{reg} as a fixed register; generated code
9800 should never refer to it (except perhaps as a stack pointer, frame
9801 pointer or in some other fixed role).
9803 @var{reg} must be the name of a register. The register names accepted
9804 are machine-specific and are defined in the @code{REGISTER_NAMES}
9805 macro in the machine description macro file.
9807 This flag does not have a negative form, because it specifies a
9810 @item -fcall-used-@var{reg}
9812 Treat the register named @var{reg} as an allocable register that is
9813 clobbered by function calls. It may be allocated for temporaries or
9814 variables that do not live across a call. Functions compiled this way
9815 will not save and restore the register @var{reg}.
9817 It is an error to used this flag with the frame pointer or stack pointer.
9818 Use of this flag for other registers that have fixed pervasive roles in
9819 the machine's execution model will produce disastrous results.
9821 This flag does not have a negative form, because it specifies a
9824 @item -fcall-saved-@var{reg}
9825 @opindex fcall-saved
9826 Treat the register named @var{reg} as an allocable register saved by
9827 functions. It may be allocated even for temporaries or variables that
9828 live across a call. Functions compiled this way will save and restore
9829 the register @var{reg} if they use it.
9831 It is an error to used this flag with the frame pointer or stack pointer.
9832 Use of this flag for other registers that have fixed pervasive roles in
9833 the machine's execution model will produce disastrous results.
9835 A different sort of disaster will result from the use of this flag for
9836 a register in which function values may be returned.
9838 This flag does not have a negative form, because it specifies a
9842 @opindex fpack-struct
9843 Pack all structure members together without holes. Usually you would
9844 not want to use this option, since it makes the code suboptimal, and
9845 the offsets of structure members won't agree with system libraries.
9847 @item -finstrument-functions
9848 @opindex finstrument-functions
9849 Generate instrumentation calls for entry and exit to functions. Just
9850 after function entry and just before function exit, the following
9851 profiling functions will be called with the address of the current
9852 function and its call site. (On some platforms,
9853 @code{__builtin_return_address} does not work beyond the current
9854 function, so the call site information may not be available to the
9855 profiling functions otherwise.)
9858 void __cyg_profile_func_enter (void *this_fn,
9860 void __cyg_profile_func_exit (void *this_fn,
9864 The first argument is the address of the start of the current function,
9865 which may be looked up exactly in the symbol table.
9867 This instrumentation is also done for functions expanded inline in other
9868 functions. The profiling calls will indicate where, conceptually, the
9869 inline function is entered and exited. This means that addressable
9870 versions of such functions must be available. If all your uses of a
9871 function are expanded inline, this may mean an additional expansion of
9872 code size. If you use @samp{extern inline} in your C code, an
9873 addressable version of such functions must be provided. (This is
9874 normally the case anyways, but if you get lucky and the optimizer always
9875 expands the functions inline, you might have gotten away without
9876 providing static copies.)
9878 A function may be given the attribute @code{no_instrument_function}, in
9879 which case this instrumentation will not be done. This can be used, for
9880 example, for the profiling functions listed above, high-priority
9881 interrupt routines, and any functions from which the profiling functions
9882 cannot safely be called (perhaps signal handlers, if the profiling
9883 routines generate output or allocate memory).
9886 @opindex fstack-check
9887 Generate code to verify that you do not go beyond the boundary of the
9888 stack. You should specify this flag if you are running in an
9889 environment with multiple threads, but only rarely need to specify it in
9890 a single-threaded environment since stack overflow is automatically
9891 detected on nearly all systems if there is only one stack.
9893 Note that this switch does not actually cause checking to be done; the
9894 operating system must do that. The switch causes generation of code
9895 to ensure that the operating system sees the stack being extended.
9897 @item -fstack-limit-register=@var{reg}
9898 @itemx -fstack-limit-symbol=@var{sym}
9899 @itemx -fno-stack-limit
9900 @opindex fstack-limit-register
9901 @opindex fstack-limit-symbol
9902 @opindex fno-stack-limit
9903 Generate code to ensure that the stack does not grow beyond a certain value,
9904 either the value of a register or the address of a symbol. If the stack
9905 would grow beyond the value, a signal is raised. For most targets,
9906 the signal is raised before the stack overruns the boundary, so
9907 it is possible to catch the signal without taking special precautions.
9909 For instance, if the stack starts at absolute address @samp{0x80000000}
9910 and grows downwards, you can use the flags
9911 @option{-fstack-limit-symbol=__stack_limit} and
9912 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
9913 of 128KB@. Note that this may only work with the GNU linker.
9915 @cindex aliasing of parameters
9916 @cindex parameters, aliased
9917 @item -fargument-alias
9918 @itemx -fargument-noalias
9919 @itemx -fargument-noalias-global
9920 @opindex fargument-alias
9921 @opindex fargument-noalias
9922 @opindex fargument-noalias-global
9923 Specify the possible relationships among parameters and between
9924 parameters and global data.
9926 @option{-fargument-alias} specifies that arguments (parameters) may
9927 alias each other and may alias global storage.@*
9928 @option{-fargument-noalias} specifies that arguments do not alias
9929 each other, but may alias global storage.@*
9930 @option{-fargument-noalias-global} specifies that arguments do not
9931 alias each other and do not alias global storage.
9933 Each language will automatically use whatever option is required by
9934 the language standard. You should not need to use these options yourself.
9936 @item -fleading-underscore
9937 @opindex fleading-underscore
9938 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
9939 change the way C symbols are represented in the object file. One use
9940 is to help link with legacy assembly code.
9942 Be warned that you should know what you are doing when invoking this
9943 option, and that not all targets provide complete support for it.
9948 @node Environment Variables
9949 @section Environment Variables Affecting GCC
9950 @cindex environment variables
9952 @c man begin ENVIRONMENT
9954 This section describes several environment variables that affect how GCC
9955 operates. Some of them work by specifying directories or prefixes to use
9956 when searching for various kinds of files. Some are used to specify other
9957 aspects of the compilation environment.
9959 Note that you can also specify places to search using options such as
9960 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
9961 take precedence over places specified using environment variables, which
9962 in turn take precedence over those specified by the configuration of GCC@.
9963 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
9964 GNU Compiler Collection (GCC) Internals}.
9969 @c @itemx LC_COLLATE
9971 @c @itemx LC_MONETARY
9972 @c @itemx LC_NUMERIC
9977 @c @findex LC_COLLATE
9979 @c @findex LC_MONETARY
9980 @c @findex LC_NUMERIC
9984 These environment variables control the way that GCC uses
9985 localization information that allow GCC to work with different
9986 national conventions. GCC inspects the locale categories
9987 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
9988 so. These locale categories can be set to any value supported by your
9989 installation. A typical value is @samp{en_UK} for English in the United
9992 The @env{LC_CTYPE} environment variable specifies character
9993 classification. GCC uses it to determine the character boundaries in
9994 a string; this is needed for some multibyte encodings that contain quote
9995 and escape characters that would otherwise be interpreted as a string
9998 The @env{LC_MESSAGES} environment variable specifies the language to
9999 use in diagnostic messages.
10001 If the @env{LC_ALL} environment variable is set, it overrides the value
10002 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10003 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10004 environment variable. If none of these variables are set, GCC
10005 defaults to traditional C English behavior.
10009 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10010 files. GCC uses temporary files to hold the output of one stage of
10011 compilation which is to be used as input to the next stage: for example,
10012 the output of the preprocessor, which is the input to the compiler
10015 @item GCC_EXEC_PREFIX
10016 @findex GCC_EXEC_PREFIX
10017 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10018 names of the subprograms executed by the compiler. No slash is added
10019 when this prefix is combined with the name of a subprogram, but you can
10020 specify a prefix that ends with a slash if you wish.
10022 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10023 an appropriate prefix to use based on the pathname it was invoked with.
10025 If GCC cannot find the subprogram using the specified prefix, it
10026 tries looking in the usual places for the subprogram.
10028 The default value of @env{GCC_EXEC_PREFIX} is
10029 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10030 of @code{prefix} when you ran the @file{configure} script.
10032 Other prefixes specified with @option{-B} take precedence over this prefix.
10034 This prefix is also used for finding files such as @file{crt0.o} that are
10037 In addition, the prefix is used in an unusual way in finding the
10038 directories to search for header files. For each of the standard
10039 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10040 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10041 replacing that beginning with the specified prefix to produce an
10042 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10043 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10044 These alternate directories are searched first; the standard directories
10047 @item COMPILER_PATH
10048 @findex COMPILER_PATH
10049 The value of @env{COMPILER_PATH} is a colon-separated list of
10050 directories, much like @env{PATH}. GCC tries the directories thus
10051 specified when searching for subprograms, if it can't find the
10052 subprograms using @env{GCC_EXEC_PREFIX}.
10055 @findex LIBRARY_PATH
10056 The value of @env{LIBRARY_PATH} is a colon-separated list of
10057 directories, much like @env{PATH}. When configured as a native compiler,
10058 GCC tries the directories thus specified when searching for special
10059 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10060 using GCC also uses these directories when searching for ordinary
10061 libraries for the @option{-l} option (but directories specified with
10062 @option{-L} come first).
10064 @item C_INCLUDE_PATH
10065 @itemx CPLUS_INCLUDE_PATH
10066 @itemx OBJC_INCLUDE_PATH
10067 @findex C_INCLUDE_PATH
10068 @findex CPLUS_INCLUDE_PATH
10069 @findex OBJC_INCLUDE_PATH
10070 @c @itemx OBJCPLUS_INCLUDE_PATH
10071 These environment variables pertain to particular languages. Each
10072 variable's value is a colon-separated list of directories, much like
10073 @env{PATH}. When GCC searches for header files, it tries the
10074 directories listed in the variable for the language you are using, after
10075 the directories specified with @option{-I} but before the standard header
10078 @item DEPENDENCIES_OUTPUT
10079 @findex DEPENDENCIES_OUTPUT
10080 @cindex dependencies for make as output
10081 If this variable is set, its value specifies how to output dependencies
10082 for Make based on the header files processed by the compiler. This
10083 output looks much like the output from the @option{-M} option
10084 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
10085 in addition to the usual results of compilation.
10087 The value of @env{DEPENDENCIES_OUTPUT} can be just a file name, in
10088 which case the Make rules are written to that file, guessing the target
10089 name from the source file name. Or the value can have the form
10090 @samp{@var{file} @var{target}}, in which case the rules are written to
10091 file @var{file} using @var{target} as the target name.
10095 @cindex locale definition
10096 This variable is used to pass locale information to the compiler. One way in
10097 which this information is used is to determine the character set to be used
10098 when character literals, string literals and comments are parsed in C and C++.
10099 When the compiler is configured to allow multibyte characters,
10100 the following values for @env{LANG} are recognized:
10104 Recognize JIS characters.
10106 Recognize SJIS characters.
10108 Recognize EUCJP characters.
10111 If @env{LANG} is not defined, or if it has some other value, then the
10112 compiler will use mblen and mbtowc as defined by the default locale to
10113 recognize and translate multibyte characters.
10118 @node Running Protoize
10119 @section Running Protoize
10121 The program @code{protoize} is an optional part of GCC@. You can use
10122 it to add prototypes to a program, thus converting the program to ISO
10123 C in one respect. The companion program @code{unprotoize} does the
10124 reverse: it removes argument types from any prototypes that are found.
10126 When you run these programs, you must specify a set of source files as
10127 command line arguments. The conversion programs start out by compiling
10128 these files to see what functions they define. The information gathered
10129 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10131 After scanning comes actual conversion. The specified files are all
10132 eligible to be converted; any files they include (whether sources or
10133 just headers) are eligible as well.
10135 But not all the eligible files are converted. By default,
10136 @code{protoize} and @code{unprotoize} convert only source and header
10137 files in the current directory. You can specify additional directories
10138 whose files should be converted with the @option{-d @var{directory}}
10139 option. You can also specify particular files to exclude with the
10140 @option{-x @var{file}} option. A file is converted if it is eligible, its
10141 directory name matches one of the specified directory names, and its
10142 name within the directory has not been excluded.
10144 Basic conversion with @code{protoize} consists of rewriting most
10145 function definitions and function declarations to specify the types of
10146 the arguments. The only ones not rewritten are those for varargs
10149 @code{protoize} optionally inserts prototype declarations at the
10150 beginning of the source file, to make them available for any calls that
10151 precede the function's definition. Or it can insert prototype
10152 declarations with block scope in the blocks where undeclared functions
10155 Basic conversion with @code{unprotoize} consists of rewriting most
10156 function declarations to remove any argument types, and rewriting
10157 function definitions to the old-style pre-ISO form.
10159 Both conversion programs print a warning for any function declaration or
10160 definition that they can't convert. You can suppress these warnings
10163 The output from @code{protoize} or @code{unprotoize} replaces the
10164 original source file. The original file is renamed to a name ending
10165 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10166 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10167 for DOS) file already exists, then the source file is simply discarded.
10169 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10170 scan the program and collect information about the functions it uses.
10171 So neither of these programs will work until GCC is installed.
10173 Here is a table of the options you can use with @code{protoize} and
10174 @code{unprotoize}. Each option works with both programs unless
10178 @item -B @var{directory}
10179 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10180 usual directory (normally @file{/usr/local/lib}). This file contains
10181 prototype information about standard system functions. This option
10182 applies only to @code{protoize}.
10184 @item -c @var{compilation-options}
10185 Use @var{compilation-options} as the options when running @code{gcc} to
10186 produce the @samp{.X} files. The special option @option{-aux-info} is
10187 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10189 Note that the compilation options must be given as a single argument to
10190 @code{protoize} or @code{unprotoize}. If you want to specify several
10191 @code{gcc} options, you must quote the entire set of compilation options
10192 to make them a single word in the shell.
10194 There are certain @code{gcc} arguments that you cannot use, because they
10195 would produce the wrong kind of output. These include @option{-g},
10196 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10197 the @var{compilation-options}, they are ignored.
10200 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10201 systems) instead of @samp{.c}. This is convenient if you are converting
10202 a C program to C++. This option applies only to @code{protoize}.
10205 Add explicit global declarations. This means inserting explicit
10206 declarations at the beginning of each source file for each function
10207 that is called in the file and was not declared. These declarations
10208 precede the first function definition that contains a call to an
10209 undeclared function. This option applies only to @code{protoize}.
10211 @item -i @var{string}
10212 Indent old-style parameter declarations with the string @var{string}.
10213 This option applies only to @code{protoize}.
10215 @code{unprotoize} converts prototyped function definitions to old-style
10216 function definitions, where the arguments are declared between the
10217 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10218 uses five spaces as the indentation. If you want to indent with just
10219 one space instead, use @option{-i " "}.
10222 Keep the @samp{.X} files. Normally, they are deleted after conversion
10226 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10227 a prototype declaration for each function in each block which calls the
10228 function without any declaration. This option applies only to
10232 Make no real changes. This mode just prints information about the conversions
10233 that would have been done without @option{-n}.
10236 Make no @samp{.save} files. The original files are simply deleted.
10237 Use this option with caution.
10239 @item -p @var{program}
10240 Use the program @var{program} as the compiler. Normally, the name
10241 @file{gcc} is used.
10244 Work quietly. Most warnings are suppressed.
10247 Print the version number, just like @option{-v} for @code{gcc}.
10250 If you need special compiler options to compile one of your program's
10251 source files, then you should generate that file's @samp{.X} file
10252 specially, by running @code{gcc} on that source file with the
10253 appropriate options and the option @option{-aux-info}. Then run
10254 @code{protoize} on the entire set of files. @code{protoize} will use
10255 the existing @samp{.X} file because it is newer than the source file.
10259 gcc -Dfoo=bar file1.c -aux-info file1.X
10264 You need to include the special files along with the rest in the
10265 @code{protoize} command, even though their @samp{.X} files already
10266 exist, because otherwise they won't get converted.
10268 @xref{Protoize Caveats}, for more information on how to use
10269 @code{protoize} successfully.